5 Changes between 0.9.7e and 0.9.8 [xx XXX xxxx]
7 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
10 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
11 [Andy Polyakov and a number of other people]
13 *) Improved PowerPC platform support. Most notably BIGNUM assembler
14 implementation contributed by IBM.
15 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
17 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
18 exponent rather than 'unsigned long'. There is a corresponding change to
19 the new 'rsa_keygen' element of the RSA_METHOD structure.
20 [Jelte Jansen, Geoff Thorpe]
22 *) Functionality for creating the initial serial number file is now
23 moved from CA.pl to the 'ca' utility with a new option -create_serial.
25 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
26 number file to 1, which is bound to cause problems. To avoid
27 the problems while respecting compatibility between different 0.9.7
28 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
29 CA.pl for serial number initialization. With the new release 0.9.8,
30 we can fix the problem directly in the 'ca' utility.)
33 *) Reduced header interdepencies by declaring more opaque objects in
34 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
35 give fewer recursive includes, which could break lazy source code - so
36 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
37 developers should define this symbol when building and using openssl to
38 ensure they track the recommended behaviour, interfaces, [etc], but
39 backwards-compatible behaviour prevails when this isn't defined.
42 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
45 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
46 This will generate a random key of the appropriate length based on the
47 cipher context. The EVP_CIPHER can provide its own random key generation
48 routine to support keys of a specific form. This is used in the des and
49 3des routines to generate a key of the correct parity. Update S/MIME
50 code to use new functions and hence generate correct parity DES keys.
51 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
52 valid (weak or incorrect parity).
55 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
56 as looking them up. This is useful when the verified structure may contain
57 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
58 present unless the new PKCS7_NO_CRL flag is asserted.
61 *) Extend ASN1 oid configuration module. It now additionally accepts the
64 shortName = some long name, 1.2.3.4
67 *) Reimplemented the BN_CTX implementation. There is now no more static
68 limitation on the number of variables it can handle nor the depth of the
69 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
70 information can now expand as required, and rather than having a single
71 static array of bignums, BN_CTX now uses a linked-list of such arrays
72 allowing it to expand on demand whilst maintaining the usefulness of
76 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
77 to allow all RSA operations to function using a single BN_CTX.
80 *) Preliminary support for certificate policy evaluation and checking. This
81 is initially intended to pass the tests outlined in "Conformance Testing
82 of Relying Party Client Certificate Path Processing Logic" v1.07.
85 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
86 remained unused and not that useful. A variety of other little bignum
87 tweaks and fixes have also been made continuing on from the audit (see
91 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
92 associated ASN1, EVP and SSL functions and old ASN1 macros.
95 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
96 and this should never fail. So the return value from the use of
97 BN_set_word() (which can fail due to needless expansion) is now deprecated;
98 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
101 *) BN_CTX_get() should return zero-valued bignums, providing the same
102 initialised value as BN_new().
103 [Geoff Thorpe, suggested by Ulf Möller]
105 *) Support for inhibitAnyPolicy certificate extension.
108 *) An audit of the BIGNUM code is underway, for which debugging code is
109 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
110 is considered valid when processing BIGNUMs, and causes execution to
111 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
112 further steps are taken to deliberately pollute unused data in BIGNUM
113 structures to try and expose faulty code further on. For now, openssl will
114 (in its default mode of operation) continue to tolerate the inconsistent
115 forms that it has tolerated in the past, but authors and packagers should
116 consider trying openssl and their own applications when compiled with
117 these debugging symbols defined. It will help highlight potential bugs in
118 their own code, and will improve the test coverage for OpenSSL itself. At
119 some point, these tighter rules will become openssl's default to improve
120 maintainability, though the assert()s and other overheads will remain only
121 in debugging configurations. See bn.h for more details.
122 [Geoff Thorpe, Nils Larsch, Ulf Möller]
124 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
125 that can only be obtained through BN_CTX_new() (which implicitly
126 initialises it). The presence of this function only made it possible
127 to overwrite an existing structure (and cause memory leaks).
130 *) Because of the callback-based approach for implementing LHASH as a
131 template type, lh_insert() adds opaque objects to hash-tables and
132 lh_doall() or lh_doall_arg() are typically used with a destructor callback
133 to clean up those corresponding objects before destroying the hash table
134 (and losing the object pointers). So some over-zealous constifications in
135 LHASH have been relaxed so that lh_insert() does not take (nor store) the
136 objects as "const" and the lh_doall[_arg] callback wrappers are not
137 prototyped to have "const" restrictions on the object pointers they are
138 given (and so aren't required to cast them away any more).
141 *) The tmdiff.h API was so ugly and minimal that our own timing utility
142 (speed) prefers to use its own implementation. The two implementations
143 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
144 its object type properly exposed (MS_TM) instead of casting to/from "char
145 *". This may still change yet if someone realises MS_TM and "ms_time_***"
146 aren't necessarily the greatest nomenclatures - but this is what was used
147 internally to the implementation so I've used that for now.
150 *) Ensure that deprecated functions do not get compiled when
151 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
152 the self-tests were still using deprecated key-generation functions so
153 these have been updated also.
156 *) Reorganise PKCS#7 code to separate the digest location functionality
157 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
158 New function PKCS7_set_digest() to set the digest type for PKCS#7
159 digestedData type. Add additional code to correctly generate the
160 digestedData type and add support for this type in PKCS7 initialization
164 *) New function PKCS7_set0_type_other() this initializes a PKCS7
165 structure of type "other".
168 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
169 sure the loop does correctly stop and breaking ("division by zero")
170 modulus operations are not performed. The (pre-generated) prime
171 table crypto/bn/bn_prime.h was already correct, but it could not be
172 re-generated on some platforms because of the "division by zero"
173 situation in the script.
174 [Ralf S. Engelschall]
176 *) Update support for ECC-based TLS ciphersuites according to
177 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
178 SHA-1 now is only used for "small" curves (where the
179 representation of a field element takes up to 24 bytes); for
180 larger curves, the field element resulting from ECDH is directly
181 used as premaster secret.
182 [Douglas Stebila (Sun Microsystems Laboratories)]
184 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
185 curve secp160r1 to the tests.
186 [Douglas Stebila (Sun Microsystems Laboratories)]
188 *) Add the possibility to load symbols globally with DSO.
189 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
191 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
192 control of the error stack.
195 *) Add support for STORE in ENGINE.
198 *) Add the STORE type. The intention is to provide a common interface
199 to certificate and key stores, be they simple file-based stores, or
200 HSM-type store, or LDAP stores, or...
201 NOTE: The code is currently UNTESTED and isn't really used anywhere.
204 *) Add a generic structure called OPENSSL_ITEM. This can be used to
205 pass a list of arguments to any function as well as provide a way
206 for a function to pass data back to the caller.
209 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
210 works like BUF_strdup() but can be used to duplicate a portion of
211 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
215 *) Add the function sk_find_ex() which works like sk_find(), but will
216 return an index to an element even if an exact match couldn't be
217 found. The index is guaranteed to point at the element where the
218 searched-for key would be inserted to preserve sorting order.
221 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
222 takes an extra flags argument for optional functionality. Currently,
223 the following flags are defined:
225 OBJ_BSEARCH_VALUE_ON_NOMATCH
226 This one gets OBJ_bsearch_ex() to return a pointer to the first
227 element where the comparing function returns a negative or zero
230 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
231 This one gets OBJ_bsearch_ex() to return a pointer to the first
232 element where the comparing function returns zero. This is useful
233 if there are more than one element where the comparing function
237 *) Make it possible to create self-signed certificates with 'openssl ca'
238 in such a way that the self-signed certificate becomes part of the
239 CA database and uses the same mechanisms for serial number generation
240 as all other certificate signing. The new flag '-selfsign' enables
241 this functionality. Adapt CA.sh and CA.pl.in.
244 *) Add functionality to check the public key of a certificate request
245 against a given private. This is useful to check that a certificate
246 request can be signed by that key (self-signing).
249 *) Make it possible to have multiple active certificates with the same
250 subject in the CA index file. This is done only if the keyword
251 'unique_subject' is set to 'no' in the main CA section (default
252 if 'CA_default') of the configuration file. The value is saved
253 with the database itself in a separate index attribute file,
254 named like the index file with '.attr' appended to the name.
257 *) Generate muti valued AVAs using '+' notation in config files for
261 *) Support for nameConstraints certificate extension.
264 *) Support for policyConstraints certificate extension.
267 *) Support for policyMappings certificate extension.
270 *) Make sure the default DSA_METHOD implementation only uses its
271 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
272 and change its own handlers to be NULL so as to remove unnecessary
273 indirection. This lets alternative implementations fallback to the
274 default implementation more easily.
277 *) Support for directoryName in GeneralName related extensions
281 *) Make it possible to link applications using Makefile.shared.
282 Make that possible even when linking against static libraries!
285 *) Support for single pass processing for S/MIME signing. This now
286 means that S/MIME signing can be done from a pipe, in addition
287 cleartext signing (multipart/signed type) is effectively streaming
288 and the signed data does not need to be all held in memory.
290 This is done with a new flag PKCS7_STREAM. When this flag is set
291 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
292 is done after the data is output (and digests calculated) in
296 *) Add full support for -rpath/-R, both in shared libraries and
297 applications, at least on the platforms where it's known how
301 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
302 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
303 will now compute a table of multiples of the generator that
304 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
305 faster (notably in the case of a single point multiplication,
307 [Nils Larsch, Bodo Moeller]
309 *) IPv6 support for certificate extensions. The various extensions
310 which use the IP:a.b.c.d can now take IPv6 addresses using the
311 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
315 *) Added an ENGINE that implements RSA by performing private key
316 exponentiations with the GMP library. The conversions to and from
317 GMP's mpz_t format aren't optimised nor are any montgomery forms
318 cached, and on x86 it appears OpenSSL's own performance has caught up.
319 However there are likely to be other architectures where GMP could
320 provide a boost. This ENGINE is not built in by default, but it can be
321 specified at Configure time and should be accompanied by the necessary
322 linker additions, eg;
323 ./config -DOPENSSL_USE_GMP -lgmp
326 *) "openssl engine" will not display ENGINE/DSO load failure errors when
327 testing availability of engines with "-t" - the old behaviour is
328 produced by increasing the feature's verbosity with "-tt".
331 *) ECDSA routines: under certain error conditions uninitialized BN objects
332 could be freed. Solution: make sure initialization is performed early
333 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
337 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
338 and DH_METHOD (eg. by ENGINE implementations) to override the normal
339 software implementations. For DSA and DH, parameter generation can
340 also be overriden by providing the appropriate method callbacks.
343 *) Change the "progress" mechanism used in key-generation and
344 primality testing to functions that take a new BN_GENCB pointer in
345 place of callback/argument pairs. The new API functions have "_ex"
346 postfixes and the older functions are reimplemented as wrappers for
347 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
348 declarations of the old functions to help (graceful) attempts to
349 migrate to the new functions. Also, the new key-generation API
350 functions operate on a caller-supplied key-structure and return
351 success/failure rather than returning a key or NULL - this is to
352 help make "keygen" another member function of RSA_METHOD etc.
354 Example for using the new callback interface:
356 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
360 BN_GENCB_set(&my_cb, my_callback, my_arg);
362 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
363 /* For the meaning of a, b in calls to my_callback(), see the
364 * documentation of the function that calls the callback.
365 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
366 * my_callback should return 1 if it wants BN_is_prime_ex()
367 * to continue, or 0 to stop.
372 *) Change the ZLIB compression method to be stateful, and make it
373 available to TLS with the number defined in
374 draft-ietf-tls-compression-04.txt.
377 *) Add the ASN.1 structures and functions for CertificatePair, which
378 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
380 CertificatePair ::= SEQUENCE {
381 forward [0] Certificate OPTIONAL,
382 reverse [1] Certificate OPTIONAL,
383 -- at least one of the pair shall be present -- }
385 Also implement the PEM functions to read and write certificate
386 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
388 This needed to be defined, mostly for the sake of the LDAP
389 attribute crossCertificatePair, but may prove useful elsewhere as
393 *) Make it possible to inhibit symlinking of shared libraries in
394 Makefile.shared, for Cygwin's sake.
397 *) Extend the BIGNUM API by creating new macros that behave like
400 void BN_set_sign(BIGNUM *a, int neg);
401 int BN_get_sign(const BIGNUM *a);
403 and avoid the need to access 'a->neg' directly in applications.
404 [Nils Larsch <nla@trustcenter.de>]
406 *) Implement fast modular reduction for pseudo-Mersenne primes
407 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
408 EC_GROUP_new_curve_GFp() will now automatically use this
410 [Nils Larsch <nla@trustcenter.de>]
412 *) Add new lock type (CRYPTO_LOCK_BN).
415 *) Change the ENGINE framework to automatically load engines
416 dynamically from specific directories unless they could be
417 found to already be built in or loaded. Move all the
418 current engines except for the cryptodev one to a new
420 The engines in engines/ are built as shared libraries if
421 the "shared" options was given to ./Configure or ./config.
422 Otherwise, they are inserted in libcrypto.a.
423 /usr/local/ssl/engines is the default directory for dynamic
424 engines, but that can be overriden at configure time through
425 the usual use of --prefix and/or --openssldir, and at run
426 time with the environment variable OPENSSL_ENGINES.
427 [Geoff Thorpe and Richard Levitte]
429 *) Add Makefile.shared, a helper makefile to build shared
430 libraries. Addapt Makefile.org.
433 *) Add version info to Win32 DLLs.
434 [Peter 'Luna' Runestig" <peter@runestig.com>]
436 *) Add new 'medium level' PKCS#12 API. Certificates and keys
437 can be added using this API to created arbitrary PKCS#12
438 files while avoiding the low level API.
440 New options to PKCS12_create(), key or cert can be NULL and
441 will then be omitted from the output file. The encryption
442 algorithm NIDs can be set to -1 for no encryption, the mac
443 iteration count can be set to 0 to omit the mac.
445 Enhance pkcs12 utility by making the -nokeys and -nocerts
446 options work when creating a PKCS#12 file. New option -nomac
447 to omit the mac, NONE can be set for an encryption algorithm.
448 New code is modified to use the enhanced PKCS12_create()
449 instead of the low level API.
452 *) Extend ASN1 encoder to support indefinite length constructed
453 encoding. This can output sequences tags and octet strings in
454 this form. Modify pk7_asn1.c to support indefinite length
455 encoding. This is experimental and needs additional code to
456 be useful, such as an ASN1 bio and some enhanced streaming
459 Extend template encode functionality so that tagging is passed
460 down to the template encoder.
463 *) Let 'openssl req' fail if an argument to '-newkey' is not
464 recognized instead of using RSA as a default.
467 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
468 As these are not official, they are not included in "ALL";
469 the "ECCdraft" ciphersuite group alias can be used to select them.
470 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
472 *) Add ECDH engine support.
473 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
475 *) Add ECDH in new directory crypto/ecdh/.
476 [Douglas Stebila (Sun Microsystems Laboratories)]
478 *) Let BN_rand_range() abort with an error after 100 iterations
479 without success (which indicates a broken PRNG).
482 *) Change BN_mod_sqrt() so that it verifies that the input value
483 is really the square of the return value. (Previously,
484 BN_mod_sqrt would show GIGO behaviour.)
487 *) Add named elliptic curves over binary fields from X9.62, SECG,
488 and WAP/WTLS; add OIDs that were still missing.
490 [Sheueling Chang Shantz and Douglas Stebila
491 (Sun Microsystems Laboratories)]
493 *) Extend the EC library for elliptic curves over binary fields
494 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
497 EC_GF2m_simple_method
501 EC_GROUP_new_curve_GF2m
502 EC_GROUP_set_curve_GF2m
503 EC_GROUP_get_curve_GF2m
504 EC_POINT_set_affine_coordinates_GF2m
505 EC_POINT_get_affine_coordinates_GF2m
506 EC_POINT_set_compressed_coordinates_GF2m
508 Point compression for binary fields is disabled by default for
509 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
512 As binary polynomials are represented as BIGNUMs, various members
513 of the EC_GROUP and EC_POINT data structures can be shared
514 between the implementations for prime fields and binary fields;
515 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
516 are essentially identical to their ..._GFp counterparts.
517 (For simplicity, the '..._GFp' prefix has been dropped from
518 various internal method names.)
520 An internal 'field_div' method (similar to 'field_mul' and
521 'field_sqr') has been added; this is used only for binary fields.
523 [Sheueling Chang Shantz and Douglas Stebila
524 (Sun Microsystems Laboratories)]
526 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
527 through methods ('mul', 'precompute_mult').
529 The generic implementations (now internally called 'ec_wNAF_mul'
530 and 'ec_wNAF_precomputed_mult') remain the default if these
531 methods are undefined.
533 [Sheueling Chang Shantz and Douglas Stebila
534 (Sun Microsystems Laboratories)]
536 *) New function EC_GROUP_get_degree, which is defined through
537 EC_METHOD. For curves over prime fields, this returns the bit
538 length of the modulus.
540 [Sheueling Chang Shantz and Douglas Stebila
541 (Sun Microsystems Laboratories)]
543 *) New functions EC_GROUP_dup, EC_POINT_dup.
544 (These simply call ..._new and ..._copy).
546 [Sheueling Chang Shantz and Douglas Stebila
547 (Sun Microsystems Laboratories)]
549 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
550 Polynomials are represented as BIGNUMs (where the sign bit is not
551 used) in the following functions [macros]:
554 BN_GF2m_sub [= BN_GF2m_add]
555 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
556 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
557 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
559 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
560 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
561 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
562 BN_GF2m_cmp [= BN_ucmp]
564 (Note that only the 'mod' functions are actually for fields GF(2^m).
565 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
567 For some functions, an the irreducible polynomial defining a
568 field can be given as an 'unsigned int[]' with strictly
569 decreasing elements giving the indices of those bits that are set;
570 i.e., p[] represents the polynomial
571 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
573 p[0] > p[1] > ... > p[k] = 0.
574 This applies to the following functions:
579 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
580 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
583 BN_GF2m_mod_solve_quad_arr
587 Conversion can be performed by the following functions:
592 bntest.c has additional tests for binary polynomial arithmetic.
594 Two implementations for BN_GF2m_mod_div() are available.
595 The default algorithm simply uses BN_GF2m_mod_inv() and
596 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
597 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
598 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
600 [Sheueling Chang Shantz and Douglas Stebila
601 (Sun Microsystems Laboratories)]
603 *) Add new error code 'ERR_R_DISABLED' that can be used when some
604 functionality is disabled at compile-time.
605 [Douglas Stebila <douglas.stebila@sun.com>]
607 *) Change default behaviour of 'openssl asn1parse' so that more
608 information is visible when viewing, e.g., a certificate:
610 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
611 mode the content of non-printable OCTET STRINGs is output in a
612 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
613 avoid the appearance of a printable string.
614 [Nils Larsch <nla@trustcenter.de>]
616 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
618 EC_GROUP_set_asn1_flag()
619 EC_GROUP_get_asn1_flag()
620 EC_GROUP_set_point_conversion_form()
621 EC_GROUP_get_point_conversion_form()
622 These control ASN1 encoding details:
623 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
624 has been set to OPENSSL_EC_NAMED_CURVE.
625 - Points are encoded in uncompressed form by default; options for
626 asn1_for are as for point2oct, namely
627 POINT_CONVERSION_COMPRESSED
628 POINT_CONVERSION_UNCOMPRESSED
629 POINT_CONVERSION_HYBRID
631 Also add 'seed' and 'seed_len' members to EC_GROUP with access
635 EC_GROUP_get_seed_len()
636 This is used only for ASN1 purposes (so far).
637 [Nils Larsch <nla@trustcenter.de>]
639 *) Add 'field_type' member to EC_METHOD, which holds the NID
640 of the appropriate field type OID. The new function
641 EC_METHOD_get_field_type() returns this value.
642 [Nils Larsch <nla@trustcenter.de>]
649 providing useful interfaces to EC_POINT_point2oct() and
650 EC_POINT_oct2point().
651 [Nils Larsch <nla@trustcenter.de>]
653 *) Change internals of the EC library so that the functions
654 EC_GROUP_set_generator()
655 EC_GROUP_get_generator()
657 EC_GROUP_get_cofactor()
658 are implemented directly in crypto/ec/ec_lib.c and not dispatched
659 to methods, which would lead to unnecessary code duplication when
660 adding different types of curves.
661 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
663 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
664 arithmetic, and such that modified wNAFs are generated
665 (which avoid length expansion in many cases).
668 *) Add a function EC_GROUP_check_discriminant() (defined via
669 EC_METHOD) that verifies that the curve discriminant is non-zero.
671 Add a function EC_GROUP_check() that makes some sanity tests
672 on a EC_GROUP, its generator and order. This includes
673 EC_GROUP_check_discriminant().
674 [Nils Larsch <nla@trustcenter.de>]
676 *) Add ECDSA in new directory crypto/ecdsa/.
678 Add applications 'openssl ecparam' and 'openssl ecdsa'
679 (these are based on 'openssl dsaparam' and 'openssl dsa').
681 ECDSA support is also included in various other files across the
682 library. Most notably,
683 - 'openssl req' now has a '-newkey ecdsa:file' option;
684 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
685 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
686 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
687 them suitable for ECDSA where domain parameters must be
688 extracted before the specific public key;
689 - ECDSA engine support has been added.
690 [Nils Larsch <nla@trustcenter.de>]
692 *) Include some named elliptic curves, and add OIDs from X9.62,
693 SECG, and WAP/WTLS. Each curve can be obtained from the new
695 EC_GROUP_new_by_nid(),
696 and the list of available named curves can be obtained with
697 EC_get_builtin_curves().
698 Also add a 'curve_name' member to EC_GROUP objects, which can be
702 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
704 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
705 was actually never needed) and in BN_mul(). The removal in BN_mul()
706 required a small change in bn_mul_part_recursive() and the addition
707 of the functions bn_cmp_part_words(), bn_sub_part_words() and
708 bn_add_part_words(), which do the same thing as bn_cmp_words(),
709 bn_sub_words() and bn_add_words() except they take arrays with
713 Changes between 0.9.7d and 0.9.7e [XX xxx XXXX]
715 *) Add Delta CRL to the extension code.
718 *) Various fixes to s3_pkt.c so alerts are sent properly.
719 [David Holmes <d.holmes@f5.com>]
721 *) Reduce the chances of duplicate issuer name and serial numbers (in
722 violation of RFC3280) using the OpenSSL certificate creation utilities.
723 This is done by creating a random 64 bit value for the initial serial
724 number when a serial number file is created or when a self signed
725 certificate is created using 'openssl req -x509'. The initial serial
726 number file is created using 'openssl x509 -next_serial' in CA.pl
727 rather than being initialized to 1.
730 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
732 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
733 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
734 [Joe Orton, Steve Henson]
736 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
738 [Joe Orton, Steve Henson]
740 *) Make it possible to have multiple active certificates with the same
741 subject in the CA index file. This is done only if the keyword
742 'unique_subject' is set to 'no' in the main CA section (default
743 if 'CA_default') of the configuration file. The value is saved
744 with the database itself in a separate index attribute file,
745 named like the index file with '.attr' appended to the name.
748 *) X509 verify fixes. Disable broken certificate workarounds when
749 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
750 keyUsage extension present. Don't accept CRLs with unhandled critical
751 extensions: since verify currently doesn't process CRL extensions this
752 rejects a CRL with *any* critical extensions. Add new verify error codes
756 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
757 A clarification of RFC2560 will require the use of OCTET STRINGs and
758 some implementations cannot handle the current raw format. Since OpenSSL
759 copies and compares OCSP nonces as opaque blobs without any attempt at
760 parsing them this should not create any compatibility issues.
763 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
764 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
765 this HMAC (and other) operations are several times slower than OpenSSL
769 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
770 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
772 *) Use the correct content when signing type "other".
775 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
777 *) Fix various bugs revealed by running the NISCC test suite:
779 Stop out of bounds reads in the ASN1 code when presented with
780 invalid tags (CAN-2003-0543 and CAN-2003-0544).
782 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
784 If verify callback ignores invalid public key errors don't try to check
785 certificate signature with the NULL public key.
789 *) New -ignore_err option in ocsp application to stop the server
790 exiting on the first error in a request.
793 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
794 if the server requested one: as stated in TLS 1.0 and SSL 3.0
798 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
799 extra data after the compression methods not only for TLS 1.0
800 but also for SSL 3.0 (as required by the specification).
801 [Bodo Moeller; problem pointed out by Matthias Loepfe]
803 *) Change X509_certificate_type() to mark the key as exported/exportable
804 when it's 512 *bits* long, not 512 bytes.
807 *) Change AES_cbc_encrypt() so it outputs exact multiple of
808 blocks during encryption.
811 *) Various fixes to base64 BIO and non blocking I/O. On write
812 flushes were not handled properly if the BIO retried. On read
813 data was not being buffered properly and had various logic bugs.
814 This also affects blocking I/O when the data being decoded is a
818 *) Various S/MIME bugfixes and compatibility changes:
819 output correct application/pkcs7 MIME type if
820 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
821 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
822 of files as .eml work). Correctly handle very long lines in MIME
826 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
828 *) Countermeasure against the Klima-Pokorny-Rosa extension of
829 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
830 a protocol version number mismatch like a decryption error
831 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
834 *) Turn on RSA blinding by default in the default implementation
835 to avoid a timing attack. Applications that don't want it can call
836 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
837 They would be ill-advised to do so in most cases.
838 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
840 *) Change RSA blinding code so that it works when the PRNG is not
841 seeded (in this case, the secret RSA exponent is abused as
842 an unpredictable seed -- if it is not unpredictable, there
843 is no point in blinding anyway). Make RSA blinding thread-safe
844 by remembering the creator's thread ID in rsa->blinding and
845 having all other threads use local one-time blinding factors
846 (this requires more computation than sharing rsa->blinding, but
847 avoids excessive locking; and if an RSA object is not shared
848 between threads, blinding will still be very fast).
