5 Changes between 0.9.7e and 0.9.8 [xx XXX xxxx]
7 *) Change ./Configure so that certain algorithms can be disabled by default.
8 The new counterpiece to "no-xxx" is "enable-xxx".
10 The patented RC5 and MDC2 algorithms will now be disabled unless
11 "enable-rc5" and "enable-mdc2", respectively, are specified.
13 (IDEA remains enabled despite being patented. This is because IDEA
14 is frequently required for interoperability, and there is no license
15 fee for non-commercial use. As before, "no-idea" can be used to
16 avoid this algorithm.)
20 *) Add processing of proxy certificates (see RFC 3820). This work was
21 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
22 EGEE (Enabling Grids for E-science in Europe).
25 *) RC4 performance overhaul on modern architectures/implementations, such
26 as Intel P4, IA-64 and AMD64.
29 *) New utility extract-section.pl. This can be used specify an alternative
30 section number in a pod file instead of having to treat each file as
31 a separate case in Makefile. This can be done by adding two lines to the
34 =for comment openssl_section:XXX
36 The blank line is mandatory.
40 *) New arguments -certform, -keyform and -pass for s_client and s_server
41 to allow alternative format key and certificate files and passphrase
45 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
46 update associated structures and add various utility functions.
48 Add new policy related verify parameters, include policy checking in
49 standard verify code. Enhance 'smime' application with extra parameters
50 to support policy checking and print out.
53 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
54 Nehemiah processors. These extensions support AES encryption in hardware
55 as well as RNG (though RNG support is currently disabled).
56 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
58 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
61 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
62 [Andy Polyakov and a number of other people]
64 *) Improved PowerPC platform support. Most notably BIGNUM assembler
65 implementation contributed by IBM.
66 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
68 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
69 exponent rather than 'unsigned long'. There is a corresponding change to
70 the new 'rsa_keygen' element of the RSA_METHOD structure.
71 [Jelte Jansen, Geoff Thorpe]
73 *) Functionality for creating the initial serial number file is now
74 moved from CA.pl to the 'ca' utility with a new option -create_serial.
76 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
77 number file to 1, which is bound to cause problems. To avoid
78 the problems while respecting compatibility between different 0.9.7
79 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
80 CA.pl for serial number initialization. With the new release 0.9.8,
81 we can fix the problem directly in the 'ca' utility.)
84 *) Reduced header interdepencies by declaring more opaque objects in
85 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
86 give fewer recursive includes, which could break lazy source code - so
87 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
88 developers should define this symbol when building and using openssl to
89 ensure they track the recommended behaviour, interfaces, [etc], but
90 backwards-compatible behaviour prevails when this isn't defined.
93 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
96 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
97 This will generate a random key of the appropriate length based on the
98 cipher context. The EVP_CIPHER can provide its own random key generation
99 routine to support keys of a specific form. This is used in the des and
100 3des routines to generate a key of the correct parity. Update S/MIME
101 code to use new functions and hence generate correct parity DES keys.
102 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
103 valid (weak or incorrect parity).
106 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
107 as looking them up. This is useful when the verified structure may contain
108 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
109 present unless the new PKCS7_NO_CRL flag is asserted.
112 *) Extend ASN1 oid configuration module. It now additionally accepts the
115 shortName = some long name, 1.2.3.4
118 *) Reimplemented the BN_CTX implementation. There is now no more static
119 limitation on the number of variables it can handle nor the depth of the
120 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
121 information can now expand as required, and rather than having a single
122 static array of bignums, BN_CTX now uses a linked-list of such arrays
123 allowing it to expand on demand whilst maintaining the usefulness of
127 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
128 to allow all RSA operations to function using a single BN_CTX.
131 *) Preliminary support for certificate policy evaluation and checking. This
132 is initially intended to pass the tests outlined in "Conformance Testing
133 of Relying Party Client Certificate Path Processing Logic" v1.07.
136 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
137 remained unused and not that useful. A variety of other little bignum
138 tweaks and fixes have also been made continuing on from the audit (see
142 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
143 associated ASN1, EVP and SSL functions and old ASN1 macros.
146 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
147 and this should never fail. So the return value from the use of
148 BN_set_word() (which can fail due to needless expansion) is now deprecated;
149 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
152 *) BN_CTX_get() should return zero-valued bignums, providing the same
153 initialised value as BN_new().
154 [Geoff Thorpe, suggested by Ulf Möller]
156 *) Support for inhibitAnyPolicy certificate extension.
159 *) An audit of the BIGNUM code is underway, for which debugging code is
160 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
161 is considered valid when processing BIGNUMs, and causes execution to
162 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
163 further steps are taken to deliberately pollute unused data in BIGNUM
164 structures to try and expose faulty code further on. For now, openssl will
165 (in its default mode of operation) continue to tolerate the inconsistent
166 forms that it has tolerated in the past, but authors and packagers should
167 consider trying openssl and their own applications when compiled with
168 these debugging symbols defined. It will help highlight potential bugs in
169 their own code, and will improve the test coverage for OpenSSL itself. At
170 some point, these tighter rules will become openssl's default to improve
171 maintainability, though the assert()s and other overheads will remain only
172 in debugging configurations. See bn.h for more details.
173 [Geoff Thorpe, Nils Larsch, Ulf Möller]
175 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
176 that can only be obtained through BN_CTX_new() (which implicitly
177 initialises it). The presence of this function only made it possible
178 to overwrite an existing structure (and cause memory leaks).
181 *) Because of the callback-based approach for implementing LHASH as a
182 template type, lh_insert() adds opaque objects to hash-tables and
183 lh_doall() or lh_doall_arg() are typically used with a destructor callback
184 to clean up those corresponding objects before destroying the hash table
185 (and losing the object pointers). So some over-zealous constifications in
186 LHASH have been relaxed so that lh_insert() does not take (nor store) the
187 objects as "const" and the lh_doall[_arg] callback wrappers are not
188 prototyped to have "const" restrictions on the object pointers they are
189 given (and so aren't required to cast them away any more).
192 *) The tmdiff.h API was so ugly and minimal that our own timing utility
193 (speed) prefers to use its own implementation. The two implementations
194 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
195 its object type properly exposed (MS_TM) instead of casting to/from "char
196 *". This may still change yet if someone realises MS_TM and "ms_time_***"
197 aren't necessarily the greatest nomenclatures - but this is what was used
198 internally to the implementation so I've used that for now.
201 *) Ensure that deprecated functions do not get compiled when
202 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
203 the self-tests were still using deprecated key-generation functions so
204 these have been updated also.
207 *) Reorganise PKCS#7 code to separate the digest location functionality
208 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
209 New function PKCS7_set_digest() to set the digest type for PKCS#7
210 digestedData type. Add additional code to correctly generate the
211 digestedData type and add support for this type in PKCS7 initialization
215 *) New function PKCS7_set0_type_other() this initializes a PKCS7
216 structure of type "other".
219 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
220 sure the loop does correctly stop and breaking ("division by zero")
221 modulus operations are not performed. The (pre-generated) prime
222 table crypto/bn/bn_prime.h was already correct, but it could not be
223 re-generated on some platforms because of the "division by zero"
224 situation in the script.
225 [Ralf S. Engelschall]
227 *) Update support for ECC-based TLS ciphersuites according to
228 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
229 SHA-1 now is only used for "small" curves (where the
230 representation of a field element takes up to 24 bytes); for
231 larger curves, the field element resulting from ECDH is directly
232 used as premaster secret.
233 [Douglas Stebila (Sun Microsystems Laboratories)]
235 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
236 curve secp160r1 to the tests.
237 [Douglas Stebila (Sun Microsystems Laboratories)]
239 *) Add the possibility to load symbols globally with DSO.
240 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
242 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
243 control of the error stack.
246 *) Add support for STORE in ENGINE.
249 *) Add the STORE type. The intention is to provide a common interface
250 to certificate and key stores, be they simple file-based stores, or
251 HSM-type store, or LDAP stores, or...
252 NOTE: The code is currently UNTESTED and isn't really used anywhere.
255 *) Add a generic structure called OPENSSL_ITEM. This can be used to
256 pass a list of arguments to any function as well as provide a way
257 for a function to pass data back to the caller.
260 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
261 works like BUF_strdup() but can be used to duplicate a portion of
262 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
266 *) Add the function sk_find_ex() which works like sk_find(), but will
267 return an index to an element even if an exact match couldn't be
268 found. The index is guaranteed to point at the element where the
269 searched-for key would be inserted to preserve sorting order.
272 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
273 takes an extra flags argument for optional functionality. Currently,
274 the following flags are defined:
276 OBJ_BSEARCH_VALUE_ON_NOMATCH
277 This one gets OBJ_bsearch_ex() to return a pointer to the first
278 element where the comparing function returns a negative or zero
281 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
282 This one gets OBJ_bsearch_ex() to return a pointer to the first
283 element where the comparing function returns zero. This is useful
284 if there are more than one element where the comparing function
288 *) Make it possible to create self-signed certificates with 'openssl ca'
289 in such a way that the self-signed certificate becomes part of the
290 CA database and uses the same mechanisms for serial number generation
291 as all other certificate signing. The new flag '-selfsign' enables
292 this functionality. Adapt CA.sh and CA.pl.in.
295 *) Add functionality to check the public key of a certificate request
296 against a given private. This is useful to check that a certificate
297 request can be signed by that key (self-signing).
300 *) Make it possible to have multiple active certificates with the same
301 subject in the CA index file. This is done only if the keyword
302 'unique_subject' is set to 'no' in the main CA section (default
303 if 'CA_default') of the configuration file. The value is saved
304 with the database itself in a separate index attribute file,
305 named like the index file with '.attr' appended to the name.
308 *) Generate muti valued AVAs using '+' notation in config files for
312 *) Support for nameConstraints certificate extension.
315 *) Support for policyConstraints certificate extension.
318 *) Support for policyMappings certificate extension.
321 *) Make sure the default DSA_METHOD implementation only uses its
322 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
323 and change its own handlers to be NULL so as to remove unnecessary
324 indirection. This lets alternative implementations fallback to the
325 default implementation more easily.
328 *) Support for directoryName in GeneralName related extensions
332 *) Make it possible to link applications using Makefile.shared.
333 Make that possible even when linking against static libraries!
336 *) Support for single pass processing for S/MIME signing. This now
337 means that S/MIME signing can be done from a pipe, in addition
338 cleartext signing (multipart/signed type) is effectively streaming
339 and the signed data does not need to be all held in memory.
341 This is done with a new flag PKCS7_STREAM. When this flag is set
342 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
343 is done after the data is output (and digests calculated) in
347 *) Add full support for -rpath/-R, both in shared libraries and
348 applications, at least on the platforms where it's known how
352 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
353 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
354 will now compute a table of multiples of the generator that
355 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
356 faster (notably in the case of a single point multiplication,
358 [Nils Larsch, Bodo Moeller]
360 *) IPv6 support for certificate extensions. The various extensions
361 which use the IP:a.b.c.d can now take IPv6 addresses using the
362 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
366 *) Added an ENGINE that implements RSA by performing private key
367 exponentiations with the GMP library. The conversions to and from
368 GMP's mpz_t format aren't optimised nor are any montgomery forms
369 cached, and on x86 it appears OpenSSL's own performance has caught up.
370 However there are likely to be other architectures where GMP could
371 provide a boost. This ENGINE is not built in by default, but it can be
372 specified at Configure time and should be accompanied by the necessary
373 linker additions, eg;
374 ./config -DOPENSSL_USE_GMP -lgmp
377 *) "openssl engine" will not display ENGINE/DSO load failure errors when
378 testing availability of engines with "-t" - the old behaviour is
379 produced by increasing the feature's verbosity with "-tt".
382 *) ECDSA routines: under certain error conditions uninitialized BN objects
383 could be freed. Solution: make sure initialization is performed early
384 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
388 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
389 and DH_METHOD (eg. by ENGINE implementations) to override the normal
390 software implementations. For DSA and DH, parameter generation can
391 also be overriden by providing the appropriate method callbacks.
394 *) Change the "progress" mechanism used in key-generation and
395 primality testing to functions that take a new BN_GENCB pointer in
396 place of callback/argument pairs. The new API functions have "_ex"
397 postfixes and the older functions are reimplemented as wrappers for
398 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
399 declarations of the old functions to help (graceful) attempts to
400 migrate to the new functions. Also, the new key-generation API
401 functions operate on a caller-supplied key-structure and return
402 success/failure rather than returning a key or NULL - this is to
403 help make "keygen" another member function of RSA_METHOD etc.
405 Example for using the new callback interface:
407 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
411 BN_GENCB_set(&my_cb, my_callback, my_arg);
413 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
414 /* For the meaning of a, b in calls to my_callback(), see the
415 * documentation of the function that calls the callback.
416 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
417 * my_callback should return 1 if it wants BN_is_prime_ex()
418 * to continue, or 0 to stop.
423 *) Change the ZLIB compression method to be stateful, and make it
424 available to TLS with the number defined in
425 draft-ietf-tls-compression-04.txt.
428 *) Add the ASN.1 structures and functions for CertificatePair, which
429 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
431 CertificatePair ::= SEQUENCE {
432 forward [0] Certificate OPTIONAL,
433 reverse [1] Certificate OPTIONAL,
434 -- at least one of the pair shall be present -- }
436 Also implement the PEM functions to read and write certificate
437 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
439 This needed to be defined, mostly for the sake of the LDAP
440 attribute crossCertificatePair, but may prove useful elsewhere as
444 *) Make it possible to inhibit symlinking of shared libraries in
445 Makefile.shared, for Cygwin's sake.
448 *) Extend the BIGNUM API by creating new macros that behave like
451 void BN_set_sign(BIGNUM *a, int neg);
452 int BN_get_sign(const BIGNUM *a);
454 and avoid the need to access 'a->neg' directly in applications.
455 [Nils Larsch <nla@trustcenter.de>]
457 *) Implement fast modular reduction for pseudo-Mersenne primes
458 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
459 EC_GROUP_new_curve_GFp() will now automatically use this
461 [Nils Larsch <nla@trustcenter.de>]
463 *) Add new lock type (CRYPTO_LOCK_BN).
466 *) Change the ENGINE framework to automatically load engines
467 dynamically from specific directories unless they could be
468 found to already be built in or loaded. Move all the
469 current engines except for the cryptodev one to a new
471 The engines in engines/ are built as shared libraries if
472 the "shared" options was given to ./Configure or ./config.
473 Otherwise, they are inserted in libcrypto.a.
474 /usr/local/ssl/engines is the default directory for dynamic
475 engines, but that can be overriden at configure time through
476 the usual use of --prefix and/or --openssldir, and at run
477 time with the environment variable OPENSSL_ENGINES.
478 [Geoff Thorpe and Richard Levitte]
480 *) Add Makefile.shared, a helper makefile to build shared
481 libraries. Addapt Makefile.org.
484 *) Add version info to Win32 DLLs.
485 [Peter 'Luna' Runestig" <peter@runestig.com>]
487 *) Add new 'medium level' PKCS#12 API. Certificates and keys
488 can be added using this API to created arbitrary PKCS#12
489 files while avoiding the low level API.
491 New options to PKCS12_create(), key or cert can be NULL and
492 will then be omitted from the output file. The encryption
493 algorithm NIDs can be set to -1 for no encryption, the mac
494 iteration count can be set to 0 to omit the mac.
496 Enhance pkcs12 utility by making the -nokeys and -nocerts
497 options work when creating a PKCS#12 file. New option -nomac
498 to omit the mac, NONE can be set for an encryption algorithm.
499 New code is modified to use the enhanced PKCS12_create()
500 instead of the low level API.
503 *) Extend ASN1 encoder to support indefinite length constructed
504 encoding. This can output sequences tags and octet strings in
505 this form. Modify pk7_asn1.c to support indefinite length
506 encoding. This is experimental and needs additional code to
507 be useful, such as an ASN1 bio and some enhanced streaming
510 Extend template encode functionality so that tagging is passed
511 down to the template encoder.
514 *) Let 'openssl req' fail if an argument to '-newkey' is not
515 recognized instead of using RSA as a default.
518 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
519 As these are not official, they are not included in "ALL";
520 the "ECCdraft" ciphersuite group alias can be used to select them.
521 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
523 *) Add ECDH engine support.
524 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
526 *) Add ECDH in new directory crypto/ecdh/.
527 [Douglas Stebila (Sun Microsystems Laboratories)]
529 *) Let BN_rand_range() abort with an error after 100 iterations
530 without success (which indicates a broken PRNG).
533 *) Change BN_mod_sqrt() so that it verifies that the input value
534 is really the square of the return value. (Previously,
535 BN_mod_sqrt would show GIGO behaviour.)
538 *) Add named elliptic curves over binary fields from X9.62, SECG,
539 and WAP/WTLS; add OIDs that were still missing.
541 [Sheueling Chang Shantz and Douglas Stebila
542 (Sun Microsystems Laboratories)]
544 *) Extend the EC library for elliptic curves over binary fields
545 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
548 EC_GF2m_simple_method
552 EC_GROUP_new_curve_GF2m
553 EC_GROUP_set_curve_GF2m
554 EC_GROUP_get_curve_GF2m
555 EC_POINT_set_affine_coordinates_GF2m
556 EC_POINT_get_affine_coordinates_GF2m
557 EC_POINT_set_compressed_coordinates_GF2m
559 Point compression for binary fields is disabled by default for
560 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
563 As binary polynomials are represented as BIGNUMs, various members
564 of the EC_GROUP and EC_POINT data structures can be shared
565 between the implementations for prime fields and binary fields;
566 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
567 are essentially identical to their ..._GFp counterparts.
568 (For simplicity, the '..._GFp' prefix has been dropped from
569 various internal method names.)
571 An internal 'field_div' method (similar to 'field_mul' and
572 'field_sqr') has been added; this is used only for binary fields.
574 [Sheueling Chang Shantz and Douglas Stebila
575 (Sun Microsystems Laboratories)]
577 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
578 through methods ('mul', 'precompute_mult').
580 The generic implementations (now internally called 'ec_wNAF_mul'
581 and 'ec_wNAF_precomputed_mult') remain the default if these
582 methods are undefined.
584 [Sheueling Chang Shantz and Douglas Stebila
585 (Sun Microsystems Laboratories)]
587 *) New function EC_GROUP_get_degree, which is defined through
588 EC_METHOD. For curves over prime fields, this returns the bit
589 length of the modulus.
591 [Sheueling Chang Shantz and Douglas Stebila
592 (Sun Microsystems Laboratories)]
594 *) New functions EC_GROUP_dup, EC_POINT_dup.
595 (These simply call ..._new and ..._copy).
597 [Sheueling Chang Shantz and Douglas Stebila
598 (Sun Microsystems Laboratories)]
600 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
601 Polynomials are represented as BIGNUMs (where the sign bit is not
602 used) in the following functions [macros]:
605 BN_GF2m_sub [= BN_GF2m_add]
606 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
607 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
608 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
610 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
611 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
612 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
613 BN_GF2m_cmp [= BN_ucmp]
615 (Note that only the 'mod' functions are actually for fields GF(2^m).
616 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
618 For some functions, an the irreducible polynomial defining a
619 field can be given as an 'unsigned int[]' with strictly
620 decreasing elements giving the indices of those bits that are set;
621 i.e., p[] represents the polynomial
622 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
624 p[0] > p[1] > ... > p[k] = 0.
625 This applies to the following functions:
630 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
631 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
634 BN_GF2m_mod_solve_quad_arr
638 Conversion can be performed by the following functions:
643 bntest.c has additional tests for binary polynomial arithmetic.
645 Two implementations for BN_GF2m_mod_div() are available.
646 The default algorithm simply uses BN_GF2m_mod_inv() and
647 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
648 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
649 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
651 [Sheueling Chang Shantz and Douglas Stebila
652 (Sun Microsystems Laboratories)]
654 *) Add new error code 'ERR_R_DISABLED' that can be used when some
655 functionality is disabled at compile-time.
656 [Douglas Stebila <douglas.stebila@sun.com>]
658 *) Change default behaviour of 'openssl asn1parse' so that more
659 information is visible when viewing, e.g., a certificate:
661 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
662 mode the content of non-printable OCTET STRINGs is output in a
663 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
664 avoid the appearance of a printable string.
665 [Nils Larsch <nla@trustcenter.de>]
667 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
669 EC_GROUP_set_asn1_flag()
670 EC_GROUP_get_asn1_flag()
671 EC_GROUP_set_point_conversion_form()
672 EC_GROUP_get_point_conversion_form()
673 These control ASN1 encoding details:
674 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
675 has been set to OPENSSL_EC_NAMED_CURVE.
676 - Points are encoded in uncompressed form by default; options for
677 asn1_for are as for point2oct, namely
678 POINT_CONVERSION_COMPRESSED
679 POINT_CONVERSION_UNCOMPRESSED
680 POINT_CONVERSION_HYBRID
682 Also add 'seed' and 'seed_len' members to EC_GROUP with access
686 EC_GROUP_get_seed_len()
687 This is used only for ASN1 purposes (so far).
688 [Nils Larsch <nla@trustcenter.de>]
690 *) Add 'field_type' member to EC_METHOD, which holds the NID
691 of the appropriate field type OID. The new function
692 EC_METHOD_get_field_type() returns this value.
693 [Nils Larsch <nla@trustcenter.de>]
700 providing useful interfaces to EC_POINT_point2oct() and
701 EC_POINT_oct2point().
702 [Nils Larsch <nla@trustcenter.de>]
704 *) Change internals of the EC library so that the functions
705 EC_GROUP_set_generator()
706 EC_GROUP_get_generator()
708 EC_GROUP_get_cofactor()
709 are implemented directly in crypto/ec/ec_lib.c and not dispatched
710 to methods, which would lead to unnecessary code duplication when
711 adding different types of curves.
712 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
714 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
715 arithmetic, and such that modified wNAFs are generated
716 (which avoid length expansion in many cases).
719 *) Add a function EC_GROUP_check_discriminant() (defined via
720 EC_METHOD) that verifies that the curve discriminant is non-zero.
722 Add a function EC_GROUP_check() that makes some sanity tests
723 on a EC_GROUP, its generator and order. This includes
724 EC_GROUP_check_discriminant().
725 [Nils Larsch <nla@trustcenter.de>]
727 *) Add ECDSA in new directory crypto/ecdsa/.
729 Add applications 'openssl ecparam' and 'openssl ecdsa'
730 (these are based on 'openssl dsaparam' and 'openssl dsa').
732 ECDSA support is also included in various other files across the
733 library. Most notably,
734 - 'openssl req' now has a '-newkey ecdsa:file' option;
735 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
736 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
737 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
738 them suitable for ECDSA where domain parameters must be
739 extracted before the specific public key;
740 - ECDSA engine support has been added.
741 [Nils Larsch <nla@trustcenter.de>]
743 *) Include some named elliptic curves, and add OIDs from X9.62,
744 SECG, and WAP/WTLS. Each curve can be obtained from the new
746 EC_GROUP_new_by_nid(),
747 and the list of available named curves can be obtained with
748 EC_get_builtin_curves().
749 Also add a 'curve_name' member to EC_GROUP objects, which can be
753 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
755 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
756 was actually never needed) and in BN_mul(). The removal in BN_mul()
757 required a small change in bn_mul_part_recursive() and the addition
758 of the functions bn_cmp_part_words(), bn_sub_part_words() and
759 bn_add_part_words(), which do the same thing as bn_cmp_words(),
760 bn_sub_words() and bn_add_words() except they take arrays with
764 Changes between 0.9.7e and 0.9.7f [XX xxx XXXX]
766 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
767 server and client random values. Previously
768 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
769 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
771 This change has negligible security impact because:
773 1. Server and client random values still have 24 bytes of pseudo random
776 2. Server and client random values are sent in the clear in the initial
779 3. The master secret is derived using the premaster secret (48 bytes in
780 size for static RSA ciphersuites) as well as client server and random
783 The OpenSSL team would like to thank the UK NISCC for bringing this issue
786 [Stephen Henson, reported by UK NISCC]
788 *) Use Windows randomness collection on Cygwin.
791 *) Fix hang in EGD/PRNGD query when communication socket is closed
792 prematurely by EGD/PRNGD.
793 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
795 *) Prompt for pass phrases when appropriate for PKCS12 input format.
798 *) Add lots of checks for memory allocation failure, error codes to indicate
799 failure and freeing up memory if a failure occurs.
800 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
802 *) Add new -passin argument to dgst.
805 *) Perform some character comparisons of different types in X509_NAME_cmp:
806 this is needed for some certificates that reencode DNs into UTF8Strings
807 (in violation of RFC3280) and can't or wont issue name rollover
811 *) Make an explicit check during certificate validation to see that
812 the CA setting in each certificate on the chain is correct. As a
813 side effect always do the following basic checks on extensions,
814 not just when there's an associated purpose to the check:
816 - if there is an unhandled critical extension (unless the user
817 has chosen to ignore this fault)
818 - if the path length has been exceeded (if one is set at all)
819 - that certain extensions fit the associated purpose (if one has
823 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
825 *) Avoid a race condition when CRLs are checked in a multi threaded
826 environment. This would happen due to the reordering of the revoked
827 entries during signature checking and serial number lookup. Now the
828 encoding is cached and the serial number sort performed under a lock.
829 Add new STACK function sk_is_sorted().
832 *) Add Delta CRL to the extension code.
835 *) Various fixes to s3_pkt.c so alerts are sent properly.
836 [David Holmes <d.holmes@f5.com>]
838 *) Reduce the chances of duplicate issuer name and serial numbers (in
839 violation of RFC3280) using the OpenSSL certificate creation utilities.
840 This is done by creating a random 64 bit value for the initial serial
841 number when a serial number file is created or when a self signed
842 certificate is created using 'openssl req -x509'. The initial serial
843 number file is created using 'openssl x509 -next_serial' in CA.pl
844 rather than being initialized to 1.
847 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
849 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
850 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
851 [Joe Orton, Steve Henson]
853 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
855 [Joe Orton, Steve Henson]
857 *) Make it possible to have multiple active certificates with the same
858 subject in the CA index file. This is done only if the keyword
859 'unique_subject' is set to 'no' in the main CA section (default
860 if 'CA_default') of the configuration file. The value is saved
861 with the database itself in a separate index attribute file,
862 named like the index file with '.attr' appended to the name.
