5 Changes between 0.9.7c and 0.9.8 [xx XXX xxxx]
7 *) Reimplemented the BN_CTX implementation. There is now no more static
8 limitation on the number of variables it can handle nor the depth of the
9 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
10 information can now expand as required, and rather than having a single
11 static array of bignums, BN_CTX now uses a linked-list of such arrays
12 allowing it to expand on demand whilst maintaining the usefulness of
16 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
17 to allow all RSA operations to function using a single BN_CTX.
20 *) Preliminary support for certificate policy evaluation and checking. This
21 is initially intended to pass the tests outlined in "Conformance Testing
22 of Relying Party Client Certificate Path Processing Logic" v1.07.
25 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
26 remained unused and not that useful. A variety of other little bignum
27 tweaks and fixes have also been made continuing on from the audit (see
31 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
32 associated ASN1, EVP and SSL functions and old ASN1 macros.
35 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
36 and this should never fail. So the return value from the use of
37 BN_set_word() (which can fail due to needless expansion) is now deprecated;
38 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
41 *) BN_CTX_get() should return zero-valued bignums, providing the same
42 initialised value as BN_new().
43 [Geoff Thorpe, suggested by Ulf Möller]
45 *) Support for inhibitAnyPolicy certificate extension.
48 *) An audit of the BIGNUM code is underway, for which debugging code is
49 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
50 is considered valid when processing BIGNUMs, and causes execution to
51 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
52 further steps are taken to deliberately pollute unused data in BIGNUM
53 structures to try and expose faulty code further on. For now, openssl will
54 (in its default mode of operation) continue to tolerate the inconsistent
55 forms that it has tolerated in the past, but authors and packagers should
56 consider trying openssl and their own applications when compiled with
57 these debugging symbols defined. It will help highlight potential bugs in
58 their own code, and will improve the test coverage for OpenSSL itself. At
59 some point, these tighter rules will become openssl's default to improve
60 maintainability, though the assert()s and other overheads will remain only
61 in debugging configurations. See bn.h for more details.
62 [Geoff Thorpe, Nils Larsch, Ulf Möller]
64 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
65 that can only be obtained through BN_CTX_new() (which implicitly
66 initialises it). The presence of this function only made it possible
67 to overwrite an existing structure (and cause memory leaks).
70 *) Because of the callback-based approach for implementing LHASH as a
71 template type, lh_insert() adds opaque objects to hash-tables and
72 lh_doall() or lh_doall_arg() are typically used with a destructor callback
73 to clean up those corresponding objects before destroying the hash table
74 (and losing the object pointers). So some over-zealous constifications in
75 LHASH have been relaxed so that lh_insert() does not take (nor store) the
76 objects as "const" and the lh_doall[_arg] callback wrappers are not
77 prototyped to have "const" restrictions on the object pointers they are
78 given (and so aren't required to cast them away any more).
81 *) The tmdiff.h API was so ugly and minimal that our own timing utility
82 (speed) prefers to use its own implementation. The two implementations
83 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
84 its object type properly exposed (MS_TM) instead of casting to/from "char
85 *". This may still change yet if someone realises MS_TM and "ms_time_***"
86 aren't necessarily the greatest nomenclatures - but this is what was used
87 internally to the implementation so I've used that for now.
90 *) Ensure that deprecated functions do not get compiled when
91 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
92 the self-tests were still using deprecated key-generation functions so
93 these have been updated also.
96 *) Reorganise PKCS#7 code to separate the digest location functionality
97 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
98 New function PKCS7_set_digest() to set the digest type for PKCS#7
99 digestedData type. Add additional code to correctly generate the
100 digestedData type and add support for this type in PKCS7 initialization
104 *) New function PKCS7_set0_type_other() this initializes a PKCS7
105 structure of type "other".
108 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
109 sure the loop does correctly stop and breaking ("division by zero")
110 modulus operations are not performed. The (pre-generated) prime
111 table crypto/bn/bn_prime.h was already correct, but it could not be
112 re-generated on some platforms because of the "division by zero"
113 situation in the script.
114 [Ralf S. Engelschall]
116 *) Update support for ECC-based TLS ciphersuites according to
117 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
118 SHA-1 now is only used for "small" curves (where the
119 representation of a field element takes up to 24 bytes); for
120 larger curves, the field element resulting from ECDH is directly
121 used as premaster secret.
122 [Douglas Stebila (Sun Microsystems Laboratories)]
124 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
125 curve secp160r1 to the tests.
126 [Douglas Stebila (Sun Microsystems Laboratories)]
128 *) Add the possibility to load symbols globally with DSO.
129 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
131 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
132 control of the error stack.
135 *) Add support for STORE in ENGINE.
138 *) Add the STORE type. The intention is to provide a common interface
139 to certificate and key stores, be they simple file-based stores, or
140 HSM-type store, or LDAP stores, or...
141 NOTE: The code is currently UNTESTED and isn't really used anywhere.
144 *) Add a generic structure called OPENSSL_ITEM. This can be used to
145 pass a list of arguments to any function as well as provide a way
146 for a function to pass data back to the caller.
149 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
150 works like BUF_strdup() but can be used to duplicate a portion of
151 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
155 *) Add the function sk_find_ex() which works like sk_find(), but will
156 return an index to an element even if an exact match couldn't be
157 found. The index is guaranteed to point at the element where the
158 searched-for key would be inserted to preserve sorting order.
161 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
162 takes an extra flags argument for optional functionality. Currently,
163 the following flags are defined:
165 OBJ_BSEARCH_VALUE_ON_NOMATCH
166 This one gets OBJ_bsearch_ex() to return a pointer to the first
167 element where the comparing function returns a negative or zero
170 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
171 This one gets OBJ_bsearch_ex() to return a pointer to the first
172 element where the comparing function returns zero. This is useful
173 if there are more than one element where the comparing function
177 *) Make it possible to create self-signed certificates with 'openssl ca'
178 in such a way that the self-signed certificate becomes part of the
179 CA database and uses the same mechanisms for serial number generation
180 as all other certificate signing. The new flag '-selfsign' enables
181 this functionality. Adapt CA.sh and CA.pl.in.
184 *) Add functionality to check the public key of a certificate request
185 against a given private. This is useful to check that a certificate
186 request can be signed by that key (self-signing).
189 *) Make it possible to have multiple active certificates with the same
190 subject in the CA index file. This is done only if the keyword
191 'unique_subject' is set to 'no' in the main CA section (default
192 if 'CA_default') of the configuration file. The value is saved
193 with the database itself in a separate index attribute file,
194 named like the index file with '.attr' appended to the name.
197 *) Generate muti valued AVAs using '+' notation in config files for
201 *) Support for nameConstraints certificate extension.
204 *) Support for policyConstraints certificate extension.
207 *) Support for policyMappings certificate extension.
210 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
211 ENGINE as defaults for all supported algorithms irrespective of
212 the 'flags' parameter. 'flags' is now honoured, so applications
213 should make sure they are passing it correctly.
216 *) Make sure the default DSA_METHOD implementation only uses its
217 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
218 and change its own handlers to be NULL so as to remove unnecessary
219 indirection. This lets alternative implementations fallback to the
220 default implementation more easily.
223 *) Support for directoryName in GeneralName related extensions
227 *) Make it possible to link applications using Makefile.shared.
228 Make that possible even when linking against static libraries!
231 *) Support for single pass processing for S/MIME signing. This now
232 means that S/MIME signing can be done from a pipe, in addition
233 cleartext signing (multipart/signed type) is effectively streaming
234 and the signed data does not need to be all held in memory.
236 This is done with a new flag PKCS7_STREAM. When this flag is set
237 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
238 is done after the data is output (and digests calculated) in
242 *) Add full support for -rpath/-R, both in shared libraries and
243 applications, at least on the platforms where it's known how
247 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
248 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
249 will now compute a table of multiples of the generator that
250 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
251 faster (notably in the case of a single point multiplication,
253 [Nils Larsch, Bodo Moeller]
255 *) IPv6 support for certificate extensions. The various extensions
256 which use the IP:a.b.c.d can now take IPv6 addresses using the
257 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
261 *) Added an ENGINE that implements RSA by performing private key
262 exponentiations with the GMP library. The conversions to and from
263 GMP's mpz_t format aren't optimised nor are any montgomery forms
264 cached, and on x86 it appears OpenSSL's own performance has caught up.
265 However there are likely to be other architectures where GMP could
266 provide a boost. This ENGINE is not built in by default, but it can be
267 specified at Configure time and should be accompanied by the necessary
268 linker additions, eg;
269 ./config -DOPENSSL_USE_GMP -lgmp
272 *) "openssl engine" will not display ENGINE/DSO load failure errors when
273 testing availability of engines with "-t" - the old behaviour is
274 produced by increasing the feature's verbosity with "-tt".
277 *) ECDSA routines: under certain error conditions uninitialized BN objects
278 could be freed. Solution: make sure initialization is performed early
279 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
283 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
284 and DH_METHOD (eg. by ENGINE implementations) to override the normal
285 software implementations. For DSA and DH, parameter generation can
286 also be overriden by providing the appropriate method callbacks.
289 *) Change the "progress" mechanism used in key-generation and
290 primality testing to functions that take a new BN_GENCB pointer in
291 place of callback/argument pairs. The new API functions have "_ex"
292 postfixes and the older functions are reimplemented as wrappers for
293 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
294 declarations of the old functions to help (graceful) attempts to
295 migrate to the new functions. Also, the new key-generation API
296 functions operate on a caller-supplied key-structure and return
297 success/failure rather than returning a key or NULL - this is to
298 help make "keygen" another member function of RSA_METHOD etc.
300 Example for using the new callback interface:
302 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
306 BN_GENCB_set(&my_cb, my_callback, my_arg);
308 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
309 /* For the meaning of a, b in calls to my_callback(), see the
310 * documentation of the function that calls the callback.
311 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
312 * my_callback should return 1 if it wants BN_is_prime_ex()
313 * to continue, or 0 to stop.
318 *) Change the ZLIB compression method to be stateful, and make it
319 available to TLS with the number defined in
320 draft-ietf-tls-compression-04.txt.
323 *) Add the ASN.1 structures and functions for CertificatePair, which
324 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
326 CertificatePair ::= SEQUENCE {
327 forward [0] Certificate OPTIONAL,
328 reverse [1] Certificate OPTIONAL,
329 -- at least one of the pair shall be present -- }
331 Also implement the PEM functions to read and write certificate
332 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
334 This needed to be defined, mostly for the sake of the LDAP
335 attribute crossCertificatePair, but may prove useful elsewhere as
339 *) Make it possible to inhibit symlinking of shared libraries in
340 Makefile.shared, for Cygwin's sake.
343 *) Extend the BIGNUM API by creating new macros that behave like
346 void BN_set_sign(BIGNUM *a, int neg);
347 int BN_get_sign(const BIGNUM *a);
349 and avoid the need to access 'a->neg' directly in applications.
350 [Nils Larsch <nla@trustcenter.de>]
352 *) Implement fast modular reduction for pseudo-Mersenne primes
353 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
354 EC_GROUP_new_curve_GFp() will now automatically use this
356 [Nils Larsch <nla@trustcenter.de>]
358 *) Add new lock type (CRYPTO_LOCK_BN).
361 *) Change the ENGINE framework to automatically load engines
362 dynamically from specific directories unless they could be
363 found to already be built in or loaded. Move all the
364 current engines except for the cryptodev one to a new
366 The engines in engines/ are built as shared libraries if
367 the "shared" options was given to ./Configure or ./config.
368 Otherwise, they are inserted in libcrypto.a.
369 /usr/local/ssl/engines is the default directory for dynamic
370 engines, but that can be overriden at configure time through
371 the usual use of --prefix and/or --openssldir, and at run
372 time with the environment variable OPENSSL_ENGINES.
373 [Geoff Thorpe and Richard Levitte]
375 *) Add Makefile.shared, a helper makefile to build shared
376 libraries. Addapt Makefile.org.
379 *) Add version info to Win32 DLLs.
380 [Peter 'Luna' Runestig" <peter@runestig.com>]
382 *) Add new 'medium level' PKCS#12 API. Certificates and keys
383 can be added using this API to created arbitrary PKCS#12
384 files while avoiding the low level API.
386 New options to PKCS12_create(), key or cert can be NULL and
387 will then be omitted from the output file. The encryption
388 algorithm NIDs can be set to -1 for no encryption, the mac
389 iteration count can be set to 0 to omit the mac.
391 Enhance pkcs12 utility by making the -nokeys and -nocerts
392 options work when creating a PKCS#12 file. New option -nomac
393 to omit the mac, NONE can be set for an encryption algorithm.
394 New code is modified to use the enhanced PKCS12_create()
395 instead of the low level API.
398 *) Extend ASN1 encoder to support indefinite length constructed
399 encoding. This can output sequences tags and octet strings in
400 this form. Modify pk7_asn1.c to support indefinite length
401 encoding. This is experimental and needs additional code to
402 be useful, such as an ASN1 bio and some enhanced streaming
405 Extend template encode functionality so that tagging is passed
406 down to the template encoder.
409 *) Let 'openssl req' fail if an argument to '-newkey' is not
410 recognized instead of using RSA as a default.
413 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
414 As these are not official, they are not included in "ALL";
415 the "ECCdraft" ciphersuite group alias can be used to select them.
416 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
418 *) Add ECDH engine support.
419 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
421 *) Add ECDH in new directory crypto/ecdh/.
422 [Douglas Stebila (Sun Microsystems Laboratories)]
424 *) Let BN_rand_range() abort with an error after 100 iterations
425 without success (which indicates a broken PRNG).
428 *) Change BN_mod_sqrt() so that it verifies that the input value
429 is really the square of the return value. (Previously,
430 BN_mod_sqrt would show GIGO behaviour.)
433 *) Add named elliptic curves over binary fields from X9.62, SECG,
434 and WAP/WTLS; add OIDs that were still missing.
436 [Sheueling Chang Shantz and Douglas Stebila
437 (Sun Microsystems Laboratories)]
439 *) Extend the EC library for elliptic curves over binary fields
440 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
443 EC_GF2m_simple_method
447 EC_GROUP_new_curve_GF2m
448 EC_GROUP_set_curve_GF2m
449 EC_GROUP_get_curve_GF2m
450 EC_POINT_set_affine_coordinates_GF2m
451 EC_POINT_get_affine_coordinates_GF2m
452 EC_POINT_set_compressed_coordinates_GF2m
454 Point compression for binary fields is disabled by default for
455 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
458 As binary polynomials are represented as BIGNUMs, various members
459 of the EC_GROUP and EC_POINT data structures can be shared
460 between the implementations for prime fields and binary fields;
461 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
462 are essentially identical to their ..._GFp counterparts.
463 (For simplicity, the '..._GFp' prefix has been dropped from
464 various internal method names.)
466 An internal 'field_div' method (similar to 'field_mul' and
467 'field_sqr') has been added; this is used only for binary fields.
469 [Sheueling Chang Shantz and Douglas Stebila
470 (Sun Microsystems Laboratories)]
472 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
473 through methods ('mul', 'precompute_mult').
475 The generic implementations (now internally called 'ec_wNAF_mul'
476 and 'ec_wNAF_precomputed_mult') remain the default if these
477 methods are undefined.
479 [Sheueling Chang Shantz and Douglas Stebila
480 (Sun Microsystems Laboratories)]
482 *) New function EC_GROUP_get_degree, which is defined through
483 EC_METHOD. For curves over prime fields, this returns the bit
484 length of the modulus.
486 [Sheueling Chang Shantz and Douglas Stebila
487 (Sun Microsystems Laboratories)]
489 *) New functions EC_GROUP_dup, EC_POINT_dup.
490 (These simply call ..._new and ..._copy).
492 [Sheueling Chang Shantz and Douglas Stebila
493 (Sun Microsystems Laboratories)]
495 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
496 Polynomials are represented as BIGNUMs (where the sign bit is not
497 used) in the following functions [macros]:
500 BN_GF2m_sub [= BN_GF2m_add]
501 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
502 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
503 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
505 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
506 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
507 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
508 BN_GF2m_cmp [= BN_ucmp]
510 (Note that only the 'mod' functions are actually for fields GF(2^m).
511 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
513 For some functions, an the irreducible polynomial defining a
514 field can be given as an 'unsigned int[]' with strictly
515 decreasing elements giving the indices of those bits that are set;
516 i.e., p[] represents the polynomial
517 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
519 p[0] > p[1] > ... > p[k] = 0.
520 This applies to the following functions:
525 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
526 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
529 BN_GF2m_mod_solve_quad_arr
533 Conversion can be performed by the following functions:
538 bntest.c has additional tests for binary polynomial arithmetic.
540 Two implementations for BN_GF2m_mod_div() are available.
541 The default algorithm simply uses BN_GF2m_mod_inv() and
542 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
543 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
544 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
546 [Sheueling Chang Shantz and Douglas Stebila
547 (Sun Microsystems Laboratories)]
549 *) Add new error code 'ERR_R_DISABLED' that can be used when some
550 functionality is disabled at compile-time.
551 [Douglas Stebila <douglas.stebila@sun.com>]
553 *) Change default behaviour of 'openssl asn1parse' so that more
554 information is visible when viewing, e.g., a certificate:
556 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
557 mode the content of non-printable OCTET STRINGs is output in a
558 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
559 avoid the appearance of a printable string.
560 [Nils Larsch <nla@trustcenter.de>]
562 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
564 EC_GROUP_set_asn1_flag()
565 EC_GROUP_get_asn1_flag()
566 EC_GROUP_set_point_conversion_form()
567 EC_GROUP_get_point_conversion_form()
568 These control ASN1 encoding details:
569 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
570 has been set to OPENSSL_EC_NAMED_CURVE.
571 - Points are encoded in uncompressed form by default; options for
572 asn1_for are as for point2oct, namely
573 POINT_CONVERSION_COMPRESSED
574 POINT_CONVERSION_UNCOMPRESSED
575 POINT_CONVERSION_HYBRID
577 Also add 'seed' and 'seed_len' members to EC_GROUP with access
581 EC_GROUP_get_seed_len()
582 This is used only for ASN1 purposes (so far).
583 [Nils Larsch <nla@trustcenter.de>]
585 *) Add 'field_type' member to EC_METHOD, which holds the NID
586 of the appropriate field type OID. The new function
587 EC_METHOD_get_field_type() returns this value.
588 [Nils Larsch <nla@trustcenter.de>]
595 providing useful interfaces to EC_POINT_point2oct() and
596 EC_POINT_oct2point().
597 [Nils Larsch <nla@trustcenter.de>]
599 *) Change internals of the EC library so that the functions
600 EC_GROUP_set_generator()
601 EC_GROUP_get_generator()
603 EC_GROUP_get_cofactor()
604 are implemented directly in crypto/ec/ec_lib.c and not dispatched
605 to methods, which would lead to unnecessary code duplication when
606 adding different types of curves.
607 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
609 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
610 arithmetic, and such that modified wNAFs are generated
611 (which avoid length expansion in many cases).
614 *) Add a function EC_GROUP_check_discriminant() (defined via
615 EC_METHOD) that verifies that the curve discriminant is non-zero.
617 Add a function EC_GROUP_check() that makes some sanity tests
618 on a EC_GROUP, its generator and order. This includes
619 EC_GROUP_check_discriminant().
620 [Nils Larsch <nla@trustcenter.de>]
622 *) Add ECDSA in new directory crypto/ecdsa/.
624 Add applications 'openssl ecparam' and 'openssl ecdsa'
625 (these are based on 'openssl dsaparam' and 'openssl dsa').
627 ECDSA support is also included in various other files across the
628 library. Most notably,
629 - 'openssl req' now has a '-newkey ecdsa:file' option;
630 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
631 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
632 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
633 them suitable for ECDSA where domain parameters must be
634 extracted before the specific public key;
635 - ECDSA engine support has been added.
636 [Nils Larsch <nla@trustcenter.de>]
638 *) Include some named elliptic curves, and add OIDs from X9.62,
639 SECG, and WAP/WTLS. Each curve can be obtained from the new
641 EC_GROUP_new_by_nid(),
642 and the list of available named curves can be obtained with
643 EC_get_builtin_curves().
644 Also add a 'curve_name' member to EC_GROUP objects, which can be
648 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
650 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
651 was actually never needed) and in BN_mul(). The removal in BN_mul()
652 required a small change in bn_mul_part_recursive() and the addition
653 of the functions bn_cmp_part_words(), bn_sub_part_words() and
654 bn_add_part_words(), which do the same thing as bn_cmp_words(),
655 bn_sub_words() and bn_add_words() except they take arrays with
659 Changes between 0.9.7c and 0.9.7d [xx XXX XXXX]
661 *) X509 verify fixes. Disable broken certificate workarounds when
662 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
663 keyUsage extension present. Don't accept CRLs with unhandled critical
664 extensions: since verify currently doesn't process CRL extensions this
665 rejects a CRL with *any* critical extensions. Add new verify error codes
669 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
670 A clarification of RFC2560 will require the use of OCTET STRINGs and
671 some implementations cannot handle the current raw format. Since OpenSSL
672 copies and compares OCSP nonces as opaque blobs without any attempt at
673 parsing them this should not create any compatibility issues.
676 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
677 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
678 this HMAC (and other) operations are several times slower than OpenSSL
682 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
683 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
685 *) Use the correct content when signing type "other".
688 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
690 *) Fix various bugs revealed by running the NISCC test suite:
692 Stop out of bounds reads in the ASN1 code when presented with
693 invalid tags (CAN-2003-0543 and CAN-2003-0544).
695 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
697 If verify callback ignores invalid public key errors don't try to check
698 certificate signature with the NULL public key.
702 *) New -ignore_err option in ocsp application to stop the server
703 exiting on the first error in a request.
706 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
707 if the server requested one: as stated in TLS 1.0 and SSL 3.0
711 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
712 extra data after the compression methods not only for TLS 1.0
713 but also for SSL 3.0 (as required by the specification).
714 [Bodo Moeller; problem pointed out by Matthias Loepfe]
716 *) Change X509_certificate_type() to mark the key as exported/exportable
717 when it's 512 *bits* long, not 512 bytes.
720 *) Change AES_cbc_encrypt() so it outputs exact multiple of
721 blocks during encryption.
724 *) Various fixes to base64 BIO and non blocking I/O. On write
725 flushes were not handled properly if the BIO retried. On read
726 data was not being buffered properly and had various logic bugs.
727 This also affects blocking I/O when the data being decoded is a
731 *) Various S/MIME bugfixes and compatibility changes:
732 output correct application/pkcs7 MIME type if
733 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
734 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
735 of files as .eml work). Correctly handle very long lines in MIME
739 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
741 *) Countermeasure against the Klima-Pokorny-Rosa extension of
742 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
743 a protocol version number mismatch like a decryption error
744 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
747 *) Turn on RSA blinding by default in the default implementation
748 to avoid a timing attack. Applications that don't want it can call
749 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
750 They would be ill-advised to do so in most cases.
751 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
753 *) Change RSA blinding code so that it works when the PRNG is not
754 seeded (in this case, the secret RSA exponent is abused as
755 an unpredictable seed -- if it is not unpredictable, there
756 is no point in blinding anyway). Make RSA blinding thread-safe
757 by remembering the creator's thread ID in rsa->blinding and
758 having all other threads use local one-time blinding factors
759 (this requires more computation than sharing rsa->blinding, but
760 avoids excessive locking; and if an RSA object is not shared
761 between threads, blinding will still be very fast).
764 yet to be integrated into this CVS branch:
765 - Geoff's ENGINE_set_default() fix
767 *) Target "mingw" now allows native Windows code to be generated in
768 the Cygwin environment as well as with the MinGW compiler.
771 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
773 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
774 via timing by performing a MAC computation even if incorrrect
775 block cipher padding has been found. This is a countermeasure
776 against active attacks where the attacker has to distinguish
777 between bad padding and a MAC verification error. (CAN-2003-0078)
779 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
780 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
781 Martin Vuagnoux (EPFL, Ilion)]
783 *) Make the no-err option work as intended. The intention with no-err
784 is not to have the whole error stack handling routines removed from
785 libcrypto, it's only intended to remove all the function name and
786 reason texts, thereby removing some of the footprint that may not
787 be interesting if those errors aren't displayed anyway.
789 NOTE: it's still possible for any application or module to have it's
790 own set of error texts inserted. The routines are there, just not
791 used by default when no-err is given.
794 *) Add support for FreeBSD on IA64.
795 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
797 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
798 Kerberos function mit_des_cbc_cksum(). Before this change,
799 the value returned by DES_cbc_cksum() was like the one from
800 mit_des_cbc_cksum(), except the bytes were swapped.
801 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
803 *) Allow an application to disable the automatic SSL chain building.
804 Before this a rather primitive chain build was always performed in
805 ssl3_output_cert_chain(): an application had no way to send the
806 correct chain if the automatic operation produced an incorrect result.
808 Now the chain builder is disabled if either:
810 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
812 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
814 The reasoning behind this is that an application would not want the
815 auto chain building to take place if extra chain certificates are
816 present and it might also want a means of sending no additional
817 certificates (for example the chain has two certificates and the
821 *) Add the possibility to build without the ENGINE framework.
822 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
824 *) Under Win32 gmtime() can return NULL: check return value in
825 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
828 *) DSA routines: under certain error conditions uninitialized BN objects
829 could be freed. Solution: make sure initialization is performed early
830 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
831 Nils Larsch <nla@trustcenter.de> via PR#459)
834 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
835 checked on reconnect on the client side, therefore session resumption
836 could still fail with a "ssl session id is different" error. This
837 behaviour is masked when SSL_OP_ALL is used due to
838 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
839 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
843 *) IA-32 assembler support enhancements: unified ELF targets, support
844 for SCO/Caldera platforms, fix for Cygwin shared build.