851 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
852 ENGINE as defaults for all supported algorithms irrespective of
853 the 'flags' parameter. 'flags' is now honoured, so applications
854 should make sure they are passing it correctly.
857 *) Target "mingw" now allows native Windows code to be generated in
858 the Cygwin environment as well as with the MinGW compiler.
861 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
863 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
864 via timing by performing a MAC computation even if incorrrect
865 block cipher padding has been found. This is a countermeasure
866 against active attacks where the attacker has to distinguish
867 between bad padding and a MAC verification error. (CAN-2003-0078)
869 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
870 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
871 Martin Vuagnoux (EPFL, Ilion)]
873 *) Make the no-err option work as intended. The intention with no-err
874 is not to have the whole error stack handling routines removed from
875 libcrypto, it's only intended to remove all the function name and
876 reason texts, thereby removing some of the footprint that may not
877 be interesting if those errors aren't displayed anyway.
879 NOTE: it's still possible for any application or module to have it's
880 own set of error texts inserted. The routines are there, just not
881 used by default when no-err is given.
884 *) Add support for FreeBSD on IA64.
885 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
887 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
888 Kerberos function mit_des_cbc_cksum(). Before this change,
889 the value returned by DES_cbc_cksum() was like the one from
890 mit_des_cbc_cksum(), except the bytes were swapped.
891 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
893 *) Allow an application to disable the automatic SSL chain building.
894 Before this a rather primitive chain build was always performed in
895 ssl3_output_cert_chain(): an application had no way to send the
896 correct chain if the automatic operation produced an incorrect result.
898 Now the chain builder is disabled if either:
900 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
902 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
904 The reasoning behind this is that an application would not want the
905 auto chain building to take place if extra chain certificates are
906 present and it might also want a means of sending no additional
907 certificates (for example the chain has two certificates and the
911 *) Add the possibility to build without the ENGINE framework.
912 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
914 *) Under Win32 gmtime() can return NULL: check return value in
915 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
918 *) DSA routines: under certain error conditions uninitialized BN objects
919 could be freed. Solution: make sure initialization is performed early
920 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
921 Nils Larsch <nla@trustcenter.de> via PR#459)
924 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
925 checked on reconnect on the client side, therefore session resumption
926 could still fail with a "ssl session id is different" error. This
927 behaviour is masked when SSL_OP_ALL is used due to
928 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
929 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
933 *) IA-32 assembler support enhancements: unified ELF targets, support
934 for SCO/Caldera platforms, fix for Cygwin shared build.
937 *) Add support for FreeBSD on sparc64. As a consequence, support for
938 FreeBSD on non-x86 processors is separate from x86 processors on
939 the config script, much like the NetBSD support.
940 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
942 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
944 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
947 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
948 code (06) was taken as the first octet of the session ID and the last
949 octet was ignored consequently. As a result SSLv2 client side session
950 caching could not have worked due to the session ID mismatch between
952 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
956 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
957 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
961 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
962 seems that in spite of existing for more than a year, many application
963 author have done nothing to provide the necessary callbacks, which
964 means that this particular engine will not work properly anywhere.
965 This is a very unfortunate situation which forces us, in the name
966 of usability, to give the hw_ncipher.c a static lock, which is part
968 NOTE: This is for the 0.9.7 series ONLY. This hack will never
969 appear in 0.9.8 or later. We EXPECT application authors to have
970 dealt properly with this when 0.9.8 is released (unless we actually
971 make such changes in the libcrypto locking code that changes will
972 have to be made anyway).
975 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
976 octets have been read, EOF or an error occurs. Without this change
977 some truncated ASN1 structures will not produce an error.
980 *) Disable Heimdal support, since it hasn't been fully implemented.
981 Still give the possibility to force the use of Heimdal, but with
982 warnings and a request that patches get sent to openssl-dev.
985 *) Add the VC-CE target, introduce the WINCE sysname, and add
986 INSTALL.WCE and appropriate conditionals to make it build.
987 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
989 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
990 cygssl-x.y.z.dll, where x, y and z are the major, minor and
991 edit numbers of the version.
992 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
994 *) Introduce safe string copy and catenation functions
995 (BUF_strlcpy() and BUF_strlcat()).
996 [Ben Laurie (CHATS) and Richard Levitte]
998 *) Avoid using fixed-size buffers for one-line DNs.
1001 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1002 resizing buffers containing secrets, and use where appropriate.
1003 [Ben Laurie (CHATS)]
1005 *) Avoid using fixed size buffers for configuration file location.
1006 [Ben Laurie (CHATS)]
1008 *) Avoid filename truncation for various CA files.
1009 [Ben Laurie (CHATS)]
1011 *) Use sizeof in preference to magic numbers.
1012 [Ben Laurie (CHATS)]
1014 *) Avoid filename truncation in cert requests.
1015 [Ben Laurie (CHATS)]
1017 *) Add assertions to check for (supposedly impossible) buffer
1019 [Ben Laurie (CHATS)]
1021 *) Don't cache truncated DNS entries in the local cache (this could
1022 potentially lead to a spoofing attack).
1023 [Ben Laurie (CHATS)]
1025 *) Fix various buffers to be large enough for hex/decimal
1026 representations in a platform independent manner.
1027 [Ben Laurie (CHATS)]
1029 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1030 resizing buffers containing secrets, and use where appropriate.
1031 [Ben Laurie (CHATS)]
1033 *) Add BIO_indent() to avoid much slightly worrying code to do
1035 [Ben Laurie (CHATS)]
1037 *) Convert sprintf()/BIO_puts() to BIO_printf().
1038 [Ben Laurie (CHATS)]
1040 *) buffer_gets() could terminate with the buffer only half
1042 [Ben Laurie (CHATS)]
1044 *) Add assertions to prevent user-supplied crypto functions from
1045 overflowing internal buffers by having large block sizes, etc.
1046 [Ben Laurie (CHATS)]
1048 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1050 [Ben Laurie (CHATS)]
1052 *) Eliminate unused copy of key in RC4.
1053 [Ben Laurie (CHATS)]
1055 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1056 [Ben Laurie (CHATS)]
1058 *) Fix off-by-one error in EGD path.
1059 [Ben Laurie (CHATS)]
1061 *) If RANDFILE path is too long, ignore instead of truncating.
1062 [Ben Laurie (CHATS)]
1064 *) Eliminate unused and incorrectly sized X.509 structure
1066 [Ben Laurie (CHATS)]
1068 *) Eliminate unused and dangerous function knumber().
1069 [Ben Laurie (CHATS)]
1071 *) Eliminate unused and dangerous structure, KSSL_ERR.
1072 [Ben Laurie (CHATS)]
1074 *) Protect against overlong session ID context length in an encoded
1075 session object. Since these are local, this does not appear to be
1077 [Ben Laurie (CHATS)]
1079 *) Change from security patch (see 0.9.6e below) that did not affect
1080 the 0.9.6 release series:
1082 Remote buffer overflow in SSL3 protocol - an attacker could
1083 supply an oversized master key in Kerberos-enabled versions.
1085 [Ben Laurie (CHATS)]
1087 *) Change the SSL kerb5 codes to match RFC 2712.
1090 *) Make -nameopt work fully for req and add -reqopt switch.
1091 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1093 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1094 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1096 *) Make sure tests can be performed even if the corresponding algorithms
1097 have been removed entirely. This was also the last step to make
1098 OpenSSL compilable with DJGPP under all reasonable conditions.
1099 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1101 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1102 to allow version independent disabling of normally unselected ciphers,
1103 which may be activated as a side-effect of selecting a single cipher.
1105 (E.g., cipher list string "RSA" enables ciphersuites that are left
1106 out of "ALL" because they do not provide symmetric encryption.
1107 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1108 [Lutz Jaenicke, Bodo Moeller]
1110 *) Add appropriate support for separate platform-dependent build
1111 directories. The recommended way to make a platform-dependent
1112 build directory is the following (tested on Linux), maybe with
1115 # Place yourself outside of the OpenSSL source tree. In
1116 # this example, the environment variable OPENSSL_SOURCE
1117 # is assumed to contain the absolute OpenSSL source directory.
1118 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1119 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1120 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1121 mkdir -p `dirname $F`
1122 ln -s $OPENSSL_SOURCE/$F $F
1125 To be absolutely sure not to disturb the source tree, a "make clean"
1126 is a good thing. If it isn't successfull, don't worry about it,
1127 it probably means the source directory is very clean.
1130 *) Make sure any ENGINE control commands make local copies of string
1131 pointers passed to them whenever necessary. Otherwise it is possible
1132 the caller may have overwritten (or deallocated) the original string
1133 data when a later ENGINE operation tries to use the stored values.
1134 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1136 *) Improve diagnostics in file reading and command-line digests.
1137 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1139 *) Add AES modes CFB and OFB to the object database. Correct an
1140 error in AES-CFB decryption.
1143 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1144 allows existing EVP_CIPHER_CTX structures to be reused after
1145 calling EVP_*Final(). This behaviour is used by encryption
1146 BIOs and some applications. This has the side effect that
1147 applications must explicitly clean up cipher contexts with
1148 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1151 *) Check the values of dna and dnb in bn_mul_recursive before calling
1152 bn_mul_comba (a non zero value means the a or b arrays do not contain
1153 n2 elements) and fallback to bn_mul_normal if either is not zero.
1156 *) Fix escaping of non-ASCII characters when using the -subj option
1157 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1160 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1161 form for "surname", serialNumber has no short form.
1162 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1163 therefore remove "mail" short name for "internet 7".
1164 The OID for unique identifiers in X509 certificates is
1165 x500UniqueIdentifier, not uniqueIdentifier.
1166 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1169 *) Add an "init" command to the ENGINE config module and auto initialize
1170 ENGINEs. Without any "init" command the ENGINE will be initialized
1171 after all ctrl commands have been executed on it. If init=1 the
1172 ENGINE is initailized at that point (ctrls before that point are run
1173 on the uninitialized ENGINE and after on the initialized one). If
1174 init=0 then the ENGINE will not be iniatialized at all.
1177 *) Fix the 'app_verify_callback' interface so that the user-defined
1178 argument is actually passed to the callback: In the
1179 SSL_CTX_set_cert_verify_callback() prototype, the callback
1180 declaration has been changed from
1183 int (*cb)(X509_STORE_CTX *,void *);
1184 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1185 i=s->ctx->app_verify_callback(&ctx)
1186 has been changed into
1187 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1189 To update applications using SSL_CTX_set_cert_verify_callback(),
1190 a dummy argument can be added to their callback functions.
1191 [D. K. Smetters <smetters@parc.xerox.com>]
1193 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1194 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1196 *) Add and OPENSSL_LOAD_CONF define which will cause
1197 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1198 This allows older applications to transparently support certain
1199 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1200 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1201 load the config file and OPENSSL_add_all_algorithms_conf() which will
1202 always load it have also been added.
1205 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1206 Adjust NIDs and EVP layer.
1207 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1209 *) Config modules support in openssl utility.
1211 Most commands now load modules from the config file,
1212 though in a few (such as version) this isn't done
1213 because it couldn't be used for anything.
1215 In the case of ca and req the config file used is
1216 the same as the utility itself: that is the -config
1217 command line option can be used to specify an
1221 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1222 use "openssl_conf" if filename is NULL use default openssl config file.
1225 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1226 config section name. Add a new flag to tolerate a missing config file
1227 and move code to CONF_modules_load_file().
1230 *) Support for crypto accelerator cards from Accelerated Encryption
1231 Processing, www.aep.ie. (Use engine 'aep')
1232 The support was copied from 0.9.6c [engine] and adapted/corrected
1233 to work with the new engine framework.
1234 [AEP Inc. and Richard Levitte]
1236 *) Support for SureWare crypto accelerator cards from Baltimore
1237 Technologies. (Use engine 'sureware')
1238 The support was copied from 0.9.6c [engine] and adapted
1239 to work with the new engine framework.
1242 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1243 make the newer ENGINE framework commands for the CHIL engine work.
1244 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1246 *) Make it possible to produce shared libraries on ReliantUNIX.
1247 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1249 *) Add the configuration target debug-linux-ppro.
1250 Make 'openssl rsa' use the general key loading routines
1251 implemented in apps.c, and make those routines able to
1252 handle the key format FORMAT_NETSCAPE and the variant
1254 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1256 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1257 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1259 *) Add -keyform to rsautl, and document -engine.
1260 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1262 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1263 BIO_R_NO_SUCH_FILE error code rather than the generic
1264 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1267 *) Add new functions
1269 ERR_peek_last_error_line
1270 ERR_peek_last_error_line_data.
1271 These are similar to
1274 ERR_peek_error_line_data,
1275 but report on the latest error recorded rather than the first one
1276 still in the error queue.
1277 [Ben Laurie, Bodo Moeller]
1279 *) default_algorithms option in ENGINE config module. This allows things
1281 default_algorithms = ALL
1282 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1285 *) Prelminary ENGINE config module.
1288 *) New experimental application configuration code.
1291 *) Change the AES code to follow the same name structure as all other
1292 symmetric ciphers, and behave the same way. Move everything to
1293 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1294 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1296 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1297 [Ben Laurie and Theo de Raadt]
1299 *) Add option to output public keys in req command.
1300 [Massimiliano Pala madwolf@openca.org]
1302 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1303 (up to about 10% better than before for P-192 and P-224).
1306 *) New functions/macros
1308 SSL_CTX_set_msg_callback(ctx, cb)
1309 SSL_CTX_set_msg_callback_arg(ctx, arg)
1310 SSL_set_msg_callback(ssl, cb)
1311 SSL_set_msg_callback_arg(ssl, arg)
1313 to request calling a callback function
1315 void cb(int write_p, int version, int content_type,
1316 const void *buf, size_t len, SSL *ssl, void *arg)
1318 whenever a protocol message has been completely received
1319 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1320 protocol version according to which the SSL library interprets
1321 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1322 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1323 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1324 specification (change_cipher_spec(20), alert(21), handshake(22)).
1325 'buf' and 'len' point to the actual message, 'ssl' to the
1326 SSL object, and 'arg' is the application-defined value set by
1327 SSL[_CTX]_set_msg_callback_arg().
1329 'openssl s_client' and 'openssl s_server' have new '-msg' options
1330 to enable a callback that displays all protocol messages.
1333 *) Change the shared library support so shared libraries are built as
1334 soon as the corresponding static library is finished, and thereby get
1335 openssl and the test programs linked against the shared library.
1336 This still only happens when the keyword "shard" has been given to
1337 the configuration scripts.
1339 NOTE: shared library support is still an experimental thing, and
1340 backward binary compatibility is still not guaranteed.
1341 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1343 *) Add support for Subject Information Access extension.
1344 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1346 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1347 additional bytes when new memory had to be allocated, not just
1348 when reusing an existing buffer.
1351 *) New command line and configuration option 'utf8' for the req command.
1352 This allows field values to be specified as UTF8 strings.
1355 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1356 runs for the former and machine-readable output for the latter.
1359 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1360 of the e-mail address in the DN (i.e., it will go into a certificate
1361 extension only). The new configuration file option 'email_in_dn = no'
1362 has the same effect.
1363 [Massimiliano Pala madwolf@openca.org]
1365 *) Change all functions with names starting with des_ to be starting
1366 with DES_ instead. Add wrappers that are compatible with libdes,
1367 but are named _ossl_old_des_*. Finally, add macros that map the
1368 des_* symbols to the corresponding _ossl_old_des_* if libdes
1369 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1370 desired, the des_* symbols will be mapped to DES_*, with one
1373 Since we provide two compatibility mappings, the user needs to
1374 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1375 compatibility is desired. The default (i.e., when that macro
1376 isn't defined) is OpenSSL 0.9.6c compatibility.
1378 There are also macros that enable and disable the support of old
1379 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1380 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1381 are defined, the default will apply: to support the old des routines.
1383 In either case, one must include openssl/des.h to get the correct
1384 definitions. Do not try to just include openssl/des_old.h, that
1387 NOTE: This is a major break of an old API into a new one. Software
1388 authors are encouraged to switch to the DES_ style functions. Some
1389 time in the future, des_old.h and the libdes compatibility functions
1390 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1391 default), and then completely removed.
1394 *) Test for certificates which contain unsupported critical extensions.
1395 If such a certificate is found during a verify operation it is
1396 rejected by default: this behaviour can be overridden by either
1397 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1398 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1399 X509_supported_extension() has also been added which returns 1 if a
1400 particular extension is supported.
1403 *) Modify the behaviour of EVP cipher functions in similar way to digests
1404 to retain compatibility with existing code.
1407 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1408 compatibility with existing code. In particular the 'ctx' parameter does
1409 not have to be to be initialized before the call to EVP_DigestInit() and
1410 it is tidied up after a call to EVP_DigestFinal(). New function
1411 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1412 EVP_MD_CTX_copy() changed to not require the destination to be
1413 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1414 requires the destination to be valid.
1416 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1417 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1420 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1421 so that complete 'Handshake' protocol structures are kept in memory
1422 instead of overwriting 'msg_type' and 'length' with 'body' data.
1425 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1426 [Massimo Santin via Richard Levitte]
1428 *) Major restructuring to the underlying ENGINE code. This includes
1429 reduction of linker bloat, separation of pure "ENGINE" manipulation
1430 (initialisation, etc) from functionality dealing with implementations
1431 of specific crypto iterfaces. This change also introduces integrated
1432 support for symmetric ciphers and digest implementations - so ENGINEs
1433 can now accelerate these by providing EVP_CIPHER and EVP_MD
1434 implementations of their own. This is detailed in crypto/engine/README
1435 as it couldn't be adequately described here. However, there are a few
1436 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1437 were changed in the original introduction of ENGINE code have now
1438 reverted back - the hooking from this code to ENGINE is now a good
1439 deal more passive and at run-time, operations deal directly with
1440 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1441 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1442 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1443 they were not being used by the framework as there is no concept of a
1444 BIGNUM_METHOD and they could not be generalised to the new
1445 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1446 ENGINE_cpy() has been removed as it cannot be consistently defined in
1450 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1453 *) Change mkdef.pl to sort symbols that get the same entry number,
1454 and make sure the automatically generated functions ERR_load_*
1455 become part of libeay.num as well.
1458 *) New function SSL_renegotiate_pending(). This returns true once
1459 renegotiation has been requested (either SSL_renegotiate() call
1460 or HelloRequest/ClientHello receveived from the peer) and becomes
1461 false once a handshake has been completed.
1462 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1463 sends a HelloRequest, but does not ensure that a handshake takes
1464 place. SSL_renegotiate_pending() is useful for checking if the
1465 client has followed the request.)
1468 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1469 By default, clients may request session resumption even during
1470 renegotiation (if session ID contexts permit); with this option,
1471 session resumption is possible only in the first handshake.
1473 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1474 more bits available for options that should not be part of
1475 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1478 *) Add some demos for certificate and certificate request creation.
1481 *) Make maximum certificate chain size accepted from the peer application
1482 settable (SSL*_get/set_max_cert_list()), as proposed by
1483 "Douglas E. Engert" <deengert@anl.gov>.
1486 *) Add support for shared libraries for Unixware-7
1487 (Boyd Lynn Gerber <gerberb@zenez.com>).
1490 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1491 be done prior to destruction. Use this to unload error strings from
1492 ENGINEs that load their own error strings. NB: This adds two new API
1493 functions to "get" and "set" this destroy handler in an ENGINE.
1496 *) Alter all existing ENGINE implementations (except "openssl" and
1497 "openbsd") to dynamically instantiate their own error strings. This
1498 makes them more flexible to be built both as statically-linked ENGINEs
1499 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1500 Also, add stub code to each that makes building them as self-contained
1501 shared-libraries easier (see README.ENGINE).
1504 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1505 implementations into applications that are completely implemented in
1506 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1507 commands that can be used to configure what shared-library to load and
1508 to control aspects of the way it is handled. Also, made an update to
1509 the README.ENGINE file that brings its information up-to-date and
1510 provides some information and instructions on the "dynamic" ENGINE
1511 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1514 *) Make it possible to unload ranges of ERR strings with a new
1515 "ERR_unload_strings" function.
1518 *) Add a copy() function to EVP_MD.
1521 *) Make EVP_MD routines take a context pointer instead of just the
1522 md_data void pointer.
1525 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1526 that the digest can only process a single chunk of data
1527 (typically because it is provided by a piece of
1528 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1529 is only going to provide a single chunk of data, and hence the
1530 framework needn't accumulate the data for oneshot drivers.
1533 *) As with "ERR", make it possible to replace the underlying "ex_data"
1534 functions. This change also alters the storage and management of global
1535 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1536 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1537 index counters. The API functions that use this state have been changed
1538 to take a "class_index" rather than pointers to the class's local STACK
1539 and counter, and there is now an API function to dynamically create new
1540 classes. This centralisation allows us to (a) plug a lot of the
1541 thread-safety problems that existed, and (b) makes it possible to clean
1542 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1543 such data would previously have always leaked in application code and
1544 workarounds were in place to make the memory debugging turn a blind eye
1545 to it. Application code that doesn't use this new function will still
1546 leak as before, but their memory debugging output will announce it now
1547 rather than letting it slide.
1549 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1550 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1551 has a return value to indicate success or failure.
1554 *) Make it possible to replace the underlying "ERR" functions such that the
1555 global state (2 LHASH tables and 2 locks) is only used by the "default"
1556 implementation. This change also adds two functions to "get" and "set"
1557 the implementation prior to it being automatically set the first time
1558 any other ERR function takes place. Ie. an application can call "get",
1559 pass the return value to a module it has just loaded, and that module
1560 can call its own "set" function using that value. This means the
1561 module's "ERR" operations will use (and modify) the error state in the
1562 application and not in its own statically linked copy of OpenSSL code.
1565 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1566 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1567 the operation, and provides a more encapsulated way for external code
1568 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1569 to use these functions rather than manually incrementing the counts.
1571 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1574 *) Add EVP test program.
1577 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1580 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1581 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1582 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1583 These allow a CRL to be built without having to access X509_CRL fields
1584 directly. Modify 'ca' application to use new functions.
1587 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1588 bug workarounds. Rollback attack detection is a security feature.
1589 The problem will only arise on OpenSSL servers when TLSv1 is not
1590 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1591 Software authors not wanting to support TLSv1 will have special reasons
1592 for their choice and can explicitly enable this option.
1593 [Bodo Moeller, Lutz Jaenicke]
1595 *) Rationalise EVP so it can be extended: don't include a union of
1596 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1597 (similar to those existing for EVP_CIPHER_CTX).
1602 EVP_MD_CTX_init(&md); /* new function call */
1603 EVP_DigestInit(&md, EVP_sha1());
1604 EVP_DigestUpdate(&md, in, len);
1605 EVP_DigestFinal(&md, out, NULL);
1606 EVP_MD_CTX_cleanup(&md); /* new function call */
1610 *) Make DES key schedule conform to the usual scheme, as well as
1611 correcting its structure. This means that calls to DES functions
1612 now have to pass a pointer to a des_key_schedule instead of a
1613 plain des_key_schedule (which was actually always a pointer
1616 des_key_schedule ks;
1618 des_set_key_checked(..., &ks);
1619 des_ncbc_encrypt(..., &ks, ...);
1621 (Note that a later change renames 'des_...' into 'DES_...'.)
1624 *) Initial reduction of linker bloat: the use of some functions, such as
1625 PEM causes large amounts of unused functions to be linked in due to
1626 poor organisation. For example pem_all.c contains every PEM function
1627 which has a knock on effect of linking in large amounts of (unused)
1628 ASN1 code. Grouping together similar functions and splitting unrelated
1629 functions prevents this.
1632 *) Cleanup of EVP macros.
1635 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1636 correct _ecb suffix.
1639 *) Add initial OCSP responder support to ocsp application. The
1640 revocation information is handled using the text based index
1641 use by the ca application. The responder can either handle
1642 requests generated internally, supplied in files (for example
1643 via a CGI script) or using an internal minimal server.
1646 *) Add configuration choices to get zlib compression for TLS.
1649 *) Changes to Kerberos SSL for RFC 2712 compliance:
1650 1. Implemented real KerberosWrapper, instead of just using
1651 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1652 2. Implemented optional authenticator field of KerberosWrapper.
1654 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1655 and authenticator structs; see crypto/krb5/.
1657 Generalized Kerberos calls to support multiple Kerberos libraries.
1658 [Vern Staats <staatsvr@asc.hpc.mil>,
1659 Jeffrey Altman <jaltman@columbia.edu>
1660 via Richard Levitte]
1662 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1663 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1664 values for each of the key sizes rather than having just
1665 parameters (and 'speed' generating keys each time).
1668 *) Speed up EVP routines.
1671 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1672 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1673 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1674 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1676 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1677 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1678 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1681 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1683 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1686 *) Added the OS2-EMX target.
1687 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1689 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1690 to support NCONF routines in extension code. New function CONF_set_nconf()
1691 to allow functions which take an NCONF to also handle the old LHASH
1692 structure: this means that the old CONF compatible routines can be
1693 retained (in particular wrt extensions) without having to duplicate the
1694 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1697 *) Enhance the general user interface with mechanisms for inner control
1698 and with possibilities to have yes/no kind of prompts.
1701 *) Change all calls to low level digest routines in the library and
1702 applications to use EVP. Add missing calls to HMAC_cleanup() and
1703 don't assume HMAC_CTX can be copied using memcpy().
1704 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1706 *) Add the possibility to control engines through control names but with
1707 arbitrary arguments instead of just a string.
1708 Change the key loaders to take a UI_METHOD instead of a callback
1709 function pointer. NOTE: this breaks binary compatibility with earlier
1710 versions of OpenSSL [engine].
1711 Adapt the nCipher code for these new conditions and add a card insertion
1715 *) Enhance the general user interface with mechanisms to better support
1716 dialog box interfaces, application-defined prompts, the possibility
1717 to use defaults (for example default passwords from somewhere else)
1718 and interrupts/cancellations.
1721 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1722 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1725 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1726 tidy up some unnecessarily weird code in 'sk_new()').