865 *) X509 verify fixes. Disable broken certificate workarounds when
866 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
867 keyUsage extension present. Don't accept CRLs with unhandled critical
868 extensions: since verify currently doesn't process CRL extensions this
869 rejects a CRL with *any* critical extensions. Add new verify error codes
873 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
874 A clarification of RFC2560 will require the use of OCTET STRINGs and
875 some implementations cannot handle the current raw format. Since OpenSSL
876 copies and compares OCSP nonces as opaque blobs without any attempt at
877 parsing them this should not create any compatibility issues.
880 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
881 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
882 this HMAC (and other) operations are several times slower than OpenSSL
886 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
887 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
889 *) Use the correct content when signing type "other".
892 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
894 *) Fix various bugs revealed by running the NISCC test suite:
896 Stop out of bounds reads in the ASN1 code when presented with
897 invalid tags (CAN-2003-0543 and CAN-2003-0544).
899 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
901 If verify callback ignores invalid public key errors don't try to check
902 certificate signature with the NULL public key.
906 *) New -ignore_err option in ocsp application to stop the server
907 exiting on the first error in a request.
910 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
911 if the server requested one: as stated in TLS 1.0 and SSL 3.0
915 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
916 extra data after the compression methods not only for TLS 1.0
917 but also for SSL 3.0 (as required by the specification).
918 [Bodo Moeller; problem pointed out by Matthias Loepfe]
920 *) Change X509_certificate_type() to mark the key as exported/exportable
921 when it's 512 *bits* long, not 512 bytes.
924 *) Change AES_cbc_encrypt() so it outputs exact multiple of
925 blocks during encryption.
928 *) Various fixes to base64 BIO and non blocking I/O. On write
929 flushes were not handled properly if the BIO retried. On read
930 data was not being buffered properly and had various logic bugs.
931 This also affects blocking I/O when the data being decoded is a
935 *) Various S/MIME bugfixes and compatibility changes:
936 output correct application/pkcs7 MIME type if
937 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
938 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
939 of files as .eml work). Correctly handle very long lines in MIME
943 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
945 *) Countermeasure against the Klima-Pokorny-Rosa extension of
946 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
947 a protocol version number mismatch like a decryption error
948 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
951 *) Turn on RSA blinding by default in the default implementation
952 to avoid a timing attack. Applications that don't want it can call
953 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
954 They would be ill-advised to do so in most cases.
955 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
957 *) Change RSA blinding code so that it works when the PRNG is not
958 seeded (in this case, the secret RSA exponent is abused as
959 an unpredictable seed -- if it is not unpredictable, there
960 is no point in blinding anyway). Make RSA blinding thread-safe
961 by remembering the creator's thread ID in rsa->blinding and
962 having all other threads use local one-time blinding factors
963 (this requires more computation than sharing rsa->blinding, but
964 avoids excessive locking; and if an RSA object is not shared
965 between threads, blinding will still be very fast).
968 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
969 ENGINE as defaults for all supported algorithms irrespective of
970 the 'flags' parameter. 'flags' is now honoured, so applications
971 should make sure they are passing it correctly.
974 *) Target "mingw" now allows native Windows code to be generated in
975 the Cygwin environment as well as with the MinGW compiler.
978 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
980 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
981 via timing by performing a MAC computation even if incorrrect
982 block cipher padding has been found. This is a countermeasure
983 against active attacks where the attacker has to distinguish
984 between bad padding and a MAC verification error. (CAN-2003-0078)
986 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
987 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
988 Martin Vuagnoux (EPFL, Ilion)]
990 *) Make the no-err option work as intended. The intention with no-err
991 is not to have the whole error stack handling routines removed from
992 libcrypto, it's only intended to remove all the function name and
993 reason texts, thereby removing some of the footprint that may not
994 be interesting if those errors aren't displayed anyway.
996 NOTE: it's still possible for any application or module to have it's
997 own set of error texts inserted. The routines are there, just not
998 used by default when no-err is given.
1001 *) Add support for FreeBSD on IA64.
1002 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1004 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1005 Kerberos function mit_des_cbc_cksum(). Before this change,
1006 the value returned by DES_cbc_cksum() was like the one from
1007 mit_des_cbc_cksum(), except the bytes were swapped.
1008 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1010 *) Allow an application to disable the automatic SSL chain building.
1011 Before this a rather primitive chain build was always performed in
1012 ssl3_output_cert_chain(): an application had no way to send the
1013 correct chain if the automatic operation produced an incorrect result.
1015 Now the chain builder is disabled if either:
1017 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1019 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1021 The reasoning behind this is that an application would not want the
1022 auto chain building to take place if extra chain certificates are
1023 present and it might also want a means of sending no additional
1024 certificates (for example the chain has two certificates and the
1028 *) Add the possibility to build without the ENGINE framework.
1029 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1031 *) Under Win32 gmtime() can return NULL: check return value in
1032 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1035 *) DSA routines: under certain error conditions uninitialized BN objects
1036 could be freed. Solution: make sure initialization is performed early
1037 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1038 Nils Larsch <nla@trustcenter.de> via PR#459)
1041 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1042 checked on reconnect on the client side, therefore session resumption
1043 could still fail with a "ssl session id is different" error. This
1044 behaviour is masked when SSL_OP_ALL is used due to
1045 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1046 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1047 followup to PR #377.
1050 *) IA-32 assembler support enhancements: unified ELF targets, support
1051 for SCO/Caldera platforms, fix for Cygwin shared build.
1054 *) Add support for FreeBSD on sparc64. As a consequence, support for
1055 FreeBSD on non-x86 processors is separate from x86 processors on
1056 the config script, much like the NetBSD support.
1057 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1059 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1061 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1064 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1065 code (06) was taken as the first octet of the session ID and the last
1066 octet was ignored consequently. As a result SSLv2 client side session
1067 caching could not have worked due to the session ID mismatch between
1069 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1073 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1074 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1078 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1079 seems that in spite of existing for more than a year, many application
1080 author have done nothing to provide the necessary callbacks, which
1081 means that this particular engine will not work properly anywhere.
1082 This is a very unfortunate situation which forces us, in the name
1083 of usability, to give the hw_ncipher.c a static lock, which is part
1085 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1086 appear in 0.9.8 or later. We EXPECT application authors to have
1087 dealt properly with this when 0.9.8 is released (unless we actually
1088 make such changes in the libcrypto locking code that changes will
1089 have to be made anyway).
1092 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1093 octets have been read, EOF or an error occurs. Without this change
1094 some truncated ASN1 structures will not produce an error.
1097 *) Disable Heimdal support, since it hasn't been fully implemented.
1098 Still give the possibility to force the use of Heimdal, but with
1099 warnings and a request that patches get sent to openssl-dev.
1102 *) Add the VC-CE target, introduce the WINCE sysname, and add
1103 INSTALL.WCE and appropriate conditionals to make it build.
1104 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1106 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1107 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1108 edit numbers of the version.
1109 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1111 *) Introduce safe string copy and catenation functions
1112 (BUF_strlcpy() and BUF_strlcat()).
1113 [Ben Laurie (CHATS) and Richard Levitte]
1115 *) Avoid using fixed-size buffers for one-line DNs.
1116 [Ben Laurie (CHATS)]
1118 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1119 resizing buffers containing secrets, and use where appropriate.
1120 [Ben Laurie (CHATS)]
1122 *) Avoid using fixed size buffers for configuration file location.
1123 [Ben Laurie (CHATS)]
1125 *) Avoid filename truncation for various CA files.
1126 [Ben Laurie (CHATS)]
1128 *) Use sizeof in preference to magic numbers.
1129 [Ben Laurie (CHATS)]
1131 *) Avoid filename truncation in cert requests.
1132 [Ben Laurie (CHATS)]
1134 *) Add assertions to check for (supposedly impossible) buffer
1136 [Ben Laurie (CHATS)]
1138 *) Don't cache truncated DNS entries in the local cache (this could
1139 potentially lead to a spoofing attack).
1140 [Ben Laurie (CHATS)]
1142 *) Fix various buffers to be large enough for hex/decimal
1143 representations in a platform independent manner.
1144 [Ben Laurie (CHATS)]
1146 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1147 resizing buffers containing secrets, and use where appropriate.
1148 [Ben Laurie (CHATS)]
1150 *) Add BIO_indent() to avoid much slightly worrying code to do
1152 [Ben Laurie (CHATS)]
1154 *) Convert sprintf()/BIO_puts() to BIO_printf().
1155 [Ben Laurie (CHATS)]
1157 *) buffer_gets() could terminate with the buffer only half
1159 [Ben Laurie (CHATS)]
1161 *) Add assertions to prevent user-supplied crypto functions from
1162 overflowing internal buffers by having large block sizes, etc.
1163 [Ben Laurie (CHATS)]
1165 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1167 [Ben Laurie (CHATS)]
1169 *) Eliminate unused copy of key in RC4.
1170 [Ben Laurie (CHATS)]
1172 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1173 [Ben Laurie (CHATS)]
1175 *) Fix off-by-one error in EGD path.
1176 [Ben Laurie (CHATS)]
1178 *) If RANDFILE path is too long, ignore instead of truncating.
1179 [Ben Laurie (CHATS)]
1181 *) Eliminate unused and incorrectly sized X.509 structure
1183 [Ben Laurie (CHATS)]
1185 *) Eliminate unused and dangerous function knumber().
1186 [Ben Laurie (CHATS)]
1188 *) Eliminate unused and dangerous structure, KSSL_ERR.
1189 [Ben Laurie (CHATS)]
1191 *) Protect against overlong session ID context length in an encoded
1192 session object. Since these are local, this does not appear to be
1194 [Ben Laurie (CHATS)]
1196 *) Change from security patch (see 0.9.6e below) that did not affect
1197 the 0.9.6 release series:
1199 Remote buffer overflow in SSL3 protocol - an attacker could
1200 supply an oversized master key in Kerberos-enabled versions.
1202 [Ben Laurie (CHATS)]
1204 *) Change the SSL kerb5 codes to match RFC 2712.
1207 *) Make -nameopt work fully for req and add -reqopt switch.
1208 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1210 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1211 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1213 *) Make sure tests can be performed even if the corresponding algorithms
1214 have been removed entirely. This was also the last step to make
1215 OpenSSL compilable with DJGPP under all reasonable conditions.
1216 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1218 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1219 to allow version independent disabling of normally unselected ciphers,
1220 which may be activated as a side-effect of selecting a single cipher.
1222 (E.g., cipher list string "RSA" enables ciphersuites that are left
1223 out of "ALL" because they do not provide symmetric encryption.
1224 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1225 [Lutz Jaenicke, Bodo Moeller]
1227 *) Add appropriate support for separate platform-dependent build
1228 directories. The recommended way to make a platform-dependent
1229 build directory is the following (tested on Linux), maybe with
1232 # Place yourself outside of the OpenSSL source tree. In
1233 # this example, the environment variable OPENSSL_SOURCE
1234 # is assumed to contain the absolute OpenSSL source directory.
1235 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1236 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1237 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1238 mkdir -p `dirname $F`
1239 ln -s $OPENSSL_SOURCE/$F $F
1242 To be absolutely sure not to disturb the source tree, a "make clean"
1243 is a good thing. If it isn't successfull, don't worry about it,
1244 it probably means the source directory is very clean.
1247 *) Make sure any ENGINE control commands make local copies of string
1248 pointers passed to them whenever necessary. Otherwise it is possible
1249 the caller may have overwritten (or deallocated) the original string
1250 data when a later ENGINE operation tries to use the stored values.
1251 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1253 *) Improve diagnostics in file reading and command-line digests.
1254 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1256 *) Add AES modes CFB and OFB to the object database. Correct an
1257 error in AES-CFB decryption.
1260 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1261 allows existing EVP_CIPHER_CTX structures to be reused after
1262 calling EVP_*Final(). This behaviour is used by encryption
1263 BIOs and some applications. This has the side effect that
1264 applications must explicitly clean up cipher contexts with
1265 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1268 *) Check the values of dna and dnb in bn_mul_recursive before calling
1269 bn_mul_comba (a non zero value means the a or b arrays do not contain
1270 n2 elements) and fallback to bn_mul_normal if either is not zero.
1273 *) Fix escaping of non-ASCII characters when using the -subj option
1274 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1277 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1278 form for "surname", serialNumber has no short form.
1279 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1280 therefore remove "mail" short name for "internet 7".
1281 The OID for unique identifiers in X509 certificates is
1282 x500UniqueIdentifier, not uniqueIdentifier.
1283 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1286 *) Add an "init" command to the ENGINE config module and auto initialize
1287 ENGINEs. Without any "init" command the ENGINE will be initialized
1288 after all ctrl commands have been executed on it. If init=1 the
1289 ENGINE is initailized at that point (ctrls before that point are run
1290 on the uninitialized ENGINE and after on the initialized one). If
1291 init=0 then the ENGINE will not be iniatialized at all.
1294 *) Fix the 'app_verify_callback' interface so that the user-defined
1295 argument is actually passed to the callback: In the
1296 SSL_CTX_set_cert_verify_callback() prototype, the callback
1297 declaration has been changed from
1300 int (*cb)(X509_STORE_CTX *,void *);
1301 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1302 i=s->ctx->app_verify_callback(&ctx)
1303 has been changed into
1304 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1306 To update applications using SSL_CTX_set_cert_verify_callback(),
1307 a dummy argument can be added to their callback functions.
1308 [D. K. Smetters <smetters@parc.xerox.com>]
1310 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1311 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1313 *) Add and OPENSSL_LOAD_CONF define which will cause
1314 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1315 This allows older applications to transparently support certain
1316 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1317 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1318 load the config file and OPENSSL_add_all_algorithms_conf() which will
1319 always load it have also been added.
1322 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1323 Adjust NIDs and EVP layer.
1324 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1326 *) Config modules support in openssl utility.
1328 Most commands now load modules from the config file,
1329 though in a few (such as version) this isn't done
1330 because it couldn't be used for anything.
1332 In the case of ca and req the config file used is
1333 the same as the utility itself: that is the -config
1334 command line option can be used to specify an
1338 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1339 use "openssl_conf" if filename is NULL use default openssl config file.
1342 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1343 config section name. Add a new flag to tolerate a missing config file
1344 and move code to CONF_modules_load_file().
1347 *) Support for crypto accelerator cards from Accelerated Encryption
1348 Processing, www.aep.ie. (Use engine 'aep')
1349 The support was copied from 0.9.6c [engine] and adapted/corrected
1350 to work with the new engine framework.
1351 [AEP Inc. and Richard Levitte]
1353 *) Support for SureWare crypto accelerator cards from Baltimore
1354 Technologies. (Use engine 'sureware')
1355 The support was copied from 0.9.6c [engine] and adapted
1356 to work with the new engine framework.
1359 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1360 make the newer ENGINE framework commands for the CHIL engine work.
1361 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1363 *) Make it possible to produce shared libraries on ReliantUNIX.
1364 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1366 *) Add the configuration target debug-linux-ppro.
1367 Make 'openssl rsa' use the general key loading routines
1368 implemented in apps.c, and make those routines able to
1369 handle the key format FORMAT_NETSCAPE and the variant
1371 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1373 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1374 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1376 *) Add -keyform to rsautl, and document -engine.
1377 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1379 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1380 BIO_R_NO_SUCH_FILE error code rather than the generic
1381 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1384 *) Add new functions
1386 ERR_peek_last_error_line
1387 ERR_peek_last_error_line_data.
1388 These are similar to
1391 ERR_peek_error_line_data,
1392 but report on the latest error recorded rather than the first one
1393 still in the error queue.
1394 [Ben Laurie, Bodo Moeller]
1396 *) default_algorithms option in ENGINE config module. This allows things
1398 default_algorithms = ALL
1399 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1402 *) Prelminary ENGINE config module.
1405 *) New experimental application configuration code.
1408 *) Change the AES code to follow the same name structure as all other
1409 symmetric ciphers, and behave the same way. Move everything to
1410 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1411 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1413 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1414 [Ben Laurie and Theo de Raadt]
1416 *) Add option to output public keys in req command.
1417 [Massimiliano Pala madwolf@openca.org]
1419 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1420 (up to about 10% better than before for P-192 and P-224).
1423 *) New functions/macros
1425 SSL_CTX_set_msg_callback(ctx, cb)
1426 SSL_CTX_set_msg_callback_arg(ctx, arg)
1427 SSL_set_msg_callback(ssl, cb)
1428 SSL_set_msg_callback_arg(ssl, arg)
1430 to request calling a callback function
1432 void cb(int write_p, int version, int content_type,
1433 const void *buf, size_t len, SSL *ssl, void *arg)
1435 whenever a protocol message has been completely received
1436 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1437 protocol version according to which the SSL library interprets
1438 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1439 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1440 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1441 specification (change_cipher_spec(20), alert(21), handshake(22)).
1442 'buf' and 'len' point to the actual message, 'ssl' to the
1443 SSL object, and 'arg' is the application-defined value set by
1444 SSL[_CTX]_set_msg_callback_arg().
1446 'openssl s_client' and 'openssl s_server' have new '-msg' options
1447 to enable a callback that displays all protocol messages.
1450 *) Change the shared library support so shared libraries are built as
1451 soon as the corresponding static library is finished, and thereby get
1452 openssl and the test programs linked against the shared library.
1453 This still only happens when the keyword "shard" has been given to
1454 the configuration scripts.
1456 NOTE: shared library support is still an experimental thing, and
1457 backward binary compatibility is still not guaranteed.
1458 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1460 *) Add support for Subject Information Access extension.
1461 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1463 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1464 additional bytes when new memory had to be allocated, not just
1465 when reusing an existing buffer.
1468 *) New command line and configuration option 'utf8' for the req command.
1469 This allows field values to be specified as UTF8 strings.
1472 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1473 runs for the former and machine-readable output for the latter.
1476 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1477 of the e-mail address in the DN (i.e., it will go into a certificate
1478 extension only). The new configuration file option 'email_in_dn = no'
1479 has the same effect.
1480 [Massimiliano Pala madwolf@openca.org]
1482 *) Change all functions with names starting with des_ to be starting
1483 with DES_ instead. Add wrappers that are compatible with libdes,
1484 but are named _ossl_old_des_*. Finally, add macros that map the
1485 des_* symbols to the corresponding _ossl_old_des_* if libdes
1486 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1487 desired, the des_* symbols will be mapped to DES_*, with one
1490 Since we provide two compatibility mappings, the user needs to
1491 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1492 compatibility is desired. The default (i.e., when that macro
1493 isn't defined) is OpenSSL 0.9.6c compatibility.
1495 There are also macros that enable and disable the support of old
1496 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1497 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1498 are defined, the default will apply: to support the old des routines.
1500 In either case, one must include openssl/des.h to get the correct
1501 definitions. Do not try to just include openssl/des_old.h, that
1504 NOTE: This is a major break of an old API into a new one. Software
1505 authors are encouraged to switch to the DES_ style functions. Some
1506 time in the future, des_old.h and the libdes compatibility functions
1507 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1508 default), and then completely removed.
1511 *) Test for certificates which contain unsupported critical extensions.
1512 If such a certificate is found during a verify operation it is
1513 rejected by default: this behaviour can be overridden by either
1514 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1515 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1516 X509_supported_extension() has also been added which returns 1 if a
1517 particular extension is supported.
1520 *) Modify the behaviour of EVP cipher functions in similar way to digests
1521 to retain compatibility with existing code.
1524 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1525 compatibility with existing code. In particular the 'ctx' parameter does
1526 not have to be to be initialized before the call to EVP_DigestInit() and
1527 it is tidied up after a call to EVP_DigestFinal(). New function
1528 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1529 EVP_MD_CTX_copy() changed to not require the destination to be
1530 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1531 requires the destination to be valid.
1533 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1534 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1537 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1538 so that complete 'Handshake' protocol structures are kept in memory
1539 instead of overwriting 'msg_type' and 'length' with 'body' data.
1542 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1543 [Massimo Santin via Richard Levitte]
1545 *) Major restructuring to the underlying ENGINE code. This includes
1546 reduction of linker bloat, separation of pure "ENGINE" manipulation
1547 (initialisation, etc) from functionality dealing with implementations
1548 of specific crypto iterfaces. This change also introduces integrated
1549 support for symmetric ciphers and digest implementations - so ENGINEs
1550 can now accelerate these by providing EVP_CIPHER and EVP_MD
1551 implementations of their own. This is detailed in crypto/engine/README
1552 as it couldn't be adequately described here. However, there are a few
1553 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1554 were changed in the original introduction of ENGINE code have now
1555 reverted back - the hooking from this code to ENGINE is now a good
1556 deal more passive and at run-time, operations deal directly with
1557 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1558 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1559 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1560 they were not being used by the framework as there is no concept of a
1561 BIGNUM_METHOD and they could not be generalised to the new
1562 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1563 ENGINE_cpy() has been removed as it cannot be consistently defined in
1567 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1570 *) Change mkdef.pl to sort symbols that get the same entry number,
1571 and make sure the automatically generated functions ERR_load_*
1572 become part of libeay.num as well.
1575 *) New function SSL_renegotiate_pending(). This returns true once
1576 renegotiation has been requested (either SSL_renegotiate() call
1577 or HelloRequest/ClientHello receveived from the peer) and becomes
1578 false once a handshake has been completed.
1579 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1580 sends a HelloRequest, but does not ensure that a handshake takes
1581 place. SSL_renegotiate_pending() is useful for checking if the
1582 client has followed the request.)
1585 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1586 By default, clients may request session resumption even during
1587 renegotiation (if session ID contexts permit); with this option,
1588 session resumption is possible only in the first handshake.
1590 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1591 more bits available for options that should not be part of
1592 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1595 *) Add some demos for certificate and certificate request creation.
1598 *) Make maximum certificate chain size accepted from the peer application
1599 settable (SSL*_get/set_max_cert_list()), as proposed by
1600 "Douglas E. Engert" <deengert@anl.gov>.
1603 *) Add support for shared libraries for Unixware-7
1604 (Boyd Lynn Gerber <gerberb@zenez.com>).
1607 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1608 be done prior to destruction. Use this to unload error strings from
1609 ENGINEs that load their own error strings. NB: This adds two new API
1610 functions to "get" and "set" this destroy handler in an ENGINE.
1613 *) Alter all existing ENGINE implementations (except "openssl" and
1614 "openbsd") to dynamically instantiate their own error strings. This
1615 makes them more flexible to be built both as statically-linked ENGINEs
1616 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1617 Also, add stub code to each that makes building them as self-contained
1618 shared-libraries easier (see README.ENGINE).
1621 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1622 implementations into applications that are completely implemented in
1623 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1624 commands that can be used to configure what shared-library to load and
1625 to control aspects of the way it is handled. Also, made an update to
1626 the README.ENGINE file that brings its information up-to-date and
1627 provides some information and instructions on the "dynamic" ENGINE
1628 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1631 *) Make it possible to unload ranges of ERR strings with a new
1632 "ERR_unload_strings" function.
1635 *) Add a copy() function to EVP_MD.
1638 *) Make EVP_MD routines take a context pointer instead of just the
1639 md_data void pointer.
1642 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1643 that the digest can only process a single chunk of data
1644 (typically because it is provided by a piece of
1645 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1646 is only going to provide a single chunk of data, and hence the
1647 framework needn't accumulate the data for oneshot drivers.
1650 *) As with "ERR", make it possible to replace the underlying "ex_data"
1651 functions. This change also alters the storage and management of global
1652 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1653 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1654 index counters. The API functions that use this state have been changed
1655 to take a "class_index" rather than pointers to the class's local STACK
1656 and counter, and there is now an API function to dynamically create new
1657 classes. This centralisation allows us to (a) plug a lot of the
1658 thread-safety problems that existed, and (b) makes it possible to clean
1659 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1660 such data would previously have always leaked in application code and
1661 workarounds were in place to make the memory debugging turn a blind eye
1662 to it. Application code that doesn't use this new function will still
1663 leak as before, but their memory debugging output will announce it now
1664 rather than letting it slide.
1666 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1667 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1668 has a return value to indicate success or failure.
1671 *) Make it possible to replace the underlying "ERR" functions such that the
1672 global state (2 LHASH tables and 2 locks) is only used by the "default"
1673 implementation. This change also adds two functions to "get" and "set"
1674 the implementation prior to it being automatically set the first time
1675 any other ERR function takes place. Ie. an application can call "get",
1676 pass the return value to a module it has just loaded, and that module
1677 can call its own "set" function using that value. This means the
1678 module's "ERR" operations will use (and modify) the error state in the
1679 application and not in its own statically linked copy of OpenSSL code.
1682 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1683 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1684 the operation, and provides a more encapsulated way for external code
1685 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1686 to use these functions rather than manually incrementing the counts.
1688 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1691 *) Add EVP test program.
1694 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1697 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1698 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1699 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1700 These allow a CRL to be built without having to access X509_CRL fields
1701 directly. Modify 'ca' application to use new functions.
1704 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1705 bug workarounds. Rollback attack detection is a security feature.
1706 The problem will only arise on OpenSSL servers when TLSv1 is not
1707 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1708 Software authors not wanting to support TLSv1 will have special reasons
1709 for their choice and can explicitly enable this option.
1710 [Bodo Moeller, Lutz Jaenicke]
1712 *) Rationalise EVP so it can be extended: don't include a union of
1713 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1714 (similar to those existing for EVP_CIPHER_CTX).
1719 EVP_MD_CTX_init(&md); /* new function call */
1720 EVP_DigestInit(&md, EVP_sha1());
1721 EVP_DigestUpdate(&md, in, len);
1722 EVP_DigestFinal(&md, out, NULL);
1723 EVP_MD_CTX_cleanup(&md); /* new function call */
1727 *) Make DES key schedule conform to the usual scheme, as well as
1728 correcting its structure. This means that calls to DES functions
1729 now have to pass a pointer to a des_key_schedule instead of a
1730 plain des_key_schedule (which was actually always a pointer
1733 des_key_schedule ks;
1735 des_set_key_checked(..., &ks);
1736 des_ncbc_encrypt(..., &ks, ...);
1738 (Note that a later change renames 'des_...' into 'DES_...'.)
1741 *) Initial reduction of linker bloat: the use of some functions, such as
1742 PEM causes large amounts of unused functions to be linked in due to
1743 poor organisation. For example pem_all.c contains every PEM function
1744 which has a knock on effect of linking in large amounts of (unused)
1745 ASN1 code. Grouping together similar functions and splitting unrelated
1746 functions prevents this.