847 *) Add support for FreeBSD on sparc64. As a consequence, support for
848 FreeBSD on non-x86 processors is separate from x86 processors on
849 the config script, much like the NetBSD support.
850 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
852 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
854 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
857 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
858 code (06) was taken as the first octet of the session ID and the last
859 octet was ignored consequently. As a result SSLv2 client side session
860 caching could not have worked due to the session ID mismatch between
862 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
866 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
867 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
871 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
872 seems that in spite of existing for more than a year, many application
873 author have done nothing to provide the necessary callbacks, which
874 means that this particular engine will not work properly anywhere.
875 This is a very unfortunate situation which forces us, in the name
876 of usability, to give the hw_ncipher.c a static lock, which is part
878 NOTE: This is for the 0.9.7 series ONLY. This hack will never
879 appear in 0.9.8 or later. We EXPECT application authors to have
880 dealt properly with this when 0.9.8 is released (unless we actually
881 make such changes in the libcrypto locking code that changes will
882 have to be made anyway).
885 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
886 octets have been read, EOF or an error occurs. Without this change
887 some truncated ASN1 structures will not produce an error.
890 *) Disable Heimdal support, since it hasn't been fully implemented.
891 Still give the possibility to force the use of Heimdal, but with
892 warnings and a request that patches get sent to openssl-dev.
895 *) Add the VC-CE target, introduce the WINCE sysname, and add
896 INSTALL.WCE and appropriate conditionals to make it build.
897 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
899 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
900 cygssl-x.y.z.dll, where x, y and z are the major, minor and
901 edit numbers of the version.
902 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
904 *) Introduce safe string copy and catenation functions
905 (BUF_strlcpy() and BUF_strlcat()).
906 [Ben Laurie (CHATS) and Richard Levitte]
908 *) Avoid using fixed-size buffers for one-line DNs.
911 *) Add BUF_MEM_grow_clean() to avoid information leakage when
912 resizing buffers containing secrets, and use where appropriate.
915 *) Avoid using fixed size buffers for configuration file location.
918 *) Avoid filename truncation for various CA files.
921 *) Use sizeof in preference to magic numbers.
924 *) Avoid filename truncation in cert requests.
927 *) Add assertions to check for (supposedly impossible) buffer
931 *) Don't cache truncated DNS entries in the local cache (this could
932 potentially lead to a spoofing attack).
935 *) Fix various buffers to be large enough for hex/decimal
936 representations in a platform independent manner.
939 *) Add CRYPTO_realloc_clean() to avoid information leakage when
940 resizing buffers containing secrets, and use where appropriate.
943 *) Add BIO_indent() to avoid much slightly worrying code to do
947 *) Convert sprintf()/BIO_puts() to BIO_printf().
950 *) buffer_gets() could terminate with the buffer only half
954 *) Add assertions to prevent user-supplied crypto functions from
955 overflowing internal buffers by having large block sizes, etc.
958 *) New OPENSSL_assert() macro (similar to assert(), but enabled
962 *) Eliminate unused copy of key in RC4.
965 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
968 *) Fix off-by-one error in EGD path.
971 *) If RANDFILE path is too long, ignore instead of truncating.
974 *) Eliminate unused and incorrectly sized X.509 structure
978 *) Eliminate unused and dangerous function knumber().
981 *) Eliminate unused and dangerous structure, KSSL_ERR.
984 *) Protect against overlong session ID context length in an encoded
985 session object. Since these are local, this does not appear to be
989 *) Change from security patch (see 0.9.6e below) that did not affect
990 the 0.9.6 release series:
992 Remote buffer overflow in SSL3 protocol - an attacker could
993 supply an oversized master key in Kerberos-enabled versions.
997 *) Change the SSL kerb5 codes to match RFC 2712.
1000 *) Make -nameopt work fully for req and add -reqopt switch.
1001 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1003 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1004 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1006 *) Make sure tests can be performed even if the corresponding algorithms
1007 have been removed entirely. This was also the last step to make
1008 OpenSSL compilable with DJGPP under all reasonable conditions.
1009 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1011 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1012 to allow version independent disabling of normally unselected ciphers,
1013 which may be activated as a side-effect of selecting a single cipher.
1015 (E.g., cipher list string "RSA" enables ciphersuites that are left
1016 out of "ALL" because they do not provide symmetric encryption.
1017 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1018 [Lutz Jaenicke, Bodo Moeller]
1020 *) Add appropriate support for separate platform-dependent build
1021 directories. The recommended way to make a platform-dependent
1022 build directory is the following (tested on Linux), maybe with
1025 # Place yourself outside of the OpenSSL source tree. In
1026 # this example, the environment variable OPENSSL_SOURCE
1027 # is assumed to contain the absolute OpenSSL source directory.
1028 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1029 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1030 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1031 mkdir -p `dirname $F`
1032 ln -s $OPENSSL_SOURCE/$F $F
1035 To be absolutely sure not to disturb the source tree, a "make clean"
1036 is a good thing. If it isn't successfull, don't worry about it,
1037 it probably means the source directory is very clean.
1040 *) Make sure any ENGINE control commands make local copies of string
1041 pointers passed to them whenever necessary. Otherwise it is possible
1042 the caller may have overwritten (or deallocated) the original string
1043 data when a later ENGINE operation tries to use the stored values.
1044 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1046 *) Improve diagnostics in file reading and command-line digests.
1047 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1049 *) Add AES modes CFB and OFB to the object database. Correct an
1050 error in AES-CFB decryption.
1053 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1054 allows existing EVP_CIPHER_CTX structures to be reused after
1055 calling EVP_*Final(). This behaviour is used by encryption
1056 BIOs and some applications. This has the side effect that
1057 applications must explicitly clean up cipher contexts with
1058 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1061 *) Check the values of dna and dnb in bn_mul_recursive before calling
1062 bn_mul_comba (a non zero value means the a or b arrays do not contain
1063 n2 elements) and fallback to bn_mul_normal if either is not zero.
1066 *) Fix escaping of non-ASCII characters when using the -subj option
1067 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1070 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1071 form for "surname", serialNumber has no short form.
1072 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1073 therefore remove "mail" short name for "internet 7".
1074 The OID for unique identifiers in X509 certificates is
1075 x500UniqueIdentifier, not uniqueIdentifier.
1076 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1079 *) Add an "init" command to the ENGINE config module and auto initialize
1080 ENGINEs. Without any "init" command the ENGINE will be initialized
1081 after all ctrl commands have been executed on it. If init=1 the
1082 ENGINE is initailized at that point (ctrls before that point are run
1083 on the uninitialized ENGINE and after on the initialized one). If
1084 init=0 then the ENGINE will not be iniatialized at all.
1087 *) Fix the 'app_verify_callback' interface so that the user-defined
1088 argument is actually passed to the callback: In the
1089 SSL_CTX_set_cert_verify_callback() prototype, the callback
1090 declaration has been changed from
1093 int (*cb)(X509_STORE_CTX *,void *);
1094 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1095 i=s->ctx->app_verify_callback(&ctx)
1096 has been changed into
1097 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1099 To update applications using SSL_CTX_set_cert_verify_callback(),
1100 a dummy argument can be added to their callback functions.
1101 [D. K. Smetters <smetters@parc.xerox.com>]
1103 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1104 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1106 *) Add and OPENSSL_LOAD_CONF define which will cause
1107 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1108 This allows older applications to transparently support certain
1109 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1110 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1111 load the config file and OPENSSL_add_all_algorithms_conf() which will
1112 always load it have also been added.
1115 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1116 Adjust NIDs and EVP layer.
1117 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1119 *) Config modules support in openssl utility.
1121 Most commands now load modules from the config file,
1122 though in a few (such as version) this isn't done
1123 because it couldn't be used for anything.
1125 In the case of ca and req the config file used is
1126 the same as the utility itself: that is the -config
1127 command line option can be used to specify an
1131 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1132 use "openssl_conf" if filename is NULL use default openssl config file.
1135 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1136 config section name. Add a new flag to tolerate a missing config file
1137 and move code to CONF_modules_load_file().
1140 *) Support for crypto accelerator cards from Accelerated Encryption
1141 Processing, www.aep.ie. (Use engine 'aep')
1142 The support was copied from 0.9.6c [engine] and adapted/corrected
1143 to work with the new engine framework.
1144 [AEP Inc. and Richard Levitte]
1146 *) Support for SureWare crypto accelerator cards from Baltimore
1147 Technologies. (Use engine 'sureware')
1148 The support was copied from 0.9.6c [engine] and adapted
1149 to work with the new engine framework.
1152 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1153 make the newer ENGINE framework commands for the CHIL engine work.
1154 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1156 *) Make it possible to produce shared libraries on ReliantUNIX.
1157 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1159 *) Add the configuration target debug-linux-ppro.
1160 Make 'openssl rsa' use the general key loading routines
1161 implemented in apps.c, and make those routines able to
1162 handle the key format FORMAT_NETSCAPE and the variant
1164 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1166 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1167 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1169 *) Add -keyform to rsautl, and document -engine.
1170 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1172 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1173 BIO_R_NO_SUCH_FILE error code rather than the generic
1174 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1177 *) Add new functions
1179 ERR_peek_last_error_line
1180 ERR_peek_last_error_line_data.
1181 These are similar to
1184 ERR_peek_error_line_data,
1185 but report on the latest error recorded rather than the first one
1186 still in the error queue.
1187 [Ben Laurie, Bodo Moeller]
1189 *) default_algorithms option in ENGINE config module. This allows things
1191 default_algorithms = ALL
1192 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1195 *) Prelminary ENGINE config module.
1198 *) New experimental application configuration code.
1201 *) Change the AES code to follow the same name structure as all other
1202 symmetric ciphers, and behave the same way. Move everything to
1203 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1204 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1206 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1207 [Ben Laurie and Theo de Raadt]
1209 *) Add option to output public keys in req command.
1210 [Massimiliano Pala madwolf@openca.org]
1212 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1213 (up to about 10% better than before for P-192 and P-224).
1216 *) New functions/macros
1218 SSL_CTX_set_msg_callback(ctx, cb)
1219 SSL_CTX_set_msg_callback_arg(ctx, arg)
1220 SSL_set_msg_callback(ssl, cb)
1221 SSL_set_msg_callback_arg(ssl, arg)
1223 to request calling a callback function
1225 void cb(int write_p, int version, int content_type,
1226 const void *buf, size_t len, SSL *ssl, void *arg)
1228 whenever a protocol message has been completely received
1229 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1230 protocol version according to which the SSL library interprets
1231 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1232 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1233 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1234 specification (change_cipher_spec(20), alert(21), handshake(22)).
1235 'buf' and 'len' point to the actual message, 'ssl' to the
1236 SSL object, and 'arg' is the application-defined value set by
1237 SSL[_CTX]_set_msg_callback_arg().
1239 'openssl s_client' and 'openssl s_server' have new '-msg' options
1240 to enable a callback that displays all protocol messages.
1243 *) Change the shared library support so shared libraries are built as
1244 soon as the corresponding static library is finished, and thereby get
1245 openssl and the test programs linked against the shared library.
1246 This still only happens when the keyword "shard" has been given to
1247 the configuration scripts.
1249 NOTE: shared library support is still an experimental thing, and
1250 backward binary compatibility is still not guaranteed.
1251 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1253 *) Add support for Subject Information Access extension.
1254 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1256 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1257 additional bytes when new memory had to be allocated, not just
1258 when reusing an existing buffer.
1261 *) New command line and configuration option 'utf8' for the req command.
1262 This allows field values to be specified as UTF8 strings.
1265 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1266 runs for the former and machine-readable output for the latter.
1269 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1270 of the e-mail address in the DN (i.e., it will go into a certificate
1271 extension only). The new configuration file option 'email_in_dn = no'
1272 has the same effect.
1273 [Massimiliano Pala madwolf@openca.org]
1275 *) Change all functions with names starting with des_ to be starting
1276 with DES_ instead. Add wrappers that are compatible with libdes,
1277 but are named _ossl_old_des_*. Finally, add macros that map the
1278 des_* symbols to the corresponding _ossl_old_des_* if libdes
1279 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1280 desired, the des_* symbols will be mapped to DES_*, with one
1283 Since we provide two compatibility mappings, the user needs to
1284 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1285 compatibility is desired. The default (i.e., when that macro
1286 isn't defined) is OpenSSL 0.9.6c compatibility.
1288 There are also macros that enable and disable the support of old
1289 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1290 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1291 are defined, the default will apply: to support the old des routines.
1293 In either case, one must include openssl/des.h to get the correct
1294 definitions. Do not try to just include openssl/des_old.h, that
1297 NOTE: This is a major break of an old API into a new one. Software
1298 authors are encouraged to switch to the DES_ style functions. Some
1299 time in the future, des_old.h and the libdes compatibility functions
1300 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1301 default), and then completely removed.
1304 *) Test for certificates which contain unsupported critical extensions.
1305 If such a certificate is found during a verify operation it is
1306 rejected by default: this behaviour can be overridden by either
1307 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1308 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1309 X509_supported_extension() has also been added which returns 1 if a
1310 particular extension is supported.
1313 *) Modify the behaviour of EVP cipher functions in similar way to digests
1314 to retain compatibility with existing code.
1317 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1318 compatibility with existing code. In particular the 'ctx' parameter does
1319 not have to be to be initialized before the call to EVP_DigestInit() and
1320 it is tidied up after a call to EVP_DigestFinal(). New function
1321 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1322 EVP_MD_CTX_copy() changed to not require the destination to be
1323 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1324 requires the destination to be valid.
1326 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1327 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1330 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1331 so that complete 'Handshake' protocol structures are kept in memory
1332 instead of overwriting 'msg_type' and 'length' with 'body' data.
1335 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1336 [Massimo Santin via Richard Levitte]
1338 *) Major restructuring to the underlying ENGINE code. This includes
1339 reduction of linker bloat, separation of pure "ENGINE" manipulation
1340 (initialisation, etc) from functionality dealing with implementations
1341 of specific crypto iterfaces. This change also introduces integrated
1342 support for symmetric ciphers and digest implementations - so ENGINEs
1343 can now accelerate these by providing EVP_CIPHER and EVP_MD
1344 implementations of their own. This is detailed in crypto/engine/README
1345 as it couldn't be adequately described here. However, there are a few
1346 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1347 were changed in the original introduction of ENGINE code have now
1348 reverted back - the hooking from this code to ENGINE is now a good
1349 deal more passive and at run-time, operations deal directly with
1350 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1351 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1352 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1353 they were not being used by the framework as there is no concept of a
1354 BIGNUM_METHOD and they could not be generalised to the new
1355 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1356 ENGINE_cpy() has been removed as it cannot be consistently defined in
1360 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1363 *) Change mkdef.pl to sort symbols that get the same entry number,
1364 and make sure the automatically generated functions ERR_load_*
1365 become part of libeay.num as well.
1368 *) New function SSL_renegotiate_pending(). This returns true once
1369 renegotiation has been requested (either SSL_renegotiate() call
1370 or HelloRequest/ClientHello receveived from the peer) and becomes
1371 false once a handshake has been completed.
1372 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1373 sends a HelloRequest, but does not ensure that a handshake takes
1374 place. SSL_renegotiate_pending() is useful for checking if the
1375 client has followed the request.)
1378 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1379 By default, clients may request session resumption even during
1380 renegotiation (if session ID contexts permit); with this option,
1381 session resumption is possible only in the first handshake.
1383 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1384 more bits available for options that should not be part of
1385 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1388 *) Add some demos for certificate and certificate request creation.
1391 *) Make maximum certificate chain size accepted from the peer application
1392 settable (SSL*_get/set_max_cert_list()), as proposed by
1393 "Douglas E. Engert" <deengert@anl.gov>.
1396 *) Add support for shared libraries for Unixware-7
1397 (Boyd Lynn Gerber <gerberb@zenez.com>).
1400 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1401 be done prior to destruction. Use this to unload error strings from
1402 ENGINEs that load their own error strings. NB: This adds two new API
1403 functions to "get" and "set" this destroy handler in an ENGINE.
1406 *) Alter all existing ENGINE implementations (except "openssl" and
1407 "openbsd") to dynamically instantiate their own error strings. This
1408 makes them more flexible to be built both as statically-linked ENGINEs
1409 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1410 Also, add stub code to each that makes building them as self-contained
1411 shared-libraries easier (see README.ENGINE).
1414 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1415 implementations into applications that are completely implemented in
1416 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1417 commands that can be used to configure what shared-library to load and
1418 to control aspects of the way it is handled. Also, made an update to
1419 the README.ENGINE file that brings its information up-to-date and
1420 provides some information and instructions on the "dynamic" ENGINE
1421 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1424 *) Make it possible to unload ranges of ERR strings with a new
1425 "ERR_unload_strings" function.
1428 *) Add a copy() function to EVP_MD.
1431 *) Make EVP_MD routines take a context pointer instead of just the
1432 md_data void pointer.
1435 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1436 that the digest can only process a single chunk of data
1437 (typically because it is provided by a piece of
1438 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1439 is only going to provide a single chunk of data, and hence the
1440 framework needn't accumulate the data for oneshot drivers.
1443 *) As with "ERR", make it possible to replace the underlying "ex_data"
1444 functions. This change also alters the storage and management of global
1445 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1446 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1447 index counters. The API functions that use this state have been changed
1448 to take a "class_index" rather than pointers to the class's local STACK
1449 and counter, and there is now an API function to dynamically create new
1450 classes. This centralisation allows us to (a) plug a lot of the
1451 thread-safety problems that existed, and (b) makes it possible to clean
1452 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1453 such data would previously have always leaked in application code and
1454 workarounds were in place to make the memory debugging turn a blind eye
1455 to it. Application code that doesn't use this new function will still
1456 leak as before, but their memory debugging output will announce it now
1457 rather than letting it slide.
1459 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1460 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1461 has a return value to indicate success or failure.
1464 *) Make it possible to replace the underlying "ERR" functions such that the
1465 global state (2 LHASH tables and 2 locks) is only used by the "default"
1466 implementation. This change also adds two functions to "get" and "set"
1467 the implementation prior to it being automatically set the first time
1468 any other ERR function takes place. Ie. an application can call "get",
1469 pass the return value to a module it has just loaded, and that module
1470 can call its own "set" function using that value. This means the
1471 module's "ERR" operations will use (and modify) the error state in the
1472 application and not in its own statically linked copy of OpenSSL code.
1475 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1476 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1477 the operation, and provides a more encapsulated way for external code
1478 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1479 to use these functions rather than manually incrementing the counts.
1481 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1484 *) Add EVP test program.
1487 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1490 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1491 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1492 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1493 These allow a CRL to be built without having to access X509_CRL fields
1494 directly. Modify 'ca' application to use new functions.
1497 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1498 bug workarounds. Rollback attack detection is a security feature.
1499 The problem will only arise on OpenSSL servers when TLSv1 is not
1500 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1501 Software authors not wanting to support TLSv1 will have special reasons
1502 for their choice and can explicitly enable this option.
1503 [Bodo Moeller, Lutz Jaenicke]
1505 *) Rationalise EVP so it can be extended: don't include a union of
1506 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1507 (similar to those existing for EVP_CIPHER_CTX).
1512 EVP_MD_CTX_init(&md); /* new function call */
1513 EVP_DigestInit(&md, EVP_sha1());
1514 EVP_DigestUpdate(&md, in, len);
1515 EVP_DigestFinal(&md, out, NULL);
1516 EVP_MD_CTX_cleanup(&md); /* new function call */
1520 *) Make DES key schedule conform to the usual scheme, as well as
1521 correcting its structure. This means that calls to DES functions
1522 now have to pass a pointer to a des_key_schedule instead of a
1523 plain des_key_schedule (which was actually always a pointer
1526 des_key_schedule ks;
1528 des_set_key_checked(..., &ks);
1529 des_ncbc_encrypt(..., &ks, ...);
1531 (Note that a later change renames 'des_...' into 'DES_...'.)
1534 *) Initial reduction of linker bloat: the use of some functions, such as
1535 PEM causes large amounts of unused functions to be linked in due to
1536 poor organisation. For example pem_all.c contains every PEM function
1537 which has a knock on effect of linking in large amounts of (unused)
1538 ASN1 code. Grouping together similar functions and splitting unrelated
1539 functions prevents this.
1542 *) Cleanup of EVP macros.
1545 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1546 correct _ecb suffix.
1549 *) Add initial OCSP responder support to ocsp application. The
1550 revocation information is handled using the text based index
1551 use by the ca application. The responder can either handle
1552 requests generated internally, supplied in files (for example
1553 via a CGI script) or using an internal minimal server.
1556 *) Add configuration choices to get zlib compression for TLS.
1559 *) Changes to Kerberos SSL for RFC 2712 compliance:
1560 1. Implemented real KerberosWrapper, instead of just using
1561 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1562 2. Implemented optional authenticator field of KerberosWrapper.
1564 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1565 and authenticator structs; see crypto/krb5/.
1567 Generalized Kerberos calls to support multiple Kerberos libraries.
1568 [Vern Staats <staatsvr@asc.hpc.mil>,
1569 Jeffrey Altman <jaltman@columbia.edu>
1570 via Richard Levitte]
1572 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1573 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1574 values for each of the key sizes rather than having just
1575 parameters (and 'speed' generating keys each time).
1578 *) Speed up EVP routines.
1581 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1582 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1583 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1584 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1586 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1587 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1588 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1591 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1593 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1596 *) Added the OS2-EMX target.
1597 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1599 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1600 to support NCONF routines in extension code. New function CONF_set_nconf()
1601 to allow functions which take an NCONF to also handle the old LHASH
1602 structure: this means that the old CONF compatible routines can be
1603 retained (in particular wrt extensions) without having to duplicate the
1604 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1607 *) Enhance the general user interface with mechanisms for inner control
1608 and with possibilities to have yes/no kind of prompts.
1611 *) Change all calls to low level digest routines in the library and
1612 applications to use EVP. Add missing calls to HMAC_cleanup() and
1613 don't assume HMAC_CTX can be copied using memcpy().
1614 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1616 *) Add the possibility to control engines through control names but with
1617 arbitrary arguments instead of just a string.
1618 Change the key loaders to take a UI_METHOD instead of a callback
1619 function pointer. NOTE: this breaks binary compatibility with earlier
1620 versions of OpenSSL [engine].
1621 Adapt the nCipher code for these new conditions and add a card insertion
1625 *) Enhance the general user interface with mechanisms to better support
1626 dialog box interfaces, application-defined prompts, the possibility
1627 to use defaults (for example default passwords from somewhere else)
1628 and interrupts/cancellations.
1631 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1632 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1635 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1636 tidy up some unnecessarily weird code in 'sk_new()').
1637 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1639 *) Change the key loading routines for ENGINEs to use the same kind
1640 callback (pem_password_cb) as all other routines that need this
1644 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1645 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1646 than this minimum value is recommended.
1649 *) New random seeder for OpenVMS, using the system process statistics
1650 that are easily reachable.
1653 *) Windows apparently can't transparently handle global
1654 variables defined in DLLs. Initialisations such as:
1656 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1658 wont compile. This is used by the any applications that need to
1659 declare their own ASN1 modules. This was fixed by adding the option
1660 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1661 needed for static libraries under Win32.
1664 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1665 setting of purpose and trust fields. New X509_STORE trust and
1666 purpose functions and tidy up setting in other SSL functions.
1669 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1670 structure. These are inherited by X509_STORE_CTX when it is
1671 initialised. This allows various defaults to be set in the
1672 X509_STORE structure (such as flags for CRL checking and custom
1673 purpose or trust settings) for functions which only use X509_STORE_CTX
1674 internally such as S/MIME.
1676 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1677 trust settings if they are not set in X509_STORE. This allows X509_STORE
1678 purposes and trust (in S/MIME for example) to override any set by default.
1680 Add command line options for CRL checking to smime, s_client and s_server
1684 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1685 are set then the CRL is looked up in the X509_STORE structure and
1686 its validity and signature checked, then if the certificate is found
1687 in the CRL the verify fails with a revoked error.
1689 Various new CRL related callbacks added to X509_STORE_CTX structure.
1691 Command line options added to 'verify' application to support this.
1693 This needs some additional work, such as being able to handle multiple
1694 CRLs with different times, extension based lookup (rather than just
1695 by subject name) and ultimately more complete V2 CRL extension
1699 *) Add a general user interface API (crypto/ui/). This is designed
1700 to replace things like des_read_password and friends (backward
1701 compatibility functions using this new API are provided).
1702 The purpose is to remove prompting functions from the DES code
1703 section as well as provide for prompting through dialog boxes in
1704 a window system and the like.
1707 *) Add "ex_data" support to ENGINE so implementations can add state at a
1708 per-structure level rather than having to store it globally.
1711 *) Make it possible for ENGINE structures to be copied when retrieved by
1712 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1713 This causes the "original" ENGINE structure to act like a template,
1714 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1715 operational state can be localised to each ENGINE structure, despite the
1716 fact they all share the same "methods". New ENGINE structures returned in
1717 this case have no functional references and the return value is the single
1718 structural reference. This matches the single structural reference returned
1719 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1723 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1724 needs to match any other type at all we need to manually clear the
1728 *) Changes to the "openssl engine" utility to include;
1729 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1730 about an ENGINE's available control commands.