1727 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1729 *) Change the key loading routines for ENGINEs to use the same kind
1730 callback (pem_password_cb) as all other routines that need this
1734 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1735 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1736 than this minimum value is recommended.
1739 *) New random seeder for OpenVMS, using the system process statistics
1740 that are easily reachable.
1743 *) Windows apparently can't transparently handle global
1744 variables defined in DLLs. Initialisations such as:
1746 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1748 wont compile. This is used by the any applications that need to
1749 declare their own ASN1 modules. This was fixed by adding the option
1750 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1751 needed for static libraries under Win32.
1754 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1755 setting of purpose and trust fields. New X509_STORE trust and
1756 purpose functions and tidy up setting in other SSL functions.
1759 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1760 structure. These are inherited by X509_STORE_CTX when it is
1761 initialised. This allows various defaults to be set in the
1762 X509_STORE structure (such as flags for CRL checking and custom
1763 purpose or trust settings) for functions which only use X509_STORE_CTX
1764 internally such as S/MIME.
1766 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1767 trust settings if they are not set in X509_STORE. This allows X509_STORE
1768 purposes and trust (in S/MIME for example) to override any set by default.
1770 Add command line options for CRL checking to smime, s_client and s_server
1774 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1775 are set then the CRL is looked up in the X509_STORE structure and
1776 its validity and signature checked, then if the certificate is found
1777 in the CRL the verify fails with a revoked error.
1779 Various new CRL related callbacks added to X509_STORE_CTX structure.
1781 Command line options added to 'verify' application to support this.
1783 This needs some additional work, such as being able to handle multiple
1784 CRLs with different times, extension based lookup (rather than just
1785 by subject name) and ultimately more complete V2 CRL extension
1789 *) Add a general user interface API (crypto/ui/). This is designed
1790 to replace things like des_read_password and friends (backward
1791 compatibility functions using this new API are provided).
1792 The purpose is to remove prompting functions from the DES code
1793 section as well as provide for prompting through dialog boxes in
1794 a window system and the like.
1797 *) Add "ex_data" support to ENGINE so implementations can add state at a
1798 per-structure level rather than having to store it globally.
1801 *) Make it possible for ENGINE structures to be copied when retrieved by
1802 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1803 This causes the "original" ENGINE structure to act like a template,
1804 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1805 operational state can be localised to each ENGINE structure, despite the
1806 fact they all share the same "methods". New ENGINE structures returned in
1807 this case have no functional references and the return value is the single
1808 structural reference. This matches the single structural reference returned
1809 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1813 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1814 needs to match any other type at all we need to manually clear the
1818 *) Changes to the "openssl engine" utility to include;
1819 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1820 about an ENGINE's available control commands.
1821 - executing control commands from command line arguments using the
1822 '-pre' and '-post' switches. '-post' is only used if '-t' is
1823 specified and the ENGINE is successfully initialised. The syntax for
1824 the individual commands are colon-separated, for example;
1825 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1828 *) New dynamic control command support for ENGINEs. ENGINEs can now
1829 declare their own commands (numbers), names (strings), descriptions,
1830 and input types for run-time discovery by calling applications. A
1831 subset of these commands are implicitly classed as "executable"
1832 depending on their input type, and only these can be invoked through
1833 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1834 can be based on user input, config files, etc). The distinction is
1835 that "executable" commands cannot return anything other than a boolean
1836 result and can only support numeric or string input, whereas some
1837 discoverable commands may only be for direct use through
1838 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1839 pointers, or other custom uses. The "executable" commands are to
1840 support parameterisations of ENGINE behaviour that can be
1841 unambiguously defined by ENGINEs and used consistently across any
1842 OpenSSL-based application. Commands have been added to all the
1843 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1844 control over shared-library paths without source code alterations.
1847 *) Changed all ENGINE implementations to dynamically allocate their
1848 ENGINEs rather than declaring them statically. Apart from this being
1849 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1850 this also allows the implementations to compile without using the
1851 internal engine_int.h header.
1854 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1855 'const' value. Any code that should be able to modify a RAND_METHOD
1856 should already have non-const pointers to it (ie. they should only
1857 modify their own ones).
1860 *) Made a variety of little tweaks to the ENGINE code.
1861 - "atalla" and "ubsec" string definitions were moved from header files
1862 to C code. "nuron" string definitions were placed in variables
1863 rather than hard-coded - allowing parameterisation of these values
1864 later on via ctrl() commands.
1865 - Removed unused "#if 0"'d code.
1866 - Fixed engine list iteration code so it uses ENGINE_free() to release
1867 structural references.
1868 - Constified the RAND_METHOD element of ENGINE structures.
1869 - Constified various get/set functions as appropriate and added
1870 missing functions (including a catch-all ENGINE_cpy that duplicates
1871 all ENGINE values onto a new ENGINE except reference counts/state).
1872 - Removed NULL parameter checks in get/set functions. Setting a method
1873 or function to NULL is a way of cancelling out a previously set
1874 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1875 and doesn't justify the extra error symbols and code.
1876 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1877 flags from engine_int.h to engine.h.
1878 - Changed prototypes for ENGINE handler functions (init(), finish(),
1879 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1882 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1883 to the algorithm using long division. The binary algorithm can be
1884 used only if the modulus is odd. On 32-bit systems, it is faster
1885 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1886 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1887 up to 450 bits. In 64-bit environments, the binary algorithm
1888 appears to be advantageous for much longer moduli; here we use it
1889 for moduli up to 2048 bits.
1892 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1893 could not support the combine flag in choice fields.
1896 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1897 extensions from a certificate request to the certificate.
1900 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1901 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1902 file: this allows the display of the certificate about to be
1903 signed to be customised, to allow certain fields to be included
1904 or excluded and extension details. The old system didn't display
1905 multicharacter strings properly, omitted fields not in the policy
1906 and couldn't display additional details such as extensions.
1909 *) Function EC_POINTs_mul for multiple scalar multiplication
1910 of an arbitrary number of elliptic curve points
1911 \sum scalars[i]*points[i],
1912 optionally including the generator defined for the EC_GROUP:
1913 scalar*generator + \sum scalars[i]*points[i].
1915 EC_POINT_mul is a simple wrapper function for the typical case
1916 that the point list has just one item (besides the optional
1920 *) First EC_METHODs for curves over GF(p):
1922 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1923 operations and provides various method functions that can also
1924 operate with faster implementations of modular arithmetic.
1926 EC_GFp_mont_method() reuses most functions that are part of
1927 EC_GFp_simple_method, but uses Montgomery arithmetic.
1929 [Bodo Moeller; point addition and point doubling
1930 implementation directly derived from source code provided by
1931 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1933 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1934 crypto/ec/ec_lib.c):
1936 Curves are EC_GROUP objects (with an optional group generator)
1937 based on EC_METHODs that are built into the library.
1939 Points are EC_POINT objects based on EC_GROUP objects.
1941 Most of the framework would be able to handle curves over arbitrary
1942 finite fields, but as there are no obvious types for fields other
1943 than GF(p), some functions are limited to that for now.
1946 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1947 that the file contains a complete HTTP response.
1950 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1951 change the def and num file printf format specifier from "%-40sXXX"
1952 to "%-39s XXX". The latter will always guarantee a space after the
1953 field while the former will cause them to run together if the field
1954 is 40 of more characters long.
1957 *) Constify the cipher and digest 'method' functions and structures
1958 and modify related functions to take constant EVP_MD and EVP_CIPHER
1962 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1963 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1966 *) Modify EVP_Digest*() routines so they now return values. Although the
1967 internal software routines can never fail additional hardware versions
1971 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1973 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1974 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1977 ERR_R_NESTED_ASN1_ERROR
1979 ERR_R_MISSING_ASN1_EOS
1980 were 4 .. 9, conflicting with
1981 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1983 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1984 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1986 Add new error code 'ERR_R_INTERNAL_ERROR'.
1989 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1993 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1994 sets the subject name for a new request or supersedes the
1995 subject name in a given request. Formats that can be parsed are
1996 'CN=Some Name, OU=myOU, C=IT'
1998 'CN=Some Name/OU=myOU/C=IT'.
2000 Add options '-batch' and '-verbose' to 'openssl req'.
2001 [Massimiliano Pala <madwolf@hackmasters.net>]
2003 *) Introduce the possibility to access global variables through
2004 functions on platform were that's the best way to handle exporting
2005 global variables in shared libraries. To enable this functionality,
2006 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2007 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2008 is normally done by Configure or something similar).
2010 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2011 in the source file (foo.c) like this:
2013 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2014 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2016 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2017 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2019 OPENSSL_DECLARE_GLOBAL(int,foo);
2020 #define foo OPENSSL_GLOBAL_REF(foo)
2021 OPENSSL_DECLARE_GLOBAL(double,bar);
2022 #define bar OPENSSL_GLOBAL_REF(bar)
2024 The #defines are very important, and therefore so is including the
2025 header file everywhere where the defined globals are used.
2027 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2028 of ASN.1 items, but that structure is a bit different.
2030 The largest change is in util/mkdef.pl which has been enhanced with
2031 better and easier to understand logic to choose which symbols should
2032 go into the Windows .def files as well as a number of fixes and code
2033 cleanup (among others, algorithm keywords are now sorted
2034 lexicographically to avoid constant rewrites).
2037 *) In BN_div() keep a copy of the sign of 'num' before writing the
2038 result to 'rm' because if rm==num the value will be overwritten
2039 and produce the wrong result if 'num' is negative: this caused
2040 problems with BN_mod() and BN_nnmod().
2043 *) Function OCSP_request_verify(). This checks the signature on an
2044 OCSP request and verifies the signer certificate. The signer
2045 certificate is just checked for a generic purpose and OCSP request
2049 *) Add OCSP_check_validity() function to check the validity of OCSP
2050 responses. OCSP responses are prepared in real time and may only
2051 be a few seconds old. Simply checking that the current time lies
2052 between thisUpdate and nextUpdate max reject otherwise valid responses
2053 caused by either OCSP responder or client clock inaccuracy. Instead
2054 we allow thisUpdate and nextUpdate to fall within a certain period of
2055 the current time. The age of the response can also optionally be
2056 checked. Two new options -validity_period and -status_age added to
2060 *) If signature or public key algorithm is unrecognized print out its
2061 OID rather that just UNKNOWN.
2064 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2065 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2066 ID to be generated from the issuer certificate alone which can then be
2067 passed to OCSP_id_issuer_cmp().
2070 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2071 ASN1 modules to export functions returning ASN1_ITEM pointers
2072 instead of the ASN1_ITEM structures themselves. This adds several
2073 new macros which allow the underlying ASN1 function/structure to
2074 be accessed transparently. As a result code should not use ASN1_ITEM
2075 references directly (such as &X509_it) but instead use the relevant
2076 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2077 use of the new ASN1 code on platforms where exporting structures
2078 is problematical (for example in shared libraries) but exporting
2079 functions returning pointers to structures is not.
2082 *) Add support for overriding the generation of SSL/TLS session IDs.
2083 These callbacks can be registered either in an SSL_CTX or per SSL.
2084 The purpose of this is to allow applications to control, if they wish,
2085 the arbitrary values chosen for use as session IDs, particularly as it
2086 can be useful for session caching in multiple-server environments. A
2087 command-line switch for testing this (and any client code that wishes
2088 to use such a feature) has been added to "s_server".
2089 [Geoff Thorpe, Lutz Jaenicke]
2091 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2092 of the form '#if defined(...) || defined(...) || ...' and
2093 '#if !defined(...) && !defined(...) && ...'. This also avoids
2094 the growing number of special cases it was previously handling.
2097 *) Make all configuration macros available for application by making
2098 sure they are available in opensslconf.h, by giving them names starting
2099 with "OPENSSL_" to avoid conflicts with other packages and by making
2100 sure e_os2.h will cover all platform-specific cases together with
2102 Additionally, it is now possible to define configuration/platform-
2103 specific names (called "system identities"). In the C code, these
2104 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2105 macro with the name beginning with "OPENSSL_SYS_", which is determined
2106 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2110 *) New option -set_serial to 'req' and 'x509' this allows the serial
2111 number to use to be specified on the command line. Previously self
2112 signed certificates were hard coded with serial number 0 and the
2113 CA options of 'x509' had to use a serial number in a file which was
2117 *) New options to 'ca' utility to support V2 CRL entry extensions.
2118 Currently CRL reason, invalidity date and hold instruction are
2119 supported. Add new CRL extensions to V3 code and some new objects.
2122 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2123 disable standard block padding (aka PKCS#5 padding) in the EVP
2124 API, which was previously mandatory. This means that the data is
2125 not padded in any way and so the total length much be a multiple
2126 of the block size, otherwise an error occurs.
2129 *) Initial (incomplete) OCSP SSL support.
2132 *) New function OCSP_parse_url(). This splits up a URL into its host,
2133 port and path components: primarily to parse OCSP URLs. New -url
2134 option to ocsp utility.
2137 *) New nonce behavior. The return value of OCSP_check_nonce() now
2138 reflects the various checks performed. Applications can decide
2139 whether to tolerate certain situations such as an absent nonce
2140 in a response when one was present in a request: the ocsp application
2141 just prints out a warning. New function OCSP_add1_basic_nonce()
2142 this is to allow responders to include a nonce in a response even if
2143 the request is nonce-less.
2146 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2147 skipped when using openssl x509 multiple times on a single input file,
2148 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2151 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2152 set string type: to handle setting ASN1_TIME structures. Fix ca
2153 utility to correctly initialize revocation date of CRLs.
2156 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2157 the clients preferred ciphersuites and rather use its own preferences.
2158 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2159 Internet Explorer by ensuring unchanged hash method during stepup.
2160 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2163 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2164 to aes and add a new 'exist' option to print out symbols that don't
2168 *) Additional options to ocsp utility to allow flags to be set and
2169 additional certificates supplied.
2172 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2173 OCSP client a number of certificate to only verify the response
2177 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2178 handle the new API. Currently only ECB, CBC modes supported. Add new
2181 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2182 Encryption Standard (AES) Ciphersuites for Transport Layer
2183 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2184 not enabled by default and were not part of the "ALL" ciphersuite
2185 alias because they were not yet official; they could be
2186 explicitly requested by specifying the "AESdraft" ciphersuite
2187 group alias. In the final release of OpenSSL 0.9.7, the group
2188 alias is called "AES" and is part of "ALL".)
2189 [Ben Laurie, Steve Henson, Bodo Moeller]
2191 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2192 request to response.
2195 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2196 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2197 extract information from a certificate request. OCSP_response_create()
2198 creates a response and optionally adds a basic response structure.
2199 OCSP_basic_add1_status() adds a complete single response to a basic
2200 response and returns the OCSP_SINGLERESP structure just added (to allow
2201 extensions to be included for example). OCSP_basic_add1_cert() adds a
2202 certificate to a basic response and OCSP_basic_sign() signs a basic
2203 response with various flags. New helper functions ASN1_TIME_check()
2204 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2205 (converts ASN1_TIME to GeneralizedTime).
2208 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2209 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2210 structure from a certificate. X509_pubkey_digest() digests the public_key
2211 contents: this is used in various key identifiers.
2214 *) Make sk_sort() tolerate a NULL argument.
2215 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2217 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2218 passed by the function are trusted implicitly. If any of them signed the
2219 response then it is assumed to be valid and is not verified.
2222 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2223 to data. This was previously part of the PKCS7 ASN1 code. This
2224 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2225 [Steve Henson, reported by Kenneth R. Robinette
2226 <support@securenetterm.com>]
2228 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2229 routines: without these tracing memory leaks is very painful.
2230 Fix leaks in PKCS12 and PKCS7 routines.
2233 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2234 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2235 effectively meant GeneralizedTime would never be used. Now it
2236 is initialised to -1 but X509_time_adj() now has to check the value
2237 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2238 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2239 [Steve Henson, reported by Kenneth R. Robinette
2240 <support@securenetterm.com>]
2242 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2243 result in a zero length in the ASN1_INTEGER structure which was
2244 not consistent with the structure when d2i_ASN1_INTEGER() was used
2245 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2246 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2247 where it did not print out a minus for negative ASN1_INTEGER.
2250 *) Add summary printout to ocsp utility. The various functions which
2251 convert status values to strings have been renamed to:
2252 OCSP_response_status_str(), OCSP_cert_status_str() and
2253 OCSP_crl_reason_str() and are no longer static. New options
2254 to verify nonce values and to disable verification. OCSP response
2255 printout format cleaned up.
2258 *) Add additional OCSP certificate checks. These are those specified
2259 in RFC2560. This consists of two separate checks: the CA of the
2260 certificate being checked must either be the OCSP signer certificate
2261 or the issuer of the OCSP signer certificate. In the latter case the
2262 OCSP signer certificate must contain the OCSP signing extended key
2263 usage. This check is performed by attempting to match the OCSP
2264 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2265 in the OCSP_CERTID structures of the response.
2268 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2269 and related routines. This uses the standard OpenSSL certificate
2270 verify routines to perform initial checks (just CA validity) and
2271 to obtain the certificate chain. Then additional checks will be
2272 performed on the chain. Currently the root CA is checked to see
2273 if it is explicitly trusted for OCSP signing. This is used to set
2274 a root CA as a global signing root: that is any certificate that
2275 chains to that CA is an acceptable OCSP signing certificate.
2278 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2279 extensions from a separate configuration file.
2280 As when reading extensions from the main configuration file,
2281 the '-extensions ...' option may be used for specifying the
2283 [Massimiliano Pala <madwolf@comune.modena.it>]
2285 *) New OCSP utility. Allows OCSP requests to be generated or
2286 read. The request can be sent to a responder and the output
2287 parsed, outputed or printed in text form. Not complete yet:
2288 still needs to check the OCSP response validity.
2291 *) New subcommands for 'openssl ca':
2292 'openssl ca -status <serial>' prints the status of the cert with
2293 the given serial number (according to the index file).
2294 'openssl ca -updatedb' updates the expiry status of certificates
2296 [Massimiliano Pala <madwolf@comune.modena.it>]
2298 *) New '-newreq-nodes' command option to CA.pl. This is like
2299 '-newreq', but calls 'openssl req' with the '-nodes' option
2300 so that the resulting key is not encrypted.
2301 [Damien Miller <djm@mindrot.org>]
2303 *) New configuration for the GNU Hurd.
2304 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2306 *) Initial code to implement OCSP basic response verify. This
2307 is currently incomplete. Currently just finds the signer's
2308 certificate and verifies the signature on the response.
2311 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2312 value of OPENSSLDIR. This is available via the new '-d' option
2313 to 'openssl version', and is also included in 'openssl version -a'.
2316 *) Allowing defining memory allocation callbacks that will be given
2317 file name and line number information in additional arguments
2318 (a const char* and an int). The basic functionality remains, as
2319 well as the original possibility to just replace malloc(),
2320 realloc() and free() by functions that do not know about these
2321 additional arguments. To register and find out the current
2322 settings for extended allocation functions, the following
2323 functions are provided:
2325 CRYPTO_set_mem_ex_functions
2326 CRYPTO_set_locked_mem_ex_functions
2327 CRYPTO_get_mem_ex_functions
2328 CRYPTO_get_locked_mem_ex_functions
2330 These work the same way as CRYPTO_set_mem_functions and friends.
2331 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2332 extended allocation function is enabled.
2333 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2334 a conventional allocation function is enabled.
2335 [Richard Levitte, Bodo Moeller]
2337 *) Finish off removing the remaining LHASH function pointer casts.
2338 There should no longer be any prototype-casting required when using
2339 the LHASH abstraction, and any casts that remain are "bugs". See
2340 the callback types and macros at the head of lhash.h for details
2341 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2344 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2345 If /dev/[u]random devices are not available or do not return enough
2346 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2348 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2349 /etc/entropy will be queried once each in this sequence, quering stops
2350 when enough entropy was collected without querying more sockets.
2353 *) Change the Unix RAND_poll() variant to be able to poll several
2354 random devices, as specified by DEVRANDOM, until a sufficient amount
2355 of data has been collected. We spend at most 10 ms on each file
2356 (select timeout) and read in non-blocking mode. DEVRANDOM now
2357 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2358 (previously it was just the string "/dev/urandom"), so on typical
2359 platforms the 10 ms delay will never occur.
2360 Also separate out the Unix variant to its own file, rand_unix.c.
2361 For VMS, there's a currently-empty rand_vms.c.
2364 *) Move OCSP client related routines to ocsp_cl.c. These
2365 provide utility functions which an application needing
2366 to issue a request to an OCSP responder and analyse the
2367 response will typically need: as opposed to those which an
2368 OCSP responder itself would need which will be added later.
2370 OCSP_request_sign() signs an OCSP request with an API similar
2371 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2372 response. OCSP_response_get1_basic() extracts basic response
2373 from response. OCSP_resp_find_status(): finds and extracts status
2374 information from an OCSP_CERTID structure (which will be created
2375 when the request structure is built). These are built from lower
2376 level functions which work on OCSP_SINGLERESP structures but
2377 wont normally be used unless the application wishes to examine
2378 extensions in the OCSP response for example.
2380 Replace nonce routines with a pair of functions.
2381 OCSP_request_add1_nonce() adds a nonce value and optionally
2382 generates a random value. OCSP_check_nonce() checks the
2383 validity of the nonce in an OCSP response.
2386 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2387 This doesn't copy the supplied OCSP_CERTID and avoids the
2388 need to free up the newly created id. Change return type
2389 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2390 This can then be used to add extensions to the request.
2391 Deleted OCSP_request_new(), since most of its functionality
2392 is now in OCSP_REQUEST_new() (and the case insensitive name
2393 clash) apart from the ability to set the request name which
2394 will be added elsewhere.
2397 *) Update OCSP API. Remove obsolete extensions argument from
2398 various functions. Extensions are now handled using the new
2399 OCSP extension code. New simple OCSP HTTP function which
2400 can be used to send requests and parse the response.
2403 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2404 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2405 uses the special reorder version of SET OF to sort the attributes
2406 and reorder them to match the encoded order. This resolves a long
2407 standing problem: a verify on a PKCS7 structure just after signing
2408 it used to fail because the attribute order did not match the
2409 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2410 it uses the received order. This is necessary to tolerate some broken
2411 software that does not order SET OF. This is handled by encoding
2412 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2413 to produce the required SET OF.
2416 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2417 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2418 files to get correct declarations of the ASN.1 item variables.
2421 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2422 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2423 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2424 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2425 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2426 ASN1_ITEM and no wrapper functions.
2429 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2430 replace the old function pointer based I/O routines. Change most of
2431 the *_d2i_bio() and *_d2i_fp() functions to use these.
2434 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2435 lines, recognice more "algorithms" that can be deselected, and make
2436 it complain about algorithm deselection that isn't recognised.
2439 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2440 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2441 to use new functions. Add NO_ASN1_OLD which can be set to remove
2442 some old style ASN1 functions: this can be used to determine if old
2443 code will still work when these eventually go away.
2446 *) New extension functions for OCSP structures, these follow the
2447 same conventions as certificates and CRLs.
2450 *) New function X509V3_add1_i2d(). This automatically encodes and
2451 adds an extension. Its behaviour can be customised with various
2452 flags to append, replace or delete. Various wrappers added for
2453 certifcates and CRLs.
2456 *) Fix to avoid calling the underlying ASN1 print routine when
2457 an extension cannot be parsed. Correct a typo in the
2458 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2461 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2462 entries for variables.
2465 *) Add functionality to apps/openssl.c for detecting locking
2466 problems: As the program is single-threaded, all we have
2467 to do is register a locking callback using an array for
2468 storing which locks are currently held by the program.
2471 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2472 SSL_get_ex_data_X509_STORE_idx(), which is used in
2473 ssl_verify_cert_chain() and thus can be called at any time
2474 during TLS/SSL handshakes so that thread-safety is essential.
2475 Unfortunately, the ex_data design is not at all suited
2476 for multi-threaded use, so it probably should be abolished.
2479 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2480 [Broadcom, tweaked and integrated by Geoff Thorpe]
2482 *) Move common extension printing code to new function
2483 X509V3_print_extensions(). Reorganise OCSP print routines and
2484 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2487 *) New function X509_signature_print() to remove duplication in some
2491 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2492 set (this was treated exactly the same as SET OF previously). This
2493 is used to reorder the STACK representing the structure to match the
2494 encoding. This will be used to get round a problem where a PKCS7
2495 structure which was signed could not be verified because the STACK
2496 order did not reflect the encoded order.
2499 *) Reimplement the OCSP ASN1 module using the new code.
2502 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2503 for its ASN1 operations. The old style function pointers still exist
2504 for now but they will eventually go away.
2507 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2508 completely replaces the old ASN1 functionality with a table driven
2509 encoder and decoder which interprets an ASN1_ITEM structure describing
2510 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2511 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2512 has also been converted to the new form.
2515 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2516 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2517 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2518 for negative moduli.
2521 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2522 of not touching the result's sign bit.
2525 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2529 *) Changed the LHASH code to use prototypes for callbacks, and created
2530 macros to declare and implement thin (optionally static) functions
2531 that provide type-safety and avoid function pointer casting for the
2532 type-specific callbacks.
2535 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2537 [Veers Staats <staatsvr@asc.hpc.mil>,
2538 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2540 *) Reformat the FAQ so the different questions and answers can be divided
2541 in sections depending on the subject.
2544 *) Have the zlib compression code load ZLIB.DLL dynamically under
2548 *) New function BN_mod_sqrt for computing square roots modulo a prime
2549 (using the probabilistic Tonelli-Shanks algorithm unless
2550 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2551 be handled deterministically).
2552 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2554 *) Make BN_mod_inverse faster by explicitly handling small quotients
2555 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2556 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2559 *) New function BN_kronecker.
2562 *) Fix BN_gcd so that it works on negative inputs; the result is
2563 positive unless both parameters are zero.
2564 Previously something reasonably close to an infinite loop was
2565 possible because numbers could be growing instead of shrinking
2566 in the implementation of Euclid's algorithm.