1749 *) Cleanup of EVP macros.
1752 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1753 correct _ecb suffix.
1756 *) Add initial OCSP responder support to ocsp application. The
1757 revocation information is handled using the text based index
1758 use by the ca application. The responder can either handle
1759 requests generated internally, supplied in files (for example
1760 via a CGI script) or using an internal minimal server.
1763 *) Add configuration choices to get zlib compression for TLS.
1766 *) Changes to Kerberos SSL for RFC 2712 compliance:
1767 1. Implemented real KerberosWrapper, instead of just using
1768 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1769 2. Implemented optional authenticator field of KerberosWrapper.
1771 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1772 and authenticator structs; see crypto/krb5/.
1774 Generalized Kerberos calls to support multiple Kerberos libraries.
1775 [Vern Staats <staatsvr@asc.hpc.mil>,
1776 Jeffrey Altman <jaltman@columbia.edu>
1777 via Richard Levitte]
1779 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1780 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1781 values for each of the key sizes rather than having just
1782 parameters (and 'speed' generating keys each time).
1785 *) Speed up EVP routines.
1788 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1789 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1790 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1791 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1793 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1794 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1795 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1798 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1800 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1803 *) Added the OS2-EMX target.
1804 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1806 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1807 to support NCONF routines in extension code. New function CONF_set_nconf()
1808 to allow functions which take an NCONF to also handle the old LHASH
1809 structure: this means that the old CONF compatible routines can be
1810 retained (in particular wrt extensions) without having to duplicate the
1811 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1814 *) Enhance the general user interface with mechanisms for inner control
1815 and with possibilities to have yes/no kind of prompts.
1818 *) Change all calls to low level digest routines in the library and
1819 applications to use EVP. Add missing calls to HMAC_cleanup() and
1820 don't assume HMAC_CTX can be copied using memcpy().
1821 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1823 *) Add the possibility to control engines through control names but with
1824 arbitrary arguments instead of just a string.
1825 Change the key loaders to take a UI_METHOD instead of a callback
1826 function pointer. NOTE: this breaks binary compatibility with earlier
1827 versions of OpenSSL [engine].
1828 Adapt the nCipher code for these new conditions and add a card insertion
1832 *) Enhance the general user interface with mechanisms to better support
1833 dialog box interfaces, application-defined prompts, the possibility
1834 to use defaults (for example default passwords from somewhere else)
1835 and interrupts/cancellations.
1838 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1839 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1842 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1843 tidy up some unnecessarily weird code in 'sk_new()').
1844 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1846 *) Change the key loading routines for ENGINEs to use the same kind
1847 callback (pem_password_cb) as all other routines that need this
1851 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1852 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1853 than this minimum value is recommended.
1856 *) New random seeder for OpenVMS, using the system process statistics
1857 that are easily reachable.
1860 *) Windows apparently can't transparently handle global
1861 variables defined in DLLs. Initialisations such as:
1863 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1865 wont compile. This is used by the any applications that need to
1866 declare their own ASN1 modules. This was fixed by adding the option
1867 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1868 needed for static libraries under Win32.
1871 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1872 setting of purpose and trust fields. New X509_STORE trust and
1873 purpose functions and tidy up setting in other SSL functions.
1876 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1877 structure. These are inherited by X509_STORE_CTX when it is
1878 initialised. This allows various defaults to be set in the
1879 X509_STORE structure (such as flags for CRL checking and custom
1880 purpose or trust settings) for functions which only use X509_STORE_CTX
1881 internally such as S/MIME.
1883 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1884 trust settings if they are not set in X509_STORE. This allows X509_STORE
1885 purposes and trust (in S/MIME for example) to override any set by default.
1887 Add command line options for CRL checking to smime, s_client and s_server
1891 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1892 are set then the CRL is looked up in the X509_STORE structure and
1893 its validity and signature checked, then if the certificate is found
1894 in the CRL the verify fails with a revoked error.
1896 Various new CRL related callbacks added to X509_STORE_CTX structure.
1898 Command line options added to 'verify' application to support this.
1900 This needs some additional work, such as being able to handle multiple
1901 CRLs with different times, extension based lookup (rather than just
1902 by subject name) and ultimately more complete V2 CRL extension
1906 *) Add a general user interface API (crypto/ui/). This is designed
1907 to replace things like des_read_password and friends (backward
1908 compatibility functions using this new API are provided).
1909 The purpose is to remove prompting functions from the DES code
1910 section as well as provide for prompting through dialog boxes in
1911 a window system and the like.
1914 *) Add "ex_data" support to ENGINE so implementations can add state at a
1915 per-structure level rather than having to store it globally.
1918 *) Make it possible for ENGINE structures to be copied when retrieved by
1919 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1920 This causes the "original" ENGINE structure to act like a template,
1921 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1922 operational state can be localised to each ENGINE structure, despite the
1923 fact they all share the same "methods". New ENGINE structures returned in
1924 this case have no functional references and the return value is the single
1925 structural reference. This matches the single structural reference returned
1926 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1930 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1931 needs to match any other type at all we need to manually clear the
1935 *) Changes to the "openssl engine" utility to include;
1936 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1937 about an ENGINE's available control commands.
1938 - executing control commands from command line arguments using the
1939 '-pre' and '-post' switches. '-post' is only used if '-t' is
1940 specified and the ENGINE is successfully initialised. The syntax for
1941 the individual commands are colon-separated, for example;
1942 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1945 *) New dynamic control command support for ENGINEs. ENGINEs can now
1946 declare their own commands (numbers), names (strings), descriptions,
1947 and input types for run-time discovery by calling applications. A
1948 subset of these commands are implicitly classed as "executable"
1949 depending on their input type, and only these can be invoked through
1950 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1951 can be based on user input, config files, etc). The distinction is
1952 that "executable" commands cannot return anything other than a boolean
1953 result and can only support numeric or string input, whereas some
1954 discoverable commands may only be for direct use through
1955 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1956 pointers, or other custom uses. The "executable" commands are to
1957 support parameterisations of ENGINE behaviour that can be
1958 unambiguously defined by ENGINEs and used consistently across any
1959 OpenSSL-based application. Commands have been added to all the
1960 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1961 control over shared-library paths without source code alterations.
1964 *) Changed all ENGINE implementations to dynamically allocate their
1965 ENGINEs rather than declaring them statically. Apart from this being
1966 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1967 this also allows the implementations to compile without using the
1968 internal engine_int.h header.
1971 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1972 'const' value. Any code that should be able to modify a RAND_METHOD
1973 should already have non-const pointers to it (ie. they should only
1974 modify their own ones).
1977 *) Made a variety of little tweaks to the ENGINE code.
1978 - "atalla" and "ubsec" string definitions were moved from header files
1979 to C code. "nuron" string definitions were placed in variables
1980 rather than hard-coded - allowing parameterisation of these values
1981 later on via ctrl() commands.
1982 - Removed unused "#if 0"'d code.
1983 - Fixed engine list iteration code so it uses ENGINE_free() to release
1984 structural references.
1985 - Constified the RAND_METHOD element of ENGINE structures.
1986 - Constified various get/set functions as appropriate and added
1987 missing functions (including a catch-all ENGINE_cpy that duplicates
1988 all ENGINE values onto a new ENGINE except reference counts/state).
1989 - Removed NULL parameter checks in get/set functions. Setting a method
1990 or function to NULL is a way of cancelling out a previously set
1991 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1992 and doesn't justify the extra error symbols and code.
1993 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1994 flags from engine_int.h to engine.h.
1995 - Changed prototypes for ENGINE handler functions (init(), finish(),
1996 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1999 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2000 to the algorithm using long division. The binary algorithm can be
2001 used only if the modulus is odd. On 32-bit systems, it is faster
2002 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2003 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2004 up to 450 bits. In 64-bit environments, the binary algorithm
2005 appears to be advantageous for much longer moduli; here we use it
2006 for moduli up to 2048 bits.
2009 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2010 could not support the combine flag in choice fields.
2013 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2014 extensions from a certificate request to the certificate.
2017 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2018 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2019 file: this allows the display of the certificate about to be
2020 signed to be customised, to allow certain fields to be included
2021 or excluded and extension details. The old system didn't display
2022 multicharacter strings properly, omitted fields not in the policy
2023 and couldn't display additional details such as extensions.
2026 *) Function EC_POINTs_mul for multiple scalar multiplication
2027 of an arbitrary number of elliptic curve points
2028 \sum scalars[i]*points[i],
2029 optionally including the generator defined for the EC_GROUP:
2030 scalar*generator + \sum scalars[i]*points[i].
2032 EC_POINT_mul is a simple wrapper function for the typical case
2033 that the point list has just one item (besides the optional
2037 *) First EC_METHODs for curves over GF(p):
2039 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2040 operations and provides various method functions that can also
2041 operate with faster implementations of modular arithmetic.
2043 EC_GFp_mont_method() reuses most functions that are part of
2044 EC_GFp_simple_method, but uses Montgomery arithmetic.
2046 [Bodo Moeller; point addition and point doubling
2047 implementation directly derived from source code provided by
2048 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2050 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2051 crypto/ec/ec_lib.c):
2053 Curves are EC_GROUP objects (with an optional group generator)
2054 based on EC_METHODs that are built into the library.
2056 Points are EC_POINT objects based on EC_GROUP objects.
2058 Most of the framework would be able to handle curves over arbitrary
2059 finite fields, but as there are no obvious types for fields other
2060 than GF(p), some functions are limited to that for now.
2063 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2064 that the file contains a complete HTTP response.
2067 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2068 change the def and num file printf format specifier from "%-40sXXX"
2069 to "%-39s XXX". The latter will always guarantee a space after the
2070 field while the former will cause them to run together if the field
2071 is 40 of more characters long.
2074 *) Constify the cipher and digest 'method' functions and structures
2075 and modify related functions to take constant EVP_MD and EVP_CIPHER
2079 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2080 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2083 *) Modify EVP_Digest*() routines so they now return values. Although the
2084 internal software routines can never fail additional hardware versions
2088 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2090 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2091 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2094 ERR_R_NESTED_ASN1_ERROR
2096 ERR_R_MISSING_ASN1_EOS
2097 were 4 .. 9, conflicting with
2098 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2100 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2101 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2103 Add new error code 'ERR_R_INTERNAL_ERROR'.
2106 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2110 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2111 sets the subject name for a new request or supersedes the
2112 subject name in a given request. Formats that can be parsed are
2113 'CN=Some Name, OU=myOU, C=IT'
2115 'CN=Some Name/OU=myOU/C=IT'.
2117 Add options '-batch' and '-verbose' to 'openssl req'.
2118 [Massimiliano Pala <madwolf@hackmasters.net>]
2120 *) Introduce the possibility to access global variables through
2121 functions on platform were that's the best way to handle exporting
2122 global variables in shared libraries. To enable this functionality,
2123 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2124 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2125 is normally done by Configure or something similar).
2127 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2128 in the source file (foo.c) like this:
2130 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2131 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2133 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2134 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2136 OPENSSL_DECLARE_GLOBAL(int,foo);
2137 #define foo OPENSSL_GLOBAL_REF(foo)
2138 OPENSSL_DECLARE_GLOBAL(double,bar);
2139 #define bar OPENSSL_GLOBAL_REF(bar)
2141 The #defines are very important, and therefore so is including the
2142 header file everywhere where the defined globals are used.
2144 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2145 of ASN.1 items, but that structure is a bit different.
2147 The largest change is in util/mkdef.pl which has been enhanced with
2148 better and easier to understand logic to choose which symbols should
2149 go into the Windows .def files as well as a number of fixes and code
2150 cleanup (among others, algorithm keywords are now sorted
2151 lexicographically to avoid constant rewrites).
2154 *) In BN_div() keep a copy of the sign of 'num' before writing the
2155 result to 'rm' because if rm==num the value will be overwritten
2156 and produce the wrong result if 'num' is negative: this caused
2157 problems with BN_mod() and BN_nnmod().
2160 *) Function OCSP_request_verify(). This checks the signature on an
2161 OCSP request and verifies the signer certificate. The signer
2162 certificate is just checked for a generic purpose and OCSP request
2166 *) Add OCSP_check_validity() function to check the validity of OCSP
2167 responses. OCSP responses are prepared in real time and may only
2168 be a few seconds old. Simply checking that the current time lies
2169 between thisUpdate and nextUpdate max reject otherwise valid responses
2170 caused by either OCSP responder or client clock inaccuracy. Instead
2171 we allow thisUpdate and nextUpdate to fall within a certain period of
2172 the current time. The age of the response can also optionally be
2173 checked. Two new options -validity_period and -status_age added to
2177 *) If signature or public key algorithm is unrecognized print out its
2178 OID rather that just UNKNOWN.
2181 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2182 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2183 ID to be generated from the issuer certificate alone which can then be
2184 passed to OCSP_id_issuer_cmp().
2187 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2188 ASN1 modules to export functions returning ASN1_ITEM pointers
2189 instead of the ASN1_ITEM structures themselves. This adds several
2190 new macros which allow the underlying ASN1 function/structure to
2191 be accessed transparently. As a result code should not use ASN1_ITEM
2192 references directly (such as &X509_it) but instead use the relevant
2193 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2194 use of the new ASN1 code on platforms where exporting structures
2195 is problematical (for example in shared libraries) but exporting
2196 functions returning pointers to structures is not.
2199 *) Add support for overriding the generation of SSL/TLS session IDs.
2200 These callbacks can be registered either in an SSL_CTX or per SSL.
2201 The purpose of this is to allow applications to control, if they wish,
2202 the arbitrary values chosen for use as session IDs, particularly as it
2203 can be useful for session caching in multiple-server environments. A
2204 command-line switch for testing this (and any client code that wishes
2205 to use such a feature) has been added to "s_server".
2206 [Geoff Thorpe, Lutz Jaenicke]
2208 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2209 of the form '#if defined(...) || defined(...) || ...' and
2210 '#if !defined(...) && !defined(...) && ...'. This also avoids
2211 the growing number of special cases it was previously handling.
2214 *) Make all configuration macros available for application by making
2215 sure they are available in opensslconf.h, by giving them names starting
2216 with "OPENSSL_" to avoid conflicts with other packages and by making
2217 sure e_os2.h will cover all platform-specific cases together with
2219 Additionally, it is now possible to define configuration/platform-
2220 specific names (called "system identities"). In the C code, these
2221 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2222 macro with the name beginning with "OPENSSL_SYS_", which is determined
2223 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2227 *) New option -set_serial to 'req' and 'x509' this allows the serial
2228 number to use to be specified on the command line. Previously self
2229 signed certificates were hard coded with serial number 0 and the
2230 CA options of 'x509' had to use a serial number in a file which was
2234 *) New options to 'ca' utility to support V2 CRL entry extensions.
2235 Currently CRL reason, invalidity date and hold instruction are
2236 supported. Add new CRL extensions to V3 code and some new objects.
2239 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2240 disable standard block padding (aka PKCS#5 padding) in the EVP
2241 API, which was previously mandatory. This means that the data is
2242 not padded in any way and so the total length much be a multiple
2243 of the block size, otherwise an error occurs.
2246 *) Initial (incomplete) OCSP SSL support.
2249 *) New function OCSP_parse_url(). This splits up a URL into its host,
2250 port and path components: primarily to parse OCSP URLs. New -url
2251 option to ocsp utility.
2254 *) New nonce behavior. The return value of OCSP_check_nonce() now
2255 reflects the various checks performed. Applications can decide
2256 whether to tolerate certain situations such as an absent nonce
2257 in a response when one was present in a request: the ocsp application
2258 just prints out a warning. New function OCSP_add1_basic_nonce()
2259 this is to allow responders to include a nonce in a response even if
2260 the request is nonce-less.
2263 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2264 skipped when using openssl x509 multiple times on a single input file,
2265 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2268 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2269 set string type: to handle setting ASN1_TIME structures. Fix ca
2270 utility to correctly initialize revocation date of CRLs.
2273 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2274 the clients preferred ciphersuites and rather use its own preferences.
2275 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2276 Internet Explorer by ensuring unchanged hash method during stepup.
2277 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2280 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2281 to aes and add a new 'exist' option to print out symbols that don't
2285 *) Additional options to ocsp utility to allow flags to be set and
2286 additional certificates supplied.
2289 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2290 OCSP client a number of certificate to only verify the response
2294 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2295 handle the new API. Currently only ECB, CBC modes supported. Add new
2298 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2299 Encryption Standard (AES) Ciphersuites for Transport Layer
2300 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2301 not enabled by default and were not part of the "ALL" ciphersuite
2302 alias because they were not yet official; they could be
2303 explicitly requested by specifying the "AESdraft" ciphersuite
2304 group alias. In the final release of OpenSSL 0.9.7, the group
2305 alias is called "AES" and is part of "ALL".)
2306 [Ben Laurie, Steve Henson, Bodo Moeller]
2308 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2309 request to response.
2312 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2313 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2314 extract information from a certificate request. OCSP_response_create()
2315 creates a response and optionally adds a basic response structure.
2316 OCSP_basic_add1_status() adds a complete single response to a basic
2317 response and returns the OCSP_SINGLERESP structure just added (to allow
2318 extensions to be included for example). OCSP_basic_add1_cert() adds a
2319 certificate to a basic response and OCSP_basic_sign() signs a basic
2320 response with various flags. New helper functions ASN1_TIME_check()
2321 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2322 (converts ASN1_TIME to GeneralizedTime).
2325 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2326 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2327 structure from a certificate. X509_pubkey_digest() digests the public_key
2328 contents: this is used in various key identifiers.
2331 *) Make sk_sort() tolerate a NULL argument.
2332 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2334 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2335 passed by the function are trusted implicitly. If any of them signed the
2336 response then it is assumed to be valid and is not verified.
2339 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2340 to data. This was previously part of the PKCS7 ASN1 code. This
2341 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2342 [Steve Henson, reported by Kenneth R. Robinette
2343 <support@securenetterm.com>]
2345 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2346 routines: without these tracing memory leaks is very painful.
2347 Fix leaks in PKCS12 and PKCS7 routines.
2350 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2351 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2352 effectively meant GeneralizedTime would never be used. Now it
2353 is initialised to -1 but X509_time_adj() now has to check the value
2354 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2355 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2356 [Steve Henson, reported by Kenneth R. Robinette
2357 <support@securenetterm.com>]
2359 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2360 result in a zero length in the ASN1_INTEGER structure which was
2361 not consistent with the structure when d2i_ASN1_INTEGER() was used
2362 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2363 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2364 where it did not print out a minus for negative ASN1_INTEGER.
2367 *) Add summary printout to ocsp utility. The various functions which
2368 convert status values to strings have been renamed to:
2369 OCSP_response_status_str(), OCSP_cert_status_str() and
2370 OCSP_crl_reason_str() and are no longer static. New options
2371 to verify nonce values and to disable verification. OCSP response
2372 printout format cleaned up.
2375 *) Add additional OCSP certificate checks. These are those specified
2376 in RFC2560. This consists of two separate checks: the CA of the
2377 certificate being checked must either be the OCSP signer certificate
2378 or the issuer of the OCSP signer certificate. In the latter case the
2379 OCSP signer certificate must contain the OCSP signing extended key
2380 usage. This check is performed by attempting to match the OCSP
2381 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2382 in the OCSP_CERTID structures of the response.
2385 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2386 and related routines. This uses the standard OpenSSL certificate
2387 verify routines to perform initial checks (just CA validity) and
2388 to obtain the certificate chain. Then additional checks will be
2389 performed on the chain. Currently the root CA is checked to see
2390 if it is explicitly trusted for OCSP signing. This is used to set
2391 a root CA as a global signing root: that is any certificate that
2392 chains to that CA is an acceptable OCSP signing certificate.
2395 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2396 extensions from a separate configuration file.
2397 As when reading extensions from the main configuration file,
2398 the '-extensions ...' option may be used for specifying the
2400 [Massimiliano Pala <madwolf@comune.modena.it>]
2402 *) New OCSP utility. Allows OCSP requests to be generated or
2403 read. The request can be sent to a responder and the output
2404 parsed, outputed or printed in text form. Not complete yet:
2405 still needs to check the OCSP response validity.
2408 *) New subcommands for 'openssl ca':
2409 'openssl ca -status <serial>' prints the status of the cert with
2410 the given serial number (according to the index file).
2411 'openssl ca -updatedb' updates the expiry status of certificates
2413 [Massimiliano Pala <madwolf@comune.modena.it>]
2415 *) New '-newreq-nodes' command option to CA.pl. This is like
2416 '-newreq', but calls 'openssl req' with the '-nodes' option
2417 so that the resulting key is not encrypted.
2418 [Damien Miller <djm@mindrot.org>]
2420 *) New configuration for the GNU Hurd.
2421 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2423 *) Initial code to implement OCSP basic response verify. This
2424 is currently incomplete. Currently just finds the signer's
2425 certificate and verifies the signature on the response.
2428 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2429 value of OPENSSLDIR. This is available via the new '-d' option
2430 to 'openssl version', and is also included in 'openssl version -a'.
2433 *) Allowing defining memory allocation callbacks that will be given
2434 file name and line number information in additional arguments
2435 (a const char* and an int). The basic functionality remains, as
2436 well as the original possibility to just replace malloc(),
2437 realloc() and free() by functions that do not know about these
2438 additional arguments. To register and find out the current
2439 settings for extended allocation functions, the following
2440 functions are provided:
2442 CRYPTO_set_mem_ex_functions
2443 CRYPTO_set_locked_mem_ex_functions
2444 CRYPTO_get_mem_ex_functions
2445 CRYPTO_get_locked_mem_ex_functions
2447 These work the same way as CRYPTO_set_mem_functions and friends.
2448 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2449 extended allocation function is enabled.
2450 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2451 a conventional allocation function is enabled.
2452 [Richard Levitte, Bodo Moeller]
2454 *) Finish off removing the remaining LHASH function pointer casts.
2455 There should no longer be any prototype-casting required when using
2456 the LHASH abstraction, and any casts that remain are "bugs". See
2457 the callback types and macros at the head of lhash.h for details
2458 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2461 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2462 If /dev/[u]random devices are not available or do not return enough
2463 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2465 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2466 /etc/entropy will be queried once each in this sequence, quering stops
2467 when enough entropy was collected without querying more sockets.
2470 *) Change the Unix RAND_poll() variant to be able to poll several
2471 random devices, as specified by DEVRANDOM, until a sufficient amount
2472 of data has been collected. We spend at most 10 ms on each file
2473 (select timeout) and read in non-blocking mode. DEVRANDOM now
2474 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2475 (previously it was just the string "/dev/urandom"), so on typical
2476 platforms the 10 ms delay will never occur.
2477 Also separate out the Unix variant to its own file, rand_unix.c.
2478 For VMS, there's a currently-empty rand_vms.c.
2481 *) Move OCSP client related routines to ocsp_cl.c. These
2482 provide utility functions which an application needing
2483 to issue a request to an OCSP responder and analyse the
2484 response will typically need: as opposed to those which an
2485 OCSP responder itself would need which will be added later.
2487 OCSP_request_sign() signs an OCSP request with an API similar
2488 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2489 response. OCSP_response_get1_basic() extracts basic response
2490 from response. OCSP_resp_find_status(): finds and extracts status
2491 information from an OCSP_CERTID structure (which will be created
2492 when the request structure is built). These are built from lower
2493 level functions which work on OCSP_SINGLERESP structures but
2494 wont normally be used unless the application wishes to examine
2495 extensions in the OCSP response for example.
2497 Replace nonce routines with a pair of functions.
2498 OCSP_request_add1_nonce() adds a nonce value and optionally
2499 generates a random value. OCSP_check_nonce() checks the
2500 validity of the nonce in an OCSP response.
2503 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2504 This doesn't copy the supplied OCSP_CERTID and avoids the
2505 need to free up the newly created id. Change return type
2506 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2507 This can then be used to add extensions to the request.
2508 Deleted OCSP_request_new(), since most of its functionality
2509 is now in OCSP_REQUEST_new() (and the case insensitive name
2510 clash) apart from the ability to set the request name which
2511 will be added elsewhere.
2514 *) Update OCSP API. Remove obsolete extensions argument from
2515 various functions. Extensions are now handled using the new
2516 OCSP extension code. New simple OCSP HTTP function which
2517 can be used to send requests and parse the response.
2520 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2521 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2522 uses the special reorder version of SET OF to sort the attributes
2523 and reorder them to match the encoded order. This resolves a long
2524 standing problem: a verify on a PKCS7 structure just after signing
2525 it used to fail because the attribute order did not match the
2526 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2527 it uses the received order. This is necessary to tolerate some broken
2528 software that does not order SET OF. This is handled by encoding
2529 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2530 to produce the required SET OF.
2533 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2534 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2535 files to get correct declarations of the ASN.1 item variables.
2538 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2539 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2540 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2541 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2542 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2543 ASN1_ITEM and no wrapper functions.
2546 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2547 replace the old function pointer based I/O routines. Change most of
2548 the *_d2i_bio() and *_d2i_fp() functions to use these.
2551 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2552 lines, recognice more "algorithms" that can be deselected, and make
2553 it complain about algorithm deselection that isn't recognised.
2556 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2557 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2558 to use new functions. Add NO_ASN1_OLD which can be set to remove
2559 some old style ASN1 functions: this can be used to determine if old
2560 code will still work when these eventually go away.
2563 *) New extension functions for OCSP structures, these follow the
2564 same conventions as certificates and CRLs.
2567 *) New function X509V3_add1_i2d(). This automatically encodes and
2568 adds an extension. Its behaviour can be customised with various
2569 flags to append, replace or delete. Various wrappers added for
2570 certifcates and CRLs.
2573 *) Fix to avoid calling the underlying ASN1 print routine when
2574 an extension cannot be parsed. Correct a typo in the
2575 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2578 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2579 entries for variables.
2582 *) Add functionality to apps/openssl.c for detecting locking
2583 problems: As the program is single-threaded, all we have
2584 to do is register a locking callback using an array for
2585 storing which locks are currently held by the program.
2588 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2589 SSL_get_ex_data_X509_STORE_idx(), which is used in
2590 ssl_verify_cert_chain() and thus can be called at any time
2591 during TLS/SSL handshakes so that thread-safety is essential.
2592 Unfortunately, the ex_data design is not at all suited
2593 for multi-threaded use, so it probably should be abolished.