1731 - executing control commands from command line arguments using the
1732 '-pre' and '-post' switches. '-post' is only used if '-t' is
1733 specified and the ENGINE is successfully initialised. The syntax for
1734 the individual commands are colon-separated, for example;
1735 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1738 *) New dynamic control command support for ENGINEs. ENGINEs can now
1739 declare their own commands (numbers), names (strings), descriptions,
1740 and input types for run-time discovery by calling applications. A
1741 subset of these commands are implicitly classed as "executable"
1742 depending on their input type, and only these can be invoked through
1743 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1744 can be based on user input, config files, etc). The distinction is
1745 that "executable" commands cannot return anything other than a boolean
1746 result and can only support numeric or string input, whereas some
1747 discoverable commands may only be for direct use through
1748 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1749 pointers, or other custom uses. The "executable" commands are to
1750 support parameterisations of ENGINE behaviour that can be
1751 unambiguously defined by ENGINEs and used consistently across any
1752 OpenSSL-based application. Commands have been added to all the
1753 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1754 control over shared-library paths without source code alterations.
1757 *) Changed all ENGINE implementations to dynamically allocate their
1758 ENGINEs rather than declaring them statically. Apart from this being
1759 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1760 this also allows the implementations to compile without using the
1761 internal engine_int.h header.
1764 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1765 'const' value. Any code that should be able to modify a RAND_METHOD
1766 should already have non-const pointers to it (ie. they should only
1767 modify their own ones).
1770 *) Made a variety of little tweaks to the ENGINE code.
1771 - "atalla" and "ubsec" string definitions were moved from header files
1772 to C code. "nuron" string definitions were placed in variables
1773 rather than hard-coded - allowing parameterisation of these values
1774 later on via ctrl() commands.
1775 - Removed unused "#if 0"'d code.
1776 - Fixed engine list iteration code so it uses ENGINE_free() to release
1777 structural references.
1778 - Constified the RAND_METHOD element of ENGINE structures.
1779 - Constified various get/set functions as appropriate and added
1780 missing functions (including a catch-all ENGINE_cpy that duplicates
1781 all ENGINE values onto a new ENGINE except reference counts/state).
1782 - Removed NULL parameter checks in get/set functions. Setting a method
1783 or function to NULL is a way of cancelling out a previously set
1784 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1785 and doesn't justify the extra error symbols and code.
1786 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1787 flags from engine_int.h to engine.h.
1788 - Changed prototypes for ENGINE handler functions (init(), finish(),
1789 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1792 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1793 to the algorithm using long division. The binary algorithm can be
1794 used only if the modulus is odd. On 32-bit systems, it is faster
1795 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1796 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1797 up to 450 bits. In 64-bit environments, the binary algorithm
1798 appears to be advantageous for much longer moduli; here we use it
1799 for moduli up to 2048 bits.
1802 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1803 could not support the combine flag in choice fields.
1806 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1807 extensions from a certificate request to the certificate.
1810 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1811 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1812 file: this allows the display of the certificate about to be
1813 signed to be customised, to allow certain fields to be included
1814 or excluded and extension details. The old system didn't display
1815 multicharacter strings properly, omitted fields not in the policy
1816 and couldn't display additional details such as extensions.
1819 *) Function EC_POINTs_mul for multiple scalar multiplication
1820 of an arbitrary number of elliptic curve points
1821 \sum scalars[i]*points[i],
1822 optionally including the generator defined for the EC_GROUP:
1823 scalar*generator + \sum scalars[i]*points[i].
1825 EC_POINT_mul is a simple wrapper function for the typical case
1826 that the point list has just one item (besides the optional
1830 *) First EC_METHODs for curves over GF(p):
1832 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1833 operations and provides various method functions that can also
1834 operate with faster implementations of modular arithmetic.
1836 EC_GFp_mont_method() reuses most functions that are part of
1837 EC_GFp_simple_method, but uses Montgomery arithmetic.
1839 [Bodo Moeller; point addition and point doubling
1840 implementation directly derived from source code provided by
1841 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1843 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1844 crypto/ec/ec_lib.c):
1846 Curves are EC_GROUP objects (with an optional group generator)
1847 based on EC_METHODs that are built into the library.
1849 Points are EC_POINT objects based on EC_GROUP objects.
1851 Most of the framework would be able to handle curves over arbitrary
1852 finite fields, but as there are no obvious types for fields other
1853 than GF(p), some functions are limited to that for now.
1856 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1857 that the file contains a complete HTTP response.
1860 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1861 change the def and num file printf format specifier from "%-40sXXX"
1862 to "%-39s XXX". The latter will always guarantee a space after the
1863 field while the former will cause them to run together if the field
1864 is 40 of more characters long.
1867 *) Constify the cipher and digest 'method' functions and structures
1868 and modify related functions to take constant EVP_MD and EVP_CIPHER
1872 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1873 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1876 *) Modify EVP_Digest*() routines so they now return values. Although the
1877 internal software routines can never fail additional hardware versions
1881 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1883 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1884 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1887 ERR_R_NESTED_ASN1_ERROR
1889 ERR_R_MISSING_ASN1_EOS
1890 were 4 .. 9, conflicting with
1891 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1893 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1894 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1896 Add new error code 'ERR_R_INTERNAL_ERROR'.
1899 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1903 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1904 sets the subject name for a new request or supersedes the
1905 subject name in a given request. Formats that can be parsed are
1906 'CN=Some Name, OU=myOU, C=IT'
1908 'CN=Some Name/OU=myOU/C=IT'.
1910 Add options '-batch' and '-verbose' to 'openssl req'.
1911 [Massimiliano Pala <madwolf@hackmasters.net>]
1913 *) Introduce the possibility to access global variables through
1914 functions on platform were that's the best way to handle exporting
1915 global variables in shared libraries. To enable this functionality,
1916 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1917 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1918 is normally done by Configure or something similar).
1920 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1921 in the source file (foo.c) like this:
1923 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1924 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1926 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1927 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1929 OPENSSL_DECLARE_GLOBAL(int,foo);
1930 #define foo OPENSSL_GLOBAL_REF(foo)
1931 OPENSSL_DECLARE_GLOBAL(double,bar);
1932 #define bar OPENSSL_GLOBAL_REF(bar)
1934 The #defines are very important, and therefore so is including the
1935 header file everywhere where the defined globals are used.
1937 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1938 of ASN.1 items, but that structure is a bit different.
1940 The largest change is in util/mkdef.pl which has been enhanced with
1941 better and easier to understand logic to choose which symbols should
1942 go into the Windows .def files as well as a number of fixes and code
1943 cleanup (among others, algorithm keywords are now sorted
1944 lexicographically to avoid constant rewrites).
1947 *) In BN_div() keep a copy of the sign of 'num' before writing the
1948 result to 'rm' because if rm==num the value will be overwritten
1949 and produce the wrong result if 'num' is negative: this caused
1950 problems with BN_mod() and BN_nnmod().
1953 *) Function OCSP_request_verify(). This checks the signature on an
1954 OCSP request and verifies the signer certificate. The signer
1955 certificate is just checked for a generic purpose and OCSP request
1959 *) Add OCSP_check_validity() function to check the validity of OCSP
1960 responses. OCSP responses are prepared in real time and may only
1961 be a few seconds old. Simply checking that the current time lies
1962 between thisUpdate and nextUpdate max reject otherwise valid responses
1963 caused by either OCSP responder or client clock inaccuracy. Instead
1964 we allow thisUpdate and nextUpdate to fall within a certain period of
1965 the current time. The age of the response can also optionally be
1966 checked. Two new options -validity_period and -status_age added to
1970 *) If signature or public key algorithm is unrecognized print out its
1971 OID rather that just UNKNOWN.
1974 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1975 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1976 ID to be generated from the issuer certificate alone which can then be
1977 passed to OCSP_id_issuer_cmp().
1980 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1981 ASN1 modules to export functions returning ASN1_ITEM pointers
1982 instead of the ASN1_ITEM structures themselves. This adds several
1983 new macros which allow the underlying ASN1 function/structure to
1984 be accessed transparently. As a result code should not use ASN1_ITEM
1985 references directly (such as &X509_it) but instead use the relevant
1986 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1987 use of the new ASN1 code on platforms where exporting structures
1988 is problematical (for example in shared libraries) but exporting
1989 functions returning pointers to structures is not.
1992 *) Add support for overriding the generation of SSL/TLS session IDs.
1993 These callbacks can be registered either in an SSL_CTX or per SSL.
1994 The purpose of this is to allow applications to control, if they wish,
1995 the arbitrary values chosen for use as session IDs, particularly as it
1996 can be useful for session caching in multiple-server environments. A
1997 command-line switch for testing this (and any client code that wishes
1998 to use such a feature) has been added to "s_server".
1999 [Geoff Thorpe, Lutz Jaenicke]
2001 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2002 of the form '#if defined(...) || defined(...) || ...' and
2003 '#if !defined(...) && !defined(...) && ...'. This also avoids
2004 the growing number of special cases it was previously handling.
2007 *) Make all configuration macros available for application by making
2008 sure they are available in opensslconf.h, by giving them names starting
2009 with "OPENSSL_" to avoid conflicts with other packages and by making
2010 sure e_os2.h will cover all platform-specific cases together with
2012 Additionally, it is now possible to define configuration/platform-
2013 specific names (called "system identities"). In the C code, these
2014 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2015 macro with the name beginning with "OPENSSL_SYS_", which is determined
2016 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2020 *) New option -set_serial to 'req' and 'x509' this allows the serial
2021 number to use to be specified on the command line. Previously self
2022 signed certificates were hard coded with serial number 0 and the
2023 CA options of 'x509' had to use a serial number in a file which was
2027 *) New options to 'ca' utility to support V2 CRL entry extensions.
2028 Currently CRL reason, invalidity date and hold instruction are
2029 supported. Add new CRL extensions to V3 code and some new objects.
2032 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2033 disable standard block padding (aka PKCS#5 padding) in the EVP
2034 API, which was previously mandatory. This means that the data is
2035 not padded in any way and so the total length much be a multiple
2036 of the block size, otherwise an error occurs.
2039 *) Initial (incomplete) OCSP SSL support.
2042 *) New function OCSP_parse_url(). This splits up a URL into its host,
2043 port and path components: primarily to parse OCSP URLs. New -url
2044 option to ocsp utility.
2047 *) New nonce behavior. The return value of OCSP_check_nonce() now
2048 reflects the various checks performed. Applications can decide
2049 whether to tolerate certain situations such as an absent nonce
2050 in a response when one was present in a request: the ocsp application
2051 just prints out a warning. New function OCSP_add1_basic_nonce()
2052 this is to allow responders to include a nonce in a response even if
2053 the request is nonce-less.
2056 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2057 skipped when using openssl x509 multiple times on a single input file,
2058 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2061 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2062 set string type: to handle setting ASN1_TIME structures. Fix ca
2063 utility to correctly initialize revocation date of CRLs.
2066 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2067 the clients preferred ciphersuites and rather use its own preferences.
2068 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2069 Internet Explorer by ensuring unchanged hash method during stepup.
2070 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2073 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2074 to aes and add a new 'exist' option to print out symbols that don't
2078 *) Additional options to ocsp utility to allow flags to be set and
2079 additional certificates supplied.
2082 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2083 OCSP client a number of certificate to only verify the response
2087 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2088 handle the new API. Currently only ECB, CBC modes supported. Add new
2091 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2092 Encryption Standard (AES) Ciphersuites for Transport Layer
2093 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2094 not enabled by default and were not part of the "ALL" ciphersuite
2095 alias because they were not yet official; they could be
2096 explicitly requested by specifying the "AESdraft" ciphersuite
2097 group alias. In the final release of OpenSSL 0.9.7, the group
2098 alias is called "AES" and is part of "ALL".)
2099 [Ben Laurie, Steve Henson, Bodo Moeller]
2101 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2102 request to response.
2105 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2106 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2107 extract information from a certificate request. OCSP_response_create()
2108 creates a response and optionally adds a basic response structure.
2109 OCSP_basic_add1_status() adds a complete single response to a basic
2110 response and returns the OCSP_SINGLERESP structure just added (to allow
2111 extensions to be included for example). OCSP_basic_add1_cert() adds a
2112 certificate to a basic response and OCSP_basic_sign() signs a basic
2113 response with various flags. New helper functions ASN1_TIME_check()
2114 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2115 (converts ASN1_TIME to GeneralizedTime).
2118 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2119 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2120 structure from a certificate. X509_pubkey_digest() digests the public_key
2121 contents: this is used in various key identifiers.
2124 *) Make sk_sort() tolerate a NULL argument.
2125 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2127 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2128 passed by the function are trusted implicitly. If any of them signed the
2129 response then it is assumed to be valid and is not verified.
2132 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2133 to data. This was previously part of the PKCS7 ASN1 code. This
2134 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2135 [Steve Henson, reported by Kenneth R. Robinette
2136 <support@securenetterm.com>]
2138 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2139 routines: without these tracing memory leaks is very painful.
2140 Fix leaks in PKCS12 and PKCS7 routines.
2143 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2144 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2145 effectively meant GeneralizedTime would never be used. Now it
2146 is initialised to -1 but X509_time_adj() now has to check the value
2147 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2148 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2149 [Steve Henson, reported by Kenneth R. Robinette
2150 <support@securenetterm.com>]
2152 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2153 result in a zero length in the ASN1_INTEGER structure which was
2154 not consistent with the structure when d2i_ASN1_INTEGER() was used
2155 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2156 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2157 where it did not print out a minus for negative ASN1_INTEGER.
2160 *) Add summary printout to ocsp utility. The various functions which
2161 convert status values to strings have been renamed to:
2162 OCSP_response_status_str(), OCSP_cert_status_str() and
2163 OCSP_crl_reason_str() and are no longer static. New options
2164 to verify nonce values and to disable verification. OCSP response
2165 printout format cleaned up.
2168 *) Add additional OCSP certificate checks. These are those specified
2169 in RFC2560. This consists of two separate checks: the CA of the
2170 certificate being checked must either be the OCSP signer certificate
2171 or the issuer of the OCSP signer certificate. In the latter case the
2172 OCSP signer certificate must contain the OCSP signing extended key
2173 usage. This check is performed by attempting to match the OCSP
2174 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2175 in the OCSP_CERTID structures of the response.
2178 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2179 and related routines. This uses the standard OpenSSL certificate
2180 verify routines to perform initial checks (just CA validity) and
2181 to obtain the certificate chain. Then additional checks will be
2182 performed on the chain. Currently the root CA is checked to see
2183 if it is explicitly trusted for OCSP signing. This is used to set
2184 a root CA as a global signing root: that is any certificate that
2185 chains to that CA is an acceptable OCSP signing certificate.
2188 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2189 extensions from a separate configuration file.
2190 As when reading extensions from the main configuration file,
2191 the '-extensions ...' option may be used for specifying the
2193 [Massimiliano Pala <madwolf@comune.modena.it>]
2195 *) New OCSP utility. Allows OCSP requests to be generated or
2196 read. The request can be sent to a responder and the output
2197 parsed, outputed or printed in text form. Not complete yet:
2198 still needs to check the OCSP response validity.
2201 *) New subcommands for 'openssl ca':
2202 'openssl ca -status <serial>' prints the status of the cert with
2203 the given serial number (according to the index file).
2204 'openssl ca -updatedb' updates the expiry status of certificates
2206 [Massimiliano Pala <madwolf@comune.modena.it>]
2208 *) New '-newreq-nodes' command option to CA.pl. This is like
2209 '-newreq', but calls 'openssl req' with the '-nodes' option
2210 so that the resulting key is not encrypted.
2211 [Damien Miller <djm@mindrot.org>]
2213 *) New configuration for the GNU Hurd.
2214 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2216 *) Initial code to implement OCSP basic response verify. This
2217 is currently incomplete. Currently just finds the signer's
2218 certificate and verifies the signature on the response.
2221 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2222 value of OPENSSLDIR. This is available via the new '-d' option
2223 to 'openssl version', and is also included in 'openssl version -a'.
2226 *) Allowing defining memory allocation callbacks that will be given
2227 file name and line number information in additional arguments
2228 (a const char* and an int). The basic functionality remains, as
2229 well as the original possibility to just replace malloc(),
2230 realloc() and free() by functions that do not know about these
2231 additional arguments. To register and find out the current
2232 settings for extended allocation functions, the following
2233 functions are provided:
2235 CRYPTO_set_mem_ex_functions
2236 CRYPTO_set_locked_mem_ex_functions
2237 CRYPTO_get_mem_ex_functions
2238 CRYPTO_get_locked_mem_ex_functions
2240 These work the same way as CRYPTO_set_mem_functions and friends.
2241 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2242 extended allocation function is enabled.
2243 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2244 a conventional allocation function is enabled.
2245 [Richard Levitte, Bodo Moeller]
2247 *) Finish off removing the remaining LHASH function pointer casts.
2248 There should no longer be any prototype-casting required when using
2249 the LHASH abstraction, and any casts that remain are "bugs". See
2250 the callback types and macros at the head of lhash.h for details
2251 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2254 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2255 If /dev/[u]random devices are not available or do not return enough
2256 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2258 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2259 /etc/entropy will be queried once each in this sequence, quering stops
2260 when enough entropy was collected without querying more sockets.
2263 *) Change the Unix RAND_poll() variant to be able to poll several
2264 random devices, as specified by DEVRANDOM, until a sufficient amount
2265 of data has been collected. We spend at most 10 ms on each file
2266 (select timeout) and read in non-blocking mode. DEVRANDOM now
2267 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2268 (previously it was just the string "/dev/urandom"), so on typical
2269 platforms the 10 ms delay will never occur.
2270 Also separate out the Unix variant to its own file, rand_unix.c.
2271 For VMS, there's a currently-empty rand_vms.c.
2274 *) Move OCSP client related routines to ocsp_cl.c. These
2275 provide utility functions which an application needing
2276 to issue a request to an OCSP responder and analyse the
2277 response will typically need: as opposed to those which an
2278 OCSP responder itself would need which will be added later.
2280 OCSP_request_sign() signs an OCSP request with an API similar
2281 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2282 response. OCSP_response_get1_basic() extracts basic response
2283 from response. OCSP_resp_find_status(): finds and extracts status
2284 information from an OCSP_CERTID structure (which will be created
2285 when the request structure is built). These are built from lower
2286 level functions which work on OCSP_SINGLERESP structures but
2287 wont normally be used unless the application wishes to examine
2288 extensions in the OCSP response for example.
2290 Replace nonce routines with a pair of functions.
2291 OCSP_request_add1_nonce() adds a nonce value and optionally
2292 generates a random value. OCSP_check_nonce() checks the
2293 validity of the nonce in an OCSP response.
2296 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2297 This doesn't copy the supplied OCSP_CERTID and avoids the
2298 need to free up the newly created id. Change return type
2299 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2300 This can then be used to add extensions to the request.
2301 Deleted OCSP_request_new(), since most of its functionality
2302 is now in OCSP_REQUEST_new() (and the case insensitive name
2303 clash) apart from the ability to set the request name which
2304 will be added elsewhere.
2307 *) Update OCSP API. Remove obsolete extensions argument from
2308 various functions. Extensions are now handled using the new
2309 OCSP extension code. New simple OCSP HTTP function which
2310 can be used to send requests and parse the response.
2313 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2314 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2315 uses the special reorder version of SET OF to sort the attributes
2316 and reorder them to match the encoded order. This resolves a long
2317 standing problem: a verify on a PKCS7 structure just after signing
2318 it used to fail because the attribute order did not match the
2319 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2320 it uses the received order. This is necessary to tolerate some broken
2321 software that does not order SET OF. This is handled by encoding
2322 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2323 to produce the required SET OF.
2326 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2327 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2328 files to get correct declarations of the ASN.1 item variables.
2331 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2332 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2333 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2334 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2335 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2336 ASN1_ITEM and no wrapper functions.
2339 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2340 replace the old function pointer based I/O routines. Change most of
2341 the *_d2i_bio() and *_d2i_fp() functions to use these.
2344 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2345 lines, recognice more "algorithms" that can be deselected, and make
2346 it complain about algorithm deselection that isn't recognised.
2349 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2350 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2351 to use new functions. Add NO_ASN1_OLD which can be set to remove
2352 some old style ASN1 functions: this can be used to determine if old
2353 code will still work when these eventually go away.
2356 *) New extension functions for OCSP structures, these follow the
2357 same conventions as certificates and CRLs.
2360 *) New function X509V3_add1_i2d(). This automatically encodes and
2361 adds an extension. Its behaviour can be customised with various
2362 flags to append, replace or delete. Various wrappers added for
2363 certifcates and CRLs.
2366 *) Fix to avoid calling the underlying ASN1 print routine when
2367 an extension cannot be parsed. Correct a typo in the
2368 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2371 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2372 entries for variables.
2375 *) Add functionality to apps/openssl.c for detecting locking
2376 problems: As the program is single-threaded, all we have
2377 to do is register a locking callback using an array for
2378 storing which locks are currently held by the program.
2381 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2382 SSL_get_ex_data_X509_STORE_idx(), which is used in
2383 ssl_verify_cert_chain() and thus can be called at any time
2384 during TLS/SSL handshakes so that thread-safety is essential.
2385 Unfortunately, the ex_data design is not at all suited
2386 for multi-threaded use, so it probably should be abolished.
2389 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2390 [Broadcom, tweaked and integrated by Geoff Thorpe]
2392 *) Move common extension printing code to new function
2393 X509V3_print_extensions(). Reorganise OCSP print routines and
2394 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2397 *) New function X509_signature_print() to remove duplication in some
2401 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2402 set (this was treated exactly the same as SET OF previously). This
2403 is used to reorder the STACK representing the structure to match the
2404 encoding. This will be used to get round a problem where a PKCS7
2405 structure which was signed could not be verified because the STACK
2406 order did not reflect the encoded order.
2409 *) Reimplement the OCSP ASN1 module using the new code.
2412 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2413 for its ASN1 operations. The old style function pointers still exist
2414 for now but they will eventually go away.
2417 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2418 completely replaces the old ASN1 functionality with a table driven
2419 encoder and decoder which interprets an ASN1_ITEM structure describing
2420 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2421 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2422 has also been converted to the new form.
2425 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2426 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2427 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2428 for negative moduli.
2431 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2432 of not touching the result's sign bit.
2435 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2439 *) Changed the LHASH code to use prototypes for callbacks, and created
2440 macros to declare and implement thin (optionally static) functions
2441 that provide type-safety and avoid function pointer casting for the
2442 type-specific callbacks.
2445 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2447 [Veers Staats <staatsvr@asc.hpc.mil>,
2448 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2450 *) Reformat the FAQ so the different questions and answers can be divided
2451 in sections depending on the subject.
2454 *) Have the zlib compression code load ZLIB.DLL dynamically under
2458 *) New function BN_mod_sqrt for computing square roots modulo a prime
2459 (using the probabilistic Tonelli-Shanks algorithm unless
2460 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2461 be handled deterministically).
2462 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2464 *) Make BN_mod_inverse faster by explicitly handling small quotients
2465 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2466 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2469 *) New function BN_kronecker.
2472 *) Fix BN_gcd so that it works on negative inputs; the result is
2473 positive unless both parameters are zero.
2474 Previously something reasonably close to an infinite loop was
2475 possible because numbers could be growing instead of shrinking
2476 in the implementation of Euclid's algorithm.
2479 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2480 sign of the number in question.
2482 Fix BN_is_word(a,w) to work correctly for w == 0.
2484 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2485 because its test if the absolute value of 'a' equals 'w'.
2486 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2487 it exists mostly for use in the implementations of BN_is_zero(),
2488 BN_is_one(), and BN_is_word().
2491 *) New function BN_swap.
2494 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2495 the exponentiation functions are more likely to produce reasonable
2496 results on negative inputs.
2499 *) Change BN_mod_mul so that the result is always non-negative.
2500 Previously, it could be negative if one of the factors was negative;
2501 I don't think anyone really wanted that behaviour.
2504 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2505 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2506 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2507 and add new functions:
2516 BN_mod_lshift1_quick
2520 These functions always generate non-negative results.
2522 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2523 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2525 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2526 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2527 be reduced modulo m.
2528 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2531 The following entry accidentily appeared in the CHANGES file
2532 distributed with OpenSSL 0.9.7. The modifications described in
2533 it do *not* apply to OpenSSL 0.9.7.
2535 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2536 was actually never needed) and in BN_mul(). The removal in BN_mul()
2537 required a small change in bn_mul_part_recursive() and the addition
2538 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2539 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2540 bn_sub_words() and bn_add_words() except they take arrays with
2545 *) In 'openssl passwd', verify passwords read from the terminal
2546 unless the '-salt' option is used (which usually means that
2547 verification would just waste user's time since the resulting
2548 hash is going to be compared with some given password hash)
2549 or the new '-noverify' option is used.
2551 This is an incompatible change, but it does not affect
2552 non-interactive use of 'openssl passwd' (passwords on the command
2553 line, '-stdin' option, '-in ...' option) and thus should not
2557 *) Remove all references to RSAref, since there's no more need for it.
2560 *) Make DSO load along a path given through an environment variable
2561 (SHLIB_PATH) with shl_load().
2564 *) Constify the ENGINE code as a result of BIGNUM constification.
2565 Also constify the RSA code and most things related to it. In a
2566 few places, most notable in the depth of the ASN.1 code, ugly
2567 casts back to non-const were required (to be solved at a later
2571 *) Make it so the openssl application has all engines loaded by default.
2574 *) Constify the BIGNUM routines a little more.
2577 *) Add the following functions:
2579 ENGINE_load_cswift()
2581 ENGINE_load_atalla()
2583 ENGINE_load_builtin_engines()
2585 That way, an application can itself choose if external engines that
2586 are built-in in OpenSSL shall ever be used or not. The benefit is
2587 that applications won't have to be linked with libdl or other dso
2588 libraries unless it's really needed.