2569 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2570 sign of the number in question.
2572 Fix BN_is_word(a,w) to work correctly for w == 0.
2574 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2575 because its test if the absolute value of 'a' equals 'w'.
2576 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2577 it exists mostly for use in the implementations of BN_is_zero(),
2578 BN_is_one(), and BN_is_word().
2581 *) New function BN_swap.
2584 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2585 the exponentiation functions are more likely to produce reasonable
2586 results on negative inputs.
2589 *) Change BN_mod_mul so that the result is always non-negative.
2590 Previously, it could be negative if one of the factors was negative;
2591 I don't think anyone really wanted that behaviour.
2594 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2595 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2596 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2597 and add new functions:
2606 BN_mod_lshift1_quick
2610 These functions always generate non-negative results.
2612 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2613 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2615 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2616 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2617 be reduced modulo m.
2618 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2621 The following entry accidentily appeared in the CHANGES file
2622 distributed with OpenSSL 0.9.7. The modifications described in
2623 it do *not* apply to OpenSSL 0.9.7.
2625 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2626 was actually never needed) and in BN_mul(). The removal in BN_mul()
2627 required a small change in bn_mul_part_recursive() and the addition
2628 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2629 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2630 bn_sub_words() and bn_add_words() except they take arrays with
2635 *) In 'openssl passwd', verify passwords read from the terminal
2636 unless the '-salt' option is used (which usually means that
2637 verification would just waste user's time since the resulting
2638 hash is going to be compared with some given password hash)
2639 or the new '-noverify' option is used.
2641 This is an incompatible change, but it does not affect
2642 non-interactive use of 'openssl passwd' (passwords on the command
2643 line, '-stdin' option, '-in ...' option) and thus should not
2647 *) Remove all references to RSAref, since there's no more need for it.
2650 *) Make DSO load along a path given through an environment variable
2651 (SHLIB_PATH) with shl_load().
2654 *) Constify the ENGINE code as a result of BIGNUM constification.
2655 Also constify the RSA code and most things related to it. In a
2656 few places, most notable in the depth of the ASN.1 code, ugly
2657 casts back to non-const were required (to be solved at a later
2661 *) Make it so the openssl application has all engines loaded by default.
2664 *) Constify the BIGNUM routines a little more.
2667 *) Add the following functions:
2669 ENGINE_load_cswift()
2671 ENGINE_load_atalla()
2673 ENGINE_load_builtin_engines()
2675 That way, an application can itself choose if external engines that
2676 are built-in in OpenSSL shall ever be used or not. The benefit is
2677 that applications won't have to be linked with libdl or other dso
2678 libraries unless it's really needed.
2680 Changed 'openssl engine' to load all engines on demand.
2681 Changed the engine header files to avoid the duplication of some
2682 declarations (they differed!).
2685 *) 'openssl engine' can now list capabilities.
2688 *) Better error reporting in 'openssl engine'.
2691 *) Never call load_dh_param(NULL) in s_server.
2694 *) Add engine application. It can currently list engines by name and
2695 identity, and test if they are actually available.
2698 *) Improve RPM specification file by forcing symbolic linking and making
2699 sure the installed documentation is also owned by root.root.
2700 [Damien Miller <djm@mindrot.org>]
2702 *) Give the OpenSSL applications more possibilities to make use of
2703 keys (public as well as private) handled by engines.
2706 *) Add OCSP code that comes from CertCo.
2709 *) Add VMS support for the Rijndael code.
2712 *) Added untested support for Nuron crypto accelerator.
2715 *) Add support for external cryptographic devices. This code was
2716 previously distributed separately as the "engine" branch.
2717 [Geoff Thorpe, Richard Levitte]
2719 *) Rework the filename-translation in the DSO code. It is now possible to
2720 have far greater control over how a "name" is turned into a filename
2721 depending on the operating environment and any oddities about the
2722 different shared library filenames on each system.
2725 *) Support threads on FreeBSD-elf in Configure.
2728 *) Fix for SHA1 assembly problem with MASM: it produces
2729 warnings about corrupt line number information when assembling
2730 with debugging information. This is caused by the overlapping
2732 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2735 NCONF_get_number() has no error checking at all. As a replacement,
2736 NCONF_get_number_e() is defined (_e for "error checking") and is
2737 promoted strongly. The old NCONF_get_number is kept around for
2738 binary backward compatibility.
2739 Make it possible for methods to load from something other than a BIO,
2740 by providing a function pointer that is given a name instead of a BIO.
2741 For example, this could be used to load configuration data from an
2745 *) Fix for non blocking accept BIOs. Added new I/O special reason
2746 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2747 with non blocking I/O was not possible because no retry code was
2748 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2752 *) Added the beginnings of Rijndael support.
2755 *) Fix for bug in DirectoryString mask setting. Add support for
2756 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2757 to allow certificate printing to more controllable, additional
2758 'certopt' option to 'x509' to allow new printing options to be
2762 *) Clean old EAY MD5 hack from e_os.h.
2765 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2767 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2768 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2769 [Joe Orton, Steve Henson]
2771 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2773 *) Fix additional bug revealed by the NISCC test suite:
2775 Stop bug triggering large recursion when presented with
2776 certain ASN.1 tags (CAN-2003-0851)
2779 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2781 *) Fix various bugs revealed by running the NISCC test suite:
2783 Stop out of bounds reads in the ASN1 code when presented with
2784 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2786 If verify callback ignores invalid public key errors don't try to check
2787 certificate signature with the NULL public key.
2791 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2792 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2796 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2797 extra data after the compression methods not only for TLS 1.0
2798 but also for SSL 3.0 (as required by the specification).
2799 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2801 *) Change X509_certificate_type() to mark the key as exported/exportable
2802 when it's 512 *bits* long, not 512 bytes.
2805 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2807 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2808 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2809 a protocol version number mismatch like a decryption error
2810 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2813 *) Turn on RSA blinding by default in the default implementation
2814 to avoid a timing attack. Applications that don't want it can call
2815 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2816 They would be ill-advised to do so in most cases.
2817 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2819 *) Change RSA blinding code so that it works when the PRNG is not
2820 seeded (in this case, the secret RSA exponent is abused as
2821 an unpredictable seed -- if it is not unpredictable, there
2822 is no point in blinding anyway). Make RSA blinding thread-safe
2823 by remembering the creator's thread ID in rsa->blinding and
2824 having all other threads use local one-time blinding factors
2825 (this requires more computation than sharing rsa->blinding, but
2826 avoids excessive locking; and if an RSA object is not shared
2827 between threads, blinding will still be very fast).
2830 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2832 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2833 via timing by performing a MAC computation even if incorrrect
2834 block cipher padding has been found. This is a countermeasure
2835 against active attacks where the attacker has to distinguish
2836 between bad padding and a MAC verification error. (CAN-2003-0078)
2838 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2839 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2840 Martin Vuagnoux (EPFL, Ilion)]
2842 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2844 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2845 memory from it's contents. This is done with a counter that will
2846 place alternating values in each byte. This can be used to solve
2847 two issues: 1) the removal of calls to memset() by highly optimizing
2848 compilers, and 2) cleansing with other values than 0, since those can
2849 be read through on certain media, for example a swap space on disk.
2852 *) Bugfix: client side session caching did not work with external caching,
2853 because the session->cipher setting was not restored when reloading
2854 from the external cache. This problem was masked, when
2855 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2856 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2859 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2860 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2861 [Zeev Lieber <zeev-l@yahoo.com>]
2863 *) Undo an undocumented change introduced in 0.9.6e which caused
2864 repeated calls to OpenSSL_add_all_ciphers() and
2865 OpenSSL_add_all_digests() to be ignored, even after calling
2869 *) Change the default configuration reader to deal with last line not
2870 being properly terminated.
2873 *) Change X509_NAME_cmp() so it applies the special rules on handling
2874 DN values that are of type PrintableString, as well as RDNs of type
2875 emailAddress where the value has the type ia5String.
2876 [stefank@valicert.com via Richard Levitte]
2878 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2879 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2880 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2881 the bitwise-OR of the two for use by the majority of applications
2882 wanting this behaviour, and update the docs. The documented
2883 behaviour and actual behaviour were inconsistent and had been
2884 changing anyway, so this is more a bug-fix than a behavioural
2886 [Geoff Thorpe, diagnosed by Nadav Har'El]
2888 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2889 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2892 *) Fix initialization code race conditions in
2893 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2894 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2895 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2896 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2897 ssl2_get_cipher_by_char(),
2898 ssl3_get_cipher_by_char().
2899 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2901 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2902 the cached sessions are flushed, as the remove_cb() might use ex_data
2903 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2904 (see [openssl.org #212]).
2905 [Geoff Thorpe, Lutz Jaenicke]
2907 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2908 length, instead of the encoding length to d2i_ASN1_OBJECT.
2911 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2913 *) [In 0.9.6g-engine release:]
2914 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2915 [Lynn Gazis <lgazis@rainbow.com>]
2917 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2919 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2920 and get fix the header length calculation.
2921 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2922 Alon Kantor <alonk@checkpoint.com> (and others),
2925 *) Use proper error handling instead of 'assertions' in buffer
2926 overflow checks added in 0.9.6e. This prevents DoS (the
2927 assertions could call abort()).
2928 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2930 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2932 *) Add various sanity checks to asn1_get_length() to reject
2933 the ASN1 length bytes if they exceed sizeof(long), will appear
2934 negative or the content length exceeds the length of the
2936 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2938 *) Fix cipher selection routines: ciphers without encryption had no flags
2939 for the cipher strength set and where therefore not handled correctly
2940 by the selection routines (PR #130).
2943 *) Fix EVP_dsa_sha macro.
2947 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2948 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2949 that was added in OpenSSL 0.9.6d.
2951 As the countermeasure turned out to be incompatible with some
2952 broken SSL implementations, the new option is part of SSL_OP_ALL.
2953 SSL_OP_ALL is usually employed when compatibility with weird SSL
2954 implementations is desired (e.g. '-bugs' option to 's_client' and
2955 's_server'), so the new option is automatically set in many
2959 *) Changes in security patch:
2961 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2962 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2963 Air Force Materiel Command, USAF, under agreement number
2966 *) Add various sanity checks to asn1_get_length() to reject
2967 the ASN1 length bytes if they exceed sizeof(long), will appear
2968 negative or the content length exceeds the length of the
2969 supplied buffer. (CAN-2002-0659)
2970 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2972 *) Assertions for various potential buffer overflows, not known to
2974 [Ben Laurie (CHATS)]
2976 *) Various temporary buffers to hold ASCII versions of integers were
2977 too small for 64 bit platforms. (CAN-2002-0655)
2978 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2980 *) Remote buffer overflow in SSL3 protocol - an attacker could
2981 supply an oversized session ID to a client. (CAN-2002-0656)
2982 [Ben Laurie (CHATS)]
2984 *) Remote buffer overflow in SSL2 protocol - an attacker could
2985 supply an oversized client master key. (CAN-2002-0656)
2986 [Ben Laurie (CHATS)]
2988 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2990 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2991 encoded as NULL) with id-dsa-with-sha1.
2992 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2994 *) Check various X509_...() return values in apps/req.c.
2995 [Nils Larsch <nla@trustcenter.de>]
2997 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2998 an end-of-file condition would erronously be flagged, when the CRLF
2999 was just at the end of a processed block. The bug was discovered when
3000 processing data through a buffering memory BIO handing the data to a
3001 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3002 <ptsekov@syntrex.com> and Nedelcho Stanev.
3005 *) Implement a countermeasure against a vulnerability recently found
3006 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3007 before application data chunks to avoid the use of known IVs
3008 with data potentially chosen by the attacker.
3011 *) Fix length checks in ssl3_get_client_hello().
3014 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3015 to prevent ssl3_read_internal() from incorrectly assuming that
3016 ssl3_read_bytes() found application data while handshake
3017 processing was enabled when in fact s->s3->in_read_app_data was
3018 merely automatically cleared during the initial handshake.
3019 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3021 *) Fix object definitions for Private and Enterprise: they were not
3022 recognized in their shortname (=lowercase) representation. Extend
3023 obj_dat.pl to issue an error when using undefined keywords instead
3024 of silently ignoring the problem (Svenning Sorensen
3025 <sss@sss.dnsalias.net>).
3028 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3029 generators, i.e. generators other than 2 and 5. (Previously, the
3030 code did not properly initialise the 'add' and 'rem' values to
3031 BN_generate_prime().)
3033 In the new general case, we do not insist that 'generator' is
3034 actually a primitive root: This requirement is rather pointless;
3035 a generator of the order-q subgroup is just as good, if not
3039 *) Map new X509 verification errors to alerts. Discovered and submitted by
3040 Tom Wu <tom@arcot.com>.
3043 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3044 returning non-zero before the data has been completely received
3045 when using non-blocking I/O.
3046 [Bodo Moeller; problem pointed out by John Hughes]
3048 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3049 [Ben Laurie, Lutz Jaenicke]
3051 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3052 Yoram Zahavi <YoramZ@gilian.com>).
3055 *) Add information about CygWin 1.3 and on, and preserve proper
3056 configuration for the versions before that.
3057 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3059 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3060 check whether we deal with a copy of a session and do not delete from
3061 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3062 <izhar@checkpoint.com>.
3065 *) Do not store session data into the internal session cache, if it
3066 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3067 flag is set). Proposed by Aslam <aslam@funk.com>.
3070 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3074 *) [In 0.9.6d-engine release:]
3075 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3076 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3078 *) Add the configuration target linux-s390x.
3079 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3081 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3082 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3083 variable as an indication that a ClientHello message has been
3084 received. As the flag value will be lost between multiple
3085 invocations of ssl3_accept when using non-blocking I/O, the
3086 function may not be aware that a handshake has actually taken
3087 place, thus preventing a new session from being added to the
3090 To avoid this problem, we now set s->new_session to 2 instead of
3091 using a local variable.
3092 [Lutz Jaenicke, Bodo Moeller]
3094 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3095 if the SSL_R_LENGTH_MISMATCH error is detected.
3096 [Geoff Thorpe, Bodo Moeller]
3098 *) New 'shared_ldflag' column in Configure platform table.
3101 *) Fix EVP_CIPHER_mode macro.
3102 ["Dan S. Camper" <dan@bti.net>]
3104 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3105 type, we must throw them away by setting rr->length to 0.
3106 [D P Chang <dpc@qualys.com>]
3108 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3110 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3111 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3112 worked incorrectly for those cases where range = 10..._2 and
3113 3*range is two bits longer than range.)
3116 *) Only add signing time to PKCS7 structures if it is not already
3120 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3121 OBJ_ld_ce should be OBJ_id_ce.
3122 Also some ip-pda OIDs in crypto/objects/objects.txt were
3123 incorrect (cf. RFC 3039).
3124 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3126 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3127 returns early because it has nothing to do.
3128 [Andy Schneider <andy.schneider@bjss.co.uk>]
3130 *) [In 0.9.6c-engine release:]
3131 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3132 [Andy Schneider <andy.schneider@bjss.co.uk>]
3134 *) [In 0.9.6c-engine release:]
3135 Add support for Cryptographic Appliance's keyserver technology.
3136 (Use engine 'keyclient')
3137 [Cryptographic Appliances and Geoff Thorpe]
3139 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3140 is called via tools/c89.sh because arguments have to be
3141 rearranged (all '-L' options must appear before the first object
3143 [Richard Shapiro <rshapiro@abinitio.com>]
3145 *) [In 0.9.6c-engine release:]
3146 Add support for Broadcom crypto accelerator cards, backported
3148 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3150 *) [In 0.9.6c-engine release:]
3151 Add support for SureWare crypto accelerator cards from
3152 Baltimore Technologies. (Use engine 'sureware')
3153 [Baltimore Technologies and Mark Cox]
3155 *) [In 0.9.6c-engine release:]
3156 Add support for crypto accelerator cards from Accelerated
3157 Encryption Processing, www.aep.ie. (Use engine 'aep')
3158 [AEP Inc. and Mark Cox]
3160 *) Add a configuration entry for gcc on UnixWare.
3161 [Gary Benson <gbenson@redhat.com>]
3163 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3164 messages are stored in a single piece (fixed-length part and
3165 variable-length part combined) and fix various bugs found on the way.
3168 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3169 instead. BIO_gethostbyname() does not know what timeouts are
3170 appropriate, so entries would stay in cache even when they have
3172 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3174 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3175 faced with a pathologically small ClientHello fragment that does
3176 not contain client_version: Instead of aborting with an error,
3177 simply choose the highest available protocol version (i.e.,
3178 TLS 1.0 unless it is disabled). In practice, ClientHello
3179 messages are never sent like this, but this change gives us
3180 strictly correct behaviour at least for TLS.
3183 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3184 never resets s->method to s->ctx->method when called from within
3185 one of the SSL handshake functions.
3186 [Bodo Moeller; problem pointed out by Niko Baric]
3188 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3189 (sent using the client's version number) if client_version is
3190 smaller than the protocol version in use. Also change
3191 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3192 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3193 the client will at least see that alert.
3196 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3200 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3201 client receives HelloRequest while in a handshake.
3202 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3204 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3205 should end in 'break', not 'goto end' which circuments various
3206 cleanups done in state SSL_ST_OK. But session related stuff
3207 must be disabled for SSL_ST_OK in the case that we just sent a
3210 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3211 before just sending a HelloRequest.
3212 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3214 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3215 reveal whether illegal block cipher padding was found or a MAC
3216 verification error occured. (Neither SSLerr() codes nor alerts
3217 are directly visible to potential attackers, but the information
3218 may leak via logfiles.)
3220 Similar changes are not required for the SSL 2.0 implementation
3221 because the number of padding bytes is sent in clear for SSL 2.0,
3222 and the extra bytes are just ignored. However ssl/s2_pkt.c
3223 failed to verify that the purported number of padding bytes is in
3227 *) Add OpenUNIX-8 support including shared libraries
3228 (Boyd Lynn Gerber <gerberb@zenez.com>).
3231 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3232 'wristwatch attack' using huge encoding parameters (cf.
3233 James H. Manger's CRYPTO 2001 paper). Note that the
3234 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3235 encoding parameters and hence was not vulnerable.
3238 *) BN_sqr() bug fix.
3239 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3241 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3242 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3243 followed by modular reduction.
3244 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3246 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3247 equivalent based on BN_pseudo_rand() instead of BN_rand().
3250 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3251 This function was broken, as the check for a new client hello message
3252 to handle SGC did not allow these large messages.
3253 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3256 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3259 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3260 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3263 *) Rework the configuration and shared library support for Tru64 Unix.
3264 The configuration part makes use of modern compiler features and
3265 still retains old compiler behavior for those that run older versions
3266 of the OS. The shared library support part includes a variant that
3267 uses the RPATH feature, and is available through the special
3268 configuration target "alpha-cc-rpath", which will never be selected
3270 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3272 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3273 with the same message size as in ssl3_get_certificate_request().
3274 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3275 messages might inadvertently be reject as too long.
3276 [Petr Lampa <lampa@fee.vutbr.cz>]
3278 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3281 *) Modified SSL library such that the verify_callback that has been set
3282 specificly for an SSL object with SSL_set_verify() is actually being
3283 used. Before the change, a verify_callback set with this function was
3284 ignored and the verify_callback() set in the SSL_CTX at the time of
3285 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3286 to allow the necessary settings.
3289 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3290 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3291 done automatically (in contradiction to the requirements of the C
3292 standard). This made problems when used from OpenSSH.
3295 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3296 dh->length and always used
3298 BN_rand_range(priv_key, dh->p).
3300 BN_rand_range() is not necessary for Diffie-Hellman, and this
3301 specific range makes Diffie-Hellman unnecessarily inefficient if
3302 dh->length (recommended exponent length) is much smaller than the
3303 length of dh->p. We could use BN_rand_range() if the order of
3304 the subgroup was stored in the DH structure, but we only have
3309 BN_rand(priv_key, l, ...)
3311 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3317 RSA_eay_public_encrypt
3318 RSA_eay_private_decrypt
3319 RSA_eay_private_encrypt (signing)
3320 RSA_eay_public_decrypt (signature verification)
3322 (default implementations for RSA_public_encrypt,
3323 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3324 always reject numbers >= n.
3327 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3328 to synchronize access to 'locking_thread'. This is necessary on
3329 systems where access to 'locking_thread' (an 'unsigned long'
3330 variable) is not atomic.
3333 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3334 *before* setting the 'crypto_lock_rand' flag. The previous code had
3335 a race condition if 0 is a valid thread ID.
3336 [Travis Vitek <vitek@roguewave.com>]
3338 *) Add support for shared libraries under Irix.
3339 [Albert Chin-A-Young <china@thewrittenword.com>]
3341 *) Add configuration option to build on Linux on both big-endian and
3343 [Ralf Baechle <ralf@uni-koblenz.de>]
3345 *) Add the possibility to create shared libraries on HP-UX.
3348 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3350 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3351 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3352 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3353 PRNG state recovery was possible based on the output of
3354 one PRNG request appropriately sized to gain knowledge on
3355 'md' followed by enough consecutive 1-byte PRNG requests
3356 to traverse all of 'state'.
3358 1. When updating 'md_local' (the current thread's copy of 'md')
3359 during PRNG output generation, hash all of the previous
3360 'md_local' value, not just the half used for PRNG output.
3362 2. Make the number of bytes from 'state' included into the hash
3363 independent from the number of PRNG bytes requested.
3365 The first measure alone would be sufficient to avoid
3366 Markku-Juhani's attack. (Actually it had never occurred
3367 to me that the half of 'md_local' used for chaining was the
3368 half from which PRNG output bytes were taken -- I had always
3369 assumed that the secret half would be used.) The second
3370 measure makes sure that additional data from 'state' is never
3371 mixed into 'md_local' in small portions; this heuristically
3372 further strengthens the PRNG.
3375 *) Fix crypto/bn/asm/mips3.s.
3378 *) When only the key is given to "enc", the IV is undefined. Print out
3379 an error message in this case.
3382 *) Handle special case when X509_NAME is empty in X509 printing routines.
3385 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3386 positive and less than q.
3389 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3390 used: it isn't thread safe and the add_lock_callback should handle
3392 [Paul Rose <Paul.Rose@bridge.com>]
3394 *) Verify that incoming data obeys the block size in
3395 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3399 [Ulf Möller, Bodo Möller]
3401 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3402 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3403 when fixing the server behaviour for backwards-compatible 'client
3404 hello' messages. (Note that the attack is impractical against
3405 SSL 3.0 and TLS 1.0 anyway because length and version checking
3406 means that the probability of guessing a valid ciphertext is
3407 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3410 Before 0.9.5, the countermeasure (hide the error by generating a
3411 random 'decryption result') did not work properly because
3412 ERR_clear_error() was missing, meaning that SSL_get_error() would
3413 detect the supposedly ignored error.
3415 Both problems are now fixed.
3418 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3419 (previously it was 1024).
3422 *) Fix for compatibility mode trust settings: ignore trust settings
3423 unless some valid trust or reject settings are present.
3426 *) Fix for blowfish EVP: its a variable length cipher.
3429 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3430 parameters in DSA public key structures and return an error in the
3431 DSA routines if parameters are absent.
3434 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3435 in the current directory if neither $RANDFILE nor $HOME was set.
3436 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3437 caused some confusion to Windows users who haven't defined $HOME.
3438 Thus RAND_file_name() is changed again: e_os.h can define a
3439 DEFAULT_HOME, which will be used if $HOME is not set.
3440 For Windows, we use "C:"; on other platforms, we still require
3441 environment variables.
3443 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3444 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3445 having multiple threads call RAND_poll() concurrently.
3448 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3449 combination of a flag and a thread ID variable.
3450 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3451 flag), *other* threads can enter ssleay_add_bytes without obeying
3452 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3453 that they do not hold after the first thread unsets add_do_not_lock).
3456 *) Change bctest again: '-x' expressions are not available in all
3460 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3462 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3463 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3465 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3466 the default extension for executables, if any. Also, make the perl
3467 scripts that use symlink() to test if it really exists and use "cp"
3468 if it doesn't. All this made OpenSSL compilable and installable in
3472 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3473 If SEQUENCE is length is indefinite just set c->slen to the total
3474 amount of data available.
3475 [Steve Henson, reported by shige@FreeBSD.org]
3476 [This change does not apply to 0.9.7.]
3478 *) Change bctest to avoid here-documents inside command substitution
3479 (workaround for FreeBSD /bin/sh bug).
3480 For compatibility with Ultrix, avoid shell functions (introduced
3481 in the bctest version that searches along $PATH).
3484 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3485 with des_encrypt() defined on some operating systems, like Solaris
3489 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3490 On the Importance of Eliminating Errors in Cryptographic
3491 Computations, J. Cryptology 14 (2001) 2, 101-119,
3492 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3495 *) MIPS assembler BIGNUM division bug fix.
3498 *) Disabled incorrect Alpha assembler code.
3501 *) Fix PKCS#7 decode routines so they correctly update the length
3502 after reading an EOC for the EXPLICIT tag.
3504 [This change does not apply to 0.9.7.]
3506 *) Fix bug in PKCS#12 key generation routines. This was triggered
3507 if a 3DES key was generated with a 0 initial byte. Include
3508 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3509 (but broken) behaviour.
3512 *) Enhance bctest to search for a working bc along $PATH and print
3514 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3516 *) Fix memory leaks in err.c: free err_data string if necessary;
3517 don't write to the wrong index in ERR_set_error_data.
3520 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3524 *) Replace rdtsc with _emit statements for VC++ version 5.
3525 [Jeremy Cooper <jeremy@baymoo.org>]
3527 *) Make it possible to reuse SSLv2 sessions.
3530 *) In copy_email() check for >= 0 as a return value for
3531 X509_NAME_get_index_by_NID() since 0 is a valid index.
3532 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3534 *) Avoid coredump with unsupported or invalid public keys by checking if
3535 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3536 PKCS7_verify() fails with non detached data.
3539 *) Don't use getenv in library functions when run as setuid/setgid.