2596 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2597 [Broadcom, tweaked and integrated by Geoff Thorpe]
2599 *) Move common extension printing code to new function
2600 X509V3_print_extensions(). Reorganise OCSP print routines and
2601 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2604 *) New function X509_signature_print() to remove duplication in some
2608 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2609 set (this was treated exactly the same as SET OF previously). This
2610 is used to reorder the STACK representing the structure to match the
2611 encoding. This will be used to get round a problem where a PKCS7
2612 structure which was signed could not be verified because the STACK
2613 order did not reflect the encoded order.
2616 *) Reimplement the OCSP ASN1 module using the new code.
2619 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2620 for its ASN1 operations. The old style function pointers still exist
2621 for now but they will eventually go away.
2624 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2625 completely replaces the old ASN1 functionality with a table driven
2626 encoder and decoder which interprets an ASN1_ITEM structure describing
2627 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2628 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2629 has also been converted to the new form.
2632 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2633 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2634 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2635 for negative moduli.
2638 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2639 of not touching the result's sign bit.
2642 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2646 *) Changed the LHASH code to use prototypes for callbacks, and created
2647 macros to declare and implement thin (optionally static) functions
2648 that provide type-safety and avoid function pointer casting for the
2649 type-specific callbacks.
2652 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2654 [Veers Staats <staatsvr@asc.hpc.mil>,
2655 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2657 *) Reformat the FAQ so the different questions and answers can be divided
2658 in sections depending on the subject.
2661 *) Have the zlib compression code load ZLIB.DLL dynamically under
2665 *) New function BN_mod_sqrt for computing square roots modulo a prime
2666 (using the probabilistic Tonelli-Shanks algorithm unless
2667 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2668 be handled deterministically).
2669 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2671 *) Make BN_mod_inverse faster by explicitly handling small quotients
2672 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2673 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2676 *) New function BN_kronecker.
2679 *) Fix BN_gcd so that it works on negative inputs; the result is
2680 positive unless both parameters are zero.
2681 Previously something reasonably close to an infinite loop was
2682 possible because numbers could be growing instead of shrinking
2683 in the implementation of Euclid's algorithm.
2686 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2687 sign of the number in question.
2689 Fix BN_is_word(a,w) to work correctly for w == 0.
2691 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2692 because its test if the absolute value of 'a' equals 'w'.
2693 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2694 it exists mostly for use in the implementations of BN_is_zero(),
2695 BN_is_one(), and BN_is_word().
2698 *) New function BN_swap.
2701 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2702 the exponentiation functions are more likely to produce reasonable
2703 results on negative inputs.
2706 *) Change BN_mod_mul so that the result is always non-negative.
2707 Previously, it could be negative if one of the factors was negative;
2708 I don't think anyone really wanted that behaviour.
2711 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2712 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2713 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2714 and add new functions:
2723 BN_mod_lshift1_quick
2727 These functions always generate non-negative results.
2729 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2730 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2732 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2733 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2734 be reduced modulo m.
2735 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2738 The following entry accidentily appeared in the CHANGES file
2739 distributed with OpenSSL 0.9.7. The modifications described in
2740 it do *not* apply to OpenSSL 0.9.7.
2742 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2743 was actually never needed) and in BN_mul(). The removal in BN_mul()
2744 required a small change in bn_mul_part_recursive() and the addition
2745 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2746 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2747 bn_sub_words() and bn_add_words() except they take arrays with
2752 *) In 'openssl passwd', verify passwords read from the terminal
2753 unless the '-salt' option is used (which usually means that
2754 verification would just waste user's time since the resulting
2755 hash is going to be compared with some given password hash)
2756 or the new '-noverify' option is used.
2758 This is an incompatible change, but it does not affect
2759 non-interactive use of 'openssl passwd' (passwords on the command
2760 line, '-stdin' option, '-in ...' option) and thus should not
2764 *) Remove all references to RSAref, since there's no more need for it.
2767 *) Make DSO load along a path given through an environment variable
2768 (SHLIB_PATH) with shl_load().
2771 *) Constify the ENGINE code as a result of BIGNUM constification.
2772 Also constify the RSA code and most things related to it. In a
2773 few places, most notable in the depth of the ASN.1 code, ugly
2774 casts back to non-const were required (to be solved at a later
2778 *) Make it so the openssl application has all engines loaded by default.
2781 *) Constify the BIGNUM routines a little more.
2784 *) Add the following functions:
2786 ENGINE_load_cswift()
2788 ENGINE_load_atalla()
2790 ENGINE_load_builtin_engines()
2792 That way, an application can itself choose if external engines that
2793 are built-in in OpenSSL shall ever be used or not. The benefit is
2794 that applications won't have to be linked with libdl or other dso
2795 libraries unless it's really needed.
2797 Changed 'openssl engine' to load all engines on demand.
2798 Changed the engine header files to avoid the duplication of some
2799 declarations (they differed!).
2802 *) 'openssl engine' can now list capabilities.
2805 *) Better error reporting in 'openssl engine'.
2808 *) Never call load_dh_param(NULL) in s_server.
2811 *) Add engine application. It can currently list engines by name and
2812 identity, and test if they are actually available.
2815 *) Improve RPM specification file by forcing symbolic linking and making
2816 sure the installed documentation is also owned by root.root.
2817 [Damien Miller <djm@mindrot.org>]
2819 *) Give the OpenSSL applications more possibilities to make use of
2820 keys (public as well as private) handled by engines.
2823 *) Add OCSP code that comes from CertCo.
2826 *) Add VMS support for the Rijndael code.
2829 *) Added untested support for Nuron crypto accelerator.
2832 *) Add support for external cryptographic devices. This code was
2833 previously distributed separately as the "engine" branch.
2834 [Geoff Thorpe, Richard Levitte]
2836 *) Rework the filename-translation in the DSO code. It is now possible to
2837 have far greater control over how a "name" is turned into a filename
2838 depending on the operating environment and any oddities about the
2839 different shared library filenames on each system.
2842 *) Support threads on FreeBSD-elf in Configure.
2845 *) Fix for SHA1 assembly problem with MASM: it produces
2846 warnings about corrupt line number information when assembling
2847 with debugging information. This is caused by the overlapping
2849 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2852 NCONF_get_number() has no error checking at all. As a replacement,
2853 NCONF_get_number_e() is defined (_e for "error checking") and is
2854 promoted strongly. The old NCONF_get_number is kept around for
2855 binary backward compatibility.
2856 Make it possible for methods to load from something other than a BIO,
2857 by providing a function pointer that is given a name instead of a BIO.
2858 For example, this could be used to load configuration data from an
2862 *) Fix for non blocking accept BIOs. Added new I/O special reason
2863 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2864 with non blocking I/O was not possible because no retry code was
2865 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2869 *) Added the beginnings of Rijndael support.
2872 *) Fix for bug in DirectoryString mask setting. Add support for
2873 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2874 to allow certificate printing to more controllable, additional
2875 'certopt' option to 'x509' to allow new printing options to be
2879 *) Clean old EAY MD5 hack from e_os.h.
2882 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2884 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2885 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2886 [Joe Orton, Steve Henson]
2888 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2890 *) Fix additional bug revealed by the NISCC test suite:
2892 Stop bug triggering large recursion when presented with
2893 certain ASN.1 tags (CAN-2003-0851)
2896 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2898 *) Fix various bugs revealed by running the NISCC test suite:
2900 Stop out of bounds reads in the ASN1 code when presented with
2901 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2903 If verify callback ignores invalid public key errors don't try to check
2904 certificate signature with the NULL public key.
2908 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2909 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2913 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2914 extra data after the compression methods not only for TLS 1.0
2915 but also for SSL 3.0 (as required by the specification).
2916 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2918 *) Change X509_certificate_type() to mark the key as exported/exportable
2919 when it's 512 *bits* long, not 512 bytes.
2922 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2924 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2925 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2926 a protocol version number mismatch like a decryption error
2927 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2930 *) Turn on RSA blinding by default in the default implementation
2931 to avoid a timing attack. Applications that don't want it can call
2932 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2933 They would be ill-advised to do so in most cases.
2934 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2936 *) Change RSA blinding code so that it works when the PRNG is not
2937 seeded (in this case, the secret RSA exponent is abused as
2938 an unpredictable seed -- if it is not unpredictable, there
2939 is no point in blinding anyway). Make RSA blinding thread-safe
2940 by remembering the creator's thread ID in rsa->blinding and
2941 having all other threads use local one-time blinding factors
2942 (this requires more computation than sharing rsa->blinding, but
2943 avoids excessive locking; and if an RSA object is not shared
2944 between threads, blinding will still be very fast).
2947 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2949 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2950 via timing by performing a MAC computation even if incorrrect
2951 block cipher padding has been found. This is a countermeasure
2952 against active attacks where the attacker has to distinguish
2953 between bad padding and a MAC verification error. (CAN-2003-0078)
2955 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2956 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2957 Martin Vuagnoux (EPFL, Ilion)]
2959 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2961 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2962 memory from it's contents. This is done with a counter that will
2963 place alternating values in each byte. This can be used to solve
2964 two issues: 1) the removal of calls to memset() by highly optimizing
2965 compilers, and 2) cleansing with other values than 0, since those can
2966 be read through on certain media, for example a swap space on disk.
2969 *) Bugfix: client side session caching did not work with external caching,
2970 because the session->cipher setting was not restored when reloading
2971 from the external cache. This problem was masked, when
2972 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2973 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2976 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2977 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2978 [Zeev Lieber <zeev-l@yahoo.com>]
2980 *) Undo an undocumented change introduced in 0.9.6e which caused
2981 repeated calls to OpenSSL_add_all_ciphers() and
2982 OpenSSL_add_all_digests() to be ignored, even after calling
2986 *) Change the default configuration reader to deal with last line not
2987 being properly terminated.
2990 *) Change X509_NAME_cmp() so it applies the special rules on handling
2991 DN values that are of type PrintableString, as well as RDNs of type
2992 emailAddress where the value has the type ia5String.
2993 [stefank@valicert.com via Richard Levitte]
2995 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2996 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2997 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2998 the bitwise-OR of the two for use by the majority of applications
2999 wanting this behaviour, and update the docs. The documented
3000 behaviour and actual behaviour were inconsistent and had been
3001 changing anyway, so this is more a bug-fix than a behavioural
3003 [Geoff Thorpe, diagnosed by Nadav Har'El]
3005 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3006 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3009 *) Fix initialization code race conditions in
3010 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3011 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3012 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3013 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3014 ssl2_get_cipher_by_char(),
3015 ssl3_get_cipher_by_char().
3016 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3018 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3019 the cached sessions are flushed, as the remove_cb() might use ex_data
3020 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3021 (see [openssl.org #212]).
3022 [Geoff Thorpe, Lutz Jaenicke]
3024 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3025 length, instead of the encoding length to d2i_ASN1_OBJECT.
3028 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3030 *) [In 0.9.6g-engine release:]
3031 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3032 [Lynn Gazis <lgazis@rainbow.com>]
3034 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3036 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3037 and get fix the header length calculation.
3038 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3039 Alon Kantor <alonk@checkpoint.com> (and others),
3042 *) Use proper error handling instead of 'assertions' in buffer
3043 overflow checks added in 0.9.6e. This prevents DoS (the
3044 assertions could call abort()).
3045 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3047 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3049 *) Add various sanity checks to asn1_get_length() to reject
3050 the ASN1 length bytes if they exceed sizeof(long), will appear
3051 negative or the content length exceeds the length of the
3053 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3055 *) Fix cipher selection routines: ciphers without encryption had no flags
3056 for the cipher strength set and where therefore not handled correctly
3057 by the selection routines (PR #130).
3060 *) Fix EVP_dsa_sha macro.
3064 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3065 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3066 that was added in OpenSSL 0.9.6d.
3068 As the countermeasure turned out to be incompatible with some
3069 broken SSL implementations, the new option is part of SSL_OP_ALL.
3070 SSL_OP_ALL is usually employed when compatibility with weird SSL
3071 implementations is desired (e.g. '-bugs' option to 's_client' and
3072 's_server'), so the new option is automatically set in many
3076 *) Changes in security patch:
3078 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3079 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3080 Air Force Materiel Command, USAF, under agreement number
3083 *) Add various sanity checks to asn1_get_length() to reject
3084 the ASN1 length bytes if they exceed sizeof(long), will appear
3085 negative or the content length exceeds the length of the
3086 supplied buffer. (CAN-2002-0659)
3087 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3089 *) Assertions for various potential buffer overflows, not known to
3091 [Ben Laurie (CHATS)]
3093 *) Various temporary buffers to hold ASCII versions of integers were
3094 too small for 64 bit platforms. (CAN-2002-0655)
3095 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3097 *) Remote buffer overflow in SSL3 protocol - an attacker could
3098 supply an oversized session ID to a client. (CAN-2002-0656)
3099 [Ben Laurie (CHATS)]
3101 *) Remote buffer overflow in SSL2 protocol - an attacker could
3102 supply an oversized client master key. (CAN-2002-0656)
3103 [Ben Laurie (CHATS)]
3105 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3107 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3108 encoded as NULL) with id-dsa-with-sha1.
3109 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3111 *) Check various X509_...() return values in apps/req.c.
3112 [Nils Larsch <nla@trustcenter.de>]
3114 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3115 an end-of-file condition would erronously be flagged, when the CRLF
3116 was just at the end of a processed block. The bug was discovered when
3117 processing data through a buffering memory BIO handing the data to a
3118 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3119 <ptsekov@syntrex.com> and Nedelcho Stanev.
3122 *) Implement a countermeasure against a vulnerability recently found
3123 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3124 before application data chunks to avoid the use of known IVs
3125 with data potentially chosen by the attacker.
3128 *) Fix length checks in ssl3_get_client_hello().
3131 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3132 to prevent ssl3_read_internal() from incorrectly assuming that
3133 ssl3_read_bytes() found application data while handshake
3134 processing was enabled when in fact s->s3->in_read_app_data was
3135 merely automatically cleared during the initial handshake.
3136 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3138 *) Fix object definitions for Private and Enterprise: they were not
3139 recognized in their shortname (=lowercase) representation. Extend
3140 obj_dat.pl to issue an error when using undefined keywords instead
3141 of silently ignoring the problem (Svenning Sorensen
3142 <sss@sss.dnsalias.net>).
3145 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3146 generators, i.e. generators other than 2 and 5. (Previously, the
3147 code did not properly initialise the 'add' and 'rem' values to
3148 BN_generate_prime().)
3150 In the new general case, we do not insist that 'generator' is
3151 actually a primitive root: This requirement is rather pointless;
3152 a generator of the order-q subgroup is just as good, if not
3156 *) Map new X509 verification errors to alerts. Discovered and submitted by
3157 Tom Wu <tom@arcot.com>.
3160 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3161 returning non-zero before the data has been completely received
3162 when using non-blocking I/O.
3163 [Bodo Moeller; problem pointed out by John Hughes]
3165 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3166 [Ben Laurie, Lutz Jaenicke]
3168 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3169 Yoram Zahavi <YoramZ@gilian.com>).
3172 *) Add information about CygWin 1.3 and on, and preserve proper
3173 configuration for the versions before that.
3174 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3176 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3177 check whether we deal with a copy of a session and do not delete from
3178 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3179 <izhar@checkpoint.com>.
3182 *) Do not store session data into the internal session cache, if it
3183 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3184 flag is set). Proposed by Aslam <aslam@funk.com>.
3187 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3191 *) [In 0.9.6d-engine release:]
3192 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3193 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3195 *) Add the configuration target linux-s390x.
3196 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3198 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3199 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3200 variable as an indication that a ClientHello message has been
3201 received. As the flag value will be lost between multiple
3202 invocations of ssl3_accept when using non-blocking I/O, the
3203 function may not be aware that a handshake has actually taken
3204 place, thus preventing a new session from being added to the
3207 To avoid this problem, we now set s->new_session to 2 instead of
3208 using a local variable.
3209 [Lutz Jaenicke, Bodo Moeller]
3211 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3212 if the SSL_R_LENGTH_MISMATCH error is detected.
3213 [Geoff Thorpe, Bodo Moeller]
3215 *) New 'shared_ldflag' column in Configure platform table.
3218 *) Fix EVP_CIPHER_mode macro.
3219 ["Dan S. Camper" <dan@bti.net>]
3221 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3222 type, we must throw them away by setting rr->length to 0.
3223 [D P Chang <dpc@qualys.com>]
3225 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3227 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3228 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3229 worked incorrectly for those cases where range = 10..._2 and
3230 3*range is two bits longer than range.)
3233 *) Only add signing time to PKCS7 structures if it is not already
3237 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3238 OBJ_ld_ce should be OBJ_id_ce.
3239 Also some ip-pda OIDs in crypto/objects/objects.txt were
3240 incorrect (cf. RFC 3039).
3241 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3243 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3244 returns early because it has nothing to do.
3245 [Andy Schneider <andy.schneider@bjss.co.uk>]
3247 *) [In 0.9.6c-engine release:]
3248 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3249 [Andy Schneider <andy.schneider@bjss.co.uk>]
3251 *) [In 0.9.6c-engine release:]
3252 Add support for Cryptographic Appliance's keyserver technology.
3253 (Use engine 'keyclient')
3254 [Cryptographic Appliances and Geoff Thorpe]
3256 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3257 is called via tools/c89.sh because arguments have to be
3258 rearranged (all '-L' options must appear before the first object
3260 [Richard Shapiro <rshapiro@abinitio.com>]
3262 *) [In 0.9.6c-engine release:]
3263 Add support for Broadcom crypto accelerator cards, backported
3265 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3267 *) [In 0.9.6c-engine release:]
3268 Add support for SureWare crypto accelerator cards from
3269 Baltimore Technologies. (Use engine 'sureware')
3270 [Baltimore Technologies and Mark Cox]
3272 *) [In 0.9.6c-engine release:]
3273 Add support for crypto accelerator cards from Accelerated
3274 Encryption Processing, www.aep.ie. (Use engine 'aep')
3275 [AEP Inc. and Mark Cox]
3277 *) Add a configuration entry for gcc on UnixWare.
3278 [Gary Benson <gbenson@redhat.com>]
3280 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3281 messages are stored in a single piece (fixed-length part and
3282 variable-length part combined) and fix various bugs found on the way.
3285 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3286 instead. BIO_gethostbyname() does not know what timeouts are
3287 appropriate, so entries would stay in cache even when they have
3289 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3291 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3292 faced with a pathologically small ClientHello fragment that does
3293 not contain client_version: Instead of aborting with an error,
3294 simply choose the highest available protocol version (i.e.,
3295 TLS 1.0 unless it is disabled). In practice, ClientHello
3296 messages are never sent like this, but this change gives us
3297 strictly correct behaviour at least for TLS.
3300 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3301 never resets s->method to s->ctx->method when called from within
3302 one of the SSL handshake functions.
3303 [Bodo Moeller; problem pointed out by Niko Baric]
3305 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3306 (sent using the client's version number) if client_version is
3307 smaller than the protocol version in use. Also change
3308 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3309 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3310 the client will at least see that alert.
3313 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3317 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3318 client receives HelloRequest while in a handshake.
3319 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3321 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3322 should end in 'break', not 'goto end' which circuments various
3323 cleanups done in state SSL_ST_OK. But session related stuff
3324 must be disabled for SSL_ST_OK in the case that we just sent a
3327 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3328 before just sending a HelloRequest.
3329 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3331 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3332 reveal whether illegal block cipher padding was found or a MAC
3333 verification error occured. (Neither SSLerr() codes nor alerts
3334 are directly visible to potential attackers, but the information
3335 may leak via logfiles.)
3337 Similar changes are not required for the SSL 2.0 implementation
3338 because the number of padding bytes is sent in clear for SSL 2.0,
3339 and the extra bytes are just ignored. However ssl/s2_pkt.c
3340 failed to verify that the purported number of padding bytes is in
3344 *) Add OpenUNIX-8 support including shared libraries
3345 (Boyd Lynn Gerber <gerberb@zenez.com>).
3348 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3349 'wristwatch attack' using huge encoding parameters (cf.
3350 James H. Manger's CRYPTO 2001 paper). Note that the
3351 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3352 encoding parameters and hence was not vulnerable.
3355 *) BN_sqr() bug fix.
3356 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3358 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3359 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3360 followed by modular reduction.
3361 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3363 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3364 equivalent based on BN_pseudo_rand() instead of BN_rand().
3367 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3368 This function was broken, as the check for a new client hello message
3369 to handle SGC did not allow these large messages.
3370 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3373 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3376 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3377 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3380 *) Rework the configuration and shared library support for Tru64 Unix.
3381 The configuration part makes use of modern compiler features and
3382 still retains old compiler behavior for those that run older versions
3383 of the OS. The shared library support part includes a variant that
3384 uses the RPATH feature, and is available through the special
3385 configuration target "alpha-cc-rpath", which will never be selected
3387 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3389 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3390 with the same message size as in ssl3_get_certificate_request().
3391 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3392 messages might inadvertently be reject as too long.
3393 [Petr Lampa <lampa@fee.vutbr.cz>]
3395 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3398 *) Modified SSL library such that the verify_callback that has been set
3399 specificly for an SSL object with SSL_set_verify() is actually being
3400 used. Before the change, a verify_callback set with this function was
3401 ignored and the verify_callback() set in the SSL_CTX at the time of
3402 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3403 to allow the necessary settings.
3406 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3407 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3408 done automatically (in contradiction to the requirements of the C
3409 standard). This made problems when used from OpenSSH.
3412 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3413 dh->length and always used
3415 BN_rand_range(priv_key, dh->p).
3417 BN_rand_range() is not necessary for Diffie-Hellman, and this
3418 specific range makes Diffie-Hellman unnecessarily inefficient if
3419 dh->length (recommended exponent length) is much smaller than the
3420 length of dh->p. We could use BN_rand_range() if the order of
3421 the subgroup was stored in the DH structure, but we only have
3426 BN_rand(priv_key, l, ...)
3428 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3434 RSA_eay_public_encrypt
3435 RSA_eay_private_decrypt
3436 RSA_eay_private_encrypt (signing)
3437 RSA_eay_public_decrypt (signature verification)
3439 (default implementations for RSA_public_encrypt,
3440 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3441 always reject numbers >= n.
3444 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3445 to synchronize access to 'locking_thread'. This is necessary on
3446 systems where access to 'locking_thread' (an 'unsigned long'
3447 variable) is not atomic.
3450 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3451 *before* setting the 'crypto_lock_rand' flag. The previous code had
3452 a race condition if 0 is a valid thread ID.
3453 [Travis Vitek <vitek@roguewave.com>]
3455 *) Add support for shared libraries under Irix.
3456 [Albert Chin-A-Young <china@thewrittenword.com>]
3458 *) Add configuration option to build on Linux on both big-endian and
3460 [Ralf Baechle <ralf@uni-koblenz.de>]
3462 *) Add the possibility to create shared libraries on HP-UX.
3465 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3467 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3468 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3469 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3470 PRNG state recovery was possible based on the output of
3471 one PRNG request appropriately sized to gain knowledge on
3472 'md' followed by enough consecutive 1-byte PRNG requests
3473 to traverse all of 'state'.
3475 1. When updating 'md_local' (the current thread's copy of 'md')
3476 during PRNG output generation, hash all of the previous
3477 'md_local' value, not just the half used for PRNG output.
3479 2. Make the number of bytes from 'state' included into the hash
3480 independent from the number of PRNG bytes requested.
3482 The first measure alone would be sufficient to avoid
3483 Markku-Juhani's attack. (Actually it had never occurred
3484 to me that the half of 'md_local' used for chaining was the
3485 half from which PRNG output bytes were taken -- I had always
3486 assumed that the secret half would be used.) The second
3487 measure makes sure that additional data from 'state' is never
3488 mixed into 'md_local' in small portions; this heuristically
3489 further strengthens the PRNG.
3492 *) Fix crypto/bn/asm/mips3.s.
3495 *) When only the key is given to "enc", the IV is undefined. Print out
3496 an error message in this case.
3499 *) Handle special case when X509_NAME is empty in X509 printing routines.
3502 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3503 positive and less than q.
3506 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3507 used: it isn't thread safe and the add_lock_callback should handle
3509 [Paul Rose <Paul.Rose@bridge.com>]
3511 *) Verify that incoming data obeys the block size in
3512 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3516 [Ulf Möller, Bodo Möller]
3518 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3519 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3520 when fixing the server behaviour for backwards-compatible 'client
3521 hello' messages. (Note that the attack is impractical against
3522 SSL 3.0 and TLS 1.0 anyway because length and version checking
3523 means that the probability of guessing a valid ciphertext is
3524 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3527 Before 0.9.5, the countermeasure (hide the error by generating a
3528 random 'decryption result') did not work properly because
3529 ERR_clear_error() was missing, meaning that SSL_get_error() would
3530 detect the supposedly ignored error.
3532 Both problems are now fixed.
3535 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3536 (previously it was 1024).
3539 *) Fix for compatibility mode trust settings: ignore trust settings
3540 unless some valid trust or reject settings are present.
3543 *) Fix for blowfish EVP: its a variable length cipher.
3546 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3547 parameters in DSA public key structures and return an error in the
3548 DSA routines if parameters are absent.
3551 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3552 in the current directory if neither $RANDFILE nor $HOME was set.
3553 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3554 caused some confusion to Windows users who haven't defined $HOME.
3555 Thus RAND_file_name() is changed again: e_os.h can define a
3556 DEFAULT_HOME, which will be used if $HOME is not set.
3557 For Windows, we use "C:"; on other platforms, we still require
3558 environment variables.
3560 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3561 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3562 having multiple threads call RAND_poll() concurrently.
3565 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3566 combination of a flag and a thread ID variable.
3567 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3568 flag), *other* threads can enter ssleay_add_bytes without obeying
3569 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3570 that they do not hold after the first thread unsets add_do_not_lock).
3573 *) Change bctest again: '-x' expressions are not available in all
3577 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3579 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3580 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3582 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3583 the default extension for executables, if any. Also, make the perl
3584 scripts that use symlink() to test if it really exists and use "cp"
3585 if it doesn't. All this made OpenSSL compilable and installable in
3589 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3590 If SEQUENCE is length is indefinite just set c->slen to the total
3591 amount of data available.
3592 [Steve Henson, reported by shige@FreeBSD.org]
3593 [This change does not apply to 0.9.7.]
