5 Changes between 0.9.8 and 0.9.9 [xx XXX xxxx]
7 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
8 place for the (very old) "NETSCAPE" format certificates which are now
9 handled using new ASN1 code equivalents.
11 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
12 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
13 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
16 *) Modify CRL distribution points extension code to print out previously
17 unsupported fields. Enhance extension setting code to allow setting of
21 *) Add print and set support for Issuing Distribution Point CRL extension.
24 Changes between 0.9.8 and 0.9.8a [XX xxx XXXX]
26 *) Extended Windows CE support.
27 [Satoshi Nakamura and Andy Polyakov]
29 *) Initialize SSL_METHOD structures at compile time instead of during
30 runtime, thus removing the need for a lock.
33 *) Make PKCS7_decrypt() work even if no certificate is supplied by
34 attempting to decrypt each encrypted key in turn. Add support to
38 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
40 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
43 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
44 key into the same file any more.
47 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
50 *) Add -utf8 command line and config file option to 'ca'.
51 [Stefan <stf@udoma.org]
53 *) Removed the macro des_crypt(), as it seems to conflict with some
54 libraries. Use DES_crypt().
57 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
58 involves renaming the source and generated shared-libs for
59 both. The engines will accept the corrected or legacy ids
60 ('ncipher' and '4758_cca' respectively) when binding. NB,
61 this only applies when building 'shared'.
62 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
64 *) Add attribute functions to EVP_PKEY structure. Modify
65 PKCS12_create() to recognize a CSP name attribute and
66 use it. Make -CSP option work again in pkcs12 utility.
69 *) Add new functionality to the bn blinding code:
70 - automatic re-creation of the BN_BLINDING parameters after
71 a fixed number of uses (currently 32)
72 - add new function for parameter creation
73 - introduce flags to control the update behaviour of the
74 BN_BLINDING parameters
75 - hide BN_BLINDING structure
76 Add a second BN_BLINDING slot to the RSA structure to improve
77 performance when a single RSA object is shared among several
81 *) Add support for DTLS.
82 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
84 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
85 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
88 *) Remove buggy and incompletet DH cert support from
89 ssl/ssl_rsa.c and ssl/s3_both.c
92 *) Use SHA-1 instead of MD5 as the default digest algorithm for
93 the apps/openssl applications.
96 *) Compile clean with "-Wall -Wmissing-prototypes
97 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
98 DEBUG_SAFESTACK must also be set.
101 *) Change ./Configure so that certain algorithms can be disabled by default.
102 The new counterpiece to "no-xxx" is "enable-xxx".
104 The patented RC5 and MDC2 algorithms will now be disabled unless
105 "enable-rc5" and "enable-mdc2", respectively, are specified.
107 (IDEA remains enabled despite being patented. This is because IDEA
108 is frequently required for interoperability, and there is no license
109 fee for non-commercial use. As before, "no-idea" can be used to
110 avoid this algorithm.)
114 *) Add processing of proxy certificates (see RFC 3820). This work was
115 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
116 EGEE (Enabling Grids for E-science in Europe).
119 *) RC4 performance overhaul on modern architectures/implementations, such
120 as Intel P4, IA-64 and AMD64.
123 *) New utility extract-section.pl. This can be used specify an alternative
124 section number in a pod file instead of having to treat each file as
125 a separate case in Makefile. This can be done by adding two lines to the
128 =for comment openssl_section:XXX
130 The blank line is mandatory.
134 *) New arguments -certform, -keyform and -pass for s_client and s_server
135 to allow alternative format key and certificate files and passphrase
139 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
140 update associated structures and add various utility functions.
142 Add new policy related verify parameters, include policy checking in
143 standard verify code. Enhance 'smime' application with extra parameters
144 to support policy checking and print out.
147 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
148 Nehemiah processors. These extensions support AES encryption in hardware
149 as well as RNG (though RNG support is currently disabled).
150 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
152 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
155 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
156 [Andy Polyakov and a number of other people]
158 *) Improved PowerPC platform support. Most notably BIGNUM assembler
159 implementation contributed by IBM.
160 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
162 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
163 exponent rather than 'unsigned long'. There is a corresponding change to
164 the new 'rsa_keygen' element of the RSA_METHOD structure.
165 [Jelte Jansen, Geoff Thorpe]
167 *) Functionality for creating the initial serial number file is now
168 moved from CA.pl to the 'ca' utility with a new option -create_serial.
170 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
171 number file to 1, which is bound to cause problems. To avoid
172 the problems while respecting compatibility between different 0.9.7
173 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
174 CA.pl for serial number initialization. With the new release 0.9.8,
175 we can fix the problem directly in the 'ca' utility.)
178 *) Reduced header interdepencies by declaring more opaque objects in
179 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
180 give fewer recursive includes, which could break lazy source code - so
181 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
182 developers should define this symbol when building and using openssl to
183 ensure they track the recommended behaviour, interfaces, [etc], but
184 backwards-compatible behaviour prevails when this isn't defined.
187 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
190 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
191 This will generate a random key of the appropriate length based on the
192 cipher context. The EVP_CIPHER can provide its own random key generation
193 routine to support keys of a specific form. This is used in the des and
194 3des routines to generate a key of the correct parity. Update S/MIME
195 code to use new functions and hence generate correct parity DES keys.
196 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
197 valid (weak or incorrect parity).
200 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
201 as looking them up. This is useful when the verified structure may contain
202 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
203 present unless the new PKCS7_NO_CRL flag is asserted.
206 *) Extend ASN1 oid configuration module. It now additionally accepts the
209 shortName = some long name, 1.2.3.4
212 *) Reimplemented the BN_CTX implementation. There is now no more static
213 limitation on the number of variables it can handle nor the depth of the
214 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
215 information can now expand as required, and rather than having a single
216 static array of bignums, BN_CTX now uses a linked-list of such arrays
217 allowing it to expand on demand whilst maintaining the usefulness of
221 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
222 to allow all RSA operations to function using a single BN_CTX.
225 *) Preliminary support for certificate policy evaluation and checking. This
226 is initially intended to pass the tests outlined in "Conformance Testing
227 of Relying Party Client Certificate Path Processing Logic" v1.07.
230 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
231 remained unused and not that useful. A variety of other little bignum
232 tweaks and fixes have also been made continuing on from the audit (see
236 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
237 associated ASN1, EVP and SSL functions and old ASN1 macros.
240 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
241 and this should never fail. So the return value from the use of
242 BN_set_word() (which can fail due to needless expansion) is now deprecated;
243 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
246 *) BN_CTX_get() should return zero-valued bignums, providing the same
247 initialised value as BN_new().
248 [Geoff Thorpe, suggested by Ulf Möller]
250 *) Support for inhibitAnyPolicy certificate extension.
253 *) An audit of the BIGNUM code is underway, for which debugging code is
254 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
255 is considered valid when processing BIGNUMs, and causes execution to
256 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
257 further steps are taken to deliberately pollute unused data in BIGNUM
258 structures to try and expose faulty code further on. For now, openssl will
259 (in its default mode of operation) continue to tolerate the inconsistent
260 forms that it has tolerated in the past, but authors and packagers should
261 consider trying openssl and their own applications when compiled with
262 these debugging symbols defined. It will help highlight potential bugs in
263 their own code, and will improve the test coverage for OpenSSL itself. At
264 some point, these tighter rules will become openssl's default to improve
265 maintainability, though the assert()s and other overheads will remain only
266 in debugging configurations. See bn.h for more details.
267 [Geoff Thorpe, Nils Larsch, Ulf Möller]
269 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
270 that can only be obtained through BN_CTX_new() (which implicitly
271 initialises it). The presence of this function only made it possible
272 to overwrite an existing structure (and cause memory leaks).
275 *) Because of the callback-based approach for implementing LHASH as a
276 template type, lh_insert() adds opaque objects to hash-tables and
277 lh_doall() or lh_doall_arg() are typically used with a destructor callback
278 to clean up those corresponding objects before destroying the hash table
279 (and losing the object pointers). So some over-zealous constifications in
280 LHASH have been relaxed so that lh_insert() does not take (nor store) the
281 objects as "const" and the lh_doall[_arg] callback wrappers are not
282 prototyped to have "const" restrictions on the object pointers they are
283 given (and so aren't required to cast them away any more).
286 *) The tmdiff.h API was so ugly and minimal that our own timing utility
287 (speed) prefers to use its own implementation. The two implementations
288 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
289 its object type properly exposed (MS_TM) instead of casting to/from "char
290 *". This may still change yet if someone realises MS_TM and "ms_time_***"
291 aren't necessarily the greatest nomenclatures - but this is what was used
292 internally to the implementation so I've used that for now.
295 *) Ensure that deprecated functions do not get compiled when
296 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
297 the self-tests were still using deprecated key-generation functions so
298 these have been updated also.
301 *) Reorganise PKCS#7 code to separate the digest location functionality
302 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
303 New function PKCS7_set_digest() to set the digest type for PKCS#7
304 digestedData type. Add additional code to correctly generate the
305 digestedData type and add support for this type in PKCS7 initialization
309 *) New function PKCS7_set0_type_other() this initializes a PKCS7
310 structure of type "other".
313 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
314 sure the loop does correctly stop and breaking ("division by zero")
315 modulus operations are not performed. The (pre-generated) prime
316 table crypto/bn/bn_prime.h was already correct, but it could not be
317 re-generated on some platforms because of the "division by zero"
318 situation in the script.
319 [Ralf S. Engelschall]
321 *) Update support for ECC-based TLS ciphersuites according to
322 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
323 SHA-1 now is only used for "small" curves (where the
324 representation of a field element takes up to 24 bytes); for
325 larger curves, the field element resulting from ECDH is directly
326 used as premaster secret.
327 [Douglas Stebila (Sun Microsystems Laboratories)]
329 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
330 curve secp160r1 to the tests.
331 [Douglas Stebila (Sun Microsystems Laboratories)]
333 *) Add the possibility to load symbols globally with DSO.
334 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
336 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
337 control of the error stack.
340 *) Add support for STORE in ENGINE.
343 *) Add the STORE type. The intention is to provide a common interface
344 to certificate and key stores, be they simple file-based stores, or
345 HSM-type store, or LDAP stores, or...
346 NOTE: The code is currently UNTESTED and isn't really used anywhere.
349 *) Add a generic structure called OPENSSL_ITEM. This can be used to
350 pass a list of arguments to any function as well as provide a way
351 for a function to pass data back to the caller.
354 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
355 works like BUF_strdup() but can be used to duplicate a portion of
356 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
360 *) Add the function sk_find_ex() which works like sk_find(), but will
361 return an index to an element even if an exact match couldn't be
362 found. The index is guaranteed to point at the element where the
363 searched-for key would be inserted to preserve sorting order.
366 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
367 takes an extra flags argument for optional functionality. Currently,
368 the following flags are defined:
370 OBJ_BSEARCH_VALUE_ON_NOMATCH
371 This one gets OBJ_bsearch_ex() to return a pointer to the first
372 element where the comparing function returns a negative or zero
375 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
376 This one gets OBJ_bsearch_ex() to return a pointer to the first
377 element where the comparing function returns zero. This is useful
378 if there are more than one element where the comparing function
382 *) Make it possible to create self-signed certificates with 'openssl ca'
383 in such a way that the self-signed certificate becomes part of the
384 CA database and uses the same mechanisms for serial number generation
385 as all other certificate signing. The new flag '-selfsign' enables
386 this functionality. Adapt CA.sh and CA.pl.in.
389 *) Add functionality to check the public key of a certificate request
390 against a given private. This is useful to check that a certificate
391 request can be signed by that key (self-signing).
394 *) Make it possible to have multiple active certificates with the same
395 subject in the CA index file. This is done only if the keyword
396 'unique_subject' is set to 'no' in the main CA section (default
397 if 'CA_default') of the configuration file. The value is saved
398 with the database itself in a separate index attribute file,
399 named like the index file with '.attr' appended to the name.
402 *) Generate muti valued AVAs using '+' notation in config files for
406 *) Support for nameConstraints certificate extension.
409 *) Support for policyConstraints certificate extension.
412 *) Support for policyMappings certificate extension.
415 *) Make sure the default DSA_METHOD implementation only uses its
416 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
417 and change its own handlers to be NULL so as to remove unnecessary
418 indirection. This lets alternative implementations fallback to the
419 default implementation more easily.
422 *) Support for directoryName in GeneralName related extensions
426 *) Make it possible to link applications using Makefile.shared.
427 Make that possible even when linking against static libraries!
430 *) Support for single pass processing for S/MIME signing. This now
431 means that S/MIME signing can be done from a pipe, in addition
432 cleartext signing (multipart/signed type) is effectively streaming
433 and the signed data does not need to be all held in memory.
435 This is done with a new flag PKCS7_STREAM. When this flag is set
436 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
437 is done after the data is output (and digests calculated) in
441 *) Add full support for -rpath/-R, both in shared libraries and
442 applications, at least on the platforms where it's known how
446 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
447 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
448 will now compute a table of multiples of the generator that
449 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
450 faster (notably in the case of a single point multiplication,
452 [Nils Larsch, Bodo Moeller]
454 *) IPv6 support for certificate extensions. The various extensions
455 which use the IP:a.b.c.d can now take IPv6 addresses using the
456 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
460 *) Added an ENGINE that implements RSA by performing private key
461 exponentiations with the GMP library. The conversions to and from
462 GMP's mpz_t format aren't optimised nor are any montgomery forms
463 cached, and on x86 it appears OpenSSL's own performance has caught up.
464 However there are likely to be other architectures where GMP could
465 provide a boost. This ENGINE is not built in by default, but it can be
466 specified at Configure time and should be accompanied by the necessary
467 linker additions, eg;
468 ./config -DOPENSSL_USE_GMP -lgmp
471 *) "openssl engine" will not display ENGINE/DSO load failure errors when
472 testing availability of engines with "-t" - the old behaviour is
473 produced by increasing the feature's verbosity with "-tt".
476 *) ECDSA routines: under certain error conditions uninitialized BN objects
477 could be freed. Solution: make sure initialization is performed early
478 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
482 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
483 and DH_METHOD (eg. by ENGINE implementations) to override the normal
484 software implementations. For DSA and DH, parameter generation can
485 also be overriden by providing the appropriate method callbacks.
488 *) Change the "progress" mechanism used in key-generation and
489 primality testing to functions that take a new BN_GENCB pointer in
490 place of callback/argument pairs. The new API functions have "_ex"
491 postfixes and the older functions are reimplemented as wrappers for
492 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
493 declarations of the old functions to help (graceful) attempts to
494 migrate to the new functions. Also, the new key-generation API
495 functions operate on a caller-supplied key-structure and return
496 success/failure rather than returning a key or NULL - this is to
497 help make "keygen" another member function of RSA_METHOD etc.
499 Example for using the new callback interface:
501 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
505 BN_GENCB_set(&my_cb, my_callback, my_arg);
507 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
508 /* For the meaning of a, b in calls to my_callback(), see the
509 * documentation of the function that calls the callback.
510 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
511 * my_callback should return 1 if it wants BN_is_prime_ex()
512 * to continue, or 0 to stop.
517 *) Change the ZLIB compression method to be stateful, and make it
518 available to TLS with the number defined in
519 draft-ietf-tls-compression-04.txt.
522 *) Add the ASN.1 structures and functions for CertificatePair, which
523 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
525 CertificatePair ::= SEQUENCE {
526 forward [0] Certificate OPTIONAL,
527 reverse [1] Certificate OPTIONAL,
528 -- at least one of the pair shall be present -- }
530 Also implement the PEM functions to read and write certificate
531 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
533 This needed to be defined, mostly for the sake of the LDAP
534 attribute crossCertificatePair, but may prove useful elsewhere as
538 *) Make it possible to inhibit symlinking of shared libraries in
539 Makefile.shared, for Cygwin's sake.
542 *) Extend the BIGNUM API by creating a function
543 void BN_set_negative(BIGNUM *a, int neg);
544 and a macro that behave like
545 int BN_is_negative(const BIGNUM *a);
547 to avoid the need to access 'a->neg' directly in applications.
550 *) Implement fast modular reduction for pseudo-Mersenne primes
551 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
552 EC_GROUP_new_curve_GFp() will now automatically use this
554 [Nils Larsch <nla@trustcenter.de>]
556 *) Add new lock type (CRYPTO_LOCK_BN).
559 *) Change the ENGINE framework to automatically load engines
560 dynamically from specific directories unless they could be
561 found to already be built in or loaded. Move all the
562 current engines except for the cryptodev one to a new
564 The engines in engines/ are built as shared libraries if
565 the "shared" options was given to ./Configure or ./config.
566 Otherwise, they are inserted in libcrypto.a.
567 /usr/local/ssl/engines is the default directory for dynamic
568 engines, but that can be overriden at configure time through
569 the usual use of --prefix and/or --openssldir, and at run
570 time with the environment variable OPENSSL_ENGINES.
571 [Geoff Thorpe and Richard Levitte]
573 *) Add Makefile.shared, a helper makefile to build shared
574 libraries. Addapt Makefile.org.
577 *) Add version info to Win32 DLLs.
578 [Peter 'Luna' Runestig" <peter@runestig.com>]
580 *) Add new 'medium level' PKCS#12 API. Certificates and keys
581 can be added using this API to created arbitrary PKCS#12
582 files while avoiding the low level API.
584 New options to PKCS12_create(), key or cert can be NULL and
585 will then be omitted from the output file. The encryption
586 algorithm NIDs can be set to -1 for no encryption, the mac
587 iteration count can be set to 0 to omit the mac.
589 Enhance pkcs12 utility by making the -nokeys and -nocerts
590 options work when creating a PKCS#12 file. New option -nomac
591 to omit the mac, NONE can be set for an encryption algorithm.
592 New code is modified to use the enhanced PKCS12_create()
593 instead of the low level API.
596 *) Extend ASN1 encoder to support indefinite length constructed
597 encoding. This can output sequences tags and octet strings in
598 this form. Modify pk7_asn1.c to support indefinite length
599 encoding. This is experimental and needs additional code to
600 be useful, such as an ASN1 bio and some enhanced streaming
603 Extend template encode functionality so that tagging is passed
604 down to the template encoder.
607 *) Let 'openssl req' fail if an argument to '-newkey' is not
608 recognized instead of using RSA as a default.
611 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
612 As these are not official, they are not included in "ALL";
613 the "ECCdraft" ciphersuite group alias can be used to select them.
614 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
616 *) Add ECDH engine support.
617 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
619 *) Add ECDH in new directory crypto/ecdh/.
620 [Douglas Stebila (Sun Microsystems Laboratories)]
622 *) Let BN_rand_range() abort with an error after 100 iterations
623 without success (which indicates a broken PRNG).
626 *) Change BN_mod_sqrt() so that it verifies that the input value
627 is really the square of the return value. (Previously,
628 BN_mod_sqrt would show GIGO behaviour.)
631 *) Add named elliptic curves over binary fields from X9.62, SECG,
632 and WAP/WTLS; add OIDs that were still missing.
634 [Sheueling Chang Shantz and Douglas Stebila
635 (Sun Microsystems Laboratories)]
637 *) Extend the EC library for elliptic curves over binary fields
638 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
641 EC_GF2m_simple_method
645 EC_GROUP_new_curve_GF2m
646 EC_GROUP_set_curve_GF2m
647 EC_GROUP_get_curve_GF2m
648 EC_POINT_set_affine_coordinates_GF2m
649 EC_POINT_get_affine_coordinates_GF2m
650 EC_POINT_set_compressed_coordinates_GF2m
652 Point compression for binary fields is disabled by default for
653 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
656 As binary polynomials are represented as BIGNUMs, various members
657 of the EC_GROUP and EC_POINT data structures can be shared
658 between the implementations for prime fields and binary fields;
659 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
660 are essentially identical to their ..._GFp counterparts.
661 (For simplicity, the '..._GFp' prefix has been dropped from
662 various internal method names.)
664 An internal 'field_div' method (similar to 'field_mul' and
665 'field_sqr') has been added; this is used only for binary fields.
667 [Sheueling Chang Shantz and Douglas Stebila
668 (Sun Microsystems Laboratories)]
670 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
671 through methods ('mul', 'precompute_mult').
673 The generic implementations (now internally called 'ec_wNAF_mul'
674 and 'ec_wNAF_precomputed_mult') remain the default if these
675 methods are undefined.
677 [Sheueling Chang Shantz and Douglas Stebila
678 (Sun Microsystems Laboratories)]
680 *) New function EC_GROUP_get_degree, which is defined through
681 EC_METHOD. For curves over prime fields, this returns the bit
682 length of the modulus.
684 [Sheueling Chang Shantz and Douglas Stebila
685 (Sun Microsystems Laboratories)]
687 *) New functions EC_GROUP_dup, EC_POINT_dup.
688 (These simply call ..._new and ..._copy).
690 [Sheueling Chang Shantz and Douglas Stebila
691 (Sun Microsystems Laboratories)]
693 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
694 Polynomials are represented as BIGNUMs (where the sign bit is not
695 used) in the following functions [macros]:
698 BN_GF2m_sub [= BN_GF2m_add]
699 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
700 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
701 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
703 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
704 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
705 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
706 BN_GF2m_cmp [= BN_ucmp]
708 (Note that only the 'mod' functions are actually for fields GF(2^m).
709 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
711 For some functions, an the irreducible polynomial defining a
712 field can be given as an 'unsigned int[]' with strictly
713 decreasing elements giving the indices of those bits that are set;
714 i.e., p[] represents the polynomial
715 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
717 p[0] > p[1] > ... > p[k] = 0.
718 This applies to the following functions:
723 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
724 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
727 BN_GF2m_mod_solve_quad_arr
731 Conversion can be performed by the following functions:
736 bntest.c has additional tests for binary polynomial arithmetic.
738 Two implementations for BN_GF2m_mod_div() are available.
739 The default algorithm simply uses BN_GF2m_mod_inv() and
740 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
741 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
742 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
744 [Sheueling Chang Shantz and Douglas Stebila
745 (Sun Microsystems Laboratories)]
747 *) Add new error code 'ERR_R_DISABLED' that can be used when some
748 functionality is disabled at compile-time.
749 [Douglas Stebila <douglas.stebila@sun.com>]
751 *) Change default behaviour of 'openssl asn1parse' so that more
752 information is visible when viewing, e.g., a certificate:
754 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
755 mode the content of non-printable OCTET STRINGs is output in a
756 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
757 avoid the appearance of a printable string.
758 [Nils Larsch <nla@trustcenter.de>]
760 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
762 EC_GROUP_set_asn1_flag()
763 EC_GROUP_get_asn1_flag()
764 EC_GROUP_set_point_conversion_form()
765 EC_GROUP_get_point_conversion_form()
766 These control ASN1 encoding details:
767 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
768 has been set to OPENSSL_EC_NAMED_CURVE.
769 - Points are encoded in uncompressed form by default; options for
770 asn1_for are as for point2oct, namely
771 POINT_CONVERSION_COMPRESSED
772 POINT_CONVERSION_UNCOMPRESSED
773 POINT_CONVERSION_HYBRID
775 Also add 'seed' and 'seed_len' members to EC_GROUP with access
779 EC_GROUP_get_seed_len()
780 This is used only for ASN1 purposes (so far).
781 [Nils Larsch <nla@trustcenter.de>]
783 *) Add 'field_type' member to EC_METHOD, which holds the NID
784 of the appropriate field type OID. The new function
785 EC_METHOD_get_field_type() returns this value.
786 [Nils Larsch <nla@trustcenter.de>]
793 providing useful interfaces to EC_POINT_point2oct() and
794 EC_POINT_oct2point().
795 [Nils Larsch <nla@trustcenter.de>]
797 *) Change internals of the EC library so that the functions
798 EC_GROUP_set_generator()
799 EC_GROUP_get_generator()
801 EC_GROUP_get_cofactor()
802 are implemented directly in crypto/ec/ec_lib.c and not dispatched
803 to methods, which would lead to unnecessary code duplication when
804 adding different types of curves.
805 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
807 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
808 arithmetic, and such that modified wNAFs are generated
809 (which avoid length expansion in many cases).
812 *) Add a function EC_GROUP_check_discriminant() (defined via
813 EC_METHOD) that verifies that the curve discriminant is non-zero.
815 Add a function EC_GROUP_check() that makes some sanity tests
816 on a EC_GROUP, its generator and order. This includes
817 EC_GROUP_check_discriminant().
818 [Nils Larsch <nla@trustcenter.de>]
820 *) Add ECDSA in new directory crypto/ecdsa/.
822 Add applications 'openssl ecparam' and 'openssl ecdsa'
823 (these are based on 'openssl dsaparam' and 'openssl dsa').
825 ECDSA support is also included in various other files across the
826 library. Most notably,
827 - 'openssl req' now has a '-newkey ecdsa:file' option;
828 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
829 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
830 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
831 them suitable for ECDSA where domain parameters must be
832 extracted before the specific public key;
833 - ECDSA engine support has been added.
834 [Nils Larsch <nla@trustcenter.de>]
836 *) Include some named elliptic curves, and add OIDs from X9.62,
837 SECG, and WAP/WTLS. Each curve can be obtained from the new
839 EC_GROUP_new_by_curve_name(),
840 and the list of available named curves can be obtained with
841 EC_get_builtin_curves().
842 Also add a 'curve_name' member to EC_GROUP objects, which can be
844 EC_GROUP_set_curve_name()
845 EC_GROUP_get_curve_name()
846 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
848 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
849 was actually never needed) and in BN_mul(). The removal in BN_mul()
850 required a small change in bn_mul_part_recursive() and the addition
851 of the functions bn_cmp_part_words(), bn_sub_part_words() and
852 bn_add_part_words(), which do the same thing as bn_cmp_words(),
853 bn_sub_words() and bn_add_words() except they take arrays with
857 Changes between 0.9.7g and 0.9.7h [XX xxx XXXX]
859 *) Minimal support for X9.31 signatures and PSS padding modes. This is
860 mainly for FIPS compliance and not fully integrated at this stage.
863 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
864 the exponentiation using a fixed-length exponent. (Otherwise,
865 the information leaked through timing could expose the secret key
866 after many signatures; cf. Bleichenbacher's attack on DSA with
870 *) Make a new fixed-window mod_exp implementation the default for
871 RSA, DSA, and DH private-key operations so that the sequence of
872 squares and multiplies and the memory access pattern are
873 independent of the particular secret key. This will mitigate
874 cache-timing and potential related attacks.
876 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
877 and this is automatically used by BN_mod_exp_mont() if the new flag
878 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
879 will use this BN flag for private exponents unless the flag
880 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
881 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
883 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
885 *) Change the client implementation for SSLv23_method() and
886 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
887 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
888 (Previously, the SSL 2.0 backwards compatible Client Hello
889 message format would be used even with SSL_OP_NO_SSLv2.)
892 *) Add support for smime-type MIME parameter in S/MIME messages which some
896 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
897 a threadsafe manner. Modify rsa code to use new function and add calls
898 to dsa and dh code (which had race conditions before).
901 *) Include the fixed error library code in the C error file definitions
902 instead of fixing them up at runtime. This keeps the error code
906 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
908 *) Fixes for newer kerberos headers. NB: the casts are needed because
909 the 'length' field is signed on one version and unsigned on another
910 with no (?) obvious way to tell the difference, without these VC++
911 complains. Also the "definition" of FAR (blank) is no longer included
912 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
913 some needed definitions.
916 *) Undo Cygwin change.
919 *) Added support for proxy certificates according to RFC 3820.
920 Because they may be a security thread to unaware applications,
921 they must be explicitely allowed in run-time. See
922 docs/HOWTO/proxy_certificates.txt for further information.
925 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
927 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
928 server and client random values. Previously
929 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
930 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
932 This change has negligible security impact because:
934 1. Server and client random values still have 24 bytes of pseudo random
937 2. Server and client random values are sent in the clear in the initial
940 3. The master secret is derived using the premaster secret (48 bytes in
941 size for static RSA ciphersuites) as well as client server and random
944 The OpenSSL team would like to thank the UK NISCC for bringing this issue
947 [Stephen Henson, reported by UK NISCC]
949 *) Use Windows randomness collection on Cygwin.
952 *) Fix hang in EGD/PRNGD query when communication socket is closed
953 prematurely by EGD/PRNGD.
954 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
956 *) Prompt for pass phrases when appropriate for PKCS12 input format.
959 *) Back-port of selected performance improvements from development
960 branch, as well as improved support for PowerPC platforms.
963 *) Add lots of checks for memory allocation failure, error codes to indicate
964 failure and freeing up memory if a failure occurs.
965 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
967 *) Add new -passin argument to dgst.
970 *) Perform some character comparisons of different types in X509_NAME_cmp:
971 this is needed for some certificates that reencode DNs into UTF8Strings
972 (in violation of RFC3280) and can't or wont issue name rollover
976 *) Make an explicit check during certificate validation to see that
977 the CA setting in each certificate on the chain is correct. As a
978 side effect always do the following basic checks on extensions,
979 not just when there's an associated purpose to the check:
981 - if there is an unhandled critical extension (unless the user
982 has chosen to ignore this fault)
983 - if the path length has been exceeded (if one is set at all)
984 - that certain extensions fit the associated purpose (if one has
988 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
990 *) Avoid a race condition when CRLs are checked in a multi threaded
991 environment. This would happen due to the reordering of the revoked
992 entries during signature checking and serial number lookup. Now the
993 encoding is cached and the serial number sort performed under a lock.
994 Add new STACK function sk_is_sorted().
997 *) Add Delta CRL to the extension code.
1000 *) Various fixes to s3_pkt.c so alerts are sent properly.
1001 [David Holmes <d.holmes@f5.com>]
1003 *) Reduce the chances of duplicate issuer name and serial numbers (in
1004 violation of RFC3280) using the OpenSSL certificate creation utilities.
1005 This is done by creating a random 64 bit value for the initial serial
1006 number when a serial number file is created or when a self signed
1007 certificate is created using 'openssl req -x509'. The initial serial
1008 number file is created using 'openssl x509 -next_serial' in CA.pl
1009 rather than being initialized to 1.
1012 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1014 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1015 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
1016 [Joe Orton, Steve Henson]
1018 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1020 [Joe Orton, Steve Henson]
1022 *) Make it possible to have multiple active certificates with the same
1023 subject in the CA index file. This is done only if the keyword
1024 'unique_subject' is set to 'no' in the main CA section (default
1025 if 'CA_default') of the configuration file. The value is saved
1026 with the database itself in a separate index attribute file,
1027 named like the index file with '.attr' appended to the name.
1030 *) X509 verify fixes. Disable broken certificate workarounds when
1031 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1032 keyUsage extension present. Don't accept CRLs with unhandled critical
1033 extensions: since verify currently doesn't process CRL extensions this
1034 rejects a CRL with *any* critical extensions. Add new verify error codes
1038 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1039 A clarification of RFC2560 will require the use of OCTET STRINGs and
1040 some implementations cannot handle the current raw format. Since OpenSSL
1041 copies and compares OCSP nonces as opaque blobs without any attempt at
1042 parsing them this should not create any compatibility issues.
1045 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1046 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1047 this HMAC (and other) operations are several times slower than OpenSSL
1051 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1052 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1054 *) Use the correct content when signing type "other".
1057 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1059 *) Fix various bugs revealed by running the NISCC test suite:
1061 Stop out of bounds reads in the ASN1 code when presented with
1062 invalid tags (CAN-2003-0543 and CAN-2003-0544).
1064 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
1066 If verify callback ignores invalid public key errors don't try to check
1067 certificate signature with the NULL public key.
1071 *) New -ignore_err option in ocsp application to stop the server
1072 exiting on the first error in a request.
1075 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1076 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1080 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1081 extra data after the compression methods not only for TLS 1.0
1082 but also for SSL 3.0 (as required by the specification).
1083 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1085 *) Change X509_certificate_type() to mark the key as exported/exportable
1086 when it's 512 *bits* long, not 512 bytes.
1089 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1090 blocks during encryption.
1093 *) Various fixes to base64 BIO and non blocking I/O. On write
1094 flushes were not handled properly if the BIO retried. On read
1095 data was not being buffered properly and had various logic bugs.
1096 This also affects blocking I/O when the data being decoded is a
1100 *) Various S/MIME bugfixes and compatibility changes:
1101 output correct application/pkcs7 MIME type if
1102 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1103 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1104 of files as .eml work). Correctly handle very long lines in MIME
1108 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1110 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1111 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1112 a protocol version number mismatch like a decryption error
1113 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1116 *) Turn on RSA blinding by default in the default implementation
1117 to avoid a timing attack. Applications that don't want it can call
1118 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1119 They would be ill-advised to do so in most cases.
1120 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1122 *) Change RSA blinding code so that it works when the PRNG is not
1123 seeded (in this case, the secret RSA exponent is abused as
1124 an unpredictable seed -- if it is not unpredictable, there
1125 is no point in blinding anyway). Make RSA blinding thread-safe
1126 by remembering the creator's thread ID in rsa->blinding and
1127 having all other threads use local one-time blinding factors
1128 (this requires more computation than sharing rsa->blinding, but
1129 avoids excessive locking; and if an RSA object is not shared
1130 between threads, blinding will still be very fast).
1133 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1134 ENGINE as defaults for all supported algorithms irrespective of
1135 the 'flags' parameter. 'flags' is now honoured, so applications
1136 should make sure they are passing it correctly.
1139 *) Target "mingw" now allows native Windows code to be generated in
1140 the Cygwin environment as well as with the MinGW compiler.
1143 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1145 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1146 via timing by performing a MAC computation even if incorrrect
1147 block cipher padding has been found. This is a countermeasure
1148 against active attacks where the attacker has to distinguish
1149 between bad padding and a MAC verification error. (CAN-2003-0078)
1151 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1152 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1153 Martin Vuagnoux (EPFL, Ilion)]
1155 *) Make the no-err option work as intended. The intention with no-err
1156 is not to have the whole error stack handling routines removed from
1157 libcrypto, it's only intended to remove all the function name and
1158 reason texts, thereby removing some of the footprint that may not
1159 be interesting if those errors aren't displayed anyway.
1161 NOTE: it's still possible for any application or module to have it's
1162 own set of error texts inserted. The routines are there, just not
1163 used by default when no-err is given.
1166 *) Add support for FreeBSD on IA64.
1167 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1169 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1170 Kerberos function mit_des_cbc_cksum(). Before this change,
1171 the value returned by DES_cbc_cksum() was like the one from
1172 mit_des_cbc_cksum(), except the bytes were swapped.
1173 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1175 *) Allow an application to disable the automatic SSL chain building.
1176 Before this a rather primitive chain build was always performed in
1177 ssl3_output_cert_chain(): an application had no way to send the
1178 correct chain if the automatic operation produced an incorrect result.
1180 Now the chain builder is disabled if either:
1182 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1184 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1186 The reasoning behind this is that an application would not want the
1187 auto chain building to take place if extra chain certificates are
1188 present and it might also want a means of sending no additional
1189 certificates (for example the chain has two certificates and the
1193 *) Add the possibility to build without the ENGINE framework.
1194 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1196 *) Under Win32 gmtime() can return NULL: check return value in
1197 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1200 *) DSA routines: under certain error conditions uninitialized BN objects
1201 could be freed. Solution: make sure initialization is performed early
1202 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1203 Nils Larsch <nla@trustcenter.de> via PR#459)
1206 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1207 checked on reconnect on the client side, therefore session resumption
1208 could still fail with a "ssl session id is different" error. This
1209 behaviour is masked when SSL_OP_ALL is used due to
1210 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1211 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1212 followup to PR #377.
1215 *) IA-32 assembler support enhancements: unified ELF targets, support
1216 for SCO/Caldera platforms, fix for Cygwin shared build.
1219 *) Add support for FreeBSD on sparc64. As a consequence, support for
1220 FreeBSD on non-x86 processors is separate from x86 processors on
1221 the config script, much like the NetBSD support.
1222 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1224 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1226 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1229 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1230 code (06) was taken as the first octet of the session ID and the last
1231 octet was ignored consequently. As a result SSLv2 client side session
1232 caching could not have worked due to the session ID mismatch between
1234 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1238 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1239 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1243 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1244 seems that in spite of existing for more than a year, many application
1245 author have done nothing to provide the necessary callbacks, which
1246 means that this particular engine will not work properly anywhere.
1247 This is a very unfortunate situation which forces us, in the name
1248 of usability, to give the hw_ncipher.c a static lock, which is part
1250 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1251 appear in 0.9.8 or later. We EXPECT application authors to have
1252 dealt properly with this when 0.9.8 is released (unless we actually
1253 make such changes in the libcrypto locking code that changes will
1254 have to be made anyway).
1257 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1258 octets have been read, EOF or an error occurs. Without this change
1259 some truncated ASN1 structures will not produce an error.
1262 *) Disable Heimdal support, since it hasn't been fully implemented.
1263 Still give the possibility to force the use of Heimdal, but with
1264 warnings and a request that patches get sent to openssl-dev.
1267 *) Add the VC-CE target, introduce the WINCE sysname, and add
1268 INSTALL.WCE and appropriate conditionals to make it build.
1269 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1271 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1272 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1273 edit numbers of the version.
1274 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1276 *) Introduce safe string copy and catenation functions
1277 (BUF_strlcpy() and BUF_strlcat()).
1278 [Ben Laurie (CHATS) and Richard Levitte]
1280 *) Avoid using fixed-size buffers for one-line DNs.
1281 [Ben Laurie (CHATS)]
1283 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1284 resizing buffers containing secrets, and use where appropriate.
1285 [Ben Laurie (CHATS)]
1287 *) Avoid using fixed size buffers for configuration file location.
1288 [Ben Laurie (CHATS)]
1290 *) Avoid filename truncation for various CA files.
1291 [Ben Laurie (CHATS)]
1293 *) Use sizeof in preference to magic numbers.
1294 [Ben Laurie (CHATS)]
1296 *) Avoid filename truncation in cert requests.
1297 [Ben Laurie (CHATS)]
1299 *) Add assertions to check for (supposedly impossible) buffer
1301 [Ben Laurie (CHATS)]
1303 *) Don't cache truncated DNS entries in the local cache (this could
1304 potentially lead to a spoofing attack).
1305 [Ben Laurie (CHATS)]
1307 *) Fix various buffers to be large enough for hex/decimal
1308 representations in a platform independent manner.
1309 [Ben Laurie (CHATS)]
1311 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1312 resizing buffers containing secrets, and use where appropriate.
1313 [Ben Laurie (CHATS)]
1315 *) Add BIO_indent() to avoid much slightly worrying code to do
1317 [Ben Laurie (CHATS)]
1319 *) Convert sprintf()/BIO_puts() to BIO_printf().
1320 [Ben Laurie (CHATS)]
1322 *) buffer_gets() could terminate with the buffer only half
1324 [Ben Laurie (CHATS)]
1326 *) Add assertions to prevent user-supplied crypto functions from
1327 overflowing internal buffers by having large block sizes, etc.
1328 [Ben Laurie (CHATS)]
1330 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1332 [Ben Laurie (CHATS)]
1334 *) Eliminate unused copy of key in RC4.
1335 [Ben Laurie (CHATS)]
1337 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1338 [Ben Laurie (CHATS)]
1340 *) Fix off-by-one error in EGD path.
1341 [Ben Laurie (CHATS)]
1343 *) If RANDFILE path is too long, ignore instead of truncating.
1344 [Ben Laurie (CHATS)]
1346 *) Eliminate unused and incorrectly sized X.509 structure
1348 [Ben Laurie (CHATS)]
1350 *) Eliminate unused and dangerous function knumber().
1351 [Ben Laurie (CHATS)]
1353 *) Eliminate unused and dangerous structure, KSSL_ERR.
1354 [Ben Laurie (CHATS)]
1356 *) Protect against overlong session ID context length in an encoded
1357 session object. Since these are local, this does not appear to be
1359 [Ben Laurie (CHATS)]
1361 *) Change from security patch (see 0.9.6e below) that did not affect
1362 the 0.9.6 release series:
1364 Remote buffer overflow in SSL3 protocol - an attacker could
1365 supply an oversized master key in Kerberos-enabled versions.
1367 [Ben Laurie (CHATS)]
1369 *) Change the SSL kerb5 codes to match RFC 2712.
1372 *) Make -nameopt work fully for req and add -reqopt switch.
1373 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1375 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1376 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1378 *) Make sure tests can be performed even if the corresponding algorithms
1379 have been removed entirely. This was also the last step to make
1380 OpenSSL compilable with DJGPP under all reasonable conditions.
1381 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1383 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1384 to allow version independent disabling of normally unselected ciphers,
1385 which may be activated as a side-effect of selecting a single cipher.
1387 (E.g., cipher list string "RSA" enables ciphersuites that are left
1388 out of "ALL" because they do not provide symmetric encryption.
1389 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1390 [Lutz Jaenicke, Bodo Moeller]
1392 *) Add appropriate support for separate platform-dependent build
1393 directories. The recommended way to make a platform-dependent
1394 build directory is the following (tested on Linux), maybe with
1397 # Place yourself outside of the OpenSSL source tree. In
1398 # this example, the environment variable OPENSSL_SOURCE
1399 # is assumed to contain the absolute OpenSSL source directory.
1400 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1401 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1402 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1403 mkdir -p `dirname $F`
1404 ln -s $OPENSSL_SOURCE/$F $F
1407 To be absolutely sure not to disturb the source tree, a "make clean"
1408 is a good thing. If it isn't successfull, don't worry about it,
1409 it probably means the source directory is very clean.
1412 *) Make sure any ENGINE control commands make local copies of string
1413 pointers passed to them whenever necessary. Otherwise it is possible
1414 the caller may have overwritten (or deallocated) the original string
1415 data when a later ENGINE operation tries to use the stored values.
1416 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1418 *) Improve diagnostics in file reading and command-line digests.
1419 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1421 *) Add AES modes CFB and OFB to the object database. Correct an
1422 error in AES-CFB decryption.
1425 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1426 allows existing EVP_CIPHER_CTX structures to be reused after
1427 calling EVP_*Final(). This behaviour is used by encryption
1428 BIOs and some applications. This has the side effect that
1429 applications must explicitly clean up cipher contexts with
1430 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1433 *) Check the values of dna and dnb in bn_mul_recursive before calling
1434 bn_mul_comba (a non zero value means the a or b arrays do not contain
1435 n2 elements) and fallback to bn_mul_normal if either is not zero.
1438 *) Fix escaping of non-ASCII characters when using the -subj option
1439 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1442 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1443 form for "surname", serialNumber has no short form.
1444 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1445 therefore remove "mail" short name for "internet 7".
1446 The OID for unique identifiers in X509 certificates is
1447 x500UniqueIdentifier, not uniqueIdentifier.
1448 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1451 *) Add an "init" command to the ENGINE config module and auto initialize
1452 ENGINEs. Without any "init" command the ENGINE will be initialized
1453 after all ctrl commands have been executed on it. If init=1 the
1454 ENGINE is initailized at that point (ctrls before that point are run
1455 on the uninitialized ENGINE and after on the initialized one). If
1456 init=0 then the ENGINE will not be iniatialized at all.
1459 *) Fix the 'app_verify_callback' interface so that the user-defined
1460 argument is actually passed to the callback: In the
1461 SSL_CTX_set_cert_verify_callback() prototype, the callback
1462 declaration has been changed from
1465 int (*cb)(X509_STORE_CTX *,void *);
1466 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1467 i=s->ctx->app_verify_callback(&ctx)
1468 has been changed into
1469 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1471 To update applications using SSL_CTX_set_cert_verify_callback(),
1472 a dummy argument can be added to their callback functions.
1473 [D. K. Smetters <smetters@parc.xerox.com>]
1475 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1476 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1478 *) Add and OPENSSL_LOAD_CONF define which will cause
1479 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1480 This allows older applications to transparently support certain
1481 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1482 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1483 load the config file and OPENSSL_add_all_algorithms_conf() which will
1484 always load it have also been added.
1487 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1488 Adjust NIDs and EVP layer.
1489 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1491 *) Config modules support in openssl utility.
1493 Most commands now load modules from the config file,
1494 though in a few (such as version) this isn't done
1495 because it couldn't be used for anything.
1497 In the case of ca and req the config file used is
1498 the same as the utility itself: that is the -config
1499 command line option can be used to specify an
1503 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1504 use "openssl_conf" if filename is NULL use default openssl config file.
1507 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1508 config section name. Add a new flag to tolerate a missing config file
1509 and move code to CONF_modules_load_file().
1512 *) Support for crypto accelerator cards from Accelerated Encryption
1513 Processing, www.aep.ie. (Use engine 'aep')
1514 The support was copied from 0.9.6c [engine] and adapted/corrected
1515 to work with the new engine framework.
1516 [AEP Inc. and Richard Levitte]
1518 *) Support for SureWare crypto accelerator cards from Baltimore
1519 Technologies. (Use engine 'sureware')
1520 The support was copied from 0.9.6c [engine] and adapted
1521 to work with the new engine framework.
1524 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1525 make the newer ENGINE framework commands for the CHIL engine work.
1526 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1528 *) Make it possible to produce shared libraries on ReliantUNIX.
1529 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1531 *) Add the configuration target debug-linux-ppro.
1532 Make 'openssl rsa' use the general key loading routines
1533 implemented in apps.c, and make those routines able to
1534 handle the key format FORMAT_NETSCAPE and the variant
1536 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1538 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1539 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1541 *) Add -keyform to rsautl, and document -engine.
1542 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1544 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1545 BIO_R_NO_SUCH_FILE error code rather than the generic
1546 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1549 *) Add new functions
1551 ERR_peek_last_error_line
1552 ERR_peek_last_error_line_data.
1553 These are similar to
1556 ERR_peek_error_line_data,
1557 but report on the latest error recorded rather than the first one
1558 still in the error queue.
1559 [Ben Laurie, Bodo Moeller]
1561 *) default_algorithms option in ENGINE config module. This allows things
1563 default_algorithms = ALL
1564 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1567 *) Prelminary ENGINE config module.
1570 *) New experimental application configuration code.
1573 *) Change the AES code to follow the same name structure as all other
1574 symmetric ciphers, and behave the same way. Move everything to
1575 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1576 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1578 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1579 [Ben Laurie and Theo de Raadt]
1581 *) Add option to output public keys in req command.
1582 [Massimiliano Pala madwolf@openca.org]
1584 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1585 (up to about 10% better than before for P-192 and P-224).
1588 *) New functions/macros
1590 SSL_CTX_set_msg_callback(ctx, cb)
1591 SSL_CTX_set_msg_callback_arg(ctx, arg)
1592 SSL_set_msg_callback(ssl, cb)
1593 SSL_set_msg_callback_arg(ssl, arg)
1595 to request calling a callback function
1597 void cb(int write_p, int version, int content_type,
1598 const void *buf, size_t len, SSL *ssl, void *arg)
1600 whenever a protocol message has been completely received
1601 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1602 protocol version according to which the SSL library interprets
1603 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1604 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1605 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1606 specification (change_cipher_spec(20), alert(21), handshake(22)).
1607 'buf' and 'len' point to the actual message, 'ssl' to the
1608 SSL object, and 'arg' is the application-defined value set by
1609 SSL[_CTX]_set_msg_callback_arg().
1611 'openssl s_client' and 'openssl s_server' have new '-msg' options
1612 to enable a callback that displays all protocol messages.
1615 *) Change the shared library support so shared libraries are built as
1616 soon as the corresponding static library is finished, and thereby get
1617 openssl and the test programs linked against the shared library.
1618 This still only happens when the keyword "shard" has been given to
1619 the configuration scripts.
1621 NOTE: shared library support is still an experimental thing, and
1622 backward binary compatibility is still not guaranteed.
1623 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1625 *) Add support for Subject Information Access extension.
1626 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1628 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1629 additional bytes when new memory had to be allocated, not just
1630 when reusing an existing buffer.
1633 *) New command line and configuration option 'utf8' for the req command.
1634 This allows field values to be specified as UTF8 strings.
1637 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1638 runs for the former and machine-readable output for the latter.
1641 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1642 of the e-mail address in the DN (i.e., it will go into a certificate
1643 extension only). The new configuration file option 'email_in_dn = no'
1644 has the same effect.
1645 [Massimiliano Pala madwolf@openca.org]
1647 *) Change all functions with names starting with des_ to be starting
1648 with DES_ instead. Add wrappers that are compatible with libdes,
1649 but are named _ossl_old_des_*. Finally, add macros that map the
1650 des_* symbols to the corresponding _ossl_old_des_* if libdes
1651 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1652 desired, the des_* symbols will be mapped to DES_*, with one
1655 Since we provide two compatibility mappings, the user needs to
1656 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1657 compatibility is desired. The default (i.e., when that macro
1658 isn't defined) is OpenSSL 0.9.6c compatibility.
1660 There are also macros that enable and disable the support of old
1661 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1662 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1663 are defined, the default will apply: to support the old des routines.
1665 In either case, one must include openssl/des.h to get the correct
1666 definitions. Do not try to just include openssl/des_old.h, that
1669 NOTE: This is a major break of an old API into a new one. Software
1670 authors are encouraged to switch to the DES_ style functions. Some
1671 time in the future, des_old.h and the libdes compatibility functions
1672 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1673 default), and then completely removed.
1676 *) Test for certificates which contain unsupported critical extensions.
1677 If such a certificate is found during a verify operation it is
1678 rejected by default: this behaviour can be overridden by either
1679 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1680 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1681 X509_supported_extension() has also been added which returns 1 if a
1682 particular extension is supported.
1685 *) Modify the behaviour of EVP cipher functions in similar way to digests
1686 to retain compatibility with existing code.
1689 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1690 compatibility with existing code. In particular the 'ctx' parameter does
1691 not have to be to be initialized before the call to EVP_DigestInit() and
1692 it is tidied up after a call to EVP_DigestFinal(). New function
1693 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1694 EVP_MD_CTX_copy() changed to not require the destination to be
1695 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1696 requires the destination to be valid.
1698 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1699 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1702 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1703 so that complete 'Handshake' protocol structures are kept in memory
1704 instead of overwriting 'msg_type' and 'length' with 'body' data.
1707 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1708 [Massimo Santin via Richard Levitte]
1710 *) Major restructuring to the underlying ENGINE code. This includes
1711 reduction of linker bloat, separation of pure "ENGINE" manipulation
1712 (initialisation, etc) from functionality dealing with implementations
1713 of specific crypto iterfaces. This change also introduces integrated
1714 support for symmetric ciphers and digest implementations - so ENGINEs
1715 can now accelerate these by providing EVP_CIPHER and EVP_MD
1716 implementations of their own. This is detailed in crypto/engine/README
1717 as it couldn't be adequately described here. However, there are a few
1718 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1719 were changed in the original introduction of ENGINE code have now
1720 reverted back - the hooking from this code to ENGINE is now a good
1721 deal more passive and at run-time, operations deal directly with
1722 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1723 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1724 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1725 they were not being used by the framework as there is no concept of a
1726 BIGNUM_METHOD and they could not be generalised to the new
1727 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1728 ENGINE_cpy() has been removed as it cannot be consistently defined in
1732 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1735 *) Change mkdef.pl to sort symbols that get the same entry number,
1736 and make sure the automatically generated functions ERR_load_*
1737 become part of libeay.num as well.
1740 *) New function SSL_renegotiate_pending(). This returns true once
1741 renegotiation has been requested (either SSL_renegotiate() call
1742 or HelloRequest/ClientHello receveived from the peer) and becomes
1743 false once a handshake has been completed.
1744 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1745 sends a HelloRequest, but does not ensure that a handshake takes
1746 place. SSL_renegotiate_pending() is useful for checking if the
1747 client has followed the request.)
1750 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1751 By default, clients may request session resumption even during
1752 renegotiation (if session ID contexts permit); with this option,
1753 session resumption is possible only in the first handshake.
1755 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1756 more bits available for options that should not be part of
1757 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1760 *) Add some demos for certificate and certificate request creation.
1763 *) Make maximum certificate chain size accepted from the peer application
1764 settable (SSL*_get/set_max_cert_list()), as proposed by
1765 "Douglas E. Engert" <deengert@anl.gov>.
1768 *) Add support for shared libraries for Unixware-7
1769 (Boyd Lynn Gerber <gerberb@zenez.com>).
1772 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1773 be done prior to destruction. Use this to unload error strings from
1774 ENGINEs that load their own error strings. NB: This adds two new API
1775 functions to "get" and "set" this destroy handler in an ENGINE.
1778 *) Alter all existing ENGINE implementations (except "openssl" and
1779 "openbsd") to dynamically instantiate their own error strings. This
1780 makes them more flexible to be built both as statically-linked ENGINEs
1781 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1782 Also, add stub code to each that makes building them as self-contained
1783 shared-libraries easier (see README.ENGINE).
1786 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1787 implementations into applications that are completely implemented in
1788 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1789 commands that can be used to configure what shared-library to load and
1790 to control aspects of the way it is handled. Also, made an update to
1791 the README.ENGINE file that brings its information up-to-date and
1792 provides some information and instructions on the "dynamic" ENGINE
1793 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1796 *) Make it possible to unload ranges of ERR strings with a new
1797 "ERR_unload_strings" function.
1800 *) Add a copy() function to EVP_MD.
1803 *) Make EVP_MD routines take a context pointer instead of just the
1804 md_data void pointer.
1807 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1808 that the digest can only process a single chunk of data
1809 (typically because it is provided by a piece of
1810 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1811 is only going to provide a single chunk of data, and hence the
1812 framework needn't accumulate the data for oneshot drivers.
1815 *) As with "ERR", make it possible to replace the underlying "ex_data"
1816 functions. This change also alters the storage and management of global
1817 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1818 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1819 index counters. The API functions that use this state have been changed
1820 to take a "class_index" rather than pointers to the class's local STACK
1821 and counter, and there is now an API function to dynamically create new
1822 classes. This centralisation allows us to (a) plug a lot of the
1823 thread-safety problems that existed, and (b) makes it possible to clean
1824 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1825 such data would previously have always leaked in application code and
1826 workarounds were in place to make the memory debugging turn a blind eye
1827 to it. Application code that doesn't use this new function will still
1828 leak as before, but their memory debugging output will announce it now
1829 rather than letting it slide.
1831 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1832 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1833 has a return value to indicate success or failure.
1836 *) Make it possible to replace the underlying "ERR" functions such that the
1837 global state (2 LHASH tables and 2 locks) is only used by the "default"
1838 implementation. This change also adds two functions to "get" and "set"
1839 the implementation prior to it being automatically set the first time
1840 any other ERR function takes place. Ie. an application can call "get",
1841 pass the return value to a module it has just loaded, and that module
1842 can call its own "set" function using that value. This means the
1843 module's "ERR" operations will use (and modify) the error state in the
1844 application and not in its own statically linked copy of OpenSSL code.
1847 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1848 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1849 the operation, and provides a more encapsulated way for external code
1850 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1851 to use these functions rather than manually incrementing the counts.
1853 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1856 *) Add EVP test program.
1859 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1862 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1863 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1864 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1865 These allow a CRL to be built without having to access X509_CRL fields
1866 directly. Modify 'ca' application to use new functions.
1869 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1870 bug workarounds. Rollback attack detection is a security feature.
1871 The problem will only arise on OpenSSL servers when TLSv1 is not
1872 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1873 Software authors not wanting to support TLSv1 will have special reasons
1874 for their choice and can explicitly enable this option.
1875 [Bodo Moeller, Lutz Jaenicke]
1877 *) Rationalise EVP so it can be extended: don't include a union of
1878 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1879 (similar to those existing for EVP_CIPHER_CTX).
1884 EVP_MD_CTX_init(&md); /* new function call */
1885 EVP_DigestInit(&md, EVP_sha1());
1886 EVP_DigestUpdate(&md, in, len);
1887 EVP_DigestFinal(&md, out, NULL);
1888 EVP_MD_CTX_cleanup(&md); /* new function call */
1892 *) Make DES key schedule conform to the usual scheme, as well as
1893 correcting its structure. This means that calls to DES functions
1894 now have to pass a pointer to a des_key_schedule instead of a
1895 plain des_key_schedule (which was actually always a pointer
1898 des_key_schedule ks;
1900 des_set_key_checked(..., &ks);
1901 des_ncbc_encrypt(..., &ks, ...);
1903 (Note that a later change renames 'des_...' into 'DES_...'.)
1906 *) Initial reduction of linker bloat: the use of some functions, such as
1907 PEM causes large amounts of unused functions to be linked in due to
1908 poor organisation. For example pem_all.c contains every PEM function
1909 which has a knock on effect of linking in large amounts of (unused)
1910 ASN1 code. Grouping together similar functions and splitting unrelated
1911 functions prevents this.
1914 *) Cleanup of EVP macros.
1917 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1918 correct _ecb suffix.
1921 *) Add initial OCSP responder support to ocsp application. The
1922 revocation information is handled using the text based index
1923 use by the ca application. The responder can either handle
1924 requests generated internally, supplied in files (for example
1925 via a CGI script) or using an internal minimal server.
1928 *) Add configuration choices to get zlib compression for TLS.
1931 *) Changes to Kerberos SSL for RFC 2712 compliance:
1932 1. Implemented real KerberosWrapper, instead of just using
1933 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1934 2. Implemented optional authenticator field of KerberosWrapper.
1936 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1937 and authenticator structs; see crypto/krb5/.
1939 Generalized Kerberos calls to support multiple Kerberos libraries.
1940 [Vern Staats <staatsvr@asc.hpc.mil>,
1941 Jeffrey Altman <jaltman@columbia.edu>
1942 via Richard Levitte]
1944 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1945 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1946 values for each of the key sizes rather than having just
1947 parameters (and 'speed' generating keys each time).
1950 *) Speed up EVP routines.
1953 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1954 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1955 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1956 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1958 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1959 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1960 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1963 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1965 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1968 *) Added the OS2-EMX target.
1969 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1971 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1972 to support NCONF routines in extension code. New function CONF_set_nconf()
1973 to allow functions which take an NCONF to also handle the old LHASH
1974 structure: this means that the old CONF compatible routines can be
1975 retained (in particular wrt extensions) without having to duplicate the
1976 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1979 *) Enhance the general user interface with mechanisms for inner control
1980 and with possibilities to have yes/no kind of prompts.
1983 *) Change all calls to low level digest routines in the library and
1984 applications to use EVP. Add missing calls to HMAC_cleanup() and
1985 don't assume HMAC_CTX can be copied using memcpy().
1986 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1988 *) Add the possibility to control engines through control names but with
1989 arbitrary arguments instead of just a string.
1990 Change the key loaders to take a UI_METHOD instead of a callback
1991 function pointer. NOTE: this breaks binary compatibility with earlier
1992 versions of OpenSSL [engine].
1993 Adapt the nCipher code for these new conditions and add a card insertion
1997 *) Enhance the general user interface with mechanisms to better support
1998 dialog box interfaces, application-defined prompts, the possibility
1999 to use defaults (for example default passwords from somewhere else)
2000 and interrupts/cancellations.
2003 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2004 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2007 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2008 tidy up some unnecessarily weird code in 'sk_new()').
2009 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2011 *) Change the key loading routines for ENGINEs to use the same kind
2012 callback (pem_password_cb) as all other routines that need this
2016 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2017 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2018 than this minimum value is recommended.
2021 *) New random seeder for OpenVMS, using the system process statistics
2022 that are easily reachable.
2025 *) Windows apparently can't transparently handle global
2026 variables defined in DLLs. Initialisations such as:
2028 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2030 wont compile. This is used by the any applications that need to
2031 declare their own ASN1 modules. This was fixed by adding the option
2032 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2033 needed for static libraries under Win32.
2036 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2037 setting of purpose and trust fields. New X509_STORE trust and
2038 purpose functions and tidy up setting in other SSL functions.
2041 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2042 structure. These are inherited by X509_STORE_CTX when it is
2043 initialised. This allows various defaults to be set in the
2044 X509_STORE structure (such as flags for CRL checking and custom
2045 purpose or trust settings) for functions which only use X509_STORE_CTX
2046 internally such as S/MIME.
2048 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2049 trust settings if they are not set in X509_STORE. This allows X509_STORE
2050 purposes and trust (in S/MIME for example) to override any set by default.
2052 Add command line options for CRL checking to smime, s_client and s_server
2056 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2057 are set then the CRL is looked up in the X509_STORE structure and
2058 its validity and signature checked, then if the certificate is found
2059 in the CRL the verify fails with a revoked error.
2061 Various new CRL related callbacks added to X509_STORE_CTX structure.
2063 Command line options added to 'verify' application to support this.
2065 This needs some additional work, such as being able to handle multiple
2066 CRLs with different times, extension based lookup (rather than just
2067 by subject name) and ultimately more complete V2 CRL extension
2071 *) Add a general user interface API (crypto/ui/). This is designed
2072 to replace things like des_read_password and friends (backward
2073 compatibility functions using this new API are provided).
2074 The purpose is to remove prompting functions from the DES code
2075 section as well as provide for prompting through dialog boxes in
2076 a window system and the like.
2079 *) Add "ex_data" support to ENGINE so implementations can add state at a
2080 per-structure level rather than having to store it globally.
2083 *) Make it possible for ENGINE structures to be copied when retrieved by
2084 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2085 This causes the "original" ENGINE structure to act like a template,
2086 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2087 operational state can be localised to each ENGINE structure, despite the
2088 fact they all share the same "methods". New ENGINE structures returned in
2089 this case have no functional references and the return value is the single
2090 structural reference. This matches the single structural reference returned
2091 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2095 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2096 needs to match any other type at all we need to manually clear the
2100 *) Changes to the "openssl engine" utility to include;
2101 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2102 about an ENGINE's available control commands.
2103 - executing control commands from command line arguments using the
2104 '-pre' and '-post' switches. '-post' is only used if '-t' is
2105 specified and the ENGINE is successfully initialised. The syntax for
2106 the individual commands are colon-separated, for example;
2107 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2110 *) New dynamic control command support for ENGINEs. ENGINEs can now
2111 declare their own commands (numbers), names (strings), descriptions,
2112 and input types for run-time discovery by calling applications. A
2113 subset of these commands are implicitly classed as "executable"
2114 depending on their input type, and only these can be invoked through
2115 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2116 can be based on user input, config files, etc). The distinction is
2117 that "executable" commands cannot return anything other than a boolean
2118 result and can only support numeric or string input, whereas some
2119 discoverable commands may only be for direct use through
2120 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2121 pointers, or other custom uses. The "executable" commands are to
2122 support parameterisations of ENGINE behaviour that can be
2123 unambiguously defined by ENGINEs and used consistently across any
2124 OpenSSL-based application. Commands have been added to all the
2125 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2126 control over shared-library paths without source code alterations.
2129 *) Changed all ENGINE implementations to dynamically allocate their
2130 ENGINEs rather than declaring them statically. Apart from this being
2131 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2132 this also allows the implementations to compile without using the
2133 internal engine_int.h header.
2136 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2137 'const' value. Any code that should be able to modify a RAND_METHOD
2138 should already have non-const pointers to it (ie. they should only
2139 modify their own ones).
2142 *) Made a variety of little tweaks to the ENGINE code.
2143 - "atalla" and "ubsec" string definitions were moved from header files
2144 to C code. "nuron" string definitions were placed in variables
2145 rather than hard-coded - allowing parameterisation of these values
2146 later on via ctrl() commands.
2147 - Removed unused "#if 0"'d code.
2148 - Fixed engine list iteration code so it uses ENGINE_free() to release
2149 structural references.
2150 - Constified the RAND_METHOD element of ENGINE structures.
2151 - Constified various get/set functions as appropriate and added
2152 missing functions (including a catch-all ENGINE_cpy that duplicates
2153 all ENGINE values onto a new ENGINE except reference counts/state).
2154 - Removed NULL parameter checks in get/set functions. Setting a method
2155 or function to NULL is a way of cancelling out a previously set
2156 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2157 and doesn't justify the extra error symbols and code.
2158 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2159 flags from engine_int.h to engine.h.
2160 - Changed prototypes for ENGINE handler functions (init(), finish(),
2161 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2164 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2165 to the algorithm using long division. The binary algorithm can be
2166 used only if the modulus is odd. On 32-bit systems, it is faster
2167 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2168 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2169 up to 450 bits. In 64-bit environments, the binary algorithm
2170 appears to be advantageous for much longer moduli; here we use it
2171 for moduli up to 2048 bits.
2174 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2175 could not support the combine flag in choice fields.
2178 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2179 extensions from a certificate request to the certificate.
2182 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2183 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2184 file: this allows the display of the certificate about to be
2185 signed to be customised, to allow certain fields to be included
2186 or excluded and extension details. The old system didn't display
2187 multicharacter strings properly, omitted fields not in the policy
2188 and couldn't display additional details such as extensions.
2191 *) Function EC_POINTs_mul for multiple scalar multiplication
2192 of an arbitrary number of elliptic curve points
2193 \sum scalars[i]*points[i],
2194 optionally including the generator defined for the EC_GROUP:
2195 scalar*generator + \sum scalars[i]*points[i].
2197 EC_POINT_mul is a simple wrapper function for the typical case
2198 that the point list has just one item (besides the optional
2202 *) First EC_METHODs for curves over GF(p):
2204 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2205 operations and provides various method functions that can also
2206 operate with faster implementations of modular arithmetic.
2208 EC_GFp_mont_method() reuses most functions that are part of
2209 EC_GFp_simple_method, but uses Montgomery arithmetic.
2211 [Bodo Moeller; point addition and point doubling
2212 implementation directly derived from source code provided by
2213 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2215 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2216 crypto/ec/ec_lib.c):
2218 Curves are EC_GROUP objects (with an optional group generator)
2219 based on EC_METHODs that are built into the library.
2221 Points are EC_POINT objects based on EC_GROUP objects.
2223 Most of the framework would be able to handle curves over arbitrary
2224 finite fields, but as there are no obvious types for fields other
2225 than GF(p), some functions are limited to that for now.
2228 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2229 that the file contains a complete HTTP response.
2232 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2233 change the def and num file printf format specifier from "%-40sXXX"
2234 to "%-39s XXX". The latter will always guarantee a space after the
2235 field while the former will cause them to run together if the field
2236 is 40 of more characters long.
2239 *) Constify the cipher and digest 'method' functions and structures
2240 and modify related functions to take constant EVP_MD and EVP_CIPHER
2244 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2245 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2248 *) Modify EVP_Digest*() routines so they now return values. Although the
2249 internal software routines can never fail additional hardware versions
2253 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2255 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2256 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2259 ERR_R_NESTED_ASN1_ERROR
2261 ERR_R_MISSING_ASN1_EOS
2262 were 4 .. 9, conflicting with
2263 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2265 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2266 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2268 Add new error code 'ERR_R_INTERNAL_ERROR'.
2271 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2275 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2276 sets the subject name for a new request or supersedes the
2277 subject name in a given request. Formats that can be parsed are
2278 'CN=Some Name, OU=myOU, C=IT'
2280 'CN=Some Name/OU=myOU/C=IT'.
2282 Add options '-batch' and '-verbose' to 'openssl req'.
2283 [Massimiliano Pala <madwolf@hackmasters.net>]
2285 *) Introduce the possibility to access global variables through
2286 functions on platform were that's the best way to handle exporting
2287 global variables in shared libraries. To enable this functionality,
2288 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2289 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2290 is normally done by Configure or something similar).
2292 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2293 in the source file (foo.c) like this:
2295 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2296 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2298 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2299 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2301 OPENSSL_DECLARE_GLOBAL(int,foo);
2302 #define foo OPENSSL_GLOBAL_REF(foo)
2303 OPENSSL_DECLARE_GLOBAL(double,bar);
2304 #define bar OPENSSL_GLOBAL_REF(bar)
2306 The #defines are very important, and therefore so is including the
2307 header file everywhere where the defined globals are used.
2309 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2310 of ASN.1 items, but that structure is a bit different.
2312 The largest change is in util/mkdef.pl which has been enhanced with
2313 better and easier to understand logic to choose which symbols should
2314 go into the Windows .def files as well as a number of fixes and code
2315 cleanup (among others, algorithm keywords are now sorted
2316 lexicographically to avoid constant rewrites).
2319 *) In BN_div() keep a copy of the sign of 'num' before writing the
2320 result to 'rm' because if rm==num the value will be overwritten
2321 and produce the wrong result if 'num' is negative: this caused
2322 problems with BN_mod() and BN_nnmod().
2325 *) Function OCSP_request_verify(). This checks the signature on an
2326 OCSP request and verifies the signer certificate. The signer
2327 certificate is just checked for a generic purpose and OCSP request
2331 *) Add OCSP_check_validity() function to check the validity of OCSP
2332 responses. OCSP responses are prepared in real time and may only
2333 be a few seconds old. Simply checking that the current time lies
2334 between thisUpdate and nextUpdate max reject otherwise valid responses
2335 caused by either OCSP responder or client clock inaccuracy. Instead
2336 we allow thisUpdate and nextUpdate to fall within a certain period of
2337 the current time. The age of the response can also optionally be
2338 checked. Two new options -validity_period and -status_age added to
2342 *) If signature or public key algorithm is unrecognized print out its
2343 OID rather that just UNKNOWN.
2346 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2347 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2348 ID to be generated from the issuer certificate alone which can then be
2349 passed to OCSP_id_issuer_cmp().
2352 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2353 ASN1 modules to export functions returning ASN1_ITEM pointers
2354 instead of the ASN1_ITEM structures themselves. This adds several
2355 new macros which allow the underlying ASN1 function/structure to
2356 be accessed transparently. As a result code should not use ASN1_ITEM
2357 references directly (such as &X509_it) but instead use the relevant
2358 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2359 use of the new ASN1 code on platforms where exporting structures
2360 is problematical (for example in shared libraries) but exporting
2361 functions returning pointers to structures is not.
2364 *) Add support for overriding the generation of SSL/TLS session IDs.
2365 These callbacks can be registered either in an SSL_CTX or per SSL.
2366 The purpose of this is to allow applications to control, if they wish,
2367 the arbitrary values chosen for use as session IDs, particularly as it
2368 can be useful for session caching in multiple-server environments. A
2369 command-line switch for testing this (and any client code that wishes
2370 to use such a feature) has been added to "s_server".
2371 [Geoff Thorpe, Lutz Jaenicke]
2373 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2374 of the form '#if defined(...) || defined(...) || ...' and
2375 '#if !defined(...) && !defined(...) && ...'. This also avoids
2376 the growing number of special cases it was previously handling.
2379 *) Make all configuration macros available for application by making
2380 sure they are available in opensslconf.h, by giving them names starting
2381 with "OPENSSL_" to avoid conflicts with other packages and by making
2382 sure e_os2.h will cover all platform-specific cases together with
2384 Additionally, it is now possible to define configuration/platform-
2385 specific names (called "system identities"). In the C code, these
2386 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2387 macro with the name beginning with "OPENSSL_SYS_", which is determined
2388 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2392 *) New option -set_serial to 'req' and 'x509' this allows the serial
2393 number to use to be specified on the command line. Previously self
2394 signed certificates were hard coded with serial number 0 and the
2395 CA options of 'x509' had to use a serial number in a file which was
2399 *) New options to 'ca' utility to support V2 CRL entry extensions.
2400 Currently CRL reason, invalidity date and hold instruction are
2401 supported. Add new CRL extensions to V3 code and some new objects.
2404 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2405 disable standard block padding (aka PKCS#5 padding) in the EVP
2406 API, which was previously mandatory. This means that the data is
2407 not padded in any way and so the total length much be a multiple
2408 of the block size, otherwise an error occurs.
2411 *) Initial (incomplete) OCSP SSL support.
2414 *) New function OCSP_parse_url(). This splits up a URL into its host,
2415 port and path components: primarily to parse OCSP URLs. New -url
2416 option to ocsp utility.
2419 *) New nonce behavior. The return value of OCSP_check_nonce() now
2420 reflects the various checks performed. Applications can decide
2421 whether to tolerate certain situations such as an absent nonce
2422 in a response when one was present in a request: the ocsp application
2423 just prints out a warning. New function OCSP_add1_basic_nonce()
2424 this is to allow responders to include a nonce in a response even if
2425 the request is nonce-less.
2428 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2429 skipped when using openssl x509 multiple times on a single input file,
2430 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2433 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2434 set string type: to handle setting ASN1_TIME structures. Fix ca
2435 utility to correctly initialize revocation date of CRLs.
2438 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2439 the clients preferred ciphersuites and rather use its own preferences.
2440 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2441 Internet Explorer by ensuring unchanged hash method during stepup.
2442 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2445 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2446 to aes and add a new 'exist' option to print out symbols that don't
2450 *) Additional options to ocsp utility to allow flags to be set and
2451 additional certificates supplied.
2454 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2455 OCSP client a number of certificate to only verify the response
2459 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2460 handle the new API. Currently only ECB, CBC modes supported. Add new
2463 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2464 Encryption Standard (AES) Ciphersuites for Transport Layer
2465 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2466 not enabled by default and were not part of the "ALL" ciphersuite
2467 alias because they were not yet official; they could be
2468 explicitly requested by specifying the "AESdraft" ciphersuite
2469 group alias. In the final release of OpenSSL 0.9.7, the group
2470 alias is called "AES" and is part of "ALL".)
2471 [Ben Laurie, Steve Henson, Bodo Moeller]
2473 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2474 request to response.
2477 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2478 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2479 extract information from a certificate request. OCSP_response_create()
2480 creates a response and optionally adds a basic response structure.
2481 OCSP_basic_add1_status() adds a complete single response to a basic
2482 response and returns the OCSP_SINGLERESP structure just added (to allow
2483 extensions to be included for example). OCSP_basic_add1_cert() adds a
2484 certificate to a basic response and OCSP_basic_sign() signs a basic
2485 response with various flags. New helper functions ASN1_TIME_check()
2486 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2487 (converts ASN1_TIME to GeneralizedTime).
2490 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2491 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2492 structure from a certificate. X509_pubkey_digest() digests the public_key
2493 contents: this is used in various key identifiers.
2496 *) Make sk_sort() tolerate a NULL argument.
2497 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2499 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2500 passed by the function are trusted implicitly. If any of them signed the
2501 response then it is assumed to be valid and is not verified.
2504 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2505 to data. This was previously part of the PKCS7 ASN1 code. This
2506 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2507 [Steve Henson, reported by Kenneth R. Robinette
2508 <support@securenetterm.com>]
2510 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2511 routines: without these tracing memory leaks is very painful.
2512 Fix leaks in PKCS12 and PKCS7 routines.
2515 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2516 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2517 effectively meant GeneralizedTime would never be used. Now it
2518 is initialised to -1 but X509_time_adj() now has to check the value
2519 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2520 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2521 [Steve Henson, reported by Kenneth R. Robinette
2522 <support@securenetterm.com>]
2524 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2525 result in a zero length in the ASN1_INTEGER structure which was
2526 not consistent with the structure when d2i_ASN1_INTEGER() was used
2527 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2528 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2529 where it did not print out a minus for negative ASN1_INTEGER.
2532 *) Add summary printout to ocsp utility. The various functions which
2533 convert status values to strings have been renamed to:
2534 OCSP_response_status_str(), OCSP_cert_status_str() and
2535 OCSP_crl_reason_str() and are no longer static. New options
2536 to verify nonce values and to disable verification. OCSP response
2537 printout format cleaned up.
2540 *) Add additional OCSP certificate checks. These are those specified
2541 in RFC2560. This consists of two separate checks: the CA of the
2542 certificate being checked must either be the OCSP signer certificate
2543 or the issuer of the OCSP signer certificate. In the latter case the
2544 OCSP signer certificate must contain the OCSP signing extended key
2545 usage. This check is performed by attempting to match the OCSP
2546 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2547 in the OCSP_CERTID structures of the response.
2550 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2551 and related routines. This uses the standard OpenSSL certificate
2552 verify routines to perform initial checks (just CA validity) and
2553 to obtain the certificate chain. Then additional checks will be
2554 performed on the chain. Currently the root CA is checked to see
2555 if it is explicitly trusted for OCSP signing. This is used to set
2556 a root CA as a global signing root: that is any certificate that
2557 chains to that CA is an acceptable OCSP signing certificate.
2560 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2561 extensions from a separate configuration file.
2562 As when reading extensions from the main configuration file,
2563 the '-extensions ...' option may be used for specifying the
2565 [Massimiliano Pala <madwolf@comune.modena.it>]
2567 *) New OCSP utility. Allows OCSP requests to be generated or
2568 read. The request can be sent to a responder and the output
2569 parsed, outputed or printed in text form. Not complete yet:
2570 still needs to check the OCSP response validity.
2573 *) New subcommands for 'openssl ca':
2574 'openssl ca -status <serial>' prints the status of the cert with
2575 the given serial number (according to the index file).
2576 'openssl ca -updatedb' updates the expiry status of certificates
2578 [Massimiliano Pala <madwolf@comune.modena.it>]
2580 *) New '-newreq-nodes' command option to CA.pl. This is like
2581 '-newreq', but calls 'openssl req' with the '-nodes' option
2582 so that the resulting key is not encrypted.
2583 [Damien Miller <djm@mindrot.org>]
2585 *) New configuration for the GNU Hurd.
2586 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2588 *) Initial code to implement OCSP basic response verify. This
2589 is currently incomplete. Currently just finds the signer's
2590 certificate and verifies the signature on the response.
2593 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2594 value of OPENSSLDIR. This is available via the new '-d' option
2595 to 'openssl version', and is also included in 'openssl version -a'.
2598 *) Allowing defining memory allocation callbacks that will be given
2599 file name and line number information in additional arguments
2600 (a const char* and an int). The basic functionality remains, as
2601 well as the original possibility to just replace malloc(),
2602 realloc() and free() by functions that do not know about these
2603 additional arguments. To register and find out the current
2604 settings for extended allocation functions, the following
2605 functions are provided:
2607 CRYPTO_set_mem_ex_functions
2608 CRYPTO_set_locked_mem_ex_functions
2609 CRYPTO_get_mem_ex_functions
2610 CRYPTO_get_locked_mem_ex_functions
2612 These work the same way as CRYPTO_set_mem_functions and friends.
2613 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2614 extended allocation function is enabled.
2615 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2616 a conventional allocation function is enabled.
2617 [Richard Levitte, Bodo Moeller]
2619 *) Finish off removing the remaining LHASH function pointer casts.
2620 There should no longer be any prototype-casting required when using
2621 the LHASH abstraction, and any casts that remain are "bugs". See
2622 the callback types and macros at the head of lhash.h for details
2623 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2626 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2627 If /dev/[u]random devices are not available or do not return enough
2628 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2630 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2631 /etc/entropy will be queried once each in this sequence, quering stops
2632 when enough entropy was collected without querying more sockets.
2635 *) Change the Unix RAND_poll() variant to be able to poll several
2636 random devices, as specified by DEVRANDOM, until a sufficient amount
2637 of data has been collected. We spend at most 10 ms on each file
2638 (select timeout) and read in non-blocking mode. DEVRANDOM now
2639 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2640 (previously it was just the string "/dev/urandom"), so on typical
2641 platforms the 10 ms delay will never occur.
2642 Also separate out the Unix variant to its own file, rand_unix.c.
2643 For VMS, there's a currently-empty rand_vms.c.
2646 *) Move OCSP client related routines to ocsp_cl.c. These
2647 provide utility functions which an application needing
2648 to issue a request to an OCSP responder and analyse the
2649 response will typically need: as opposed to those which an
2650 OCSP responder itself would need which will be added later.
2652 OCSP_request_sign() signs an OCSP request with an API similar
2653 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2654 response. OCSP_response_get1_basic() extracts basic response
2655 from response. OCSP_resp_find_status(): finds and extracts status
2656 information from an OCSP_CERTID structure (which will be created
2657 when the request structure is built). These are built from lower
2658 level functions which work on OCSP_SINGLERESP structures but
2659 wont normally be used unless the application wishes to examine
2660 extensions in the OCSP response for example.
2662 Replace nonce routines with a pair of functions.
2663 OCSP_request_add1_nonce() adds a nonce value and optionally
2664 generates a random value. OCSP_check_nonce() checks the
2665 validity of the nonce in an OCSP response.
2668 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2669 This doesn't copy the supplied OCSP_CERTID and avoids the
2670 need to free up the newly created id. Change return type
2671 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2672 This can then be used to add extensions to the request.
2673 Deleted OCSP_request_new(), since most of its functionality
2674 is now in OCSP_REQUEST_new() (and the case insensitive name
2675 clash) apart from the ability to set the request name which
2676 will be added elsewhere.
2679 *) Update OCSP API. Remove obsolete extensions argument from
2680 various functions. Extensions are now handled using the new
2681 OCSP extension code. New simple OCSP HTTP function which
2682 can be used to send requests and parse the response.
2685 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2686 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2687 uses the special reorder version of SET OF to sort the attributes
2688 and reorder them to match the encoded order. This resolves a long
2689 standing problem: a verify on a PKCS7 structure just after signing
2690 it used to fail because the attribute order did not match the
2691 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2692 it uses the received order. This is necessary to tolerate some broken
2693 software that does not order SET OF. This is handled by encoding
2694 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2695 to produce the required SET OF.
2698 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2699 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2700 files to get correct declarations of the ASN.1 item variables.
2703 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2704 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2705 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2706 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2707 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2708 ASN1_ITEM and no wrapper functions.
2711 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2712 replace the old function pointer based I/O routines. Change most of
2713 the *_d2i_bio() and *_d2i_fp() functions to use these.
2716 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2717 lines, recognice more "algorithms" that can be deselected, and make
2718 it complain about algorithm deselection that isn't recognised.
2721 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2722 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2723 to use new functions. Add NO_ASN1_OLD which can be set to remove
2724 some old style ASN1 functions: this can be used to determine if old
2725 code will still work when these eventually go away.
2728 *) New extension functions for OCSP structures, these follow the
2729 same conventions as certificates and CRLs.
2732 *) New function X509V3_add1_i2d(). This automatically encodes and
2733 adds an extension. Its behaviour can be customised with various
2734 flags to append, replace or delete. Various wrappers added for
2735 certifcates and CRLs.
2738 *) Fix to avoid calling the underlying ASN1 print routine when
2739 an extension cannot be parsed. Correct a typo in the
2740 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2743 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2744 entries for variables.
2747 *) Add functionality to apps/openssl.c for detecting locking
2748 problems: As the program is single-threaded, all we have
2749 to do is register a locking callback using an array for
2750 storing which locks are currently held by the program.
2753 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2754 SSL_get_ex_data_X509_STORE_idx(), which is used in
2755 ssl_verify_cert_chain() and thus can be called at any time
2756 during TLS/SSL handshakes so that thread-safety is essential.
2757 Unfortunately, the ex_data design is not at all suited
2758 for multi-threaded use, so it probably should be abolished.
2761 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2762 [Broadcom, tweaked and integrated by Geoff Thorpe]
2764 *) Move common extension printing code to new function
2765 X509V3_print_extensions(). Reorganise OCSP print routines and
2766 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2769 *) New function X509_signature_print() to remove duplication in some
2773 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2774 set (this was treated exactly the same as SET OF previously). This
2775 is used to reorder the STACK representing the structure to match the
2776 encoding. This will be used to get round a problem where a PKCS7
2777 structure which was signed could not be verified because the STACK
2778 order did not reflect the encoded order.
2781 *) Reimplement the OCSP ASN1 module using the new code.
2784 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2785 for its ASN1 operations. The old style function pointers still exist
2786 for now but they will eventually go away.
2789 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2790 completely replaces the old ASN1 functionality with a table driven
2791 encoder and decoder which interprets an ASN1_ITEM structure describing
2792 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2793 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2794 has also been converted to the new form.
2797 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2798 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2799 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2800 for negative moduli.
2803 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2804 of not touching the result's sign bit.
2807 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2811 *) Changed the LHASH code to use prototypes for callbacks, and created
2812 macros to declare and implement thin (optionally static) functions
2813 that provide type-safety and avoid function pointer casting for the
2814 type-specific callbacks.
2817 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2819 [Veers Staats <staatsvr@asc.hpc.mil>,
2820 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2822 *) Reformat the FAQ so the different questions and answers can be divided
2823 in sections depending on the subject.
2826 *) Have the zlib compression code load ZLIB.DLL dynamically under
2830 *) New function BN_mod_sqrt for computing square roots modulo a prime
2831 (using the probabilistic Tonelli-Shanks algorithm unless
2832 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2833 be handled deterministically).
2834 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2836 *) Make BN_mod_inverse faster by explicitly handling small quotients
2837 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2838 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2841 *) New function BN_kronecker.
2844 *) Fix BN_gcd so that it works on negative inputs; the result is
2845 positive unless both parameters are zero.
2846 Previously something reasonably close to an infinite loop was
2847 possible because numbers could be growing instead of shrinking
2848 in the implementation of Euclid's algorithm.
2851 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2852 sign of the number in question.
2854 Fix BN_is_word(a,w) to work correctly for w == 0.
2856 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2857 because its test if the absolute value of 'a' equals 'w'.
2858 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2859 it exists mostly for use in the implementations of BN_is_zero(),
2860 BN_is_one(), and BN_is_word().
2863 *) New function BN_swap.
2866 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2867 the exponentiation functions are more likely to produce reasonable
2868 results on negative inputs.
2871 *) Change BN_mod_mul so that the result is always non-negative.
2872 Previously, it could be negative if one of the factors was negative;
2873 I don't think anyone really wanted that behaviour.
2876 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2877 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2878 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2879 and add new functions:
2888 BN_mod_lshift1_quick
2892 These functions always generate non-negative results.
2894 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2895 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2897 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2898 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2899 be reduced modulo m.
2900 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2903 The following entry accidentily appeared in the CHANGES file
2904 distributed with OpenSSL 0.9.7. The modifications described in
2905 it do *not* apply to OpenSSL 0.9.7.
2907 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2908 was actually never needed) and in BN_mul(). The removal in BN_mul()
2909 required a small change in bn_mul_part_recursive() and the addition
2910 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2911 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2912 bn_sub_words() and bn_add_words() except they take arrays with
2917 *) In 'openssl passwd', verify passwords read from the terminal
2918 unless the '-salt' option is used (which usually means that
2919 verification would just waste user's time since the resulting
2920 hash is going to be compared with some given password hash)
2921 or the new '-noverify' option is used.
2923 This is an incompatible change, but it does not affect
2924 non-interactive use of 'openssl passwd' (passwords on the command
2925 line, '-stdin' option, '-in ...' option) and thus should not
2929 *) Remove all references to RSAref, since there's no more need for it.
2932 *) Make DSO load along a path given through an environment variable
2933 (SHLIB_PATH) with shl_load().
2936 *) Constify the ENGINE code as a result of BIGNUM constification.
2937 Also constify the RSA code and most things related to it. In a
2938 few places, most notable in the depth of the ASN.1 code, ugly
2939 casts back to non-const were required (to be solved at a later
2943 *) Make it so the openssl application has all engines loaded by default.
2946 *) Constify the BIGNUM routines a little more.
2949 *) Add the following functions:
2951 ENGINE_load_cswift()
2953 ENGINE_load_atalla()
2955 ENGINE_load_builtin_engines()
2957 That way, an application can itself choose if external engines that
2958 are built-in in OpenSSL shall ever be used or not. The benefit is
2959 that applications won't have to be linked with libdl or other dso
2960 libraries unless it's really needed.
2962 Changed 'openssl engine' to load all engines on demand.
2963 Changed the engine header files to avoid the duplication of some
2964 declarations (they differed!).
2967 *) 'openssl engine' can now list capabilities.
2970 *) Better error reporting in 'openssl engine'.
2973 *) Never call load_dh_param(NULL) in s_server.
2976 *) Add engine application. It can currently list engines by name and
2977 identity, and test if they are actually available.
2980 *) Improve RPM specification file by forcing symbolic linking and making
2981 sure the installed documentation is also owned by root.root.
2982 [Damien Miller <djm@mindrot.org>]
2984 *) Give the OpenSSL applications more possibilities to make use of
2985 keys (public as well as private) handled by engines.
2988 *) Add OCSP code that comes from CertCo.
2991 *) Add VMS support for the Rijndael code.
2994 *) Added untested support for Nuron crypto accelerator.
2997 *) Add support for external cryptographic devices. This code was
2998 previously distributed separately as the "engine" branch.
2999 [Geoff Thorpe, Richard Levitte]
3001 *) Rework the filename-translation in the DSO code. It is now possible to
3002 have far greater control over how a "name" is turned into a filename
3003 depending on the operating environment and any oddities about the
3004 different shared library filenames on each system.
3007 *) Support threads on FreeBSD-elf in Configure.
3010 *) Fix for SHA1 assembly problem with MASM: it produces
3011 warnings about corrupt line number information when assembling
3012 with debugging information. This is caused by the overlapping
3014 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3017 NCONF_get_number() has no error checking at all. As a replacement,
3018 NCONF_get_number_e() is defined (_e for "error checking") and is
3019 promoted strongly. The old NCONF_get_number is kept around for
3020 binary backward compatibility.
3021 Make it possible for methods to load from something other than a BIO,
3022 by providing a function pointer that is given a name instead of a BIO.
3023 For example, this could be used to load configuration data from an
3027 *) Fix for non blocking accept BIOs. Added new I/O special reason
3028 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3029 with non blocking I/O was not possible because no retry code was
3030 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3034 *) Added the beginnings of Rijndael support.
3037 *) Fix for bug in DirectoryString mask setting. Add support for
3038 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3039 to allow certificate printing to more controllable, additional
3040 'certopt' option to 'x509' to allow new printing options to be
3044 *) Clean old EAY MD5 hack from e_os.h.
3047 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3049 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3050 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
3051 [Joe Orton, Steve Henson]
3053 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3055 *) Fix additional bug revealed by the NISCC test suite:
3057 Stop bug triggering large recursion when presented with
3058 certain ASN.1 tags (CAN-2003-0851)
3061 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3063 *) Fix various bugs revealed by running the NISCC test suite:
3065 Stop out of bounds reads in the ASN1 code when presented with
3066 invalid tags (CAN-2003-0543 and CAN-2003-0544).
3068 If verify callback ignores invalid public key errors don't try to check
3069 certificate signature with the NULL public key.
3073 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3074 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3078 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3079 extra data after the compression methods not only for TLS 1.0
3080 but also for SSL 3.0 (as required by the specification).
3081 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3083 *) Change X509_certificate_type() to mark the key as exported/exportable
3084 when it's 512 *bits* long, not 512 bytes.
3087 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3089 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3090 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3091 a protocol version number mismatch like a decryption error
3092 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3095 *) Turn on RSA blinding by default in the default implementation
3096 to avoid a timing attack. Applications that don't want it can call
3097 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3098 They would be ill-advised to do so in most cases.
3099 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3101 *) Change RSA blinding code so that it works when the PRNG is not
3102 seeded (in this case, the secret RSA exponent is abused as
3103 an unpredictable seed -- if it is not unpredictable, there
3104 is no point in blinding anyway). Make RSA blinding thread-safe
3105 by remembering the creator's thread ID in rsa->blinding and
3106 having all other threads use local one-time blinding factors
3107 (this requires more computation than sharing rsa->blinding, but
3108 avoids excessive locking; and if an RSA object is not shared
3109 between threads, blinding will still be very fast).
3112 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3114 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3115 via timing by performing a MAC computation even if incorrrect
3116 block cipher padding has been found. This is a countermeasure
3117 against active attacks where the attacker has to distinguish
3118 between bad padding and a MAC verification error. (CAN-2003-0078)
3120 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3121 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3122 Martin Vuagnoux (EPFL, Ilion)]
3124 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3126 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3127 memory from it's contents. This is done with a counter that will
3128 place alternating values in each byte. This can be used to solve
3129 two issues: 1) the removal of calls to memset() by highly optimizing
3130 compilers, and 2) cleansing with other values than 0, since those can
3131 be read through on certain media, for example a swap space on disk.
3134 *) Bugfix: client side session caching did not work with external caching,
3135 because the session->cipher setting was not restored when reloading
3136 from the external cache. This problem was masked, when
3137 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3138 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3141 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3142 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3143 [Zeev Lieber <zeev-l@yahoo.com>]
3145 *) Undo an undocumented change introduced in 0.9.6e which caused
3146 repeated calls to OpenSSL_add_all_ciphers() and
3147 OpenSSL_add_all_digests() to be ignored, even after calling
3151 *) Change the default configuration reader to deal with last line not
3152 being properly terminated.
3155 *) Change X509_NAME_cmp() so it applies the special rules on handling
3156 DN values that are of type PrintableString, as well as RDNs of type
3157 emailAddress where the value has the type ia5String.
3158 [stefank@valicert.com via Richard Levitte]
3160 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3161 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3162 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3163 the bitwise-OR of the two for use by the majority of applications
3164 wanting this behaviour, and update the docs. The documented
3165 behaviour and actual behaviour were inconsistent and had been
3166 changing anyway, so this is more a bug-fix than a behavioural
3168 [Geoff Thorpe, diagnosed by Nadav Har'El]
3170 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3171 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3174 *) Fix initialization code race conditions in
3175 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3176 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3177 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3178 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3179 ssl2_get_cipher_by_char(),
3180 ssl3_get_cipher_by_char().
3181 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3183 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3184 the cached sessions are flushed, as the remove_cb() might use ex_data
3185 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3186 (see [openssl.org #212]).
3187 [Geoff Thorpe, Lutz Jaenicke]
3189 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3190 length, instead of the encoding length to d2i_ASN1_OBJECT.
3193 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3195 *) [In 0.9.6g-engine release:]
3196 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3197 [Lynn Gazis <lgazis@rainbow.com>]
3199 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3201 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3202 and get fix the header length calculation.
3203 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3204 Alon Kantor <alonk@checkpoint.com> (and others),
3207 *) Use proper error handling instead of 'assertions' in buffer
3208 overflow checks added in 0.9.6e. This prevents DoS (the
3209 assertions could call abort()).
3210 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3212 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3214 *) Add various sanity checks to asn1_get_length() to reject
3215 the ASN1 length bytes if they exceed sizeof(long), will appear
3216 negative or the content length exceeds the length of the
3218 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3220 *) Fix cipher selection routines: ciphers without encryption had no flags
3221 for the cipher strength set and where therefore not handled correctly
3222 by the selection routines (PR #130).
3225 *) Fix EVP_dsa_sha macro.
3229 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3230 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3231 that was added in OpenSSL 0.9.6d.
3233 As the countermeasure turned out to be incompatible with some
3234 broken SSL implementations, the new option is part of SSL_OP_ALL.
3235 SSL_OP_ALL is usually employed when compatibility with weird SSL
3236 implementations is desired (e.g. '-bugs' option to 's_client' and
3237 's_server'), so the new option is automatically set in many
3241 *) Changes in security patch:
3243 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3244 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3245 Air Force Materiel Command, USAF, under agreement number
3248 *) Add various sanity checks to asn1_get_length() to reject
3249 the ASN1 length bytes if they exceed sizeof(long), will appear
3250 negative or the content length exceeds the length of the
3251 supplied buffer. (CAN-2002-0659)
3252 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3254 *) Assertions for various potential buffer overflows, not known to
3256 [Ben Laurie (CHATS)]
3258 *) Various temporary buffers to hold ASCII versions of integers were
3259 too small for 64 bit platforms. (CAN-2002-0655)
3260 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3262 *) Remote buffer overflow in SSL3 protocol - an attacker could
3263 supply an oversized session ID to a client. (CAN-2002-0656)
3264 [Ben Laurie (CHATS)]
3266 *) Remote buffer overflow in SSL2 protocol - an attacker could
3267 supply an oversized client master key. (CAN-2002-0656)
3268 [Ben Laurie (CHATS)]
3270 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3272 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3273 encoded as NULL) with id-dsa-with-sha1.
3274 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3276 *) Check various X509_...() return values in apps/req.c.
3277 [Nils Larsch <nla@trustcenter.de>]
3279 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3280 an end-of-file condition would erronously be flagged, when the CRLF
3281 was just at the end of a processed block. The bug was discovered when
3282 processing data through a buffering memory BIO handing the data to a
3283 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3284 <ptsekov@syntrex.com> and Nedelcho Stanev.
3287 *) Implement a countermeasure against a vulnerability recently found
3288 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3289 before application data chunks to avoid the use of known IVs
3290 with data potentially chosen by the attacker.
3293 *) Fix length checks in ssl3_get_client_hello().
3296 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3297 to prevent ssl3_read_internal() from incorrectly assuming that
3298 ssl3_read_bytes() found application data while handshake
3299 processing was enabled when in fact s->s3->in_read_app_data was
3300 merely automatically cleared during the initial handshake.
3301 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3303 *) Fix object definitions for Private and Enterprise: they were not
3304 recognized in their shortname (=lowercase) representation. Extend
3305 obj_dat.pl to issue an error when using undefined keywords instead
3306 of silently ignoring the problem (Svenning Sorensen
3307 <sss@sss.dnsalias.net>).
3310 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3311 generators, i.e. generators other than 2 and 5. (Previously, the
3312 code did not properly initialise the 'add' and 'rem' values to
3313 BN_generate_prime().)
3315 In the new general case, we do not insist that 'generator' is
3316 actually a primitive root: This requirement is rather pointless;
3317 a generator of the order-q subgroup is just as good, if not
3321 *) Map new X509 verification errors to alerts. Discovered and submitted by
3322 Tom Wu <tom@arcot.com>.
3325 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3326 returning non-zero before the data has been completely received
3327 when using non-blocking I/O.
3328 [Bodo Moeller; problem pointed out by John Hughes]
3330 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3331 [Ben Laurie, Lutz Jaenicke]
3333 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3334 Yoram Zahavi <YoramZ@gilian.com>).
3337 *) Add information about CygWin 1.3 and on, and preserve proper
3338 configuration for the versions before that.
3339 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3341 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3342 check whether we deal with a copy of a session and do not delete from
3343 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3344 <izhar@checkpoint.com>.
3347 *) Do not store session data into the internal session cache, if it
3348 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3349 flag is set). Proposed by Aslam <aslam@funk.com>.
3352 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3356 *) [In 0.9.6d-engine release:]
3357 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3358 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3360 *) Add the configuration target linux-s390x.
3361 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3363 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3364 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3365 variable as an indication that a ClientHello message has been
3366 received. As the flag value will be lost between multiple
3367 invocations of ssl3_accept when using non-blocking I/O, the
3368 function may not be aware that a handshake has actually taken
3369 place, thus preventing a new session from being added to the
3372 To avoid this problem, we now set s->new_session to 2 instead of
3373 using a local variable.
3374 [Lutz Jaenicke, Bodo Moeller]
3376 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3377 if the SSL_R_LENGTH_MISMATCH error is detected.
3378 [Geoff Thorpe, Bodo Moeller]
3380 *) New 'shared_ldflag' column in Configure platform table.
3383 *) Fix EVP_CIPHER_mode macro.
3384 ["Dan S. Camper" <dan@bti.net>]
3386 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3387 type, we must throw them away by setting rr->length to 0.
3388 [D P Chang <dpc@qualys.com>]
3390 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3392 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3393 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3394 worked incorrectly for those cases where range = 10..._2 and
3395 3*range is two bits longer than range.)
3398 *) Only add signing time to PKCS7 structures if it is not already
3402 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3403 OBJ_ld_ce should be OBJ_id_ce.
3404 Also some ip-pda OIDs in crypto/objects/objects.txt were
3405 incorrect (cf. RFC 3039).
3406 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3408 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3409 returns early because it has nothing to do.
3410 [Andy Schneider <andy.schneider@bjss.co.uk>]
3412 *) [In 0.9.6c-engine release:]
3413 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3414 [Andy Schneider <andy.schneider@bjss.co.uk>]
3416 *) [In 0.9.6c-engine release:]
3417 Add support for Cryptographic Appliance's keyserver technology.
3418 (Use engine 'keyclient')
3419 [Cryptographic Appliances and Geoff Thorpe]
3421 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3422 is called via tools/c89.sh because arguments have to be
3423 rearranged (all '-L' options must appear before the first object
3425 [Richard Shapiro <rshapiro@abinitio.com>]
3427 *) [In 0.9.6c-engine release:]
3428 Add support for Broadcom crypto accelerator cards, backported
3430 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3432 *) [In 0.9.6c-engine release:]
3433 Add support for SureWare crypto accelerator cards from
3434 Baltimore Technologies. (Use engine 'sureware')
3435 [Baltimore Technologies and Mark Cox]
3437 *) [In 0.9.6c-engine release:]
3438 Add support for crypto accelerator cards from Accelerated
3439 Encryption Processing, www.aep.ie. (Use engine 'aep')
3440 [AEP Inc. and Mark Cox]
3442 *) Add a configuration entry for gcc on UnixWare.
3443 [Gary Benson <gbenson@redhat.com>]
3445 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3446 messages are stored in a single piece (fixed-length part and
3447 variable-length part combined) and fix various bugs found on the way.
3450 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3451 instead. BIO_gethostbyname() does not know what timeouts are
3452 appropriate, so entries would stay in cache even when they have
3454 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3456 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3457 faced with a pathologically small ClientHello fragment that does
3458 not contain client_version: Instead of aborting with an error,
3459 simply choose the highest available protocol version (i.e.,
3460 TLS 1.0 unless it is disabled). In practice, ClientHello
3461 messages are never sent like this, but this change gives us
3462 strictly correct behaviour at least for TLS.
3465 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3466 never resets s->method to s->ctx->method when called from within
3467 one of the SSL handshake functions.
3468 [Bodo Moeller; problem pointed out by Niko Baric]
3470 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3471 (sent using the client's version number) if client_version is
3472 smaller than the protocol version in use. Also change
3473 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3474 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3475 the client will at least see that alert.
3478 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3482 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3483 client receives HelloRequest while in a handshake.
3484 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3486 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3487 should end in 'break', not 'goto end' which circuments various
3488 cleanups done in state SSL_ST_OK. But session related stuff
3489 must be disabled for SSL_ST_OK in the case that we just sent a
3492 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3493 before just sending a HelloRequest.
3494 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3496 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3497 reveal whether illegal block cipher padding was found or a MAC
3498 verification error occured. (Neither SSLerr() codes nor alerts
3499 are directly visible to potential attackers, but the information
3500 may leak via logfiles.)
3502 Similar changes are not required for the SSL 2.0 implementation
3503 because the number of padding bytes is sent in clear for SSL 2.0,
3504 and the extra bytes are just ignored. However ssl/s2_pkt.c
3505 failed to verify that the purported number of padding bytes is in
3509 *) Add OpenUNIX-8 support including shared libraries
3510 (Boyd Lynn Gerber <gerberb@zenez.com>).
3513 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3514 'wristwatch attack' using huge encoding parameters (cf.
3515 James H. Manger's CRYPTO 2001 paper). Note that the
3516 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3517 encoding parameters and hence was not vulnerable.
3520 *) BN_sqr() bug fix.
3521 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3523 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3524 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3525 followed by modular reduction.
3526 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3528 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3529 equivalent based on BN_pseudo_rand() instead of BN_rand().
3532 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3533 This function was broken, as the check for a new client hello message
3534 to handle SGC did not allow these large messages.
3535 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3538 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3541 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3542 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3545 *) Rework the configuration and shared library support for Tru64 Unix.
3546 The configuration part makes use of modern compiler features and
3547 still retains old compiler behavior for those that run older versions
3548 of the OS. The shared library support part includes a variant that
3549 uses the RPATH feature, and is available through the special
3550 configuration target "alpha-cc-rpath", which will never be selected
3552 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3554 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3555 with the same message size as in ssl3_get_certificate_request().
3556 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3557 messages might inadvertently be reject as too long.
3558 [Petr Lampa <lampa@fee.vutbr.cz>]
3560 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3563 *) Modified SSL library such that the verify_callback that has been set
3564 specificly for an SSL object with SSL_set_verify() is actually being
3565 used. Before the change, a verify_callback set with this function was
3566 ignored and the verify_callback() set in the SSL_CTX at the time of
3567 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3568 to allow the necessary settings.
3571 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3572 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3573 done automatically (in contradiction to the requirements of the C
3574 standard). This made problems when used from OpenSSH.
3577 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3578 dh->length and always used
3580 BN_rand_range(priv_key, dh->p).
3582 BN_rand_range() is not necessary for Diffie-Hellman, and this
3583 specific range makes Diffie-Hellman unnecessarily inefficient if
3584 dh->length (recommended exponent length) is much smaller than the
3585 length of dh->p. We could use BN_rand_range() if the order of
3586 the subgroup was stored in the DH structure, but we only have
3591 BN_rand(priv_key, l, ...)
3593 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3599 RSA_eay_public_encrypt
3600 RSA_eay_private_decrypt
3601 RSA_eay_private_encrypt (signing)
3602 RSA_eay_public_decrypt (signature verification)
3604 (default implementations for RSA_public_encrypt,
3605 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3606 always reject numbers >= n.
3609 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3610 to synchronize access to 'locking_thread'. This is necessary on
3611 systems where access to 'locking_thread' (an 'unsigned long'
3612 variable) is not atomic.
3615 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3616 *before* setting the 'crypto_lock_rand' flag. The previous code had
3617 a race condition if 0 is a valid thread ID.
3618 [Travis Vitek <vitek@roguewave.com>]
3620 *) Add support for shared libraries under Irix.
3621 [Albert Chin-A-Young <china@thewrittenword.com>]
3623 *) Add configuration option to build on Linux on both big-endian and
3625 [Ralf Baechle <ralf@uni-koblenz.de>]
3627 *) Add the possibility to create shared libraries on HP-UX.
3630 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3632 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3633 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3634 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3635 PRNG state recovery was possible based on the output of
3636 one PRNG request appropriately sized to gain knowledge on
3637 'md' followed by enough consecutive 1-byte PRNG requests
3638 to traverse all of 'state'.
3640 1. When updating 'md_local' (the current thread's copy of 'md')
3641 during PRNG output generation, hash all of the previous
3642 'md_local' value, not just the half used for PRNG output.
3644 2. Make the number of bytes from 'state' included into the hash
3645 independent from the number of PRNG bytes requested.
3647 The first measure alone would be sufficient to avoid
3648 Markku-Juhani's attack. (Actually it had never occurred
3649 to me that the half of 'md_local' used for chaining was the
3650 half from which PRNG output bytes were taken -- I had always
3651 assumed that the secret half would be used.) The second
3652 measure makes sure that additional data from 'state' is never
3653 mixed into 'md_local' in small portions; this heuristically
3654 further strengthens the PRNG.
3657 *) Fix crypto/bn/asm/mips3.s.
3660 *) When only the key is given to "enc", the IV is undefined. Print out
3661 an error message in this case.
3664 *) Handle special case when X509_NAME is empty in X509 printing routines.
3667 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3668 positive and less than q.