2590 Changed 'openssl engine' to load all engines on demand.
2591 Changed the engine header files to avoid the duplication of some
2592 declarations (they differed!).
2595 *) 'openssl engine' can now list capabilities.
2598 *) Better error reporting in 'openssl engine'.
2601 *) Never call load_dh_param(NULL) in s_server.
2604 *) Add engine application. It can currently list engines by name and
2605 identity, and test if they are actually available.
2608 *) Improve RPM specification file by forcing symbolic linking and making
2609 sure the installed documentation is also owned by root.root.
2610 [Damien Miller <djm@mindrot.org>]
2612 *) Give the OpenSSL applications more possibilities to make use of
2613 keys (public as well as private) handled by engines.
2616 *) Add OCSP code that comes from CertCo.
2619 *) Add VMS support for the Rijndael code.
2622 *) Added untested support for Nuron crypto accelerator.
2625 *) Add support for external cryptographic devices. This code was
2626 previously distributed separately as the "engine" branch.
2627 [Geoff Thorpe, Richard Levitte]
2629 *) Rework the filename-translation in the DSO code. It is now possible to
2630 have far greater control over how a "name" is turned into a filename
2631 depending on the operating environment and any oddities about the
2632 different shared library filenames on each system.
2635 *) Support threads on FreeBSD-elf in Configure.
2638 *) Fix for SHA1 assembly problem with MASM: it produces
2639 warnings about corrupt line number information when assembling
2640 with debugging information. This is caused by the overlapping
2642 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2645 NCONF_get_number() has no error checking at all. As a replacement,
2646 NCONF_get_number_e() is defined (_e for "error checking") and is
2647 promoted strongly. The old NCONF_get_number is kept around for
2648 binary backward compatibility.
2649 Make it possible for methods to load from something other than a BIO,
2650 by providing a function pointer that is given a name instead of a BIO.
2651 For example, this could be used to load configuration data from an
2655 *) Fix for non blocking accept BIOs. Added new I/O special reason
2656 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2657 with non blocking I/O was not possible because no retry code was
2658 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2662 *) Added the beginnings of Rijndael support.
2665 *) Fix for bug in DirectoryString mask setting. Add support for
2666 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2667 to allow certificate printing to more controllable, additional
2668 'certopt' option to 'x509' to allow new printing options to be
2672 *) Clean old EAY MD5 hack from e_os.h.
2675 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2677 *) Fix various bugs revealed by running the NISCC test suite:
2679 Stop out of bounds reads in the ASN1 code when presented with
2680 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2682 If verify callback ignores invalid public key errors don't try to check
2683 certificate signature with the NULL public key.
2687 *) Fix various bugs revealed by running the NISCC test suite:
2689 Stop out of bounds reads in the ASN1 code when presented with
2690 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2692 If verify callback ignores invalid public key errors don't try to check
2693 certificate signature with the NULL public key.
2697 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2698 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2702 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2703 extra data after the compression methods not only for TLS 1.0
2704 but also for SSL 3.0 (as required by the specification).
2705 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2707 *) Change X509_certificate_type() to mark the key as exported/exportable
2708 when it's 512 *bits* long, not 512 bytes.
2711 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2713 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2714 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2715 a protocol version number mismatch like a decryption error
2716 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2719 *) Turn on RSA blinding by default in the default implementation
2720 to avoid a timing attack. Applications that don't want it can call
2721 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2722 They would be ill-advised to do so in most cases.
2723 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2725 *) Change RSA blinding code so that it works when the PRNG is not
2726 seeded (in this case, the secret RSA exponent is abused as
2727 an unpredictable seed -- if it is not unpredictable, there
2728 is no point in blinding anyway). Make RSA blinding thread-safe
2729 by remembering the creator's thread ID in rsa->blinding and
2730 having all other threads use local one-time blinding factors
2731 (this requires more computation than sharing rsa->blinding, but
2732 avoids excessive locking; and if an RSA object is not shared
2733 between threads, blinding will still be very fast).
2736 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2738 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2739 via timing by performing a MAC computation even if incorrrect
2740 block cipher padding has been found. This is a countermeasure
2741 against active attacks where the attacker has to distinguish
2742 between bad padding and a MAC verification error. (CAN-2003-0078)
2744 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2745 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2746 Martin Vuagnoux (EPFL, Ilion)]
2748 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2750 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2751 memory from it's contents. This is done with a counter that will
2752 place alternating values in each byte. This can be used to solve
2753 two issues: 1) the removal of calls to memset() by highly optimizing
2754 compilers, and 2) cleansing with other values than 0, since those can
2755 be read through on certain media, for example a swap space on disk.
2758 *) Bugfix: client side session caching did not work with external caching,
2759 because the session->cipher setting was not restored when reloading
2760 from the external cache. This problem was masked, when
2761 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2762 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2765 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2766 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2767 [Zeev Lieber <zeev-l@yahoo.com>]
2769 *) Undo an undocumented change introduced in 0.9.6e which caused
2770 repeated calls to OpenSSL_add_all_ciphers() and
2771 OpenSSL_add_all_digests() to be ignored, even after calling
2775 *) Change the default configuration reader to deal with last line not
2776 being properly terminated.
2779 *) Change X509_NAME_cmp() so it applies the special rules on handling
2780 DN values that are of type PrintableString, as well as RDNs of type
2781 emailAddress where the value has the type ia5String.
2782 [stefank@valicert.com via Richard Levitte]
2784 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2785 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2786 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2787 the bitwise-OR of the two for use by the majority of applications
2788 wanting this behaviour, and update the docs. The documented
2789 behaviour and actual behaviour were inconsistent and had been
2790 changing anyway, so this is more a bug-fix than a behavioural
2792 [Geoff Thorpe, diagnosed by Nadav Har'El]
2794 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2795 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2798 *) Fix initialization code race conditions in
2799 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2800 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2801 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2802 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2803 ssl2_get_cipher_by_char(),
2804 ssl3_get_cipher_by_char().
2805 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2807 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2808 the cached sessions are flushed, as the remove_cb() might use ex_data
2809 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2810 (see [openssl.org #212]).
2811 [Geoff Thorpe, Lutz Jaenicke]
2813 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2814 length, instead of the encoding length to d2i_ASN1_OBJECT.
2817 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2819 *) [In 0.9.6g-engine release:]
2820 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2821 [Lynn Gazis <lgazis@rainbow.com>]
2823 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2825 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2826 and get fix the header length calculation.
2827 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2828 Alon Kantor <alonk@checkpoint.com> (and others),
2831 *) Use proper error handling instead of 'assertions' in buffer
2832 overflow checks added in 0.9.6e. This prevents DoS (the
2833 assertions could call abort()).
2834 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2836 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2838 *) Add various sanity checks to asn1_get_length() to reject
2839 the ASN1 length bytes if they exceed sizeof(long), will appear
2840 negative or the content length exceeds the length of the
2842 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2844 *) Fix cipher selection routines: ciphers without encryption had no flags
2845 for the cipher strength set and where therefore not handled correctly
2846 by the selection routines (PR #130).
2849 *) Fix EVP_dsa_sha macro.
2853 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2854 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2855 that was added in OpenSSL 0.9.6d.
2857 As the countermeasure turned out to be incompatible with some
2858 broken SSL implementations, the new option is part of SSL_OP_ALL.
2859 SSL_OP_ALL is usually employed when compatibility with weird SSL
2860 implementations is desired (e.g. '-bugs' option to 's_client' and
2861 's_server'), so the new option is automatically set in many
2865 *) Changes in security patch:
2867 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2868 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2869 Air Force Materiel Command, USAF, under agreement number
2872 *) Add various sanity checks to asn1_get_length() to reject
2873 the ASN1 length bytes if they exceed sizeof(long), will appear
2874 negative or the content length exceeds the length of the
2875 supplied buffer. (CAN-2002-0659)
2876 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2878 *) Assertions for various potential buffer overflows, not known to
2880 [Ben Laurie (CHATS)]
2882 *) Various temporary buffers to hold ASCII versions of integers were
2883 too small for 64 bit platforms. (CAN-2002-0655)
2884 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
2886 *) Remote buffer overflow in SSL3 protocol - an attacker could
2887 supply an oversized session ID to a client. (CAN-2002-0656)
2888 [Ben Laurie (CHATS)]
2890 *) Remote buffer overflow in SSL2 protocol - an attacker could
2891 supply an oversized client master key. (CAN-2002-0656)
2892 [Ben Laurie (CHATS)]
2894 Changes between 0.9.6c and 0.9.6d [9 May 2002]
2896 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
2897 encoded as NULL) with id-dsa-with-sha1.
2898 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
2900 *) Check various X509_...() return values in apps/req.c.
2901 [Nils Larsch <nla@trustcenter.de>]
2903 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
2904 an end-of-file condition would erronously be flagged, when the CRLF
2905 was just at the end of a processed block. The bug was discovered when
2906 processing data through a buffering memory BIO handing the data to a
2907 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
2908 <ptsekov@syntrex.com> and Nedelcho Stanev.
2911 *) Implement a countermeasure against a vulnerability recently found
2912 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
2913 before application data chunks to avoid the use of known IVs
2914 with data potentially chosen by the attacker.
2917 *) Fix length checks in ssl3_get_client_hello().
2920 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
2921 to prevent ssl3_read_internal() from incorrectly assuming that
2922 ssl3_read_bytes() found application data while handshake
2923 processing was enabled when in fact s->s3->in_read_app_data was
2924 merely automatically cleared during the initial handshake.
2925 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
2927 *) Fix object definitions for Private and Enterprise: they were not
2928 recognized in their shortname (=lowercase) representation. Extend
2929 obj_dat.pl to issue an error when using undefined keywords instead
2930 of silently ignoring the problem (Svenning Sorensen
2931 <sss@sss.dnsalias.net>).
2934 *) Fix DH_generate_parameters() so that it works for 'non-standard'
2935 generators, i.e. generators other than 2 and 5. (Previously, the
2936 code did not properly initialise the 'add' and 'rem' values to
2937 BN_generate_prime().)
2939 In the new general case, we do not insist that 'generator' is
2940 actually a primitive root: This requirement is rather pointless;
2941 a generator of the order-q subgroup is just as good, if not
2945 *) Map new X509 verification errors to alerts. Discovered and submitted by
2946 Tom Wu <tom@arcot.com>.
2949 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
2950 returning non-zero before the data has been completely received
2951 when using non-blocking I/O.
2952 [Bodo Moeller; problem pointed out by John Hughes]
2954 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
2955 [Ben Laurie, Lutz Jaenicke]
2957 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
2958 Yoram Zahavi <YoramZ@gilian.com>).
2961 *) Add information about CygWin 1.3 and on, and preserve proper
2962 configuration for the versions before that.
2963 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2965 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
2966 check whether we deal with a copy of a session and do not delete from
2967 the cache in this case. Problem reported by "Izhar Shoshani Levi"
2968 <izhar@checkpoint.com>.
2971 *) Do not store session data into the internal session cache, if it
2972 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
2973 flag is set). Proposed by Aslam <aslam@funk.com>.
2976 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
2980 *) [In 0.9.6d-engine release:]
2981 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2982 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2984 *) Add the configuration target linux-s390x.
2985 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
2987 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
2988 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
2989 variable as an indication that a ClientHello message has been
2990 received. As the flag value will be lost between multiple
2991 invocations of ssl3_accept when using non-blocking I/O, the
2992 function may not be aware that a handshake has actually taken
2993 place, thus preventing a new session from being added to the
2996 To avoid this problem, we now set s->new_session to 2 instead of
2997 using a local variable.
2998 [Lutz Jaenicke, Bodo Moeller]
3000 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3001 if the SSL_R_LENGTH_MISMATCH error is detected.
3002 [Geoff Thorpe, Bodo Moeller]
3004 *) New 'shared_ldflag' column in Configure platform table.
3007 *) Fix EVP_CIPHER_mode macro.
3008 ["Dan S. Camper" <dan@bti.net>]
3010 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3011 type, we must throw them away by setting rr->length to 0.
3012 [D P Chang <dpc@qualys.com>]
3014 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3016 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3017 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3018 worked incorrectly for those cases where range = 10..._2 and
3019 3*range is two bits longer than range.)
3022 *) Only add signing time to PKCS7 structures if it is not already
3026 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3027 OBJ_ld_ce should be OBJ_id_ce.
3028 Also some ip-pda OIDs in crypto/objects/objects.txt were
3029 incorrect (cf. RFC 3039).
3030 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3032 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3033 returns early because it has nothing to do.
3034 [Andy Schneider <andy.schneider@bjss.co.uk>]
3036 *) [In 0.9.6c-engine release:]
3037 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3038 [Andy Schneider <andy.schneider@bjss.co.uk>]
3040 *) [In 0.9.6c-engine release:]
3041 Add support for Cryptographic Appliance's keyserver technology.
3042 (Use engine 'keyclient')
3043 [Cryptographic Appliances and Geoff Thorpe]
3045 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3046 is called via tools/c89.sh because arguments have to be
3047 rearranged (all '-L' options must appear before the first object
3049 [Richard Shapiro <rshapiro@abinitio.com>]
3051 *) [In 0.9.6c-engine release:]
3052 Add support for Broadcom crypto accelerator cards, backported
3054 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3056 *) [In 0.9.6c-engine release:]
3057 Add support for SureWare crypto accelerator cards from
3058 Baltimore Technologies. (Use engine 'sureware')
3059 [Baltimore Technologies and Mark Cox]
3061 *) [In 0.9.6c-engine release:]
3062 Add support for crypto accelerator cards from Accelerated
3063 Encryption Processing, www.aep.ie. (Use engine 'aep')
3064 [AEP Inc. and Mark Cox]
3066 *) Add a configuration entry for gcc on UnixWare.
3067 [Gary Benson <gbenson@redhat.com>]
3069 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3070 messages are stored in a single piece (fixed-length part and
3071 variable-length part combined) and fix various bugs found on the way.
3074 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3075 instead. BIO_gethostbyname() does not know what timeouts are
3076 appropriate, so entries would stay in cache even when they have
3078 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3080 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3081 faced with a pathologically small ClientHello fragment that does
3082 not contain client_version: Instead of aborting with an error,
3083 simply choose the highest available protocol version (i.e.,
3084 TLS 1.0 unless it is disabled). In practice, ClientHello
3085 messages are never sent like this, but this change gives us
3086 strictly correct behaviour at least for TLS.
3089 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3090 never resets s->method to s->ctx->method when called from within
3091 one of the SSL handshake functions.
3092 [Bodo Moeller; problem pointed out by Niko Baric]
3094 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3095 (sent using the client's version number) if client_version is
3096 smaller than the protocol version in use. Also change
3097 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3098 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3099 the client will at least see that alert.
3102 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3106 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3107 client receives HelloRequest while in a handshake.
3108 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3110 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3111 should end in 'break', not 'goto end' which circuments various
3112 cleanups done in state SSL_ST_OK. But session related stuff
3113 must be disabled for SSL_ST_OK in the case that we just sent a
3116 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3117 before just sending a HelloRequest.
3118 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3120 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3121 reveal whether illegal block cipher padding was found or a MAC
3122 verification error occured. (Neither SSLerr() codes nor alerts
3123 are directly visible to potential attackers, but the information
3124 may leak via logfiles.)
3126 Similar changes are not required for the SSL 2.0 implementation
3127 because the number of padding bytes is sent in clear for SSL 2.0,
3128 and the extra bytes are just ignored. However ssl/s2_pkt.c
3129 failed to verify that the purported number of padding bytes is in
3133 *) Add OpenUNIX-8 support including shared libraries
3134 (Boyd Lynn Gerber <gerberb@zenez.com>).
3137 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3138 'wristwatch attack' using huge encoding parameters (cf.
3139 James H. Manger's CRYPTO 2001 paper). Note that the
3140 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3141 encoding parameters and hence was not vulnerable.
3144 *) BN_sqr() bug fix.
3145 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3147 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3148 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3149 followed by modular reduction.
3150 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3152 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3153 equivalent based on BN_pseudo_rand() instead of BN_rand().
3156 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3157 This function was broken, as the check for a new client hello message
3158 to handle SGC did not allow these large messages.
3159 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3162 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3165 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3166 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3169 *) Rework the configuration and shared library support for Tru64 Unix.
3170 The configuration part makes use of modern compiler features and
3171 still retains old compiler behavior for those that run older versions
3172 of the OS. The shared library support part includes a variant that
3173 uses the RPATH feature, and is available through the special
3174 configuration target "alpha-cc-rpath", which will never be selected
3176 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3178 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3179 with the same message size as in ssl3_get_certificate_request().
3180 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3181 messages might inadvertently be reject as too long.
3182 [Petr Lampa <lampa@fee.vutbr.cz>]
3184 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3187 *) Modified SSL library such that the verify_callback that has been set
3188 specificly for an SSL object with SSL_set_verify() is actually being
3189 used. Before the change, a verify_callback set with this function was
3190 ignored and the verify_callback() set in the SSL_CTX at the time of
3191 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3192 to allow the necessary settings.
3195 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3196 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3197 done automatically (in contradiction to the requirements of the C
3198 standard). This made problems when used from OpenSSH.
3201 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3202 dh->length and always used
3204 BN_rand_range(priv_key, dh->p).
3206 BN_rand_range() is not necessary for Diffie-Hellman, and this
3207 specific range makes Diffie-Hellman unnecessarily inefficient if
3208 dh->length (recommended exponent length) is much smaller than the
3209 length of dh->p. We could use BN_rand_range() if the order of
3210 the subgroup was stored in the DH structure, but we only have
3215 BN_rand(priv_key, l, ...)
3217 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3223 RSA_eay_public_encrypt
3224 RSA_eay_private_decrypt
3225 RSA_eay_private_encrypt (signing)
3226 RSA_eay_public_decrypt (signature verification)
3228 (default implementations for RSA_public_encrypt,
3229 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3230 always reject numbers >= n.
3233 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3234 to synchronize access to 'locking_thread'. This is necessary on
3235 systems where access to 'locking_thread' (an 'unsigned long'
3236 variable) is not atomic.
3239 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3240 *before* setting the 'crypto_lock_rand' flag. The previous code had
3241 a race condition if 0 is a valid thread ID.
3242 [Travis Vitek <vitek@roguewave.com>]
3244 *) Add support for shared libraries under Irix.
3245 [Albert Chin-A-Young <china@thewrittenword.com>]
3247 *) Add configuration option to build on Linux on both big-endian and
3249 [Ralf Baechle <ralf@uni-koblenz.de>]
3251 *) Add the possibility to create shared libraries on HP-UX.
3254 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3256 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3257 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3258 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3259 PRNG state recovery was possible based on the output of
3260 one PRNG request appropriately sized to gain knowledge on
3261 'md' followed by enough consecutive 1-byte PRNG requests
3262 to traverse all of 'state'.
3264 1. When updating 'md_local' (the current thread's copy of 'md')
3265 during PRNG output generation, hash all of the previous
3266 'md_local' value, not just the half used for PRNG output.
3268 2. Make the number of bytes from 'state' included into the hash
3269 independent from the number of PRNG bytes requested.
3271 The first measure alone would be sufficient to avoid
3272 Markku-Juhani's attack. (Actually it had never occurred
3273 to me that the half of 'md_local' used for chaining was the
3274 half from which PRNG output bytes were taken -- I had always
3275 assumed that the secret half would be used.) The second
3276 measure makes sure that additional data from 'state' is never
3277 mixed into 'md_local' in small portions; this heuristically
3278 further strengthens the PRNG.
3281 *) Fix crypto/bn/asm/mips3.s.
3284 *) When only the key is given to "enc", the IV is undefined. Print out
3285 an error message in this case.
3288 *) Handle special case when X509_NAME is empty in X509 printing routines.
3291 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3292 positive and less than q.
3295 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3296 used: it isn't thread safe and the add_lock_callback should handle
3298 [Paul Rose <Paul.Rose@bridge.com>]
3300 *) Verify that incoming data obeys the block size in
3301 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3305 [Ulf Möller, Bodo Möller]
3307 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3308 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3309 when fixing the server behaviour for backwards-compatible 'client
3310 hello' messages. (Note that the attack is impractical against
3311 SSL 3.0 and TLS 1.0 anyway because length and version checking
3312 means that the probability of guessing a valid ciphertext is
3313 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3316 Before 0.9.5, the countermeasure (hide the error by generating a
3317 random 'decryption result') did not work properly because
3318 ERR_clear_error() was missing, meaning that SSL_get_error() would
3319 detect the supposedly ignored error.
3321 Both problems are now fixed.
3324 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3325 (previously it was 1024).
3328 *) Fix for compatibility mode trust settings: ignore trust settings
3329 unless some valid trust or reject settings are present.
3332 *) Fix for blowfish EVP: its a variable length cipher.
3335 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3336 parameters in DSA public key structures and return an error in the
3337 DSA routines if parameters are absent.
3340 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3341 in the current directory if neither $RANDFILE nor $HOME was set.
3342 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3343 caused some confusion to Windows users who haven't defined $HOME.
3344 Thus RAND_file_name() is changed again: e_os.h can define a
3345 DEFAULT_HOME, which will be used if $HOME is not set.
3346 For Windows, we use "C:"; on other platforms, we still require
3347 environment variables.
3349 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3350 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3351 having multiple threads call RAND_poll() concurrently.
3354 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3355 combination of a flag and a thread ID variable.
3356 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3357 flag), *other* threads can enter ssleay_add_bytes without obeying
3358 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3359 that they do not hold after the first thread unsets add_do_not_lock).
3362 *) Change bctest again: '-x' expressions are not available in all
3366 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3368 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3369 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3371 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3372 the default extension for executables, if any. Also, make the perl
3373 scripts that use symlink() to test if it really exists and use "cp"
3374 if it doesn't. All this made OpenSSL compilable and installable in
3378 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3379 If SEQUENCE is length is indefinite just set c->slen to the total
3380 amount of data available.
3381 [Steve Henson, reported by shige@FreeBSD.org]
3382 [This change does not apply to 0.9.7.]
3384 *) Change bctest to avoid here-documents inside command substitution
3385 (workaround for FreeBSD /bin/sh bug).
3386 For compatibility with Ultrix, avoid shell functions (introduced
3387 in the bctest version that searches along $PATH).
3390 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3391 with des_encrypt() defined on some operating systems, like Solaris
3395 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3396 On the Importance of Eliminating Errors in Cryptographic
3397 Computations, J. Cryptology 14 (2001) 2, 101-119,
3398 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3401 *) MIPS assembler BIGNUM division bug fix.
3404 *) Disabled incorrect Alpha assembler code.
3407 *) Fix PKCS#7 decode routines so they correctly update the length
3408 after reading an EOC for the EXPLICIT tag.
3410 [This change does not apply to 0.9.7.]
3412 *) Fix bug in PKCS#12 key generation routines. This was triggered
3413 if a 3DES key was generated with a 0 initial byte. Include
3414 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3415 (but broken) behaviour.
3418 *) Enhance bctest to search for a working bc along $PATH and print
3420 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3422 *) Fix memory leaks in err.c: free err_data string if necessary;
3423 don't write to the wrong index in ERR_set_error_data.
3426 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3430 *) Replace rdtsc with _emit statements for VC++ version 5.
3431 [Jeremy Cooper <jeremy@baymoo.org>]
3433 *) Make it possible to reuse SSLv2 sessions.
3436 *) In copy_email() check for >= 0 as a return value for
3437 X509_NAME_get_index_by_NID() since 0 is a valid index.
3438 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3440 *) Avoid coredump with unsupported or invalid public keys by checking if
3441 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3442 PKCS7_verify() fails with non detached data.
3445 *) Don't use getenv in library functions when run as setuid/setgid.
3446 New function OPENSSL_issetugid().
3449 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3450 due to incorrect handling of multi-threading:
3452 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3454 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3456 3. Count how many times MemCheck_off() has been called so that
3457 nested use can be treated correctly. This also avoids
3458 inband-signalling in the previous code (which relied on the
3459 assumption that thread ID 0 is impossible).
3462 *) Add "-rand" option also to s_client and s_server.
3465 *) Fix CPU detection on Irix 6.x.
3466 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3467 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3469 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3472 [This change does not apply to 0.9.7.]
3474 *) Use the cached encoding of an X509_NAME structure rather than
3475 copying it. This is apparently the reason for the libsafe "errors"
3476 but the code is actually correct.
3479 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3480 Bleichenbacher's DSA attack.
3481 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3482 to be set and top=0 forces the highest bit to be set; top=-1 is new
3483 and leaves the highest bit random.
3484 [Ulf Moeller, Bodo Moeller]
3486 *) In the NCONF_...-based implementations for CONF_... queries
3487 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3488 a temporary CONF structure with the data component set to NULL
3489 (which gives segmentation faults in lh_retrieve).
3490 Instead, use NULL for the CONF pointer in CONF_get_string and
3491 CONF_get_number (which may use environment variables) and directly
3492 return NULL from CONF_get_section.
3495 *) Fix potential buffer overrun for EBCDIC.
3498 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3499 keyUsage if basicConstraints absent for a CA.
3502 *) Make SMIME_write_PKCS7() write mail header values with a format that
3503 is more generally accepted (no spaces before the semicolon), since
3504 some programs can't parse those values properly otherwise. Also make
3505 sure BIO's that break lines after each write do not create invalid
3509 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3510 macros previously used would not encode an empty SEQUENCE OF
3511 and break the signature.
3513 [This change does not apply to 0.9.7.]
3515 *) Zero the premaster secret after deriving the master secret in
3519 *) Add some EVP_add_digest_alias registrations (as found in
3520 OpenSSL_add_all_digests()) to SSL_library_init()
3521 aka OpenSSL_add_ssl_algorithms(). This provides improved
3522 compatibility with peers using X.509 certificates
3523 with unconventional AlgorithmIdentifier OIDs.