3595 *) Change bctest to avoid here-documents inside command substitution
3596 (workaround for FreeBSD /bin/sh bug).
3597 For compatibility with Ultrix, avoid shell functions (introduced
3598 in the bctest version that searches along $PATH).
3601 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3602 with des_encrypt() defined on some operating systems, like Solaris
3606 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3607 On the Importance of Eliminating Errors in Cryptographic
3608 Computations, J. Cryptology 14 (2001) 2, 101-119,
3609 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3612 *) MIPS assembler BIGNUM division bug fix.
3615 *) Disabled incorrect Alpha assembler code.
3618 *) Fix PKCS#7 decode routines so they correctly update the length
3619 after reading an EOC for the EXPLICIT tag.
3621 [This change does not apply to 0.9.7.]
3623 *) Fix bug in PKCS#12 key generation routines. This was triggered
3624 if a 3DES key was generated with a 0 initial byte. Include
3625 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3626 (but broken) behaviour.
3629 *) Enhance bctest to search for a working bc along $PATH and print
3631 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3633 *) Fix memory leaks in err.c: free err_data string if necessary;
3634 don't write to the wrong index in ERR_set_error_data.
3637 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3641 *) Replace rdtsc with _emit statements for VC++ version 5.
3642 [Jeremy Cooper <jeremy@baymoo.org>]
3644 *) Make it possible to reuse SSLv2 sessions.
3647 *) In copy_email() check for >= 0 as a return value for
3648 X509_NAME_get_index_by_NID() since 0 is a valid index.
3649 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3651 *) Avoid coredump with unsupported or invalid public keys by checking if
3652 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3653 PKCS7_verify() fails with non detached data.
3656 *) Don't use getenv in library functions when run as setuid/setgid.
3657 New function OPENSSL_issetugid().
3660 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3661 due to incorrect handling of multi-threading:
3663 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3665 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3667 3. Count how many times MemCheck_off() has been called so that
3668 nested use can be treated correctly. This also avoids
3669 inband-signalling in the previous code (which relied on the
3670 assumption that thread ID 0 is impossible).
3673 *) Add "-rand" option also to s_client and s_server.
3676 *) Fix CPU detection on Irix 6.x.
3677 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3678 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3680 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3683 [This change does not apply to 0.9.7.]
3685 *) Use the cached encoding of an X509_NAME structure rather than
3686 copying it. This is apparently the reason for the libsafe "errors"
3687 but the code is actually correct.
3690 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3691 Bleichenbacher's DSA attack.
3692 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3693 to be set and top=0 forces the highest bit to be set; top=-1 is new
3694 and leaves the highest bit random.
3695 [Ulf Moeller, Bodo Moeller]
3697 *) In the NCONF_...-based implementations for CONF_... queries
3698 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3699 a temporary CONF structure with the data component set to NULL
3700 (which gives segmentation faults in lh_retrieve).
3701 Instead, use NULL for the CONF pointer in CONF_get_string and
3702 CONF_get_number (which may use environment variables) and directly
3703 return NULL from CONF_get_section.
3706 *) Fix potential buffer overrun for EBCDIC.
3709 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3710 keyUsage if basicConstraints absent for a CA.
3713 *) Make SMIME_write_PKCS7() write mail header values with a format that
3714 is more generally accepted (no spaces before the semicolon), since
3715 some programs can't parse those values properly otherwise. Also make
3716 sure BIO's that break lines after each write do not create invalid
3720 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3721 macros previously used would not encode an empty SEQUENCE OF
3722 and break the signature.
3724 [This change does not apply to 0.9.7.]
3726 *) Zero the premaster secret after deriving the master secret in
3730 *) Add some EVP_add_digest_alias registrations (as found in
3731 OpenSSL_add_all_digests()) to SSL_library_init()
3732 aka OpenSSL_add_ssl_algorithms(). This provides improved
3733 compatibility with peers using X.509 certificates
3734 with unconventional AlgorithmIdentifier OIDs.
3737 *) Fix for Irix with NO_ASM.
3738 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3740 *) ./config script fixes.
3741 [Ulf Moeller, Richard Levitte]
3743 *) Fix 'openssl passwd -1'.
3746 *) Change PKCS12_key_gen_asc() so it can cope with non null
3747 terminated strings whose length is passed in the passlen
3748 parameter, for example from PEM callbacks. This was done
3749 by adding an extra length parameter to asc2uni().
3750 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3752 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3753 call failed, free the DSA structure.
3756 *) Fix to uni2asc() to cope with zero length Unicode strings.
3757 These are present in some PKCS#12 files.
3760 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3761 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3762 when writing a 32767 byte record.
3763 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3765 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3766 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3768 (RSA objects have a reference count access to which is protected
3769 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3770 so they are meant to be shared between threads.)
3771 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3772 "Reddie, Steven" <Steven.Reddie@ca.com>]
3774 *) Fix a deadlock in CRYPTO_mem_leaks().
3777 *) Use better test patterns in bntest.
3780 *) rand_win.c fix for Borland C.
3783 *) BN_rshift bugfix for n == 0.
3786 *) Add a 'bctest' script that checks for some known 'bc' bugs
3787 so that 'make test' does not abort just because 'bc' is broken.
3790 *) Store verify_result within SSL_SESSION also for client side to
3791 avoid potential security hole. (Re-used sessions on the client side
3792 always resulted in verify_result==X509_V_OK, not using the original
3793 result of the server certificate verification.)
3796 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3797 SSL3_RT_APPLICATION_DATA, return 0.
3798 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3802 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3803 releases, have been re-implemented by renaming the previous
3804 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3805 and ssl3_read_internal, respectively, and adding 'peek' parameters
3806 to them. The new ssl[23]_{read,peek} functions are calls to
3807 ssl[23]_read_internal with the 'peek' flag set appropriately.
3808 A 'peek' parameter has also been added to ssl3_read_bytes, which
3809 does the actual work for ssl3_read_internal.
3812 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3813 the method-specific "init()" handler. Also clean up ex_data after
3814 calling the method-specific "finish()" handler. Previously, this was
3815 happening the other way round.
3818 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3819 The previous value, 12, was not always sufficient for BN_mod_exp().
3822 *) Make sure that shared libraries get the internal name engine with
3823 the full version number and not just 0. This should mark the
3824 shared libraries as not backward compatible. Of course, this should
3825 be changed again when we can guarantee backward binary compatibility.
3828 *) Fix typo in get_cert_by_subject() in by_dir.c
3829 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3831 *) Rework the system to generate shared libraries:
3833 - Make note of the expected extension for the shared libraries and
3834 if there is a need for symbolic links from for example libcrypto.so.0
3835 to libcrypto.so.0.9.7. There is extended info in Configure for
3838 - Make as few rebuilds of the shared libraries as possible.
3840 - Still avoid linking the OpenSSL programs with the shared libraries.
3842 - When installing, install the shared libraries separately from the
3846 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3848 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3849 and not in SSL_clear because the latter is also used by the
3850 accept/connect functions; previously, the settings made by
3851 SSL_set_read_ahead would be lost during the handshake.
3852 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3854 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3855 Previously, it would create entries for disableed algorithms no
3859 *) Added several new manual pages for SSL_* function.
3862 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3864 *) In ssl23_get_client_hello, generate an error message when faced
3865 with an initial SSL 3.0/TLS record that is too small to contain the
3866 first two bytes of the ClientHello message, i.e. client_version.
3867 (Note that this is a pathologic case that probably has never happened
3868 in real life.) The previous approach was to use the version number
3869 from the record header as a substitute; but our protocol choice
3870 should not depend on that one because it is not authenticated
3871 by the Finished messages.
3874 *) More robust randomness gathering functions for Windows.
3875 [Jeffrey Altman <jaltman@columbia.edu>]
3877 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3878 not set then we don't setup the error code for issuer check errors
3879 to avoid possibly overwriting other errors which the callback does
3880 handle. If an application does set the flag then we assume it knows
3881 what it is doing and can handle the new informational codes
3885 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3886 a general "ANY" type, as such it should be able to decode anything
3887 including tagged types. However it didn't check the class so it would
3888 wrongly interpret tagged types in the same way as their universal
3889 counterpart and unknown types were just rejected. Changed so that the
3890 tagged and unknown types are handled in the same way as a SEQUENCE:
3891 that is the encoding is stored intact. There is also a new type
3892 "V_ASN1_OTHER" which is used when the class is not universal, in this
3893 case we have no idea what the actual type is so we just lump them all
3897 *) On VMS, stdout may very well lead to a file that is written to
3898 in a record-oriented fashion. That means that every write() will
3899 write a separate record, which will be read separately by the
3900 programs trying to read from it. This can be very confusing.
3902 The solution is to put a BIO filter in the way that will buffer
3903 text until a linefeed is reached, and then write everything a
3904 line at a time, so every record written will be an actual line,
3905 not chunks of lines and not (usually doesn't happen, but I've
3906 seen it once) several lines in one record. BIO_f_linebuffer() is
3909 Currently, it's a VMS-only method, because that's where it has
3910 been tested well enough.
3913 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3914 it can return incorrect results.
3915 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3916 but it was in 0.9.6-beta[12].)
3919 *) Disable the check for content being present when verifying detached
3920 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3921 include zero length content when signing messages.
3924 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3925 BIO_ctrl (for BIO pairs).
3928 *) Add DSO method for VMS.
3931 *) Bug fix: Montgomery multiplication could produce results with the
3935 *) Add RPM specification openssl.spec and modify it to build three
3936 packages. The default package contains applications, application
3937 documentation and run-time libraries. The devel package contains
3938 include files, static libraries and function documentation. The
3939 doc package contains the contents of the doc directory. The original
3940 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3943 *) Add a large number of documentation files for many SSL routines.
3944 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3946 *) Add a configuration entry for Sony News 4.
3947 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3949 *) Don't set the two most significant bits to one when generating a
3950 random number < q in the DSA library.
3953 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3954 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3955 the underlying transport is blocking) if a handshake took place.
3956 (The default behaviour is needed by applications such as s_client
3957 and s_server that use select() to determine when to use SSL_read;
3958 but for applications that know in advance when to expect data, it
3959 just makes things more complicated.)
3962 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3966 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3967 work better on such systems.
3968 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3970 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3971 Update PKCS12_parse() so it copies the friendlyName and the
3972 keyid to the certificates aux info.
3975 *) Fix bug in PKCS7_verify() which caused an infinite loop
3976 if there was more than one signature.
3977 [Sven Uszpelkat <su@celocom.de>]
3979 *) Major change in util/mkdef.pl to include extra information
3980 about each symbol, as well as presentig variables as well
3981 as functions. This change means that there's n more need
3982 to rebuild the .num files when some algorithms are excluded.
3985 *) Allow the verify time to be set by an application,
3986 rather than always using the current time.
3989 *) Phase 2 verify code reorganisation. The certificate
3990 verify code now looks up an issuer certificate by a
3991 number of criteria: subject name, authority key id
3992 and key usage. It also verifies self signed certificates
3993 by the same criteria. The main comparison function is
3994 X509_check_issued() which performs these checks.
3996 Lot of changes were necessary in order to support this
3997 without completely rewriting the lookup code.
3999 Authority and subject key identifier are now cached.
4001 The LHASH 'certs' is X509_STORE has now been replaced
4002 by a STACK_OF(X509_OBJECT). This is mainly because an
4003 LHASH can't store or retrieve multiple objects with
4004 the same hash value.
4006 As a result various functions (which were all internal
4007 use only) have changed to handle the new X509_STORE
4008 structure. This will break anything that messed round
4009 with X509_STORE internally.
4011 The functions X509_STORE_add_cert() now checks for an
4012 exact match, rather than just subject name.
4014 The X509_STORE API doesn't directly support the retrieval
4015 of multiple certificates matching a given criteria, however
4016 this can be worked round by performing a lookup first
4017 (which will fill the cache with candidate certificates)
4018 and then examining the cache for matches. This is probably
4019 the best we can do without throwing out X509_LOOKUP
4020 entirely (maybe later...).
4022 The X509_VERIFY_CTX structure has been enhanced considerably.
4024 All certificate lookup operations now go via a get_issuer()
4025 callback. Although this currently uses an X509_STORE it
4026 can be replaced by custom lookups. This is a simple way
4027 to bypass the X509_STORE hackery necessary to make this
4028 work and makes it possible to use more efficient techniques
4029 in future. A very simple version which uses a simple
4030 STACK for its trusted certificate store is also provided
4031 using X509_STORE_CTX_trusted_stack().
4033 The verify_cb() and verify() callbacks now have equivalents
4034 in the X509_STORE_CTX structure.
4036 X509_STORE_CTX also has a 'flags' field which can be used
4037 to customise the verify behaviour.
4040 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4041 excludes S/MIME capabilities.
4044 *) When a certificate request is read in keep a copy of the
4045 original encoding of the signed data and use it when outputing
4046 again. Signatures then use the original encoding rather than
4047 a decoded, encoded version which may cause problems if the
4048 request is improperly encoded.
4051 *) For consistency with other BIO_puts implementations, call
4052 buffer_write(b, ...) directly in buffer_puts instead of calling
4055 In BIO_puts, increment b->num_write as in BIO_write.
4056 [Peter.Sylvester@EdelWeb.fr]
4058 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4059 BN_zero, we may not return a BIGNUM with an array consisting of
4063 *) Avoid calling abort() from within the library when problems are
4064 detected, except if preprocessor symbols have been defined
4065 (such as REF_CHECK, BN_DEBUG etc.).
4068 *) New openssl application 'rsautl'. This utility can be
4069 used for low level RSA operations. DER public key
4070 BIO/fp routines also added.
4073 *) New Configure entry and patches for compiling on QNX 4.
4074 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4076 *) A demo state-machine implementation was sponsored by
4077 Nuron (http://www.nuron.com/) and is now available in
4078 demos/state_machine.
4081 *) New options added to the 'dgst' utility for signature
4082 generation and verification.
4085 *) Unrecognized PKCS#7 content types are now handled via a
4086 catch all ASN1_TYPE structure. This allows unsupported
4087 types to be stored as a "blob" and an application can
4088 encode and decode it manually.
4091 *) Fix various signed/unsigned issues to make a_strex.c
4093 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4095 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4096 length if passed a buffer. ASN1_INTEGER_to_BN failed
4097 if passed a NULL BN and its argument was negative.
4098 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4100 *) Modification to PKCS#7 encoding routines to output definite
4101 length encoding. Since currently the whole structures are in
4102 memory there's not real point in using indefinite length
4103 constructed encoding. However if OpenSSL is compiled with
4104 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4107 *) Added BIO_vprintf() and BIO_vsnprintf().
4110 *) Added more prefixes to parse for in the the strings written
4111 through a logging bio, to cover all the levels that are available
4112 through syslog. The prefixes are now:
4114 PANIC, EMERG, EMR => LOG_EMERG
4115 ALERT, ALR => LOG_ALERT
4116 CRIT, CRI => LOG_CRIT
4117 ERROR, ERR => LOG_ERR
4118 WARNING, WARN, WAR => LOG_WARNING
4119 NOTICE, NOTE, NOT => LOG_NOTICE
4120 INFO, INF => LOG_INFO
4121 DEBUG, DBG => LOG_DEBUG
4123 and as before, if none of those prefixes are present at the
4124 beginning of the string, LOG_ERR is chosen.
4126 On Win32, the LOG_* levels are mapped according to this:
4128 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4129 LOG_WARNING => EVENTLOG_WARNING_TYPE
4130 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4134 *) Made it possible to reconfigure with just the configuration
4135 argument "reconf" or "reconfigure". The command line arguments
4136 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4137 and are retrieved from there when reconfiguring.
4141 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4143 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4146 *) The obj_dat.pl script was messing up the sorting of object
4147 names. The reason was that it compared the quoted version
4148 of strings as a result "OCSP" > "OCSP Signing" because
4149 " > SPACE. Changed script to store unquoted versions of
4150 names and add quotes on output. It was also omitting some
4151 names from the lookup table if they were given a default
4152 value (that is if SN is missing it is given the same
4153 value as LN and vice versa), these are now added on the
4154 grounds that if an object has a name we should be able to
4155 look it up. Finally added warning output when duplicate
4156 short or long names are found.
4159 *) Changes needed for Tandem NSK.
4160 [Scott Uroff <scott@xypro.com>]
4162 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4163 RSA_padding_check_SSLv23(), special padding was never detected
4164 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4165 version rollback attacks was not effective.
4167 In s23_clnt.c, don't use special rollback-attack detection padding
4168 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4169 client; similarly, in s23_srvr.c, don't do the rollback check if
4170 SSL 2.0 is the only protocol enabled in the server.
4173 *) Make it possible to get hexdumps of unprintable data with 'openssl
4174 asn1parse'. By implication, the functions ASN1_parse_dump() and
4175 BIO_dump_indent() are added.
4178 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4179 these print out strings and name structures based on various
4180 flags including RFC2253 support and proper handling of
4181 multibyte characters. Added options to the 'x509' utility
4182 to allow the various flags to be set.
4185 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4186 Also change the functions X509_cmp_current_time() and
4187 X509_gmtime_adj() work with an ASN1_TIME structure,
4188 this will enable certificates using GeneralizedTime in validity
4189 dates to be checked.
4192 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4193 negative public key encodings) on by default,
4194 NO_NEG_PUBKEY_BUG can be set to disable it.
4197 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4198 content octets. An i2c_ASN1_OBJECT is unnecessary because
4199 the encoding can be trivially obtained from the structure.
4202 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4203 not read locks (CRYPTO_r_[un]lock).
4206 *) A first attempt at creating official support for shared
4207 libraries through configuration. I've kept it so the
4208 default is static libraries only, and the OpenSSL programs
4209 are always statically linked for now, but there are
4210 preparations for dynamic linking in place.
4211 This has been tested on Linux and Tru64.
4214 *) Randomness polling function for Win9x, as described in:
4215 Peter Gutmann, Software Generation of Practically Strong
4219 *) Fix so PRNG is seeded in req if using an already existing
4223 *) New options to smime application. -inform and -outform
4224 allow alternative formats for the S/MIME message including
4225 PEM and DER. The -content option allows the content to be
4226 specified separately. This should allow things like Netscape
4227 form signing output easier to verify.
4230 *) Fix the ASN1 encoding of tags using the 'long form'.
4233 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4234 STRING types. These convert content octets to and from the
4235 underlying type. The actual tag and length octets are
4236 already assumed to have been read in and checked. These
4237 are needed because all other string types have virtually
4238 identical handling apart from the tag. By having versions
4239 of the ASN1 functions that just operate on content octets
4240 IMPLICIT tagging can be handled properly. It also allows
4241 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4242 and ASN1_INTEGER are identical apart from the tag.
4245 *) Change the handling of OID objects as follows:
4247 - New object identifiers are inserted in objects.txt, following
4248 the syntax given in objects.README.
4249 - objects.pl is used to process obj_mac.num and create a new
4251 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4254 This is currently kind of a hack, and the perl code in objects.pl
4255 isn't very elegant, but it works as I intended. The simplest way
4256 to check that it worked correctly is to look in obj_dat.h and
4257 check the array nid_objs and make sure the objects haven't moved
4258 around (this is important!). Additions are OK, as well as
4259 consistent name changes.
4262 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4265 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4266 The given file adds to whatever has already been seeded into the
4267 random pool through the RANDFILE configuration file option or
4268 environment variable, or the default random state file.
4271 *) mkstack.pl now sorts each macro group into lexical order.
4272 Previously the output order depended on the order the files
4273 appeared in the directory, resulting in needless rewriting
4277 *) Patches to make OpenSSL compile under Win32 again. Mostly
4278 work arounds for the VC++ problem that it treats func() as
4279 func(void). Also stripped out the parts of mkdef.pl that
4280 added extra typesafe functions: these no longer exist.
4283 *) Reorganisation of the stack code. The macros are now all
4284 collected in safestack.h . Each macro is defined in terms of
4285 a "stack macro" of the form SKM_<name>(type, a, b). The
4286 DEBUG_SAFESTACK is now handled in terms of function casts,
4287 this has the advantage of retaining type safety without the
4288 use of additional functions. If DEBUG_SAFESTACK is not defined
4289 then the non typesafe macros are used instead. Also modified the
4290 mkstack.pl script to handle the new form. Needs testing to see
4291 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4292 the default if no major problems. Similar behaviour for ASN1_SET_OF
4293 and PKCS12_STACK_OF.
4296 *) When some versions of IIS use the 'NET' form of private key the
4297 key derivation algorithm is different. Normally MD5(password) is
4298 used as a 128 bit RC4 key. In the modified case
4299 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4300 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4301 as the old Netscape_RSA functions except they have an additional
4302 'sgckey' parameter which uses the modified algorithm. Also added
4303 an -sgckey command line option to the rsa utility. Thanks to
4304 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4305 algorithm to openssl-dev.
4308 *) The evp_local.h macros were using 'c.##kname' which resulted in
4309 invalid expansion on some systems (SCO 5.0.5 for example).
4310 Corrected to 'c.kname'.
4311 [Phillip Porch <root@theporch.com>]
4313 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4314 a STACK of email addresses from a certificate or request, these look
4315 in the subject name and the subject alternative name extensions and
4316 omit any duplicate addresses.
4319 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4320 This makes DSA verification about 2 % faster.
4323 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4324 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4325 plus overhead for 1024 bit moduli).
4326 This makes exponentiations about 0.5 % faster for 1024 bit
4327 exponents (as measured by "openssl speed rsa2048").
4330 *) Rename memory handling macros to avoid conflicts with other
4332 Malloc => OPENSSL_malloc
4333 Malloc_locked => OPENSSL_malloc_locked
4334 Realloc => OPENSSL_realloc
4335 Free => OPENSSL_free
4338 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4339 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4342 *) CygWin32 support.
4343 [John Jarvie <jjarvie@newsguy.com>]
4345 *) The type-safe stack code has been rejigged. It is now only compiled
4346 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4347 by default all type-specific stack functions are "#define"d back to
4348 standard stack functions. This results in more streamlined output
4349 but retains the type-safety checking possibilities of the original
4353 *) The STACK code has been cleaned up, and certain type declarations
4354 that didn't make a lot of sense have been brought in line. This has
4355 also involved a cleanup of sorts in safestack.h to more correctly
4356 map type-safe stack functions onto their plain stack counterparts.
4357 This work has also resulted in a variety of "const"ifications of
4358 lots of the code, especially "_cmp" operations which should normally
4359 be prototyped with "const" parameters anyway.
4362 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4363 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4364 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4365 where all of 'md' is used each time the PRNG is used, but 'state'
4366 is used only indexed by a cyclic counter. As entropy may not be
4367 well distributed from the beginning, 'md' is important as a
4368 chaining variable. However, the output function chains only half
4369 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4370 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4371 in all of 'state' being rewritten, with the new values depending
4372 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4375 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4376 the handshake is continued after ssl_verify_cert_chain();
4377 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4378 can lead to 'unexplainable' connection aborts later.
4379 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4381 *) Major EVP API cipher revision.
4382 Add hooks for extra EVP features. This allows various cipher
4383 parameters to be set in the EVP interface. Support added for variable
4384 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4385 setting of RC2 and RC5 parameters.
4387 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4390 Remove lots of duplicated code from the EVP library. For example *every*
4391 cipher init() function handles the 'iv' in the same way according to the
4392 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4393 for CFB and OFB modes they zero ctx->num.
4395 New functionality allows removal of S/MIME code RC2 hack.
4397 Most of the routines have the same form and so can be declared in terms
4400 By shifting this to the top level EVP_CipherInit() it can be removed from
4401 all individual ciphers. If the cipher wants to handle IVs or keys
4402 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4405 Change lots of functions like EVP_EncryptUpdate() to now return a
4406 value: although software versions of the algorithms cannot fail
4407 any installed hardware versions can.
4410 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4411 this option is set, tolerate broken clients that send the negotiated
4412 protocol version number instead of the requested protocol version
4416 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4417 i.e. non-zero for export ciphersuites, zero otherwise.
4418 Previous versions had this flag inverted, inconsistent with
4419 rsa_tmp_cb (..._TMP_RSA_CB).
4420 [Bodo Moeller; problem reported by Amit Chopra]
4422 *) Add missing DSA library text string. Work around for some IIS
4423 key files with invalid SEQUENCE encoding.
4426 *) Add a document (doc/standards.txt) that list all kinds of standards
4427 and so on that are implemented in OpenSSL.
4430 *) Enhance c_rehash script. Old version would mishandle certificates
4431 with the same subject name hash and wouldn't handle CRLs at all.
4432 Added -fingerprint option to crl utility, to support new c_rehash
4436 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4439 *) Fix for SSL server purpose checking. Server checking was
4440 rejecting certificates which had extended key usage present
4441 but no ssl client purpose.
4442 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4444 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4445 is a little unclear about how a blank password is handled.
4446 Since the password in encoded as a BMPString with terminating
4447 double NULL a zero length password would end up as just the
4448 double NULL. However no password at all is different and is
4449 handled differently in the PKCS#12 key generation code. NS
4450 treats a blank password as zero length. MSIE treats it as no
4451 password on export: but it will try both on import. We now do
4452 the same: PKCS12_parse() tries zero length and no password if
4453 the password is set to "" or NULL (NULL is now a valid password:
4454 it wasn't before) as does the pkcs12 application.
4457 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4458 perror when PEM_read_bio_X509_REQ fails, the error message must
4459 be obtained from the error queue.
4462 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4463 it in ERR_remove_state if appropriate, and change ERR_get_state
4464 accordingly to avoid race conditions (this is necessary because
4465 thread_hash is no longer constant once set).
4468 *) Bugfix for linux-elf makefile.one.
4471 *) RSA_get_default_method() will now cause a default
4472 RSA_METHOD to be chosen if one doesn't exist already.
4473 Previously this was only set during a call to RSA_new()
4474 or RSA_new_method(NULL) meaning it was possible for
4475 RSA_get_default_method() to return NULL.
4478 *) Added native name translation to the existing DSO code
4479 that will convert (if the flag to do so is set) filenames
4480 that are sufficiently small and have no path information
4481 into a canonical native form. Eg. "blah" converted to
4482 "libblah.so" or "blah.dll" etc.