3540 New function OPENSSL_issetugid().
3543 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3544 due to incorrect handling of multi-threading:
3546 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3548 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3550 3. Count how many times MemCheck_off() has been called so that
3551 nested use can be treated correctly. This also avoids
3552 inband-signalling in the previous code (which relied on the
3553 assumption that thread ID 0 is impossible).
3556 *) Add "-rand" option also to s_client and s_server.
3559 *) Fix CPU detection on Irix 6.x.
3560 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3561 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3563 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3566 [This change does not apply to 0.9.7.]
3568 *) Use the cached encoding of an X509_NAME structure rather than
3569 copying it. This is apparently the reason for the libsafe "errors"
3570 but the code is actually correct.
3573 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3574 Bleichenbacher's DSA attack.
3575 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3576 to be set and top=0 forces the highest bit to be set; top=-1 is new
3577 and leaves the highest bit random.
3578 [Ulf Moeller, Bodo Moeller]
3580 *) In the NCONF_...-based implementations for CONF_... queries
3581 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3582 a temporary CONF structure with the data component set to NULL
3583 (which gives segmentation faults in lh_retrieve).
3584 Instead, use NULL for the CONF pointer in CONF_get_string and
3585 CONF_get_number (which may use environment variables) and directly
3586 return NULL from CONF_get_section.
3589 *) Fix potential buffer overrun for EBCDIC.
3592 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3593 keyUsage if basicConstraints absent for a CA.
3596 *) Make SMIME_write_PKCS7() write mail header values with a format that
3597 is more generally accepted (no spaces before the semicolon), since
3598 some programs can't parse those values properly otherwise. Also make
3599 sure BIO's that break lines after each write do not create invalid
3603 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3604 macros previously used would not encode an empty SEQUENCE OF
3605 and break the signature.
3607 [This change does not apply to 0.9.7.]
3609 *) Zero the premaster secret after deriving the master secret in
3613 *) Add some EVP_add_digest_alias registrations (as found in
3614 OpenSSL_add_all_digests()) to SSL_library_init()
3615 aka OpenSSL_add_ssl_algorithms(). This provides improved
3616 compatibility with peers using X.509 certificates
3617 with unconventional AlgorithmIdentifier OIDs.
3620 *) Fix for Irix with NO_ASM.
3621 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3623 *) ./config script fixes.
3624 [Ulf Moeller, Richard Levitte]
3626 *) Fix 'openssl passwd -1'.
3629 *) Change PKCS12_key_gen_asc() so it can cope with non null
3630 terminated strings whose length is passed in the passlen
3631 parameter, for example from PEM callbacks. This was done
3632 by adding an extra length parameter to asc2uni().
3633 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3635 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3636 call failed, free the DSA structure.
3639 *) Fix to uni2asc() to cope with zero length Unicode strings.
3640 These are present in some PKCS#12 files.
3643 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3644 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3645 when writing a 32767 byte record.
3646 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3648 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3649 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3651 (RSA objects have a reference count access to which is protected
3652 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3653 so they are meant to be shared between threads.)
3654 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3655 "Reddie, Steven" <Steven.Reddie@ca.com>]
3657 *) Fix a deadlock in CRYPTO_mem_leaks().
3660 *) Use better test patterns in bntest.
3663 *) rand_win.c fix for Borland C.
3666 *) BN_rshift bugfix for n == 0.
3669 *) Add a 'bctest' script that checks for some known 'bc' bugs
3670 so that 'make test' does not abort just because 'bc' is broken.
3673 *) Store verify_result within SSL_SESSION also for client side to
3674 avoid potential security hole. (Re-used sessions on the client side
3675 always resulted in verify_result==X509_V_OK, not using the original
3676 result of the server certificate verification.)
3679 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3680 SSL3_RT_APPLICATION_DATA, return 0.
3681 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3685 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3686 releases, have been re-implemented by renaming the previous
3687 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3688 and ssl3_read_internal, respectively, and adding 'peek' parameters
3689 to them. The new ssl[23]_{read,peek} functions are calls to
3690 ssl[23]_read_internal with the 'peek' flag set appropriately.
3691 A 'peek' parameter has also been added to ssl3_read_bytes, which
3692 does the actual work for ssl3_read_internal.
3695 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3696 the method-specific "init()" handler. Also clean up ex_data after
3697 calling the method-specific "finish()" handler. Previously, this was
3698 happening the other way round.
3701 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3702 The previous value, 12, was not always sufficient for BN_mod_exp().
3705 *) Make sure that shared libraries get the internal name engine with
3706 the full version number and not just 0. This should mark the
3707 shared libraries as not backward compatible. Of course, this should
3708 be changed again when we can guarantee backward binary compatibility.
3711 *) Fix typo in get_cert_by_subject() in by_dir.c
3712 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3714 *) Rework the system to generate shared libraries:
3716 - Make note of the expected extension for the shared libraries and
3717 if there is a need for symbolic links from for example libcrypto.so.0
3718 to libcrypto.so.0.9.7. There is extended info in Configure for
3721 - Make as few rebuilds of the shared libraries as possible.
3723 - Still avoid linking the OpenSSL programs with the shared libraries.
3725 - When installing, install the shared libraries separately from the
3729 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3731 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3732 and not in SSL_clear because the latter is also used by the
3733 accept/connect functions; previously, the settings made by
3734 SSL_set_read_ahead would be lost during the handshake.
3735 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3737 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3738 Previously, it would create entries for disableed algorithms no
3742 *) Added several new manual pages for SSL_* function.
3745 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3747 *) In ssl23_get_client_hello, generate an error message when faced
3748 with an initial SSL 3.0/TLS record that is too small to contain the
3749 first two bytes of the ClientHello message, i.e. client_version.
3750 (Note that this is a pathologic case that probably has never happened
3751 in real life.) The previous approach was to use the version number
3752 from the record header as a substitute; but our protocol choice
3753 should not depend on that one because it is not authenticated
3754 by the Finished messages.
3757 *) More robust randomness gathering functions for Windows.
3758 [Jeffrey Altman <jaltman@columbia.edu>]
3760 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3761 not set then we don't setup the error code for issuer check errors
3762 to avoid possibly overwriting other errors which the callback does
3763 handle. If an application does set the flag then we assume it knows
3764 what it is doing and can handle the new informational codes
3768 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3769 a general "ANY" type, as such it should be able to decode anything
3770 including tagged types. However it didn't check the class so it would
3771 wrongly interpret tagged types in the same way as their universal
3772 counterpart and unknown types were just rejected. Changed so that the
3773 tagged and unknown types are handled in the same way as a SEQUENCE:
3774 that is the encoding is stored intact. There is also a new type
3775 "V_ASN1_OTHER" which is used when the class is not universal, in this
3776 case we have no idea what the actual type is so we just lump them all
3780 *) On VMS, stdout may very well lead to a file that is written to
3781 in a record-oriented fashion. That means that every write() will
3782 write a separate record, which will be read separately by the
3783 programs trying to read from it. This can be very confusing.
3785 The solution is to put a BIO filter in the way that will buffer
3786 text until a linefeed is reached, and then write everything a
3787 line at a time, so every record written will be an actual line,
3788 not chunks of lines and not (usually doesn't happen, but I've
3789 seen it once) several lines in one record. BIO_f_linebuffer() is
3792 Currently, it's a VMS-only method, because that's where it has
3793 been tested well enough.
3796 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3797 it can return incorrect results.
3798 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3799 but it was in 0.9.6-beta[12].)
3802 *) Disable the check for content being present when verifying detached
3803 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3804 include zero length content when signing messages.
3807 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3808 BIO_ctrl (for BIO pairs).
3811 *) Add DSO method for VMS.
3814 *) Bug fix: Montgomery multiplication could produce results with the
3818 *) Add RPM specification openssl.spec and modify it to build three
3819 packages. The default package contains applications, application
3820 documentation and run-time libraries. The devel package contains
3821 include files, static libraries and function documentation. The
3822 doc package contains the contents of the doc directory. The original
3823 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3826 *) Add a large number of documentation files for many SSL routines.
3827 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3829 *) Add a configuration entry for Sony News 4.
3830 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3832 *) Don't set the two most significant bits to one when generating a
3833 random number < q in the DSA library.
3836 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3837 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3838 the underlying transport is blocking) if a handshake took place.
3839 (The default behaviour is needed by applications such as s_client
3840 and s_server that use select() to determine when to use SSL_read;
3841 but for applications that know in advance when to expect data, it
3842 just makes things more complicated.)
3845 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3849 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3850 work better on such systems.
3851 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3853 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3854 Update PKCS12_parse() so it copies the friendlyName and the
3855 keyid to the certificates aux info.
3858 *) Fix bug in PKCS7_verify() which caused an infinite loop
3859 if there was more than one signature.
3860 [Sven Uszpelkat <su@celocom.de>]
3862 *) Major change in util/mkdef.pl to include extra information
3863 about each symbol, as well as presentig variables as well
3864 as functions. This change means that there's n more need
3865 to rebuild the .num files when some algorithms are excluded.
3868 *) Allow the verify time to be set by an application,
3869 rather than always using the current time.
3872 *) Phase 2 verify code reorganisation. The certificate
3873 verify code now looks up an issuer certificate by a
3874 number of criteria: subject name, authority key id
3875 and key usage. It also verifies self signed certificates
3876 by the same criteria. The main comparison function is
3877 X509_check_issued() which performs these checks.
3879 Lot of changes were necessary in order to support this
3880 without completely rewriting the lookup code.
3882 Authority and subject key identifier are now cached.
3884 The LHASH 'certs' is X509_STORE has now been replaced
3885 by a STACK_OF(X509_OBJECT). This is mainly because an
3886 LHASH can't store or retrieve multiple objects with
3887 the same hash value.
3889 As a result various functions (which were all internal
3890 use only) have changed to handle the new X509_STORE
3891 structure. This will break anything that messed round
3892 with X509_STORE internally.
3894 The functions X509_STORE_add_cert() now checks for an
3895 exact match, rather than just subject name.
3897 The X509_STORE API doesn't directly support the retrieval
3898 of multiple certificates matching a given criteria, however
3899 this can be worked round by performing a lookup first
3900 (which will fill the cache with candidate certificates)
3901 and then examining the cache for matches. This is probably
3902 the best we can do without throwing out X509_LOOKUP
3903 entirely (maybe later...).
3905 The X509_VERIFY_CTX structure has been enhanced considerably.
3907 All certificate lookup operations now go via a get_issuer()
3908 callback. Although this currently uses an X509_STORE it
3909 can be replaced by custom lookups. This is a simple way
3910 to bypass the X509_STORE hackery necessary to make this
3911 work and makes it possible to use more efficient techniques
3912 in future. A very simple version which uses a simple
3913 STACK for its trusted certificate store is also provided
3914 using X509_STORE_CTX_trusted_stack().
3916 The verify_cb() and verify() callbacks now have equivalents
3917 in the X509_STORE_CTX structure.
3919 X509_STORE_CTX also has a 'flags' field which can be used
3920 to customise the verify behaviour.
3923 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3924 excludes S/MIME capabilities.
3927 *) When a certificate request is read in keep a copy of the
3928 original encoding of the signed data and use it when outputing
3929 again. Signatures then use the original encoding rather than
3930 a decoded, encoded version which may cause problems if the
3931 request is improperly encoded.
3934 *) For consistency with other BIO_puts implementations, call
3935 buffer_write(b, ...) directly in buffer_puts instead of calling
3938 In BIO_puts, increment b->num_write as in BIO_write.
3939 [Peter.Sylvester@EdelWeb.fr]
3941 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3942 BN_zero, we may not return a BIGNUM with an array consisting of
3946 *) Avoid calling abort() from within the library when problems are
3947 detected, except if preprocessor symbols have been defined
3948 (such as REF_CHECK, BN_DEBUG etc.).
3951 *) New openssl application 'rsautl'. This utility can be
3952 used for low level RSA operations. DER public key
3953 BIO/fp routines also added.
3956 *) New Configure entry and patches for compiling on QNX 4.
3957 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3959 *) A demo state-machine implementation was sponsored by
3960 Nuron (http://www.nuron.com/) and is now available in
3961 demos/state_machine.
3964 *) New options added to the 'dgst' utility for signature
3965 generation and verification.
3968 *) Unrecognized PKCS#7 content types are now handled via a
3969 catch all ASN1_TYPE structure. This allows unsupported
3970 types to be stored as a "blob" and an application can
3971 encode and decode it manually.
3974 *) Fix various signed/unsigned issues to make a_strex.c
3976 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3978 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3979 length if passed a buffer. ASN1_INTEGER_to_BN failed
3980 if passed a NULL BN and its argument was negative.
3981 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3983 *) Modification to PKCS#7 encoding routines to output definite
3984 length encoding. Since currently the whole structures are in
3985 memory there's not real point in using indefinite length
3986 constructed encoding. However if OpenSSL is compiled with
3987 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3990 *) Added BIO_vprintf() and BIO_vsnprintf().
3993 *) Added more prefixes to parse for in the the strings written
3994 through a logging bio, to cover all the levels that are available
3995 through syslog. The prefixes are now:
3997 PANIC, EMERG, EMR => LOG_EMERG
3998 ALERT, ALR => LOG_ALERT
3999 CRIT, CRI => LOG_CRIT
4000 ERROR, ERR => LOG_ERR
4001 WARNING, WARN, WAR => LOG_WARNING
4002 NOTICE, NOTE, NOT => LOG_NOTICE
4003 INFO, INF => LOG_INFO
4004 DEBUG, DBG => LOG_DEBUG
4006 and as before, if none of those prefixes are present at the
4007 beginning of the string, LOG_ERR is chosen.
4009 On Win32, the LOG_* levels are mapped according to this:
4011 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4012 LOG_WARNING => EVENTLOG_WARNING_TYPE
4013 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4017 *) Made it possible to reconfigure with just the configuration
4018 argument "reconf" or "reconfigure". The command line arguments
4019 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4020 and are retrieved from there when reconfiguring.
4024 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4026 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4029 *) The obj_dat.pl script was messing up the sorting of object
4030 names. The reason was that it compared the quoted version
4031 of strings as a result "OCSP" > "OCSP Signing" because
4032 " > SPACE. Changed script to store unquoted versions of
4033 names and add quotes on output. It was also omitting some
4034 names from the lookup table if they were given a default
4035 value (that is if SN is missing it is given the same
4036 value as LN and vice versa), these are now added on the
4037 grounds that if an object has a name we should be able to
4038 look it up. Finally added warning output when duplicate
4039 short or long names are found.
4042 *) Changes needed for Tandem NSK.
4043 [Scott Uroff <scott@xypro.com>]
4045 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4046 RSA_padding_check_SSLv23(), special padding was never detected
4047 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4048 version rollback attacks was not effective.
4050 In s23_clnt.c, don't use special rollback-attack detection padding
4051 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4052 client; similarly, in s23_srvr.c, don't do the rollback check if
4053 SSL 2.0 is the only protocol enabled in the server.
4056 *) Make it possible to get hexdumps of unprintable data with 'openssl
4057 asn1parse'. By implication, the functions ASN1_parse_dump() and
4058 BIO_dump_indent() are added.
4061 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4062 these print out strings and name structures based on various
4063 flags including RFC2253 support and proper handling of
4064 multibyte characters. Added options to the 'x509' utility
4065 to allow the various flags to be set.
4068 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4069 Also change the functions X509_cmp_current_time() and
4070 X509_gmtime_adj() work with an ASN1_TIME structure,
4071 this will enable certificates using GeneralizedTime in validity
4072 dates to be checked.
4075 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4076 negative public key encodings) on by default,
4077 NO_NEG_PUBKEY_BUG can be set to disable it.
4080 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4081 content octets. An i2c_ASN1_OBJECT is unnecessary because
4082 the encoding can be trivially obtained from the structure.
4085 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4086 not read locks (CRYPTO_r_[un]lock).
4089 *) A first attempt at creating official support for shared
4090 libraries through configuration. I've kept it so the
4091 default is static libraries only, and the OpenSSL programs
4092 are always statically linked for now, but there are
4093 preparations for dynamic linking in place.
4094 This has been tested on Linux and Tru64.
4097 *) Randomness polling function for Win9x, as described in:
4098 Peter Gutmann, Software Generation of Practically Strong
4102 *) Fix so PRNG is seeded in req if using an already existing
4106 *) New options to smime application. -inform and -outform
4107 allow alternative formats for the S/MIME message including
4108 PEM and DER. The -content option allows the content to be
4109 specified separately. This should allow things like Netscape
4110 form signing output easier to verify.
4113 *) Fix the ASN1 encoding of tags using the 'long form'.
4116 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4117 STRING types. These convert content octets to and from the
4118 underlying type. The actual tag and length octets are
4119 already assumed to have been read in and checked. These
4120 are needed because all other string types have virtually
4121 identical handling apart from the tag. By having versions
4122 of the ASN1 functions that just operate on content octets
4123 IMPLICIT tagging can be handled properly. It also allows
4124 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4125 and ASN1_INTEGER are identical apart from the tag.
4128 *) Change the handling of OID objects as follows:
4130 - New object identifiers are inserted in objects.txt, following
4131 the syntax given in objects.README.
4132 - objects.pl is used to process obj_mac.num and create a new
4134 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4137 This is currently kind of a hack, and the perl code in objects.pl
4138 isn't very elegant, but it works as I intended. The simplest way
4139 to check that it worked correctly is to look in obj_dat.h and
4140 check the array nid_objs and make sure the objects haven't moved
4141 around (this is important!). Additions are OK, as well as
4142 consistent name changes.
4145 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4148 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4149 The given file adds to whatever has already been seeded into the
4150 random pool through the RANDFILE configuration file option or
4151 environment variable, or the default random state file.
4154 *) mkstack.pl now sorts each macro group into lexical order.
4155 Previously the output order depended on the order the files
4156 appeared in the directory, resulting in needless rewriting
4160 *) Patches to make OpenSSL compile under Win32 again. Mostly
4161 work arounds for the VC++ problem that it treats func() as
4162 func(void). Also stripped out the parts of mkdef.pl that
4163 added extra typesafe functions: these no longer exist.
4166 *) Reorganisation of the stack code. The macros are now all
4167 collected in safestack.h . Each macro is defined in terms of
4168 a "stack macro" of the form SKM_<name>(type, a, b). The
4169 DEBUG_SAFESTACK is now handled in terms of function casts,
4170 this has the advantage of retaining type safety without the
4171 use of additional functions. If DEBUG_SAFESTACK is not defined
4172 then the non typesafe macros are used instead. Also modified the
4173 mkstack.pl script to handle the new form. Needs testing to see
4174 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4175 the default if no major problems. Similar behaviour for ASN1_SET_OF
4176 and PKCS12_STACK_OF.
4179 *) When some versions of IIS use the 'NET' form of private key the
4180 key derivation algorithm is different. Normally MD5(password) is
4181 used as a 128 bit RC4 key. In the modified case
4182 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4183 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4184 as the old Netscape_RSA functions except they have an additional
4185 'sgckey' parameter which uses the modified algorithm. Also added
4186 an -sgckey command line option to the rsa utility. Thanks to
4187 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4188 algorithm to openssl-dev.
4191 *) The evp_local.h macros were using 'c.##kname' which resulted in
4192 invalid expansion on some systems (SCO 5.0.5 for example).
4193 Corrected to 'c.kname'.
4194 [Phillip Porch <root@theporch.com>]
4196 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4197 a STACK of email addresses from a certificate or request, these look
4198 in the subject name and the subject alternative name extensions and
4199 omit any duplicate addresses.
4202 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4203 This makes DSA verification about 2 % faster.
4206 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4207 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4208 plus overhead for 1024 bit moduli).
4209 This makes exponentiations about 0.5 % faster for 1024 bit
4210 exponents (as measured by "openssl speed rsa2048").
4213 *) Rename memory handling macros to avoid conflicts with other
4215 Malloc => OPENSSL_malloc
4216 Malloc_locked => OPENSSL_malloc_locked
4217 Realloc => OPENSSL_realloc
4218 Free => OPENSSL_free
4221 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4222 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4225 *) CygWin32 support.
4226 [John Jarvie <jjarvie@newsguy.com>]
4228 *) The type-safe stack code has been rejigged. It is now only compiled
4229 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4230 by default all type-specific stack functions are "#define"d back to
4231 standard stack functions. This results in more streamlined output
4232 but retains the type-safety checking possibilities of the original
4236 *) The STACK code has been cleaned up, and certain type declarations
4237 that didn't make a lot of sense have been brought in line. This has
4238 also involved a cleanup of sorts in safestack.h to more correctly
4239 map type-safe stack functions onto their plain stack counterparts.
4240 This work has also resulted in a variety of "const"ifications of
4241 lots of the code, especially "_cmp" operations which should normally
4242 be prototyped with "const" parameters anyway.
4245 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4246 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4247 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4248 where all of 'md' is used each time the PRNG is used, but 'state'
4249 is used only indexed by a cyclic counter. As entropy may not be
4250 well distributed from the beginning, 'md' is important as a
4251 chaining variable. However, the output function chains only half
4252 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4253 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4254 in all of 'state' being rewritten, with the new values depending
4255 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4258 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4259 the handshake is continued after ssl_verify_cert_chain();
4260 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4261 can lead to 'unexplainable' connection aborts later.
4262 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4264 *) Major EVP API cipher revision.
4265 Add hooks for extra EVP features. This allows various cipher
4266 parameters to be set in the EVP interface. Support added for variable
4267 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4268 setting of RC2 and RC5 parameters.
4270 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4273 Remove lots of duplicated code from the EVP library. For example *every*
4274 cipher init() function handles the 'iv' in the same way according to the
4275 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4276 for CFB and OFB modes they zero ctx->num.
4278 New functionality allows removal of S/MIME code RC2 hack.
4280 Most of the routines have the same form and so can be declared in terms
4283 By shifting this to the top level EVP_CipherInit() it can be removed from
4284 all individual ciphers. If the cipher wants to handle IVs or keys
4285 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4288 Change lots of functions like EVP_EncryptUpdate() to now return a
4289 value: although software versions of the algorithms cannot fail
4290 any installed hardware versions can.
4293 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4294 this option is set, tolerate broken clients that send the negotiated
4295 protocol version number instead of the requested protocol version
4299 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4300 i.e. non-zero for export ciphersuites, zero otherwise.
4301 Previous versions had this flag inverted, inconsistent with
4302 rsa_tmp_cb (..._TMP_RSA_CB).
4303 [Bodo Moeller; problem reported by Amit Chopra]
4305 *) Add missing DSA library text string. Work around for some IIS
4306 key files with invalid SEQUENCE encoding.
4309 *) Add a document (doc/standards.txt) that list all kinds of standards
4310 and so on that are implemented in OpenSSL.
4313 *) Enhance c_rehash script. Old version would mishandle certificates
4314 with the same subject name hash and wouldn't handle CRLs at all.
4315 Added -fingerprint option to crl utility, to support new c_rehash
4319 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4322 *) Fix for SSL server purpose checking. Server checking was
4323 rejecting certificates which had extended key usage present
4324 but no ssl client purpose.
4325 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4327 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4328 is a little unclear about how a blank password is handled.
4329 Since the password in encoded as a BMPString with terminating
4330 double NULL a zero length password would end up as just the
4331 double NULL. However no password at all is different and is
4332 handled differently in the PKCS#12 key generation code. NS
4333 treats a blank password as zero length. MSIE treats it as no
4334 password on export: but it will try both on import. We now do
4335 the same: PKCS12_parse() tries zero length and no password if
4336 the password is set to "" or NULL (NULL is now a valid password:
4337 it wasn't before) as does the pkcs12 application.
4340 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4341 perror when PEM_read_bio_X509_REQ fails, the error message must
4342 be obtained from the error queue.
4345 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4346 it in ERR_remove_state if appropriate, and change ERR_get_state
4347 accordingly to avoid race conditions (this is necessary because
4348 thread_hash is no longer constant once set).
4351 *) Bugfix for linux-elf makefile.one.
4354 *) RSA_get_default_method() will now cause a default
4355 RSA_METHOD to be chosen if one doesn't exist already.
4356 Previously this was only set during a call to RSA_new()
4357 or RSA_new_method(NULL) meaning it was possible for
4358 RSA_get_default_method() to return NULL.
4361 *) Added native name translation to the existing DSO code
4362 that will convert (if the flag to do so is set) filenames
4363 that are sufficiently small and have no path information
4364 into a canonical native form. Eg. "blah" converted to
4365 "libblah.so" or "blah.dll" etc.
4368 *) New function ERR_error_string_n(e, buf, len) which is like
4369 ERR_error_string(e, buf), but writes at most 'len' bytes
4370 including the 0 terminator. For ERR_error_string_n, 'buf'
4372 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4374 *) CONF library reworked to become more general. A new CONF
4375 configuration file reader "class" is implemented as well as a
4376 new functions (NCONF_*, for "New CONF") to handle it. The now
4377 old CONF_* functions are still there, but are reimplemented to
4378 work in terms of the new functions. Also, a set of functions
4379 to handle the internal storage of the configuration data is
4380 provided to make it easier to write new configuration file
4381 reader "classes" (I can definitely see something reading a
4382 configuration file in XML format, for example), called _CONF_*,
4383 or "the configuration storage API"...
4385 The new configuration file reading functions are:
4387 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4388 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4390 NCONF_default, NCONF_WIN32
4392 NCONF_dump_fp, NCONF_dump_bio
4394 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4395 NCONF_new creates a new CONF object. This works in the same way
4396 as other interfaces in OpenSSL, like the BIO interface.
4397 NCONF_dump_* dump the internal storage of the configuration file,
4398 which is useful for debugging. All other functions take the same
4399 arguments as the old CONF_* functions wth the exception of the
4400 first that must be a `CONF *' instead of a `LHASH *'.
4402 To make it easer to use the new classes with the old CONF_* functions,
4403 the function CONF_set_default_method is provided.
4406 *) Add '-tls1' option to 'openssl ciphers', which was already
4407 mentioned in the documentation but had not been implemented.
4408 (This option is not yet really useful because even the additional
4409 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4412 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4413 OpenSSL-based applications) load shared libraries and bind to
4414 them in a portable way.
4415 [Geoff Thorpe, with contributions from Richard Levitte]
4417 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4419 *) Make sure _lrotl and _lrotr are only used with MSVC.
4421 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4422 (the default implementation of RAND_status).
4424 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4425 to '-clrext' (= clear extensions), as intended and documented.
4426 [Bodo Moeller; inconsistency pointed out by Michael Attili
4429 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4430 was larger than the MD block size.
4431 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4433 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4434 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4435 using the passed key: if the passed key was a private key the result
4436 of X509_print(), for example, would be to print out all the private key
4440 *) des_quad_cksum() byte order bug fix.
4441 [Ulf Möller, using the problem description in krb4-0.9.7, where
4442 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4444 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4446 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4448 *) For easily testing in shell scripts whether some command
4449 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4450 returns with exit code 0 iff no command of the given name is available.
4451 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4452 the output goes to stdout and nothing is printed to stderr.
4453 Additional arguments are always ignored.
4455 Since for each cipher there is a command of the same name,
4456 the 'no-cipher' compilation switches can be tested this way.
4458 ('openssl no-XXX' is not able to detect pseudo-commands such
4459 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4462 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4465 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4466 is set; it will be thrown away anyway because each handshake creates
4468 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4469 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4470 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4471 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4474 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4475 'Q' and 'R' lose their special meanings (quit/renegotiate).
4476 This is part of what -quiet does; unlike -quiet, -ign_eof
4477 does not suppress any output.
4480 *) Add compatibility options to the purpose and trust code. The
4481 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4482 accepts a certificate or CA, this was the previous behaviour,
4483 with all the associated security issues.
4485 X509_TRUST_COMPAT is the old trust behaviour: only and
4486 automatically trust self signed roots in certificate store. A
4487 new trust setting X509_TRUST_DEFAULT is used to specify that
4488 a purpose has no associated trust setting and it should instead
4489 use the value in the default purpose.
4492 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4493 and fix a memory leak.
4496 *) In util/mkerr.pl (which implements 'make errors'), preserve
4497 reason strings from the previous version of the .c file, as
4498 the default to have only downcase letters (and digits) in
4499 automatically generated reasons codes is not always appropriate.
4502 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4503 using strerror. Previously, ERR_reason_error_string() returned
4504 library names as reason strings for SYSerr; but SYSerr is a special
4505 case where small numbers are errno values, not library numbers.
4508 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4509 converts DSA parameters into DH parameters. (When creating parameters,
4510 DSA_generate_parameters is used.)
4513 *) Include 'length' (recommended exponent length) in C code generated
4514 by 'openssl dhparam -C'.
4517 *) The second argument to set_label in perlasm was already being used
4518 so couldn't be used as a "file scope" flag. Moved to third argument
4522 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4523 instead of RAND_bytes for encryption IVs and salts.
4526 *) Include RAND_status() into RAND_METHOD instead of implementing
4527 it only for md_rand.c Otherwise replacing the PRNG by calling
4528 RAND_set_rand_method would be impossible.
4531 *) Don't let DSA_generate_key() enter an infinite loop if the random
4532 number generation fails.
4535 *) New 'rand' application for creating pseudo-random output.
4538 *) Added configuration support for Linux/IA64
4539 [Rolf Haberrecker <rolf@suse.de>]
4541 *) Assembler module support for Mingw32.
4544 *) Shared library support for HPUX (in shlib/).
4545 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4547 *) Shared library support for Solaris gcc.
4548 [Lutz Behnke <behnke@trustcenter.de>]
4550 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4552 *) PKCS7_encrypt() was adding text MIME headers twice because they
4553 were added manually and by SMIME_crlf_copy().
4556 *) In bntest.c don't call BN_rand with zero bits argument.
4557 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4559 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4560 case was implemented. This caused BN_div_recp() to fail occasionally.
4563 *) Add an optional second argument to the set_label() in the perl
4564 assembly language builder. If this argument exists and is set
4565 to 1 it signals that the assembler should use a symbol whose
4566 scope is the entire file, not just the current function. This
4567 is needed with MASM which uses the format label:: for this scope.
4568 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4570 *) Change the ASN1 types so they are typedefs by default. Before
4571 almost all types were #define'd to ASN1_STRING which was causing
4572 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4576 *) Change names of new functions to the new get1/get0 naming
4577 convention: After 'get1', the caller owns a reference count
4578 and has to call ..._free; 'get0' returns a pointer to some
4579 data structure without incrementing reference counters.
4580 (Some of the existing 'get' functions increment a reference
4581 counter, some don't.)
4582 Similarly, 'set1' and 'add1' functions increase reference
4583 counters or duplicate objects.
4586 *) Allow for the possibility of temp RSA key generation failure:
4587 the code used to assume it always worked and crashed on failure.
4590 *) Fix potential buffer overrun problem in BIO_printf().
4591 [Ulf Möller, using public domain code by Patrick Powell; problem
4592 pointed out by David Sacerdote <das33@cornell.edu>]
4594 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4595 RAND_egd() and RAND_status(). In the command line application,
4596 the EGD socket can be specified like a seed file using RANDFILE
4600 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4601 Some CAs (e.g. Verisign) distribute certificates in this form.
4604 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4605 list to exclude them. This means that no special compilation option
4606 is needed to use anonymous DH: it just needs to be included in the
4610 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4611 EVP_MD_type. The old functionality is available in a new macro called
4612 EVP_MD_md(). Change code that uses it and update docs.
4615 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4616 where the 'void *' argument is replaced by a function pointer argument.
4617 Previously 'void *' was abused to point to functions, which works on
4618 many platforms, but is not correct. As these functions are usually
4619 called by macros defined in OpenSSL header files, most source code
4620 should work without changes.
4623 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4624 sections with information on -D... compiler switches used for
4625 compiling the library so that applications can see them. To enable
4626 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4627 must be defined. E.g.,
4628 #define OPENSSL_ALGORITHM_DEFINES
4629 #include <openssl/opensslconf.h>
4630 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4631 [Richard Levitte, Ulf and Bodo Möller]
4633 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4637 *) Change the 'other' type in certificate aux info to a STACK_OF
4638 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4639 the required ASN1 format: arbitrary types determined by an OID.
4642 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4643 argument to 'req'. This is not because the function is newer or
4644 better than others it just uses the work 'NEW' in the certificate
4645 request header lines. Some software needs this.
4648 *) Reorganise password command line arguments: now passwords can be
4649 obtained from various sources. Delete the PEM_cb function and make
4650 it the default behaviour: i.e. if the callback is NULL and the
4651 usrdata argument is not NULL interpret it as a null terminated pass
4652 phrase. If usrdata and the callback are NULL then the pass phrase
4653 is prompted for as usual.
4656 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4657 the support is automatically enabled. The resulting binaries will
4658 autodetect the card and use it if present.
4659 [Ben Laurie and Compaq Inc.]
4661 *) Work around for Netscape hang bug. This sends certificate request
4662 and server done in one record. Since this is perfectly legal in the
4663 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4664 the bugs/SSLv3 entry for more info.
4667 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4670 *) Add -rand argument to smime and pkcs12 applications and read/write
4674 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4677 *) Add command line password options to the remaining applications.
4680 *) Bug fix for BN_div_recp() for numerators with an even number of
4684 *) More tests in bntest.c, and changed test_bn output.
4687 *) ./config recognizes MacOS X now.
4690 *) Bug fix for BN_div() when the first words of num and divsor are
4691 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4694 *) Add support for various broken PKCS#8 formats, and command line
4695 options to produce them.
4698 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4699 get temporary BIGNUMs from a BN_CTX.
4702 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4706 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4707 include a #define from the old name to the new. The original intent
4708 was that statically linked binaries could for example just call
4709 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4710 link with digests. This never worked becayse SSLeay_add_all_digests()
4711 and SSLeay_add_all_ciphers() were in the same source file so calling
4712 one would link with the other. They are now in separate source files.
4715 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4718 *) Use a less unusual form of the Miller-Rabin primality test (it used
4719 a binary algorithm for exponentiation integrated into the Miller-Rabin
4720 loop, our standard modexp algorithms are faster).
4723 *) Support for the EBCDIC character set completed.
4724 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4726 *) Source code cleanups: use const where appropriate, eliminate casts,
4727 use void * instead of char * in lhash.
4730 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4731 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4732 this the server could overwrite ephemeral keys that the client
4736 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4737 using 50 iterations of the Rabin-Miller test.
4739 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4740 iterations of the Rabin-Miller test as required by the appendix
4741 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4742 As BN_is_prime_fasttest includes trial division, DSA parameter
4743 generation becomes much faster.
4745 This implies a change for the callback functions in DSA_is_prime
4746 and DSA_generate_parameters: The callback function is called once
4747 for each positive witness in the Rabin-Miller test, not just
4748 occasionally in the inner loop; and the parameters to the
4749 callback function now provide an iteration count for the outer
4750 loop rather than for the current invocation of the inner loop.
4751 DSA_generate_parameters additionally can call the callback
4752 function with an 'iteration count' of -1, meaning that a
4753 candidate has passed the trial division test (when q is generated
4754 from an application-provided seed, trial division is skipped).
4757 *) New function BN_is_prime_fasttest that optionally does trial
4758 division before starting the Rabin-Miller test and has
4759 an additional BN_CTX * argument (whereas BN_is_prime always
4760 has to allocate at least one BN_CTX).
4761 'callback(1, -1, cb_arg)' is called when a number has passed the
4762 trial division stage.
4765 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4769 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4772 *) New function BN_pseudo_rand().
4775 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4776 bignum version of BN_from_montgomery() with the working code from
4777 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4781 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4782 made it impossible to use the same SSL_SESSION data structure in
4783 SSL2 clients in multiple threads.
4786 *) The return value of RAND_load_file() no longer counts bytes obtained
4787 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4788 to seed the PRNG (previously an explicit byte count was required).
4789 [Ulf Möller, Bodo Möller]
4791 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4792 used (char *) instead of (void *) and had casts all over the place.
4795 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4798 *) Retain source code compatibility for BN_prime_checks macro:
4799 BN_is_prime(..., BN_prime_checks, ...) now uses
4800 BN_prime_checks_for_size to determine the appropriate number of
4801 Rabin-Miller iterations.
4804 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4805 DH_CHECK_P_NOT_SAFE_PRIME.
4806 (Check if this is true? OpenPGP calls them "strong".)
4809 *) Merge the functionality of "dh" and "gendh" programs into a new program
4810 "dhparam". The old programs are retained for now but will handle DH keys
4811 (instead of parameters) in future.
4814 *) Make the ciphers, s_server and s_client programs check the return values
4815 when a new cipher list is set.
4818 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4819 ciphers. Before when the 56bit ciphers were enabled the sorting was
4822 The syntax for the cipher sorting has been extended to support sorting by
4823 cipher-strength (using the strength_bits hard coded in the tables).
4824 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4826 Fix a bug in the cipher-command parser: when supplying a cipher command
4827 string with an "undefined" symbol (neither command nor alphanumeric
4828 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4829 an error is flagged.
4831 Due to the strength-sorting extension, the code of the
4832 ssl_create_cipher_list() function was completely rearranged. I hope that
4833 the readability was also increased :-)
4834 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4836 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4837 for the first serial number and places 2 in the serial number file. This
4838 avoids problems when the root CA is created with serial number zero and
4839 the first user certificate has the same issuer name and serial number
4843 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4844 the new code. Add documentation for this stuff.
4847 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4848 X509_*() to X509at_*() on the grounds that they don't handle X509
4849 structures and behave in an analagous way to the X509v3 functions:
4850 they shouldn't be called directly but wrapper functions should be used
4853 So we also now have some wrapper functions that call the X509at functions
4854 when passed certificate requests. (TO DO: similar things can be done with
4855 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4856 things. Some of these need some d2i or i2d and print functionality
4857 because they handle more complex structures.)
4860 *) Add missing #ifndefs that caused missing symbols when building libssl
4861 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4862 NO_RSA in ssl/s2*.c.
4863 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4865 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4866 has a return value which indicates the quality of the random data
4867 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4868 error queue. New function RAND_pseudo_bytes() generates output that is
4869 guaranteed to be unique but not unpredictable. RAND_add is like
4870 RAND_seed, but takes an extra argument for an entropy estimate
4871 (RAND_seed always assumes full entropy).
4874 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4875 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4876 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4877 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4878 false-positive rate of at most 2^-80 for random input.
4881 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4884 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4885 in the 0.9.5 release), this returns the chain
4886 from an X509_CTX structure with a dup of the stack and all
4887 the X509 reference counts upped: so the stack will exist
4888 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4891 Also make SSL_SESSION_print() print out the verify return
4895 *) Add manpage for the pkcs12 command. Also change the default
4896 behaviour so MAC iteration counts are used unless the new
4897 -nomaciter option is used. This improves file security and
4898 only older versions of MSIE (4.0 for example) need it.
4901 *) Honor the no-xxx Configure options when creating .DEF files.
4904 *) Add PKCS#10 attributes to field table: challengePassword,
4905 unstructuredName and unstructuredAddress. These are taken from
4906 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4907 international characters are used.
4909 More changes to X509_ATTRIBUTE code: allow the setting of types
4910 based on strings. Remove the 'loc' parameter when adding
4911 attributes because these will be a SET OF encoding which is sorted
4915 *) Initial changes to the 'req' utility to allow request generation
4916 automation. This will allow an application to just generate a template
4917 file containing all the field values and have req construct the
4920 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4921 used all over the place including certificate requests and PKCS#7
4922 structures. They are currently handled manually where necessary with
4923 some primitive wrappers for PKCS#7. The new functions behave in a
4924 manner analogous to the X509 extension functions: they allow
4925 attributes to be looked up by NID and added.
4927 Later something similar to the X509V3 code would be desirable to
4928 automatically handle the encoding, decoding and printing of the
4929 more complex types. The string types like challengePassword can
4930 be handled by the string table functions.
4932 Also modified the multi byte string table handling. Now there is
4933 a 'global mask' which masks out certain types. The table itself
4934 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4935 is useful when for example there is only one permissible type
4936 (as in countryName) and using the mask might result in no valid
4940 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4941 SSL_get_peer_finished to allow applications to obtain the latest
4942 Finished messages sent to the peer or expected from the peer,
4943 respectively. (SSL_get_peer_finished is usually the Finished message
4944 actually received from the peer, otherwise the protocol will be aborted.)
4946 As the Finished message are message digests of the complete handshake
4947 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4948 be used for external authentication procedures when the authentication
4949 provided by SSL/TLS is not desired or is not enough.
4952 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4953 the host supports BWX extension and if Compaq C is present on the
4954 $PATH. Just exploiting of the BWX extension results in 20-30%
4955 performance kick for some algorithms, e.g. DES and RC4 to mention
4956 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4960 *) Add support for MS "fast SGC". This is arguably a violation of the
4961 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4962 weak crypto and after checking the certificate is SGC a second one
4963 with strong crypto. MS SGC stops the first handshake after receiving
4964 the server certificate message and sends a second client hello. Since
4965 a server will typically do all the time consuming operations before
4966 expecting any further messages from the client (server key exchange
4967 is the most expensive) there is little difference between the two.
4969 To get OpenSSL to support MS SGC we have to permit a second client
4970 hello message after we have sent server done. In addition we have to
4971 reset the MAC if we do get this second client hello.
4974 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4975 if a DER encoded private key is RSA or DSA traditional format. Changed
4976 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4977 format DER encoded private key. Newer code should use PKCS#8 format which
4978 has the key type encoded in the ASN1 structure. Added DER private key
4979 support to pkcs8 application.
4982 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4983 ciphersuites has been selected (as required by the SSL 3/TLS 1
4984 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4985 is set, we interpret this as a request to violate the specification
4986 (the worst that can happen is a handshake failure, and 'correct'
4987 behaviour would result in a handshake failure anyway).
4990 *) In SSL_CTX_add_session, take into account that there might be multiple
4991 SSL_SESSION structures with the same session ID (e.g. when two threads
4992 concurrently obtain them from an external cache).
4993 The internal cache can handle only one SSL_SESSION with a given ID,
4994 so if there's a conflict, we now throw out the old one to achieve
4998 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4999 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5000 some routines that use cipher OIDs: some ciphers do not have OIDs
5001 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5005 *) Simplify the trust setting structure and code. Now we just have
5006 two sequences of OIDs for trusted and rejected settings. These will
5007 typically have values the same as the extended key usage extension
5008 and any application specific purposes.
5010 The trust checking code now has a default behaviour: it will just
5011 check for an object with the same NID as the passed id. Functions can
5012 be provided to override either the default behaviour or the behaviour
5013 for a given id. SSL client, server and email already have functions
5014 in place for compatibility: they check the NID and also return "trusted"
5015 if the certificate is self signed.
5018 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5019 traditional format into an EVP_PKEY structure.
5022 *) Add a password callback function PEM_cb() which either prompts for
5023 a password if usr_data is NULL or otherwise assumes it is a null
5024 terminated password. Allow passwords to be passed on command line
5025 environment or config files in a few more utilities.
5028 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5029 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5030 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5031 Update documentation.
5034 *) Support for ASN1 "NULL" type. This could be handled before by using
5035 ASN1_TYPE but there wasn't any function that would try to read a NULL
5036 and produce an error if it couldn't. For compatibility we also have
5037 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5038 don't allocate anything because they don't need to.
5041 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5043 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5045 *) Rebuild of the memory allocation routines used by OpenSSL code and
5046 possibly others as well. The purpose is to make an interface that
5047 provide hooks so anyone can build a separate set of allocation and
5048 deallocation routines to be used by OpenSSL, for example memory
5049 pool implementations, or something else, which was previously hard
5050 since Malloc(), Realloc() and Free() were defined as macros having
5051 the values malloc, realloc and free, respectively (except for Win32
5052 compilations). The same is provided for memory debugging code.
5053 OpenSSL already comes with functionality to find memory leaks, but
5054 this gives people a chance to debug other memory problems.
5056 With these changes, a new set of functions and macros have appeared:
5058 CRYPTO_set_mem_debug_functions() [F]
5059 CRYPTO_get_mem_debug_functions() [F]
5060 CRYPTO_dbg_set_options() [F]
5061 CRYPTO_dbg_get_options() [F]
5062 CRYPTO_malloc_debug_init() [M]
5064 The memory debug functions are NULL by default, unless the library
5065 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5066 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5067 gives the standard debugging functions that come with OpenSSL) or
5068 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5069 provided by the library user) must be used. When the standard
5070 debugging functions are used, CRYPTO_dbg_set_options can be used to
5071 request additional information:
5072 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5073 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5075 Also, things like CRYPTO_set_mem_functions will always give the
5076 expected result (the new set of functions is used for allocation
5077 and deallocation) at all times, regardless of platform and compiler
5080 To finish it up, some functions that were never use in any other
5081 way than through macros have a new API and new semantic:
5084 CRYPTO_dbg_realloc()
5087 All macros of value have retained their old syntax.
5088 [Richard Levitte and Bodo Moeller]
5090 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5091 ordering of SMIMECapabilities wasn't in "strength order" and there
5092 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5096 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5097 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5098 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5100 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5101 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5102 functionality to handle multipart/signed properly) and a utility
5103 called 'smime' to call all this stuff. This is based on code I
5104 originally wrote for Celo who have kindly allowed it to be
5105 included in OpenSSL.
5108 *) Add variants des_set_key_checked and des_set_key_unchecked of
5109 des_set_key (aka des_key_sched). Global variable des_check_key
5110 decides which of these is called by des_set_key; this way
5111 des_check_key behaves as it always did, but applications and
5112 the library itself, which was buggy for des_check_key == 1,
5113 have a cleaner way to pick the version they need.
5116 *) New function PKCS12_newpass() which changes the password of a
5120 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5121 dynamic mix. In both cases the ids can be used as an index into the
5122 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5123 functions so they accept a list of the field values and the
5124 application doesn't need to directly manipulate the X509_TRUST
5128 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5132 *) Modify the way the V3 extension code looks up extensions. This now
5133 works in a similar way to the object code: we have some "standard"
5134 extensions in a static table which is searched with OBJ_bsearch()
5135 and the application can add dynamic ones if needed. The file
5136 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5137 updated whenever a new extension is added to the core code and kept
5138 in ext_nid order. There is a simple program 'tabtest.c' which checks
5139 this. New extensions are not added too often so this file can readily
5140 be maintained manually.
5142 There are two big advantages in doing things this way. The extensions
5143 can be looked up immediately and no longer need to be "added" using
5144 X509V3_add_standard_extensions(): this function now does nothing.
5145 [Side note: I get *lots* of email saying the extension code doesn't
5146 work because people forget to call this function]
5147 Also no dynamic allocation is done unless new extensions are added:
5148 so if we don't add custom extensions there is no need to call
5149 X509V3_EXT_cleanup().
5152 *) Modify enc utility's salting as follows: make salting the default. Add a
5153 magic header, so unsalted files fail gracefully instead of just decrypting
5154 to garbage. This is because not salting is a big security hole, so people
5155 should be discouraged from doing it.
5158 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5159 digest to be passed on the command line but it only used this
5160 parameter when signing a certificate. Modified so all relevant
5161 operations are affected by the digest parameter including the
5162 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5163 DSA key was used because it didn't fix the digest.
5166 *) Initial certificate chain verify code. Currently tests the untrusted
5167 certificates for consistency with the verify purpose (which is set
5168 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5170 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5171 this is because it will reject chains with invalid extensions whereas
5172 every previous version of OpenSSL and SSLeay made no checks at all.
5174 Trust code: checks the root CA for the relevant trust settings. Trust
5175 settings have an initial value consistent with the verify purpose: e.g.
5176 if the verify purpose is for SSL client use it expects the CA to be
5177 trusted for SSL client use. However the default value can be changed to
5178 permit custom trust settings: one example of this would be to only trust
5179 certificates from a specific "secure" set of CAs.
5181 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5182 which should be used for version portability: especially since the
5183 verify structure is likely to change more often now.
5185 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5186 to set them. If not set then assume SSL clients will verify SSL servers
5189 Two new options to the verify program: -untrusted allows a set of
5190 untrusted certificates to be passed in and -purpose which sets the
5191 intended purpose of the certificate. If a purpose is set then the
5192 new chain verify code is used to check extension consistency.
5195 *) Support for the authority information access extension.
5198 *) Modify RSA and DSA PEM read routines to transparently handle
5199 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5200 public keys in a format compatible with certificate
5201 SubjectPublicKeyInfo structures. Unfortunately there were already
5202 functions called *_PublicKey_* which used various odd formats so
5203 these are retained for compatibility: however the DSA variants were
5204 never in a public release so they have been deleted. Changed dsa/rsa
5205 utilities to handle the new format: note no releases ever handled public
5206 keys so we should be OK.
5208 The primary motivation for this change is to avoid the same fiasco
5209 that dogs private keys: there are several incompatible private key
5210 formats some of which are standard and some OpenSSL specific and
5211 require various evil hacks to allow partial transparent handling and
5212 even then it doesn't work with DER formats. Given the option anything
5213 other than PKCS#8 should be dumped: but the other formats have to
5214 stay in the name of compatibility.
5216 With public keys and the benefit of hindsight one standard format
5217 is used which works with EVP_PKEY, RSA or DSA structures: though
5218 it clearly returns an error if you try to read the wrong kind of key.
5220 Added a -pubkey option to the 'x509' utility to output the public key.
5221 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5222 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5223 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5224 that do the same as the EVP_PKEY_assign_*() except they up the
5225 reference count of the added key (they don't "swallow" the
5229 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5230 CRLs would fail if the file contained no certificates or no CRLs:
5231 added a new function to read in both types and return the number
5232 read: this means that if none are read it will be an error. The
5233 DER versions of the certificate and CRL reader would always fail
5234 because it isn't possible to mix certificates and CRLs in DER format
5235 without choking one or the other routine. Changed this to just read
5236 a certificate: this is the best we can do. Also modified the code
5237 in apps/verify.c to take notice of return codes: it was previously
5238 attempting to read in certificates from NULL pointers and ignoring
5239 any errors: this is one reason why the cert and CRL reader seemed
5240 to work. It doesn't check return codes from the default certificate
5241 routines: these may well fail if the certificates aren't installed.
5244 *) Code to support otherName option in GeneralName.
5247 *) First update to verify code. Change the verify utility
5248 so it warns if it is passed a self signed certificate:
5249 for consistency with the normal behaviour. X509_verify
5250 has been modified to it will now verify a self signed
5251 certificate if *exactly* the same certificate appears
5252 in the store: it was previously impossible to trust a
5253 single self signed certificate. This means that:
5254 openssl verify ss.pem
5255 now gives a warning about a self signed certificate but
5256 openssl verify -CAfile ss.pem ss.pem
5260 *) For servers, store verify_result in SSL_SESSION data structure
5261 (and add it to external session representation).
5262 This is needed when client certificate verifications fails,
5263 but an application-provided verification callback (set by
5264 SSL_CTX_set_cert_verify_callback) allows accepting the session
5265 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5266 but returns 1): When the session is reused, we have to set
5267 ssl->verify_result to the appropriate error code to avoid
5269 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5271 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5272 case in PKCS7_dataInit() where the signed PKCS7 structure
5273 didn't contain any existing data because it was being created.
5274 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5276 *) Add a salt to the key derivation routines in enc.c. This
5277 forms the first 8 bytes of the encrypted file. Also add a
5278 -S option to allow a salt to be input on the command line.
5281 *) New function X509_cmp(). Oddly enough there wasn't a function
5282 to compare two certificates. We do this by working out the SHA1
5283 hash and comparing that. X509_cmp() will be needed by the trust
5287 *) SSL_get1_session() is like SSL_get_session(), but increments
5288 the reference count in the SSL_SESSION returned.
5289 [Geoff Thorpe <geoff@eu.c2.net>]
5291 *) Fix for 'req': it was adding a null to request attributes.
5292 Also change the X509_LOOKUP and X509_INFO code to handle
5293 certificate auxiliary information.
5296 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5300 *) Add the possibility to add extra information to the memory leak
5301 detecting output, to form tracebacks, showing from where each
5302 allocation was originated: CRYPTO_push_info("constant string") adds
5303 the string plus current file name and line number to a per-thread
5304 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5305 is like calling CYRPTO_pop_info() until the stack is empty.
5306 Also updated memory leak detection code to be multi-thread-safe.
5309 *) Add options -text and -noout to pkcs7 utility and delete the
5310 encryption options which never did anything. Update docs.
5313 *) Add options to some of the utilities to allow the pass phrase
5314 to be included on either the command line (not recommended on
5315 OSes like Unix) or read from the environment. Update the
5316 manpages and fix a few bugs.
5319 *) Add a few manpages for some of the openssl commands.
5322 *) Fix the -revoke option in ca. It was freeing up memory twice,
5323 leaking and not finding already revoked certificates.
5326 *) Extensive changes to support certificate auxiliary information.
5327 This involves the use of X509_CERT_AUX structure and X509_AUX
5328 functions. An X509_AUX function such as PEM_read_X509_AUX()
5329 can still read in a certificate file in the usual way but it
5330 will also read in any additional "auxiliary information". By
5331 doing things this way a fair degree of compatibility can be
5332 retained: existing certificates can have this information added
5333 using the new 'x509' options.
5335 Current auxiliary information includes an "alias" and some trust
5336 settings. The trust settings will ultimately be used in enhanced
5337 certificate chain verification routines: currently a certificate
5338 can only be trusted if it is self signed and then it is trusted
5342 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5343 The problem was that one of the replacement routines had not been working
5344 since SSLeay releases. For now the offending routine has been replaced
5345 with non-optimised assembler. Even so, this now gives around 95%
5346 performance improvement for 1024 bit RSA signs.
5349 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5350 handling. Most clients have the effective key size in bits equal to
5351 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5352 A few however don't do this and instead use the size of the decrypted key
5353 to determine the RC2 key length and the AlgorithmIdentifier to determine
5354 the effective key length. In this case the effective key length can still
5355 be 40 bits but the key length can be 168 bits for example. This is fixed
5356 by manually forcing an RC2 key into the EVP_PKEY structure because the
5357 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5358 the key length and effective key length are equal.
5361 *) Add a bunch of functions that should simplify the creation of
5362 X509_NAME structures. Now you should be able to do:
5363 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5364 and have it automatically work out the correct field type and fill in
5365 the structures. The more adventurous can try:
5366 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5367 and it will (hopefully) work out the correct multibyte encoding.
5370 *) Change the 'req' utility to use the new field handling and multibyte
5371 copy routines. Before the DN field creation was handled in an ad hoc
5372 way in req, ca, and x509 which was rather broken and didn't support
5373 BMPStrings or UTF8Strings. Since some software doesn't implement
5374 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5375 using the dirstring_type option. See the new comment in the default
5376 openssl.cnf for more info.
5379 *) Make crypto/rand/md_rand.c more robust:
5380 - Assure unique random numbers after fork().
5381 - Make sure that concurrent threads access the global counter and
5382 md serializably so that we never lose entropy in them
5383 or use exactly the same state in multiple threads.
5384 Access to the large state is not always serializable because
5385 the additional locking could be a performance killer, and
5386 md should be large enough anyway.
5389 *) New file apps/app_rand.c with commonly needed functionality
5390 for handling the random seed file.
5392 Use the random seed file in some applications that previously did not:
5394 dsaparam -genkey (which also ignored its '-rand' option),
5397 x509 (when signing).
5398 Except on systems with /dev/urandom, it is crucial to have a random
5399 seed file at least for key creation, DSA signing, and for DH exchanges;
5400 for RSA signatures we could do without one.
5402 gendh and gendsa (unlike genrsa) used to read only the first byte
5403 of each file listed in the '-rand' option. The function as previously
5404 found in genrsa is now in app_rand.c and is used by all programs
5405 that support '-rand'.
5408 *) In RAND_write_file, use mode 0600 for creating files;
5409 don't just chmod when it may be too late.
5412 *) Report an error from X509_STORE_load_locations
5413 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5416 *) New function ASN1_mbstring_copy() this copies a string in either
5417 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5418 into an ASN1_STRING type. A mask of permissible types is passed
5419 and it chooses the "minimal" type to use or an error if not type
5423 *) Add function equivalents to the various macros in asn1.h. The old
5424 macros are retained with an M_ prefix. Code inside the library can
5425 use the M_ macros. External code (including the openssl utility)
5426 should *NOT* in order to be "shared library friendly".
5429 *) Add various functions that can check a certificate's extensions
5430 to see if it usable for various purposes such as SSL client,
5431 server or S/MIME and CAs of these types. This is currently
5432 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5433 verification. Also added a -purpose flag to x509 utility to
5434 print out all the purposes.
5437 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5441 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5442 for, obtain and decode and extension and obtain its critical flag.
5443 This allows all the necessary extension code to be handled in a
5444 single function call.
5447 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5448 platforms. See crypto/rc4/rc4_enc.c for further details.
5451 *) New -noout option to asn1parse. This causes no output to be produced
5452 its main use is when combined with -strparse and -out to extract data
5453 from a file (which may not be in ASN.1 format).
5456 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5457 when producing the local key id.
5458 [Richard Levitte <levitte@stacken.kth.se>]
5460 *) New option -dhparam in s_server. This allows a DH parameter file to be
5461 stated explicitly. If it is not stated then it tries the first server
5462 certificate file. The previous behaviour hard coded the filename
5466 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5467 a public key to be input or output. For example:
5468 openssl rsa -in key.pem -pubout -out pubkey.pem
5469 Also added necessary DSA public key functions to handle this.
5472 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5473 in the message. This was handled by allowing
5474 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5475 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5477 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5478 to the end of the strings whereas this didn't. This would cause problems
5479 if strings read with d2i_ASN1_bytes() were later modified.
5480 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5482 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5483 data and it contains EOF it will end up returning an error. This is
5484 caused by input 46 bytes long. The cause is due to the way base64
5485 BIOs find the start of base64 encoded data. They do this by trying a
5486 trial decode on each line until they find one that works. When they
5487 do a flag is set and it starts again knowing it can pass all the
5488 data directly through the decoder. Unfortunately it doesn't reset
5489 the context it uses. This means that if EOF is reached an attempt
5490 is made to pass two EOFs through the context and this causes the
5491 resulting error. This can also cause other problems as well. As is
5492 usual with these problems it takes *ages* to find and the fix is
5493 trivial: move one line.
5494 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5496 *) Ugly workaround to get s_client and s_server working under Windows. The
5497 old code wouldn't work because it needed to select() on sockets and the
5498 tty (for keypresses and to see if data could be written). Win32 only
5499 supports select() on sockets so we select() with a 1s timeout on the
5500 sockets and then see if any characters are waiting to be read, if none
5501 are present then we retry, we also assume we can always write data to
5502 the tty. This isn't nice because the code then blocks until we've
5503 received a complete line of data and it is effectively polling the
5504 keyboard at 1s intervals: however it's quite a bit better than not
5505 working at all :-) A dedicated Windows application might handle this
5506 with an event loop for example.
5509 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5510 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5511 will be called when RSA_sign() and RSA_verify() are used. This is useful
5512 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5513 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5514 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5515 This necessitated the support of an extra signature type NID_md5_sha1
5516 for SSL signatures and modifications to the SSL library to use it instead
5517 of calling RSA_public_decrypt() and RSA_private_encrypt().
5520 *) Add new -verify -CAfile and -CApath options to the crl program, these
5521 will lookup a CRL issuers certificate and verify the signature in a
5522 similar way to the verify program. Tidy up the crl program so it
5523 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5524 less strict. It will now permit CRL extensions even if it is not
5525 a V2 CRL: this will allow it to tolerate some broken CRLs.
5528 *) Initialize all non-automatic variables each time one of the openssl
5529 sub-programs is started (this is necessary as they may be started
5530 multiple times from the "OpenSSL>" prompt).
5531 [Lennart Bang, Bodo Moeller]
5533 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5534 removing all other RSA functionality (this is what NO_RSA does). This
5535 is so (for example) those in the US can disable those operations covered
5536 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5540 *) Non-copying interface to BIO pairs.
5541 (still largely untested)
5544 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5545 ASCII string. This was handled independently in various places before.
5548 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5549 UTF8 strings a character at a time.
5552 *) Use client_version from client hello to select the protocol
5553 (s23_srvr.c) and for RSA client key exchange verification
5554 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5557 *) Add various utility functions to handle SPKACs, these were previously
5558 handled by poking round in the structure internals. Added new function
5559 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5560 print, verify and generate SPKACs. Based on an original idea from
5561 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5564 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5567 *) Allow the config file extension section to be overwritten on the
5568 command line. Based on an original idea from Massimiliano Pala
5569 <madwolf@comune.modena.it>. The new option is called -extensions
5570 and can be applied to ca, req and x509. Also -reqexts to override
5571 the request extensions in req and -crlexts to override the crl extensions
5575 *) Add new feature to the SPKAC handling in ca. Now you can include
5576 the same field multiple times by preceding it by "XXXX." for example:
5579 this is the same syntax as used in the req config file.
5582 *) Allow certificate extensions to be added to certificate requests. These
5583 are specified in a 'req_extensions' option of the req section of the
5584 config file. They can be printed out with the -text option to req but
5585 are otherwise ignored at present.
5588 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5589 data read consists of only the final block it would not decrypted because
5590 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5591 A misplaced 'break' also meant the decrypted final block might not be
5592 copied until the next read.
5595 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5596 a few extra parameters to the DH structure: these will be useful if
5597 for example we want the value of 'q' or implement X9.42 DH.
5600 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5601 provides hooks that allow the default DSA functions or functions on a
5602 "per key" basis to be replaced. This allows hardware acceleration and
5603 hardware key storage to be handled without major modification to the
5604 library. Also added low level modexp hooks and CRYPTO_EX structure and
5605 associated functions.
5608 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5609 as "read only": it can't be written to and the buffer it points to will
5610 not be freed. Reading from a read only BIO is much more efficient than
5611 a normal memory BIO. This was added because there are several times when
5612 an area of memory needs to be read from a BIO. The previous method was
5613 to create a memory BIO and write the data to it, this results in two
5614 copies of the data and an O(n^2) reading algorithm. There is a new
5615 function BIO_new_mem_buf() which creates a read only memory BIO from
5616 an area of memory. Also modified the PKCS#7 routines to use read only
5620 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5621 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5622 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5623 but a retry condition occured while trying to read the rest.
5626 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5627 NID_pkcs7_encrypted by default: this was wrong since this should almost
5628 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5629 the encrypted data type: this is a more sensible place to put it and it
5630 allows the PKCS#12 code to be tidied up that duplicated this
5634 *) Changed obj_dat.pl script so it takes its input and output files on
5635 the command line. This should avoid shell escape redirection problems
5639 *) Initial support for certificate extension requests, these are included
5640 in things like Xenroll certificate requests. Included functions to allow
5641 extensions to be obtained and added.
5644 *) -crlf option to s_client and s_server for sending newlines as
5645 CRLF (as required by many protocols).
5648 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5650 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5651 [Ralf S. Engelschall]
5653 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5654 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5656 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5660 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5661 DH parameters/keys (q is lost during that conversion, but the resulting
5662 DH parameters contain its length).
5664 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5665 much faster than DH_generate_parameters (which creates parameters
5666 where p = 2*q + 1), and also the smaller q makes DH computations
5667 much more efficient (160-bit exponentiation instead of 1024-bit
5668 exponentiation); so this provides a convenient way to support DHE
5669 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5670 utter importance to use
5671 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5673 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5674 when such DH parameters are used, because otherwise small subgroup
5675 attacks may become possible!
5678 *) Avoid memory leak in i2d_DHparams.
5681 *) Allow the -k option to be used more than once in the enc program:
5682 this allows the same encrypted message to be read by multiple recipients.
5685 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5686 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5687 it will always use the numerical form of the OID, even if it has a short
5691 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5692 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5693 otherwise bn_mod_exp was called. In the case of hardware keys for example
5694 no private key components need be present and it might store extra data
5695 in the RSA structure, which cannot be accessed from bn_mod_exp.
5696 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5697 private key operations.
5700 *) Added support for SPARC Linux.
5703 *) pem_password_cb function type incompatibly changed from
5704 typedef int pem_password_cb(char *buf, int size, int rwflag);
5706 ....(char *buf, int size, int rwflag, void *userdata);
5707 so that applications can pass data to their callbacks:
5708 The PEM[_ASN1]_{read,write}... functions and macros now take an
5709 additional void * argument, which is just handed through whenever
5710 the password callback is called.
5711 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5713 New function SSL_CTX_set_default_passwd_cb_userdata.
5715 Compatibility note: As many C implementations push function arguments
5716 onto the stack in reverse order, the new library version is likely to
5717 interoperate with programs that have been compiled with the old
5718 pem_password_cb definition (PEM_whatever takes some data that
5719 happens to be on the stack as its last argument, and the callback
5720 just ignores this garbage); but there is no guarantee whatsoever that
5723 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5724 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5725 problems not only on Windows, but also on some Unix platforms.
5726 To avoid problematic command lines, these definitions are now in an
5727 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5728 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5731 *) MIPS III/IV assembler module is reimplemented.
5734 *) More DES library cleanups: remove references to srand/rand and
5735 delete an unused file.
5738 *) Add support for the the free Netwide assembler (NASM) under Win32,
5739 since not many people have MASM (ml) and it can be hard to obtain.
5740 This is currently experimental but it seems to work OK and pass all
5741 the tests. Check out INSTALL.W32 for info.
5744 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5745 without temporary keys kept an extra copy of the server key,
5746 and connections with temporary keys did not free everything in case
5750 *) New function RSA_check_key and new openssl rsa option -check
5751 for verifying the consistency of RSA keys.
5752 [Ulf Moeller, Bodo Moeller]
5754 *) Various changes to make Win32 compile work:
5755 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5756 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5757 comparison" warnings.
5758 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5761 *) Add a debugging option to PKCS#5 v2 key generation function: when
5762 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5763 derived keys are printed to stderr.
5766 *) Copy the flags in ASN1_STRING_dup().
5767 [Roman E. Pavlov <pre@mo.msk.ru>]
5769 *) The x509 application mishandled signing requests containing DSA
5770 keys when the signing key was also DSA and the parameters didn't match.
5772 It was supposed to omit the parameters when they matched the signing key:
5773 the verifying software was then supposed to automatically use the CA's
5774 parameters if they were absent from the end user certificate.
5776 Omitting parameters is no longer recommended. The test was also
5777 the wrong way round! This was probably due to unusual behaviour in
5778 EVP_cmp_parameters() which returns 1 if the parameters match.
5779 This meant that parameters were omitted when they *didn't* match and
5780 the certificate was useless. Certificates signed with 'ca' didn't have
5782 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5784 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5785 The interface is as follows:
5786 Applications can use
5787 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5788 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5789 "off" is now the default.
5790 The library internally uses
5791 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5792 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5793 to disable memory-checking temporarily.
5795 Some inconsistent states that previously were possible (and were
5796 even the default) are now avoided.
5798 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5799 with each memory chunk allocated; this is occasionally more helpful
5800 than just having a counter.
5802 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5804 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5808 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5809 which largely parallels "options", but is for changing API behaviour,
5810 whereas "options" are about protocol behaviour.
5811 Initial "mode" flags are:
5813 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5814 a single record has been written.
5815 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5816 retries use the same buffer location.
5817 (But all of the contents must be
5821 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5824 *) Fix problems with no-hmac etc.
5825 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5827 *) New functions RSA_get_default_method(), RSA_set_method() and
5828 RSA_get_method(). These allows replacement of RSA_METHODs without having
5829 to mess around with the internals of an RSA structure.
5832 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5833 Also really enable memory leak checks in openssl.c and in some
5835 [Chad C. Mulligan, Bodo Moeller]
5837 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5838 up the length of negative integers. This has now been simplified to just
5839 store the length when it is first determined and use it later, rather
5840 than trying to keep track of where data is copied and updating it to
5842 [Steve Henson, reported by Brien Wheeler
5843 <bwheeler@authentica-security.com>]
5845 *) Add a new function PKCS7_signatureVerify. This allows the verification
5846 of a PKCS#7 signature but with the signing certificate passed to the
5847 function itself. This contrasts with PKCS7_dataVerify which assumes the
5848 certificate is present in the PKCS#7 structure. This isn't always the
5849 case: certificates can be omitted from a PKCS#7 structure and be
5850 distributed by "out of band" means (such as a certificate database).
5853 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5854 function prototypes in pem.h, also change util/mkdef.pl to add the
5855 necessary function names.
5858 *) mk1mf.pl (used by Windows builds) did not properly read the
5859 options set by Configure in the top level Makefile, and Configure
5860 was not even able to write more than one option correctly.
5861 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5864 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5865 file to be loaded from a BIO or FILE pointer. The BIO version will
5866 for example allow memory BIOs to contain config info.
5869 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5870 Whoever hopes to achieve shared-library compatibility across versions
5871 must use this, not the compile-time macro.
5872 (Exercise 0.9.4: Which is the minimum library version required by
5874 Note: All this applies only to multi-threaded programs, others don't
5878 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5879 through a BIO pair triggered the default case, i.e.
5880 SSLerr(...,SSL_R_UNKNOWN_STATE).
5883 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5884 can use the SSL library even if none of the specific BIOs is
5888 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5889 for the encoded length.
5890 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5892 *) Add initial documentation of the X509V3 functions.
5895 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5896 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5897 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5898 secure PKCS#8 private key format with a high iteration count.
5901 *) Fix determination of Perl interpreter: A perl or perl5
5902 _directory_ in $PATH was also accepted as the interpreter.
5903 [Ralf S. Engelschall]
5905 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5906 wrong with it but it was very old and did things like calling
5907 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5911 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5912 to use the new extension code.
5915 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5916 with macros. This should make it easier to change their form, add extra
5917 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5921 *) Add to configuration table a new entry that can specify an alternative
5922 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5923 according to Mark Crispin <MRC@Panda.COM>.
5927 *) DES CBC did not update the IV. Weird.
5930 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5931 Changing the behaviour of the former might break existing programs --
5932 where IV updating is needed, des_ncbc_encrypt can be used.
5935 *) When bntest is run from "make test" it drives bc to check its
5936 calculations, as well as internally checking them. If an internal check
5937 fails, it needs to cause bc to give a non-zero result or make test carries
5938 on without noticing the failure. Fixed.
5941 *) DES library cleanups.
5944 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5945 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5946 ciphers. NOTE: although the key derivation function has been verified
5947 against some published test vectors it has not been extensively tested
5948 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5952 *) Instead of "mkdir -p", which is not fully portable, use new
5953 Perl script "util/mkdir-p.pl".
5956 *) Rewrite the way password based encryption (PBE) is handled. It used to
5957 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5958 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5959 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5960 the 'parameter' field of the AlgorithmIdentifier is passed to the
5961 underlying key generation function so it must do its own ASN1 parsing.
5962 This has also changed the EVP_PBE_CipherInit() function which now has a
5963 'parameter' argument instead of literal salt and iteration count values
5964 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5967 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5968 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5969 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5970 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5971 value was just used as a "magic string" and not used directly its
5972 value doesn't matter.
5975 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5979 *) "linux-sparc64" configuration (ultrapenguin).
5980 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5981 "linux-sparc" configuration.
5982 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5984 *) config now generates no-xxx options for missing ciphers.
5987 *) Support the EBCDIC character set (work in progress).
5988 File ebcdic.c not yet included because it has a different license.
5989 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5991 *) Support BS2000/OSD-POSIX.
5992 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5994 *) Make callbacks for key generation use void * instead of char *.
5997 *) Make S/MIME samples compile (not yet tested).
6000 *) Additional typesafe stacks.
6003 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6007 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6009 *) New configuration variant "sco5-gcc".
6011 *) Updated some demos.
6012 [Sean O Riordain, Wade Scholine]
6014 *) Add missing BIO_free at exit of pkcs12 application.
6017 *) Fix memory leak in conf.c.
6020 *) Updates for Win32 to assembler version of MD5.
6023 *) Set #! path to perl in apps/der_chop to where we found it
6024 instead of using a fixed path.
6027 *) SHA library changes for irix64-mips4-cc.
6030 *) Improvements for VMS support.
6034 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6036 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6037 This also avoids the problems with SC4.2 and unpatched SC5.
6038 [Andy Polyakov <appro@fy.chalmers.se>]
6040 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6041 These are required because of the typesafe stack would otherwise break
6042 existing code. If old code used a structure member which used to be STACK
6043 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6044 sk_num or sk_value it would produce an error because the num, data members
6045 are not present in STACK_OF. Now it just produces a warning. sk_set
6046 replaces the old method of assigning a value to sk_value
6047 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6048 that does this will no longer work (and should use sk_set instead) but
6049 this could be regarded as a "questionable" behaviour anyway.
6052 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6053 correctly handle encrypted S/MIME data.
6056 *) Change type of various DES function arguments from des_cblock
6057 (which means, in function argument declarations, pointer to char)
6058 to des_cblock * (meaning pointer to array with 8 char elements),
6059 which allows the compiler to do more typechecking; it was like
6060 that back in SSLeay, but with lots of ugly casts.
6062 Introduce new type const_des_cblock.
6065 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6066 problems: find RecipientInfo structure that matches recipient certificate
6067 and initialise the ASN1 structures properly based on passed cipher.
6070 *) Belatedly make the BN tests actually check the results.
6073 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6074 to and from BNs: it was completely broken. New compilation option
6075 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6076 key elements as negative integers.
6079 *) Reorganize and speed up MD5.
6080 [Andy Polyakov <appro@fy.chalmers.se>]
6083 [Richard Levitte <richard@levitte.org>]
6085 *) New option -out to asn1parse to allow the parsed structure to be
6086 output to a file. This is most useful when combined with the -strparse
6087 option to examine the output of things like OCTET STRINGS.
6090 *) Make SSL library a little more fool-proof by not requiring any longer
6091 that SSL_set_{accept,connect}_state be called before
6092 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6093 in many applications because usually everything *appeared* to work as
6094 intended anyway -- now it really works as intended).
6097 *) Move openssl.cnf out of lib/.
6100 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6101 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6102 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6103 [Ralf S. Engelschall]
6105 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6106 handle PKCS#7 enveloped data properly.
6107 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6109 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6110 copying pointers. The cert_st handling is changed by this in
6111 various ways (and thus what used to be known as ctx->default_cert
6112 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6113 any longer when s->cert does not give us what we need).
6114 ssl_cert_instantiate becomes obsolete by this change.
6115 As soon as we've got the new code right (possibly it already is?),
6116 we have solved a couple of bugs of the earlier code where s->cert
6117 was used as if it could not have been shared with other SSL structures.
6119 Note that using the SSL API in certain dirty ways now will result
6120 in different behaviour than observed with earlier library versions:
6121 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6122 does not influence s as it used to.
6124 In order to clean up things more thoroughly, inside SSL_SESSION
6125 we don't use CERT any longer, but a new structure SESS_CERT
6126 that holds per-session data (if available); currently, this is
6127 the peer's certificate chain and, for clients, the server's certificate
6128 and temporary key. CERT holds only those values that can have
6129 meaningful defaults in an SSL_CTX.
6132 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6133 from the internal representation. Various PKCS#7 fixes: remove some
6134 evil casts and set the enc_dig_alg field properly based on the signing
6138 *) Allow PKCS#12 password to be set from the command line or the
6139 environment. Let 'ca' get its config file name from the environment
6140 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6144 *) Allow certificate policies extension to use an IA5STRING for the
6145 organization field. This is contrary to the PKIX definition but
6146 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6150 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6151 without disallowing inline assembler and the like for non-pedantic builds.
6154 *) Support Borland C++ builder.
6155 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6160 *) SHA-1 cleanups and performance enhancements.
6161 [Andy Polyakov <appro@fy.chalmers.se>]
6163 *) Sparc v8plus assembler for the bignum library.
6164 [Andy Polyakov <appro@fy.chalmers.se>]
6166 *) Accept any -xxx and +xxx compiler options in Configure.
6169 *) Update HPUX configuration.
6172 *) Add missing sk_<type>_unshift() function to safestack.h
6173 [Ralf S. Engelschall]
6175 *) New function SSL_CTX_use_certificate_chain_file that sets the
6176 "extra_cert"s in addition to the certificate. (This makes sense
6177 only for "PEM" format files, as chains as a whole are not
6181 *) Support verify_depth from the SSL API.
6182 x509_vfy.c had what can be considered an off-by-one-error:
6183 Its depth (which was not part of the external interface)
6184 was actually counting the number of certificates in a chain;
6185 now it really counts the depth.
6188 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6189 instead of X509err, which often resulted in confusing error
6190 messages since the error codes are not globally unique
6191 (e.g. an alleged error in ssl3_accept when a certificate
6192 didn't match the private key).
6194 *) New function SSL_CTX_set_session_id_context that allows to set a default
6195 value (so that you don't need SSL_set_session_id_context for each
6196 connection using the SSL_CTX).
6199 *) OAEP decoding bug fix.
6202 *) Support INSTALL_PREFIX for package builders, as proposed by
6206 *) New Configure options "threads" and "no-threads". For systems
6207 where the proper compiler options are known (currently Solaris
6208 and Linux), "threads" is the default.
6211 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6214 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6215 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6216 such as /usr/local/bin.
6219 *) "make linux-shared" to build shared libraries.
6220 [Niels Poppe <niels@netbox.org>]
6222 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6225 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6226 extension adding in x509 utility.
6229 *) Remove NOPROTO sections and error code comments.
6232 *) Partial rewrite of the DEF file generator to now parse the ANSI
6236 *) New Configure options --prefix=DIR and --openssldir=DIR.
6239 *) Complete rewrite of the error code script(s). It is all now handled
6240 by one script at the top level which handles error code gathering,
6241 header rewriting and C source file generation. It should be much better
6242 than the old method: it now uses a modified version of Ulf's parser to
6243 read the ANSI prototypes in all header files (thus the old K&R definitions
6244 aren't needed for error creation any more) and do a better job of
6245 translating function codes into names. The old 'ASN1 error code imbedded
6246 in a comment' is no longer necessary and it doesn't use .err files which
6247 have now been deleted. Also the error code call doesn't have to appear all
6248 on one line (which resulted in some large lines...).
6251 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6254 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6255 0 (which usually indicates a closed connection), but continue reading.
6258 *) Fix some race conditions.
6261 *) Add support for CRL distribution points extension. Add Certificate
6262 Policies and CRL distribution points documentation.
6265 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6268 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6269 8 of keying material. Merlin has also confirmed interop with this fix
6270 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6271 [Merlin Hughes <merlin@baltimore.ie>]
6273 *) Fix lots of warnings.
6274 [Richard Levitte <levitte@stacken.kth.se>]
6276 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6277 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6278 [Richard Levitte <levitte@stacken.kth.se>]
6280 *) Fix problems with sizeof(long) == 8.
6281 [Andy Polyakov <appro@fy.chalmers.se>]
6283 *) Change functions to ANSI C.
6286 *) Fix typos in error codes.
6287 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6289 *) Remove defunct assembler files from Configure.
6292 *) SPARC v8 assembler BIGNUM implementation.
6293 [Andy Polyakov <appro@fy.chalmers.se>]
6295 *) Support for Certificate Policies extension: both print and set.
6296 Various additions to support the r2i method this uses.
6299 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6300 return a const string when you are expecting an allocated buffer.
6303 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6304 types DirectoryString and DisplayText.
6307 *) Add code to allow r2i extensions to access the configuration database,
6308 add an LHASH database driver and add several ctx helper functions.
6311 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6312 fail when they extended the size of a BIGNUM.
6315 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6316 support typesafe stack.
6319 *) Fix typo in SSL_[gs]et_options().
6320 [Nils Frostberg <nils@medcom.se>]
6322 *) Delete various functions and files that belonged to the (now obsolete)
6323 old X509V3 handling code.
6326 *) New Configure option "rsaref".
6329 *) Don't auto-generate pem.h.
6332 *) Introduce type-safe ASN.1 SETs.
6335 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6336 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6338 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6339 that links with OpenSSL (well at least cause lots of warnings), but fear
6340 not: the conversion is trivial, and it eliminates loads of evil casts. A
6341 few STACKed things have been converted already. Feel free to convert more.
6342 In the fullness of time, I'll do away with the STACK type altogether.
6345 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6346 specified in <certfile> by updating the entry in the index.txt file.
6347 This way one no longer has to edit the index.txt file manually for
6348 revoking a certificate. The -revoke option does the gory details now.
6349 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6351 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6352 `-text' option at all and this way the `-noout -text' combination was
6353 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6354 [Ralf S. Engelschall]
6356 *) Make sure a corresponding plain text error message exists for the
6357 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6358 verify callback function determined that a certificate was revoked.
6359 [Ralf S. Engelschall]
6361 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6362 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6363 all available cipers including rc5, which was forgotten until now.
6364 In order to let the testing shell script know which algorithms
6365 are available, a new (up to now undocumented) command
6366 "openssl list-cipher-commands" is used.
6369 *) Bugfix: s_client occasionally would sleep in select() when
6370 it should have checked SSL_pending() first.
6373 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6374 the raw DSA values prior to ASN.1 encoding.
6377 *) Tweaks to Configure
6378 [Niels Poppe <niels@netbox.org>]
6380 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6384 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6387 *) New config option to avoid instructions that are illegal on the 80386.
6388 The default code is faster, but requires at least a 486.
6391 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6392 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6393 same as SSL2_VERSION anyway.
6396 *) New "-showcerts" option for s_client.
6399 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6400 application. Various cleanups and fixes.
6403 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6404 modify error routines to work internally. Add error codes and PBE init
6405 to library startup routines.
6408 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6409 packing functions to asn1 and evp. Changed function names and error
6410 codes along the way.
6413 *) PKCS12 integration: and so it begins... First of several patches to
6414 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6415 objects to objects.h
6418 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6419 and display support for Thawte strong extranet extension.
6422 *) Add LinuxPPC support.
6423 [Jeff Dubrule <igor@pobox.org>]
6425 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6426 bn_div_words in alpha.s.
6427 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6429 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6430 OAEP isn't supported when OpenSSL is built with RSAref.
6431 [Ulf Moeller <ulf@fitug.de>]
6433 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6434 so they no longer are missing under -DNOPROTO.
6435 [Soren S. Jorvang <soren@t.dk>]
6438 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6440 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6441 doesn't work when the session is reused. Coming soon!
6444 *) Fix a security hole, that allows sessions to be reused in the wrong
6445 context thus bypassing client cert protection! All software that uses
6446 client certs and session caches in multiple contexts NEEDS PATCHING to
6447 allow session reuse! A fuller solution is in the works.
6448 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6450 *) Some more source tree cleanups (removed obsolete files
6451 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6452 permission on "config" script to be executable) and a fix for the INSTALL
6454 [Ulf Moeller <ulf@fitug.de>]
6456 *) Remove some legacy and erroneous uses of malloc, free instead of
6458 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6460 *) Make rsa_oaep_test return non-zero on error.
6461 [Ulf Moeller <ulf@fitug.de>]
6463 *) Add support for native Solaris shared libraries. Configure
6464 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6465 if someone would make that last step automatic.
6466 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6468 *) ctx_size was not built with the right compiler during "make links". Fixed.
6471 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6472 except NULL ciphers". This means the default cipher list will no longer
6473 enable NULL ciphers. They need to be specifically enabled e.g. with
6474 the string "DEFAULT:eNULL".
6477 *) Fix to RSA private encryption routines: if p < q then it would
6478 occasionally produce an invalid result. This will only happen with
6479 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6482 *) Be less restrictive and allow also `perl util/perlpath.pl
6483 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6484 because this way one can also use an interpreter named `perl5' (which is
6485 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6486 installed as `perl').
6487 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6489 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6490 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6492 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6493 advapi32.lib to Win32 build and change the pem test comparision
6494 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6495 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6496 and crypto/des/ede_cbcm_enc.c.
6499 *) DES quad checksum was broken on big-endian architectures. Fixed.
6502 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6503 Win32 test batch file so it (might) work again. The Win32 test batch file
6504 is horrible: I feel ill....
6507 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6508 in e_os.h. Audit of header files to check ANSI and non ANSI
6509 sections: 10 functions were absent from non ANSI section and not exported
6510 from Windows DLLs. Fixed up libeay.num for new functions.
6513 *) Make `openssl version' output lines consistent.
6514 [Ralf S. Engelschall]
6516 *) Fix Win32 symbol export lists for BIO functions: Added
6517 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6518 to ms/libeay{16,32}.def.
6519 [Ralf S. Engelschall]
6521 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6522 fine under Unix and passes some trivial tests I've now added. But the
6523 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6524 added to make sure no one expects that this stuff really works in the
6525 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6526 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6528 [Ralf S. Engelschall]
6530 *) Fix the generation of two part addresses in perl.
6531 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6533 *) Add config entry for Linux on MIPS.
6534 [John Tobey <jtobey@channel1.com>]
6536 *) Make links whenever Configure is run, unless we are on Windoze.
6539 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6540 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6544 *) Add a useful kludge to allow package maintainers to specify compiler and
6545 other platforms details on the command line without having to patch the
6546 Configure script everytime: One now can use ``perl Configure
6547 <id>:<details>'', i.e. platform ids are allowed to have details appended
6548 to them (seperated by colons). This is treated as there would be a static
6549 pre-configured entry in Configure's %table under key <id> with value
6550 <details> and ``perl Configure <id>'' is called. So, when you want to
6551 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6552 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6553 now, which overrides the FreeBSD-elf entry on-the-fly.
6554 [Ralf S. Engelschall]
6556 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6559 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6560 on the `perl Configure ...' command line. This way one can compile
6561 OpenSSL libraries with Position Independent Code (PIC) which is needed
6562 for linking it into DSOs.
6563 [Ralf S. Engelschall]
6565 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6569 *) Cleaned up the LICENSE document: The official contact for any license
6570 questions now is the OpenSSL core team under openssl-core@openssl.org.
6571 And add a paragraph about the dual-license situation to make sure people
6572 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6573 to the OpenSSL toolkit.
6574 [Ralf S. Engelschall]
6576 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6577 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6578 Additonally cleaned up the `make links' target: Remove unnecessary
6579 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6580 to speed processing and no longer clutter the display with confusing
6581 stuff. Instead only the actually done links are displayed.
6582 [Ralf S. Engelschall]
6584 *) Permit null encryption ciphersuites, used for authentication only. It used
6585 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6586 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6590 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6591 signed attributes when verifying signatures (this would break them),
6592 the detached data encoding was wrong and public keys obtained using
6593 X509_get_pubkey() weren't freed.
6596 *) Add text documentation for the BUFFER functions. Also added a work around
6597 to a Win95 console bug. This was triggered by the password read stuff: the
6598 last character typed gets carried over to the next fread(). If you were
6599 generating a new cert request using 'req' for example then the last
6600 character of the passphrase would be CR which would then enter the first
6604 *) Added the new `Includes OpenSSL Cryptography Software' button as
6605 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6606 button and can be used by applications based on OpenSSL to show the
6607 relationship to the OpenSSL project.
6608 [Ralf S. Engelschall]
6610 *) Remove confusing variables in function signatures in files
6611 ssl/ssl_lib.c and ssl/ssl.h.
6612 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6614 *) Don't install bss_file.c under PREFIX/include/
6615 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6617 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6618 functions that return function pointers and has support for NT specific
6619 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6620 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6621 unsigned to signed types: this was killing the Win32 compile.
6624 *) Add new certificate file to stack functions,
6625 SSL_add_dir_cert_subjects_to_stack() and
6626 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6627 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6628 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6629 This means that Apache-SSL and similar packages don't have to mess around
6630 to add as many CAs as they want to the preferred list.
6633 *) Experiment with doxygen documentation. Currently only partially applied to
6635 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6636 openssl.doxy as the configuration file.
6639 *) Get rid of remaining C++-style comments which strict C compilers hate.
6640 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6642 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6643 compiled in by default: it has problems with large keys.
6646 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6647 DH private keys and/or callback functions which directly correspond to
6648 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6649 is needed for applications which have to configure certificates on a
6650 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6652 For the RSA certificate situation is makes no difference, but
6653 for the DSA certificate situation this fixes the "no shared cipher"
6654 problem where the OpenSSL cipher selection procedure failed because the
6655 temporary keys were not overtaken from the context and the API provided
6656 no way to reconfigure them.
6657 The new functions now let applications reconfigure the stuff and they
6658 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6659 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6660 non-public-API function ssl_cert_instantiate() is used as a helper
6661 function and also to reduce code redundancy inside ssl_rsa.c.
6662 [Ralf S. Engelschall]
6664 *) Move s_server -dcert and -dkey options out of the undocumented feature
6665 area because they are useful for the DSA situation and should be
6666 recognized by the users.
6667 [Ralf S. Engelschall]
6669 *) Fix the cipher decision scheme for export ciphers: the export bits are
6670 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6671 SSL_EXP_MASK. So, the original variable has to be used instead of the
6672 already masked variable.
6673 [Richard Levitte <levitte@stacken.kth.se>]
6675 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6676 [Richard Levitte <levitte@stacken.kth.se>]
6678 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6679 from `int' to `unsigned int' because it's a length and initialized by
6680 EVP_DigestFinal() which expects an `unsigned int *'.
6681 [Richard Levitte <levitte@stacken.kth.se>]
6683 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6684 script. Instead use the usual Shell->Perl transition trick.
6685 [Ralf S. Engelschall]
6687 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6688 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6689 -noout -modulus' as it's already the case for `openssl rsa -noout
6690 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6691 currently the public key is printed (a decision which was already done by
6692 `openssl dsa -modulus' in the past) which serves a similar purpose.
6693 Additionally the NO_RSA no longer completely removes the whole -modulus
6694 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6696 [Ralf S. Engelschall]
6698 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6699 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6700 [Arne Ansper <arne@ats.cyber.ee>]
6702 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6703 to be added. Now both 'req' and 'ca' can use new objects defined in the
6707 *) Add cool BIO that does syslog (or event log on NT).
6708 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6710 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6711 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6712 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6713 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6716 *) Add preliminary config info for new extension code.
6719 *) Make RSA_NO_PADDING really use no padding.
6720 [Ulf Moeller <ulf@fitug.de>]
6722 *) Generate errors when private/public key check is done.
6725 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6726 for some CRL extensions and new objects added.
6729 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6730 key usage extension and fuller support for authority key id.
6733 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6734 padding method for RSA, which is recommended for new applications in PKCS
6735 #1 v2.0 (RFC 2437, October 1998).
6736 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6737 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6738 against Bleichbacher's attack on RSA.
6739 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6742 *) Updates to the new SSL compression code
6743 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6745 *) Fix so that the version number in the master secret, when passed
6746 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6747 (because the server will not accept higher), that the version number
6748 is 0x03,0x01, not 0x03,0x00
6749 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6751 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6752 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6753 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6756 *) Support for RAW extensions where an arbitrary extension can be
6757 created by including its DER encoding. See apps/openssl.cnf for
6761 *) Make sure latest Perl versions don't interpret some generated C array
6762 code as Perl array code in the crypto/err/err_genc.pl script.
6763 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6765 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6766 not many people have the assembler. Various Win32 compilation fixes and
6767 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6771 *) Modify configure script 'Configure' to automatically create crypto/date.h
6772 file under Win32 and also build pem.h from pem.org. New script
6773 util/mkfiles.pl to create the MINFO file on environments that can't do a
6774 'make files': perl util/mkfiles.pl >MINFO should work.
6777 *) Major rework of DES function declarations, in the pursuit of correctness
6778 and purity. As a result, many evil casts evaporated, and some weirdness,
6779 too. You may find this causes warnings in your code. Zapping your evil
6780 casts will probably fix them. Mostly.
6783 *) Fix for a typo in asn1.h. Bug fix to object creation script
6784 obj_dat.pl. It considered a zero in an object definition to mean
6785 "end of object": none of the objects in objects.h have any zeros
6786 so it wasn't spotted.
6787 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6789 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6790 Masking (CBCM). In the absence of test vectors, the best I have been able
6791 to do is check that the decrypt undoes the encrypt, so far. Send me test
6792 vectors if you have them.
6795 *) Correct calculation of key length for export ciphers (too much space was
6796 allocated for null ciphers). This has not been tested!
6799 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6800 message is now correct (it understands "crypto" and "ssl" on its
6801 command line). There is also now an "update" option. This will update
6802 the util/ssleay.num and util/libeay.num files with any new functions.
6804 perl util/mkdef.pl crypto ssl update
6805 it will update them.
6808 *) Overhauled the Perl interface (perl/*):
6809 - ported BN stuff to OpenSSL's different BN library
6810 - made the perl/ source tree CVS-aware
6811 - renamed the package from SSLeay to OpenSSL (the files still contain
6812 their history because I've copied them in the repository)
6813 - removed obsolete files (the test scripts will be replaced
6814 by better Test::Harness variants in the future)
6815 [Ralf S. Engelschall]
6817 *) First cut for a very conservative source tree cleanup:
6818 1. merge various obsolete readme texts into doc/ssleay.txt
6819 where we collect the old documents and readme texts.
6820 2. remove the first part of files where I'm already sure that we no
6821 longer need them because of three reasons: either they are just temporary
6822 files which were left by Eric or they are preserved original files where
6823 I've verified that the diff is also available in the CVS via "cvs diff
6824 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6825 the crypto/md/ stuff).
6826 [Ralf S. Engelschall]
6828 *) More extension code. Incomplete support for subject and issuer alt
6829 name, issuer and authority key id. Change the i2v function parameters
6830 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6831 what that's for :-) Fix to ASN1 macro which messed up
6832 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6835 *) Preliminary support for ENUMERATED type. This is largely copied from the
6839 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6840 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6842 *) Make sure `make rehash' target really finds the `openssl' program.
6843 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6845 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6846 like to hear about it if this slows down other processors.
6849 *) Add CygWin32 platform information to Configure script.
6850 [Alan Batie <batie@aahz.jf.intel.com>]
6852 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6853 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6855 *) New program nseq to manipulate netscape certificate sequences
6858 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6862 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6863 but the BN code had some problems that would cause failures when
6864 doing certificate verification and some other functions.
6865 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6867 *) Add ASN1 and PEM code to support netscape certificate sequences.
6870 *) Add ASN1 and PEM code to support netscape certificate sequences.
6873 *) Add several PKIX and private extended key usage OIDs.
6876 *) Modify the 'ca' program to handle the new extension code. Modify
6877 openssl.cnf for new extension format, add comments.
6880 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6881 and add a sample to openssl.cnf so req -x509 now adds appropriate
6885 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6886 error code, add initial support to X509_print() and x509 application.
6889 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6890 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6891 stuff is currently isolated and isn't even compiled yet.
6894 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6895 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6896 Removed the versions check from X509 routines when loading extensions:
6897 this allows certain broken certificates that don't set the version
6898 properly to be processed.
6901 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6902 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6903 can still be regenerated with "make depend".
6906 *) Spelling mistake in C version of CAST-128.
6907 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6909 *) Changes to the error generation code. The perl script err-code.pl
6910 now reads in the old error codes and retains the old numbers, only
6911 adding new ones if necessary. It also only changes the .err files if new
6912 codes are added. The makefiles have been modified to only insert errors
6913 when needed (to avoid needlessly modifying header files). This is done
6914 by only inserting errors if the .err file is newer than the auto generated
6915 C file. To rebuild all the error codes from scratch (the old behaviour)
6916 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6917 or delete all the .err files.
6920 *) CAST-128 was incorrectly implemented for short keys. The C version has
6921 been fixed, but is untested. The assembler versions are also fixed, but
6922 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6923 to regenerate it if needed.
6924 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6925 Hagino <itojun@kame.net>]
6927 *) File was opened incorrectly in randfile.c.
6928 [Ulf Möller <ulf@fitug.de>]
6930 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6931 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6932 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6933 al: it's just almost always a UTCTime. Note this patch adds new error
6934 codes so do a "make errors" if there are problems.
6937 *) Correct Linux 1 recognition in config.
6938 [Ulf Möller <ulf@fitug.de>]
6940 *) Remove pointless MD5 hash when using DSA keys in ca.
6941 [Anonymous <nobody@replay.com>]
6943 *) Generate an error if given an empty string as a cert directory. Also
6944 generate an error if handed NULL (previously returned 0 to indicate an
6945 error, but didn't set one).
6946 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6948 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6951 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6952 parameters. This was causing a warning which killed off the Win32 compile.
6955 *) Remove C++ style comments from crypto/bn/bn_local.h.
6956 [Neil Costigan <neil.costigan@celocom.com>]
6958 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6959 based on a text string, looking up short and long names and finally
6960 "dot" format. The "dot" format stuff didn't work. Added new function
6961 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6962 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6963 OID is not part of the table.
6966 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6967 X509_LOOKUP_by_alias().
6970 *) Sort openssl functions by name.
6973 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6974 encryption from sample DSA keys (in case anyone is interested the password
6978 *) Make _all_ *_free functions accept a NULL pointer.
6979 [Frans Heymans <fheymans@isaserver.be>]
6981 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6983 [Anonymous <nobody@replay.com>]
6985 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6986 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6988 *) Don't blow it for numeric -newkey arguments to apps/req.
6989 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6991 *) Temp key "for export" tests were wrong in s3_srvr.c.
6992 [Anonymous <nobody@replay.com>]
6994 *) Add prototype for temp key callback functions
6995 SSL_CTX_set_tmp_{rsa,dh}_callback().
6998 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6999 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7002 *) X509_name_add_entry() freed the wrong thing after an error.
7003 [Arne Ansper <arne@ats.cyber.ee>]
7005 *) rsa_eay.c would attempt to free a NULL context.
7006 [Arne Ansper <arne@ats.cyber.ee>]
7008 *) BIO_s_socket() had a broken should_retry() on Windoze.
7009 [Arne Ansper <arne@ats.cyber.ee>]
7011 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7012 [Arne Ansper <arne@ats.cyber.ee>]
7014 *) Make sure the already existing X509_STORE->depth variable is initialized
7015 in X509_STORE_new(), but document the fact that this variable is still
7016 unused in the certificate verification process.
7017 [Ralf S. Engelschall]
7019 *) Fix the various library and apps files to free up pkeys obtained from
7020 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7023 *) Fix reference counting in X509_PUBKEY_get(). This makes
7024 demos/maurice/example2.c work, amongst others, probably.
7025 [Steve Henson and Ben Laurie]
7027 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7028 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7029 are no longer created. This way we have a single and consistent command
7030 line interface `openssl <command>', similar to `cvs <command>'.
7031 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7033 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7034 BIT STRING wrapper always have zero unused bits.
7037 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7040 *) Make the top-level INSTALL documentation easier to understand.
7043 *) Makefiles updated to exit if an error occurs in a sub-directory
7044 make (including if user presses ^C) [Paul Sutton]
7046 *) Make Montgomery context stuff explicit in RSA data structure.
7049 *) Fix build order of pem and err to allow for generated pem.h.
7052 *) Fix renumbering bug in X509_NAME_delete_entry().
7055 *) Enhanced the err-ins.pl script so it makes the error library number
7056 global and can add a library name. This is needed for external ASN1 and
7057 other error libraries.
7060 *) Fixed sk_insert which never worked properly.
7063 *) Fix ASN1 macros so they can handle indefinite length construted
7064 EXPLICIT tags. Some non standard certificates use these: they can now
7068 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7069 into a single doc/ssleay.txt bundle. This way the information is still
7070 preserved but no longer messes up this directory. Now it's new room for
7071 the new set of documenation files.
7072 [Ralf S. Engelschall]
7074 *) SETs were incorrectly DER encoded. This was a major pain, because they
7075 shared code with SEQUENCEs, which aren't coded the same. This means that
7076 almost everything to do with SETs or SEQUENCEs has either changed name or
7077 number of arguments.
7078 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7080 *) Fix test data to work with the above.
7083 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7084 was already fixed by Eric for 0.9.1 it seems.
7085 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7087 *) Autodetect FreeBSD3.
7090 *) Fix various bugs in Configure. This affects the following platforms:
7094 unixware-2.0-pentium
7098 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7099 before they are needed.
7102 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7106 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7108 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7109 changed SSLeay to OpenSSL in version strings.
7110 [Ralf S. Engelschall]
7112 *) Some fixups to the top-level documents.
7115 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7116 because the symlink to include/ was missing.
7117 [Ralf S. Engelschall]
7119 *) Incorporated the popular no-RSA/DSA-only patches
7120 which allow to compile a RSA-free SSLeay.
7121 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7123 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7124 when "ssleay" is still not found.
7125 [Ralf S. Engelschall]
7127 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7128 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7130 *) Updated the README file.
7131 [Ralf S. Engelschall]
7133 *) Added various .cvsignore files in the CVS repository subdirs
7134 to make a "cvs update" really silent.
7135 [Ralf S. Engelschall]
7137 *) Recompiled the error-definition header files and added
7138 missing symbols to the Win32 linker tables.
7139 [Ralf S. Engelschall]
7141 *) Cleaned up the top-level documents;
7142 o new files: CHANGES and LICENSE
7143 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7144 o merged COPYRIGHT into LICENSE
7145 o removed obsolete TODO file
7146 o renamed MICROSOFT to INSTALL.W32
7147 [Ralf S. Engelschall]
7149 *) Removed dummy files from the 0.9.1b source tree:
7150 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7151 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7152 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7153 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7154 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7155 [Ralf S. Engelschall]
7157 *) Added various platform portability fixes.
7160 *) The Genesis of the OpenSSL rpject:
7161 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7162 Young and Tim J. Hudson created while they were working for C2Net until
7164 [The OpenSSL Project]
7167 Changes between 0.9.0b and 0.9.1b [not released]
7169 *) Updated a few CA certificates under certs/
7172 *) Changed some BIGNUM api stuff.
7175 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7176 DGUX x86, Linux Alpha, etc.
7179 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7180 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7184 *) Add -strparse option to asn1pars program which parses nested
7186 [Dr Stephen Henson <shenson@bigfoot.com>]
7188 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7191 *) DSA fix for "ca" program.
7194 *) Added "-genkey" option to "dsaparam" program.
7197 *) Added RIPE MD160 (rmd160) message digest.
7200 *) Added -a (all) option to "ssleay version" command.
7203 *) Added PLATFORM define which is the id given to Configure.
7206 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7209 *) Extended the ASN.1 parser routines.
7212 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7215 *) Added a BN_CTX to the BN library.
7218 *) Fixed the weak key values in DES library
7221 *) Changed API in EVP library for cipher aliases.
7224 *) Added support for RC2/64bit cipher.
7227 *) Converted the lhash library to the crypto/mem.c functions.
7230 *) Added more recognized ASN.1 object ids.
7233 *) Added more RSA padding checks for SSL/TLS.
7236 *) Added BIO proxy/filter functionality.
7239 *) Added extra_certs to SSL_CTX which can be used
7240 send extra CA certificates to the client in the CA cert chain sending
7241 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7244 *) Now Fortezza is denied in the authentication phase because
7245 this is key exchange mechanism is not supported by SSLeay at all.
7248 *) Additional PKCS1 checks.
7251 *) Support the string "TLSv1" for all TLS v1 ciphers.
7254 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7255 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7258 *) Fixed a few memory leaks.
7261 *) Fixed various code and comment typos.
7264 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7265 bytes sent in the client random.
7266 [Edward Bishop <ebishop@spyglass.com>]