3526 *) Fix for Irix with NO_ASM.
3527 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3529 *) ./config script fixes.
3530 [Ulf Moeller, Richard Levitte]
3532 *) Fix 'openssl passwd -1'.
3535 *) Change PKCS12_key_gen_asc() so it can cope with non null
3536 terminated strings whose length is passed in the passlen
3537 parameter, for example from PEM callbacks. This was done
3538 by adding an extra length parameter to asc2uni().
3539 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3541 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3542 call failed, free the DSA structure.
3545 *) Fix to uni2asc() to cope with zero length Unicode strings.
3546 These are present in some PKCS#12 files.
3549 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3550 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3551 when writing a 32767 byte record.
3552 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3554 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3555 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3557 (RSA objects have a reference count access to which is protected
3558 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3559 so they are meant to be shared between threads.)
3560 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3561 "Reddie, Steven" <Steven.Reddie@ca.com>]
3563 *) Fix a deadlock in CRYPTO_mem_leaks().
3566 *) Use better test patterns in bntest.
3569 *) rand_win.c fix for Borland C.
3572 *) BN_rshift bugfix for n == 0.
3575 *) Add a 'bctest' script that checks for some known 'bc' bugs
3576 so that 'make test' does not abort just because 'bc' is broken.
3579 *) Store verify_result within SSL_SESSION also for client side to
3580 avoid potential security hole. (Re-used sessions on the client side
3581 always resulted in verify_result==X509_V_OK, not using the original
3582 result of the server certificate verification.)
3585 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3586 SSL3_RT_APPLICATION_DATA, return 0.
3587 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3591 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3592 releases, have been re-implemented by renaming the previous
3593 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3594 and ssl3_read_internal, respectively, and adding 'peek' parameters
3595 to them. The new ssl[23]_{read,peek} functions are calls to
3596 ssl[23]_read_internal with the 'peek' flag set appropriately.
3597 A 'peek' parameter has also been added to ssl3_read_bytes, which
3598 does the actual work for ssl3_read_internal.
3601 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3602 the method-specific "init()" handler. Also clean up ex_data after
3603 calling the method-specific "finish()" handler. Previously, this was
3604 happening the other way round.
3607 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3608 The previous value, 12, was not always sufficient for BN_mod_exp().
3611 *) Make sure that shared libraries get the internal name engine with
3612 the full version number and not just 0. This should mark the
3613 shared libraries as not backward compatible. Of course, this should
3614 be changed again when we can guarantee backward binary compatibility.
3617 *) Fix typo in get_cert_by_subject() in by_dir.c
3618 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3620 *) Rework the system to generate shared libraries:
3622 - Make note of the expected extension for the shared libraries and
3623 if there is a need for symbolic links from for example libcrypto.so.0
3624 to libcrypto.so.0.9.7. There is extended info in Configure for
3627 - Make as few rebuilds of the shared libraries as possible.
3629 - Still avoid linking the OpenSSL programs with the shared libraries.
3631 - When installing, install the shared libraries separately from the
3635 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3637 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3638 and not in SSL_clear because the latter is also used by the
3639 accept/connect functions; previously, the settings made by
3640 SSL_set_read_ahead would be lost during the handshake.
3641 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3643 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3644 Previously, it would create entries for disableed algorithms no
3648 *) Added several new manual pages for SSL_* function.
3651 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3653 *) In ssl23_get_client_hello, generate an error message when faced
3654 with an initial SSL 3.0/TLS record that is too small to contain the
3655 first two bytes of the ClientHello message, i.e. client_version.
3656 (Note that this is a pathologic case that probably has never happened
3657 in real life.) The previous approach was to use the version number
3658 from the record header as a substitute; but our protocol choice
3659 should not depend on that one because it is not authenticated
3660 by the Finished messages.
3663 *) More robust randomness gathering functions for Windows.
3664 [Jeffrey Altman <jaltman@columbia.edu>]
3666 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3667 not set then we don't setup the error code for issuer check errors
3668 to avoid possibly overwriting other errors which the callback does
3669 handle. If an application does set the flag then we assume it knows
3670 what it is doing and can handle the new informational codes
3674 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3675 a general "ANY" type, as such it should be able to decode anything
3676 including tagged types. However it didn't check the class so it would
3677 wrongly interpret tagged types in the same way as their universal
3678 counterpart and unknown types were just rejected. Changed so that the
3679 tagged and unknown types are handled in the same way as a SEQUENCE:
3680 that is the encoding is stored intact. There is also a new type
3681 "V_ASN1_OTHER" which is used when the class is not universal, in this
3682 case we have no idea what the actual type is so we just lump them all
3686 *) On VMS, stdout may very well lead to a file that is written to
3687 in a record-oriented fashion. That means that every write() will
3688 write a separate record, which will be read separately by the
3689 programs trying to read from it. This can be very confusing.
3691 The solution is to put a BIO filter in the way that will buffer
3692 text until a linefeed is reached, and then write everything a
3693 line at a time, so every record written will be an actual line,
3694 not chunks of lines and not (usually doesn't happen, but I've
3695 seen it once) several lines in one record. BIO_f_linebuffer() is
3698 Currently, it's a VMS-only method, because that's where it has
3699 been tested well enough.
3702 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3703 it can return incorrect results.
3704 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3705 but it was in 0.9.6-beta[12].)
3708 *) Disable the check for content being present when verifying detached
3709 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3710 include zero length content when signing messages.
3713 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3714 BIO_ctrl (for BIO pairs).
3717 *) Add DSO method for VMS.
3720 *) Bug fix: Montgomery multiplication could produce results with the
3724 *) Add RPM specification openssl.spec and modify it to build three
3725 packages. The default package contains applications, application
3726 documentation and run-time libraries. The devel package contains
3727 include files, static libraries and function documentation. The
3728 doc package contains the contents of the doc directory. The original
3729 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3732 *) Add a large number of documentation files for many SSL routines.
3733 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3735 *) Add a configuration entry for Sony News 4.
3736 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3738 *) Don't set the two most significant bits to one when generating a
3739 random number < q in the DSA library.
3742 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3743 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3744 the underlying transport is blocking) if a handshake took place.
3745 (The default behaviour is needed by applications such as s_client
3746 and s_server that use select() to determine when to use SSL_read;
3747 but for applications that know in advance when to expect data, it
3748 just makes things more complicated.)
3751 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3755 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3756 work better on such systems.
3757 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3759 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3760 Update PKCS12_parse() so it copies the friendlyName and the
3761 keyid to the certificates aux info.
3764 *) Fix bug in PKCS7_verify() which caused an infinite loop
3765 if there was more than one signature.
3766 [Sven Uszpelkat <su@celocom.de>]
3768 *) Major change in util/mkdef.pl to include extra information
3769 about each symbol, as well as presentig variables as well
3770 as functions. This change means that there's n more need
3771 to rebuild the .num files when some algorithms are excluded.
3774 *) Allow the verify time to be set by an application,
3775 rather than always using the current time.
3778 *) Phase 2 verify code reorganisation. The certificate
3779 verify code now looks up an issuer certificate by a
3780 number of criteria: subject name, authority key id
3781 and key usage. It also verifies self signed certificates
3782 by the same criteria. The main comparison function is
3783 X509_check_issued() which performs these checks.
3785 Lot of changes were necessary in order to support this
3786 without completely rewriting the lookup code.
3788 Authority and subject key identifier are now cached.
3790 The LHASH 'certs' is X509_STORE has now been replaced
3791 by a STACK_OF(X509_OBJECT). This is mainly because an
3792 LHASH can't store or retrieve multiple objects with
3793 the same hash value.
3795 As a result various functions (which were all internal
3796 use only) have changed to handle the new X509_STORE
3797 structure. This will break anything that messed round
3798 with X509_STORE internally.
3800 The functions X509_STORE_add_cert() now checks for an
3801 exact match, rather than just subject name.
3803 The X509_STORE API doesn't directly support the retrieval
3804 of multiple certificates matching a given criteria, however
3805 this can be worked round by performing a lookup first
3806 (which will fill the cache with candidate certificates)
3807 and then examining the cache for matches. This is probably
3808 the best we can do without throwing out X509_LOOKUP
3809 entirely (maybe later...).
3811 The X509_VERIFY_CTX structure has been enhanced considerably.
3813 All certificate lookup operations now go via a get_issuer()
3814 callback. Although this currently uses an X509_STORE it
3815 can be replaced by custom lookups. This is a simple way
3816 to bypass the X509_STORE hackery necessary to make this
3817 work and makes it possible to use more efficient techniques
3818 in future. A very simple version which uses a simple
3819 STACK for its trusted certificate store is also provided
3820 using X509_STORE_CTX_trusted_stack().
3822 The verify_cb() and verify() callbacks now have equivalents
3823 in the X509_STORE_CTX structure.
3825 X509_STORE_CTX also has a 'flags' field which can be used
3826 to customise the verify behaviour.
3829 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3830 excludes S/MIME capabilities.
3833 *) When a certificate request is read in keep a copy of the
3834 original encoding of the signed data and use it when outputing
3835 again. Signatures then use the original encoding rather than
3836 a decoded, encoded version which may cause problems if the
3837 request is improperly encoded.
3840 *) For consistency with other BIO_puts implementations, call
3841 buffer_write(b, ...) directly in buffer_puts instead of calling
3844 In BIO_puts, increment b->num_write as in BIO_write.
3845 [Peter.Sylvester@EdelWeb.fr]
3847 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3848 BN_zero, we may not return a BIGNUM with an array consisting of
3852 *) Avoid calling abort() from within the library when problems are
3853 detected, except if preprocessor symbols have been defined
3854 (such as REF_CHECK, BN_DEBUG etc.).
3857 *) New openssl application 'rsautl'. This utility can be
3858 used for low level RSA operations. DER public key
3859 BIO/fp routines also added.
3862 *) New Configure entry and patches for compiling on QNX 4.
3863 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3865 *) A demo state-machine implementation was sponsored by
3866 Nuron (http://www.nuron.com/) and is now available in
3867 demos/state_machine.
3870 *) New options added to the 'dgst' utility for signature
3871 generation and verification.
3874 *) Unrecognized PKCS#7 content types are now handled via a
3875 catch all ASN1_TYPE structure. This allows unsupported
3876 types to be stored as a "blob" and an application can
3877 encode and decode it manually.
3880 *) Fix various signed/unsigned issues to make a_strex.c
3882 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
3884 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
3885 length if passed a buffer. ASN1_INTEGER_to_BN failed
3886 if passed a NULL BN and its argument was negative.
3887 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
3889 *) Modification to PKCS#7 encoding routines to output definite
3890 length encoding. Since currently the whole structures are in
3891 memory there's not real point in using indefinite length
3892 constructed encoding. However if OpenSSL is compiled with
3893 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
3896 *) Added BIO_vprintf() and BIO_vsnprintf().
3899 *) Added more prefixes to parse for in the the strings written
3900 through a logging bio, to cover all the levels that are available
3901 through syslog. The prefixes are now:
3903 PANIC, EMERG, EMR => LOG_EMERG
3904 ALERT, ALR => LOG_ALERT
3905 CRIT, CRI => LOG_CRIT
3906 ERROR, ERR => LOG_ERR
3907 WARNING, WARN, WAR => LOG_WARNING
3908 NOTICE, NOTE, NOT => LOG_NOTICE
3909 INFO, INF => LOG_INFO
3910 DEBUG, DBG => LOG_DEBUG
3912 and as before, if none of those prefixes are present at the
3913 beginning of the string, LOG_ERR is chosen.
3915 On Win32, the LOG_* levels are mapped according to this:
3917 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
3918 LOG_WARNING => EVENTLOG_WARNING_TYPE
3919 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
3923 *) Made it possible to reconfigure with just the configuration
3924 argument "reconf" or "reconfigure". The command line arguments
3925 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
3926 and are retrieved from there when reconfiguring.
3930 [Assar Westerlund <assar@sics.se>, Richard Levitte]
3932 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
3935 *) The obj_dat.pl script was messing up the sorting of object
3936 names. The reason was that it compared the quoted version
3937 of strings as a result "OCSP" > "OCSP Signing" because
3938 " > SPACE. Changed script to store unquoted versions of
3939 names and add quotes on output. It was also omitting some
3940 names from the lookup table if they were given a default
3941 value (that is if SN is missing it is given the same
3942 value as LN and vice versa), these are now added on the
3943 grounds that if an object has a name we should be able to
3944 look it up. Finally added warning output when duplicate
3945 short or long names are found.
3948 *) Changes needed for Tandem NSK.
3949 [Scott Uroff <scott@xypro.com>]
3951 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
3952 RSA_padding_check_SSLv23(), special padding was never detected
3953 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
3954 version rollback attacks was not effective.
3956 In s23_clnt.c, don't use special rollback-attack detection padding
3957 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
3958 client; similarly, in s23_srvr.c, don't do the rollback check if
3959 SSL 2.0 is the only protocol enabled in the server.
3962 *) Make it possible to get hexdumps of unprintable data with 'openssl
3963 asn1parse'. By implication, the functions ASN1_parse_dump() and
3964 BIO_dump_indent() are added.
3967 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
3968 these print out strings and name structures based on various
3969 flags including RFC2253 support and proper handling of
3970 multibyte characters. Added options to the 'x509' utility
3971 to allow the various flags to be set.
3974 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
3975 Also change the functions X509_cmp_current_time() and
3976 X509_gmtime_adj() work with an ASN1_TIME structure,
3977 this will enable certificates using GeneralizedTime in validity
3978 dates to be checked.
3981 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
3982 negative public key encodings) on by default,
3983 NO_NEG_PUBKEY_BUG can be set to disable it.
3986 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
3987 content octets. An i2c_ASN1_OBJECT is unnecessary because
3988 the encoding can be trivially obtained from the structure.
3991 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
3992 not read locks (CRYPTO_r_[un]lock).
3995 *) A first attempt at creating official support for shared
3996 libraries through configuration. I've kept it so the
3997 default is static libraries only, and the OpenSSL programs
3998 are always statically linked for now, but there are
3999 preparations for dynamic linking in place.
4000 This has been tested on Linux and Tru64.
4003 *) Randomness polling function for Win9x, as described in:
4004 Peter Gutmann, Software Generation of Practically Strong
4008 *) Fix so PRNG is seeded in req if using an already existing
4012 *) New options to smime application. -inform and -outform
4013 allow alternative formats for the S/MIME message including
4014 PEM and DER. The -content option allows the content to be
4015 specified separately. This should allow things like Netscape
4016 form signing output easier to verify.
4019 *) Fix the ASN1 encoding of tags using the 'long form'.
4022 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4023 STRING types. These convert content octets to and from the
4024 underlying type. The actual tag and length octets are
4025 already assumed to have been read in and checked. These
4026 are needed because all other string types have virtually
4027 identical handling apart from the tag. By having versions
4028 of the ASN1 functions that just operate on content octets
4029 IMPLICIT tagging can be handled properly. It also allows
4030 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4031 and ASN1_INTEGER are identical apart from the tag.
4034 *) Change the handling of OID objects as follows:
4036 - New object identifiers are inserted in objects.txt, following
4037 the syntax given in objects.README.
4038 - objects.pl is used to process obj_mac.num and create a new
4040 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4043 This is currently kind of a hack, and the perl code in objects.pl
4044 isn't very elegant, but it works as I intended. The simplest way
4045 to check that it worked correctly is to look in obj_dat.h and
4046 check the array nid_objs and make sure the objects haven't moved
4047 around (this is important!). Additions are OK, as well as
4048 consistent name changes.
4051 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4054 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4055 The given file adds to whatever has already been seeded into the
4056 random pool through the RANDFILE configuration file option or
4057 environment variable, or the default random state file.
4060 *) mkstack.pl now sorts each macro group into lexical order.
4061 Previously the output order depended on the order the files
4062 appeared in the directory, resulting in needless rewriting
4066 *) Patches to make OpenSSL compile under Win32 again. Mostly
4067 work arounds for the VC++ problem that it treats func() as
4068 func(void). Also stripped out the parts of mkdef.pl that
4069 added extra typesafe functions: these no longer exist.
4072 *) Reorganisation of the stack code. The macros are now all
4073 collected in safestack.h . Each macro is defined in terms of
4074 a "stack macro" of the form SKM_<name>(type, a, b). The
4075 DEBUG_SAFESTACK is now handled in terms of function casts,
4076 this has the advantage of retaining type safety without the
4077 use of additional functions. If DEBUG_SAFESTACK is not defined
4078 then the non typesafe macros are used instead. Also modified the
4079 mkstack.pl script to handle the new form. Needs testing to see
4080 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4081 the default if no major problems. Similar behaviour for ASN1_SET_OF
4082 and PKCS12_STACK_OF.
4085 *) When some versions of IIS use the 'NET' form of private key the
4086 key derivation algorithm is different. Normally MD5(password) is
4087 used as a 128 bit RC4 key. In the modified case
4088 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4089 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4090 as the old Netscape_RSA functions except they have an additional
4091 'sgckey' parameter which uses the modified algorithm. Also added
4092 an -sgckey command line option to the rsa utility. Thanks to
4093 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4094 algorithm to openssl-dev.
4097 *) The evp_local.h macros were using 'c.##kname' which resulted in
4098 invalid expansion on some systems (SCO 5.0.5 for example).
4099 Corrected to 'c.kname'.
4100 [Phillip Porch <root@theporch.com>]
4102 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4103 a STACK of email addresses from a certificate or request, these look
4104 in the subject name and the subject alternative name extensions and
4105 omit any duplicate addresses.
4108 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4109 This makes DSA verification about 2 % faster.
4112 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4113 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4114 plus overhead for 1024 bit moduli).
4115 This makes exponentiations about 0.5 % faster for 1024 bit
4116 exponents (as measured by "openssl speed rsa2048").
4119 *) Rename memory handling macros to avoid conflicts with other
4121 Malloc => OPENSSL_malloc
4122 Malloc_locked => OPENSSL_malloc_locked
4123 Realloc => OPENSSL_realloc
4124 Free => OPENSSL_free
4127 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4128 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4131 *) CygWin32 support.
4132 [John Jarvie <jjarvie@newsguy.com>]
4134 *) The type-safe stack code has been rejigged. It is now only compiled
4135 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4136 by default all type-specific stack functions are "#define"d back to
4137 standard stack functions. This results in more streamlined output
4138 but retains the type-safety checking possibilities of the original
4142 *) The STACK code has been cleaned up, and certain type declarations
4143 that didn't make a lot of sense have been brought in line. This has
4144 also involved a cleanup of sorts in safestack.h to more correctly
4145 map type-safe stack functions onto their plain stack counterparts.
4146 This work has also resulted in a variety of "const"ifications of
4147 lots of the code, especially "_cmp" operations which should normally
4148 be prototyped with "const" parameters anyway.
4151 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4152 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4153 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4154 where all of 'md' is used each time the PRNG is used, but 'state'
4155 is used only indexed by a cyclic counter. As entropy may not be
4156 well distributed from the beginning, 'md' is important as a
4157 chaining variable. However, the output function chains only half
4158 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4159 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4160 in all of 'state' being rewritten, with the new values depending
4161 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4164 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4165 the handshake is continued after ssl_verify_cert_chain();
4166 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4167 can lead to 'unexplainable' connection aborts later.
4168 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4170 *) Major EVP API cipher revision.
4171 Add hooks for extra EVP features. This allows various cipher
4172 parameters to be set in the EVP interface. Support added for variable
4173 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4174 setting of RC2 and RC5 parameters.
4176 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4179 Remove lots of duplicated code from the EVP library. For example *every*
4180 cipher init() function handles the 'iv' in the same way according to the
4181 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4182 for CFB and OFB modes they zero ctx->num.
4184 New functionality allows removal of S/MIME code RC2 hack.
4186 Most of the routines have the same form and so can be declared in terms
4189 By shifting this to the top level EVP_CipherInit() it can be removed from
4190 all individual ciphers. If the cipher wants to handle IVs or keys
4191 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4194 Change lots of functions like EVP_EncryptUpdate() to now return a
4195 value: although software versions of the algorithms cannot fail
4196 any installed hardware versions can.
4199 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4200 this option is set, tolerate broken clients that send the negotiated
4201 protocol version number instead of the requested protocol version
4205 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4206 i.e. non-zero for export ciphersuites, zero otherwise.
4207 Previous versions had this flag inverted, inconsistent with
4208 rsa_tmp_cb (..._TMP_RSA_CB).
4209 [Bodo Moeller; problem reported by Amit Chopra]
4211 *) Add missing DSA library text string. Work around for some IIS
4212 key files with invalid SEQUENCE encoding.
4215 *) Add a document (doc/standards.txt) that list all kinds of standards
4216 and so on that are implemented in OpenSSL.
4219 *) Enhance c_rehash script. Old version would mishandle certificates
4220 with the same subject name hash and wouldn't handle CRLs at all.
4221 Added -fingerprint option to crl utility, to support new c_rehash
4225 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4228 *) Fix for SSL server purpose checking. Server checking was
4229 rejecting certificates which had extended key usage present
4230 but no ssl client purpose.
4231 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4233 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4234 is a little unclear about how a blank password is handled.
4235 Since the password in encoded as a BMPString with terminating
4236 double NULL a zero length password would end up as just the
4237 double NULL. However no password at all is different and is
4238 handled differently in the PKCS#12 key generation code. NS
4239 treats a blank password as zero length. MSIE treats it as no
4240 password on export: but it will try both on import. We now do
4241 the same: PKCS12_parse() tries zero length and no password if
4242 the password is set to "" or NULL (NULL is now a valid password:
4243 it wasn't before) as does the pkcs12 application.
4246 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4247 perror when PEM_read_bio_X509_REQ fails, the error message must
4248 be obtained from the error queue.
4251 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4252 it in ERR_remove_state if appropriate, and change ERR_get_state
4253 accordingly to avoid race conditions (this is necessary because
4254 thread_hash is no longer constant once set).
4257 *) Bugfix for linux-elf makefile.one.
4260 *) RSA_get_default_method() will now cause a default
4261 RSA_METHOD to be chosen if one doesn't exist already.
4262 Previously this was only set during a call to RSA_new()
4263 or RSA_new_method(NULL) meaning it was possible for
4264 RSA_get_default_method() to return NULL.
4267 *) Added native name translation to the existing DSO code
4268 that will convert (if the flag to do so is set) filenames
4269 that are sufficiently small and have no path information
4270 into a canonical native form. Eg. "blah" converted to
4271 "libblah.so" or "blah.dll" etc.
4274 *) New function ERR_error_string_n(e, buf, len) which is like
4275 ERR_error_string(e, buf), but writes at most 'len' bytes
4276 including the 0 terminator. For ERR_error_string_n, 'buf'
4278 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4280 *) CONF library reworked to become more general. A new CONF
4281 configuration file reader "class" is implemented as well as a
4282 new functions (NCONF_*, for "New CONF") to handle it. The now
4283 old CONF_* functions are still there, but are reimplemented to
4284 work in terms of the new functions. Also, a set of functions
4285 to handle the internal storage of the configuration data is
4286 provided to make it easier to write new configuration file
4287 reader "classes" (I can definitely see something reading a
4288 configuration file in XML format, for example), called _CONF_*,
4289 or "the configuration storage API"...
4291 The new configuration file reading functions are:
4293 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4294 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4296 NCONF_default, NCONF_WIN32
4298 NCONF_dump_fp, NCONF_dump_bio
4300 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4301 NCONF_new creates a new CONF object. This works in the same way
4302 as other interfaces in OpenSSL, like the BIO interface.
4303 NCONF_dump_* dump the internal storage of the configuration file,
4304 which is useful for debugging. All other functions take the same
4305 arguments as the old CONF_* functions wth the exception of the
4306 first that must be a `CONF *' instead of a `LHASH *'.
4308 To make it easer to use the new classes with the old CONF_* functions,
4309 the function CONF_set_default_method is provided.
4312 *) Add '-tls1' option to 'openssl ciphers', which was already
4313 mentioned in the documentation but had not been implemented.
4314 (This option is not yet really useful because even the additional
4315 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4318 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4319 OpenSSL-based applications) load shared libraries and bind to
4320 them in a portable way.
4321 [Geoff Thorpe, with contributions from Richard Levitte]
4323 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4325 *) Make sure _lrotl and _lrotr are only used with MSVC.
4327 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4328 (the default implementation of RAND_status).
4330 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4331 to '-clrext' (= clear extensions), as intended and documented.
4332 [Bodo Moeller; inconsistency pointed out by Michael Attili
4335 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4336 was larger than the MD block size.
4337 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4339 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4340 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4341 using the passed key: if the passed key was a private key the result
4342 of X509_print(), for example, would be to print out all the private key
4346 *) des_quad_cksum() byte order bug fix.
4347 [Ulf Möller, using the problem description in krb4-0.9.7, where
4348 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4350 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4352 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4354 *) For easily testing in shell scripts whether some command
4355 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4356 returns with exit code 0 iff no command of the given name is available.
4357 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4358 the output goes to stdout and nothing is printed to stderr.
4359 Additional arguments are always ignored.
4361 Since for each cipher there is a command of the same name,
4362 the 'no-cipher' compilation switches can be tested this way.
4364 ('openssl no-XXX' is not able to detect pseudo-commands such
4365 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4368 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4371 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4372 is set; it will be thrown away anyway because each handshake creates
4374 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4375 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4376 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4377 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4380 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4381 'Q' and 'R' lose their special meanings (quit/renegotiate).
4382 This is part of what -quiet does; unlike -quiet, -ign_eof
4383 does not suppress any output.