4485 *) New function ERR_error_string_n(e, buf, len) which is like
4486 ERR_error_string(e, buf), but writes at most 'len' bytes
4487 including the 0 terminator. For ERR_error_string_n, 'buf'
4489 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4491 *) CONF library reworked to become more general. A new CONF
4492 configuration file reader "class" is implemented as well as a
4493 new functions (NCONF_*, for "New CONF") to handle it. The now
4494 old CONF_* functions are still there, but are reimplemented to
4495 work in terms of the new functions. Also, a set of functions
4496 to handle the internal storage of the configuration data is
4497 provided to make it easier to write new configuration file
4498 reader "classes" (I can definitely see something reading a
4499 configuration file in XML format, for example), called _CONF_*,
4500 or "the configuration storage API"...
4502 The new configuration file reading functions are:
4504 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4505 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4507 NCONF_default, NCONF_WIN32
4509 NCONF_dump_fp, NCONF_dump_bio
4511 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4512 NCONF_new creates a new CONF object. This works in the same way
4513 as other interfaces in OpenSSL, like the BIO interface.
4514 NCONF_dump_* dump the internal storage of the configuration file,
4515 which is useful for debugging. All other functions take the same
4516 arguments as the old CONF_* functions wth the exception of the
4517 first that must be a `CONF *' instead of a `LHASH *'.
4519 To make it easer to use the new classes with the old CONF_* functions,
4520 the function CONF_set_default_method is provided.
4523 *) Add '-tls1' option to 'openssl ciphers', which was already
4524 mentioned in the documentation but had not been implemented.
4525 (This option is not yet really useful because even the additional
4526 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4529 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4530 OpenSSL-based applications) load shared libraries and bind to
4531 them in a portable way.
4532 [Geoff Thorpe, with contributions from Richard Levitte]
4534 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4536 *) Make sure _lrotl and _lrotr are only used with MSVC.
4538 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4539 (the default implementation of RAND_status).
4541 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4542 to '-clrext' (= clear extensions), as intended and documented.
4543 [Bodo Moeller; inconsistency pointed out by Michael Attili
4546 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4547 was larger than the MD block size.
4548 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4550 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4551 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4552 using the passed key: if the passed key was a private key the result
4553 of X509_print(), for example, would be to print out all the private key
4557 *) des_quad_cksum() byte order bug fix.
4558 [Ulf Möller, using the problem description in krb4-0.9.7, where
4559 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4561 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4563 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4565 *) For easily testing in shell scripts whether some command
4566 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4567 returns with exit code 0 iff no command of the given name is available.
4568 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4569 the output goes to stdout and nothing is printed to stderr.
4570 Additional arguments are always ignored.
4572 Since for each cipher there is a command of the same name,
4573 the 'no-cipher' compilation switches can be tested this way.
4575 ('openssl no-XXX' is not able to detect pseudo-commands such
4576 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4579 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4582 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4583 is set; it will be thrown away anyway because each handshake creates
4585 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4586 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4587 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4588 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4591 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4592 'Q' and 'R' lose their special meanings (quit/renegotiate).
4593 This is part of what -quiet does; unlike -quiet, -ign_eof
4594 does not suppress any output.
4597 *) Add compatibility options to the purpose and trust code. The
4598 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4599 accepts a certificate or CA, this was the previous behaviour,
4600 with all the associated security issues.
4602 X509_TRUST_COMPAT is the old trust behaviour: only and
4603 automatically trust self signed roots in certificate store. A
4604 new trust setting X509_TRUST_DEFAULT is used to specify that
4605 a purpose has no associated trust setting and it should instead
4606 use the value in the default purpose.
4609 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4610 and fix a memory leak.
4613 *) In util/mkerr.pl (which implements 'make errors'), preserve
4614 reason strings from the previous version of the .c file, as
4615 the default to have only downcase letters (and digits) in
4616 automatically generated reasons codes is not always appropriate.
4619 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4620 using strerror. Previously, ERR_reason_error_string() returned
4621 library names as reason strings for SYSerr; but SYSerr is a special
4622 case where small numbers are errno values, not library numbers.
4625 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4626 converts DSA parameters into DH parameters. (When creating parameters,
4627 DSA_generate_parameters is used.)
4630 *) Include 'length' (recommended exponent length) in C code generated
4631 by 'openssl dhparam -C'.
4634 *) The second argument to set_label in perlasm was already being used
4635 so couldn't be used as a "file scope" flag. Moved to third argument
4639 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4640 instead of RAND_bytes for encryption IVs and salts.
4643 *) Include RAND_status() into RAND_METHOD instead of implementing
4644 it only for md_rand.c Otherwise replacing the PRNG by calling
4645 RAND_set_rand_method would be impossible.
4648 *) Don't let DSA_generate_key() enter an infinite loop if the random
4649 number generation fails.
4652 *) New 'rand' application for creating pseudo-random output.
4655 *) Added configuration support for Linux/IA64
4656 [Rolf Haberrecker <rolf@suse.de>]
4658 *) Assembler module support for Mingw32.
4661 *) Shared library support for HPUX (in shlib/).
4662 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4664 *) Shared library support for Solaris gcc.
4665 [Lutz Behnke <behnke@trustcenter.de>]
4667 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4669 *) PKCS7_encrypt() was adding text MIME headers twice because they
4670 were added manually and by SMIME_crlf_copy().
4673 *) In bntest.c don't call BN_rand with zero bits argument.
4674 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4676 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4677 case was implemented. This caused BN_div_recp() to fail occasionally.
4680 *) Add an optional second argument to the set_label() in the perl
4681 assembly language builder. If this argument exists and is set
4682 to 1 it signals that the assembler should use a symbol whose
4683 scope is the entire file, not just the current function. This
4684 is needed with MASM which uses the format label:: for this scope.
4685 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4687 *) Change the ASN1 types so they are typedefs by default. Before
4688 almost all types were #define'd to ASN1_STRING which was causing
4689 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4693 *) Change names of new functions to the new get1/get0 naming
4694 convention: After 'get1', the caller owns a reference count
4695 and has to call ..._free; 'get0' returns a pointer to some
4696 data structure without incrementing reference counters.
4697 (Some of the existing 'get' functions increment a reference
4698 counter, some don't.)
4699 Similarly, 'set1' and 'add1' functions increase reference
4700 counters or duplicate objects.
4703 *) Allow for the possibility of temp RSA key generation failure:
4704 the code used to assume it always worked and crashed on failure.
4707 *) Fix potential buffer overrun problem in BIO_printf().
4708 [Ulf Möller, using public domain code by Patrick Powell; problem
4709 pointed out by David Sacerdote <das33@cornell.edu>]
4711 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4712 RAND_egd() and RAND_status(). In the command line application,
4713 the EGD socket can be specified like a seed file using RANDFILE
4717 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4718 Some CAs (e.g. Verisign) distribute certificates in this form.
4721 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4722 list to exclude them. This means that no special compilation option
4723 is needed to use anonymous DH: it just needs to be included in the
4727 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4728 EVP_MD_type. The old functionality is available in a new macro called
4729 EVP_MD_md(). Change code that uses it and update docs.
4732 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4733 where the 'void *' argument is replaced by a function pointer argument.
4734 Previously 'void *' was abused to point to functions, which works on
4735 many platforms, but is not correct. As these functions are usually
4736 called by macros defined in OpenSSL header files, most source code
4737 should work without changes.
4740 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4741 sections with information on -D... compiler switches used for
4742 compiling the library so that applications can see them. To enable
4743 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4744 must be defined. E.g.,
4745 #define OPENSSL_ALGORITHM_DEFINES
4746 #include <openssl/opensslconf.h>
4747 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4748 [Richard Levitte, Ulf and Bodo Möller]
4750 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4754 *) Change the 'other' type in certificate aux info to a STACK_OF
4755 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4756 the required ASN1 format: arbitrary types determined by an OID.
4759 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4760 argument to 'req'. This is not because the function is newer or
4761 better than others it just uses the work 'NEW' in the certificate
4762 request header lines. Some software needs this.
4765 *) Reorganise password command line arguments: now passwords can be
4766 obtained from various sources. Delete the PEM_cb function and make
4767 it the default behaviour: i.e. if the callback is NULL and the
4768 usrdata argument is not NULL interpret it as a null terminated pass
4769 phrase. If usrdata and the callback are NULL then the pass phrase
4770 is prompted for as usual.
4773 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4774 the support is automatically enabled. The resulting binaries will
4775 autodetect the card and use it if present.
4776 [Ben Laurie and Compaq Inc.]
4778 *) Work around for Netscape hang bug. This sends certificate request
4779 and server done in one record. Since this is perfectly legal in the
4780 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4781 the bugs/SSLv3 entry for more info.
4784 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4787 *) Add -rand argument to smime and pkcs12 applications and read/write
4791 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4794 *) Add command line password options to the remaining applications.
4797 *) Bug fix for BN_div_recp() for numerators with an even number of
4801 *) More tests in bntest.c, and changed test_bn output.
4804 *) ./config recognizes MacOS X now.
4807 *) Bug fix for BN_div() when the first words of num and divsor are
4808 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4811 *) Add support for various broken PKCS#8 formats, and command line
4812 options to produce them.
4815 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4816 get temporary BIGNUMs from a BN_CTX.
4819 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4823 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4824 include a #define from the old name to the new. The original intent
4825 was that statically linked binaries could for example just call
4826 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4827 link with digests. This never worked becayse SSLeay_add_all_digests()
4828 and SSLeay_add_all_ciphers() were in the same source file so calling
4829 one would link with the other. They are now in separate source files.
4832 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4835 *) Use a less unusual form of the Miller-Rabin primality test (it used
4836 a binary algorithm for exponentiation integrated into the Miller-Rabin
4837 loop, our standard modexp algorithms are faster).
4840 *) Support for the EBCDIC character set completed.
4841 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4843 *) Source code cleanups: use const where appropriate, eliminate casts,
4844 use void * instead of char * in lhash.
4847 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4848 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4849 this the server could overwrite ephemeral keys that the client
4853 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4854 using 50 iterations of the Rabin-Miller test.
4856 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4857 iterations of the Rabin-Miller test as required by the appendix
4858 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4859 As BN_is_prime_fasttest includes trial division, DSA parameter
4860 generation becomes much faster.
4862 This implies a change for the callback functions in DSA_is_prime
4863 and DSA_generate_parameters: The callback function is called once
4864 for each positive witness in the Rabin-Miller test, not just
4865 occasionally in the inner loop; and the parameters to the
4866 callback function now provide an iteration count for the outer
4867 loop rather than for the current invocation of the inner loop.
4868 DSA_generate_parameters additionally can call the callback
4869 function with an 'iteration count' of -1, meaning that a
4870 candidate has passed the trial division test (when q is generated
4871 from an application-provided seed, trial division is skipped).
4874 *) New function BN_is_prime_fasttest that optionally does trial
4875 division before starting the Rabin-Miller test and has
4876 an additional BN_CTX * argument (whereas BN_is_prime always
4877 has to allocate at least one BN_CTX).
4878 'callback(1, -1, cb_arg)' is called when a number has passed the
4879 trial division stage.
4882 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4886 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4889 *) New function BN_pseudo_rand().
4892 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4893 bignum version of BN_from_montgomery() with the working code from
4894 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4898 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4899 made it impossible to use the same SSL_SESSION data structure in
4900 SSL2 clients in multiple threads.
4903 *) The return value of RAND_load_file() no longer counts bytes obtained
4904 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4905 to seed the PRNG (previously an explicit byte count was required).
4906 [Ulf Möller, Bodo Möller]
4908 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4909 used (char *) instead of (void *) and had casts all over the place.
4912 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4915 *) Retain source code compatibility for BN_prime_checks macro:
4916 BN_is_prime(..., BN_prime_checks, ...) now uses
4917 BN_prime_checks_for_size to determine the appropriate number of
4918 Rabin-Miller iterations.
4921 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4922 DH_CHECK_P_NOT_SAFE_PRIME.
4923 (Check if this is true? OpenPGP calls them "strong".)
4926 *) Merge the functionality of "dh" and "gendh" programs into a new program
4927 "dhparam". The old programs are retained for now but will handle DH keys
4928 (instead of parameters) in future.
4931 *) Make the ciphers, s_server and s_client programs check the return values
4932 when a new cipher list is set.
4935 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4936 ciphers. Before when the 56bit ciphers were enabled the sorting was
4939 The syntax for the cipher sorting has been extended to support sorting by
4940 cipher-strength (using the strength_bits hard coded in the tables).
4941 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4943 Fix a bug in the cipher-command parser: when supplying a cipher command
4944 string with an "undefined" symbol (neither command nor alphanumeric
4945 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4946 an error is flagged.
4948 Due to the strength-sorting extension, the code of the
4949 ssl_create_cipher_list() function was completely rearranged. I hope that
4950 the readability was also increased :-)
4951 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4953 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4954 for the first serial number and places 2 in the serial number file. This
4955 avoids problems when the root CA is created with serial number zero and
4956 the first user certificate has the same issuer name and serial number
4960 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4961 the new code. Add documentation for this stuff.
4964 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4965 X509_*() to X509at_*() on the grounds that they don't handle X509
4966 structures and behave in an analagous way to the X509v3 functions:
4967 they shouldn't be called directly but wrapper functions should be used
4970 So we also now have some wrapper functions that call the X509at functions
4971 when passed certificate requests. (TO DO: similar things can be done with
4972 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4973 things. Some of these need some d2i or i2d and print functionality
4974 because they handle more complex structures.)
4977 *) Add missing #ifndefs that caused missing symbols when building libssl
4978 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4979 NO_RSA in ssl/s2*.c.
4980 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4982 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4983 has a return value which indicates the quality of the random data
4984 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4985 error queue. New function RAND_pseudo_bytes() generates output that is
4986 guaranteed to be unique but not unpredictable. RAND_add is like
4987 RAND_seed, but takes an extra argument for an entropy estimate
4988 (RAND_seed always assumes full entropy).
4991 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4992 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4993 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4994 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4995 false-positive rate of at most 2^-80 for random input.
4998 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5001 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5002 in the 0.9.5 release), this returns the chain
5003 from an X509_CTX structure with a dup of the stack and all
5004 the X509 reference counts upped: so the stack will exist
5005 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5008 Also make SSL_SESSION_print() print out the verify return
5012 *) Add manpage for the pkcs12 command. Also change the default
5013 behaviour so MAC iteration counts are used unless the new
5014 -nomaciter option is used. This improves file security and
5015 only older versions of MSIE (4.0 for example) need it.
5018 *) Honor the no-xxx Configure options when creating .DEF files.
5021 *) Add PKCS#10 attributes to field table: challengePassword,
5022 unstructuredName and unstructuredAddress. These are taken from
5023 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5024 international characters are used.
5026 More changes to X509_ATTRIBUTE code: allow the setting of types
5027 based on strings. Remove the 'loc' parameter when adding
5028 attributes because these will be a SET OF encoding which is sorted
5032 *) Initial changes to the 'req' utility to allow request generation
5033 automation. This will allow an application to just generate a template
5034 file containing all the field values and have req construct the
5037 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5038 used all over the place including certificate requests and PKCS#7
5039 structures. They are currently handled manually where necessary with
5040 some primitive wrappers for PKCS#7. The new functions behave in a
5041 manner analogous to the X509 extension functions: they allow
5042 attributes to be looked up by NID and added.
5044 Later something similar to the X509V3 code would be desirable to
5045 automatically handle the encoding, decoding and printing of the
5046 more complex types. The string types like challengePassword can
5047 be handled by the string table functions.
5049 Also modified the multi byte string table handling. Now there is
5050 a 'global mask' which masks out certain types. The table itself
5051 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5052 is useful when for example there is only one permissible type
5053 (as in countryName) and using the mask might result in no valid
5057 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5058 SSL_get_peer_finished to allow applications to obtain the latest
5059 Finished messages sent to the peer or expected from the peer,
5060 respectively. (SSL_get_peer_finished is usually the Finished message
5061 actually received from the peer, otherwise the protocol will be aborted.)
5063 As the Finished message are message digests of the complete handshake
5064 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5065 be used for external authentication procedures when the authentication
5066 provided by SSL/TLS is not desired or is not enough.
5069 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5070 the host supports BWX extension and if Compaq C is present on the
5071 $PATH. Just exploiting of the BWX extension results in 20-30%
5072 performance kick for some algorithms, e.g. DES and RC4 to mention
5073 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5077 *) Add support for MS "fast SGC". This is arguably a violation of the
5078 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5079 weak crypto and after checking the certificate is SGC a second one
5080 with strong crypto. MS SGC stops the first handshake after receiving
5081 the server certificate message and sends a second client hello. Since
5082 a server will typically do all the time consuming operations before
5083 expecting any further messages from the client (server key exchange
5084 is the most expensive) there is little difference between the two.
5086 To get OpenSSL to support MS SGC we have to permit a second client
5087 hello message after we have sent server done. In addition we have to
5088 reset the MAC if we do get this second client hello.
5091 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5092 if a DER encoded private key is RSA or DSA traditional format. Changed
5093 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5094 format DER encoded private key. Newer code should use PKCS#8 format which
5095 has the key type encoded in the ASN1 structure. Added DER private key
5096 support to pkcs8 application.
5099 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5100 ciphersuites has been selected (as required by the SSL 3/TLS 1
5101 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5102 is set, we interpret this as a request to violate the specification
5103 (the worst that can happen is a handshake failure, and 'correct'
5104 behaviour would result in a handshake failure anyway).
5107 *) In SSL_CTX_add_session, take into account that there might be multiple
5108 SSL_SESSION structures with the same session ID (e.g. when two threads
5109 concurrently obtain them from an external cache).
5110 The internal cache can handle only one SSL_SESSION with a given ID,
5111 so if there's a conflict, we now throw out the old one to achieve
5115 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5116 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5117 some routines that use cipher OIDs: some ciphers do not have OIDs
5118 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5122 *) Simplify the trust setting structure and code. Now we just have
5123 two sequences of OIDs for trusted and rejected settings. These will
5124 typically have values the same as the extended key usage extension
5125 and any application specific purposes.
5127 The trust checking code now has a default behaviour: it will just
5128 check for an object with the same NID as the passed id. Functions can
5129 be provided to override either the default behaviour or the behaviour
5130 for a given id. SSL client, server and email already have functions
5131 in place for compatibility: they check the NID and also return "trusted"
5132 if the certificate is self signed.
5135 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5136 traditional format into an EVP_PKEY structure.
5139 *) Add a password callback function PEM_cb() which either prompts for
5140 a password if usr_data is NULL or otherwise assumes it is a null
5141 terminated password. Allow passwords to be passed on command line
5142 environment or config files in a few more utilities.
5145 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5146 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5147 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5148 Update documentation.
5151 *) Support for ASN1 "NULL" type. This could be handled before by using
5152 ASN1_TYPE but there wasn't any function that would try to read a NULL
5153 and produce an error if it couldn't. For compatibility we also have
5154 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5155 don't allocate anything because they don't need to.
5158 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5160 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5162 *) Rebuild of the memory allocation routines used by OpenSSL code and
5163 possibly others as well. The purpose is to make an interface that
5164 provide hooks so anyone can build a separate set of allocation and
5165 deallocation routines to be used by OpenSSL, for example memory
5166 pool implementations, or something else, which was previously hard
5167 since Malloc(), Realloc() and Free() were defined as macros having
5168 the values malloc, realloc and free, respectively (except for Win32
5169 compilations). The same is provided for memory debugging code.
5170 OpenSSL already comes with functionality to find memory leaks, but
5171 this gives people a chance to debug other memory problems.
5173 With these changes, a new set of functions and macros have appeared:
5175 CRYPTO_set_mem_debug_functions() [F]
5176 CRYPTO_get_mem_debug_functions() [F]
5177 CRYPTO_dbg_set_options() [F]
5178 CRYPTO_dbg_get_options() [F]
5179 CRYPTO_malloc_debug_init() [M]
5181 The memory debug functions are NULL by default, unless the library
5182 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5183 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5184 gives the standard debugging functions that come with OpenSSL) or
5185 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5186 provided by the library user) must be used. When the standard
5187 debugging functions are used, CRYPTO_dbg_set_options can be used to
5188 request additional information:
5189 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5190 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5192 Also, things like CRYPTO_set_mem_functions will always give the
5193 expected result (the new set of functions is used for allocation
5194 and deallocation) at all times, regardless of platform and compiler
5197 To finish it up, some functions that were never use in any other
5198 way than through macros have a new API and new semantic:
5201 CRYPTO_dbg_realloc()
5204 All macros of value have retained their old syntax.
5205 [Richard Levitte and Bodo Moeller]
5207 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5208 ordering of SMIMECapabilities wasn't in "strength order" and there
5209 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5213 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5214 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5215 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5217 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5218 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5219 functionality to handle multipart/signed properly) and a utility
5220 called 'smime' to call all this stuff. This is based on code I
5221 originally wrote for Celo who have kindly allowed it to be
5222 included in OpenSSL.
5225 *) Add variants des_set_key_checked and des_set_key_unchecked of
5226 des_set_key (aka des_key_sched). Global variable des_check_key
5227 decides which of these is called by des_set_key; this way
5228 des_check_key behaves as it always did, but applications and
5229 the library itself, which was buggy for des_check_key == 1,
5230 have a cleaner way to pick the version they need.
5233 *) New function PKCS12_newpass() which changes the password of a
5237 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5238 dynamic mix. In both cases the ids can be used as an index into the
5239 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5240 functions so they accept a list of the field values and the
5241 application doesn't need to directly manipulate the X509_TRUST
5245 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5249 *) Modify the way the V3 extension code looks up extensions. This now
5250 works in a similar way to the object code: we have some "standard"
5251 extensions in a static table which is searched with OBJ_bsearch()
5252 and the application can add dynamic ones if needed. The file
5253 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5254 updated whenever a new extension is added to the core code and kept
5255 in ext_nid order. There is a simple program 'tabtest.c' which checks
5256 this. New extensions are not added too often so this file can readily
5257 be maintained manually.
5259 There are two big advantages in doing things this way. The extensions
5260 can be looked up immediately and no longer need to be "added" using
5261 X509V3_add_standard_extensions(): this function now does nothing.
5262 [Side note: I get *lots* of email saying the extension code doesn't
5263 work because people forget to call this function]
5264 Also no dynamic allocation is done unless new extensions are added:
5265 so if we don't add custom extensions there is no need to call
5266 X509V3_EXT_cleanup().
5269 *) Modify enc utility's salting as follows: make salting the default. Add a
5270 magic header, so unsalted files fail gracefully instead of just decrypting
5271 to garbage. This is because not salting is a big security hole, so people
5272 should be discouraged from doing it.
5275 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5276 digest to be passed on the command line but it only used this
5277 parameter when signing a certificate. Modified so all relevant
5278 operations are affected by the digest parameter including the
5279 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5280 DSA key was used because it didn't fix the digest.
5283 *) Initial certificate chain verify code. Currently tests the untrusted
5284 certificates for consistency with the verify purpose (which is set
5285 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5287 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5288 this is because it will reject chains with invalid extensions whereas
5289 every previous version of OpenSSL and SSLeay made no checks at all.
5291 Trust code: checks the root CA for the relevant trust settings. Trust
5292 settings have an initial value consistent with the verify purpose: e.g.
5293 if the verify purpose is for SSL client use it expects the CA to be
5294 trusted for SSL client use. However the default value can be changed to
5295 permit custom trust settings: one example of this would be to only trust
5296 certificates from a specific "secure" set of CAs.
5298 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5299 which should be used for version portability: especially since the
5300 verify structure is likely to change more often now.
5302 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5303 to set them. If not set then assume SSL clients will verify SSL servers
5306 Two new options to the verify program: -untrusted allows a set of
5307 untrusted certificates to be passed in and -purpose which sets the
5308 intended purpose of the certificate. If a purpose is set then the
5309 new chain verify code is used to check extension consistency.
5312 *) Support for the authority information access extension.
5315 *) Modify RSA and DSA PEM read routines to transparently handle
5316 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5317 public keys in a format compatible with certificate
5318 SubjectPublicKeyInfo structures. Unfortunately there were already
5319 functions called *_PublicKey_* which used various odd formats so
5320 these are retained for compatibility: however the DSA variants were
5321 never in a public release so they have been deleted. Changed dsa/rsa
5322 utilities to handle the new format: note no releases ever handled public
5323 keys so we should be OK.
5325 The primary motivation for this change is to avoid the same fiasco
5326 that dogs private keys: there are several incompatible private key
5327 formats some of which are standard and some OpenSSL specific and
5328 require various evil hacks to allow partial transparent handling and
5329 even then it doesn't work with DER formats. Given the option anything
5330 other than PKCS#8 should be dumped: but the other formats have to
5331 stay in the name of compatibility.
5333 With public keys and the benefit of hindsight one standard format
5334 is used which works with EVP_PKEY, RSA or DSA structures: though
5335 it clearly returns an error if you try to read the wrong kind of key.
5337 Added a -pubkey option to the 'x509' utility to output the public key.
5338 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5339 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5340 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5341 that do the same as the EVP_PKEY_assign_*() except they up the
5342 reference count of the added key (they don't "swallow" the
5346 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5347 CRLs would fail if the file contained no certificates or no CRLs:
5348 added a new function to read in both types and return the number
5349 read: this means that if none are read it will be an error. The
5350 DER versions of the certificate and CRL reader would always fail
5351 because it isn't possible to mix certificates and CRLs in DER format
5352 without choking one or the other routine. Changed this to just read
5353 a certificate: this is the best we can do. Also modified the code
5354 in apps/verify.c to take notice of return codes: it was previously
5355 attempting to read in certificates from NULL pointers and ignoring
5356 any errors: this is one reason why the cert and CRL reader seemed
5357 to work. It doesn't check return codes from the default certificate
5358 routines: these may well fail if the certificates aren't installed.
5361 *) Code to support otherName option in GeneralName.
5364 *) First update to verify code. Change the verify utility
5365 so it warns if it is passed a self signed certificate:
5366 for consistency with the normal behaviour. X509_verify
5367 has been modified to it will now verify a self signed
5368 certificate if *exactly* the same certificate appears
5369 in the store: it was previously impossible to trust a
5370 single self signed certificate. This means that:
5371 openssl verify ss.pem
5372 now gives a warning about a self signed certificate but
5373 openssl verify -CAfile ss.pem ss.pem
5377 *) For servers, store verify_result in SSL_SESSION data structure
5378 (and add it to external session representation).