3671 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3672 used: it isn't thread safe and the add_lock_callback should handle
3674 [Paul Rose <Paul.Rose@bridge.com>]
3676 *) Verify that incoming data obeys the block size in
3677 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3681 [Ulf Möller, Bodo Möller]
3683 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3684 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3685 when fixing the server behaviour for backwards-compatible 'client
3686 hello' messages. (Note that the attack is impractical against
3687 SSL 3.0 and TLS 1.0 anyway because length and version checking
3688 means that the probability of guessing a valid ciphertext is
3689 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3692 Before 0.9.5, the countermeasure (hide the error by generating a
3693 random 'decryption result') did not work properly because
3694 ERR_clear_error() was missing, meaning that SSL_get_error() would
3695 detect the supposedly ignored error.
3697 Both problems are now fixed.
3700 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3701 (previously it was 1024).
3704 *) Fix for compatibility mode trust settings: ignore trust settings
3705 unless some valid trust or reject settings are present.
3708 *) Fix for blowfish EVP: its a variable length cipher.
3711 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3712 parameters in DSA public key structures and return an error in the
3713 DSA routines if parameters are absent.
3716 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3717 in the current directory if neither $RANDFILE nor $HOME was set.
3718 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3719 caused some confusion to Windows users who haven't defined $HOME.
3720 Thus RAND_file_name() is changed again: e_os.h can define a
3721 DEFAULT_HOME, which will be used if $HOME is not set.
3722 For Windows, we use "C:"; on other platforms, we still require
3723 environment variables.
3725 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3726 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3727 having multiple threads call RAND_poll() concurrently.
3730 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3731 combination of a flag and a thread ID variable.
3732 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3733 flag), *other* threads can enter ssleay_add_bytes without obeying
3734 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3735 that they do not hold after the first thread unsets add_do_not_lock).
3738 *) Change bctest again: '-x' expressions are not available in all
3742 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3744 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3745 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3747 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3748 the default extension for executables, if any. Also, make the perl
3749 scripts that use symlink() to test if it really exists and use "cp"
3750 if it doesn't. All this made OpenSSL compilable and installable in
3754 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3755 If SEQUENCE is length is indefinite just set c->slen to the total
3756 amount of data available.
3757 [Steve Henson, reported by shige@FreeBSD.org]
3758 [This change does not apply to 0.9.7.]
3760 *) Change bctest to avoid here-documents inside command substitution
3761 (workaround for FreeBSD /bin/sh bug).
3762 For compatibility with Ultrix, avoid shell functions (introduced
3763 in the bctest version that searches along $PATH).
3766 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3767 with des_encrypt() defined on some operating systems, like Solaris
3771 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3772 On the Importance of Eliminating Errors in Cryptographic
3773 Computations, J. Cryptology 14 (2001) 2, 101-119,
3774 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3777 *) MIPS assembler BIGNUM division bug fix.
3780 *) Disabled incorrect Alpha assembler code.
3783 *) Fix PKCS#7 decode routines so they correctly update the length
3784 after reading an EOC for the EXPLICIT tag.
3786 [This change does not apply to 0.9.7.]
3788 *) Fix bug in PKCS#12 key generation routines. This was triggered
3789 if a 3DES key was generated with a 0 initial byte. Include
3790 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3791 (but broken) behaviour.
3794 *) Enhance bctest to search for a working bc along $PATH and print
3796 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3798 *) Fix memory leaks in err.c: free err_data string if necessary;
3799 don't write to the wrong index in ERR_set_error_data.
3802 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3806 *) Replace rdtsc with _emit statements for VC++ version 5.
3807 [Jeremy Cooper <jeremy@baymoo.org>]
3809 *) Make it possible to reuse SSLv2 sessions.
3812 *) In copy_email() check for >= 0 as a return value for
3813 X509_NAME_get_index_by_NID() since 0 is a valid index.
3814 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3816 *) Avoid coredump with unsupported or invalid public keys by checking if
3817 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3818 PKCS7_verify() fails with non detached data.
3821 *) Don't use getenv in library functions when run as setuid/setgid.
3822 New function OPENSSL_issetugid().
3825 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3826 due to incorrect handling of multi-threading:
3828 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3830 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3832 3. Count how many times MemCheck_off() has been called so that
3833 nested use can be treated correctly. This also avoids
3834 inband-signalling in the previous code (which relied on the
3835 assumption that thread ID 0 is impossible).
3838 *) Add "-rand" option also to s_client and s_server.
3841 *) Fix CPU detection on Irix 6.x.
3842 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3843 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3845 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3848 [This change does not apply to 0.9.7.]
3850 *) Use the cached encoding of an X509_NAME structure rather than
3851 copying it. This is apparently the reason for the libsafe "errors"
3852 but the code is actually correct.
3855 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3856 Bleichenbacher's DSA attack.
3857 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3858 to be set and top=0 forces the highest bit to be set; top=-1 is new
3859 and leaves the highest bit random.
3860 [Ulf Moeller, Bodo Moeller]
3862 *) In the NCONF_...-based implementations for CONF_... queries
3863 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3864 a temporary CONF structure with the data component set to NULL
3865 (which gives segmentation faults in lh_retrieve).
3866 Instead, use NULL for the CONF pointer in CONF_get_string and
3867 CONF_get_number (which may use environment variables) and directly
3868 return NULL from CONF_get_section.
3871 *) Fix potential buffer overrun for EBCDIC.
3874 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3875 keyUsage if basicConstraints absent for a CA.
3878 *) Make SMIME_write_PKCS7() write mail header values with a format that
3879 is more generally accepted (no spaces before the semicolon), since
3880 some programs can't parse those values properly otherwise. Also make
3881 sure BIO's that break lines after each write do not create invalid
3885 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3886 macros previously used would not encode an empty SEQUENCE OF
3887 and break the signature.
3889 [This change does not apply to 0.9.7.]
3891 *) Zero the premaster secret after deriving the master secret in
3895 *) Add some EVP_add_digest_alias registrations (as found in
3896 OpenSSL_add_all_digests()) to SSL_library_init()
3897 aka OpenSSL_add_ssl_algorithms(). This provides improved
3898 compatibility with peers using X.509 certificates
3899 with unconventional AlgorithmIdentifier OIDs.
3902 *) Fix for Irix with NO_ASM.
3903 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3905 *) ./config script fixes.
3906 [Ulf Moeller, Richard Levitte]
3908 *) Fix 'openssl passwd -1'.
3911 *) Change PKCS12_key_gen_asc() so it can cope with non null
3912 terminated strings whose length is passed in the passlen
3913 parameter, for example from PEM callbacks. This was done
3914 by adding an extra length parameter to asc2uni().
3915 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3917 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3918 call failed, free the DSA structure.
3921 *) Fix to uni2asc() to cope with zero length Unicode strings.
3922 These are present in some PKCS#12 files.
3925 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3926 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3927 when writing a 32767 byte record.
3928 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3930 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3931 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3933 (RSA objects have a reference count access to which is protected
3934 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3935 so they are meant to be shared between threads.)
3936 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3937 "Reddie, Steven" <Steven.Reddie@ca.com>]
3939 *) Fix a deadlock in CRYPTO_mem_leaks().
3942 *) Use better test patterns in bntest.
3945 *) rand_win.c fix for Borland C.
3948 *) BN_rshift bugfix for n == 0.
3951 *) Add a 'bctest' script that checks for some known 'bc' bugs
3952 so that 'make test' does not abort just because 'bc' is broken.
3955 *) Store verify_result within SSL_SESSION also for client side to
3956 avoid potential security hole. (Re-used sessions on the client side
3957 always resulted in verify_result==X509_V_OK, not using the original
3958 result of the server certificate verification.)
3961 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3962 SSL3_RT_APPLICATION_DATA, return 0.
3963 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3967 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3968 releases, have been re-implemented by renaming the previous
3969 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3970 and ssl3_read_internal, respectively, and adding 'peek' parameters
3971 to them. The new ssl[23]_{read,peek} functions are calls to
3972 ssl[23]_read_internal with the 'peek' flag set appropriately.
3973 A 'peek' parameter has also been added to ssl3_read_bytes, which
3974 does the actual work for ssl3_read_internal.
3977 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3978 the method-specific "init()" handler. Also clean up ex_data after
3979 calling the method-specific "finish()" handler. Previously, this was
3980 happening the other way round.
3983 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3984 The previous value, 12, was not always sufficient for BN_mod_exp().
3987 *) Make sure that shared libraries get the internal name engine with
3988 the full version number and not just 0. This should mark the
3989 shared libraries as not backward compatible. Of course, this should
3990 be changed again when we can guarantee backward binary compatibility.
3993 *) Fix typo in get_cert_by_subject() in by_dir.c
3994 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3996 *) Rework the system to generate shared libraries:
3998 - Make note of the expected extension for the shared libraries and
3999 if there is a need for symbolic links from for example libcrypto.so.0
4000 to libcrypto.so.0.9.7. There is extended info in Configure for
4003 - Make as few rebuilds of the shared libraries as possible.
4005 - Still avoid linking the OpenSSL programs with the shared libraries.
4007 - When installing, install the shared libraries separately from the
4011 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4013 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4014 and not in SSL_clear because the latter is also used by the
4015 accept/connect functions; previously, the settings made by
4016 SSL_set_read_ahead would be lost during the handshake.
4017 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4019 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4020 Previously, it would create entries for disableed algorithms no
4024 *) Added several new manual pages for SSL_* function.
4027 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4029 *) In ssl23_get_client_hello, generate an error message when faced
4030 with an initial SSL 3.0/TLS record that is too small to contain the
4031 first two bytes of the ClientHello message, i.e. client_version.
4032 (Note that this is a pathologic case that probably has never happened
4033 in real life.) The previous approach was to use the version number
4034 from the record header as a substitute; but our protocol choice
4035 should not depend on that one because it is not authenticated
4036 by the Finished messages.
4039 *) More robust randomness gathering functions for Windows.
4040 [Jeffrey Altman <jaltman@columbia.edu>]
4042 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4043 not set then we don't setup the error code for issuer check errors
4044 to avoid possibly overwriting other errors which the callback does
4045 handle. If an application does set the flag then we assume it knows
4046 what it is doing and can handle the new informational codes
4050 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4051 a general "ANY" type, as such it should be able to decode anything
4052 including tagged types. However it didn't check the class so it would
4053 wrongly interpret tagged types in the same way as their universal
4054 counterpart and unknown types were just rejected. Changed so that the
4055 tagged and unknown types are handled in the same way as a SEQUENCE:
4056 that is the encoding is stored intact. There is also a new type
4057 "V_ASN1_OTHER" which is used when the class is not universal, in this
4058 case we have no idea what the actual type is so we just lump them all
4062 *) On VMS, stdout may very well lead to a file that is written to
4063 in a record-oriented fashion. That means that every write() will
4064 write a separate record, which will be read separately by the
4065 programs trying to read from it. This can be very confusing.
4067 The solution is to put a BIO filter in the way that will buffer
4068 text until a linefeed is reached, and then write everything a
4069 line at a time, so every record written will be an actual line,
4070 not chunks of lines and not (usually doesn't happen, but I've
4071 seen it once) several lines in one record. BIO_f_linebuffer() is
4074 Currently, it's a VMS-only method, because that's where it has
4075 been tested well enough.
4078 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4079 it can return incorrect results.
4080 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4081 but it was in 0.9.6-beta[12].)
4084 *) Disable the check for content being present when verifying detached
4085 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4086 include zero length content when signing messages.
4089 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4090 BIO_ctrl (for BIO pairs).
4093 *) Add DSO method for VMS.
4096 *) Bug fix: Montgomery multiplication could produce results with the
4100 *) Add RPM specification openssl.spec and modify it to build three
4101 packages. The default package contains applications, application
4102 documentation and run-time libraries. The devel package contains
4103 include files, static libraries and function documentation. The
4104 doc package contains the contents of the doc directory. The original
4105 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4108 *) Add a large number of documentation files for many SSL routines.
4109 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4111 *) Add a configuration entry for Sony News 4.
4112 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4114 *) Don't set the two most significant bits to one when generating a
4115 random number < q in the DSA library.
4118 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4119 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4120 the underlying transport is blocking) if a handshake took place.
4121 (The default behaviour is needed by applications such as s_client
4122 and s_server that use select() to determine when to use SSL_read;
4123 but for applications that know in advance when to expect data, it
4124 just makes things more complicated.)
4127 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4131 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4132 work better on such systems.
4133 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4135 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4136 Update PKCS12_parse() so it copies the friendlyName and the
4137 keyid to the certificates aux info.
4140 *) Fix bug in PKCS7_verify() which caused an infinite loop
4141 if there was more than one signature.
4142 [Sven Uszpelkat <su@celocom.de>]
4144 *) Major change in util/mkdef.pl to include extra information
4145 about each symbol, as well as presentig variables as well
4146 as functions. This change means that there's n more need
4147 to rebuild the .num files when some algorithms are excluded.
4150 *) Allow the verify time to be set by an application,
4151 rather than always using the current time.
4154 *) Phase 2 verify code reorganisation. The certificate
4155 verify code now looks up an issuer certificate by a
4156 number of criteria: subject name, authority key id
4157 and key usage. It also verifies self signed certificates
4158 by the same criteria. The main comparison function is
4159 X509_check_issued() which performs these checks.
4161 Lot of changes were necessary in order to support this
4162 without completely rewriting the lookup code.
4164 Authority and subject key identifier are now cached.
4166 The LHASH 'certs' is X509_STORE has now been replaced
4167 by a STACK_OF(X509_OBJECT). This is mainly because an
4168 LHASH can't store or retrieve multiple objects with
4169 the same hash value.
4171 As a result various functions (which were all internal
4172 use only) have changed to handle the new X509_STORE
4173 structure. This will break anything that messed round
4174 with X509_STORE internally.
4176 The functions X509_STORE_add_cert() now checks for an
4177 exact match, rather than just subject name.
4179 The X509_STORE API doesn't directly support the retrieval
4180 of multiple certificates matching a given criteria, however
4181 this can be worked round by performing a lookup first
4182 (which will fill the cache with candidate certificates)
4183 and then examining the cache for matches. This is probably
4184 the best we can do without throwing out X509_LOOKUP
4185 entirely (maybe later...).
4187 The X509_VERIFY_CTX structure has been enhanced considerably.
4189 All certificate lookup operations now go via a get_issuer()
4190 callback. Although this currently uses an X509_STORE it
4191 can be replaced by custom lookups. This is a simple way
4192 to bypass the X509_STORE hackery necessary to make this
4193 work and makes it possible to use more efficient techniques
4194 in future. A very simple version which uses a simple
4195 STACK for its trusted certificate store is also provided
4196 using X509_STORE_CTX_trusted_stack().
4198 The verify_cb() and verify() callbacks now have equivalents
4199 in the X509_STORE_CTX structure.
4201 X509_STORE_CTX also has a 'flags' field which can be used
4202 to customise the verify behaviour.
4205 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4206 excludes S/MIME capabilities.
4209 *) When a certificate request is read in keep a copy of the
4210 original encoding of the signed data and use it when outputing
4211 again. Signatures then use the original encoding rather than
4212 a decoded, encoded version which may cause problems if the
4213 request is improperly encoded.
4216 *) For consistency with other BIO_puts implementations, call
4217 buffer_write(b, ...) directly in buffer_puts instead of calling
4220 In BIO_puts, increment b->num_write as in BIO_write.
4221 [Peter.Sylvester@EdelWeb.fr]
4223 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4224 BN_zero, we may not return a BIGNUM with an array consisting of
4228 *) Avoid calling abort() from within the library when problems are
4229 detected, except if preprocessor symbols have been defined
4230 (such as REF_CHECK, BN_DEBUG etc.).
4233 *) New openssl application 'rsautl'. This utility can be
4234 used for low level RSA operations. DER public key
4235 BIO/fp routines also added.
4238 *) New Configure entry and patches for compiling on QNX 4.
4239 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4241 *) A demo state-machine implementation was sponsored by
4242 Nuron (http://www.nuron.com/) and is now available in
4243 demos/state_machine.
4246 *) New options added to the 'dgst' utility for signature
4247 generation and verification.
4250 *) Unrecognized PKCS#7 content types are now handled via a
4251 catch all ASN1_TYPE structure. This allows unsupported
4252 types to be stored as a "blob" and an application can
4253 encode and decode it manually.
4256 *) Fix various signed/unsigned issues to make a_strex.c
4258 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4260 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4261 length if passed a buffer. ASN1_INTEGER_to_BN failed
4262 if passed a NULL BN and its argument was negative.
4263 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4265 *) Modification to PKCS#7 encoding routines to output definite
4266 length encoding. Since currently the whole structures are in
4267 memory there's not real point in using indefinite length
4268 constructed encoding. However if OpenSSL is compiled with
4269 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4272 *) Added BIO_vprintf() and BIO_vsnprintf().
4275 *) Added more prefixes to parse for in the the strings written
4276 through a logging bio, to cover all the levels that are available
4277 through syslog. The prefixes are now:
4279 PANIC, EMERG, EMR => LOG_EMERG
4280 ALERT, ALR => LOG_ALERT
4281 CRIT, CRI => LOG_CRIT
4282 ERROR, ERR => LOG_ERR
4283 WARNING, WARN, WAR => LOG_WARNING
4284 NOTICE, NOTE, NOT => LOG_NOTICE
4285 INFO, INF => LOG_INFO
4286 DEBUG, DBG => LOG_DEBUG
4288 and as before, if none of those prefixes are present at the
4289 beginning of the string, LOG_ERR is chosen.
4291 On Win32, the LOG_* levels are mapped according to this:
4293 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4294 LOG_WARNING => EVENTLOG_WARNING_TYPE
4295 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4299 *) Made it possible to reconfigure with just the configuration
4300 argument "reconf" or "reconfigure". The command line arguments
4301 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4302 and are retrieved from there when reconfiguring.
4306 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4308 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4311 *) The obj_dat.pl script was messing up the sorting of object
4312 names. The reason was that it compared the quoted version
4313 of strings as a result "OCSP" > "OCSP Signing" because
4314 " > SPACE. Changed script to store unquoted versions of
4315 names and add quotes on output. It was also omitting some
4316 names from the lookup table if they were given a default
4317 value (that is if SN is missing it is given the same
4318 value as LN and vice versa), these are now added on the
4319 grounds that if an object has a name we should be able to
4320 look it up. Finally added warning output when duplicate
4321 short or long names are found.
4324 *) Changes needed for Tandem NSK.
4325 [Scott Uroff <scott@xypro.com>]
4327 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4328 RSA_padding_check_SSLv23(), special padding was never detected
4329 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4330 version rollback attacks was not effective.
4332 In s23_clnt.c, don't use special rollback-attack detection padding
4333 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4334 client; similarly, in s23_srvr.c, don't do the rollback check if
4335 SSL 2.0 is the only protocol enabled in the server.
4338 *) Make it possible to get hexdumps of unprintable data with 'openssl
4339 asn1parse'. By implication, the functions ASN1_parse_dump() and
4340 BIO_dump_indent() are added.
4343 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4344 these print out strings and name structures based on various
4345 flags including RFC2253 support and proper handling of
4346 multibyte characters. Added options to the 'x509' utility
4347 to allow the various flags to be set.
4350 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4351 Also change the functions X509_cmp_current_time() and
4352 X509_gmtime_adj() work with an ASN1_TIME structure,
4353 this will enable certificates using GeneralizedTime in validity
4354 dates to be checked.
4357 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4358 negative public key encodings) on by default,
4359 NO_NEG_PUBKEY_BUG can be set to disable it.
4362 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4363 content octets. An i2c_ASN1_OBJECT is unnecessary because
4364 the encoding can be trivially obtained from the structure.
4367 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4368 not read locks (CRYPTO_r_[un]lock).
4371 *) A first attempt at creating official support for shared
4372 libraries through configuration. I've kept it so the
4373 default is static libraries only, and the OpenSSL programs
4374 are always statically linked for now, but there are
4375 preparations for dynamic linking in place.
4376 This has been tested on Linux and Tru64.
4379 *) Randomness polling function for Win9x, as described in:
4380 Peter Gutmann, Software Generation of Practically Strong
4384 *) Fix so PRNG is seeded in req if using an already existing
4388 *) New options to smime application. -inform and -outform
4389 allow alternative formats for the S/MIME message including
4390 PEM and DER. The -content option allows the content to be
4391 specified separately. This should allow things like Netscape
4392 form signing output easier to verify.
4395 *) Fix the ASN1 encoding of tags using the 'long form'.
4398 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4399 STRING types. These convert content octets to and from the
4400 underlying type. The actual tag and length octets are
4401 already assumed to have been read in and checked. These
4402 are needed because all other string types have virtually
4403 identical handling apart from the tag. By having versions
4404 of the ASN1 functions that just operate on content octets
4405 IMPLICIT tagging can be handled properly. It also allows
4406 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4407 and ASN1_INTEGER are identical apart from the tag.
4410 *) Change the handling of OID objects as follows:
4412 - New object identifiers are inserted in objects.txt, following
4413 the syntax given in objects.README.
4414 - objects.pl is used to process obj_mac.num and create a new
4416 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4419 This is currently kind of a hack, and the perl code in objects.pl
4420 isn't very elegant, but it works as I intended. The simplest way
4421 to check that it worked correctly is to look in obj_dat.h and
4422 check the array nid_objs and make sure the objects haven't moved
4423 around (this is important!). Additions are OK, as well as
4424 consistent name changes.
4427 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4430 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4431 The given file adds to whatever has already been seeded into the
4432 random pool through the RANDFILE configuration file option or
4433 environment variable, or the default random state file.
4436 *) mkstack.pl now sorts each macro group into lexical order.
4437 Previously the output order depended on the order the files
4438 appeared in the directory, resulting in needless rewriting
4442 *) Patches to make OpenSSL compile under Win32 again. Mostly
4443 work arounds for the VC++ problem that it treats func() as
4444 func(void). Also stripped out the parts of mkdef.pl that
4445 added extra typesafe functions: these no longer exist.
4448 *) Reorganisation of the stack code. The macros are now all
4449 collected in safestack.h . Each macro is defined in terms of
4450 a "stack macro" of the form SKM_<name>(type, a, b). The
4451 DEBUG_SAFESTACK is now handled in terms of function casts,
4452 this has the advantage of retaining type safety without the
4453 use of additional functions. If DEBUG_SAFESTACK is not defined
4454 then the non typesafe macros are used instead. Also modified the
4455 mkstack.pl script to handle the new form. Needs testing to see
4456 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4457 the default if no major problems. Similar behaviour for ASN1_SET_OF
4458 and PKCS12_STACK_OF.
4461 *) When some versions of IIS use the 'NET' form of private key the
4462 key derivation algorithm is different. Normally MD5(password) is
4463 used as a 128 bit RC4 key. In the modified case
4464 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4465 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4466 as the old Netscape_RSA functions except they have an additional
4467 'sgckey' parameter which uses the modified algorithm. Also added
4468 an -sgckey command line option to the rsa utility. Thanks to
4469 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4470 algorithm to openssl-dev.
4473 *) The evp_local.h macros were using 'c.##kname' which resulted in
4474 invalid expansion on some systems (SCO 5.0.5 for example).
4475 Corrected to 'c.kname'.
4476 [Phillip Porch <root@theporch.com>]
4478 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4479 a STACK of email addresses from a certificate or request, these look
4480 in the subject name and the subject alternative name extensions and
4481 omit any duplicate addresses.
4484 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4485 This makes DSA verification about 2 % faster.
4488 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4489 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4490 plus overhead for 1024 bit moduli).
4491 This makes exponentiations about 0.5 % faster for 1024 bit
4492 exponents (as measured by "openssl speed rsa2048").
4495 *) Rename memory handling macros to avoid conflicts with other
4497 Malloc => OPENSSL_malloc
4498 Malloc_locked => OPENSSL_malloc_locked
4499 Realloc => OPENSSL_realloc
4500 Free => OPENSSL_free
4503 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4504 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4507 *) CygWin32 support.
4508 [John Jarvie <jjarvie@newsguy.com>]
4510 *) The type-safe stack code has been rejigged. It is now only compiled
4511 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4512 by default all type-specific stack functions are "#define"d back to
4513 standard stack functions. This results in more streamlined output
4514 but retains the type-safety checking possibilities of the original
4518 *) The STACK code has been cleaned up, and certain type declarations
4519 that didn't make a lot of sense have been brought in line. This has
4520 also involved a cleanup of sorts in safestack.h to more correctly
4521 map type-safe stack functions onto their plain stack counterparts.
4522 This work has also resulted in a variety of "const"ifications of
4523 lots of the code, especially "_cmp" operations which should normally
4524 be prototyped with "const" parameters anyway.
4527 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4528 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4529 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4530 where all of 'md' is used each time the PRNG is used, but 'state'
4531 is used only indexed by a cyclic counter. As entropy may not be
4532 well distributed from the beginning, 'md' is important as a
4533 chaining variable. However, the output function chains only half
4534 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4535 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4536 in all of 'state' being rewritten, with the new values depending
4537 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4540 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4541 the handshake is continued after ssl_verify_cert_chain();
4542 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4543 can lead to 'unexplainable' connection aborts later.
4544 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4546 *) Major EVP API cipher revision.
4547 Add hooks for extra EVP features. This allows various cipher
4548 parameters to be set in the EVP interface. Support added for variable
4549 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4550 setting of RC2 and RC5 parameters.
4552 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4555 Remove lots of duplicated code from the EVP library. For example *every*
4556 cipher init() function handles the 'iv' in the same way according to the
4557 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4558 for CFB and OFB modes they zero ctx->num.
4560 New functionality allows removal of S/MIME code RC2 hack.
4562 Most of the routines have the same form and so can be declared in terms
4565 By shifting this to the top level EVP_CipherInit() it can be removed from
4566 all individual ciphers. If the cipher wants to handle IVs or keys
4567 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4570 Change lots of functions like EVP_EncryptUpdate() to now return a
4571 value: although software versions of the algorithms cannot fail
4572 any installed hardware versions can.
4575 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4576 this option is set, tolerate broken clients that send the negotiated
4577 protocol version number instead of the requested protocol version
4581 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4582 i.e. non-zero for export ciphersuites, zero otherwise.
4583 Previous versions had this flag inverted, inconsistent with
4584 rsa_tmp_cb (..._TMP_RSA_CB).
4585 [Bodo Moeller; problem reported by Amit Chopra]
4587 *) Add missing DSA library text string. Work around for some IIS
4588 key files with invalid SEQUENCE encoding.
4591 *) Add a document (doc/standards.txt) that list all kinds of standards
4592 and so on that are implemented in OpenSSL.
4595 *) Enhance c_rehash script. Old version would mishandle certificates
4596 with the same subject name hash and wouldn't handle CRLs at all.
4597 Added -fingerprint option to crl utility, to support new c_rehash
4601 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4604 *) Fix for SSL server purpose checking. Server checking was
4605 rejecting certificates which had extended key usage present
4606 but no ssl client purpose.
4607 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4609 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4610 is a little unclear about how a blank password is handled.
4611 Since the password in encoded as a BMPString with terminating
4612 double NULL a zero length password would end up as just the
4613 double NULL. However no password at all is different and is
4614 handled differently in the PKCS#12 key generation code. NS
4615 treats a blank password as zero length. MSIE treats it as no
4616 password on export: but it will try both on import. We now do
4617 the same: PKCS12_parse() tries zero length and no password if
4618 the password is set to "" or NULL (NULL is now a valid password:
4619 it wasn't before) as does the pkcs12 application.
4622 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4623 perror when PEM_read_bio_X509_REQ fails, the error message must
4624 be obtained from the error queue.
4627 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4628 it in ERR_remove_state if appropriate, and change ERR_get_state
4629 accordingly to avoid race conditions (this is necessary because
4630 thread_hash is no longer constant once set).
4633 *) Bugfix for linux-elf makefile.one.
4636 *) RSA_get_default_method() will now cause a default
4637 RSA_METHOD to be chosen if one doesn't exist already.
4638 Previously this was only set during a call to RSA_new()
4639 or RSA_new_method(NULL) meaning it was possible for
4640 RSA_get_default_method() to return NULL.
4643 *) Added native name translation to the existing DSO code
4644 that will convert (if the flag to do so is set) filenames
4645 that are sufficiently small and have no path information
4646 into a canonical native form. Eg. "blah" converted to
4647 "libblah.so" or "blah.dll" etc.
4650 *) New function ERR_error_string_n(e, buf, len) which is like
4651 ERR_error_string(e, buf), but writes at most 'len' bytes
4652 including the 0 terminator. For ERR_error_string_n, 'buf'
4654 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4656 *) CONF library reworked to become more general. A new CONF
4657 configuration file reader "class" is implemented as well as a
4658 new functions (NCONF_*, for "New CONF") to handle it. The now
4659 old CONF_* functions are still there, but are reimplemented to
4660 work in terms of the new functions. Also, a set of functions
4661 to handle the internal storage of the configuration data is
4662 provided to make it easier to write new configuration file
4663 reader "classes" (I can definitely see something reading a
4664 configuration file in XML format, for example), called _CONF_*,
4665 or "the configuration storage API"...
4667 The new configuration file reading functions are:
4669 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4670 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4672 NCONF_default, NCONF_WIN32
4674 NCONF_dump_fp, NCONF_dump_bio
4676 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4677 NCONF_new creates a new CONF object. This works in the same way
4678 as other interfaces in OpenSSL, like the BIO interface.
4679 NCONF_dump_* dump the internal storage of the configuration file,
4680 which is useful for debugging. All other functions take the same
4681 arguments as the old CONF_* functions wth the exception of the
4682 first that must be a `CONF *' instead of a `LHASH *'.
4684 To make it easer to use the new classes with the old CONF_* functions,
4685 the function CONF_set_default_method is provided.
4688 *) Add '-tls1' option to 'openssl ciphers', which was already
4689 mentioned in the documentation but had not been implemented.
4690 (This option is not yet really useful because even the additional
4691 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4694 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4695 OpenSSL-based applications) load shared libraries and bind to
4696 them in a portable way.
4697 [Geoff Thorpe, with contributions from Richard Levitte]
4699 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4701 *) Make sure _lrotl and _lrotr are only used with MSVC.
4703 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4704 (the default implementation of RAND_status).
4706 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4707 to '-clrext' (= clear extensions), as intended and documented.
4708 [Bodo Moeller; inconsistency pointed out by Michael Attili
4711 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4712 was larger than the MD block size.
4713 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4715 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4716 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4717 using the passed key: if the passed key was a private key the result
4718 of X509_print(), for example, would be to print out all the private key
4722 *) des_quad_cksum() byte order bug fix.
4723 [Ulf Möller, using the problem description in krb4-0.9.7, where
4724 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4726 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4728 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4730 *) For easily testing in shell scripts whether some command
4731 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4732 returns with exit code 0 iff no command of the given name is available.
4733 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4734 the output goes to stdout and nothing is printed to stderr.
4735 Additional arguments are always ignored.
4737 Since for each cipher there is a command of the same name,
4738 the 'no-cipher' compilation switches can be tested this way.
4740 ('openssl no-XXX' is not able to detect pseudo-commands such
4741 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4744 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4747 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4748 is set; it will be thrown away anyway because each handshake creates
4750 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4751 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4752 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4753 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4756 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4757 'Q' and 'R' lose their special meanings (quit/renegotiate).
4758 This is part of what -quiet does; unlike -quiet, -ign_eof
4759 does not suppress any output.
4762 *) Add compatibility options to the purpose and trust code. The
4763 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4764 accepts a certificate or CA, this was the previous behaviour,
4765 with all the associated security issues.
4767 X509_TRUST_COMPAT is the old trust behaviour: only and
4768 automatically trust self signed roots in certificate store. A
4769 new trust setting X509_TRUST_DEFAULT is used to specify that
4770 a purpose has no associated trust setting and it should instead
4771 use the value in the default purpose.
4774 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4775 and fix a memory leak.
4778 *) In util/mkerr.pl (which implements 'make errors'), preserve
4779 reason strings from the previous version of the .c file, as
4780 the default to have only downcase letters (and digits) in
4781 automatically generated reasons codes is not always appropriate.
4784 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4785 using strerror. Previously, ERR_reason_error_string() returned
4786 library names as reason strings for SYSerr; but SYSerr is a special
4787 case where small numbers are errno values, not library numbers.
4790 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4791 converts DSA parameters into DH parameters. (When creating parameters,
4792 DSA_generate_parameters is used.)
4795 *) Include 'length' (recommended exponent length) in C code generated
4796 by 'openssl dhparam -C'.
4799 *) The second argument to set_label in perlasm was already being used
4800 so couldn't be used as a "file scope" flag. Moved to third argument
4804 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4805 instead of RAND_bytes for encryption IVs and salts.
4808 *) Include RAND_status() into RAND_METHOD instead of implementing
4809 it only for md_rand.c Otherwise replacing the PRNG by calling
4810 RAND_set_rand_method would be impossible.
4813 *) Don't let DSA_generate_key() enter an infinite loop if the random
4814 number generation fails.
4817 *) New 'rand' application for creating pseudo-random output.
4820 *) Added configuration support for Linux/IA64
4821 [Rolf Haberrecker <rolf@suse.de>]
4823 *) Assembler module support for Mingw32.
4826 *) Shared library support for HPUX (in shlib/).
4827 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4829 *) Shared library support for Solaris gcc.
4830 [Lutz Behnke <behnke@trustcenter.de>]
4832 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4834 *) PKCS7_encrypt() was adding text MIME headers twice because they
4835 were added manually and by SMIME_crlf_copy().
4838 *) In bntest.c don't call BN_rand with zero bits argument.
4839 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4841 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4842 case was implemented. This caused BN_div_recp() to fail occasionally.
4845 *) Add an optional second argument to the set_label() in the perl
4846 assembly language builder. If this argument exists and is set
4847 to 1 it signals that the assembler should use a symbol whose
4848 scope is the entire file, not just the current function. This
4849 is needed with MASM which uses the format label:: for this scope.
4850 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4852 *) Change the ASN1 types so they are typedefs by default. Before
4853 almost all types were #define'd to ASN1_STRING which was causing
4854 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4858 *) Change names of new functions to the new get1/get0 naming
4859 convention: After 'get1', the caller owns a reference count
4860 and has to call ..._free; 'get0' returns a pointer to some
4861 data structure without incrementing reference counters.
4862 (Some of the existing 'get' functions increment a reference
4863 counter, some don't.)
4864 Similarly, 'set1' and 'add1' functions increase reference
4865 counters or duplicate objects.
4868 *) Allow for the possibility of temp RSA key generation failure:
4869 the code used to assume it always worked and crashed on failure.
4872 *) Fix potential buffer overrun problem in BIO_printf().
4873 [Ulf Möller, using public domain code by Patrick Powell; problem
4874 pointed out by David Sacerdote <das33@cornell.edu>]
4876 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4877 RAND_egd() and RAND_status(). In the command line application,
4878 the EGD socket can be specified like a seed file using RANDFILE
4882 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4883 Some CAs (e.g. Verisign) distribute certificates in this form.
4886 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4887 list to exclude them. This means that no special compilation option
4888 is needed to use anonymous DH: it just needs to be included in the
4892 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4893 EVP_MD_type. The old functionality is available in a new macro called
4894 EVP_MD_md(). Change code that uses it and update docs.
4897 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4898 where the 'void *' argument is replaced by a function pointer argument.
4899 Previously 'void *' was abused to point to functions, which works on
4900 many platforms, but is not correct. As these functions are usually
4901 called by macros defined in OpenSSL header files, most source code
4902 should work without changes.
4905 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4906 sections with information on -D... compiler switches used for
4907 compiling the library so that applications can see them. To enable
4908 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4909 must be defined. E.g.,
4910 #define OPENSSL_ALGORITHM_DEFINES
4911 #include <openssl/opensslconf.h>
4912 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4913 [Richard Levitte, Ulf and Bodo Möller]
4915 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4919 *) Change the 'other' type in certificate aux info to a STACK_OF
4920 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4921 the required ASN1 format: arbitrary types determined by an OID.
4924 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4925 argument to 'req'. This is not because the function is newer or
4926 better than others it just uses the work 'NEW' in the certificate
4927 request header lines. Some software needs this.
4930 *) Reorganise password command line arguments: now passwords can be
4931 obtained from various sources. Delete the PEM_cb function and make
4932 it the default behaviour: i.e. if the callback is NULL and the
4933 usrdata argument is not NULL interpret it as a null terminated pass
4934 phrase. If usrdata and the callback are NULL then the pass phrase
4935 is prompted for as usual.
4938 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4939 the support is automatically enabled. The resulting binaries will
4940 autodetect the card and use it if present.
4941 [Ben Laurie and Compaq Inc.]
4943 *) Work around for Netscape hang bug. This sends certificate request
4944 and server done in one record. Since this is perfectly legal in the
4945 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4946 the bugs/SSLv3 entry for more info.
4949 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4952 *) Add -rand argument to smime and pkcs12 applications and read/write
4956 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4959 *) Add command line password options to the remaining applications.
4962 *) Bug fix for BN_div_recp() for numerators with an even number of
4966 *) More tests in bntest.c, and changed test_bn output.
4969 *) ./config recognizes MacOS X now.
4972 *) Bug fix for BN_div() when the first words of num and divsor are
4973 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4976 *) Add support for various broken PKCS#8 formats, and command line
4977 options to produce them.
4980 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4981 get temporary BIGNUMs from a BN_CTX.
4984 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4988 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4989 include a #define from the old name to the new. The original intent
4990 was that statically linked binaries could for example just call
4991 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4992 link with digests. This never worked becayse SSLeay_add_all_digests()
4993 and SSLeay_add_all_ciphers() were in the same source file so calling
4994 one would link with the other. They are now in separate source files.
4997 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5000 *) Use a less unusual form of the Miller-Rabin primality test (it used
5001 a binary algorithm for exponentiation integrated into the Miller-Rabin
5002 loop, our standard modexp algorithms are faster).
5005 *) Support for the EBCDIC character set completed.
5006 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5008 *) Source code cleanups: use const where appropriate, eliminate casts,
5009 use void * instead of char * in lhash.
5012 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5013 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5014 this the server could overwrite ephemeral keys that the client
5018 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5019 using 50 iterations of the Rabin-Miller test.
5021 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5022 iterations of the Rabin-Miller test as required by the appendix
5023 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5024 As BN_is_prime_fasttest includes trial division, DSA parameter
5025 generation becomes much faster.
5027 This implies a change for the callback functions in DSA_is_prime
5028 and DSA_generate_parameters: The callback function is called once
5029 for each positive witness in the Rabin-Miller test, not just
5030 occasionally in the inner loop; and the parameters to the
5031 callback function now provide an iteration count for the outer
5032 loop rather than for the current invocation of the inner loop.
5033 DSA_generate_parameters additionally can call the callback
5034 function with an 'iteration count' of -1, meaning that a
5035 candidate has passed the trial division test (when q is generated
5036 from an application-provided seed, trial division is skipped).
5039 *) New function BN_is_prime_fasttest that optionally does trial
5040 division before starting the Rabin-Miller test and has
5041 an additional BN_CTX * argument (whereas BN_is_prime always
5042 has to allocate at least one BN_CTX).
5043 'callback(1, -1, cb_arg)' is called when a number has passed the
5044 trial division stage.
5047 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5051 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5054 *) New function BN_pseudo_rand().
5057 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5058 bignum version of BN_from_montgomery() with the working code from
5059 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5063 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5064 made it impossible to use the same SSL_SESSION data structure in
5065 SSL2 clients in multiple threads.
5068 *) The return value of RAND_load_file() no longer counts bytes obtained
5069 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5070 to seed the PRNG (previously an explicit byte count was required).
5071 [Ulf Möller, Bodo Möller]
5073 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5074 used (char *) instead of (void *) and had casts all over the place.
5077 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5080 *) Retain source code compatibility for BN_prime_checks macro:
5081 BN_is_prime(..., BN_prime_checks, ...) now uses
5082 BN_prime_checks_for_size to determine the appropriate number of
5083 Rabin-Miller iterations.
5086 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5087 DH_CHECK_P_NOT_SAFE_PRIME.
5088 (Check if this is true? OpenPGP calls them "strong".)
5091 *) Merge the functionality of "dh" and "gendh" programs into a new program
5092 "dhparam". The old programs are retained for now but will handle DH keys
5093 (instead of parameters) in future.
5096 *) Make the ciphers, s_server and s_client programs check the return values
5097 when a new cipher list is set.
5100 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5101 ciphers. Before when the 56bit ciphers were enabled the sorting was
5104 The syntax for the cipher sorting has been extended to support sorting by
5105 cipher-strength (using the strength_bits hard coded in the tables).
5106 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5108 Fix a bug in the cipher-command parser: when supplying a cipher command
5109 string with an "undefined" symbol (neither command nor alphanumeric
5110 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5111 an error is flagged.
5113 Due to the strength-sorting extension, the code of the
5114 ssl_create_cipher_list() function was completely rearranged. I hope that
5115 the readability was also increased :-)
5116 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5118 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5119 for the first serial number and places 2 in the serial number file. This
5120 avoids problems when the root CA is created with serial number zero and
5121 the first user certificate has the same issuer name and serial number
5125 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5126 the new code. Add documentation for this stuff.
5129 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5130 X509_*() to X509at_*() on the grounds that they don't handle X509
5131 structures and behave in an analagous way to the X509v3 functions:
5132 they shouldn't be called directly but wrapper functions should be used
5135 So we also now have some wrapper functions that call the X509at functions
5136 when passed certificate requests. (TO DO: similar things can be done with
5137 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5138 things. Some of these need some d2i or i2d and print functionality
5139 because they handle more complex structures.)
5142 *) Add missing #ifndefs that caused missing symbols when building libssl
5143 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5144 NO_RSA in ssl/s2*.c.
5145 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5147 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5148 has a return value which indicates the quality of the random data
5149 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5150 error queue. New function RAND_pseudo_bytes() generates output that is
5151 guaranteed to be unique but not unpredictable. RAND_add is like
5152 RAND_seed, but takes an extra argument for an entropy estimate
5153 (RAND_seed always assumes full entropy).
5156 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5157 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5158 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5159 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5160 false-positive rate of at most 2^-80 for random input.
5163 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5166 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5167 in the 0.9.5 release), this returns the chain
5168 from an X509_CTX structure with a dup of the stack and all
5169 the X509 reference counts upped: so the stack will exist
5170 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5173 Also make SSL_SESSION_print() print out the verify return
5177 *) Add manpage for the pkcs12 command. Also change the default
5178 behaviour so MAC iteration counts are used unless the new
5179 -nomaciter option is used. This improves file security and
5180 only older versions of MSIE (4.0 for example) need it.
5183 *) Honor the no-xxx Configure options when creating .DEF files.
5186 *) Add PKCS#10 attributes to field table: challengePassword,
5187 unstructuredName and unstructuredAddress. These are taken from
5188 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5189 international characters are used.
5191 More changes to X509_ATTRIBUTE code: allow the setting of types
5192 based on strings. Remove the 'loc' parameter when adding
5193 attributes because these will be a SET OF encoding which is sorted
5197 *) Initial changes to the 'req' utility to allow request generation
5198 automation. This will allow an application to just generate a template
5199 file containing all the field values and have req construct the
5202 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5203 used all over the place including certificate requests and PKCS#7
5204 structures. They are currently handled manually where necessary with
5205 some primitive wrappers for PKCS#7. The new functions behave in a
5206 manner analogous to the X509 extension functions: they allow
5207 attributes to be looked up by NID and added.
5209 Later something similar to the X509V3 code would be desirable to
5210 automatically handle the encoding, decoding and printing of the
5211 more complex types. The string types like challengePassword can
5212 be handled by the string table functions.
5214 Also modified the multi byte string table handling. Now there is
5215 a 'global mask' which masks out certain types. The table itself
5216 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5217 is useful when for example there is only one permissible type
5218 (as in countryName) and using the mask might result in no valid
5222 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5223 SSL_get_peer_finished to allow applications to obtain the latest
5224 Finished messages sent to the peer or expected from the peer,
5225 respectively. (SSL_get_peer_finished is usually the Finished message
5226 actually received from the peer, otherwise the protocol will be aborted.)
5228 As the Finished message are message digests of the complete handshake
5229 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5230 be used for external authentication procedures when the authentication
5231 provided by SSL/TLS is not desired or is not enough.
5234 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5235 the host supports BWX extension and if Compaq C is present on the
5236 $PATH. Just exploiting of the BWX extension results in 20-30%
5237 performance kick for some algorithms, e.g. DES and RC4 to mention
5238 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5242 *) Add support for MS "fast SGC". This is arguably a violation of the
5243 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5244 weak crypto and after checking the certificate is SGC a second one
5245 with strong crypto. MS SGC stops the first handshake after receiving
5246 the server certificate message and sends a second client hello. Since
5247 a server will typically do all the time consuming operations before
5248 expecting any further messages from the client (server key exchange
5249 is the most expensive) there is little difference between the two.
5251 To get OpenSSL to support MS SGC we have to permit a second client
5252 hello message after we have sent server done. In addition we have to
5253 reset the MAC if we do get this second client hello.
5256 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5257 if a DER encoded private key is RSA or DSA traditional format. Changed
5258 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5259 format DER encoded private key. Newer code should use PKCS#8 format which
5260 has the key type encoded in the ASN1 structure. Added DER private key
5261 support to pkcs8 application.
5264 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5265 ciphersuites has been selected (as required by the SSL 3/TLS 1
5266 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5267 is set, we interpret this as a request to violate the specification
5268 (the worst that can happen is a handshake failure, and 'correct'
5269 behaviour would result in a handshake failure anyway).
5272 *) In SSL_CTX_add_session, take into account that there might be multiple
5273 SSL_SESSION structures with the same session ID (e.g. when two threads
5274 concurrently obtain them from an external cache).
5275 The internal cache can handle only one SSL_SESSION with a given ID,
5276 so if there's a conflict, we now throw out the old one to achieve
5280 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5281 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5282 some routines that use cipher OIDs: some ciphers do not have OIDs
5283 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5287 *) Simplify the trust setting structure and code. Now we just have
5288 two sequences of OIDs for trusted and rejected settings. These will
5289 typically have values the same as the extended key usage extension
5290 and any application specific purposes.
5292 The trust checking code now has a default behaviour: it will just
5293 check for an object with the same NID as the passed id. Functions can
5294 be provided to override either the default behaviour or the behaviour
5295 for a given id. SSL client, server and email already have functions
5296 in place for compatibility: they check the NID and also return "trusted"
5297 if the certificate is self signed.
5300 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5301 traditional format into an EVP_PKEY structure.
5304 *) Add a password callback function PEM_cb() which either prompts for
5305 a password if usr_data is NULL or otherwise assumes it is a null
5306 terminated password. Allow passwords to be passed on command line
5307 environment or config files in a few more utilities.
5310 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5311 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5312 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5313 Update documentation.
5316 *) Support for ASN1 "NULL" type. This could be handled before by using
5317 ASN1_TYPE but there wasn't any function that would try to read a NULL
5318 and produce an error if it couldn't. For compatibility we also have
5319 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5320 don't allocate anything because they don't need to.
5323 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5325 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5327 *) Rebuild of the memory allocation routines used by OpenSSL code and
5328 possibly others as well. The purpose is to make an interface that
5329 provide hooks so anyone can build a separate set of allocation and
5330 deallocation routines to be used by OpenSSL, for example memory
5331 pool implementations, or something else, which was previously hard
5332 since Malloc(), Realloc() and Free() were defined as macros having
5333 the values malloc, realloc and free, respectively (except for Win32
5334 compilations). The same is provided for memory debugging code.
5335 OpenSSL already comes with functionality to find memory leaks, but
5336 this gives people a chance to debug other memory problems.
5338 With these changes, a new set of functions and macros have appeared:
5340 CRYPTO_set_mem_debug_functions() [F]
5341 CRYPTO_get_mem_debug_functions() [F]
5342 CRYPTO_dbg_set_options() [F]
5343 CRYPTO_dbg_get_options() [F]
5344 CRYPTO_malloc_debug_init() [M]
5346 The memory debug functions are NULL by default, unless the library
5347 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5348 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5349 gives the standard debugging functions that come with OpenSSL) or
5350 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5351 provided by the library user) must be used. When the standard
5352 debugging functions are used, CRYPTO_dbg_set_options can be used to
5353 request additional information:
5354 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5355 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5357 Also, things like CRYPTO_set_mem_functions will always give the
5358 expected result (the new set of functions is used for allocation
5359 and deallocation) at all times, regardless of platform and compiler
5362 To finish it up, some functions that were never use in any other
5363 way than through macros have a new API and new semantic:
5366 CRYPTO_dbg_realloc()
5369 All macros of value have retained their old syntax.
5370 [Richard Levitte and Bodo Moeller]
5372 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5373 ordering of SMIMECapabilities wasn't in "strength order" and there
5374 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5378 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5379 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5380 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5382 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5383 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5384 functionality to handle multipart/signed properly) and a utility
5385 called 'smime' to call all this stuff. This is based on code I
5386 originally wrote for Celo who have kindly allowed it to be
5387 included in OpenSSL.
5390 *) Add variants des_set_key_checked and des_set_key_unchecked of
5391 des_set_key (aka des_key_sched). Global variable des_check_key
5392 decides which of these is called by des_set_key; this way
5393 des_check_key behaves as it always did, but applications and
5394 the library itself, which was buggy for des_check_key == 1,
5395 have a cleaner way to pick the version they need.
5398 *) New function PKCS12_newpass() which changes the password of a
5402 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5403 dynamic mix. In both cases the ids can be used as an index into the
5404 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5405 functions so they accept a list of the field values and the
5406 application doesn't need to directly manipulate the X509_TRUST
5410 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5414 *) Modify the way the V3 extension code looks up extensions. This now
5415 works in a similar way to the object code: we have some "standard"
5416 extensions in a static table which is searched with OBJ_bsearch()
5417 and the application can add dynamic ones if needed. The file
5418 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5419 updated whenever a new extension is added to the core code and kept
5420 in ext_nid order. There is a simple program 'tabtest.c' which checks
5421 this. New extensions are not added too often so this file can readily
5422 be maintained manually.
5424 There are two big advantages in doing things this way. The extensions
5425 can be looked up immediately and no longer need to be "added" using
5426 X509V3_add_standard_extensions(): this function now does nothing.
5427 [Side note: I get *lots* of email saying the extension code doesn't
5428 work because people forget to call this function]
5429 Also no dynamic allocation is done unless new extensions are added:
5430 so if we don't add custom extensions there is no need to call
5431 X509V3_EXT_cleanup().
5434 *) Modify enc utility's salting as follows: make salting the default. Add a
5435 magic header, so unsalted files fail gracefully instead of just decrypting
5436 to garbage. This is because not salting is a big security hole, so people
5437 should be discouraged from doing it.
5440 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5441 digest to be passed on the command line but it only used this
5442 parameter when signing a certificate. Modified so all relevant
5443 operations are affected by the digest parameter including the
5444 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5445 DSA key was used because it didn't fix the digest.
5448 *) Initial certificate chain verify code. Currently tests the untrusted
5449 certificates for consistency with the verify purpose (which is set
5450 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5452 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5453 this is because it will reject chains with invalid extensions whereas
5454 every previous version of OpenSSL and SSLeay made no checks at all.
5456 Trust code: checks the root CA for the relevant trust settings. Trust
5457 settings have an initial value consistent with the verify purpose: e.g.
5458 if the verify purpose is for SSL client use it expects the CA to be
5459 trusted for SSL client use. However the default value can be changed to
5460 permit custom trust settings: one example of this would be to only trust
5461 certificates from a specific "secure" set of CAs.
5463 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5464 which should be used for version portability: especially since the
5465 verify structure is likely to change more often now.
5467 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5468 to set them. If not set then assume SSL clients will verify SSL servers
5471 Two new options to the verify program: -untrusted allows a set of
5472 untrusted certificates to be passed in and -purpose which sets the
5473 intended purpose of the certificate. If a purpose is set then the
5474 new chain verify code is used to check extension consistency.
5477 *) Support for the authority information access extension.
5480 *) Modify RSA and DSA PEM read routines to transparently handle
5481 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5482 public keys in a format compatible with certificate
5483 SubjectPublicKeyInfo structures. Unfortunately there were already
5484 functions called *_PublicKey_* which used various odd formats so
5485 these are retained for compatibility: however the DSA variants were
5486 never in a public release so they have been deleted. Changed dsa/rsa
5487 utilities to handle the new format: note no releases ever handled public
5488 keys so we should be OK.
5490 The primary motivation for this change is to avoid the same fiasco
5491 that dogs private keys: there are several incompatible private key
5492 formats some of which are standard and some OpenSSL specific and
5493 require various evil hacks to allow partial transparent handling and
5494 even then it doesn't work with DER formats. Given the option anything
5495 other than PKCS#8 should be dumped: but the other formats have to
5496 stay in the name of compatibility.
5498 With public keys and the benefit of hindsight one standard format
5499 is used which works with EVP_PKEY, RSA or DSA structures: though
5500 it clearly returns an error if you try to read the wrong kind of key.
5502 Added a -pubkey option to the 'x509' utility to output the public key.
5503 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5504 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5505 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5506 that do the same as the EVP_PKEY_assign_*() except they up the
5507 reference count of the added key (they don't "swallow" the
5511 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5512 CRLs would fail if the file contained no certificates or no CRLs:
5513 added a new function to read in both types and return the number
5514 read: this means that if none are read it will be an error. The
5515 DER versions of the certificate and CRL reader would always fail
5516 because it isn't possible to mix certificates and CRLs in DER format
5517 without choking one or the other routine. Changed this to just read
5518 a certificate: this is the best we can do. Also modified the code
5519 in apps/verify.c to take notice of return codes: it was previously
5520 attempting to read in certificates from NULL pointers and ignoring
5521 any errors: this is one reason why the cert and CRL reader seemed
5522 to work. It doesn't check return codes from the default certificate
5523 routines: these may well fail if the certificates aren't installed.
5526 *) Code to support otherName option in GeneralName.
5529 *) First update to verify code. Change the verify utility
5530 so it warns if it is passed a self signed certificate:
5531 for consistency with the normal behaviour. X509_verify
5532 has been modified to it will now verify a self signed
5533 certificate if *exactly* the same certificate appears
5534 in the store: it was previously impossible to trust a
5535 single self signed certificate. This means that:
5536 openssl verify ss.pem
5537 now gives a warning about a self signed certificate but
5538 openssl verify -CAfile ss.pem ss.pem
5542 *) For servers, store verify_result in SSL_SESSION data structure
5543 (and add it to external session representation).
5544 This is needed when client certificate verifications fails,
5545 but an application-provided verification callback (set by
5546 SSL_CTX_set_cert_verify_callback) allows accepting the session
5547 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5548 but returns 1): When the session is reused, we have to set
5549 ssl->verify_result to the appropriate error code to avoid
5551 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5553 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5554 case in PKCS7_dataInit() where the signed PKCS7 structure
5555 didn't contain any existing data because it was being created.
5556 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5558 *) Add a salt to the key derivation routines in enc.c. This
5559 forms the first 8 bytes of the encrypted file. Also add a
5560 -S option to allow a salt to be input on the command line.
5563 *) New function X509_cmp(). Oddly enough there wasn't a function
5564 to compare two certificates. We do this by working out the SHA1
5565 hash and comparing that. X509_cmp() will be needed by the trust
5569 *) SSL_get1_session() is like SSL_get_session(), but increments
5570 the reference count in the SSL_SESSION returned.
5571 [Geoff Thorpe <geoff@eu.c2.net>]
5573 *) Fix for 'req': it was adding a null to request attributes.
5574 Also change the X509_LOOKUP and X509_INFO code to handle
5575 certificate auxiliary information.
5578 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5582 *) Add the possibility to add extra information to the memory leak
5583 detecting output, to form tracebacks, showing from where each
5584 allocation was originated: CRYPTO_push_info("constant string") adds
5585 the string plus current file name and line number to a per-thread
5586 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5587 is like calling CYRPTO_pop_info() until the stack is empty.
5588 Also updated memory leak detection code to be multi-thread-safe.
5591 *) Add options -text and -noout to pkcs7 utility and delete the
5592 encryption options which never did anything. Update docs.
5595 *) Add options to some of the utilities to allow the pass phrase
5596 to be included on either the command line (not recommended on
5597 OSes like Unix) or read from the environment. Update the
5598 manpages and fix a few bugs.
5601 *) Add a few manpages for some of the openssl commands.
5604 *) Fix the -revoke option in ca. It was freeing up memory twice,
5605 leaking and not finding already revoked certificates.
5608 *) Extensive changes to support certificate auxiliary information.
5609 This involves the use of X509_CERT_AUX structure and X509_AUX
5610 functions. An X509_AUX function such as PEM_read_X509_AUX()
5611 can still read in a certificate file in the usual way but it
5612 will also read in any additional "auxiliary information". By
5613 doing things this way a fair degree of compatibility can be
5614 retained: existing certificates can have this information added
5615 using the new 'x509' options.
5617 Current auxiliary information includes an "alias" and some trust
5618 settings. The trust settings will ultimately be used in enhanced
5619 certificate chain verification routines: currently a certificate
5620 can only be trusted if it is self signed and then it is trusted
5624 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5625 The problem was that one of the replacement routines had not been working
5626 since SSLeay releases. For now the offending routine has been replaced
5627 with non-optimised assembler. Even so, this now gives around 95%
5628 performance improvement for 1024 bit RSA signs.
5631 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5632 handling. Most clients have the effective key size in bits equal to
5633 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5634 A few however don't do this and instead use the size of the decrypted key
5635 to determine the RC2 key length and the AlgorithmIdentifier to determine
5636 the effective key length. In this case the effective key length can still
5637 be 40 bits but the key length can be 168 bits for example. This is fixed
5638 by manually forcing an RC2 key into the EVP_PKEY structure because the
5639 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5640 the key length and effective key length are equal.
5643 *) Add a bunch of functions that should simplify the creation of
5644 X509_NAME structures. Now you should be able to do:
5645 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5646 and have it automatically work out the correct field type and fill in
5647 the structures. The more adventurous can try:
5648 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5649 and it will (hopefully) work out the correct multibyte encoding.
5652 *) Change the 'req' utility to use the new field handling and multibyte
5653 copy routines. Before the DN field creation was handled in an ad hoc
5654 way in req, ca, and x509 which was rather broken and didn't support
5655 BMPStrings or UTF8Strings. Since some software doesn't implement
5656 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5657 using the dirstring_type option. See the new comment in the default
5658 openssl.cnf for more info.
5661 *) Make crypto/rand/md_rand.c more robust:
5662 - Assure unique random numbers after fork().
5663 - Make sure that concurrent threads access the global counter and
5664 md serializably so that we never lose entropy in them
5665 or use exactly the same state in multiple threads.
5666 Access to the large state is not always serializable because
5667 the additional locking could be a performance killer, and
5668 md should be large enough anyway.
5671 *) New file apps/app_rand.c with commonly needed functionality
5672 for handling the random seed file.
5674 Use the random seed file in some applications that previously did not:
5676 dsaparam -genkey (which also ignored its '-rand' option),
5679 x509 (when signing).
5680 Except on systems with /dev/urandom, it is crucial to have a random
5681 seed file at least for key creation, DSA signing, and for DH exchanges;
5682 for RSA signatures we could do without one.
5684 gendh and gendsa (unlike genrsa) used to read only the first byte
5685 of each file listed in the '-rand' option. The function as previously
5686 found in genrsa is now in app_rand.c and is used by all programs
5687 that support '-rand'.
5690 *) In RAND_write_file, use mode 0600 for creating files;
5691 don't just chmod when it may be too late.
5694 *) Report an error from X509_STORE_load_locations
5695 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5698 *) New function ASN1_mbstring_copy() this copies a string in either
5699 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5700 into an ASN1_STRING type. A mask of permissible types is passed
5701 and it chooses the "minimal" type to use or an error if not type
5705 *) Add function equivalents to the various macros in asn1.h. The old
5706 macros are retained with an M_ prefix. Code inside the library can
5707 use the M_ macros. External code (including the openssl utility)
5708 should *NOT* in order to be "shared library friendly".
5711 *) Add various functions that can check a certificate's extensions
5712 to see if it usable for various purposes such as SSL client,
5713 server or S/MIME and CAs of these types. This is currently
5714 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5715 verification. Also added a -purpose flag to x509 utility to
5716 print out all the purposes.
5719 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5723 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5724 for, obtain and decode and extension and obtain its critical flag.
5725 This allows all the necessary extension code to be handled in a
5726 single function call.
5729 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5730 platforms. See crypto/rc4/rc4_enc.c for further details.
5733 *) New -noout option to asn1parse. This causes no output to be produced
5734 its main use is when combined with -strparse and -out to extract data
5735 from a file (which may not be in ASN.1 format).
5738 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5739 when producing the local key id.
5740 [Richard Levitte <levitte@stacken.kth.se>]
5742 *) New option -dhparam in s_server. This allows a DH parameter file to be
5743 stated explicitly. If it is not stated then it tries the first server
5744 certificate file. The previous behaviour hard coded the filename
5748 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5749 a public key to be input or output. For example:
5750 openssl rsa -in key.pem -pubout -out pubkey.pem
5751 Also added necessary DSA public key functions to handle this.
5754 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5755 in the message. This was handled by allowing
5756 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5757 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5759 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5760 to the end of the strings whereas this didn't. This would cause problems
5761 if strings read with d2i_ASN1_bytes() were later modified.
5762 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5764 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5765 data and it contains EOF it will end up returning an error. This is
5766 caused by input 46 bytes long. The cause is due to the way base64
5767 BIOs find the start of base64 encoded data. They do this by trying a
5768 trial decode on each line until they find one that works. When they
5769 do a flag is set and it starts again knowing it can pass all the
5770 data directly through the decoder. Unfortunately it doesn't reset
5771 the context it uses. This means that if EOF is reached an attempt
5772 is made to pass two EOFs through the context and this causes the
5773 resulting error. This can also cause other problems as well. As is
5774 usual with these problems it takes *ages* to find and the fix is
5775 trivial: move one line.
5776 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5778 *) Ugly workaround to get s_client and s_server working under Windows. The
5779 old code wouldn't work because it needed to select() on sockets and the
5780 tty (for keypresses and to see if data could be written). Win32 only
5781 supports select() on sockets so we select() with a 1s timeout on the
5782 sockets and then see if any characters are waiting to be read, if none
5783 are present then we retry, we also assume we can always write data to
5784 the tty. This isn't nice because the code then blocks until we've
5785 received a complete line of data and it is effectively polling the
5786 keyboard at 1s intervals: however it's quite a bit better than not
5787 working at all :-) A dedicated Windows application might handle this
5788 with an event loop for example.
5791 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5792 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5793 will be called when RSA_sign() and RSA_verify() are used. This is useful
5794 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5795 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5796 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5797 This necessitated the support of an extra signature type NID_md5_sha1
5798 for SSL signatures and modifications to the SSL library to use it instead
5799 of calling RSA_public_decrypt() and RSA_private_encrypt().
5802 *) Add new -verify -CAfile and -CApath options to the crl program, these
5803 will lookup a CRL issuers certificate and verify the signature in a
5804 similar way to the verify program. Tidy up the crl program so it
5805 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5806 less strict. It will now permit CRL extensions even if it is not
5807 a V2 CRL: this will allow it to tolerate some broken CRLs.
5810 *) Initialize all non-automatic variables each time one of the openssl
5811 sub-programs is started (this is necessary as they may be started
5812 multiple times from the "OpenSSL>" prompt).
5813 [Lennart Bang, Bodo Moeller]
5815 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5816 removing all other RSA functionality (this is what NO_RSA does). This
5817 is so (for example) those in the US can disable those operations covered
5818 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5822 *) Non-copying interface to BIO pairs.
5823 (still largely untested)
5826 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5827 ASCII string. This was handled independently in various places before.
5830 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5831 UTF8 strings a character at a time.
5834 *) Use client_version from client hello to select the protocol
5835 (s23_srvr.c) and for RSA client key exchange verification
5836 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5839 *) Add various utility functions to handle SPKACs, these were previously
5840 handled by poking round in the structure internals. Added new function
5841 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5842 print, verify and generate SPKACs. Based on an original idea from
5843 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5846 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5849 *) Allow the config file extension section to be overwritten on the
5850 command line. Based on an original idea from Massimiliano Pala
5851 <madwolf@comune.modena.it>. The new option is called -extensions
5852 and can be applied to ca, req and x509. Also -reqexts to override
5853 the request extensions in req and -crlexts to override the crl extensions
5857 *) Add new feature to the SPKAC handling in ca. Now you can include
5858 the same field multiple times by preceding it by "XXXX." for example:
5861 this is the same syntax as used in the req config file.
5864 *) Allow certificate extensions to be added to certificate requests. These
5865 are specified in a 'req_extensions' option of the req section of the
5866 config file. They can be printed out with the -text option to req but
5867 are otherwise ignored at present.
5870 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5871 data read consists of only the final block it would not decrypted because
5872 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5873 A misplaced 'break' also meant the decrypted final block might not be
5874 copied until the next read.
5877 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5878 a few extra parameters to the DH structure: these will be useful if
5879 for example we want the value of 'q' or implement X9.42 DH.
5882 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5883 provides hooks that allow the default DSA functions or functions on a
5884 "per key" basis to be replaced. This allows hardware acceleration and
5885 hardware key storage to be handled without major modification to the
5886 library. Also added low level modexp hooks and CRYPTO_EX structure and
5887 associated functions.
5890 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5891 as "read only": it can't be written to and the buffer it points to will
5892 not be freed. Reading from a read only BIO is much more efficient than
5893 a normal memory BIO. This was added because there are several times when
5894 an area of memory needs to be read from a BIO. The previous method was
5895 to create a memory BIO and write the data to it, this results in two
5896 copies of the data and an O(n^2) reading algorithm. There is a new
5897 function BIO_new_mem_buf() which creates a read only memory BIO from
5898 an area of memory. Also modified the PKCS#7 routines to use read only
5902 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5903 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5904 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5905 but a retry condition occured while trying to read the rest.
5908 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5909 NID_pkcs7_encrypted by default: this was wrong since this should almost
5910 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5911 the encrypted data type: this is a more sensible place to put it and it
5912 allows the PKCS#12 code to be tidied up that duplicated this
5916 *) Changed obj_dat.pl script so it takes its input and output files on
5917 the command line. This should avoid shell escape redirection problems
5921 *) Initial support for certificate extension requests, these are included
5922 in things like Xenroll certificate requests. Included functions to allow
5923 extensions to be obtained and added.
5926 *) -crlf option to s_client and s_server for sending newlines as
5927 CRLF (as required by many protocols).
5930 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5932 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5933 [Ralf S. Engelschall]
5935 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5936 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5938 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5942 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5943 DH parameters/keys (q is lost during that conversion, but the resulting
5944 DH parameters contain its length).
5946 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5947 much faster than DH_generate_parameters (which creates parameters
5948 where p = 2*q + 1), and also the smaller q makes DH computations
5949 much more efficient (160-bit exponentiation instead of 1024-bit
5950 exponentiation); so this provides a convenient way to support DHE
5951 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5952 utter importance to use
5953 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5955 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5956 when such DH parameters are used, because otherwise small subgroup
5957 attacks may become possible!
5960 *) Avoid memory leak in i2d_DHparams.
5963 *) Allow the -k option to be used more than once in the enc program:
5964 this allows the same encrypted message to be read by multiple recipients.
5967 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5968 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5969 it will always use the numerical form of the OID, even if it has a short
5973 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5974 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5975 otherwise bn_mod_exp was called. In the case of hardware keys for example
5976 no private key components need be present and it might store extra data
5977 in the RSA structure, which cannot be accessed from bn_mod_exp.
5978 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5979 private key operations.
5982 *) Added support for SPARC Linux.
5985 *) pem_password_cb function type incompatibly changed from
5986 typedef int pem_password_cb(char *buf, int size, int rwflag);
5988 ....(char *buf, int size, int rwflag, void *userdata);
5989 so that applications can pass data to their callbacks:
5990 The PEM[_ASN1]_{read,write}... functions and macros now take an
5991 additional void * argument, which is just handed through whenever
5992 the password callback is called.
5993 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5995 New function SSL_CTX_set_default_passwd_cb_userdata.
5997 Compatibility note: As many C implementations push function arguments
5998 onto the stack in reverse order, the new library version is likely to
5999 interoperate with programs that have been compiled with the old
6000 pem_password_cb definition (PEM_whatever takes some data that
6001 happens to be on the stack as its last argument, and the callback
6002 just ignores this garbage); but there is no guarantee whatsoever that
6005 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6006 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6007 problems not only on Windows, but also on some Unix platforms.
6008 To avoid problematic command lines, these definitions are now in an
6009 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6010 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6013 *) MIPS III/IV assembler module is reimplemented.
6016 *) More DES library cleanups: remove references to srand/rand and
6017 delete an unused file.
6020 *) Add support for the the free Netwide assembler (NASM) under Win32,
6021 since not many people have MASM (ml) and it can be hard to obtain.
6022 This is currently experimental but it seems to work OK and pass all
6023 the tests. Check out INSTALL.W32 for info.
6026 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6027 without temporary keys kept an extra copy of the server key,
6028 and connections with temporary keys did not free everything in case
6032 *) New function RSA_check_key and new openssl rsa option -check
6033 for verifying the consistency of RSA keys.
6034 [Ulf Moeller, Bodo Moeller]
6036 *) Various changes to make Win32 compile work:
6037 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6038 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6039 comparison" warnings.
6040 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6043 *) Add a debugging option to PKCS#5 v2 key generation function: when
6044 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6045 derived keys are printed to stderr.
6048 *) Copy the flags in ASN1_STRING_dup().
6049 [Roman E. Pavlov <pre@mo.msk.ru>]
6051 *) The x509 application mishandled signing requests containing DSA
6052 keys when the signing key was also DSA and the parameters didn't match.
6054 It was supposed to omit the parameters when they matched the signing key:
6055 the verifying software was then supposed to automatically use the CA's
6056 parameters if they were absent from the end user certificate.
6058 Omitting parameters is no longer recommended. The test was also
6059 the wrong way round! This was probably due to unusual behaviour in
6060 EVP_cmp_parameters() which returns 1 if the parameters match.
6061 This meant that parameters were omitted when they *didn't* match and
6062 the certificate was useless. Certificates signed with 'ca' didn't have
6064 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6066 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6067 The interface is as follows:
6068 Applications can use
6069 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6070 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6071 "off" is now the default.
6072 The library internally uses
6073 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6074 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6075 to disable memory-checking temporarily.
6077 Some inconsistent states that previously were possible (and were
6078 even the default) are now avoided.
6080 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6081 with each memory chunk allocated; this is occasionally more helpful
6082 than just having a counter.
6084 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6086 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6090 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6091 which largely parallels "options", but is for changing API behaviour,
6092 whereas "options" are about protocol behaviour.
6093 Initial "mode" flags are:
6095 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6096 a single record has been written.
6097 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6098 retries use the same buffer location.
6099 (But all of the contents must be
6103 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6106 *) Fix problems with no-hmac etc.
6107 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6109 *) New functions RSA_get_default_method(), RSA_set_method() and
6110 RSA_get_method(). These allows replacement of RSA_METHODs without having
6111 to mess around with the internals of an RSA structure.
6114 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6115 Also really enable memory leak checks in openssl.c and in some
6117 [Chad C. Mulligan, Bodo Moeller]
6119 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6120 up the length of negative integers. This has now been simplified to just
6121 store the length when it is first determined and use it later, rather
6122 than trying to keep track of where data is copied and updating it to
6124 [Steve Henson, reported by Brien Wheeler
6125 <bwheeler@authentica-security.com>]
6127 *) Add a new function PKCS7_signatureVerify. This allows the verification
6128 of a PKCS#7 signature but with the signing certificate passed to the
6129 function itself. This contrasts with PKCS7_dataVerify which assumes the
6130 certificate is present in the PKCS#7 structure. This isn't always the
6131 case: certificates can be omitted from a PKCS#7 structure and be
6132 distributed by "out of band" means (such as a certificate database).
6135 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6136 function prototypes in pem.h, also change util/mkdef.pl to add the
6137 necessary function names.
6140 *) mk1mf.pl (used by Windows builds) did not properly read the
6141 options set by Configure in the top level Makefile, and Configure
6142 was not even able to write more than one option correctly.
6143 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6146 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6147 file to be loaded from a BIO or FILE pointer. The BIO version will
6148 for example allow memory BIOs to contain config info.
6151 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6152 Whoever hopes to achieve shared-library compatibility across versions
6153 must use this, not the compile-time macro.
6154 (Exercise 0.9.4: Which is the minimum library version required by
6156 Note: All this applies only to multi-threaded programs, others don't
6160 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6161 through a BIO pair triggered the default case, i.e.
6162 SSLerr(...,SSL_R_UNKNOWN_STATE).
6165 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6166 can use the SSL library even if none of the specific BIOs is
6170 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6171 for the encoded length.
6172 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6174 *) Add initial documentation of the X509V3 functions.
6177 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6178 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6179 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6180 secure PKCS#8 private key format with a high iteration count.
6183 *) Fix determination of Perl interpreter: A perl or perl5
6184 _directory_ in $PATH was also accepted as the interpreter.
6185 [Ralf S. Engelschall]
6187 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6188 wrong with it but it was very old and did things like calling
6189 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6193 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6194 to use the new extension code.
6197 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6198 with macros. This should make it easier to change their form, add extra
6199 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6203 *) Add to configuration table a new entry that can specify an alternative
6204 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6205 according to Mark Crispin <MRC@Panda.COM>.
6209 *) DES CBC did not update the IV. Weird.
6212 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6213 Changing the behaviour of the former might break existing programs --
6214 where IV updating is needed, des_ncbc_encrypt can be used.
6217 *) When bntest is run from "make test" it drives bc to check its
6218 calculations, as well as internally checking them. If an internal check
6219 fails, it needs to cause bc to give a non-zero result or make test carries
6220 on without noticing the failure. Fixed.
6223 *) DES library cleanups.
6226 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6227 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6228 ciphers. NOTE: although the key derivation function has been verified
6229 against some published test vectors it has not been extensively tested
6230 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6234 *) Instead of "mkdir -p", which is not fully portable, use new
6235 Perl script "util/mkdir-p.pl".
6238 *) Rewrite the way password based encryption (PBE) is handled. It used to
6239 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6240 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6241 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6242 the 'parameter' field of the AlgorithmIdentifier is passed to the
6243 underlying key generation function so it must do its own ASN1 parsing.
6244 This has also changed the EVP_PBE_CipherInit() function which now has a
6245 'parameter' argument instead of literal salt and iteration count values
6246 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6249 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6250 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6251 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6252 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6253 value was just used as a "magic string" and not used directly its
6254 value doesn't matter.
6257 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6261 *) "linux-sparc64" configuration (ultrapenguin).
6262 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6263 "linux-sparc" configuration.
6264 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6266 *) config now generates no-xxx options for missing ciphers.
6269 *) Support the EBCDIC character set (work in progress).
6270 File ebcdic.c not yet included because it has a different license.
6271 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6273 *) Support BS2000/OSD-POSIX.
6274 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6276 *) Make callbacks for key generation use void * instead of char *.
6279 *) Make S/MIME samples compile (not yet tested).
6282 *) Additional typesafe stacks.
6285 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6289 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6291 *) New configuration variant "sco5-gcc".
6293 *) Updated some demos.
6294 [Sean O Riordain, Wade Scholine]
6296 *) Add missing BIO_free at exit of pkcs12 application.
6299 *) Fix memory leak in conf.c.
6302 *) Updates for Win32 to assembler version of MD5.
6305 *) Set #! path to perl in apps/der_chop to where we found it
6306 instead of using a fixed path.
6309 *) SHA library changes for irix64-mips4-cc.
6312 *) Improvements for VMS support.
6316 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6318 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6319 This also avoids the problems with SC4.2 and unpatched SC5.
6320 [Andy Polyakov <appro@fy.chalmers.se>]
6322 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6323 These are required because of the typesafe stack would otherwise break
6324 existing code. If old code used a structure member which used to be STACK
6325 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6326 sk_num or sk_value it would produce an error because the num, data members
6327 are not present in STACK_OF. Now it just produces a warning. sk_set
6328 replaces the old method of assigning a value to sk_value
6329 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6330 that does this will no longer work (and should use sk_set instead) but
6331 this could be regarded as a "questionable" behaviour anyway.
6334 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6335 correctly handle encrypted S/MIME data.
6338 *) Change type of various DES function arguments from des_cblock
6339 (which means, in function argument declarations, pointer to char)
6340 to des_cblock * (meaning pointer to array with 8 char elements),
6341 which allows the compiler to do more typechecking; it was like
6342 that back in SSLeay, but with lots of ugly casts.
6344 Introduce new type const_des_cblock.
6347 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6348 problems: find RecipientInfo structure that matches recipient certificate
6349 and initialise the ASN1 structures properly based on passed cipher.
6352 *) Belatedly make the BN tests actually check the results.
6355 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6356 to and from BNs: it was completely broken. New compilation option
6357 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6358 key elements as negative integers.
6361 *) Reorganize and speed up MD5.
6362 [Andy Polyakov <appro@fy.chalmers.se>]
6365 [Richard Levitte <richard@levitte.org>]
6367 *) New option -out to asn1parse to allow the parsed structure to be
6368 output to a file. This is most useful when combined with the -strparse
6369 option to examine the output of things like OCTET STRINGS.
6372 *) Make SSL library a little more fool-proof by not requiring any longer
6373 that SSL_set_{accept,connect}_state be called before
6374 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6375 in many applications because usually everything *appeared* to work as
6376 intended anyway -- now it really works as intended).
6379 *) Move openssl.cnf out of lib/.
6382 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6383 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6384 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6385 [Ralf S. Engelschall]
6387 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6388 handle PKCS#7 enveloped data properly.
6389 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6391 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6392 copying pointers. The cert_st handling is changed by this in
6393 various ways (and thus what used to be known as ctx->default_cert
6394 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6395 any longer when s->cert does not give us what we need).
6396 ssl_cert_instantiate becomes obsolete by this change.
6397 As soon as we've got the new code right (possibly it already is?),
6398 we have solved a couple of bugs of the earlier code where s->cert
6399 was used as if it could not have been shared with other SSL structures.
6401 Note that using the SSL API in certain dirty ways now will result
6402 in different behaviour than observed with earlier library versions:
6403 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6404 does not influence s as it used to.
6406 In order to clean up things more thoroughly, inside SSL_SESSION
6407 we don't use CERT any longer, but a new structure SESS_CERT
6408 that holds per-session data (if available); currently, this is
6409 the peer's certificate chain and, for clients, the server's certificate
6410 and temporary key. CERT holds only those values that can have
6411 meaningful defaults in an SSL_CTX.
6414 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6415 from the internal representation. Various PKCS#7 fixes: remove some
6416 evil casts and set the enc_dig_alg field properly based on the signing
6420 *) Allow PKCS#12 password to be set from the command line or the
6421 environment. Let 'ca' get its config file name from the environment
6422 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6426 *) Allow certificate policies extension to use an IA5STRING for the
6427 organization field. This is contrary to the PKIX definition but
6428 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6432 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6433 without disallowing inline assembler and the like for non-pedantic builds.
6436 *) Support Borland C++ builder.
6437 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6442 *) SHA-1 cleanups and performance enhancements.
6443 [Andy Polyakov <appro@fy.chalmers.se>]
6445 *) Sparc v8plus assembler for the bignum library.
6446 [Andy Polyakov <appro@fy.chalmers.se>]
6448 *) Accept any -xxx and +xxx compiler options in Configure.
6451 *) Update HPUX configuration.
6454 *) Add missing sk_<type>_unshift() function to safestack.h
6455 [Ralf S. Engelschall]
6457 *) New function SSL_CTX_use_certificate_chain_file that sets the
6458 "extra_cert"s in addition to the certificate. (This makes sense
6459 only for "PEM" format files, as chains as a whole are not
6463 *) Support verify_depth from the SSL API.
6464 x509_vfy.c had what can be considered an off-by-one-error:
6465 Its depth (which was not part of the external interface)
6466 was actually counting the number of certificates in a chain;
6467 now it really counts the depth.
6470 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6471 instead of X509err, which often resulted in confusing error
6472 messages since the error codes are not globally unique
6473 (e.g. an alleged error in ssl3_accept when a certificate
6474 didn't match the private key).
6476 *) New function SSL_CTX_set_session_id_context that allows to set a default
6477 value (so that you don't need SSL_set_session_id_context for each
6478 connection using the SSL_CTX).
6481 *) OAEP decoding bug fix.
6484 *) Support INSTALL_PREFIX for package builders, as proposed by
6488 *) New Configure options "threads" and "no-threads". For systems
6489 where the proper compiler options are known (currently Solaris
6490 and Linux), "threads" is the default.
6493 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6496 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6497 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6498 such as /usr/local/bin.
6501 *) "make linux-shared" to build shared libraries.
6502 [Niels Poppe <niels@netbox.org>]
6504 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6507 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6508 extension adding in x509 utility.
6511 *) Remove NOPROTO sections and error code comments.
6514 *) Partial rewrite of the DEF file generator to now parse the ANSI
6518 *) New Configure options --prefix=DIR and --openssldir=DIR.
6521 *) Complete rewrite of the error code script(s). It is all now handled
6522 by one script at the top level which handles error code gathering,
6523 header rewriting and C source file generation. It should be much better
6524 than the old method: it now uses a modified version of Ulf's parser to
6525 read the ANSI prototypes in all header files (thus the old K&R definitions
6526 aren't needed for error creation any more) and do a better job of
6527 translating function codes into names. The old 'ASN1 error code imbedded
6528 in a comment' is no longer necessary and it doesn't use .err files which
6529 have now been deleted. Also the error code call doesn't have to appear all
6530 on one line (which resulted in some large lines...).
6533 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6536 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6537 0 (which usually indicates a closed connection), but continue reading.
6540 *) Fix some race conditions.
6543 *) Add support for CRL distribution points extension. Add Certificate
6544 Policies and CRL distribution points documentation.
6547 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6550 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6551 8 of keying material. Merlin has also confirmed interop with this fix
6552 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6553 [Merlin Hughes <merlin@baltimore.ie>]
6555 *) Fix lots of warnings.
6556 [Richard Levitte <levitte@stacken.kth.se>]
6558 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6559 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6560 [Richard Levitte <levitte@stacken.kth.se>]
6562 *) Fix problems with sizeof(long) == 8.
6563 [Andy Polyakov <appro@fy.chalmers.se>]
6565 *) Change functions to ANSI C.
6568 *) Fix typos in error codes.
6569 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6571 *) Remove defunct assembler files from Configure.
6574 *) SPARC v8 assembler BIGNUM implementation.
6575 [Andy Polyakov <appro@fy.chalmers.se>]
6577 *) Support for Certificate Policies extension: both print and set.
6578 Various additions to support the r2i method this uses.
6581 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6582 return a const string when you are expecting an allocated buffer.
6585 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6586 types DirectoryString and DisplayText.
6589 *) Add code to allow r2i extensions to access the configuration database,
6590 add an LHASH database driver and add several ctx helper functions.
6593 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6594 fail when they extended the size of a BIGNUM.
6597 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6598 support typesafe stack.
6601 *) Fix typo in SSL_[gs]et_options().
6602 [Nils Frostberg <nils@medcom.se>]
6604 *) Delete various functions and files that belonged to the (now obsolete)
6605 old X509V3 handling code.
6608 *) New Configure option "rsaref".
6611 *) Don't auto-generate pem.h.
6614 *) Introduce type-safe ASN.1 SETs.
6617 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6618 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6620 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6621 that links with OpenSSL (well at least cause lots of warnings), but fear
6622 not: the conversion is trivial, and it eliminates loads of evil casts. A
6623 few STACKed things have been converted already. Feel free to convert more.
6624 In the fullness of time, I'll do away with the STACK type altogether.
6627 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6628 specified in <certfile> by updating the entry in the index.txt file.
6629 This way one no longer has to edit the index.txt file manually for
6630 revoking a certificate. The -revoke option does the gory details now.
6631 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6633 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6634 `-text' option at all and this way the `-noout -text' combination was
6635 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6636 [Ralf S. Engelschall]
6638 *) Make sure a corresponding plain text error message exists for the
6639 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6640 verify callback function determined that a certificate was revoked.
6641 [Ralf S. Engelschall]
6643 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6644 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6645 all available cipers including rc5, which was forgotten until now.
6646 In order to let the testing shell script know which algorithms
6647 are available, a new (up to now undocumented) command
6648 "openssl list-cipher-commands" is used.
6651 *) Bugfix: s_client occasionally would sleep in select() when
6652 it should have checked SSL_pending() first.
6655 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6656 the raw DSA values prior to ASN.1 encoding.
6659 *) Tweaks to Configure
6660 [Niels Poppe <niels@netbox.org>]
6662 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6666 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6669 *) New config option to avoid instructions that are illegal on the 80386.
6670 The default code is faster, but requires at least a 486.
6673 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6674 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6675 same as SSL2_VERSION anyway.
6678 *) New "-showcerts" option for s_client.
6681 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6682 application. Various cleanups and fixes.
6685 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6686 modify error routines to work internally. Add error codes and PBE init
6687 to library startup routines.
6690 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6691 packing functions to asn1 and evp. Changed function names and error
6692 codes along the way.
6695 *) PKCS12 integration: and so it begins... First of several patches to
6696 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6697 objects to objects.h
6700 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6701 and display support for Thawte strong extranet extension.
6704 *) Add LinuxPPC support.
6705 [Jeff Dubrule <igor@pobox.org>]
6707 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6708 bn_div_words in alpha.s.
6709 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6711 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6712 OAEP isn't supported when OpenSSL is built with RSAref.
6713 [Ulf Moeller <ulf@fitug.de>]
6715 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6716 so they no longer are missing under -DNOPROTO.
6717 [Soren S. Jorvang <soren@t.dk>]
6720 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6722 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6723 doesn't work when the session is reused. Coming soon!
6726 *) Fix a security hole, that allows sessions to be reused in the wrong
6727 context thus bypassing client cert protection! All software that uses
6728 client certs and session caches in multiple contexts NEEDS PATCHING to
6729 allow session reuse! A fuller solution is in the works.
6730 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6732 *) Some more source tree cleanups (removed obsolete files
6733 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6734 permission on "config" script to be executable) and a fix for the INSTALL
6736 [Ulf Moeller <ulf@fitug.de>]
6738 *) Remove some legacy and erroneous uses of malloc, free instead of
6740 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6742 *) Make rsa_oaep_test return non-zero on error.
6743 [Ulf Moeller <ulf@fitug.de>]
6745 *) Add support for native Solaris shared libraries. Configure
6746 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6747 if someone would make that last step automatic.
6748 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6750 *) ctx_size was not built with the right compiler during "make links". Fixed.
6753 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6754 except NULL ciphers". This means the default cipher list will no longer
6755 enable NULL ciphers. They need to be specifically enabled e.g. with
6756 the string "DEFAULT:eNULL".
6759 *) Fix to RSA private encryption routines: if p < q then it would
6760 occasionally produce an invalid result. This will only happen with
6761 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6764 *) Be less restrictive and allow also `perl util/perlpath.pl
6765 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6766 because this way one can also use an interpreter named `perl5' (which is
6767 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6768 installed as `perl').
6769 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6771 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6772 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6774 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6775 advapi32.lib to Win32 build and change the pem test comparision
6776 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6777 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6778 and crypto/des/ede_cbcm_enc.c.
6781 *) DES quad checksum was broken on big-endian architectures. Fixed.
6784 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6785 Win32 test batch file so it (might) work again. The Win32 test batch file
6786 is horrible: I feel ill....
6789 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6790 in e_os.h. Audit of header files to check ANSI and non ANSI
6791 sections: 10 functions were absent from non ANSI section and not exported
6792 from Windows DLLs. Fixed up libeay.num for new functions.
6795 *) Make `openssl version' output lines consistent.
6796 [Ralf S. Engelschall]
6798 *) Fix Win32 symbol export lists for BIO functions: Added
6799 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6800 to ms/libeay{16,32}.def.
6801 [Ralf S. Engelschall]
6803 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6804 fine under Unix and passes some trivial tests I've now added. But the
6805 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6806 added to make sure no one expects that this stuff really works in the
6807 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6808 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6810 [Ralf S. Engelschall]
6812 *) Fix the generation of two part addresses in perl.
6813 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6815 *) Add config entry for Linux on MIPS.
6816 [John Tobey <jtobey@channel1.com>]
6818 *) Make links whenever Configure is run, unless we are on Windoze.
6821 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6822 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6826 *) Add a useful kludge to allow package maintainers to specify compiler and
6827 other platforms details on the command line without having to patch the
6828 Configure script everytime: One now can use ``perl Configure
6829 <id>:<details>'', i.e. platform ids are allowed to have details appended
6830 to them (seperated by colons). This is treated as there would be a static
6831 pre-configured entry in Configure's %table under key <id> with value
6832 <details> and ``perl Configure <id>'' is called. So, when you want to
6833 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6834 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6835 now, which overrides the FreeBSD-elf entry on-the-fly.
6836 [Ralf S. Engelschall]
6838 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6841 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6842 on the `perl Configure ...' command line. This way one can compile
6843 OpenSSL libraries with Position Independent Code (PIC) which is needed
6844 for linking it into DSOs.
6845 [Ralf S. Engelschall]
6847 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6851 *) Cleaned up the LICENSE document: The official contact for any license
6852 questions now is the OpenSSL core team under openssl-core@openssl.org.
6853 And add a paragraph about the dual-license situation to make sure people
6854 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6855 to the OpenSSL toolkit.
6856 [Ralf S. Engelschall]
6858 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6859 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6860 Additonally cleaned up the `make links' target: Remove unnecessary
6861 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6862 to speed processing and no longer clutter the display with confusing
6863 stuff. Instead only the actually done links are displayed.
6864 [Ralf S. Engelschall]
6866 *) Permit null encryption ciphersuites, used for authentication only. It used
6867 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6868 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6872 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6873 signed attributes when verifying signatures (this would break them),
6874 the detached data encoding was wrong and public keys obtained using
6875 X509_get_pubkey() weren't freed.
6878 *) Add text documentation for the BUFFER functions. Also added a work around
6879 to a Win95 console bug. This was triggered by the password read stuff: the
6880 last character typed gets carried over to the next fread(). If you were
6881 generating a new cert request using 'req' for example then the last
6882 character of the passphrase would be CR which would then enter the first
6886 *) Added the new `Includes OpenSSL Cryptography Software' button as
6887 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6888 button and can be used by applications based on OpenSSL to show the
6889 relationship to the OpenSSL project.
6890 [Ralf S. Engelschall]
6892 *) Remove confusing variables in function signatures in files
6893 ssl/ssl_lib.c and ssl/ssl.h.
6894 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6896 *) Don't install bss_file.c under PREFIX/include/
6897 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6899 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6900 functions that return function pointers and has support for NT specific
6901 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6902 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6903 unsigned to signed types: this was killing the Win32 compile.
6906 *) Add new certificate file to stack functions,
6907 SSL_add_dir_cert_subjects_to_stack() and
6908 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6909 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6910 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6911 This means that Apache-SSL and similar packages don't have to mess around
6912 to add as many CAs as they want to the preferred list.
6915 *) Experiment with doxygen documentation. Currently only partially applied to
6917 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6918 openssl.doxy as the configuration file.
6921 *) Get rid of remaining C++-style comments which strict C compilers hate.
6922 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6924 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6925 compiled in by default: it has problems with large keys.
6928 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6929 DH private keys and/or callback functions which directly correspond to
6930 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6931 is needed for applications which have to configure certificates on a
6932 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6934 For the RSA certificate situation is makes no difference, but
6935 for the DSA certificate situation this fixes the "no shared cipher"
6936 problem where the OpenSSL cipher selection procedure failed because the
6937 temporary keys were not overtaken from the context and the API provided
6938 no way to reconfigure them.
6939 The new functions now let applications reconfigure the stuff and they
6940 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6941 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6942 non-public-API function ssl_cert_instantiate() is used as a helper
6943 function and also to reduce code redundancy inside ssl_rsa.c.
6944 [Ralf S. Engelschall]
6946 *) Move s_server -dcert and -dkey options out of the undocumented feature
6947 area because they are useful for the DSA situation and should be
6948 recognized by the users.
6949 [Ralf S. Engelschall]
6951 *) Fix the cipher decision scheme for export ciphers: the export bits are
6952 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6953 SSL_EXP_MASK. So, the original variable has to be used instead of the
6954 already masked variable.
6955 [Richard Levitte <levitte@stacken.kth.se>]
6957 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6958 [Richard Levitte <levitte@stacken.kth.se>]
6960 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6961 from `int' to `unsigned int' because it's a length and initialized by
6962 EVP_DigestFinal() which expects an `unsigned int *'.
6963 [Richard Levitte <levitte@stacken.kth.se>]
6965 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6966 script. Instead use the usual Shell->Perl transition trick.
6967 [Ralf S. Engelschall]
6969 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6970 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6971 -noout -modulus' as it's already the case for `openssl rsa -noout
6972 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6973 currently the public key is printed (a decision which was already done by
6974 `openssl dsa -modulus' in the past) which serves a similar purpose.
6975 Additionally the NO_RSA no longer completely removes the whole -modulus
6976 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6978 [Ralf S. Engelschall]
6980 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6981 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6982 [Arne Ansper <arne@ats.cyber.ee>]
6984 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6985 to be added. Now both 'req' and 'ca' can use new objects defined in the
6989 *) Add cool BIO that does syslog (or event log on NT).
6990 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6992 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6993 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6994 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6995 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6998 *) Add preliminary config info for new extension code.
7001 *) Make RSA_NO_PADDING really use no padding.
7002 [Ulf Moeller <ulf@fitug.de>]
7004 *) Generate errors when private/public key check is done.
7007 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7008 for some CRL extensions and new objects added.
7011 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7012 key usage extension and fuller support for authority key id.
7015 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7016 padding method for RSA, which is recommended for new applications in PKCS
7017 #1 v2.0 (RFC 2437, October 1998).
7018 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7019 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7020 against Bleichbacher's attack on RSA.
7021 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7024 *) Updates to the new SSL compression code
7025 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7027 *) Fix so that the version number in the master secret, when passed
7028 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7029 (because the server will not accept higher), that the version number
7030 is 0x03,0x01, not 0x03,0x00
7031 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7033 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7034 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7035 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7038 *) Support for RAW extensions where an arbitrary extension can be
7039 created by including its DER encoding. See apps/openssl.cnf for
7043 *) Make sure latest Perl versions don't interpret some generated C array
7044 code as Perl array code in the crypto/err/err_genc.pl script.
7045 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7047 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7048 not many people have the assembler. Various Win32 compilation fixes and
7049 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7053 *) Modify configure script 'Configure' to automatically create crypto/date.h
7054 file under Win32 and also build pem.h from pem.org. New script
7055 util/mkfiles.pl to create the MINFO file on environments that can't do a
7056 'make files': perl util/mkfiles.pl >MINFO should work.
7059 *) Major rework of DES function declarations, in the pursuit of correctness
7060 and purity. As a result, many evil casts evaporated, and some weirdness,
7061 too. You may find this causes warnings in your code. Zapping your evil
7062 casts will probably fix them. Mostly.
7065 *) Fix for a typo in asn1.h. Bug fix to object creation script
7066 obj_dat.pl. It considered a zero in an object definition to mean
7067 "end of object": none of the objects in objects.h have any zeros
7068 so it wasn't spotted.
7069 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7071 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7072 Masking (CBCM). In the absence of test vectors, the best I have been able
7073 to do is check that the decrypt undoes the encrypt, so far. Send me test
7074 vectors if you have them.
7077 *) Correct calculation of key length for export ciphers (too much space was
7078 allocated for null ciphers). This has not been tested!
7081 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7082 message is now correct (it understands "crypto" and "ssl" on its
7083 command line). There is also now an "update" option. This will update
7084 the util/ssleay.num and util/libeay.num files with any new functions.
7086 perl util/mkdef.pl crypto ssl update
7087 it will update them.
7090 *) Overhauled the Perl interface (perl/*):
7091 - ported BN stuff to OpenSSL's different BN library
7092 - made the perl/ source tree CVS-aware
7093 - renamed the package from SSLeay to OpenSSL (the files still contain
7094 their history because I've copied them in the repository)
7095 - removed obsolete files (the test scripts will be replaced
7096 by better Test::Harness variants in the future)
7097 [Ralf S. Engelschall]
7099 *) First cut for a very conservative source tree cleanup:
7100 1. merge various obsolete readme texts into doc/ssleay.txt
7101 where we collect the old documents and readme texts.
7102 2. remove the first part of files where I'm already sure that we no
7103 longer need them because of three reasons: either they are just temporary
7104 files which were left by Eric or they are preserved original files where
7105 I've verified that the diff is also available in the CVS via "cvs diff
7106 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7107 the crypto/md/ stuff).
7108 [Ralf S. Engelschall]
7110 *) More extension code. Incomplete support for subject and issuer alt
7111 name, issuer and authority key id. Change the i2v function parameters
7112 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7113 what that's for :-) Fix to ASN1 macro which messed up
7114 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7117 *) Preliminary support for ENUMERATED type. This is largely copied from the
7121 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7122 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7124 *) Make sure `make rehash' target really finds the `openssl' program.
7125 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7127 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7128 like to hear about it if this slows down other processors.
7131 *) Add CygWin32 platform information to Configure script.
7132 [Alan Batie <batie@aahz.jf.intel.com>]
7134 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7135 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7137 *) New program nseq to manipulate netscape certificate sequences
7140 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7144 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7145 but the BN code had some problems that would cause failures when
7146 doing certificate verification and some other functions.
7147 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7149 *) Add ASN1 and PEM code to support netscape certificate sequences.
7152 *) Add ASN1 and PEM code to support netscape certificate sequences.
7155 *) Add several PKIX and private extended key usage OIDs.
7158 *) Modify the 'ca' program to handle the new extension code. Modify
7159 openssl.cnf for new extension format, add comments.
7162 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7163 and add a sample to openssl.cnf so req -x509 now adds appropriate
7167 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7168 error code, add initial support to X509_print() and x509 application.
7171 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7172 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7173 stuff is currently isolated and isn't even compiled yet.
7176 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7177 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7178 Removed the versions check from X509 routines when loading extensions:
7179 this allows certain broken certificates that don't set the version
7180 properly to be processed.
7183 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7184 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7185 can still be regenerated with "make depend".
7188 *) Spelling mistake in C version of CAST-128.
7189 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7191 *) Changes to the error generation code. The perl script err-code.pl
7192 now reads in the old error codes and retains the old numbers, only
7193 adding new ones if necessary. It also only changes the .err files if new
7194 codes are added. The makefiles have been modified to only insert errors
7195 when needed (to avoid needlessly modifying header files). This is done
7196 by only inserting errors if the .err file is newer than the auto generated
7197 C file. To rebuild all the error codes from scratch (the old behaviour)
7198 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7199 or delete all the .err files.
7202 *) CAST-128 was incorrectly implemented for short keys. The C version has
7203 been fixed, but is untested. The assembler versions are also fixed, but
7204 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7205 to regenerate it if needed.
7206 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7207 Hagino <itojun@kame.net>]
7209 *) File was opened incorrectly in randfile.c.
7210 [Ulf Möller <ulf@fitug.de>]
7212 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7213 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7214 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7215 al: it's just almost always a UTCTime. Note this patch adds new error
7216 codes so do a "make errors" if there are problems.
7219 *) Correct Linux 1 recognition in config.
7220 [Ulf Möller <ulf@fitug.de>]
7222 *) Remove pointless MD5 hash when using DSA keys in ca.
7223 [Anonymous <nobody@replay.com>]
7225 *) Generate an error if given an empty string as a cert directory. Also
7226 generate an error if handed NULL (previously returned 0 to indicate an
7227 error, but didn't set one).
7228 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7230 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7233 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7234 parameters. This was causing a warning which killed off the Win32 compile.
7237 *) Remove C++ style comments from crypto/bn/bn_local.h.
7238 [Neil Costigan <neil.costigan@celocom.com>]
7240 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7241 based on a text string, looking up short and long names and finally
7242 "dot" format. The "dot" format stuff didn't work. Added new function
7243 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7244 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7245 OID is not part of the table.
7248 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7249 X509_LOOKUP_by_alias().
7252 *) Sort openssl functions by name.
7255 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7256 encryption from sample DSA keys (in case anyone is interested the password
7260 *) Make _all_ *_free functions accept a NULL pointer.
7261 [Frans Heymans <fheymans@isaserver.be>]
7263 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7265 [Anonymous <nobody@replay.com>]
7267 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7268 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7270 *) Don't blow it for numeric -newkey arguments to apps/req.
7271 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7273 *) Temp key "for export" tests were wrong in s3_srvr.c.
7274 [Anonymous <nobody@replay.com>]
7276 *) Add prototype for temp key callback functions
7277 SSL_CTX_set_tmp_{rsa,dh}_callback().
7280 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7281 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7284 *) X509_name_add_entry() freed the wrong thing after an error.
7285 [Arne Ansper <arne@ats.cyber.ee>]
7287 *) rsa_eay.c would attempt to free a NULL context.
7288 [Arne Ansper <arne@ats.cyber.ee>]
7290 *) BIO_s_socket() had a broken should_retry() on Windoze.
7291 [Arne Ansper <arne@ats.cyber.ee>]
7293 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7294 [Arne Ansper <arne@ats.cyber.ee>]
7296 *) Make sure the already existing X509_STORE->depth variable is initialized
7297 in X509_STORE_new(), but document the fact that this variable is still
7298 unused in the certificate verification process.
7299 [Ralf S. Engelschall]
7301 *) Fix the various library and apps files to free up pkeys obtained from
7302 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7305 *) Fix reference counting in X509_PUBKEY_get(). This makes
7306 demos/maurice/example2.c work, amongst others, probably.
7307 [Steve Henson and Ben Laurie]
7309 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7310 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7311 are no longer created. This way we have a single and consistent command
7312 line interface `openssl <command>', similar to `cvs <command>'.
7313 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7315 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7316 BIT STRING wrapper always have zero unused bits.
7319 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7322 *) Make the top-level INSTALL documentation easier to understand.
7325 *) Makefiles updated to exit if an error occurs in a sub-directory
7326 make (including if user presses ^C) [Paul Sutton]
7328 *) Make Montgomery context stuff explicit in RSA data structure.
7331 *) Fix build order of pem and err to allow for generated pem.h.
7334 *) Fix renumbering bug in X509_NAME_delete_entry().
7337 *) Enhanced the err-ins.pl script so it makes the error library number
7338 global and can add a library name. This is needed for external ASN1 and
7339 other error libraries.
7342 *) Fixed sk_insert which never worked properly.
7345 *) Fix ASN1 macros so they can handle indefinite length construted
7346 EXPLICIT tags. Some non standard certificates use these: they can now
7350 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7351 into a single doc/ssleay.txt bundle. This way the information is still
7352 preserved but no longer messes up this directory. Now it's new room for
7353 the new set of documenation files.
7354 [Ralf S. Engelschall]
7356 *) SETs were incorrectly DER encoded. This was a major pain, because they
7357 shared code with SEQUENCEs, which aren't coded the same. This means that
7358 almost everything to do with SETs or SEQUENCEs has either changed name or
7359 number of arguments.
7360 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7362 *) Fix test data to work with the above.
7365 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7366 was already fixed by Eric for 0.9.1 it seems.
7367 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7369 *) Autodetect FreeBSD3.
7372 *) Fix various bugs in Configure. This affects the following platforms:
7376 unixware-2.0-pentium
7380 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7381 before they are needed.
7384 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7388 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7390 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7391 changed SSLeay to OpenSSL in version strings.
7392 [Ralf S. Engelschall]
7394 *) Some fixups to the top-level documents.
7397 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7398 because the symlink to include/ was missing.
7399 [Ralf S. Engelschall]
7401 *) Incorporated the popular no-RSA/DSA-only patches
7402 which allow to compile a RSA-free SSLeay.
7403 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7405 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7406 when "ssleay" is still not found.
7407 [Ralf S. Engelschall]
7409 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7410 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7412 *) Updated the README file.
7413 [Ralf S. Engelschall]
7415 *) Added various .cvsignore files in the CVS repository subdirs
7416 to make a "cvs update" really silent.
7417 [Ralf S. Engelschall]
7419 *) Recompiled the error-definition header files and added
7420 missing symbols to the Win32 linker tables.
7421 [Ralf S. Engelschall]
7423 *) Cleaned up the top-level documents;
7424 o new files: CHANGES and LICENSE
7425 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7426 o merged COPYRIGHT into LICENSE
7427 o removed obsolete TODO file
7428 o renamed MICROSOFT to INSTALL.W32
7429 [Ralf S. Engelschall]
7431 *) Removed dummy files from the 0.9.1b source tree:
7432 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7433 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7434 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7435 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7436 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7437 [Ralf S. Engelschall]
7439 *) Added various platform portability fixes.
7442 *) The Genesis of the OpenSSL rpject:
7443 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7444 Young and Tim J. Hudson created while they were working for C2Net until
7446 [The OpenSSL Project]
7449 Changes between 0.9.0b and 0.9.1b [not released]
7451 *) Updated a few CA certificates under certs/
7454 *) Changed some BIGNUM api stuff.
7457 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7458 DGUX x86, Linux Alpha, etc.
7461 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7462 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7466 *) Add -strparse option to asn1pars program which parses nested
7468 [Dr Stephen Henson <shenson@bigfoot.com>]
7470 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7473 *) DSA fix for "ca" program.
7476 *) Added "-genkey" option to "dsaparam" program.
7479 *) Added RIPE MD160 (rmd160) message digest.
7482 *) Added -a (all) option to "ssleay version" command.
7485 *) Added PLATFORM define which is the id given to Configure.
7488 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7491 *) Extended the ASN.1 parser routines.
7494 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7497 *) Added a BN_CTX to the BN library.
7500 *) Fixed the weak key values in DES library
7503 *) Changed API in EVP library for cipher aliases.
7506 *) Added support for RC2/64bit cipher.
7509 *) Converted the lhash library to the crypto/mem.c functions.
7512 *) Added more recognized ASN.1 object ids.
7515 *) Added more RSA padding checks for SSL/TLS.
7518 *) Added BIO proxy/filter functionality.
7521 *) Added extra_certs to SSL_CTX which can be used
7522 send extra CA certificates to the client in the CA cert chain sending
7523 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7526 *) Now Fortezza is denied in the authentication phase because
7527 this is key exchange mechanism is not supported by SSLeay at all.
7530 *) Additional PKCS1 checks.
7533 *) Support the string "TLSv1" for all TLS v1 ciphers.
7536 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7537 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7540 *) Fixed a few memory leaks.
7543 *) Fixed various code and comment typos.
7546 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7547 bytes sent in the client random.
7548 [Edward Bishop <ebishop@spyglass.com>]