4386 *) Add compatibility options to the purpose and trust code. The
4387 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4388 accepts a certificate or CA, this was the previous behaviour,
4389 with all the associated security issues.
4391 X509_TRUST_COMPAT is the old trust behaviour: only and
4392 automatically trust self signed roots in certificate store. A
4393 new trust setting X509_TRUST_DEFAULT is used to specify that
4394 a purpose has no associated trust setting and it should instead
4395 use the value in the default purpose.
4398 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4399 and fix a memory leak.
4402 *) In util/mkerr.pl (which implements 'make errors'), preserve
4403 reason strings from the previous version of the .c file, as
4404 the default to have only downcase letters (and digits) in
4405 automatically generated reasons codes is not always appropriate.
4408 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4409 using strerror. Previously, ERR_reason_error_string() returned
4410 library names as reason strings for SYSerr; but SYSerr is a special
4411 case where small numbers are errno values, not library numbers.
4414 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4415 converts DSA parameters into DH parameters. (When creating parameters,
4416 DSA_generate_parameters is used.)
4419 *) Include 'length' (recommended exponent length) in C code generated
4420 by 'openssl dhparam -C'.
4423 *) The second argument to set_label in perlasm was already being used
4424 so couldn't be used as a "file scope" flag. Moved to third argument
4428 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4429 instead of RAND_bytes for encryption IVs and salts.
4432 *) Include RAND_status() into RAND_METHOD instead of implementing
4433 it only for md_rand.c Otherwise replacing the PRNG by calling
4434 RAND_set_rand_method would be impossible.
4437 *) Don't let DSA_generate_key() enter an infinite loop if the random
4438 number generation fails.
4441 *) New 'rand' application for creating pseudo-random output.
4444 *) Added configuration support for Linux/IA64
4445 [Rolf Haberrecker <rolf@suse.de>]
4447 *) Assembler module support for Mingw32.
4450 *) Shared library support for HPUX (in shlib/).
4451 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4453 *) Shared library support for Solaris gcc.
4454 [Lutz Behnke <behnke@trustcenter.de>]
4456 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4458 *) PKCS7_encrypt() was adding text MIME headers twice because they
4459 were added manually and by SMIME_crlf_copy().
4462 *) In bntest.c don't call BN_rand with zero bits argument.
4463 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4465 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4466 case was implemented. This caused BN_div_recp() to fail occasionally.
4469 *) Add an optional second argument to the set_label() in the perl
4470 assembly language builder. If this argument exists and is set
4471 to 1 it signals that the assembler should use a symbol whose
4472 scope is the entire file, not just the current function. This
4473 is needed with MASM which uses the format label:: for this scope.
4474 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4476 *) Change the ASN1 types so they are typedefs by default. Before
4477 almost all types were #define'd to ASN1_STRING which was causing
4478 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4482 *) Change names of new functions to the new get1/get0 naming
4483 convention: After 'get1', the caller owns a reference count
4484 and has to call ..._free; 'get0' returns a pointer to some
4485 data structure without incrementing reference counters.
4486 (Some of the existing 'get' functions increment a reference
4487 counter, some don't.)
4488 Similarly, 'set1' and 'add1' functions increase reference
4489 counters or duplicate objects.
4492 *) Allow for the possibility of temp RSA key generation failure:
4493 the code used to assume it always worked and crashed on failure.
4496 *) Fix potential buffer overrun problem in BIO_printf().
4497 [Ulf Möller, using public domain code by Patrick Powell; problem
4498 pointed out by David Sacerdote <das33@cornell.edu>]
4500 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4501 RAND_egd() and RAND_status(). In the command line application,
4502 the EGD socket can be specified like a seed file using RANDFILE
4506 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4507 Some CAs (e.g. Verisign) distribute certificates in this form.
4510 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4511 list to exclude them. This means that no special compilation option
4512 is needed to use anonymous DH: it just needs to be included in the
4516 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4517 EVP_MD_type. The old functionality is available in a new macro called
4518 EVP_MD_md(). Change code that uses it and update docs.
4521 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4522 where the 'void *' argument is replaced by a function pointer argument.
4523 Previously 'void *' was abused to point to functions, which works on
4524 many platforms, but is not correct. As these functions are usually
4525 called by macros defined in OpenSSL header files, most source code
4526 should work without changes.
4529 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4530 sections with information on -D... compiler switches used for
4531 compiling the library so that applications can see them. To enable
4532 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4533 must be defined. E.g.,
4534 #define OPENSSL_ALGORITHM_DEFINES
4535 #include <openssl/opensslconf.h>
4536 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4537 [Richard Levitte, Ulf and Bodo Möller]
4539 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4543 *) Change the 'other' type in certificate aux info to a STACK_OF
4544 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4545 the required ASN1 format: arbitrary types determined by an OID.
4548 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4549 argument to 'req'. This is not because the function is newer or
4550 better than others it just uses the work 'NEW' in the certificate
4551 request header lines. Some software needs this.
4554 *) Reorganise password command line arguments: now passwords can be
4555 obtained from various sources. Delete the PEM_cb function and make
4556 it the default behaviour: i.e. if the callback is NULL and the
4557 usrdata argument is not NULL interpret it as a null terminated pass
4558 phrase. If usrdata and the callback are NULL then the pass phrase
4559 is prompted for as usual.
4562 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4563 the support is automatically enabled. The resulting binaries will
4564 autodetect the card and use it if present.
4565 [Ben Laurie and Compaq Inc.]
4567 *) Work around for Netscape hang bug. This sends certificate request
4568 and server done in one record. Since this is perfectly legal in the
4569 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4570 the bugs/SSLv3 entry for more info.
4573 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4576 *) Add -rand argument to smime and pkcs12 applications and read/write
4580 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4583 *) Add command line password options to the remaining applications.
4586 *) Bug fix for BN_div_recp() for numerators with an even number of
4590 *) More tests in bntest.c, and changed test_bn output.
4593 *) ./config recognizes MacOS X now.
4596 *) Bug fix for BN_div() when the first words of num and divsor are
4597 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4600 *) Add support for various broken PKCS#8 formats, and command line
4601 options to produce them.
4604 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4605 get temporary BIGNUMs from a BN_CTX.
4608 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4612 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4613 include a #define from the old name to the new. The original intent
4614 was that statically linked binaries could for example just call
4615 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4616 link with digests. This never worked becayse SSLeay_add_all_digests()
4617 and SSLeay_add_all_ciphers() were in the same source file so calling
4618 one would link with the other. They are now in separate source files.
4621 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4624 *) Use a less unusual form of the Miller-Rabin primality test (it used
4625 a binary algorithm for exponentiation integrated into the Miller-Rabin
4626 loop, our standard modexp algorithms are faster).
4629 *) Support for the EBCDIC character set completed.
4630 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4632 *) Source code cleanups: use const where appropriate, eliminate casts,
4633 use void * instead of char * in lhash.
4636 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4637 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4638 this the server could overwrite ephemeral keys that the client
4642 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4643 using 50 iterations of the Rabin-Miller test.
4645 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4646 iterations of the Rabin-Miller test as required by the appendix
4647 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4648 As BN_is_prime_fasttest includes trial division, DSA parameter
4649 generation becomes much faster.
4651 This implies a change for the callback functions in DSA_is_prime
4652 and DSA_generate_parameters: The callback function is called once
4653 for each positive witness in the Rabin-Miller test, not just
4654 occasionally in the inner loop; and the parameters to the
4655 callback function now provide an iteration count for the outer
4656 loop rather than for the current invocation of the inner loop.
4657 DSA_generate_parameters additionally can call the callback
4658 function with an 'iteration count' of -1, meaning that a
4659 candidate has passed the trial division test (when q is generated
4660 from an application-provided seed, trial division is skipped).
4663 *) New function BN_is_prime_fasttest that optionally does trial
4664 division before starting the Rabin-Miller test and has
4665 an additional BN_CTX * argument (whereas BN_is_prime always
4666 has to allocate at least one BN_CTX).
4667 'callback(1, -1, cb_arg)' is called when a number has passed the
4668 trial division stage.
4671 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4675 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4678 *) New function BN_pseudo_rand().
4681 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4682 bignum version of BN_from_montgomery() with the working code from
4683 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4687 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4688 made it impossible to use the same SSL_SESSION data structure in
4689 SSL2 clients in multiple threads.
4692 *) The return value of RAND_load_file() no longer counts bytes obtained
4693 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4694 to seed the PRNG (previously an explicit byte count was required).
4695 [Ulf Möller, Bodo Möller]
4697 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4698 used (char *) instead of (void *) and had casts all over the place.
4701 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4704 *) Retain source code compatibility for BN_prime_checks macro:
4705 BN_is_prime(..., BN_prime_checks, ...) now uses
4706 BN_prime_checks_for_size to determine the appropriate number of
4707 Rabin-Miller iterations.
4710 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4711 DH_CHECK_P_NOT_SAFE_PRIME.
4712 (Check if this is true? OpenPGP calls them "strong".)
4715 *) Merge the functionality of "dh" and "gendh" programs into a new program
4716 "dhparam". The old programs are retained for now but will handle DH keys
4717 (instead of parameters) in future.
4720 *) Make the ciphers, s_server and s_client programs check the return values
4721 when a new cipher list is set.
4724 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4725 ciphers. Before when the 56bit ciphers were enabled the sorting was
4728 The syntax for the cipher sorting has been extended to support sorting by
4729 cipher-strength (using the strength_bits hard coded in the tables).
4730 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4732 Fix a bug in the cipher-command parser: when supplying a cipher command
4733 string with an "undefined" symbol (neither command nor alphanumeric
4734 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4735 an error is flagged.
4737 Due to the strength-sorting extension, the code of the
4738 ssl_create_cipher_list() function was completely rearranged. I hope that
4739 the readability was also increased :-)
4740 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4742 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4743 for the first serial number and places 2 in the serial number file. This
4744 avoids problems when the root CA is created with serial number zero and
4745 the first user certificate has the same issuer name and serial number
4749 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4750 the new code. Add documentation for this stuff.
4753 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4754 X509_*() to X509at_*() on the grounds that they don't handle X509
4755 structures and behave in an analagous way to the X509v3 functions:
4756 they shouldn't be called directly but wrapper functions should be used
4759 So we also now have some wrapper functions that call the X509at functions
4760 when passed certificate requests. (TO DO: similar things can be done with
4761 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4762 things. Some of these need some d2i or i2d and print functionality
4763 because they handle more complex structures.)
4766 *) Add missing #ifndefs that caused missing symbols when building libssl
4767 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4768 NO_RSA in ssl/s2*.c.
4769 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4771 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4772 has a return value which indicates the quality of the random data
4773 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4774 error queue. New function RAND_pseudo_bytes() generates output that is
4775 guaranteed to be unique but not unpredictable. RAND_add is like
4776 RAND_seed, but takes an extra argument for an entropy estimate
4777 (RAND_seed always assumes full entropy).
4780 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4781 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4782 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4783 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4784 false-positive rate of at most 2^-80 for random input.
4787 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4790 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4791 in the 0.9.5 release), this returns the chain
4792 from an X509_CTX structure with a dup of the stack and all
4793 the X509 reference counts upped: so the stack will exist
4794 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4797 Also make SSL_SESSION_print() print out the verify return
4801 *) Add manpage for the pkcs12 command. Also change the default
4802 behaviour so MAC iteration counts are used unless the new
4803 -nomaciter option is used. This improves file security and
4804 only older versions of MSIE (4.0 for example) need it.
4807 *) Honor the no-xxx Configure options when creating .DEF files.
4810 *) Add PKCS#10 attributes to field table: challengePassword,
4811 unstructuredName and unstructuredAddress. These are taken from
4812 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4813 international characters are used.
4815 More changes to X509_ATTRIBUTE code: allow the setting of types
4816 based on strings. Remove the 'loc' parameter when adding
4817 attributes because these will be a SET OF encoding which is sorted
4821 *) Initial changes to the 'req' utility to allow request generation
4822 automation. This will allow an application to just generate a template
4823 file containing all the field values and have req construct the
4826 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4827 used all over the place including certificate requests and PKCS#7
4828 structures. They are currently handled manually where necessary with
4829 some primitive wrappers for PKCS#7. The new functions behave in a
4830 manner analogous to the X509 extension functions: they allow
4831 attributes to be looked up by NID and added.
4833 Later something similar to the X509V3 code would be desirable to
4834 automatically handle the encoding, decoding and printing of the
4835 more complex types. The string types like challengePassword can
4836 be handled by the string table functions.
4838 Also modified the multi byte string table handling. Now there is
4839 a 'global mask' which masks out certain types. The table itself
4840 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4841 is useful when for example there is only one permissible type
4842 (as in countryName) and using the mask might result in no valid
4846 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4847 SSL_get_peer_finished to allow applications to obtain the latest
4848 Finished messages sent to the peer or expected from the peer,
4849 respectively. (SSL_get_peer_finished is usually the Finished message
4850 actually received from the peer, otherwise the protocol will be aborted.)
4852 As the Finished message are message digests of the complete handshake
4853 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4854 be used for external authentication procedures when the authentication
4855 provided by SSL/TLS is not desired or is not enough.
4858 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4859 the host supports BWX extension and if Compaq C is present on the
4860 $PATH. Just exploiting of the BWX extension results in 20-30%
4861 performance kick for some algorithms, e.g. DES and RC4 to mention
4862 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4866 *) Add support for MS "fast SGC". This is arguably a violation of the
4867 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4868 weak crypto and after checking the certificate is SGC a second one
4869 with strong crypto. MS SGC stops the first handshake after receiving
4870 the server certificate message and sends a second client hello. Since
4871 a server will typically do all the time consuming operations before
4872 expecting any further messages from the client (server key exchange
4873 is the most expensive) there is little difference between the two.
4875 To get OpenSSL to support MS SGC we have to permit a second client
4876 hello message after we have sent server done. In addition we have to
4877 reset the MAC if we do get this second client hello.
4880 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
4881 if a DER encoded private key is RSA or DSA traditional format. Changed
4882 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
4883 format DER encoded private key. Newer code should use PKCS#8 format which
4884 has the key type encoded in the ASN1 structure. Added DER private key
4885 support to pkcs8 application.
4888 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
4889 ciphersuites has been selected (as required by the SSL 3/TLS 1
4890 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
4891 is set, we interpret this as a request to violate the specification
4892 (the worst that can happen is a handshake failure, and 'correct'
4893 behaviour would result in a handshake failure anyway).
4896 *) In SSL_CTX_add_session, take into account that there might be multiple
4897 SSL_SESSION structures with the same session ID (e.g. when two threads
4898 concurrently obtain them from an external cache).
4899 The internal cache can handle only one SSL_SESSION with a given ID,
4900 so if there's a conflict, we now throw out the old one to achieve
4904 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
4905 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
4906 some routines that use cipher OIDs: some ciphers do not have OIDs
4907 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
4911 *) Simplify the trust setting structure and code. Now we just have
4912 two sequences of OIDs for trusted and rejected settings. These will
4913 typically have values the same as the extended key usage extension
4914 and any application specific purposes.
4916 The trust checking code now has a default behaviour: it will just
4917 check for an object with the same NID as the passed id. Functions can
4918 be provided to override either the default behaviour or the behaviour
4919 for a given id. SSL client, server and email already have functions
4920 in place for compatibility: they check the NID and also return "trusted"
4921 if the certificate is self signed.
4924 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
4925 traditional format into an EVP_PKEY structure.
4928 *) Add a password callback function PEM_cb() which either prompts for
4929 a password if usr_data is NULL or otherwise assumes it is a null
4930 terminated password. Allow passwords to be passed on command line
4931 environment or config files in a few more utilities.
4934 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
4935 keys. Add some short names for PKCS#8 PBE algorithms and allow them
4936 to be specified on the command line for the pkcs8 and pkcs12 utilities.
4937 Update documentation.
4940 *) Support for ASN1 "NULL" type. This could be handled before by using
4941 ASN1_TYPE but there wasn't any function that would try to read a NULL
4942 and produce an error if it couldn't. For compatibility we also have
4943 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
4944 don't allocate anything because they don't need to.
4947 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
4949 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
4951 *) Rebuild of the memory allocation routines used by OpenSSL code and
4952 possibly others as well. The purpose is to make an interface that
4953 provide hooks so anyone can build a separate set of allocation and
4954 deallocation routines to be used by OpenSSL, for example memory
4955 pool implementations, or something else, which was previously hard
4956 since Malloc(), Realloc() and Free() were defined as macros having
4957 the values malloc, realloc and free, respectively (except for Win32
4958 compilations). The same is provided for memory debugging code.
4959 OpenSSL already comes with functionality to find memory leaks, but
4960 this gives people a chance to debug other memory problems.
4962 With these changes, a new set of functions and macros have appeared:
4964 CRYPTO_set_mem_debug_functions() [F]
4965 CRYPTO_get_mem_debug_functions() [F]
4966 CRYPTO_dbg_set_options() [F]
4967 CRYPTO_dbg_get_options() [F]
4968 CRYPTO_malloc_debug_init() [M]
4970 The memory debug functions are NULL by default, unless the library
4971 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
4972 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
4973 gives the standard debugging functions that come with OpenSSL) or
4974 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
4975 provided by the library user) must be used. When the standard
4976 debugging functions are used, CRYPTO_dbg_set_options can be used to
4977 request additional information:
4978 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
4979 the CRYPTO_MDEBUG_xxx macro when compiling the library.
4981 Also, things like CRYPTO_set_mem_functions will always give the
4982 expected result (the new set of functions is used for allocation
4983 and deallocation) at all times, regardless of platform and compiler
4986 To finish it up, some functions that were never use in any other
4987 way than through macros have a new API and new semantic:
4990 CRYPTO_dbg_realloc()
4993 All macros of value have retained their old syntax.
4994 [Richard Levitte and Bodo Moeller]
4996 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
4997 ordering of SMIMECapabilities wasn't in "strength order" and there
4998 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5002 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5003 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5004 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5006 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5007 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5008 functionality to handle multipart/signed properly) and a utility
5009 called 'smime' to call all this stuff. This is based on code I
5010 originally wrote for Celo who have kindly allowed it to be
5011 included in OpenSSL.
5014 *) Add variants des_set_key_checked and des_set_key_unchecked of
5015 des_set_key (aka des_key_sched). Global variable des_check_key
5016 decides which of these is called by des_set_key; this way
5017 des_check_key behaves as it always did, but applications and
5018 the library itself, which was buggy for des_check_key == 1,
5019 have a cleaner way to pick the version they need.
5022 *) New function PKCS12_newpass() which changes the password of a
5026 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5027 dynamic mix. In both cases the ids can be used as an index into the
5028 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5029 functions so they accept a list of the field values and the
5030 application doesn't need to directly manipulate the X509_TRUST
5034 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5038 *) Modify the way the V3 extension code looks up extensions. This now
5039 works in a similar way to the object code: we have some "standard"
5040 extensions in a static table which is searched with OBJ_bsearch()
5041 and the application can add dynamic ones if needed. The file
5042 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5043 updated whenever a new extension is added to the core code and kept
5044 in ext_nid order. There is a simple program 'tabtest.c' which checks
5045 this. New extensions are not added too often so this file can readily
5046 be maintained manually.
5048 There are two big advantages in doing things this way. The extensions
5049 can be looked up immediately and no longer need to be "added" using
5050 X509V3_add_standard_extensions(): this function now does nothing.
5051 [Side note: I get *lots* of email saying the extension code doesn't
5052 work because people forget to call this function]
5053 Also no dynamic allocation is done unless new extensions are added:
5054 so if we don't add custom extensions there is no need to call
5055 X509V3_EXT_cleanup().
5058 *) Modify enc utility's salting as follows: make salting the default. Add a
5059 magic header, so unsalted files fail gracefully instead of just decrypting
5060 to garbage. This is because not salting is a big security hole, so people
5061 should be discouraged from doing it.
5064 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5065 digest to be passed on the command line but it only used this
5066 parameter when signing a certificate. Modified so all relevant
5067 operations are affected by the digest parameter including the
5068 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5069 DSA key was used because it didn't fix the digest.
5072 *) Initial certificate chain verify code. Currently tests the untrusted
5073 certificates for consistency with the verify purpose (which is set
5074 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5076 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5077 this is because it will reject chains with invalid extensions whereas
5078 every previous version of OpenSSL and SSLeay made no checks at all.
5080 Trust code: checks the root CA for the relevant trust settings. Trust
5081 settings have an initial value consistent with the verify purpose: e.g.
5082 if the verify purpose is for SSL client use it expects the CA to be
5083 trusted for SSL client use. However the default value can be changed to
5084 permit custom trust settings: one example of this would be to only trust
5085 certificates from a specific "secure" set of CAs.
5087 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5088 which should be used for version portability: especially since the
5089 verify structure is likely to change more often now.
5091 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5092 to set them. If not set then assume SSL clients will verify SSL servers
5095 Two new options to the verify program: -untrusted allows a set of
5096 untrusted certificates to be passed in and -purpose which sets the
5097 intended purpose of the certificate. If a purpose is set then the
5098 new chain verify code is used to check extension consistency.
5101 *) Support for the authority information access extension.
5104 *) Modify RSA and DSA PEM read routines to transparently handle
5105 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5106 public keys in a format compatible with certificate
5107 SubjectPublicKeyInfo structures. Unfortunately there were already
5108 functions called *_PublicKey_* which used various odd formats so
5109 these are retained for compatibility: however the DSA variants were
5110 never in a public release so they have been deleted. Changed dsa/rsa
5111 utilities to handle the new format: note no releases ever handled public
5112 keys so we should be OK.
5114 The primary motivation for this change is to avoid the same fiasco
5115 that dogs private keys: there are several incompatible private key
5116 formats some of which are standard and some OpenSSL specific and
5117 require various evil hacks to allow partial transparent handling and
5118 even then it doesn't work with DER formats. Given the option anything
5119 other than PKCS#8 should be dumped: but the other formats have to
5120 stay in the name of compatibility.
5122 With public keys and the benefit of hindsight one standard format
5123 is used which works with EVP_PKEY, RSA or DSA structures: though
5124 it clearly returns an error if you try to read the wrong kind of key.
5126 Added a -pubkey option to the 'x509' utility to output the public key.
5127 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5128 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5129 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5130 that do the same as the EVP_PKEY_assign_*() except they up the
5131 reference count of the added key (they don't "swallow" the
5135 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5136 CRLs would fail if the file contained no certificates or no CRLs:
5137 added a new function to read in both types and return the number
5138 read: this means that if none are read it will be an error. The
5139 DER versions of the certificate and CRL reader would always fail
5140 because it isn't possible to mix certificates and CRLs in DER format
5141 without choking one or the other routine. Changed this to just read
5142 a certificate: this is the best we can do. Also modified the code
5143 in apps/verify.c to take notice of return codes: it was previously
5144 attempting to read in certificates from NULL pointers and ignoring
5145 any errors: this is one reason why the cert and CRL reader seemed
5146 to work. It doesn't check return codes from the default certificate
5147 routines: these may well fail if the certificates aren't installed.
5150 *) Code to support otherName option in GeneralName.
5153 *) First update to verify code. Change the verify utility
5154 so it warns if it is passed a self signed certificate:
5155 for consistency with the normal behaviour. X509_verify
5156 has been modified to it will now verify a self signed
5157 certificate if *exactly* the same certificate appears
5158 in the store: it was previously impossible to trust a
5159 single self signed certificate. This means that:
5160 openssl verify ss.pem
5161 now gives a warning about a self signed certificate but
5162 openssl verify -CAfile ss.pem ss.pem
5166 *) For servers, store verify_result in SSL_SESSION data structure
5167 (and add it to external session representation).
5168 This is needed when client certificate verifications fails,
5169 but an application-provided verification callback (set by
5170 SSL_CTX_set_cert_verify_callback) allows accepting the session
5171 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5172 but returns 1): When the session is reused, we have to set
5173 ssl->verify_result to the appropriate error code to avoid
5175 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5177 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5178 case in PKCS7_dataInit() where the signed PKCS7 structure
5179 didn't contain any existing data because it was being created.
5180 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5182 *) Add a salt to the key derivation routines in enc.c. This
5183 forms the first 8 bytes of the encrypted file. Also add a
5184 -S option to allow a salt to be input on the command line.
5187 *) New function X509_cmp(). Oddly enough there wasn't a function
5188 to compare two certificates. We do this by working out the SHA1
5189 hash and comparing that. X509_cmp() will be needed by the trust
5193 *) SSL_get1_session() is like SSL_get_session(), but increments
5194 the reference count in the SSL_SESSION returned.
5195 [Geoff Thorpe <geoff@eu.c2.net>]
5197 *) Fix for 'req': it was adding a null to request attributes.
5198 Also change the X509_LOOKUP and X509_INFO code to handle
5199 certificate auxiliary information.
5202 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5206 *) Add the possibility to add extra information to the memory leak
5207 detecting output, to form tracebacks, showing from where each
5208 allocation was originated: CRYPTO_push_info("constant string") adds
5209 the string plus current file name and line number to a per-thread
5210 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5211 is like calling CYRPTO_pop_info() until the stack is empty.
5212 Also updated memory leak detection code to be multi-thread-safe.
5215 *) Add options -text and -noout to pkcs7 utility and delete the
5216 encryption options which never did anything. Update docs.
5219 *) Add options to some of the utilities to allow the pass phrase
5220 to be included on either the command line (not recommended on
5221 OSes like Unix) or read from the environment. Update the
5222 manpages and fix a few bugs.
5225 *) Add a few manpages for some of the openssl commands.
5228 *) Fix the -revoke option in ca. It was freeing up memory twice,
5229 leaking and not finding already revoked certificates.