5379 This is needed when client certificate verifications fails,
5380 but an application-provided verification callback (set by
5381 SSL_CTX_set_cert_verify_callback) allows accepting the session
5382 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5383 but returns 1): When the session is reused, we have to set
5384 ssl->verify_result to the appropriate error code to avoid
5386 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5388 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5389 case in PKCS7_dataInit() where the signed PKCS7 structure
5390 didn't contain any existing data because it was being created.
5391 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5393 *) Add a salt to the key derivation routines in enc.c. This
5394 forms the first 8 bytes of the encrypted file. Also add a
5395 -S option to allow a salt to be input on the command line.
5398 *) New function X509_cmp(). Oddly enough there wasn't a function
5399 to compare two certificates. We do this by working out the SHA1
5400 hash and comparing that. X509_cmp() will be needed by the trust
5404 *) SSL_get1_session() is like SSL_get_session(), but increments
5405 the reference count in the SSL_SESSION returned.
5406 [Geoff Thorpe <geoff@eu.c2.net>]
5408 *) Fix for 'req': it was adding a null to request attributes.
5409 Also change the X509_LOOKUP and X509_INFO code to handle
5410 certificate auxiliary information.
5413 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5417 *) Add the possibility to add extra information to the memory leak
5418 detecting output, to form tracebacks, showing from where each
5419 allocation was originated: CRYPTO_push_info("constant string") adds
5420 the string plus current file name and line number to a per-thread
5421 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5422 is like calling CYRPTO_pop_info() until the stack is empty.
5423 Also updated memory leak detection code to be multi-thread-safe.
5426 *) Add options -text and -noout to pkcs7 utility and delete the
5427 encryption options which never did anything. Update docs.
5430 *) Add options to some of the utilities to allow the pass phrase
5431 to be included on either the command line (not recommended on
5432 OSes like Unix) or read from the environment. Update the
5433 manpages and fix a few bugs.
5436 *) Add a few manpages for some of the openssl commands.
5439 *) Fix the -revoke option in ca. It was freeing up memory twice,
5440 leaking and not finding already revoked certificates.
5443 *) Extensive changes to support certificate auxiliary information.
5444 This involves the use of X509_CERT_AUX structure and X509_AUX
5445 functions. An X509_AUX function such as PEM_read_X509_AUX()
5446 can still read in a certificate file in the usual way but it
5447 will also read in any additional "auxiliary information". By
5448 doing things this way a fair degree of compatibility can be
5449 retained: existing certificates can have this information added
5450 using the new 'x509' options.
5452 Current auxiliary information includes an "alias" and some trust
5453 settings. The trust settings will ultimately be used in enhanced
5454 certificate chain verification routines: currently a certificate
5455 can only be trusted if it is self signed and then it is trusted
5459 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5460 The problem was that one of the replacement routines had not been working
5461 since SSLeay releases. For now the offending routine has been replaced
5462 with non-optimised assembler. Even so, this now gives around 95%
5463 performance improvement for 1024 bit RSA signs.
5466 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5467 handling. Most clients have the effective key size in bits equal to
5468 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5469 A few however don't do this and instead use the size of the decrypted key
5470 to determine the RC2 key length and the AlgorithmIdentifier to determine
5471 the effective key length. In this case the effective key length can still
5472 be 40 bits but the key length can be 168 bits for example. This is fixed
5473 by manually forcing an RC2 key into the EVP_PKEY structure because the
5474 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5475 the key length and effective key length are equal.
5478 *) Add a bunch of functions that should simplify the creation of
5479 X509_NAME structures. Now you should be able to do:
5480 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5481 and have it automatically work out the correct field type and fill in
5482 the structures. The more adventurous can try:
5483 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5484 and it will (hopefully) work out the correct multibyte encoding.
5487 *) Change the 'req' utility to use the new field handling and multibyte
5488 copy routines. Before the DN field creation was handled in an ad hoc
5489 way in req, ca, and x509 which was rather broken and didn't support
5490 BMPStrings or UTF8Strings. Since some software doesn't implement
5491 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5492 using the dirstring_type option. See the new comment in the default
5493 openssl.cnf for more info.
5496 *) Make crypto/rand/md_rand.c more robust:
5497 - Assure unique random numbers after fork().
5498 - Make sure that concurrent threads access the global counter and
5499 md serializably so that we never lose entropy in them
5500 or use exactly the same state in multiple threads.
5501 Access to the large state is not always serializable because
5502 the additional locking could be a performance killer, and
5503 md should be large enough anyway.
5506 *) New file apps/app_rand.c with commonly needed functionality
5507 for handling the random seed file.
5509 Use the random seed file in some applications that previously did not:
5511 dsaparam -genkey (which also ignored its '-rand' option),
5514 x509 (when signing).
5515 Except on systems with /dev/urandom, it is crucial to have a random
5516 seed file at least for key creation, DSA signing, and for DH exchanges;
5517 for RSA signatures we could do without one.
5519 gendh and gendsa (unlike genrsa) used to read only the first byte
5520 of each file listed in the '-rand' option. The function as previously
5521 found in genrsa is now in app_rand.c and is used by all programs
5522 that support '-rand'.
5525 *) In RAND_write_file, use mode 0600 for creating files;
5526 don't just chmod when it may be too late.
5529 *) Report an error from X509_STORE_load_locations
5530 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5533 *) New function ASN1_mbstring_copy() this copies a string in either
5534 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5535 into an ASN1_STRING type. A mask of permissible types is passed
5536 and it chooses the "minimal" type to use or an error if not type
5540 *) Add function equivalents to the various macros in asn1.h. The old
5541 macros are retained with an M_ prefix. Code inside the library can
5542 use the M_ macros. External code (including the openssl utility)
5543 should *NOT* in order to be "shared library friendly".
5546 *) Add various functions that can check a certificate's extensions
5547 to see if it usable for various purposes such as SSL client,
5548 server or S/MIME and CAs of these types. This is currently
5549 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5550 verification. Also added a -purpose flag to x509 utility to
5551 print out all the purposes.
5554 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5558 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5559 for, obtain and decode and extension and obtain its critical flag.
5560 This allows all the necessary extension code to be handled in a
5561 single function call.
5564 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5565 platforms. See crypto/rc4/rc4_enc.c for further details.
5568 *) New -noout option to asn1parse. This causes no output to be produced
5569 its main use is when combined with -strparse and -out to extract data
5570 from a file (which may not be in ASN.1 format).
5573 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5574 when producing the local key id.
5575 [Richard Levitte <levitte@stacken.kth.se>]
5577 *) New option -dhparam in s_server. This allows a DH parameter file to be
5578 stated explicitly. If it is not stated then it tries the first server
5579 certificate file. The previous behaviour hard coded the filename
5583 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5584 a public key to be input or output. For example:
5585 openssl rsa -in key.pem -pubout -out pubkey.pem
5586 Also added necessary DSA public key functions to handle this.
5589 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5590 in the message. This was handled by allowing
5591 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5592 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5594 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5595 to the end of the strings whereas this didn't. This would cause problems
5596 if strings read with d2i_ASN1_bytes() were later modified.
5597 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5599 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5600 data and it contains EOF it will end up returning an error. This is
5601 caused by input 46 bytes long. The cause is due to the way base64
5602 BIOs find the start of base64 encoded data. They do this by trying a
5603 trial decode on each line until they find one that works. When they
5604 do a flag is set and it starts again knowing it can pass all the
5605 data directly through the decoder. Unfortunately it doesn't reset
5606 the context it uses. This means that if EOF is reached an attempt
5607 is made to pass two EOFs through the context and this causes the
5608 resulting error. This can also cause other problems as well. As is
5609 usual with these problems it takes *ages* to find and the fix is
5610 trivial: move one line.
5611 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5613 *) Ugly workaround to get s_client and s_server working under Windows. The
5614 old code wouldn't work because it needed to select() on sockets and the
5615 tty (for keypresses and to see if data could be written). Win32 only
5616 supports select() on sockets so we select() with a 1s timeout on the
5617 sockets and then see if any characters are waiting to be read, if none
5618 are present then we retry, we also assume we can always write data to
5619 the tty. This isn't nice because the code then blocks until we've
5620 received a complete line of data and it is effectively polling the
5621 keyboard at 1s intervals: however it's quite a bit better than not
5622 working at all :-) A dedicated Windows application might handle this
5623 with an event loop for example.
5626 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5627 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5628 will be called when RSA_sign() and RSA_verify() are used. This is useful
5629 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5630 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5631 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5632 This necessitated the support of an extra signature type NID_md5_sha1
5633 for SSL signatures and modifications to the SSL library to use it instead
5634 of calling RSA_public_decrypt() and RSA_private_encrypt().
5637 *) Add new -verify -CAfile and -CApath options to the crl program, these
5638 will lookup a CRL issuers certificate and verify the signature in a
5639 similar way to the verify program. Tidy up the crl program so it
5640 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5641 less strict. It will now permit CRL extensions even if it is not
5642 a V2 CRL: this will allow it to tolerate some broken CRLs.
5645 *) Initialize all non-automatic variables each time one of the openssl
5646 sub-programs is started (this is necessary as they may be started
5647 multiple times from the "OpenSSL>" prompt).
5648 [Lennart Bang, Bodo Moeller]
5650 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5651 removing all other RSA functionality (this is what NO_RSA does). This
5652 is so (for example) those in the US can disable those operations covered
5653 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5657 *) Non-copying interface to BIO pairs.
5658 (still largely untested)
5661 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5662 ASCII string. This was handled independently in various places before.
5665 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5666 UTF8 strings a character at a time.
5669 *) Use client_version from client hello to select the protocol
5670 (s23_srvr.c) and for RSA client key exchange verification
5671 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5674 *) Add various utility functions to handle SPKACs, these were previously
5675 handled by poking round in the structure internals. Added new function
5676 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5677 print, verify and generate SPKACs. Based on an original idea from
5678 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5681 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5684 *) Allow the config file extension section to be overwritten on the
5685 command line. Based on an original idea from Massimiliano Pala
5686 <madwolf@comune.modena.it>. The new option is called -extensions
5687 and can be applied to ca, req and x509. Also -reqexts to override
5688 the request extensions in req and -crlexts to override the crl extensions
5692 *) Add new feature to the SPKAC handling in ca. Now you can include
5693 the same field multiple times by preceding it by "XXXX." for example:
5696 this is the same syntax as used in the req config file.
5699 *) Allow certificate extensions to be added to certificate requests. These
5700 are specified in a 'req_extensions' option of the req section of the
5701 config file. They can be printed out with the -text option to req but
5702 are otherwise ignored at present.
5705 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5706 data read consists of only the final block it would not decrypted because
5707 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5708 A misplaced 'break' also meant the decrypted final block might not be
5709 copied until the next read.
5712 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5713 a few extra parameters to the DH structure: these will be useful if
5714 for example we want the value of 'q' or implement X9.42 DH.
5717 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5718 provides hooks that allow the default DSA functions or functions on a
5719 "per key" basis to be replaced. This allows hardware acceleration and
5720 hardware key storage to be handled without major modification to the
5721 library. Also added low level modexp hooks and CRYPTO_EX structure and
5722 associated functions.
5725 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5726 as "read only": it can't be written to and the buffer it points to will
5727 not be freed. Reading from a read only BIO is much more efficient than
5728 a normal memory BIO. This was added because there are several times when
5729 an area of memory needs to be read from a BIO. The previous method was
5730 to create a memory BIO and write the data to it, this results in two
5731 copies of the data and an O(n^2) reading algorithm. There is a new
5732 function BIO_new_mem_buf() which creates a read only memory BIO from
5733 an area of memory. Also modified the PKCS#7 routines to use read only
5737 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5738 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5739 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5740 but a retry condition occured while trying to read the rest.
5743 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5744 NID_pkcs7_encrypted by default: this was wrong since this should almost
5745 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5746 the encrypted data type: this is a more sensible place to put it and it
5747 allows the PKCS#12 code to be tidied up that duplicated this
5751 *) Changed obj_dat.pl script so it takes its input and output files on
5752 the command line. This should avoid shell escape redirection problems
5756 *) Initial support for certificate extension requests, these are included
5757 in things like Xenroll certificate requests. Included functions to allow
5758 extensions to be obtained and added.
5761 *) -crlf option to s_client and s_server for sending newlines as
5762 CRLF (as required by many protocols).
5765 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5767 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5768 [Ralf S. Engelschall]
5770 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5771 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5773 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5777 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5778 DH parameters/keys (q is lost during that conversion, but the resulting
5779 DH parameters contain its length).
5781 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5782 much faster than DH_generate_parameters (which creates parameters
5783 where p = 2*q + 1), and also the smaller q makes DH computations
5784 much more efficient (160-bit exponentiation instead of 1024-bit
5785 exponentiation); so this provides a convenient way to support DHE
5786 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5787 utter importance to use
5788 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5790 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5791 when such DH parameters are used, because otherwise small subgroup
5792 attacks may become possible!
5795 *) Avoid memory leak in i2d_DHparams.
5798 *) Allow the -k option to be used more than once in the enc program:
5799 this allows the same encrypted message to be read by multiple recipients.
5802 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5803 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5804 it will always use the numerical form of the OID, even if it has a short
5808 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5809 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5810 otherwise bn_mod_exp was called. In the case of hardware keys for example
5811 no private key components need be present and it might store extra data
5812 in the RSA structure, which cannot be accessed from bn_mod_exp.
5813 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5814 private key operations.
5817 *) Added support for SPARC Linux.
5820 *) pem_password_cb function type incompatibly changed from
5821 typedef int pem_password_cb(char *buf, int size, int rwflag);
5823 ....(char *buf, int size, int rwflag, void *userdata);
5824 so that applications can pass data to their callbacks:
5825 The PEM[_ASN1]_{read,write}... functions and macros now take an
5826 additional void * argument, which is just handed through whenever
5827 the password callback is called.
5828 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5830 New function SSL_CTX_set_default_passwd_cb_userdata.
5832 Compatibility note: As many C implementations push function arguments
5833 onto the stack in reverse order, the new library version is likely to
5834 interoperate with programs that have been compiled with the old
5835 pem_password_cb definition (PEM_whatever takes some data that
5836 happens to be on the stack as its last argument, and the callback
5837 just ignores this garbage); but there is no guarantee whatsoever that
5840 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5841 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5842 problems not only on Windows, but also on some Unix platforms.
5843 To avoid problematic command lines, these definitions are now in an
5844 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5845 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5848 *) MIPS III/IV assembler module is reimplemented.
5851 *) More DES library cleanups: remove references to srand/rand and
5852 delete an unused file.
5855 *) Add support for the the free Netwide assembler (NASM) under Win32,
5856 since not many people have MASM (ml) and it can be hard to obtain.
5857 This is currently experimental but it seems to work OK and pass all
5858 the tests. Check out INSTALL.W32 for info.
5861 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5862 without temporary keys kept an extra copy of the server key,
5863 and connections with temporary keys did not free everything in case
5867 *) New function RSA_check_key and new openssl rsa option -check
5868 for verifying the consistency of RSA keys.
5869 [Ulf Moeller, Bodo Moeller]
5871 *) Various changes to make Win32 compile work:
5872 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5873 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5874 comparison" warnings.
5875 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5878 *) Add a debugging option to PKCS#5 v2 key generation function: when
5879 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5880 derived keys are printed to stderr.
5883 *) Copy the flags in ASN1_STRING_dup().
5884 [Roman E. Pavlov <pre@mo.msk.ru>]
5886 *) The x509 application mishandled signing requests containing DSA
5887 keys when the signing key was also DSA and the parameters didn't match.
5889 It was supposed to omit the parameters when they matched the signing key:
5890 the verifying software was then supposed to automatically use the CA's
5891 parameters if they were absent from the end user certificate.
5893 Omitting parameters is no longer recommended. The test was also
5894 the wrong way round! This was probably due to unusual behaviour in
5895 EVP_cmp_parameters() which returns 1 if the parameters match.
5896 This meant that parameters were omitted when they *didn't* match and
5897 the certificate was useless. Certificates signed with 'ca' didn't have
5899 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5901 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5902 The interface is as follows:
5903 Applications can use
5904 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5905 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5906 "off" is now the default.
5907 The library internally uses
5908 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5909 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5910 to disable memory-checking temporarily.
5912 Some inconsistent states that previously were possible (and were
5913 even the default) are now avoided.
5915 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5916 with each memory chunk allocated; this is occasionally more helpful
5917 than just having a counter.
5919 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5921 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5925 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5926 which largely parallels "options", but is for changing API behaviour,
5927 whereas "options" are about protocol behaviour.
5928 Initial "mode" flags are:
5930 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5931 a single record has been written.
5932 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5933 retries use the same buffer location.
5934 (But all of the contents must be
5938 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5941 *) Fix problems with no-hmac etc.
5942 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5944 *) New functions RSA_get_default_method(), RSA_set_method() and
5945 RSA_get_method(). These allows replacement of RSA_METHODs without having
5946 to mess around with the internals of an RSA structure.
5949 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5950 Also really enable memory leak checks in openssl.c and in some
5952 [Chad C. Mulligan, Bodo Moeller]
5954 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5955 up the length of negative integers. This has now been simplified to just
5956 store the length when it is first determined and use it later, rather
5957 than trying to keep track of where data is copied and updating it to
5959 [Steve Henson, reported by Brien Wheeler
5960 <bwheeler@authentica-security.com>]
5962 *) Add a new function PKCS7_signatureVerify. This allows the verification
5963 of a PKCS#7 signature but with the signing certificate passed to the
5964 function itself. This contrasts with PKCS7_dataVerify which assumes the
5965 certificate is present in the PKCS#7 structure. This isn't always the
5966 case: certificates can be omitted from a PKCS#7 structure and be
5967 distributed by "out of band" means (such as a certificate database).
5970 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5971 function prototypes in pem.h, also change util/mkdef.pl to add the
5972 necessary function names.
5975 *) mk1mf.pl (used by Windows builds) did not properly read the
5976 options set by Configure in the top level Makefile, and Configure
5977 was not even able to write more than one option correctly.
5978 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5981 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5982 file to be loaded from a BIO or FILE pointer. The BIO version will
5983 for example allow memory BIOs to contain config info.
5986 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5987 Whoever hopes to achieve shared-library compatibility across versions
5988 must use this, not the compile-time macro.
5989 (Exercise 0.9.4: Which is the minimum library version required by
5991 Note: All this applies only to multi-threaded programs, others don't
5995 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5996 through a BIO pair triggered the default case, i.e.
5997 SSLerr(...,SSL_R_UNKNOWN_STATE).
6000 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6001 can use the SSL library even if none of the specific BIOs is
6005 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6006 for the encoded length.
6007 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6009 *) Add initial documentation of the X509V3 functions.
6012 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6013 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6014 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6015 secure PKCS#8 private key format with a high iteration count.
6018 *) Fix determination of Perl interpreter: A perl or perl5
6019 _directory_ in $PATH was also accepted as the interpreter.
6020 [Ralf S. Engelschall]
6022 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6023 wrong with it but it was very old and did things like calling
6024 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6028 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6029 to use the new extension code.
6032 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6033 with macros. This should make it easier to change their form, add extra
6034 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6038 *) Add to configuration table a new entry that can specify an alternative
6039 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6040 according to Mark Crispin <MRC@Panda.COM>.
6044 *) DES CBC did not update the IV. Weird.
6047 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6048 Changing the behaviour of the former might break existing programs --
6049 where IV updating is needed, des_ncbc_encrypt can be used.
6052 *) When bntest is run from "make test" it drives bc to check its
6053 calculations, as well as internally checking them. If an internal check
6054 fails, it needs to cause bc to give a non-zero result or make test carries
6055 on without noticing the failure. Fixed.
6058 *) DES library cleanups.
6061 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6062 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6063 ciphers. NOTE: although the key derivation function has been verified
6064 against some published test vectors it has not been extensively tested
6065 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6069 *) Instead of "mkdir -p", which is not fully portable, use new
6070 Perl script "util/mkdir-p.pl".
6073 *) Rewrite the way password based encryption (PBE) is handled. It used to
6074 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6075 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6076 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6077 the 'parameter' field of the AlgorithmIdentifier is passed to the
6078 underlying key generation function so it must do its own ASN1 parsing.
6079 This has also changed the EVP_PBE_CipherInit() function which now has a
6080 'parameter' argument instead of literal salt and iteration count values
6081 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6084 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6085 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6086 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6087 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6088 value was just used as a "magic string" and not used directly its
6089 value doesn't matter.
6092 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6096 *) "linux-sparc64" configuration (ultrapenguin).
6097 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6098 "linux-sparc" configuration.
6099 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6101 *) config now generates no-xxx options for missing ciphers.
6104 *) Support the EBCDIC character set (work in progress).
6105 File ebcdic.c not yet included because it has a different license.
6106 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6108 *) Support BS2000/OSD-POSIX.
6109 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6111 *) Make callbacks for key generation use void * instead of char *.
6114 *) Make S/MIME samples compile (not yet tested).
6117 *) Additional typesafe stacks.
6120 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6124 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6126 *) New configuration variant "sco5-gcc".
6128 *) Updated some demos.
6129 [Sean O Riordain, Wade Scholine]
6131 *) Add missing BIO_free at exit of pkcs12 application.
6134 *) Fix memory leak in conf.c.
6137 *) Updates for Win32 to assembler version of MD5.
6140 *) Set #! path to perl in apps/der_chop to where we found it
6141 instead of using a fixed path.
6144 *) SHA library changes for irix64-mips4-cc.
6147 *) Improvements for VMS support.
6151 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6153 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6154 This also avoids the problems with SC4.2 and unpatched SC5.
6155 [Andy Polyakov <appro@fy.chalmers.se>]
6157 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6158 These are required because of the typesafe stack would otherwise break
6159 existing code. If old code used a structure member which used to be STACK
6160 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6161 sk_num or sk_value it would produce an error because the num, data members
6162 are not present in STACK_OF. Now it just produces a warning. sk_set
6163 replaces the old method of assigning a value to sk_value
6164 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6165 that does this will no longer work (and should use sk_set instead) but
6166 this could be regarded as a "questionable" behaviour anyway.
6169 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6170 correctly handle encrypted S/MIME data.
6173 *) Change type of various DES function arguments from des_cblock
6174 (which means, in function argument declarations, pointer to char)
6175 to des_cblock * (meaning pointer to array with 8 char elements),
6176 which allows the compiler to do more typechecking; it was like
6177 that back in SSLeay, but with lots of ugly casts.
6179 Introduce new type const_des_cblock.
6182 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6183 problems: find RecipientInfo structure that matches recipient certificate
6184 and initialise the ASN1 structures properly based on passed cipher.
6187 *) Belatedly make the BN tests actually check the results.
6190 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6191 to and from BNs: it was completely broken. New compilation option
6192 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6193 key elements as negative integers.
6196 *) Reorganize and speed up MD5.
6197 [Andy Polyakov <appro@fy.chalmers.se>]
6200 [Richard Levitte <richard@levitte.org>]
6202 *) New option -out to asn1parse to allow the parsed structure to be
6203 output to a file. This is most useful when combined with the -strparse
6204 option to examine the output of things like OCTET STRINGS.
6207 *) Make SSL library a little more fool-proof by not requiring any longer
6208 that SSL_set_{accept,connect}_state be called before
6209 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6210 in many applications because usually everything *appeared* to work as
6211 intended anyway -- now it really works as intended).
6214 *) Move openssl.cnf out of lib/.
6217 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6218 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6219 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6220 [Ralf S. Engelschall]
6222 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6223 handle PKCS#7 enveloped data properly.
6224 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6226 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6227 copying pointers. The cert_st handling is changed by this in
6228 various ways (and thus what used to be known as ctx->default_cert
6229 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6230 any longer when s->cert does not give us what we need).
6231 ssl_cert_instantiate becomes obsolete by this change.
6232 As soon as we've got the new code right (possibly it already is?),
6233 we have solved a couple of bugs of the earlier code where s->cert
6234 was used as if it could not have been shared with other SSL structures.
6236 Note that using the SSL API in certain dirty ways now will result
6237 in different behaviour than observed with earlier library versions:
6238 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6239 does not influence s as it used to.
6241 In order to clean up things more thoroughly, inside SSL_SESSION
6242 we don't use CERT any longer, but a new structure SESS_CERT
6243 that holds per-session data (if available); currently, this is
6244 the peer's certificate chain and, for clients, the server's certificate
6245 and temporary key. CERT holds only those values that can have
6246 meaningful defaults in an SSL_CTX.
6249 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6250 from the internal representation. Various PKCS#7 fixes: remove some
6251 evil casts and set the enc_dig_alg field properly based on the signing
6255 *) Allow PKCS#12 password to be set from the command line or the
6256 environment. Let 'ca' get its config file name from the environment
6257 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6261 *) Allow certificate policies extension to use an IA5STRING for the
6262 organization field. This is contrary to the PKIX definition but
6263 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6267 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6268 without disallowing inline assembler and the like for non-pedantic builds.
6271 *) Support Borland C++ builder.
6272 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6277 *) SHA-1 cleanups and performance enhancements.
6278 [Andy Polyakov <appro@fy.chalmers.se>]
6280 *) Sparc v8plus assembler for the bignum library.
6281 [Andy Polyakov <appro@fy.chalmers.se>]
6283 *) Accept any -xxx and +xxx compiler options in Configure.
6286 *) Update HPUX configuration.
6289 *) Add missing sk_<type>_unshift() function to safestack.h
6290 [Ralf S. Engelschall]
6292 *) New function SSL_CTX_use_certificate_chain_file that sets the
6293 "extra_cert"s in addition to the certificate. (This makes sense
6294 only for "PEM" format files, as chains as a whole are not
6298 *) Support verify_depth from the SSL API.
6299 x509_vfy.c had what can be considered an off-by-one-error:
6300 Its depth (which was not part of the external interface)
6301 was actually counting the number of certificates in a chain;
6302 now it really counts the depth.
6305 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6306 instead of X509err, which often resulted in confusing error
6307 messages since the error codes are not globally unique
6308 (e.g. an alleged error in ssl3_accept when a certificate
6309 didn't match the private key).
6311 *) New function SSL_CTX_set_session_id_context that allows to set a default
6312 value (so that you don't need SSL_set_session_id_context for each
6313 connection using the SSL_CTX).
6316 *) OAEP decoding bug fix.
6319 *) Support INSTALL_PREFIX for package builders, as proposed by
6323 *) New Configure options "threads" and "no-threads". For systems
6324 where the proper compiler options are known (currently Solaris
6325 and Linux), "threads" is the default.
6328 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6331 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6332 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6333 such as /usr/local/bin.
6336 *) "make linux-shared" to build shared libraries.
6337 [Niels Poppe <niels@netbox.org>]
6339 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6342 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6343 extension adding in x509 utility.
6346 *) Remove NOPROTO sections and error code comments.
6349 *) Partial rewrite of the DEF file generator to now parse the ANSI
6353 *) New Configure options --prefix=DIR and --openssldir=DIR.
6356 *) Complete rewrite of the error code script(s). It is all now handled
6357 by one script at the top level which handles error code gathering,
6358 header rewriting and C source file generation. It should be much better
6359 than the old method: it now uses a modified version of Ulf's parser to
6360 read the ANSI prototypes in all header files (thus the old K&R definitions
6361 aren't needed for error creation any more) and do a better job of
6362 translating function codes into names. The old 'ASN1 error code imbedded
6363 in a comment' is no longer necessary and it doesn't use .err files which
6364 have now been deleted. Also the error code call doesn't have to appear all
6365 on one line (which resulted in some large lines...).
6368 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6371 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6372 0 (which usually indicates a closed connection), but continue reading.
6375 *) Fix some race conditions.
6378 *) Add support for CRL distribution points extension. Add Certificate
6379 Policies and CRL distribution points documentation.
6382 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6385 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6386 8 of keying material. Merlin has also confirmed interop with this fix
6387 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6388 [Merlin Hughes <merlin@baltimore.ie>]
6390 *) Fix lots of warnings.
6391 [Richard Levitte <levitte@stacken.kth.se>]
6393 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6394 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6395 [Richard Levitte <levitte@stacken.kth.se>]
6397 *) Fix problems with sizeof(long) == 8.
6398 [Andy Polyakov <appro@fy.chalmers.se>]
6400 *) Change functions to ANSI C.
6403 *) Fix typos in error codes.
6404 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6406 *) Remove defunct assembler files from Configure.
6409 *) SPARC v8 assembler BIGNUM implementation.
6410 [Andy Polyakov <appro@fy.chalmers.se>]
6412 *) Support for Certificate Policies extension: both print and set.
6413 Various additions to support the r2i method this uses.
6416 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6417 return a const string when you are expecting an allocated buffer.
6420 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6421 types DirectoryString and DisplayText.
6424 *) Add code to allow r2i extensions to access the configuration database,
6425 add an LHASH database driver and add several ctx helper functions.
6428 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6429 fail when they extended the size of a BIGNUM.
6432 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6433 support typesafe stack.
6436 *) Fix typo in SSL_[gs]et_options().
6437 [Nils Frostberg <nils@medcom.se>]
6439 *) Delete various functions and files that belonged to the (now obsolete)
6440 old X509V3 handling code.
6443 *) New Configure option "rsaref".
6446 *) Don't auto-generate pem.h.
6449 *) Introduce type-safe ASN.1 SETs.
6452 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6453 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6455 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6456 that links with OpenSSL (well at least cause lots of warnings), but fear
6457 not: the conversion is trivial, and it eliminates loads of evil casts. A
6458 few STACKed things have been converted already. Feel free to convert more.
6459 In the fullness of time, I'll do away with the STACK type altogether.
6462 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6463 specified in <certfile> by updating the entry in the index.txt file.
6464 This way one no longer has to edit the index.txt file manually for
6465 revoking a certificate. The -revoke option does the gory details now.
6466 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6468 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6469 `-text' option at all and this way the `-noout -text' combination was
6470 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6471 [Ralf S. Engelschall]
6473 *) Make sure a corresponding plain text error message exists for the
6474 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6475 verify callback function determined that a certificate was revoked.
6476 [Ralf S. Engelschall]
6478 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6479 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6480 all available cipers including rc5, which was forgotten until now.
6481 In order to let the testing shell script know which algorithms
6482 are available, a new (up to now undocumented) command
6483 "openssl list-cipher-commands" is used.
6486 *) Bugfix: s_client occasionally would sleep in select() when
6487 it should have checked SSL_pending() first.
6490 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6491 the raw DSA values prior to ASN.1 encoding.
6494 *) Tweaks to Configure
6495 [Niels Poppe <niels@netbox.org>]
6497 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6501 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6504 *) New config option to avoid instructions that are illegal on the 80386.
6505 The default code is faster, but requires at least a 486.
6508 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6509 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6510 same as SSL2_VERSION anyway.
6513 *) New "-showcerts" option for s_client.
6516 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6517 application. Various cleanups and fixes.
6520 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6521 modify error routines to work internally. Add error codes and PBE init
6522 to library startup routines.
6525 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6526 packing functions to asn1 and evp. Changed function names and error
6527 codes along the way.
6530 *) PKCS12 integration: and so it begins... First of several patches to
6531 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6532 objects to objects.h
6535 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6536 and display support for Thawte strong extranet extension.
6539 *) Add LinuxPPC support.
6540 [Jeff Dubrule <igor@pobox.org>]
6542 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6543 bn_div_words in alpha.s.
6544 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6546 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6547 OAEP isn't supported when OpenSSL is built with RSAref.
6548 [Ulf Moeller <ulf@fitug.de>]
6550 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6551 so they no longer are missing under -DNOPROTO.
6552 [Soren S. Jorvang <soren@t.dk>]
6555 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6557 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6558 doesn't work when the session is reused. Coming soon!
6561 *) Fix a security hole, that allows sessions to be reused in the wrong
6562 context thus bypassing client cert protection! All software that uses
6563 client certs and session caches in multiple contexts NEEDS PATCHING to
6564 allow session reuse! A fuller solution is in the works.
6565 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6567 *) Some more source tree cleanups (removed obsolete files
6568 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6569 permission on "config" script to be executable) and a fix for the INSTALL
6571 [Ulf Moeller <ulf@fitug.de>]
6573 *) Remove some legacy and erroneous uses of malloc, free instead of
6575 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6577 *) Make rsa_oaep_test return non-zero on error.
6578 [Ulf Moeller <ulf@fitug.de>]
6580 *) Add support for native Solaris shared libraries. Configure
6581 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6582 if someone would make that last step automatic.
6583 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6585 *) ctx_size was not built with the right compiler during "make links". Fixed.
6588 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6589 except NULL ciphers". This means the default cipher list will no longer
6590 enable NULL ciphers. They need to be specifically enabled e.g. with
6591 the string "DEFAULT:eNULL".
6594 *) Fix to RSA private encryption routines: if p < q then it would
6595 occasionally produce an invalid result. This will only happen with
6596 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6599 *) Be less restrictive and allow also `perl util/perlpath.pl
6600 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6601 because this way one can also use an interpreter named `perl5' (which is
6602 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6603 installed as `perl').
6604 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6606 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6607 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6609 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6610 advapi32.lib to Win32 build and change the pem test comparision
6611 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6612 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6613 and crypto/des/ede_cbcm_enc.c.
6616 *) DES quad checksum was broken on big-endian architectures. Fixed.
6619 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6620 Win32 test batch file so it (might) work again. The Win32 test batch file
6621 is horrible: I feel ill....
6624 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6625 in e_os.h. Audit of header files to check ANSI and non ANSI
6626 sections: 10 functions were absent from non ANSI section and not exported
6627 from Windows DLLs. Fixed up libeay.num for new functions.
6630 *) Make `openssl version' output lines consistent.
6631 [Ralf S. Engelschall]
6633 *) Fix Win32 symbol export lists for BIO functions: Added
6634 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6635 to ms/libeay{16,32}.def.
6636 [Ralf S. Engelschall]
6638 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6639 fine under Unix and passes some trivial tests I've now added. But the
6640 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6641 added to make sure no one expects that this stuff really works in the
6642 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6643 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6645 [Ralf S. Engelschall]
6647 *) Fix the generation of two part addresses in perl.
6648 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6650 *) Add config entry for Linux on MIPS.
6651 [John Tobey <jtobey@channel1.com>]
6653 *) Make links whenever Configure is run, unless we are on Windoze.
6656 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6657 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6661 *) Add a useful kludge to allow package maintainers to specify compiler and
6662 other platforms details on the command line without having to patch the
6663 Configure script everytime: One now can use ``perl Configure
6664 <id>:<details>'', i.e. platform ids are allowed to have details appended
6665 to them (seperated by colons). This is treated as there would be a static
6666 pre-configured entry in Configure's %table under key <id> with value
6667 <details> and ``perl Configure <id>'' is called. So, when you want to
6668 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6669 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6670 now, which overrides the FreeBSD-elf entry on-the-fly.
6671 [Ralf S. Engelschall]
6673 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6676 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6677 on the `perl Configure ...' command line. This way one can compile
6678 OpenSSL libraries with Position Independent Code (PIC) which is needed
6679 for linking it into DSOs.
6680 [Ralf S. Engelschall]
6682 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6686 *) Cleaned up the LICENSE document: The official contact for any license
6687 questions now is the OpenSSL core team under openssl-core@openssl.org.
6688 And add a paragraph about the dual-license situation to make sure people
6689 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6690 to the OpenSSL toolkit.
6691 [Ralf S. Engelschall]
6693 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6694 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6695 Additonally cleaned up the `make links' target: Remove unnecessary
6696 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6697 to speed processing and no longer clutter the display with confusing
6698 stuff. Instead only the actually done links are displayed.
6699 [Ralf S. Engelschall]
6701 *) Permit null encryption ciphersuites, used for authentication only. It used
6702 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6703 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6707 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6708 signed attributes when verifying signatures (this would break them),
6709 the detached data encoding was wrong and public keys obtained using
6710 X509_get_pubkey() weren't freed.
6713 *) Add text documentation for the BUFFER functions. Also added a work around
6714 to a Win95 console bug. This was triggered by the password read stuff: the
6715 last character typed gets carried over to the next fread(). If you were
6716 generating a new cert request using 'req' for example then the last
6717 character of the passphrase would be CR which would then enter the first
6721 *) Added the new `Includes OpenSSL Cryptography Software' button as
6722 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6723 button and can be used by applications based on OpenSSL to show the
6724 relationship to the OpenSSL project.
6725 [Ralf S. Engelschall]
6727 *) Remove confusing variables in function signatures in files
6728 ssl/ssl_lib.c and ssl/ssl.h.
6729 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6731 *) Don't install bss_file.c under PREFIX/include/
6732 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6734 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6735 functions that return function pointers and has support for NT specific
6736 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6737 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6738 unsigned to signed types: this was killing the Win32 compile.
6741 *) Add new certificate file to stack functions,
6742 SSL_add_dir_cert_subjects_to_stack() and
6743 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6744 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6745 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6746 This means that Apache-SSL and similar packages don't have to mess around
6747 to add as many CAs as they want to the preferred list.
6750 *) Experiment with doxygen documentation. Currently only partially applied to
6752 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6753 openssl.doxy as the configuration file.
6756 *) Get rid of remaining C++-style comments which strict C compilers hate.
6757 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6759 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6760 compiled in by default: it has problems with large keys.
6763 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6764 DH private keys and/or callback functions which directly correspond to
6765 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6766 is needed for applications which have to configure certificates on a
6767 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6769 For the RSA certificate situation is makes no difference, but
6770 for the DSA certificate situation this fixes the "no shared cipher"
6771 problem where the OpenSSL cipher selection procedure failed because the
6772 temporary keys were not overtaken from the context and the API provided
6773 no way to reconfigure them.
6774 The new functions now let applications reconfigure the stuff and they
6775 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6776 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6777 non-public-API function ssl_cert_instantiate() is used as a helper
6778 function and also to reduce code redundancy inside ssl_rsa.c.
6779 [Ralf S. Engelschall]
6781 *) Move s_server -dcert and -dkey options out of the undocumented feature
6782 area because they are useful for the DSA situation and should be
6783 recognized by the users.
6784 [Ralf S. Engelschall]
6786 *) Fix the cipher decision scheme for export ciphers: the export bits are
6787 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6788 SSL_EXP_MASK. So, the original variable has to be used instead of the
6789 already masked variable.
6790 [Richard Levitte <levitte@stacken.kth.se>]
6792 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6793 [Richard Levitte <levitte@stacken.kth.se>]
6795 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6796 from `int' to `unsigned int' because it's a length and initialized by
6797 EVP_DigestFinal() which expects an `unsigned int *'.
6798 [Richard Levitte <levitte@stacken.kth.se>]
6800 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6801 script. Instead use the usual Shell->Perl transition trick.
6802 [Ralf S. Engelschall]
6804 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6805 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6806 -noout -modulus' as it's already the case for `openssl rsa -noout
6807 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6808 currently the public key is printed (a decision which was already done by
6809 `openssl dsa -modulus' in the past) which serves a similar purpose.
6810 Additionally the NO_RSA no longer completely removes the whole -modulus
6811 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6813 [Ralf S. Engelschall]
6815 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6816 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6817 [Arne Ansper <arne@ats.cyber.ee>]
6819 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6820 to be added. Now both 'req' and 'ca' can use new objects defined in the
6824 *) Add cool BIO that does syslog (or event log on NT).
6825 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6827 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6828 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6829 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6830 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6833 *) Add preliminary config info for new extension code.
6836 *) Make RSA_NO_PADDING really use no padding.
6837 [Ulf Moeller <ulf@fitug.de>]
6839 *) Generate errors when private/public key check is done.
6842 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6843 for some CRL extensions and new objects added.
6846 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6847 key usage extension and fuller support for authority key id.
6850 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6851 padding method for RSA, which is recommended for new applications in PKCS
6852 #1 v2.0 (RFC 2437, October 1998).
6853 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6854 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6855 against Bleichbacher's attack on RSA.
6856 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6859 *) Updates to the new SSL compression code
6860 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6862 *) Fix so that the version number in the master secret, when passed
6863 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6864 (because the server will not accept higher), that the version number
6865 is 0x03,0x01, not 0x03,0x00
6866 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6868 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6869 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6870 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6873 *) Support for RAW extensions where an arbitrary extension can be
6874 created by including its DER encoding. See apps/openssl.cnf for
6878 *) Make sure latest Perl versions don't interpret some generated C array
6879 code as Perl array code in the crypto/err/err_genc.pl script.
6880 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6882 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6883 not many people have the assembler. Various Win32 compilation fixes and
6884 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6888 *) Modify configure script 'Configure' to automatically create crypto/date.h
6889 file under Win32 and also build pem.h from pem.org. New script
6890 util/mkfiles.pl to create the MINFO file on environments that can't do a
6891 'make files': perl util/mkfiles.pl >MINFO should work.
6894 *) Major rework of DES function declarations, in the pursuit of correctness
6895 and purity. As a result, many evil casts evaporated, and some weirdness,
6896 too. You may find this causes warnings in your code. Zapping your evil
6897 casts will probably fix them. Mostly.
6900 *) Fix for a typo in asn1.h. Bug fix to object creation script
6901 obj_dat.pl. It considered a zero in an object definition to mean
6902 "end of object": none of the objects in objects.h have any zeros
6903 so it wasn't spotted.
6904 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6906 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6907 Masking (CBCM). In the absence of test vectors, the best I have been able
6908 to do is check that the decrypt undoes the encrypt, so far. Send me test
6909 vectors if you have them.
6912 *) Correct calculation of key length for export ciphers (too much space was
6913 allocated for null ciphers). This has not been tested!
6916 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6917 message is now correct (it understands "crypto" and "ssl" on its
6918 command line). There is also now an "update" option. This will update
6919 the util/ssleay.num and util/libeay.num files with any new functions.
6921 perl util/mkdef.pl crypto ssl update
6922 it will update them.
6925 *) Overhauled the Perl interface (perl/*):
6926 - ported BN stuff to OpenSSL's different BN library
6927 - made the perl/ source tree CVS-aware
6928 - renamed the package from SSLeay to OpenSSL (the files still contain
6929 their history because I've copied them in the repository)
6930 - removed obsolete files (the test scripts will be replaced
6931 by better Test::Harness variants in the future)
6932 [Ralf S. Engelschall]
6934 *) First cut for a very conservative source tree cleanup:
6935 1. merge various obsolete readme texts into doc/ssleay.txt
6936 where we collect the old documents and readme texts.
6937 2. remove the first part of files where I'm already sure that we no
6938 longer need them because of three reasons: either they are just temporary
6939 files which were left by Eric or they are preserved original files where
6940 I've verified that the diff is also available in the CVS via "cvs diff
6941 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6942 the crypto/md/ stuff).
6943 [Ralf S. Engelschall]
6945 *) More extension code. Incomplete support for subject and issuer alt
6946 name, issuer and authority key id. Change the i2v function parameters
6947 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6948 what that's for :-) Fix to ASN1 macro which messed up
6949 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6952 *) Preliminary support for ENUMERATED type. This is largely copied from the
6956 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6957 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6959 *) Make sure `make rehash' target really finds the `openssl' program.
6960 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6962 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6963 like to hear about it if this slows down other processors.
6966 *) Add CygWin32 platform information to Configure script.
6967 [Alan Batie <batie@aahz.jf.intel.com>]
6969 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6970 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6972 *) New program nseq to manipulate netscape certificate sequences
6975 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6979 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6980 but the BN code had some problems that would cause failures when
6981 doing certificate verification and some other functions.
6982 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6984 *) Add ASN1 and PEM code to support netscape certificate sequences.
6987 *) Add ASN1 and PEM code to support netscape certificate sequences.
6990 *) Add several PKIX and private extended key usage OIDs.
6993 *) Modify the 'ca' program to handle the new extension code. Modify
6994 openssl.cnf for new extension format, add comments.
6997 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6998 and add a sample to openssl.cnf so req -x509 now adds appropriate
7002 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7003 error code, add initial support to X509_print() and x509 application.
7006 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7007 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7008 stuff is currently isolated and isn't even compiled yet.
7011 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7012 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7013 Removed the versions check from X509 routines when loading extensions:
7014 this allows certain broken certificates that don't set the version
7015 properly to be processed.
7018 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7019 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7020 can still be regenerated with "make depend".
7023 *) Spelling mistake in C version of CAST-128.
7024 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7026 *) Changes to the error generation code. The perl script err-code.pl
7027 now reads in the old error codes and retains the old numbers, only
7028 adding new ones if necessary. It also only changes the .err files if new
7029 codes are added. The makefiles have been modified to only insert errors
7030 when needed (to avoid needlessly modifying header files). This is done
7031 by only inserting errors if the .err file is newer than the auto generated
7032 C file. To rebuild all the error codes from scratch (the old behaviour)
7033 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7034 or delete all the .err files.
7037 *) CAST-128 was incorrectly implemented for short keys. The C version has
7038 been fixed, but is untested. The assembler versions are also fixed, but
7039 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7040 to regenerate it if needed.
7041 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7042 Hagino <itojun@kame.net>]
7044 *) File was opened incorrectly in randfile.c.
7045 [Ulf Möller <ulf@fitug.de>]
7047 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7048 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7049 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7050 al: it's just almost always a UTCTime. Note this patch adds new error
7051 codes so do a "make errors" if there are problems.
7054 *) Correct Linux 1 recognition in config.
7055 [Ulf Möller <ulf@fitug.de>]
7057 *) Remove pointless MD5 hash when using DSA keys in ca.
7058 [Anonymous <nobody@replay.com>]
7060 *) Generate an error if given an empty string as a cert directory. Also
7061 generate an error if handed NULL (previously returned 0 to indicate an
7062 error, but didn't set one).
7063 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7065 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7068 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7069 parameters. This was causing a warning which killed off the Win32 compile.
7072 *) Remove C++ style comments from crypto/bn/bn_local.h.
7073 [Neil Costigan <neil.costigan@celocom.com>]
7075 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7076 based on a text string, looking up short and long names and finally
7077 "dot" format. The "dot" format stuff didn't work. Added new function
7078 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7079 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7080 OID is not part of the table.
7083 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7084 X509_LOOKUP_by_alias().
7087 *) Sort openssl functions by name.
7090 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7091 encryption from sample DSA keys (in case anyone is interested the password
7095 *) Make _all_ *_free functions accept a NULL pointer.
7096 [Frans Heymans <fheymans@isaserver.be>]
7098 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7100 [Anonymous <nobody@replay.com>]
7102 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7103 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7105 *) Don't blow it for numeric -newkey arguments to apps/req.
7106 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7108 *) Temp key "for export" tests were wrong in s3_srvr.c.
7109 [Anonymous <nobody@replay.com>]
7111 *) Add prototype for temp key callback functions
7112 SSL_CTX_set_tmp_{rsa,dh}_callback().
7115 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7116 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7119 *) X509_name_add_entry() freed the wrong thing after an error.
7120 [Arne Ansper <arne@ats.cyber.ee>]
7122 *) rsa_eay.c would attempt to free a NULL context.
7123 [Arne Ansper <arne@ats.cyber.ee>]
7125 *) BIO_s_socket() had a broken should_retry() on Windoze.
7126 [Arne Ansper <arne@ats.cyber.ee>]
7128 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7129 [Arne Ansper <arne@ats.cyber.ee>]
7131 *) Make sure the already existing X509_STORE->depth variable is initialized
7132 in X509_STORE_new(), but document the fact that this variable is still
7133 unused in the certificate verification process.
7134 [Ralf S. Engelschall]
7136 *) Fix the various library and apps files to free up pkeys obtained from
7137 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7140 *) Fix reference counting in X509_PUBKEY_get(). This makes
7141 demos/maurice/example2.c work, amongst others, probably.
7142 [Steve Henson and Ben Laurie]
7144 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7145 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7146 are no longer created. This way we have a single and consistent command
7147 line interface `openssl <command>', similar to `cvs <command>'.
7148 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7150 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7151 BIT STRING wrapper always have zero unused bits.
7154 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7157 *) Make the top-level INSTALL documentation easier to understand.
7160 *) Makefiles updated to exit if an error occurs in a sub-directory
7161 make (including if user presses ^C) [Paul Sutton]
7163 *) Make Montgomery context stuff explicit in RSA data structure.
7166 *) Fix build order of pem and err to allow for generated pem.h.
7169 *) Fix renumbering bug in X509_NAME_delete_entry().
7172 *) Enhanced the err-ins.pl script so it makes the error library number
7173 global and can add a library name. This is needed for external ASN1 and
7174 other error libraries.
7177 *) Fixed sk_insert which never worked properly.
7180 *) Fix ASN1 macros so they can handle indefinite length construted
7181 EXPLICIT tags. Some non standard certificates use these: they can now
7185 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7186 into a single doc/ssleay.txt bundle. This way the information is still
7187 preserved but no longer messes up this directory. Now it's new room for
7188 the new set of documenation files.
7189 [Ralf S. Engelschall]
7191 *) SETs were incorrectly DER encoded. This was a major pain, because they
7192 shared code with SEQUENCEs, which aren't coded the same. This means that
7193 almost everything to do with SETs or SEQUENCEs has either changed name or
7194 number of arguments.
7195 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7197 *) Fix test data to work with the above.
7200 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7201 was already fixed by Eric for 0.9.1 it seems.
7202 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7204 *) Autodetect FreeBSD3.
7207 *) Fix various bugs in Configure. This affects the following platforms:
7211 unixware-2.0-pentium
7215 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7216 before they are needed.
7219 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7223 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7225 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7226 changed SSLeay to OpenSSL in version strings.
7227 [Ralf S. Engelschall]
7229 *) Some fixups to the top-level documents.
7232 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7233 because the symlink to include/ was missing.
7234 [Ralf S. Engelschall]
7236 *) Incorporated the popular no-RSA/DSA-only patches
7237 which allow to compile a RSA-free SSLeay.
7238 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7240 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7241 when "ssleay" is still not found.
7242 [Ralf S. Engelschall]
7244 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7245 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7247 *) Updated the README file.
7248 [Ralf S. Engelschall]
7250 *) Added various .cvsignore files in the CVS repository subdirs
7251 to make a "cvs update" really silent.
7252 [Ralf S. Engelschall]
7254 *) Recompiled the error-definition header files and added
7255 missing symbols to the Win32 linker tables.
7256 [Ralf S. Engelschall]
7258 *) Cleaned up the top-level documents;
7259 o new files: CHANGES and LICENSE
7260 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7261 o merged COPYRIGHT into LICENSE
7262 o removed obsolete TODO file
7263 o renamed MICROSOFT to INSTALL.W32
7264 [Ralf S. Engelschall]
7266 *) Removed dummy files from the 0.9.1b source tree:
7267 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7268 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7269 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7270 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7271 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7272 [Ralf S. Engelschall]
7274 *) Added various platform portability fixes.
7277 *) The Genesis of the OpenSSL rpject:
7278 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7279 Young and Tim J. Hudson created while they were working for C2Net until
7281 [The OpenSSL Project]
7284 Changes between 0.9.0b and 0.9.1b [not released]
7286 *) Updated a few CA certificates under certs/
7289 *) Changed some BIGNUM api stuff.
7292 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7293 DGUX x86, Linux Alpha, etc.
7296 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7297 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7301 *) Add -strparse option to asn1pars program which parses nested
7303 [Dr Stephen Henson <shenson@bigfoot.com>]
7305 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7308 *) DSA fix for "ca" program.
7311 *) Added "-genkey" option to "dsaparam" program.
7314 *) Added RIPE MD160 (rmd160) message digest.
7317 *) Added -a (all) option to "ssleay version" command.
7320 *) Added PLATFORM define which is the id given to Configure.
7323 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7326 *) Extended the ASN.1 parser routines.
7329 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7332 *) Added a BN_CTX to the BN library.
7335 *) Fixed the weak key values in DES library
7338 *) Changed API in EVP library for cipher aliases.
7341 *) Added support for RC2/64bit cipher.
7344 *) Converted the lhash library to the crypto/mem.c functions.
7347 *) Added more recognized ASN.1 object ids.
7350 *) Added more RSA padding checks for SSL/TLS.
7353 *) Added BIO proxy/filter functionality.
7356 *) Added extra_certs to SSL_CTX which can be used
7357 send extra CA certificates to the client in the CA cert chain sending
7358 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7361 *) Now Fortezza is denied in the authentication phase because
7362 this is key exchange mechanism is not supported by SSLeay at all.
7365 *) Additional PKCS1 checks.
7368 *) Support the string "TLSv1" for all TLS v1 ciphers.
7371 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7372 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7375 *) Fixed a few memory leaks.
7378 *) Fixed various code and comment typos.
7381 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7382 bytes sent in the client random.
7383 [Edward Bishop <ebishop@spyglass.com>]