5232 *) Extensive changes to support certificate auxiliary information.
5233 This involves the use of X509_CERT_AUX structure and X509_AUX
5234 functions. An X509_AUX function such as PEM_read_X509_AUX()
5235 can still read in a certificate file in the usual way but it
5236 will also read in any additional "auxiliary information". By
5237 doing things this way a fair degree of compatibility can be
5238 retained: existing certificates can have this information added
5239 using the new 'x509' options.
5241 Current auxiliary information includes an "alias" and some trust
5242 settings. The trust settings will ultimately be used in enhanced
5243 certificate chain verification routines: currently a certificate
5244 can only be trusted if it is self signed and then it is trusted
5248 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5249 The problem was that one of the replacement routines had not been working
5250 since SSLeay releases. For now the offending routine has been replaced
5251 with non-optimised assembler. Even so, this now gives around 95%
5252 performance improvement for 1024 bit RSA signs.
5255 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5256 handling. Most clients have the effective key size in bits equal to
5257 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5258 A few however don't do this and instead use the size of the decrypted key
5259 to determine the RC2 key length and the AlgorithmIdentifier to determine
5260 the effective key length. In this case the effective key length can still
5261 be 40 bits but the key length can be 168 bits for example. This is fixed
5262 by manually forcing an RC2 key into the EVP_PKEY structure because the
5263 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5264 the key length and effective key length are equal.
5267 *) Add a bunch of functions that should simplify the creation of
5268 X509_NAME structures. Now you should be able to do:
5269 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5270 and have it automatically work out the correct field type and fill in
5271 the structures. The more adventurous can try:
5272 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5273 and it will (hopefully) work out the correct multibyte encoding.
5276 *) Change the 'req' utility to use the new field handling and multibyte
5277 copy routines. Before the DN field creation was handled in an ad hoc
5278 way in req, ca, and x509 which was rather broken and didn't support
5279 BMPStrings or UTF8Strings. Since some software doesn't implement
5280 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5281 using the dirstring_type option. See the new comment in the default
5282 openssl.cnf for more info.
5285 *) Make crypto/rand/md_rand.c more robust:
5286 - Assure unique random numbers after fork().
5287 - Make sure that concurrent threads access the global counter and
5288 md serializably so that we never lose entropy in them
5289 or use exactly the same state in multiple threads.
5290 Access to the large state is not always serializable because
5291 the additional locking could be a performance killer, and
5292 md should be large enough anyway.
5295 *) New file apps/app_rand.c with commonly needed functionality
5296 for handling the random seed file.
5298 Use the random seed file in some applications that previously did not:
5300 dsaparam -genkey (which also ignored its '-rand' option),
5303 x509 (when signing).
5304 Except on systems with /dev/urandom, it is crucial to have a random
5305 seed file at least for key creation, DSA signing, and for DH exchanges;
5306 for RSA signatures we could do without one.
5308 gendh and gendsa (unlike genrsa) used to read only the first byte
5309 of each file listed in the '-rand' option. The function as previously
5310 found in genrsa is now in app_rand.c and is used by all programs
5311 that support '-rand'.
5314 *) In RAND_write_file, use mode 0600 for creating files;
5315 don't just chmod when it may be too late.
5318 *) Report an error from X509_STORE_load_locations
5319 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5322 *) New function ASN1_mbstring_copy() this copies a string in either
5323 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5324 into an ASN1_STRING type. A mask of permissible types is passed
5325 and it chooses the "minimal" type to use or an error if not type
5329 *) Add function equivalents to the various macros in asn1.h. The old
5330 macros are retained with an M_ prefix. Code inside the library can
5331 use the M_ macros. External code (including the openssl utility)
5332 should *NOT* in order to be "shared library friendly".
5335 *) Add various functions that can check a certificate's extensions
5336 to see if it usable for various purposes such as SSL client,
5337 server or S/MIME and CAs of these types. This is currently
5338 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5339 verification. Also added a -purpose flag to x509 utility to
5340 print out all the purposes.
5343 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5347 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5348 for, obtain and decode and extension and obtain its critical flag.
5349 This allows all the necessary extension code to be handled in a
5350 single function call.
5353 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5354 platforms. See crypto/rc4/rc4_enc.c for further details.
5357 *) New -noout option to asn1parse. This causes no output to be produced
5358 its main use is when combined with -strparse and -out to extract data
5359 from a file (which may not be in ASN.1 format).
5362 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5363 when producing the local key id.
5364 [Richard Levitte <levitte@stacken.kth.se>]
5366 *) New option -dhparam in s_server. This allows a DH parameter file to be
5367 stated explicitly. If it is not stated then it tries the first server
5368 certificate file. The previous behaviour hard coded the filename
5372 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5373 a public key to be input or output. For example:
5374 openssl rsa -in key.pem -pubout -out pubkey.pem
5375 Also added necessary DSA public key functions to handle this.
5378 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5379 in the message. This was handled by allowing
5380 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5381 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5383 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5384 to the end of the strings whereas this didn't. This would cause problems
5385 if strings read with d2i_ASN1_bytes() were later modified.
5386 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5388 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5389 data and it contains EOF it will end up returning an error. This is
5390 caused by input 46 bytes long. The cause is due to the way base64
5391 BIOs find the start of base64 encoded data. They do this by trying a
5392 trial decode on each line until they find one that works. When they
5393 do a flag is set and it starts again knowing it can pass all the
5394 data directly through the decoder. Unfortunately it doesn't reset
5395 the context it uses. This means that if EOF is reached an attempt
5396 is made to pass two EOFs through the context and this causes the
5397 resulting error. This can also cause other problems as well. As is
5398 usual with these problems it takes *ages* to find and the fix is
5399 trivial: move one line.
5400 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5402 *) Ugly workaround to get s_client and s_server working under Windows. The
5403 old code wouldn't work because it needed to select() on sockets and the
5404 tty (for keypresses and to see if data could be written). Win32 only
5405 supports select() on sockets so we select() with a 1s timeout on the
5406 sockets and then see if any characters are waiting to be read, if none
5407 are present then we retry, we also assume we can always write data to
5408 the tty. This isn't nice because the code then blocks until we've
5409 received a complete line of data and it is effectively polling the
5410 keyboard at 1s intervals: however it's quite a bit better than not
5411 working at all :-) A dedicated Windows application might handle this
5412 with an event loop for example.
5415 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5416 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5417 will be called when RSA_sign() and RSA_verify() are used. This is useful
5418 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5419 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5420 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5421 This necessitated the support of an extra signature type NID_md5_sha1
5422 for SSL signatures and modifications to the SSL library to use it instead
5423 of calling RSA_public_decrypt() and RSA_private_encrypt().
5426 *) Add new -verify -CAfile and -CApath options to the crl program, these
5427 will lookup a CRL issuers certificate and verify the signature in a
5428 similar way to the verify program. Tidy up the crl program so it
5429 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5430 less strict. It will now permit CRL extensions even if it is not
5431 a V2 CRL: this will allow it to tolerate some broken CRLs.
5434 *) Initialize all non-automatic variables each time one of the openssl
5435 sub-programs is started (this is necessary as they may be started
5436 multiple times from the "OpenSSL>" prompt).
5437 [Lennart Bang, Bodo Moeller]
5439 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5440 removing all other RSA functionality (this is what NO_RSA does). This
5441 is so (for example) those in the US can disable those operations covered
5442 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5446 *) Non-copying interface to BIO pairs.
5447 (still largely untested)
5450 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5451 ASCII string. This was handled independently in various places before.
5454 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5455 UTF8 strings a character at a time.
5458 *) Use client_version from client hello to select the protocol
5459 (s23_srvr.c) and for RSA client key exchange verification
5460 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5463 *) Add various utility functions to handle SPKACs, these were previously
5464 handled by poking round in the structure internals. Added new function
5465 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5466 print, verify and generate SPKACs. Based on an original idea from
5467 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5470 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5473 *) Allow the config file extension section to be overwritten on the
5474 command line. Based on an original idea from Massimiliano Pala
5475 <madwolf@comune.modena.it>. The new option is called -extensions
5476 and can be applied to ca, req and x509. Also -reqexts to override
5477 the request extensions in req and -crlexts to override the crl extensions
5481 *) Add new feature to the SPKAC handling in ca. Now you can include
5482 the same field multiple times by preceding it by "XXXX." for example:
5485 this is the same syntax as used in the req config file.
5488 *) Allow certificate extensions to be added to certificate requests. These
5489 are specified in a 'req_extensions' option of the req section of the
5490 config file. They can be printed out with the -text option to req but
5491 are otherwise ignored at present.
5494 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5495 data read consists of only the final block it would not decrypted because
5496 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5497 A misplaced 'break' also meant the decrypted final block might not be
5498 copied until the next read.
5501 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5502 a few extra parameters to the DH structure: these will be useful if
5503 for example we want the value of 'q' or implement X9.42 DH.
5506 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5507 provides hooks that allow the default DSA functions or functions on a
5508 "per key" basis to be replaced. This allows hardware acceleration and
5509 hardware key storage to be handled without major modification to the
5510 library. Also added low level modexp hooks and CRYPTO_EX structure and
5511 associated functions.
5514 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5515 as "read only": it can't be written to and the buffer it points to will
5516 not be freed. Reading from a read only BIO is much more efficient than
5517 a normal memory BIO. This was added because there are several times when
5518 an area of memory needs to be read from a BIO. The previous method was
5519 to create a memory BIO and write the data to it, this results in two
5520 copies of the data and an O(n^2) reading algorithm. There is a new
5521 function BIO_new_mem_buf() which creates a read only memory BIO from
5522 an area of memory. Also modified the PKCS#7 routines to use read only
5526 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5527 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5528 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5529 but a retry condition occured while trying to read the rest.
5532 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5533 NID_pkcs7_encrypted by default: this was wrong since this should almost
5534 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5535 the encrypted data type: this is a more sensible place to put it and it
5536 allows the PKCS#12 code to be tidied up that duplicated this
5540 *) Changed obj_dat.pl script so it takes its input and output files on
5541 the command line. This should avoid shell escape redirection problems
5545 *) Initial support for certificate extension requests, these are included
5546 in things like Xenroll certificate requests. Included functions to allow
5547 extensions to be obtained and added.
5550 *) -crlf option to s_client and s_server for sending newlines as
5551 CRLF (as required by many protocols).
5554 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5556 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5557 [Ralf S. Engelschall]
5559 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5560 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5562 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5566 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5567 DH parameters/keys (q is lost during that conversion, but the resulting
5568 DH parameters contain its length).
5570 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5571 much faster than DH_generate_parameters (which creates parameters
5572 where p = 2*q + 1), and also the smaller q makes DH computations
5573 much more efficient (160-bit exponentiation instead of 1024-bit
5574 exponentiation); so this provides a convenient way to support DHE
5575 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5576 utter importance to use
5577 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5579 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5580 when such DH parameters are used, because otherwise small subgroup
5581 attacks may become possible!
5584 *) Avoid memory leak in i2d_DHparams.
5587 *) Allow the -k option to be used more than once in the enc program:
5588 this allows the same encrypted message to be read by multiple recipients.
5591 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5592 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5593 it will always use the numerical form of the OID, even if it has a short
5597 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5598 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5599 otherwise bn_mod_exp was called. In the case of hardware keys for example
5600 no private key components need be present and it might store extra data
5601 in the RSA structure, which cannot be accessed from bn_mod_exp.
5602 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5603 private key operations.
5606 *) Added support for SPARC Linux.
5609 *) pem_password_cb function type incompatibly changed from
5610 typedef int pem_password_cb(char *buf, int size, int rwflag);
5612 ....(char *buf, int size, int rwflag, void *userdata);
5613 so that applications can pass data to their callbacks:
5614 The PEM[_ASN1]_{read,write}... functions and macros now take an
5615 additional void * argument, which is just handed through whenever
5616 the password callback is called.
5617 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5619 New function SSL_CTX_set_default_passwd_cb_userdata.
5621 Compatibility note: As many C implementations push function arguments
5622 onto the stack in reverse order, the new library version is likely to
5623 interoperate with programs that have been compiled with the old
5624 pem_password_cb definition (PEM_whatever takes some data that
5625 happens to be on the stack as its last argument, and the callback
5626 just ignores this garbage); but there is no guarantee whatsoever that
5629 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5630 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5631 problems not only on Windows, but also on some Unix platforms.
5632 To avoid problematic command lines, these definitions are now in an
5633 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5634 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5637 *) MIPS III/IV assembler module is reimplemented.
5640 *) More DES library cleanups: remove references to srand/rand and
5641 delete an unused file.
5644 *) Add support for the the free Netwide assembler (NASM) under Win32,
5645 since not many people have MASM (ml) and it can be hard to obtain.
5646 This is currently experimental but it seems to work OK and pass all
5647 the tests. Check out INSTALL.W32 for info.
5650 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5651 without temporary keys kept an extra copy of the server key,
5652 and connections with temporary keys did not free everything in case
5656 *) New function RSA_check_key and new openssl rsa option -check
5657 for verifying the consistency of RSA keys.
5658 [Ulf Moeller, Bodo Moeller]
5660 *) Various changes to make Win32 compile work:
5661 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5662 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5663 comparison" warnings.
5664 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5667 *) Add a debugging option to PKCS#5 v2 key generation function: when
5668 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5669 derived keys are printed to stderr.
5672 *) Copy the flags in ASN1_STRING_dup().
5673 [Roman E. Pavlov <pre@mo.msk.ru>]
5675 *) The x509 application mishandled signing requests containing DSA
5676 keys when the signing key was also DSA and the parameters didn't match.
5678 It was supposed to omit the parameters when they matched the signing key:
5679 the verifying software was then supposed to automatically use the CA's
5680 parameters if they were absent from the end user certificate.
5682 Omitting parameters is no longer recommended. The test was also
5683 the wrong way round! This was probably due to unusual behaviour in
5684 EVP_cmp_parameters() which returns 1 if the parameters match.
5685 This meant that parameters were omitted when they *didn't* match and
5686 the certificate was useless. Certificates signed with 'ca' didn't have
5688 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5690 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5691 The interface is as follows:
5692 Applications can use
5693 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5694 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5695 "off" is now the default.
5696 The library internally uses
5697 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5698 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5699 to disable memory-checking temporarily.
5701 Some inconsistent states that previously were possible (and were
5702 even the default) are now avoided.
5704 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5705 with each memory chunk allocated; this is occasionally more helpful
5706 than just having a counter.
5708 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5710 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5714 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5715 which largely parallels "options", but is for changing API behaviour,
5716 whereas "options" are about protocol behaviour.
5717 Initial "mode" flags are:
5719 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5720 a single record has been written.
5721 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5722 retries use the same buffer location.
5723 (But all of the contents must be
5727 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5730 *) Fix problems with no-hmac etc.
5731 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5733 *) New functions RSA_get_default_method(), RSA_set_method() and
5734 RSA_get_method(). These allows replacement of RSA_METHODs without having
5735 to mess around with the internals of an RSA structure.
5738 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5739 Also really enable memory leak checks in openssl.c and in some
5741 [Chad C. Mulligan, Bodo Moeller]
5743 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5744 up the length of negative integers. This has now been simplified to just
5745 store the length when it is first determined and use it later, rather
5746 than trying to keep track of where data is copied and updating it to
5748 [Steve Henson, reported by Brien Wheeler
5749 <bwheeler@authentica-security.com>]
5751 *) Add a new function PKCS7_signatureVerify. This allows the verification
5752 of a PKCS#7 signature but with the signing certificate passed to the
5753 function itself. This contrasts with PKCS7_dataVerify which assumes the
5754 certificate is present in the PKCS#7 structure. This isn't always the
5755 case: certificates can be omitted from a PKCS#7 structure and be
5756 distributed by "out of band" means (such as a certificate database).
5759 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5760 function prototypes in pem.h, also change util/mkdef.pl to add the
5761 necessary function names.
5764 *) mk1mf.pl (used by Windows builds) did not properly read the
5765 options set by Configure in the top level Makefile, and Configure
5766 was not even able to write more than one option correctly.
5767 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5770 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5771 file to be loaded from a BIO or FILE pointer. The BIO version will
5772 for example allow memory BIOs to contain config info.
5775 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5776 Whoever hopes to achieve shared-library compatibility across versions
5777 must use this, not the compile-time macro.
5778 (Exercise 0.9.4: Which is the minimum library version required by
5780 Note: All this applies only to multi-threaded programs, others don't
5784 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5785 through a BIO pair triggered the default case, i.e.
5786 SSLerr(...,SSL_R_UNKNOWN_STATE).
5789 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5790 can use the SSL library even if none of the specific BIOs is
5794 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5795 for the encoded length.
5796 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5798 *) Add initial documentation of the X509V3 functions.
5801 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5802 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5803 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5804 secure PKCS#8 private key format with a high iteration count.
5807 *) Fix determination of Perl interpreter: A perl or perl5
5808 _directory_ in $PATH was also accepted as the interpreter.
5809 [Ralf S. Engelschall]
5811 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5812 wrong with it but it was very old and did things like calling
5813 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5817 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5818 to use the new extension code.
5821 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5822 with macros. This should make it easier to change their form, add extra
5823 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5827 *) Add to configuration table a new entry that can specify an alternative
5828 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5829 according to Mark Crispin <MRC@Panda.COM>.
5833 *) DES CBC did not update the IV. Weird.
5836 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5837 Changing the behaviour of the former might break existing programs --
5838 where IV updating is needed, des_ncbc_encrypt can be used.
5841 *) When bntest is run from "make test" it drives bc to check its
5842 calculations, as well as internally checking them. If an internal check
5843 fails, it needs to cause bc to give a non-zero result or make test carries
5844 on without noticing the failure. Fixed.
5847 *) DES library cleanups.
5850 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5851 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5852 ciphers. NOTE: although the key derivation function has been verified
5853 against some published test vectors it has not been extensively tested
5854 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5858 *) Instead of "mkdir -p", which is not fully portable, use new
5859 Perl script "util/mkdir-p.pl".
5862 *) Rewrite the way password based encryption (PBE) is handled. It used to
5863 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5864 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5865 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5866 the 'parameter' field of the AlgorithmIdentifier is passed to the
5867 underlying key generation function so it must do its own ASN1 parsing.
5868 This has also changed the EVP_PBE_CipherInit() function which now has a
5869 'parameter' argument instead of literal salt and iteration count values
5870 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5873 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5874 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5875 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5876 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5877 value was just used as a "magic string" and not used directly its
5878 value doesn't matter.
5881 *) Introduce some semblance of const correctness to BN. Shame C doesn't
5885 *) "linux-sparc64" configuration (ultrapenguin).
5886 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
5887 "linux-sparc" configuration.
5888 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
5890 *) config now generates no-xxx options for missing ciphers.
5893 *) Support the EBCDIC character set (work in progress).
5894 File ebcdic.c not yet included because it has a different license.
5895 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5897 *) Support BS2000/OSD-POSIX.
5898 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5900 *) Make callbacks for key generation use void * instead of char *.
5903 *) Make S/MIME samples compile (not yet tested).
5906 *) Additional typesafe stacks.
5909 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
5913 Changes between 0.9.3 and 0.9.3a [29 May 1999]
5915 *) New configuration variant "sco5-gcc".
5917 *) Updated some demos.
5918 [Sean O Riordain, Wade Scholine]
5920 *) Add missing BIO_free at exit of pkcs12 application.
5923 *) Fix memory leak in conf.c.
5926 *) Updates for Win32 to assembler version of MD5.
5929 *) Set #! path to perl in apps/der_chop to where we found it
5930 instead of using a fixed path.
5933 *) SHA library changes for irix64-mips4-cc.
5936 *) Improvements for VMS support.
5940 Changes between 0.9.2b and 0.9.3 [24 May 1999]
5942 *) Bignum library bug fix. IRIX 6 passes "make test" now!
5943 This also avoids the problems with SC4.2 and unpatched SC5.
5944 [Andy Polyakov <appro@fy.chalmers.se>]
5946 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
5947 These are required because of the typesafe stack would otherwise break
5948 existing code. If old code used a structure member which used to be STACK
5949 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
5950 sk_num or sk_value it would produce an error because the num, data members
5951 are not present in STACK_OF. Now it just produces a warning. sk_set
5952 replaces the old method of assigning a value to sk_value
5953 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
5954 that does this will no longer work (and should use sk_set instead) but
5955 this could be regarded as a "questionable" behaviour anyway.
5958 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
5959 correctly handle encrypted S/MIME data.
5962 *) Change type of various DES function arguments from des_cblock
5963 (which means, in function argument declarations, pointer to char)
5964 to des_cblock * (meaning pointer to array with 8 char elements),
5965 which allows the compiler to do more typechecking; it was like
5966 that back in SSLeay, but with lots of ugly casts.
5968 Introduce new type const_des_cblock.
5971 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
5972 problems: find RecipientInfo structure that matches recipient certificate
5973 and initialise the ASN1 structures properly based on passed cipher.
5976 *) Belatedly make the BN tests actually check the results.
5979 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
5980 to and from BNs: it was completely broken. New compilation option
5981 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
5982 key elements as negative integers.
5985 *) Reorganize and speed up MD5.
5986 [Andy Polyakov <appro@fy.chalmers.se>]
5989 [Richard Levitte <richard@levitte.org>]
5991 *) New option -out to asn1parse to allow the parsed structure to be
5992 output to a file. This is most useful when combined with the -strparse
5993 option to examine the output of things like OCTET STRINGS.
5996 *) Make SSL library a little more fool-proof by not requiring any longer
5997 that SSL_set_{accept,connect}_state be called before
5998 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
5999 in many applications because usually everything *appeared* to work as
6000 intended anyway -- now it really works as intended).
6003 *) Move openssl.cnf out of lib/.
6006 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6007 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6008 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6009 [Ralf S. Engelschall]
6011 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6012 handle PKCS#7 enveloped data properly.
6013 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6015 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6016 copying pointers. The cert_st handling is changed by this in
6017 various ways (and thus what used to be known as ctx->default_cert
6018 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6019 any longer when s->cert does not give us what we need).
6020 ssl_cert_instantiate becomes obsolete by this change.
6021 As soon as we've got the new code right (possibly it already is?),
6022 we have solved a couple of bugs of the earlier code where s->cert
6023 was used as if it could not have been shared with other SSL structures.
6025 Note that using the SSL API in certain dirty ways now will result
6026 in different behaviour than observed with earlier library versions:
6027 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6028 does not influence s as it used to.
6030 In order to clean up things more thoroughly, inside SSL_SESSION
6031 we don't use CERT any longer, but a new structure SESS_CERT
6032 that holds per-session data (if available); currently, this is
6033 the peer's certificate chain and, for clients, the server's certificate
6034 and temporary key. CERT holds only those values that can have
6035 meaningful defaults in an SSL_CTX.
6038 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6039 from the internal representation. Various PKCS#7 fixes: remove some
6040 evil casts and set the enc_dig_alg field properly based on the signing
6044 *) Allow PKCS#12 password to be set from the command line or the
6045 environment. Let 'ca' get its config file name from the environment
6046 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6050 *) Allow certificate policies extension to use an IA5STRING for the
6051 organization field. This is contrary to the PKIX definition but
6052 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6056 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6057 without disallowing inline assembler and the like for non-pedantic builds.
6060 *) Support Borland C++ builder.
6061 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6066 *) SHA-1 cleanups and performance enhancements.
6067 [Andy Polyakov <appro@fy.chalmers.se>]
6069 *) Sparc v8plus assembler for the bignum library.
6070 [Andy Polyakov <appro@fy.chalmers.se>]
6072 *) Accept any -xxx and +xxx compiler options in Configure.
6075 *) Update HPUX configuration.
6078 *) Add missing sk_<type>_unshift() function to safestack.h
6079 [Ralf S. Engelschall]
6081 *) New function SSL_CTX_use_certificate_chain_file that sets the
6082 "extra_cert"s in addition to the certificate. (This makes sense
6083 only for "PEM" format files, as chains as a whole are not
6087 *) Support verify_depth from the SSL API.
6088 x509_vfy.c had what can be considered an off-by-one-error:
6089 Its depth (which was not part of the external interface)
6090 was actually counting the number of certificates in a chain;
6091 now it really counts the depth.
6094 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6095 instead of X509err, which often resulted in confusing error
6096 messages since the error codes are not globally unique
6097 (e.g. an alleged error in ssl3_accept when a certificate
6098 didn't match the private key).
6100 *) New function SSL_CTX_set_session_id_context that allows to set a default
6101 value (so that you don't need SSL_set_session_id_context for each
6102 connection using the SSL_CTX).
6105 *) OAEP decoding bug fix.
6108 *) Support INSTALL_PREFIX for package builders, as proposed by
6112 *) New Configure options "threads" and "no-threads". For systems
6113 where the proper compiler options are known (currently Solaris
6114 and Linux), "threads" is the default.
6117 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6120 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6121 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6122 such as /usr/local/bin.
6125 *) "make linux-shared" to build shared libraries.
6126 [Niels Poppe <niels@netbox.org>]
6128 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6131 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6132 extension adding in x509 utility.
6135 *) Remove NOPROTO sections and error code comments.
6138 *) Partial rewrite of the DEF file generator to now parse the ANSI
6142 *) New Configure options --prefix=DIR and --openssldir=DIR.
6145 *) Complete rewrite of the error code script(s). It is all now handled
6146 by one script at the top level which handles error code gathering,
6147 header rewriting and C source file generation. It should be much better
6148 than the old method: it now uses a modified version of Ulf's parser to
6149 read the ANSI prototypes in all header files (thus the old K&R definitions
6150 aren't needed for error creation any more) and do a better job of
6151 translating function codes into names. The old 'ASN1 error code imbedded
6152 in a comment' is no longer necessary and it doesn't use .err files which
6153 have now been deleted. Also the error code call doesn't have to appear all
6154 on one line (which resulted in some large lines...).
6157 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6160 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6161 0 (which usually indicates a closed connection), but continue reading.
6164 *) Fix some race conditions.
6167 *) Add support for CRL distribution points extension. Add Certificate
6168 Policies and CRL distribution points documentation.
6171 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6174 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6175 8 of keying material. Merlin has also confirmed interop with this fix
6176 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6177 [Merlin Hughes <merlin@baltimore.ie>]
6179 *) Fix lots of warnings.
6180 [Richard Levitte <levitte@stacken.kth.se>]
6182 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6183 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6184 [Richard Levitte <levitte@stacken.kth.se>]
6186 *) Fix problems with sizeof(long) == 8.
6187 [Andy Polyakov <appro@fy.chalmers.se>]
6189 *) Change functions to ANSI C.
6192 *) Fix typos in error codes.
6193 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6195 *) Remove defunct assembler files from Configure.
6198 *) SPARC v8 assembler BIGNUM implementation.
6199 [Andy Polyakov <appro@fy.chalmers.se>]
6201 *) Support for Certificate Policies extension: both print and set.
6202 Various additions to support the r2i method this uses.
6205 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6206 return a const string when you are expecting an allocated buffer.
6209 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6210 types DirectoryString and DisplayText.
6213 *) Add code to allow r2i extensions to access the configuration database,
6214 add an LHASH database driver and add several ctx helper functions.
6217 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6218 fail when they extended the size of a BIGNUM.
6221 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6222 support typesafe stack.
6225 *) Fix typo in SSL_[gs]et_options().
6226 [Nils Frostberg <nils@medcom.se>]
6228 *) Delete various functions and files that belonged to the (now obsolete)
6229 old X509V3 handling code.
6232 *) New Configure option "rsaref".
6235 *) Don't auto-generate pem.h.
6238 *) Introduce type-safe ASN.1 SETs.
6241 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6242 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6244 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6245 that links with OpenSSL (well at least cause lots of warnings), but fear
6246 not: the conversion is trivial, and it eliminates loads of evil casts. A
6247 few STACKed things have been converted already. Feel free to convert more.
6248 In the fullness of time, I'll do away with the STACK type altogether.
6251 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6252 specified in <certfile> by updating the entry in the index.txt file.
6253 This way one no longer has to edit the index.txt file manually for
6254 revoking a certificate. The -revoke option does the gory details now.
6255 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6257 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6258 `-text' option at all and this way the `-noout -text' combination was
6259 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6260 [Ralf S. Engelschall]
6262 *) Make sure a corresponding plain text error message exists for the
6263 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6264 verify callback function determined that a certificate was revoked.
6265 [Ralf S. Engelschall]
6267 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6268 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6269 all available cipers including rc5, which was forgotten until now.
6270 In order to let the testing shell script know which algorithms
6271 are available, a new (up to now undocumented) command
6272 "openssl list-cipher-commands" is used.
6275 *) Bugfix: s_client occasionally would sleep in select() when
6276 it should have checked SSL_pending() first.
6279 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6280 the raw DSA values prior to ASN.1 encoding.
6283 *) Tweaks to Configure
6284 [Niels Poppe <niels@netbox.org>]
6286 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6290 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6293 *) New config option to avoid instructions that are illegal on the 80386.
6294 The default code is faster, but requires at least a 486.
6297 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6298 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6299 same as SSL2_VERSION anyway.
6302 *) New "-showcerts" option for s_client.
6305 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6306 application. Various cleanups and fixes.
6309 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6310 modify error routines to work internally. Add error codes and PBE init
6311 to library startup routines.
6314 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6315 packing functions to asn1 and evp. Changed function names and error
6316 codes along the way.
6319 *) PKCS12 integration: and so it begins... First of several patches to
6320 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6321 objects to objects.h
6324 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6325 and display support for Thawte strong extranet extension.
6328 *) Add LinuxPPC support.
6329 [Jeff Dubrule <igor@pobox.org>]
6331 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6332 bn_div_words in alpha.s.
6333 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6335 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6336 OAEP isn't supported when OpenSSL is built with RSAref.
6337 [Ulf Moeller <ulf@fitug.de>]
6339 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6340 so they no longer are missing under -DNOPROTO.
6341 [Soren S. Jorvang <soren@t.dk>]
6344 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6346 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6347 doesn't work when the session is reused. Coming soon!
6350 *) Fix a security hole, that allows sessions to be reused in the wrong
6351 context thus bypassing client cert protection! All software that uses
6352 client certs and session caches in multiple contexts NEEDS PATCHING to
6353 allow session reuse! A fuller solution is in the works.
6354 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6356 *) Some more source tree cleanups (removed obsolete files
6357 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6358 permission on "config" script to be executable) and a fix for the INSTALL
6360 [Ulf Moeller <ulf@fitug.de>]
6362 *) Remove some legacy and erroneous uses of malloc, free instead of
6364 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6366 *) Make rsa_oaep_test return non-zero on error.
6367 [Ulf Moeller <ulf@fitug.de>]
6369 *) Add support for native Solaris shared libraries. Configure
6370 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6371 if someone would make that last step automatic.
6372 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6374 *) ctx_size was not built with the right compiler during "make links". Fixed.
6377 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6378 except NULL ciphers". This means the default cipher list will no longer
6379 enable NULL ciphers. They need to be specifically enabled e.g. with
6380 the string "DEFAULT:eNULL".
6383 *) Fix to RSA private encryption routines: if p < q then it would
6384 occasionally produce an invalid result. This will only happen with
6385 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6388 *) Be less restrictive and allow also `perl util/perlpath.pl
6389 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6390 because this way one can also use an interpreter named `perl5' (which is
6391 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6392 installed as `perl').
6393 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6395 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6396 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6398 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6399 advapi32.lib to Win32 build and change the pem test comparision
6400 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6401 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6402 and crypto/des/ede_cbcm_enc.c.
6405 *) DES quad checksum was broken on big-endian architectures. Fixed.
6408 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6409 Win32 test batch file so it (might) work again. The Win32 test batch file
6410 is horrible: I feel ill....
6413 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6414 in e_os.h. Audit of header files to check ANSI and non ANSI
6415 sections: 10 functions were absent from non ANSI section and not exported
6416 from Windows DLLs. Fixed up libeay.num for new functions.
6419 *) Make `openssl version' output lines consistent.
6420 [Ralf S. Engelschall]
6422 *) Fix Win32 symbol export lists for BIO functions: Added
6423 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6424 to ms/libeay{16,32}.def.
6425 [Ralf S. Engelschall]
6427 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6428 fine under Unix and passes some trivial tests I've now added. But the
6429 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6430 added to make sure no one expects that this stuff really works in the
6431 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6432 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6434 [Ralf S. Engelschall]
6436 *) Fix the generation of two part addresses in perl.
6437 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6439 *) Add config entry for Linux on MIPS.
6440 [John Tobey <jtobey@channel1.com>]
6442 *) Make links whenever Configure is run, unless we are on Windoze.
6445 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6446 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6450 *) Add a useful kludge to allow package maintainers to specify compiler and
6451 other platforms details on the command line without having to patch the
6452 Configure script everytime: One now can use ``perl Configure
6453 <id>:<details>'', i.e. platform ids are allowed to have details appended
6454 to them (seperated by colons). This is treated as there would be a static
6455 pre-configured entry in Configure's %table under key <id> with value
6456 <details> and ``perl Configure <id>'' is called. So, when you want to
6457 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6458 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6459 now, which overrides the FreeBSD-elf entry on-the-fly.
6460 [Ralf S. Engelschall]
6462 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6465 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6466 on the `perl Configure ...' command line. This way one can compile
6467 OpenSSL libraries with Position Independent Code (PIC) which is needed
6468 for linking it into DSOs.
6469 [Ralf S. Engelschall]
6471 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6475 *) Cleaned up the LICENSE document: The official contact for any license
6476 questions now is the OpenSSL core team under openssl-core@openssl.org.
6477 And add a paragraph about the dual-license situation to make sure people
6478 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6479 to the OpenSSL toolkit.
6480 [Ralf S. Engelschall]
6482 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6483 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6484 Additonally cleaned up the `make links' target: Remove unnecessary
6485 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6486 to speed processing and no longer clutter the display with confusing
6487 stuff. Instead only the actually done links are displayed.
6488 [Ralf S. Engelschall]
6490 *) Permit null encryption ciphersuites, used for authentication only. It used
6491 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6492 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6496 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6497 signed attributes when verifying signatures (this would break them),
6498 the detached data encoding was wrong and public keys obtained using
6499 X509_get_pubkey() weren't freed.
6502 *) Add text documentation for the BUFFER functions. Also added a work around
6503 to a Win95 console bug. This was triggered by the password read stuff: the
6504 last character typed gets carried over to the next fread(). If you were
6505 generating a new cert request using 'req' for example then the last
6506 character of the passphrase would be CR which would then enter the first
6510 *) Added the new `Includes OpenSSL Cryptography Software' button as
6511 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6512 button and can be used by applications based on OpenSSL to show the
6513 relationship to the OpenSSL project.
6514 [Ralf S. Engelschall]
6516 *) Remove confusing variables in function signatures in files
6517 ssl/ssl_lib.c and ssl/ssl.h.
6518 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6520 *) Don't install bss_file.c under PREFIX/include/
6521 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6523 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6524 functions that return function pointers and has support for NT specific
6525 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6526 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6527 unsigned to signed types: this was killing the Win32 compile.
6530 *) Add new certificate file to stack functions,
6531 SSL_add_dir_cert_subjects_to_stack() and
6532 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6533 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6534 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6535 This means that Apache-SSL and similar packages don't have to mess around
6536 to add as many CAs as they want to the preferred list.
6539 *) Experiment with doxygen documentation. Currently only partially applied to
6541 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6542 openssl.doxy as the configuration file.
6545 *) Get rid of remaining C++-style comments which strict C compilers hate.
6546 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6548 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6549 compiled in by default: it has problems with large keys.
6552 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6553 DH private keys and/or callback functions which directly correspond to
6554 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6555 is needed for applications which have to configure certificates on a
6556 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6558 For the RSA certificate situation is makes no difference, but
6559 for the DSA certificate situation this fixes the "no shared cipher"
6560 problem where the OpenSSL cipher selection procedure failed because the
6561 temporary keys were not overtaken from the context and the API provided
6562 no way to reconfigure them.
6563 The new functions now let applications reconfigure the stuff and they
6564 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6565 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6566 non-public-API function ssl_cert_instantiate() is used as a helper
6567 function and also to reduce code redundancy inside ssl_rsa.c.
6568 [Ralf S. Engelschall]
6570 *) Move s_server -dcert and -dkey options out of the undocumented feature
6571 area because they are useful for the DSA situation and should be
6572 recognized by the users.
6573 [Ralf S. Engelschall]
6575 *) Fix the cipher decision scheme for export ciphers: the export bits are
6576 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6577 SSL_EXP_MASK. So, the original variable has to be used instead of the
6578 already masked variable.
6579 [Richard Levitte <levitte@stacken.kth.se>]
6581 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6582 [Richard Levitte <levitte@stacken.kth.se>]
6584 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6585 from `int' to `unsigned int' because it's a length and initialized by
6586 EVP_DigestFinal() which expects an `unsigned int *'.
6587 [Richard Levitte <levitte@stacken.kth.se>]
6589 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6590 script. Instead use the usual Shell->Perl transition trick.
6591 [Ralf S. Engelschall]
6593 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6594 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6595 -noout -modulus' as it's already the case for `openssl rsa -noout
6596 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6597 currently the public key is printed (a decision which was already done by
6598 `openssl dsa -modulus' in the past) which serves a similar purpose.
6599 Additionally the NO_RSA no longer completely removes the whole -modulus
6600 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6602 [Ralf S. Engelschall]
6604 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6605 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6606 [Arne Ansper <arne@ats.cyber.ee>]
6608 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6609 to be added. Now both 'req' and 'ca' can use new objects defined in the
6613 *) Add cool BIO that does syslog (or event log on NT).
6614 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6616 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6617 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6618 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6619 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6622 *) Add preliminary config info for new extension code.
6625 *) Make RSA_NO_PADDING really use no padding.
6626 [Ulf Moeller <ulf@fitug.de>]
6628 *) Generate errors when private/public key check is done.
6631 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6632 for some CRL extensions and new objects added.
6635 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6636 key usage extension and fuller support for authority key id.
6639 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6640 padding method for RSA, which is recommended for new applications in PKCS
6641 #1 v2.0 (RFC 2437, October 1998).
6642 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6643 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6644 against Bleichbacher's attack on RSA.
6645 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6648 *) Updates to the new SSL compression code
6649 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6651 *) Fix so that the version number in the master secret, when passed
6652 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6653 (because the server will not accept higher), that the version number
6654 is 0x03,0x01, not 0x03,0x00
6655 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6657 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6658 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6659 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6662 *) Support for RAW extensions where an arbitrary extension can be
6663 created by including its DER encoding. See apps/openssl.cnf for
6667 *) Make sure latest Perl versions don't interpret some generated C array
6668 code as Perl array code in the crypto/err/err_genc.pl script.
6669 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6671 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6672 not many people have the assembler. Various Win32 compilation fixes and
6673 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6677 *) Modify configure script 'Configure' to automatically create crypto/date.h
6678 file under Win32 and also build pem.h from pem.org. New script
6679 util/mkfiles.pl to create the MINFO file on environments that can't do a
6680 'make files': perl util/mkfiles.pl >MINFO should work.
6683 *) Major rework of DES function declarations, in the pursuit of correctness
6684 and purity. As a result, many evil casts evaporated, and some weirdness,
6685 too. You may find this causes warnings in your code. Zapping your evil
6686 casts will probably fix them. Mostly.
6689 *) Fix for a typo in asn1.h. Bug fix to object creation script
6690 obj_dat.pl. It considered a zero in an object definition to mean
6691 "end of object": none of the objects in objects.h have any zeros
6692 so it wasn't spotted.
6693 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6695 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6696 Masking (CBCM). In the absence of test vectors, the best I have been able
6697 to do is check that the decrypt undoes the encrypt, so far. Send me test
6698 vectors if you have them.
6701 *) Correct calculation of key length for export ciphers (too much space was
6702 allocated for null ciphers). This has not been tested!
6705 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6706 message is now correct (it understands "crypto" and "ssl" on its
6707 command line). There is also now an "update" option. This will update
6708 the util/ssleay.num and util/libeay.num files with any new functions.
6710 perl util/mkdef.pl crypto ssl update
6711 it will update them.
6714 *) Overhauled the Perl interface (perl/*):
6715 - ported BN stuff to OpenSSL's different BN library
6716 - made the perl/ source tree CVS-aware
6717 - renamed the package from SSLeay to OpenSSL (the files still contain
6718 their history because I've copied them in the repository)
6719 - removed obsolete files (the test scripts will be replaced
6720 by better Test::Harness variants in the future)
6721 [Ralf S. Engelschall]
6723 *) First cut for a very conservative source tree cleanup:
6724 1. merge various obsolete readme texts into doc/ssleay.txt
6725 where we collect the old documents and readme texts.
6726 2. remove the first part of files where I'm already sure that we no
6727 longer need them because of three reasons: either they are just temporary
6728 files which were left by Eric or they are preserved original files where
6729 I've verified that the diff is also available in the CVS via "cvs diff
6730 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6731 the crypto/md/ stuff).
6732 [Ralf S. Engelschall]
6734 *) More extension code. Incomplete support for subject and issuer alt
6735 name, issuer and authority key id. Change the i2v function parameters
6736 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6737 what that's for :-) Fix to ASN1 macro which messed up
6738 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6741 *) Preliminary support for ENUMERATED type. This is largely copied from the
6745 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6746 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6748 *) Make sure `make rehash' target really finds the `openssl' program.
6749 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6751 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6752 like to hear about it if this slows down other processors.
6755 *) Add CygWin32 platform information to Configure script.
6756 [Alan Batie <batie@aahz.jf.intel.com>]
6758 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6759 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6761 *) New program nseq to manipulate netscape certificate sequences
6764 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6768 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6769 but the BN code had some problems that would cause failures when
6770 doing certificate verification and some other functions.
6771 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6773 *) Add ASN1 and PEM code to support netscape certificate sequences.
6776 *) Add ASN1 and PEM code to support netscape certificate sequences.
6779 *) Add several PKIX and private extended key usage OIDs.
6782 *) Modify the 'ca' program to handle the new extension code. Modify
6783 openssl.cnf for new extension format, add comments.
6786 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6787 and add a sample to openssl.cnf so req -x509 now adds appropriate
6791 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6792 error code, add initial support to X509_print() and x509 application.
6795 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6796 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6797 stuff is currently isolated and isn't even compiled yet.
6800 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6801 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6802 Removed the versions check from X509 routines when loading extensions:
6803 this allows certain broken certificates that don't set the version
6804 properly to be processed.
6807 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6808 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6809 can still be regenerated with "make depend".
6812 *) Spelling mistake in C version of CAST-128.
6813 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6815 *) Changes to the error generation code. The perl script err-code.pl
6816 now reads in the old error codes and retains the old numbers, only
6817 adding new ones if necessary. It also only changes the .err files if new
6818 codes are added. The makefiles have been modified to only insert errors
6819 when needed (to avoid needlessly modifying header files). This is done
6820 by only inserting errors if the .err file is newer than the auto generated
6821 C file. To rebuild all the error codes from scratch (the old behaviour)
6822 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6823 or delete all the .err files.
6826 *) CAST-128 was incorrectly implemented for short keys. The C version has
6827 been fixed, but is untested. The assembler versions are also fixed, but
6828 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6829 to regenerate it if needed.
6830 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6831 Hagino <itojun@kame.net>]
6833 *) File was opened incorrectly in randfile.c.
6834 [Ulf Möller <ulf@fitug.de>]
6836 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6837 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6838 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6839 al: it's just almost always a UTCTime. Note this patch adds new error
6840 codes so do a "make errors" if there are problems.
6843 *) Correct Linux 1 recognition in config.
6844 [Ulf Möller <ulf@fitug.de>]
6846 *) Remove pointless MD5 hash when using DSA keys in ca.
6847 [Anonymous <nobody@replay.com>]
6849 *) Generate an error if given an empty string as a cert directory. Also
6850 generate an error if handed NULL (previously returned 0 to indicate an
6851 error, but didn't set one).
6852 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6854 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6857 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6858 parameters. This was causing a warning which killed off the Win32 compile.
6861 *) Remove C++ style comments from crypto/bn/bn_local.h.
6862 [Neil Costigan <neil.costigan@celocom.com>]
6864 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6865 based on a text string, looking up short and long names and finally
6866 "dot" format. The "dot" format stuff didn't work. Added new function
6867 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6868 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6869 OID is not part of the table.
6872 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6873 X509_LOOKUP_by_alias().
6876 *) Sort openssl functions by name.
6879 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6880 encryption from sample DSA keys (in case anyone is interested the password
6884 *) Make _all_ *_free functions accept a NULL pointer.
6885 [Frans Heymans <fheymans@isaserver.be>]
6887 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
6889 [Anonymous <nobody@replay.com>]
6891 *) s_server should send the CAfile as acceptable CAs, not its own cert.
6892 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6894 *) Don't blow it for numeric -newkey arguments to apps/req.
6895 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
6897 *) Temp key "for export" tests were wrong in s3_srvr.c.
6898 [Anonymous <nobody@replay.com>]
6900 *) Add prototype for temp key callback functions
6901 SSL_CTX_set_tmp_{rsa,dh}_callback().
6904 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
6905 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
6908 *) X509_name_add_entry() freed the wrong thing after an error.
6909 [Arne Ansper <arne@ats.cyber.ee>]
6911 *) rsa_eay.c would attempt to free a NULL context.
6912 [Arne Ansper <arne@ats.cyber.ee>]
6914 *) BIO_s_socket() had a broken should_retry() on Windoze.
6915 [Arne Ansper <arne@ats.cyber.ee>]
6917 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
6918 [Arne Ansper <arne@ats.cyber.ee>]
6920 *) Make sure the already existing X509_STORE->depth variable is initialized
6921 in X509_STORE_new(), but document the fact that this variable is still
6922 unused in the certificate verification process.
6923 [Ralf S. Engelschall]
6925 *) Fix the various library and apps files to free up pkeys obtained from
6926 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
6929 *) Fix reference counting in X509_PUBKEY_get(). This makes
6930 demos/maurice/example2.c work, amongst others, probably.
6931 [Steve Henson and Ben Laurie]
6933 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
6934 `openssl' and second, the shortcut symlinks for the `openssl <command>'
6935 are no longer created. This way we have a single and consistent command
6936 line interface `openssl <command>', similar to `cvs <command>'.
6937 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
6939 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
6940 BIT STRING wrapper always have zero unused bits.
6943 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
6946 *) Make the top-level INSTALL documentation easier to understand.
6949 *) Makefiles updated to exit if an error occurs in a sub-directory
6950 make (including if user presses ^C) [Paul Sutton]
6952 *) Make Montgomery context stuff explicit in RSA data structure.
6955 *) Fix build order of pem and err to allow for generated pem.h.
6958 *) Fix renumbering bug in X509_NAME_delete_entry().
6961 *) Enhanced the err-ins.pl script so it makes the error library number
6962 global and can add a library name. This is needed for external ASN1 and
6963 other error libraries.
6966 *) Fixed sk_insert which never worked properly.
6969 *) Fix ASN1 macros so they can handle indefinite length construted
6970 EXPLICIT tags. Some non standard certificates use these: they can now
6974 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
6975 into a single doc/ssleay.txt bundle. This way the information is still
6976 preserved but no longer messes up this directory. Now it's new room for
6977 the new set of documenation files.
6978 [Ralf S. Engelschall]
6980 *) SETs were incorrectly DER encoded. This was a major pain, because they
6981 shared code with SEQUENCEs, which aren't coded the same. This means that
6982 almost everything to do with SETs or SEQUENCEs has either changed name or
6983 number of arguments.
6984 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
6986 *) Fix test data to work with the above.
6989 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
6990 was already fixed by Eric for 0.9.1 it seems.
6991 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
6993 *) Autodetect FreeBSD3.
6996 *) Fix various bugs in Configure. This affects the following platforms:
7000 unixware-2.0-pentium
7004 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7005 before they are needed.
7008 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7012 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7014 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7015 changed SSLeay to OpenSSL in version strings.
7016 [Ralf S. Engelschall]
7018 *) Some fixups to the top-level documents.
7021 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7022 because the symlink to include/ was missing.
7023 [Ralf S. Engelschall]
7025 *) Incorporated the popular no-RSA/DSA-only patches
7026 which allow to compile a RSA-free SSLeay.
7027 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7029 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7030 when "ssleay" is still not found.
7031 [Ralf S. Engelschall]
7033 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7034 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7036 *) Updated the README file.
7037 [Ralf S. Engelschall]
7039 *) Added various .cvsignore files in the CVS repository subdirs
7040 to make a "cvs update" really silent.
7041 [Ralf S. Engelschall]
7043 *) Recompiled the error-definition header files and added
7044 missing symbols to the Win32 linker tables.
7045 [Ralf S. Engelschall]
7047 *) Cleaned up the top-level documents;
7048 o new files: CHANGES and LICENSE
7049 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7050 o merged COPYRIGHT into LICENSE
7051 o removed obsolete TODO file
7052 o renamed MICROSOFT to INSTALL.W32
7053 [Ralf S. Engelschall]
7055 *) Removed dummy files from the 0.9.1b source tree:
7056 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7057 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7058 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7059 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7060 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7061 [Ralf S. Engelschall]
7063 *) Added various platform portability fixes.
7066 *) The Genesis of the OpenSSL rpject:
7067 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7068 Young and Tim J. Hudson created while they were working for C2Net until
7070 [The OpenSSL Project]
7073 Changes between 0.9.0b and 0.9.1b [not released]
7075 *) Updated a few CA certificates under certs/
7078 *) Changed some BIGNUM api stuff.
7081 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7082 DGUX x86, Linux Alpha, etc.
7085 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7086 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7090 *) Add -strparse option to asn1pars program which parses nested
7092 [Dr Stephen Henson <shenson@bigfoot.com>]
7094 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7097 *) DSA fix for "ca" program.
7100 *) Added "-genkey" option to "dsaparam" program.
7103 *) Added RIPE MD160 (rmd160) message digest.
7106 *) Added -a (all) option to "ssleay version" command.
7109 *) Added PLATFORM define which is the id given to Configure.
7112 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7115 *) Extended the ASN.1 parser routines.
7118 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7121 *) Added a BN_CTX to the BN library.
7124 *) Fixed the weak key values in DES library
7127 *) Changed API in EVP library for cipher aliases.
7130 *) Added support for RC2/64bit cipher.
7133 *) Converted the lhash library to the crypto/mem.c functions.
7136 *) Added more recognized ASN.1 object ids.
7139 *) Added more RSA padding checks for SSL/TLS.
7142 *) Added BIO proxy/filter functionality.
7145 *) Added extra_certs to SSL_CTX which can be used
7146 send extra CA certificates to the client in the CA cert chain sending
7147 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7150 *) Now Fortezza is denied in the authentication phase because
7151 this is key exchange mechanism is not supported by SSLeay at all.
7154 *) Additional PKCS1 checks.
7157 *) Support the string "TLSv1" for all TLS v1 ciphers.
7160 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7161 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7164 *) Fixed a few memory leaks.
7167 *) Fixed various code and comment typos.
7170 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7171 bytes sent in the client random.
7172 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob