5 Changes between 0.9.7e and 0.9.8 [xx XXX xxxx]
7 *) New arguments -certform, -keyform and -pass for s_client and s_server
8 to allow alternative format key and certificate files and passphrase
12 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
13 update associated structures and add various utility functions.
15 Add new policy related verify parameters, include policy checking in
16 standard verify code. Enhance 'smime' application with extra parameters
17 to support policy checking and print out.
20 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
21 Nehemiah processors. These extensions support AES encryption in hardware
22 as well as RNG (though RNG support is currently disabled).
23 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
25 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
28 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
29 [Andy Polyakov and a number of other people]
31 *) Improved PowerPC platform support. Most notably BIGNUM assembler
32 implementation contributed by IBM.
33 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
35 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
36 exponent rather than 'unsigned long'. There is a corresponding change to
37 the new 'rsa_keygen' element of the RSA_METHOD structure.
38 [Jelte Jansen, Geoff Thorpe]
40 *) Functionality for creating the initial serial number file is now
41 moved from CA.pl to the 'ca' utility with a new option -create_serial.
43 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
44 number file to 1, which is bound to cause problems. To avoid
45 the problems while respecting compatibility between different 0.9.7
46 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
47 CA.pl for serial number initialization. With the new release 0.9.8,
48 we can fix the problem directly in the 'ca' utility.)
51 *) Reduced header interdepencies by declaring more opaque objects in
52 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
53 give fewer recursive includes, which could break lazy source code - so
54 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
55 developers should define this symbol when building and using openssl to
56 ensure they track the recommended behaviour, interfaces, [etc], but
57 backwards-compatible behaviour prevails when this isn't defined.
60 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
63 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
64 This will generate a random key of the appropriate length based on the
65 cipher context. The EVP_CIPHER can provide its own random key generation
66 routine to support keys of a specific form. This is used in the des and
67 3des routines to generate a key of the correct parity. Update S/MIME
68 code to use new functions and hence generate correct parity DES keys.
69 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
70 valid (weak or incorrect parity).
73 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
74 as looking them up. This is useful when the verified structure may contain
75 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
76 present unless the new PKCS7_NO_CRL flag is asserted.
79 *) Extend ASN1 oid configuration module. It now additionally accepts the
82 shortName = some long name, 1.2.3.4
85 *) Reimplemented the BN_CTX implementation. There is now no more static
86 limitation on the number of variables it can handle nor the depth of the
87 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
88 information can now expand as required, and rather than having a single
89 static array of bignums, BN_CTX now uses a linked-list of such arrays
90 allowing it to expand on demand whilst maintaining the usefulness of
94 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
95 to allow all RSA operations to function using a single BN_CTX.
98 *) Preliminary support for certificate policy evaluation and checking. This
99 is initially intended to pass the tests outlined in "Conformance Testing
100 of Relying Party Client Certificate Path Processing Logic" v1.07.
103 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
104 remained unused and not that useful. A variety of other little bignum
105 tweaks and fixes have also been made continuing on from the audit (see
109 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
110 associated ASN1, EVP and SSL functions and old ASN1 macros.
113 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
114 and this should never fail. So the return value from the use of
115 BN_set_word() (which can fail due to needless expansion) is now deprecated;
116 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
119 *) BN_CTX_get() should return zero-valued bignums, providing the same
120 initialised value as BN_new().
121 [Geoff Thorpe, suggested by Ulf Möller]
123 *) Support for inhibitAnyPolicy certificate extension.
126 *) An audit of the BIGNUM code is underway, for which debugging code is
127 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
128 is considered valid when processing BIGNUMs, and causes execution to
129 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
130 further steps are taken to deliberately pollute unused data in BIGNUM
131 structures to try and expose faulty code further on. For now, openssl will
132 (in its default mode of operation) continue to tolerate the inconsistent
133 forms that it has tolerated in the past, but authors and packagers should
134 consider trying openssl and their own applications when compiled with
135 these debugging symbols defined. It will help highlight potential bugs in
136 their own code, and will improve the test coverage for OpenSSL itself. At
137 some point, these tighter rules will become openssl's default to improve
138 maintainability, though the assert()s and other overheads will remain only
139 in debugging configurations. See bn.h for more details.
140 [Geoff Thorpe, Nils Larsch, Ulf Möller]
142 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
143 that can only be obtained through BN_CTX_new() (which implicitly
144 initialises it). The presence of this function only made it possible
145 to overwrite an existing structure (and cause memory leaks).
148 *) Because of the callback-based approach for implementing LHASH as a
149 template type, lh_insert() adds opaque objects to hash-tables and
150 lh_doall() or lh_doall_arg() are typically used with a destructor callback
151 to clean up those corresponding objects before destroying the hash table
152 (and losing the object pointers). So some over-zealous constifications in
153 LHASH have been relaxed so that lh_insert() does not take (nor store) the
154 objects as "const" and the lh_doall[_arg] callback wrappers are not
155 prototyped to have "const" restrictions on the object pointers they are
156 given (and so aren't required to cast them away any more).
159 *) The tmdiff.h API was so ugly and minimal that our own timing utility
160 (speed) prefers to use its own implementation. The two implementations
161 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
162 its object type properly exposed (MS_TM) instead of casting to/from "char
163 *". This may still change yet if someone realises MS_TM and "ms_time_***"
164 aren't necessarily the greatest nomenclatures - but this is what was used
165 internally to the implementation so I've used that for now.
168 *) Ensure that deprecated functions do not get compiled when
169 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
170 the self-tests were still using deprecated key-generation functions so
171 these have been updated also.
174 *) Reorganise PKCS#7 code to separate the digest location functionality
175 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
176 New function PKCS7_set_digest() to set the digest type for PKCS#7
177 digestedData type. Add additional code to correctly generate the
178 digestedData type and add support for this type in PKCS7 initialization
182 *) New function PKCS7_set0_type_other() this initializes a PKCS7
183 structure of type "other".
186 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
187 sure the loop does correctly stop and breaking ("division by zero")
188 modulus operations are not performed. The (pre-generated) prime
189 table crypto/bn/bn_prime.h was already correct, but it could not be
190 re-generated on some platforms because of the "division by zero"
191 situation in the script.
192 [Ralf S. Engelschall]
194 *) Update support for ECC-based TLS ciphersuites according to
195 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
196 SHA-1 now is only used for "small" curves (where the
197 representation of a field element takes up to 24 bytes); for
198 larger curves, the field element resulting from ECDH is directly
199 used as premaster secret.
200 [Douglas Stebila (Sun Microsystems Laboratories)]
202 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
203 curve secp160r1 to the tests.
204 [Douglas Stebila (Sun Microsystems Laboratories)]
206 *) Add the possibility to load symbols globally with DSO.
207 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
209 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
210 control of the error stack.
213 *) Add support for STORE in ENGINE.
216 *) Add the STORE type. The intention is to provide a common interface
217 to certificate and key stores, be they simple file-based stores, or
218 HSM-type store, or LDAP stores, or...
219 NOTE: The code is currently UNTESTED and isn't really used anywhere.
222 *) Add a generic structure called OPENSSL_ITEM. This can be used to
223 pass a list of arguments to any function as well as provide a way
224 for a function to pass data back to the caller.
227 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
228 works like BUF_strdup() but can be used to duplicate a portion of
229 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
233 *) Add the function sk_find_ex() which works like sk_find(), but will
234 return an index to an element even if an exact match couldn't be
235 found. The index is guaranteed to point at the element where the
236 searched-for key would be inserted to preserve sorting order.
239 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
240 takes an extra flags argument for optional functionality. Currently,
241 the following flags are defined:
243 OBJ_BSEARCH_VALUE_ON_NOMATCH
244 This one gets OBJ_bsearch_ex() to return a pointer to the first
245 element where the comparing function returns a negative or zero
248 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
249 This one gets OBJ_bsearch_ex() to return a pointer to the first
250 element where the comparing function returns zero. This is useful
251 if there are more than one element where the comparing function
255 *) Make it possible to create self-signed certificates with 'openssl ca'
256 in such a way that the self-signed certificate becomes part of the
257 CA database and uses the same mechanisms for serial number generation
258 as all other certificate signing. The new flag '-selfsign' enables
259 this functionality. Adapt CA.sh and CA.pl.in.
262 *) Add functionality to check the public key of a certificate request
263 against a given private. This is useful to check that a certificate
264 request can be signed by that key (self-signing).
267 *) Make it possible to have multiple active certificates with the same
268 subject in the CA index file. This is done only if the keyword
269 'unique_subject' is set to 'no' in the main CA section (default
270 if 'CA_default') of the configuration file. The value is saved
271 with the database itself in a separate index attribute file,
272 named like the index file with '.attr' appended to the name.
275 *) Generate muti valued AVAs using '+' notation in config files for
279 *) Support for nameConstraints certificate extension.
282 *) Support for policyConstraints certificate extension.
285 *) Support for policyMappings certificate extension.
288 *) Make sure the default DSA_METHOD implementation only uses its
289 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
290 and change its own handlers to be NULL so as to remove unnecessary
291 indirection. This lets alternative implementations fallback to the
292 default implementation more easily.
295 *) Support for directoryName in GeneralName related extensions
299 *) Make it possible to link applications using Makefile.shared.
300 Make that possible even when linking against static libraries!
303 *) Support for single pass processing for S/MIME signing. This now
304 means that S/MIME signing can be done from a pipe, in addition
305 cleartext signing (multipart/signed type) is effectively streaming
306 and the signed data does not need to be all held in memory.
308 This is done with a new flag PKCS7_STREAM. When this flag is set
309 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
310 is done after the data is output (and digests calculated) in
314 *) Add full support for -rpath/-R, both in shared libraries and
315 applications, at least on the platforms where it's known how
319 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
320 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
321 will now compute a table of multiples of the generator that
322 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
323 faster (notably in the case of a single point multiplication,
325 [Nils Larsch, Bodo Moeller]
327 *) IPv6 support for certificate extensions. The various extensions
328 which use the IP:a.b.c.d can now take IPv6 addresses using the
329 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
333 *) Added an ENGINE that implements RSA by performing private key
334 exponentiations with the GMP library. The conversions to and from
335 GMP's mpz_t format aren't optimised nor are any montgomery forms
336 cached, and on x86 it appears OpenSSL's own performance has caught up.
337 However there are likely to be other architectures where GMP could
338 provide a boost. This ENGINE is not built in by default, but it can be
339 specified at Configure time and should be accompanied by the necessary
340 linker additions, eg;
341 ./config -DOPENSSL_USE_GMP -lgmp
344 *) "openssl engine" will not display ENGINE/DSO load failure errors when
345 testing availability of engines with "-t" - the old behaviour is
346 produced by increasing the feature's verbosity with "-tt".
349 *) ECDSA routines: under certain error conditions uninitialized BN objects
350 could be freed. Solution: make sure initialization is performed early
351 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
355 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
356 and DH_METHOD (eg. by ENGINE implementations) to override the normal
357 software implementations. For DSA and DH, parameter generation can
358 also be overriden by providing the appropriate method callbacks.
361 *) Change the "progress" mechanism used in key-generation and
362 primality testing to functions that take a new BN_GENCB pointer in
363 place of callback/argument pairs. The new API functions have "_ex"
364 postfixes and the older functions are reimplemented as wrappers for
365 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
366 declarations of the old functions to help (graceful) attempts to
367 migrate to the new functions. Also, the new key-generation API
368 functions operate on a caller-supplied key-structure and return
369 success/failure rather than returning a key or NULL - this is to
370 help make "keygen" another member function of RSA_METHOD etc.
372 Example for using the new callback interface:
374 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
378 BN_GENCB_set(&my_cb, my_callback, my_arg);
380 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
381 /* For the meaning of a, b in calls to my_callback(), see the
382 * documentation of the function that calls the callback.
383 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
384 * my_callback should return 1 if it wants BN_is_prime_ex()
385 * to continue, or 0 to stop.
390 *) Change the ZLIB compression method to be stateful, and make it
391 available to TLS with the number defined in
392 draft-ietf-tls-compression-04.txt.
395 *) Add the ASN.1 structures and functions for CertificatePair, which
396 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
398 CertificatePair ::= SEQUENCE {
399 forward [0] Certificate OPTIONAL,
400 reverse [1] Certificate OPTIONAL,
401 -- at least one of the pair shall be present -- }
403 Also implement the PEM functions to read and write certificate
404 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
406 This needed to be defined, mostly for the sake of the LDAP
407 attribute crossCertificatePair, but may prove useful elsewhere as
411 *) Make it possible to inhibit symlinking of shared libraries in
412 Makefile.shared, for Cygwin's sake.
415 *) Extend the BIGNUM API by creating new macros that behave like
418 void BN_set_sign(BIGNUM *a, int neg);
419 int BN_get_sign(const BIGNUM *a);
421 and avoid the need to access 'a->neg' directly in applications.
422 [Nils Larsch <nla@trustcenter.de>]
424 *) Implement fast modular reduction for pseudo-Mersenne primes
425 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
426 EC_GROUP_new_curve_GFp() will now automatically use this
428 [Nils Larsch <nla@trustcenter.de>]
430 *) Add new lock type (CRYPTO_LOCK_BN).
433 *) Change the ENGINE framework to automatically load engines
434 dynamically from specific directories unless they could be
435 found to already be built in or loaded. Move all the
436 current engines except for the cryptodev one to a new
438 The engines in engines/ are built as shared libraries if
439 the "shared" options was given to ./Configure or ./config.
440 Otherwise, they are inserted in libcrypto.a.
441 /usr/local/ssl/engines is the default directory for dynamic
442 engines, but that can be overriden at configure time through
443 the usual use of --prefix and/or --openssldir, and at run
444 time with the environment variable OPENSSL_ENGINES.
445 [Geoff Thorpe and Richard Levitte]
447 *) Add Makefile.shared, a helper makefile to build shared
448 libraries. Addapt Makefile.org.
451 *) Add version info to Win32 DLLs.
452 [Peter 'Luna' Runestig" <peter@runestig.com>]
454 *) Add new 'medium level' PKCS#12 API. Certificates and keys
455 can be added using this API to created arbitrary PKCS#12
456 files while avoiding the low level API.
458 New options to PKCS12_create(), key or cert can be NULL and
459 will then be omitted from the output file. The encryption
460 algorithm NIDs can be set to -1 for no encryption, the mac
461 iteration count can be set to 0 to omit the mac.
463 Enhance pkcs12 utility by making the -nokeys and -nocerts
464 options work when creating a PKCS#12 file. New option -nomac
465 to omit the mac, NONE can be set for an encryption algorithm.
466 New code is modified to use the enhanced PKCS12_create()
467 instead of the low level API.
470 *) Extend ASN1 encoder to support indefinite length constructed
471 encoding. This can output sequences tags and octet strings in
472 this form. Modify pk7_asn1.c to support indefinite length
473 encoding. This is experimental and needs additional code to
474 be useful, such as an ASN1 bio and some enhanced streaming
477 Extend template encode functionality so that tagging is passed
478 down to the template encoder.
481 *) Let 'openssl req' fail if an argument to '-newkey' is not
482 recognized instead of using RSA as a default.
485 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
486 As these are not official, they are not included in "ALL";
487 the "ECCdraft" ciphersuite group alias can be used to select them.
488 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
490 *) Add ECDH engine support.
491 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
493 *) Add ECDH in new directory crypto/ecdh/.
494 [Douglas Stebila (Sun Microsystems Laboratories)]
496 *) Let BN_rand_range() abort with an error after 100 iterations
497 without success (which indicates a broken PRNG).
500 *) Change BN_mod_sqrt() so that it verifies that the input value
501 is really the square of the return value. (Previously,
502 BN_mod_sqrt would show GIGO behaviour.)
505 *) Add named elliptic curves over binary fields from X9.62, SECG,
506 and WAP/WTLS; add OIDs that were still missing.
508 [Sheueling Chang Shantz and Douglas Stebila
509 (Sun Microsystems Laboratories)]
511 *) Extend the EC library for elliptic curves over binary fields
512 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
515 EC_GF2m_simple_method
519 EC_GROUP_new_curve_GF2m
520 EC_GROUP_set_curve_GF2m
521 EC_GROUP_get_curve_GF2m
522 EC_POINT_set_affine_coordinates_GF2m
523 EC_POINT_get_affine_coordinates_GF2m
524 EC_POINT_set_compressed_coordinates_GF2m
526 Point compression for binary fields is disabled by default for
527 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
530 As binary polynomials are represented as BIGNUMs, various members
531 of the EC_GROUP and EC_POINT data structures can be shared
532 between the implementations for prime fields and binary fields;
533 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
534 are essentially identical to their ..._GFp counterparts.
535 (For simplicity, the '..._GFp' prefix has been dropped from
536 various internal method names.)
538 An internal 'field_div' method (similar to 'field_mul' and
539 'field_sqr') has been added; this is used only for binary fields.
541 [Sheueling Chang Shantz and Douglas Stebila
542 (Sun Microsystems Laboratories)]
544 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
545 through methods ('mul', 'precompute_mult').
547 The generic implementations (now internally called 'ec_wNAF_mul'
548 and 'ec_wNAF_precomputed_mult') remain the default if these
549 methods are undefined.
551 [Sheueling Chang Shantz and Douglas Stebila
552 (Sun Microsystems Laboratories)]
554 *) New function EC_GROUP_get_degree, which is defined through
555 EC_METHOD. For curves over prime fields, this returns the bit
556 length of the modulus.
558 [Sheueling Chang Shantz and Douglas Stebila
559 (Sun Microsystems Laboratories)]
561 *) New functions EC_GROUP_dup, EC_POINT_dup.
562 (These simply call ..._new and ..._copy).
564 [Sheueling Chang Shantz and Douglas Stebila
565 (Sun Microsystems Laboratories)]
567 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
568 Polynomials are represented as BIGNUMs (where the sign bit is not
569 used) in the following functions [macros]:
572 BN_GF2m_sub [= BN_GF2m_add]
573 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
574 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
575 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
577 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
578 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
579 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
580 BN_GF2m_cmp [= BN_ucmp]
582 (Note that only the 'mod' functions are actually for fields GF(2^m).
583 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
585 For some functions, an the irreducible polynomial defining a
586 field can be given as an 'unsigned int[]' with strictly
587 decreasing elements giving the indices of those bits that are set;
588 i.e., p[] represents the polynomial
589 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
591 p[0] > p[1] > ... > p[k] = 0.
592 This applies to the following functions:
597 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
598 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
601 BN_GF2m_mod_solve_quad_arr
605 Conversion can be performed by the following functions:
610 bntest.c has additional tests for binary polynomial arithmetic.
612 Two implementations for BN_GF2m_mod_div() are available.
613 The default algorithm simply uses BN_GF2m_mod_inv() and
614 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
615 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
616 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
618 [Sheueling Chang Shantz and Douglas Stebila
619 (Sun Microsystems Laboratories)]
621 *) Add new error code 'ERR_R_DISABLED' that can be used when some
622 functionality is disabled at compile-time.
623 [Douglas Stebila <douglas.stebila@sun.com>]
625 *) Change default behaviour of 'openssl asn1parse' so that more
626 information is visible when viewing, e.g., a certificate:
628 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
629 mode the content of non-printable OCTET STRINGs is output in a
630 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
631 avoid the appearance of a printable string.
632 [Nils Larsch <nla@trustcenter.de>]
634 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
636 EC_GROUP_set_asn1_flag()
637 EC_GROUP_get_asn1_flag()
638 EC_GROUP_set_point_conversion_form()
639 EC_GROUP_get_point_conversion_form()
640 These control ASN1 encoding details:
641 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
642 has been set to OPENSSL_EC_NAMED_CURVE.
643 - Points are encoded in uncompressed form by default; options for
644 asn1_for are as for point2oct, namely
645 POINT_CONVERSION_COMPRESSED
646 POINT_CONVERSION_UNCOMPRESSED
647 POINT_CONVERSION_HYBRID
649 Also add 'seed' and 'seed_len' members to EC_GROUP with access
653 EC_GROUP_get_seed_len()
654 This is used only for ASN1 purposes (so far).
655 [Nils Larsch <nla@trustcenter.de>]
657 *) Add 'field_type' member to EC_METHOD, which holds the NID
658 of the appropriate field type OID. The new function
659 EC_METHOD_get_field_type() returns this value.
660 [Nils Larsch <nla@trustcenter.de>]
667 providing useful interfaces to EC_POINT_point2oct() and
668 EC_POINT_oct2point().
669 [Nils Larsch <nla@trustcenter.de>]
671 *) Change internals of the EC library so that the functions
672 EC_GROUP_set_generator()
673 EC_GROUP_get_generator()
675 EC_GROUP_get_cofactor()
676 are implemented directly in crypto/ec/ec_lib.c and not dispatched
677 to methods, which would lead to unnecessary code duplication when
678 adding different types of curves.
679 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
681 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
682 arithmetic, and such that modified wNAFs are generated
683 (which avoid length expansion in many cases).
686 *) Add a function EC_GROUP_check_discriminant() (defined via
687 EC_METHOD) that verifies that the curve discriminant is non-zero.
689 Add a function EC_GROUP_check() that makes some sanity tests
690 on a EC_GROUP, its generator and order. This includes
691 EC_GROUP_check_discriminant().
692 [Nils Larsch <nla@trustcenter.de>]
694 *) Add ECDSA in new directory crypto/ecdsa/.
696 Add applications 'openssl ecparam' and 'openssl ecdsa'
697 (these are based on 'openssl dsaparam' and 'openssl dsa').
699 ECDSA support is also included in various other files across the
700 library. Most notably,
701 - 'openssl req' now has a '-newkey ecdsa:file' option;
702 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
703 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
704 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
705 them suitable for ECDSA where domain parameters must be
706 extracted before the specific public key;
707 - ECDSA engine support has been added.
708 [Nils Larsch <nla@trustcenter.de>]
710 *) Include some named elliptic curves, and add OIDs from X9.62,
711 SECG, and WAP/WTLS. Each curve can be obtained from the new
713 EC_GROUP_new_by_nid(),
714 and the list of available named curves can be obtained with
715 EC_get_builtin_curves().
716 Also add a 'curve_name' member to EC_GROUP objects, which can be
720 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
722 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
723 was actually never needed) and in BN_mul(). The removal in BN_mul()
724 required a small change in bn_mul_part_recursive() and the addition
725 of the functions bn_cmp_part_words(), bn_sub_part_words() and
726 bn_add_part_words(), which do the same thing as bn_cmp_words(),
727 bn_sub_words() and bn_add_words() except they take arrays with
731 Changes between 0.9.7d and 0.9.7e [XX xxx XXXX]
733 *) Avoid a race condition when CRLs are checked in a multi threaded
734 environment. This would happen due to the reordering of the revoked
735 entries during signature checking and serial number lookup. Now the
736 encoding is cached and the serial number sort performed under a lock.
737 Add new STACK function sk_is_sorted().
740 *) Add Delta CRL to the extension code.
743 *) Various fixes to s3_pkt.c so alerts are sent properly.
744 [David Holmes <d.holmes@f5.com>]
746 *) Reduce the chances of duplicate issuer name and serial numbers (in
747 violation of RFC3280) using the OpenSSL certificate creation utilities.
748 This is done by creating a random 64 bit value for the initial serial
749 number when a serial number file is created or when a self signed
750 certificate is created using 'openssl req -x509'. The initial serial
751 number file is created using 'openssl x509 -next_serial' in CA.pl
752 rather than being initialized to 1.
755 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
757 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
758 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
759 [Joe Orton, Steve Henson]
761 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
763 [Joe Orton, Steve Henson]
765 *) Make it possible to have multiple active certificates with the same
766 subject in the CA index file. This is done only if the keyword
767 'unique_subject' is set to 'no' in the main CA section (default
768 if 'CA_default') of the configuration file. The value is saved
769 with the database itself in a separate index attribute file,
770 named like the index file with '.attr' appended to the name.
773 *) X509 verify fixes. Disable broken certificate workarounds when
774 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
775 keyUsage extension present. Don't accept CRLs with unhandled critical
776 extensions: since verify currently doesn't process CRL extensions this
777 rejects a CRL with *any* critical extensions. Add new verify error codes
781 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
782 A clarification of RFC2560 will require the use of OCTET STRINGs and
783 some implementations cannot handle the current raw format. Since OpenSSL
784 copies and compares OCSP nonces as opaque blobs without any attempt at
785 parsing them this should not create any compatibility issues.
788 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
789 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
790 this HMAC (and other) operations are several times slower than OpenSSL
794 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
795 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
797 *) Use the correct content when signing type "other".
800 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
802 *) Fix various bugs revealed by running the NISCC test suite:
804 Stop out of bounds reads in the ASN1 code when presented with
805 invalid tags (CAN-2003-0543 and CAN-2003-0544).
807 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
809 If verify callback ignores invalid public key errors don't try to check
810 certificate signature with the NULL public key.
814 *) New -ignore_err option in ocsp application to stop the server
815 exiting on the first error in a request.
818 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
819 if the server requested one: as stated in TLS 1.0 and SSL 3.0
823 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
824 extra data after the compression methods not only for TLS 1.0
825 but also for SSL 3.0 (as required by the specification).
826 [Bodo Moeller; problem pointed out by Matthias Loepfe]
828 *) Change X509_certificate_type() to mark the key as exported/exportable
829 when it's 512 *bits* long, not 512 bytes.
832 *) Change AES_cbc_encrypt() so it outputs exact multiple of
833 blocks during encryption.
836 *) Various fixes to base64 BIO and non blocking I/O. On write
837 flushes were not handled properly if the BIO retried. On read
838 data was not being buffered properly and had various logic bugs.
839 This also affects blocking I/O when the data being decoded is a
843 *) Various S/MIME bugfixes and compatibility changes:
844 output correct application/pkcs7 MIME type if
845 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
846 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
847 of files as .eml work). Correctly handle very long lines in MIME
851 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
853 *) Countermeasure against the Klima-Pokorny-Rosa extension of
854 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
855 a protocol version number mismatch like a decryption error
856 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
859 *) Turn on RSA blinding by default in the default implementation
860 to avoid a timing attack. Applications that don't want it can call
861 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
862 They would be ill-advised to do so in most cases.
863 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
865 *) Change RSA blinding code so that it works when the PRNG is not
866 seeded (in this case, the secret RSA exponent is abused as
867 an unpredictable seed -- if it is not unpredictable, there
868 is no point in blinding anyway). Make RSA blinding thread-safe
869 by remembering the creator's thread ID in rsa->blinding and
870 having all other threads use local one-time blinding factors
871 (this requires more computation than sharing rsa->blinding, but
872 avoids excessive locking; and if an RSA object is not shared
873 between threads, blinding will still be very fast).
876 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
877 ENGINE as defaults for all supported algorithms irrespective of
878 the 'flags' parameter. 'flags' is now honoured, so applications
879 should make sure they are passing it correctly.
882 *) Target "mingw" now allows native Windows code to be generated in
883 the Cygwin environment as well as with the MinGW compiler.
886 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
888 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
889 via timing by performing a MAC computation even if incorrrect
890 block cipher padding has been found. This is a countermeasure
891 against active attacks where the attacker has to distinguish
892 between bad padding and a MAC verification error. (CAN-2003-0078)
894 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
895 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
896 Martin Vuagnoux (EPFL, Ilion)]
898 *) Make the no-err option work as intended. The intention with no-err
899 is not to have the whole error stack handling routines removed from
900 libcrypto, it's only intended to remove all the function name and
901 reason texts, thereby removing some of the footprint that may not
902 be interesting if those errors aren't displayed anyway.
904 NOTE: it's still possible for any application or module to have it's
905 own set of error texts inserted. The routines are there, just not
906 used by default when no-err is given.
909 *) Add support for FreeBSD on IA64.
910 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
912 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
913 Kerberos function mit_des_cbc_cksum(). Before this change,
914 the value returned by DES_cbc_cksum() was like the one from
915 mit_des_cbc_cksum(), except the bytes were swapped.
916 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
918 *) Allow an application to disable the automatic SSL chain building.
919 Before this a rather primitive chain build was always performed in
920 ssl3_output_cert_chain(): an application had no way to send the
921 correct chain if the automatic operation produced an incorrect result.
923 Now the chain builder is disabled if either:
925 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
927 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
929 The reasoning behind this is that an application would not want the
930 auto chain building to take place if extra chain certificates are
931 present and it might also want a means of sending no additional
932 certificates (for example the chain has two certificates and the
936 *) Add the possibility to build without the ENGINE framework.
937 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
939 *) Under Win32 gmtime() can return NULL: check return value in
940 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
943 *) DSA routines: under certain error conditions uninitialized BN objects
944 could be freed. Solution: make sure initialization is performed early
945 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
946 Nils Larsch <nla@trustcenter.de> via PR#459)
949 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
950 checked on reconnect on the client side, therefore session resumption
951 could still fail with a "ssl session id is different" error. This
952 behaviour is masked when SSL_OP_ALL is used due to
953 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
954 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
958 *) IA-32 assembler support enhancements: unified ELF targets, support
959 for SCO/Caldera platforms, fix for Cygwin shared build.
962 *) Add support for FreeBSD on sparc64. As a consequence, support for
963 FreeBSD on non-x86 processors is separate from x86 processors on
964 the config script, much like the NetBSD support.
965 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
967 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
969 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
972 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
973 code (06) was taken as the first octet of the session ID and the last
974 octet was ignored consequently. As a result SSLv2 client side session
975 caching could not have worked due to the session ID mismatch between
977 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
981 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
982 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
986 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
987 seems that in spite of existing for more than a year, many application
988 author have done nothing to provide the necessary callbacks, which
989 means that this particular engine will not work properly anywhere.
990 This is a very unfortunate situation which forces us, in the name
991 of usability, to give the hw_ncipher.c a static lock, which is part
993 NOTE: This is for the 0.9.7 series ONLY. This hack will never
994 appear in 0.9.8 or later. We EXPECT application authors to have
995 dealt properly with this when 0.9.8 is released (unless we actually
996 make such changes in the libcrypto locking code that changes will
997 have to be made anyway).
1000 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1001 octets have been read, EOF or an error occurs. Without this change
1002 some truncated ASN1 structures will not produce an error.
1005 *) Disable Heimdal support, since it hasn't been fully implemented.
1006 Still give the possibility to force the use of Heimdal, but with
1007 warnings and a request that patches get sent to openssl-dev.
1010 *) Add the VC-CE target, introduce the WINCE sysname, and add
1011 INSTALL.WCE and appropriate conditionals to make it build.
1012 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1014 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1015 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1016 edit numbers of the version.
1017 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1019 *) Introduce safe string copy and catenation functions
1020 (BUF_strlcpy() and BUF_strlcat()).
1021 [Ben Laurie (CHATS) and Richard Levitte]
1023 *) Avoid using fixed-size buffers for one-line DNs.
1024 [Ben Laurie (CHATS)]
1026 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1027 resizing buffers containing secrets, and use where appropriate.
1028 [Ben Laurie (CHATS)]
1030 *) Avoid using fixed size buffers for configuration file location.
1031 [Ben Laurie (CHATS)]
1033 *) Avoid filename truncation for various CA files.
1034 [Ben Laurie (CHATS)]
1036 *) Use sizeof in preference to magic numbers.
1037 [Ben Laurie (CHATS)]
1039 *) Avoid filename truncation in cert requests.
1040 [Ben Laurie (CHATS)]
1042 *) Add assertions to check for (supposedly impossible) buffer
1044 [Ben Laurie (CHATS)]
1046 *) Don't cache truncated DNS entries in the local cache (this could
1047 potentially lead to a spoofing attack).
1048 [Ben Laurie (CHATS)]
1050 *) Fix various buffers to be large enough for hex/decimal
1051 representations in a platform independent manner.
1052 [Ben Laurie (CHATS)]
1054 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1055 resizing buffers containing secrets, and use where appropriate.
1056 [Ben Laurie (CHATS)]
1058 *) Add BIO_indent() to avoid much slightly worrying code to do
1060 [Ben Laurie (CHATS)]
1062 *) Convert sprintf()/BIO_puts() to BIO_printf().
1063 [Ben Laurie (CHATS)]
1065 *) buffer_gets() could terminate with the buffer only half
1067 [Ben Laurie (CHATS)]
1069 *) Add assertions to prevent user-supplied crypto functions from
1070 overflowing internal buffers by having large block sizes, etc.
1071 [Ben Laurie (CHATS)]
1073 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1075 [Ben Laurie (CHATS)]
1077 *) Eliminate unused copy of key in RC4.
1078 [Ben Laurie (CHATS)]
1080 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1081 [Ben Laurie (CHATS)]
1083 *) Fix off-by-one error in EGD path.
1084 [Ben Laurie (CHATS)]
1086 *) If RANDFILE path is too long, ignore instead of truncating.
1087 [Ben Laurie (CHATS)]
1089 *) Eliminate unused and incorrectly sized X.509 structure
1091 [Ben Laurie (CHATS)]
1093 *) Eliminate unused and dangerous function knumber().
1094 [Ben Laurie (CHATS)]
1096 *) Eliminate unused and dangerous structure, KSSL_ERR.
1097 [Ben Laurie (CHATS)]
1099 *) Protect against overlong session ID context length in an encoded
1100 session object. Since these are local, this does not appear to be
1102 [Ben Laurie (CHATS)]
1104 *) Change from security patch (see 0.9.6e below) that did not affect
1105 the 0.9.6 release series:
1107 Remote buffer overflow in SSL3 protocol - an attacker could
1108 supply an oversized master key in Kerberos-enabled versions.
1110 [Ben Laurie (CHATS)]
1112 *) Change the SSL kerb5 codes to match RFC 2712.
1115 *) Make -nameopt work fully for req and add -reqopt switch.
1116 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1118 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1119 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1121 *) Make sure tests can be performed even if the corresponding algorithms
1122 have been removed entirely. This was also the last step to make
1123 OpenSSL compilable with DJGPP under all reasonable conditions.
1124 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1126 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1127 to allow version independent disabling of normally unselected ciphers,
1128 which may be activated as a side-effect of selecting a single cipher.
1130 (E.g., cipher list string "RSA" enables ciphersuites that are left
1131 out of "ALL" because they do not provide symmetric encryption.
1132 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1133 [Lutz Jaenicke, Bodo Moeller]
1135 *) Add appropriate support for separate platform-dependent build
1136 directories. The recommended way to make a platform-dependent
1137 build directory is the following (tested on Linux), maybe with
1140 # Place yourself outside of the OpenSSL source tree. In
1141 # this example, the environment variable OPENSSL_SOURCE
1142 # is assumed to contain the absolute OpenSSL source directory.
1143 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1144 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1145 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1146 mkdir -p `dirname $F`
1147 ln -s $OPENSSL_SOURCE/$F $F
1150 To be absolutely sure not to disturb the source tree, a "make clean"
1151 is a good thing. If it isn't successfull, don't worry about it,
1152 it probably means the source directory is very clean.
1155 *) Make sure any ENGINE control commands make local copies of string
1156 pointers passed to them whenever necessary. Otherwise it is possible
1157 the caller may have overwritten (or deallocated) the original string
1158 data when a later ENGINE operation tries to use the stored values.
1159 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1161 *) Improve diagnostics in file reading and command-line digests.
1162 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1164 *) Add AES modes CFB and OFB to the object database. Correct an
1165 error in AES-CFB decryption.
1168 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1169 allows existing EVP_CIPHER_CTX structures to be reused after
1170 calling EVP_*Final(). This behaviour is used by encryption
1171 BIOs and some applications. This has the side effect that
1172 applications must explicitly clean up cipher contexts with
1173 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1176 *) Check the values of dna and dnb in bn_mul_recursive before calling
1177 bn_mul_comba (a non zero value means the a or b arrays do not contain
1178 n2 elements) and fallback to bn_mul_normal if either is not zero.
1181 *) Fix escaping of non-ASCII characters when using the -subj option
1182 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1185 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1186 form for "surname", serialNumber has no short form.
1187 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1188 therefore remove "mail" short name for "internet 7".
1189 The OID for unique identifiers in X509 certificates is
1190 x500UniqueIdentifier, not uniqueIdentifier.
1191 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1194 *) Add an "init" command to the ENGINE config module and auto initialize
1195 ENGINEs. Without any "init" command the ENGINE will be initialized
1196 after all ctrl commands have been executed on it. If init=1 the
1197 ENGINE is initailized at that point (ctrls before that point are run
1198 on the uninitialized ENGINE and after on the initialized one). If
1199 init=0 then the ENGINE will not be iniatialized at all.
1202 *) Fix the 'app_verify_callback' interface so that the user-defined
1203 argument is actually passed to the callback: In the
1204 SSL_CTX_set_cert_verify_callback() prototype, the callback
1205 declaration has been changed from
1208 int (*cb)(X509_STORE_CTX *,void *);
1209 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1210 i=s->ctx->app_verify_callback(&ctx)
1211 has been changed into
1212 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1214 To update applications using SSL_CTX_set_cert_verify_callback(),
1215 a dummy argument can be added to their callback functions.
1216 [D. K. Smetters <smetters@parc.xerox.com>]
1218 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1219 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1221 *) Add and OPENSSL_LOAD_CONF define which will cause
1222 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1223 This allows older applications to transparently support certain
1224 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1225 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1226 load the config file and OPENSSL_add_all_algorithms_conf() which will
1227 always load it have also been added.
1230 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1231 Adjust NIDs and EVP layer.
1232 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1234 *) Config modules support in openssl utility.
1236 Most commands now load modules from the config file,
1237 though in a few (such as version) this isn't done
1238 because it couldn't be used for anything.
1240 In the case of ca and req the config file used is
1241 the same as the utility itself: that is the -config
1242 command line option can be used to specify an
1246 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1247 use "openssl_conf" if filename is NULL use default openssl config file.
1250 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1251 config section name. Add a new flag to tolerate a missing config file
1252 and move code to CONF_modules_load_file().
1255 *) Support for crypto accelerator cards from Accelerated Encryption
1256 Processing, www.aep.ie. (Use engine 'aep')
1257 The support was copied from 0.9.6c [engine] and adapted/corrected
1258 to work with the new engine framework.
1259 [AEP Inc. and Richard Levitte]
1261 *) Support for SureWare crypto accelerator cards from Baltimore
1262 Technologies. (Use engine 'sureware')
1263 The support was copied from 0.9.6c [engine] and adapted
1264 to work with the new engine framework.
1267 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1268 make the newer ENGINE framework commands for the CHIL engine work.
1269 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1271 *) Make it possible to produce shared libraries on ReliantUNIX.
1272 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1274 *) Add the configuration target debug-linux-ppro.
1275 Make 'openssl rsa' use the general key loading routines
1276 implemented in apps.c, and make those routines able to
1277 handle the key format FORMAT_NETSCAPE and the variant
1279 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1281 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1282 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1284 *) Add -keyform to rsautl, and document -engine.
1285 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1287 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1288 BIO_R_NO_SUCH_FILE error code rather than the generic
1289 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1292 *) Add new functions
1294 ERR_peek_last_error_line
1295 ERR_peek_last_error_line_data.
1296 These are similar to
1299 ERR_peek_error_line_data,
1300 but report on the latest error recorded rather than the first one
1301 still in the error queue.
1302 [Ben Laurie, Bodo Moeller]
1304 *) default_algorithms option in ENGINE config module. This allows things
1306 default_algorithms = ALL
1307 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1310 *) Prelminary ENGINE config module.
1313 *) New experimental application configuration code.
1316 *) Change the AES code to follow the same name structure as all other
1317 symmetric ciphers, and behave the same way. Move everything to
1318 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1319 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1321 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1322 [Ben Laurie and Theo de Raadt]
1324 *) Add option to output public keys in req command.
1325 [Massimiliano Pala madwolf@openca.org]
1327 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1328 (up to about 10% better than before for P-192 and P-224).
1331 *) New functions/macros
1333 SSL_CTX_set_msg_callback(ctx, cb)
1334 SSL_CTX_set_msg_callback_arg(ctx, arg)
1335 SSL_set_msg_callback(ssl, cb)
1336 SSL_set_msg_callback_arg(ssl, arg)
1338 to request calling a callback function
1340 void cb(int write_p, int version, int content_type,
1341 const void *buf, size_t len, SSL *ssl, void *arg)
1343 whenever a protocol message has been completely received
1344 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1345 protocol version according to which the SSL library interprets
1346 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1347 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1348 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1349 specification (change_cipher_spec(20), alert(21), handshake(22)).
1350 'buf' and 'len' point to the actual message, 'ssl' to the
1351 SSL object, and 'arg' is the application-defined value set by
1352 SSL[_CTX]_set_msg_callback_arg().
1354 'openssl s_client' and 'openssl s_server' have new '-msg' options
1355 to enable a callback that displays all protocol messages.
1358 *) Change the shared library support so shared libraries are built as
1359 soon as the corresponding static library is finished, and thereby get
1360 openssl and the test programs linked against the shared library.
1361 This still only happens when the keyword "shard" has been given to
1362 the configuration scripts.
1364 NOTE: shared library support is still an experimental thing, and
1365 backward binary compatibility is still not guaranteed.
1366 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1368 *) Add support for Subject Information Access extension.
1369 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1371 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1372 additional bytes when new memory had to be allocated, not just
1373 when reusing an existing buffer.
1376 *) New command line and configuration option 'utf8' for the req command.
1377 This allows field values to be specified as UTF8 strings.
1380 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1381 runs for the former and machine-readable output for the latter.
1384 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1385 of the e-mail address in the DN (i.e., it will go into a certificate
1386 extension only). The new configuration file option 'email_in_dn = no'
1387 has the same effect.
1388 [Massimiliano Pala madwolf@openca.org]
1390 *) Change all functions with names starting with des_ to be starting
1391 with DES_ instead. Add wrappers that are compatible with libdes,
1392 but are named _ossl_old_des_*. Finally, add macros that map the
1393 des_* symbols to the corresponding _ossl_old_des_* if libdes
1394 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1395 desired, the des_* symbols will be mapped to DES_*, with one
1398 Since we provide two compatibility mappings, the user needs to
1399 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1400 compatibility is desired. The default (i.e., when that macro
1401 isn't defined) is OpenSSL 0.9.6c compatibility.
1403 There are also macros that enable and disable the support of old
1404 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1405 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1406 are defined, the default will apply: to support the old des routines.
1408 In either case, one must include openssl/des.h to get the correct
1409 definitions. Do not try to just include openssl/des_old.h, that
1412 NOTE: This is a major break of an old API into a new one. Software
1413 authors are encouraged to switch to the DES_ style functions. Some
1414 time in the future, des_old.h and the libdes compatibility functions
1415 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1416 default), and then completely removed.
1419 *) Test for certificates which contain unsupported critical extensions.
1420 If such a certificate is found during a verify operation it is
1421 rejected by default: this behaviour can be overridden by either
1422 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1423 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1424 X509_supported_extension() has also been added which returns 1 if a
1425 particular extension is supported.
1428 *) Modify the behaviour of EVP cipher functions in similar way to digests
1429 to retain compatibility with existing code.
1432 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1433 compatibility with existing code. In particular the 'ctx' parameter does
1434 not have to be to be initialized before the call to EVP_DigestInit() and
1435 it is tidied up after a call to EVP_DigestFinal(). New function
1436 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1437 EVP_MD_CTX_copy() changed to not require the destination to be
1438 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1439 requires the destination to be valid.
1441 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1442 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1445 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1446 so that complete 'Handshake' protocol structures are kept in memory
1447 instead of overwriting 'msg_type' and 'length' with 'body' data.
1450 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1451 [Massimo Santin via Richard Levitte]
1453 *) Major restructuring to the underlying ENGINE code. This includes
1454 reduction of linker bloat, separation of pure "ENGINE" manipulation
1455 (initialisation, etc) from functionality dealing with implementations
1456 of specific crypto iterfaces. This change also introduces integrated
1457 support for symmetric ciphers and digest implementations - so ENGINEs
1458 can now accelerate these by providing EVP_CIPHER and EVP_MD
1459 implementations of their own. This is detailed in crypto/engine/README
1460 as it couldn't be adequately described here. However, there are a few
1461 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1462 were changed in the original introduction of ENGINE code have now
1463 reverted back - the hooking from this code to ENGINE is now a good
1464 deal more passive and at run-time, operations deal directly with
1465 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1466 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1467 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1468 they were not being used by the framework as there is no concept of a
1469 BIGNUM_METHOD and they could not be generalised to the new
1470 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1471 ENGINE_cpy() has been removed as it cannot be consistently defined in
1475 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1478 *) Change mkdef.pl to sort symbols that get the same entry number,
1479 and make sure the automatically generated functions ERR_load_*
1480 become part of libeay.num as well.
1483 *) New function SSL_renegotiate_pending(). This returns true once
1484 renegotiation has been requested (either SSL_renegotiate() call
1485 or HelloRequest/ClientHello receveived from the peer) and becomes
1486 false once a handshake has been completed.
1487 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1488 sends a HelloRequest, but does not ensure that a handshake takes
1489 place. SSL_renegotiate_pending() is useful for checking if the
1490 client has followed the request.)
1493 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1494 By default, clients may request session resumption even during
1495 renegotiation (if session ID contexts permit); with this option,
1496 session resumption is possible only in the first handshake.
1498 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1499 more bits available for options that should not be part of
1500 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1503 *) Add some demos for certificate and certificate request creation.
1506 *) Make maximum certificate chain size accepted from the peer application
1507 settable (SSL*_get/set_max_cert_list()), as proposed by
1508 "Douglas E. Engert" <deengert@anl.gov>.
1511 *) Add support for shared libraries for Unixware-7
1512 (Boyd Lynn Gerber <gerberb@zenez.com>).
1515 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1516 be done prior to destruction. Use this to unload error strings from
1517 ENGINEs that load their own error strings. NB: This adds two new API
1518 functions to "get" and "set" this destroy handler in an ENGINE.
1521 *) Alter all existing ENGINE implementations (except "openssl" and
1522 "openbsd") to dynamically instantiate their own error strings. This
1523 makes them more flexible to be built both as statically-linked ENGINEs
1524 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1525 Also, add stub code to each that makes building them as self-contained
1526 shared-libraries easier (see README.ENGINE).
1529 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1530 implementations into applications that are completely implemented in
1531 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1532 commands that can be used to configure what shared-library to load and
1533 to control aspects of the way it is handled. Also, made an update to
1534 the README.ENGINE file that brings its information up-to-date and
1535 provides some information and instructions on the "dynamic" ENGINE
1536 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1539 *) Make it possible to unload ranges of ERR strings with a new
1540 "ERR_unload_strings" function.
1543 *) Add a copy() function to EVP_MD.
1546 *) Make EVP_MD routines take a context pointer instead of just the
1547 md_data void pointer.
1550 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1551 that the digest can only process a single chunk of data
1552 (typically because it is provided by a piece of
1553 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1554 is only going to provide a single chunk of data, and hence the
1555 framework needn't accumulate the data for oneshot drivers.
1558 *) As with "ERR", make it possible to replace the underlying "ex_data"
1559 functions. This change also alters the storage and management of global
1560 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1561 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1562 index counters. The API functions that use this state have been changed
1563 to take a "class_index" rather than pointers to the class's local STACK
1564 and counter, and there is now an API function to dynamically create new
1565 classes. This centralisation allows us to (a) plug a lot of the
1566 thread-safety problems that existed, and (b) makes it possible to clean
1567 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1568 such data would previously have always leaked in application code and
1569 workarounds were in place to make the memory debugging turn a blind eye
1570 to it. Application code that doesn't use this new function will still
1571 leak as before, but their memory debugging output will announce it now
1572 rather than letting it slide.
1574 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1575 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1576 has a return value to indicate success or failure.
1579 *) Make it possible to replace the underlying "ERR" functions such that the
1580 global state (2 LHASH tables and 2 locks) is only used by the "default"
1581 implementation. This change also adds two functions to "get" and "set"
1582 the implementation prior to it being automatically set the first time
1583 any other ERR function takes place. Ie. an application can call "get",
1584 pass the return value to a module it has just loaded, and that module
1585 can call its own "set" function using that value. This means the
1586 module's "ERR" operations will use (and modify) the error state in the
1587 application and not in its own statically linked copy of OpenSSL code.
1590 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1591 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1592 the operation, and provides a more encapsulated way for external code
1593 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1594 to use these functions rather than manually incrementing the counts.
1596 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1599 *) Add EVP test program.
1602 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1605 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1606 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1607 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1608 These allow a CRL to be built without having to access X509_CRL fields
1609 directly. Modify 'ca' application to use new functions.
1612 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1613 bug workarounds. Rollback attack detection is a security feature.
1614 The problem will only arise on OpenSSL servers when TLSv1 is not
1615 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1616 Software authors not wanting to support TLSv1 will have special reasons
1617 for their choice and can explicitly enable this option.
1618 [Bodo Moeller, Lutz Jaenicke]
1620 *) Rationalise EVP so it can be extended: don't include a union of
1621 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1622 (similar to those existing for EVP_CIPHER_CTX).
1627 EVP_MD_CTX_init(&md); /* new function call */
1628 EVP_DigestInit(&md, EVP_sha1());
1629 EVP_DigestUpdate(&md, in, len);
1630 EVP_DigestFinal(&md, out, NULL);
1631 EVP_MD_CTX_cleanup(&md); /* new function call */
1635 *) Make DES key schedule conform to the usual scheme, as well as
1636 correcting its structure. This means that calls to DES functions
1637 now have to pass a pointer to a des_key_schedule instead of a
1638 plain des_key_schedule (which was actually always a pointer
1641 des_key_schedule ks;
1643 des_set_key_checked(..., &ks);
1644 des_ncbc_encrypt(..., &ks, ...);
1646 (Note that a later change renames 'des_...' into 'DES_...'.)
1649 *) Initial reduction of linker bloat: the use of some functions, such as
1650 PEM causes large amounts of unused functions to be linked in due to
1651 poor organisation. For example pem_all.c contains every PEM function
1652 which has a knock on effect of linking in large amounts of (unused)
1653 ASN1 code. Grouping together similar functions and splitting unrelated
1654 functions prevents this.
1657 *) Cleanup of EVP macros.
1660 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1661 correct _ecb suffix.
1664 *) Add initial OCSP responder support to ocsp application. The
1665 revocation information is handled using the text based index
1666 use by the ca application. The responder can either handle
1667 requests generated internally, supplied in files (for example
1668 via a CGI script) or using an internal minimal server.
1671 *) Add configuration choices to get zlib compression for TLS.
1674 *) Changes to Kerberos SSL for RFC 2712 compliance:
1675 1. Implemented real KerberosWrapper, instead of just using
1676 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1677 2. Implemented optional authenticator field of KerberosWrapper.
1679 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1680 and authenticator structs; see crypto/krb5/.
1682 Generalized Kerberos calls to support multiple Kerberos libraries.
1683 [Vern Staats <staatsvr@asc.hpc.mil>,
1684 Jeffrey Altman <jaltman@columbia.edu>
1685 via Richard Levitte]
1687 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1688 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1689 values for each of the key sizes rather than having just
1690 parameters (and 'speed' generating keys each time).
1693 *) Speed up EVP routines.
1696 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1697 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1698 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1699 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1701 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1702 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1703 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1706 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1708 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1711 *) Added the OS2-EMX target.
1712 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1714 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1715 to support NCONF routines in extension code. New function CONF_set_nconf()
1716 to allow functions which take an NCONF to also handle the old LHASH
1717 structure: this means that the old CONF compatible routines can be
1718 retained (in particular wrt extensions) without having to duplicate the
1719 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1722 *) Enhance the general user interface with mechanisms for inner control
1723 and with possibilities to have yes/no kind of prompts.
1726 *) Change all calls to low level digest routines in the library and
1727 applications to use EVP. Add missing calls to HMAC_cleanup() and
1728 don't assume HMAC_CTX can be copied using memcpy().
1729 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1731 *) Add the possibility to control engines through control names but with
1732 arbitrary arguments instead of just a string.
1733 Change the key loaders to take a UI_METHOD instead of a callback
1734 function pointer. NOTE: this breaks binary compatibility with earlier
1735 versions of OpenSSL [engine].
1736 Adapt the nCipher code for these new conditions and add a card insertion
1740 *) Enhance the general user interface with mechanisms to better support
1741 dialog box interfaces, application-defined prompts, the possibility
1742 to use defaults (for example default passwords from somewhere else)
1743 and interrupts/cancellations.
1746 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1747 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1750 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1751 tidy up some unnecessarily weird code in 'sk_new()').
1752 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1754 *) Change the key loading routines for ENGINEs to use the same kind
1755 callback (pem_password_cb) as all other routines that need this
1759 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1760 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1761 than this minimum value is recommended.
1764 *) New random seeder for OpenVMS, using the system process statistics
1765 that are easily reachable.
1768 *) Windows apparently can't transparently handle global
1769 variables defined in DLLs. Initialisations such as:
1771 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1773 wont compile. This is used by the any applications that need to
1774 declare their own ASN1 modules. This was fixed by adding the option
1775 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1776 needed for static libraries under Win32.
1779 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1780 setting of purpose and trust fields. New X509_STORE trust and
1781 purpose functions and tidy up setting in other SSL functions.
1784 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1785 structure. These are inherited by X509_STORE_CTX when it is
1786 initialised. This allows various defaults to be set in the
1787 X509_STORE structure (such as flags for CRL checking and custom
1788 purpose or trust settings) for functions which only use X509_STORE_CTX
1789 internally such as S/MIME.
1791 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1792 trust settings if they are not set in X509_STORE. This allows X509_STORE
1793 purposes and trust (in S/MIME for example) to override any set by default.
1795 Add command line options for CRL checking to smime, s_client and s_server
1799 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1800 are set then the CRL is looked up in the X509_STORE structure and
1801 its validity and signature checked, then if the certificate is found
1802 in the CRL the verify fails with a revoked error.
1804 Various new CRL related callbacks added to X509_STORE_CTX structure.
1806 Command line options added to 'verify' application to support this.
1808 This needs some additional work, such as being able to handle multiple
1809 CRLs with different times, extension based lookup (rather than just
1810 by subject name) and ultimately more complete V2 CRL extension
1814 *) Add a general user interface API (crypto/ui/). This is designed
1815 to replace things like des_read_password and friends (backward
1816 compatibility functions using this new API are provided).
1817 The purpose is to remove prompting functions from the DES code
1818 section as well as provide for prompting through dialog boxes in
1819 a window system and the like.
1822 *) Add "ex_data" support to ENGINE so implementations can add state at a
1823 per-structure level rather than having to store it globally.
1826 *) Make it possible for ENGINE structures to be copied when retrieved by
1827 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1828 This causes the "original" ENGINE structure to act like a template,
1829 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1830 operational state can be localised to each ENGINE structure, despite the
1831 fact they all share the same "methods". New ENGINE structures returned in
1832 this case have no functional references and the return value is the single
1833 structural reference. This matches the single structural reference returned
1834 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1838 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1839 needs to match any other type at all we need to manually clear the
1843 *) Changes to the "openssl engine" utility to include;
1844 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1845 about an ENGINE's available control commands.
1846 - executing control commands from command line arguments using the
1847 '-pre' and '-post' switches. '-post' is only used if '-t' is
1848 specified and the ENGINE is successfully initialised. The syntax for
1849 the individual commands are colon-separated, for example;
1850 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1853 *) New dynamic control command support for ENGINEs. ENGINEs can now
1854 declare their own commands (numbers), names (strings), descriptions,
1855 and input types for run-time discovery by calling applications. A
1856 subset of these commands are implicitly classed as "executable"
1857 depending on their input type, and only these can be invoked through
1858 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1859 can be based on user input, config files, etc). The distinction is
1860 that "executable" commands cannot return anything other than a boolean
1861 result and can only support numeric or string input, whereas some
1862 discoverable commands may only be for direct use through
1863 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1864 pointers, or other custom uses. The "executable" commands are to
1865 support parameterisations of ENGINE behaviour that can be
1866 unambiguously defined by ENGINEs and used consistently across any
1867 OpenSSL-based application. Commands have been added to all the
1868 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1869 control over shared-library paths without source code alterations.
1872 *) Changed all ENGINE implementations to dynamically allocate their
1873 ENGINEs rather than declaring them statically. Apart from this being
1874 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1875 this also allows the implementations to compile without using the
1876 internal engine_int.h header.
1879 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1880 'const' value. Any code that should be able to modify a RAND_METHOD
1881 should already have non-const pointers to it (ie. they should only
1882 modify their own ones).
1885 *) Made a variety of little tweaks to the ENGINE code.
1886 - "atalla" and "ubsec" string definitions were moved from header files
1887 to C code. "nuron" string definitions were placed in variables
1888 rather than hard-coded - allowing parameterisation of these values
1889 later on via ctrl() commands.
1890 - Removed unused "#if 0"'d code.
1891 - Fixed engine list iteration code so it uses ENGINE_free() to release
1892 structural references.
1893 - Constified the RAND_METHOD element of ENGINE structures.
1894 - Constified various get/set functions as appropriate and added
1895 missing functions (including a catch-all ENGINE_cpy that duplicates
1896 all ENGINE values onto a new ENGINE except reference counts/state).
1897 - Removed NULL parameter checks in get/set functions. Setting a method
1898 or function to NULL is a way of cancelling out a previously set
1899 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1900 and doesn't justify the extra error symbols and code.
1901 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1902 flags from engine_int.h to engine.h.
1903 - Changed prototypes for ENGINE handler functions (init(), finish(),
1904 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1907 *) Implement binary inversion algorithm for BN_mod_inverse in addition
1908 to the algorithm using long division. The binary algorithm can be
1909 used only if the modulus is odd. On 32-bit systems, it is faster
1910 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1911 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1912 up to 450 bits. In 64-bit environments, the binary algorithm
1913 appears to be advantageous for much longer moduli; here we use it
1914 for moduli up to 2048 bits.
1917 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1918 could not support the combine flag in choice fields.
1921 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
1922 extensions from a certificate request to the certificate.
1925 *) Allow multiple 'certopt' and 'nameopt' options to be separated
1926 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1927 file: this allows the display of the certificate about to be
1928 signed to be customised, to allow certain fields to be included
1929 or excluded and extension details. The old system didn't display
1930 multicharacter strings properly, omitted fields not in the policy
1931 and couldn't display additional details such as extensions.
1934 *) Function EC_POINTs_mul for multiple scalar multiplication
1935 of an arbitrary number of elliptic curve points
1936 \sum scalars[i]*points[i],
1937 optionally including the generator defined for the EC_GROUP:
1938 scalar*generator + \sum scalars[i]*points[i].
1940 EC_POINT_mul is a simple wrapper function for the typical case
1941 that the point list has just one item (besides the optional
1945 *) First EC_METHODs for curves over GF(p):
1947 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1948 operations and provides various method functions that can also
1949 operate with faster implementations of modular arithmetic.
1951 EC_GFp_mont_method() reuses most functions that are part of
1952 EC_GFp_simple_method, but uses Montgomery arithmetic.
1954 [Bodo Moeller; point addition and point doubling
1955 implementation directly derived from source code provided by
1956 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1958 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1959 crypto/ec/ec_lib.c):
1961 Curves are EC_GROUP objects (with an optional group generator)
1962 based on EC_METHODs that are built into the library.
1964 Points are EC_POINT objects based on EC_GROUP objects.
1966 Most of the framework would be able to handle curves over arbitrary
1967 finite fields, but as there are no obvious types for fields other
1968 than GF(p), some functions are limited to that for now.
1971 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1972 that the file contains a complete HTTP response.
1975 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1976 change the def and num file printf format specifier from "%-40sXXX"
1977 to "%-39s XXX". The latter will always guarantee a space after the
1978 field while the former will cause them to run together if the field
1979 is 40 of more characters long.
1982 *) Constify the cipher and digest 'method' functions and structures
1983 and modify related functions to take constant EVP_MD and EVP_CIPHER
1987 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1988 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1991 *) Modify EVP_Digest*() routines so they now return values. Although the
1992 internal software routines can never fail additional hardware versions
1996 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1998 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1999 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2002 ERR_R_NESTED_ASN1_ERROR
2004 ERR_R_MISSING_ASN1_EOS
2005 were 4 .. 9, conflicting with
2006 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2008 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2009 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2011 Add new error code 'ERR_R_INTERNAL_ERROR'.
2014 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2018 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2019 sets the subject name for a new request or supersedes the
2020 subject name in a given request. Formats that can be parsed are
2021 'CN=Some Name, OU=myOU, C=IT'
2023 'CN=Some Name/OU=myOU/C=IT'.
2025 Add options '-batch' and '-verbose' to 'openssl req'.
2026 [Massimiliano Pala <madwolf@hackmasters.net>]
2028 *) Introduce the possibility to access global variables through
2029 functions on platform were that's the best way to handle exporting
2030 global variables in shared libraries. To enable this functionality,
2031 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2032 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2033 is normally done by Configure or something similar).
2035 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2036 in the source file (foo.c) like this:
2038 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2039 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2041 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2042 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2044 OPENSSL_DECLARE_GLOBAL(int,foo);
2045 #define foo OPENSSL_GLOBAL_REF(foo)
2046 OPENSSL_DECLARE_GLOBAL(double,bar);
2047 #define bar OPENSSL_GLOBAL_REF(bar)
2049 The #defines are very important, and therefore so is including the
2050 header file everywhere where the defined globals are used.
2052 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2053 of ASN.1 items, but that structure is a bit different.
2055 The largest change is in util/mkdef.pl which has been enhanced with
2056 better and easier to understand logic to choose which symbols should
2057 go into the Windows .def files as well as a number of fixes and code
2058 cleanup (among others, algorithm keywords are now sorted
2059 lexicographically to avoid constant rewrites).
2062 *) In BN_div() keep a copy of the sign of 'num' before writing the
2063 result to 'rm' because if rm==num the value will be overwritten
2064 and produce the wrong result if 'num' is negative: this caused
2065 problems with BN_mod() and BN_nnmod().
2068 *) Function OCSP_request_verify(). This checks the signature on an
2069 OCSP request and verifies the signer certificate. The signer
2070 certificate is just checked for a generic purpose and OCSP request
2074 *) Add OCSP_check_validity() function to check the validity of OCSP
2075 responses. OCSP responses are prepared in real time and may only
2076 be a few seconds old. Simply checking that the current time lies
2077 between thisUpdate and nextUpdate max reject otherwise valid responses
2078 caused by either OCSP responder or client clock inaccuracy. Instead
2079 we allow thisUpdate and nextUpdate to fall within a certain period of
2080 the current time. The age of the response can also optionally be
2081 checked. Two new options -validity_period and -status_age added to
2085 *) If signature or public key algorithm is unrecognized print out its
2086 OID rather that just UNKNOWN.
2089 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2090 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2091 ID to be generated from the issuer certificate alone which can then be
2092 passed to OCSP_id_issuer_cmp().
2095 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2096 ASN1 modules to export functions returning ASN1_ITEM pointers
2097 instead of the ASN1_ITEM structures themselves. This adds several
2098 new macros which allow the underlying ASN1 function/structure to
2099 be accessed transparently. As a result code should not use ASN1_ITEM
2100 references directly (such as &X509_it) but instead use the relevant
2101 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2102 use of the new ASN1 code on platforms where exporting structures
2103 is problematical (for example in shared libraries) but exporting
2104 functions returning pointers to structures is not.
2107 *) Add support for overriding the generation of SSL/TLS session IDs.
2108 These callbacks can be registered either in an SSL_CTX or per SSL.
2109 The purpose of this is to allow applications to control, if they wish,
2110 the arbitrary values chosen for use as session IDs, particularly as it
2111 can be useful for session caching in multiple-server environments. A
2112 command-line switch for testing this (and any client code that wishes
2113 to use such a feature) has been added to "s_server".
2114 [Geoff Thorpe, Lutz Jaenicke]
2116 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2117 of the form '#if defined(...) || defined(...) || ...' and
2118 '#if !defined(...) && !defined(...) && ...'. This also avoids
2119 the growing number of special cases it was previously handling.
2122 *) Make all configuration macros available for application by making
2123 sure they are available in opensslconf.h, by giving them names starting
2124 with "OPENSSL_" to avoid conflicts with other packages and by making
2125 sure e_os2.h will cover all platform-specific cases together with
2127 Additionally, it is now possible to define configuration/platform-
2128 specific names (called "system identities"). In the C code, these
2129 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2130 macro with the name beginning with "OPENSSL_SYS_", which is determined
2131 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2135 *) New option -set_serial to 'req' and 'x509' this allows the serial
2136 number to use to be specified on the command line. Previously self
2137 signed certificates were hard coded with serial number 0 and the
2138 CA options of 'x509' had to use a serial number in a file which was
2142 *) New options to 'ca' utility to support V2 CRL entry extensions.
2143 Currently CRL reason, invalidity date and hold instruction are
2144 supported. Add new CRL extensions to V3 code and some new objects.
2147 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2148 disable standard block padding (aka PKCS#5 padding) in the EVP
2149 API, which was previously mandatory. This means that the data is
2150 not padded in any way and so the total length much be a multiple
2151 of the block size, otherwise an error occurs.
2154 *) Initial (incomplete) OCSP SSL support.
2157 *) New function OCSP_parse_url(). This splits up a URL into its host,
2158 port and path components: primarily to parse OCSP URLs. New -url
2159 option to ocsp utility.
2162 *) New nonce behavior. The return value of OCSP_check_nonce() now
2163 reflects the various checks performed. Applications can decide
2164 whether to tolerate certain situations such as an absent nonce
2165 in a response when one was present in a request: the ocsp application
2166 just prints out a warning. New function OCSP_add1_basic_nonce()
2167 this is to allow responders to include a nonce in a response even if
2168 the request is nonce-less.
2171 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2172 skipped when using openssl x509 multiple times on a single input file,
2173 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2176 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2177 set string type: to handle setting ASN1_TIME structures. Fix ca
2178 utility to correctly initialize revocation date of CRLs.
2181 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2182 the clients preferred ciphersuites and rather use its own preferences.
2183 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2184 Internet Explorer by ensuring unchanged hash method during stepup.
2185 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2188 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2189 to aes and add a new 'exist' option to print out symbols that don't
2193 *) Additional options to ocsp utility to allow flags to be set and
2194 additional certificates supplied.
2197 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2198 OCSP client a number of certificate to only verify the response
2202 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2203 handle the new API. Currently only ECB, CBC modes supported. Add new
2206 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2207 Encryption Standard (AES) Ciphersuites for Transport Layer
2208 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2209 not enabled by default and were not part of the "ALL" ciphersuite
2210 alias because they were not yet official; they could be
2211 explicitly requested by specifying the "AESdraft" ciphersuite
2212 group alias. In the final release of OpenSSL 0.9.7, the group
2213 alias is called "AES" and is part of "ALL".)
2214 [Ben Laurie, Steve Henson, Bodo Moeller]
2216 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2217 request to response.
2220 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2221 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2222 extract information from a certificate request. OCSP_response_create()
2223 creates a response and optionally adds a basic response structure.
2224 OCSP_basic_add1_status() adds a complete single response to a basic
2225 response and returns the OCSP_SINGLERESP structure just added (to allow
2226 extensions to be included for example). OCSP_basic_add1_cert() adds a
2227 certificate to a basic response and OCSP_basic_sign() signs a basic
2228 response with various flags. New helper functions ASN1_TIME_check()
2229 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2230 (converts ASN1_TIME to GeneralizedTime).
2233 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2234 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2235 structure from a certificate. X509_pubkey_digest() digests the public_key
2236 contents: this is used in various key identifiers.
2239 *) Make sk_sort() tolerate a NULL argument.
2240 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2242 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2243 passed by the function are trusted implicitly. If any of them signed the
2244 response then it is assumed to be valid and is not verified.
2247 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2248 to data. This was previously part of the PKCS7 ASN1 code. This
2249 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2250 [Steve Henson, reported by Kenneth R. Robinette
2251 <support@securenetterm.com>]
2253 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2254 routines: without these tracing memory leaks is very painful.
2255 Fix leaks in PKCS12 and PKCS7 routines.
2258 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2259 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2260 effectively meant GeneralizedTime would never be used. Now it
2261 is initialised to -1 but X509_time_adj() now has to check the value
2262 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2263 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2264 [Steve Henson, reported by Kenneth R. Robinette
2265 <support@securenetterm.com>]
2267 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2268 result in a zero length in the ASN1_INTEGER structure which was
2269 not consistent with the structure when d2i_ASN1_INTEGER() was used
2270 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2271 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2272 where it did not print out a minus for negative ASN1_INTEGER.
2275 *) Add summary printout to ocsp utility. The various functions which
2276 convert status values to strings have been renamed to:
2277 OCSP_response_status_str(), OCSP_cert_status_str() and
2278 OCSP_crl_reason_str() and are no longer static. New options
2279 to verify nonce values and to disable verification. OCSP response
2280 printout format cleaned up.
2283 *) Add additional OCSP certificate checks. These are those specified
2284 in RFC2560. This consists of two separate checks: the CA of the
2285 certificate being checked must either be the OCSP signer certificate
2286 or the issuer of the OCSP signer certificate. In the latter case the
2287 OCSP signer certificate must contain the OCSP signing extended key
2288 usage. This check is performed by attempting to match the OCSP
2289 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2290 in the OCSP_CERTID structures of the response.
2293 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2294 and related routines. This uses the standard OpenSSL certificate
2295 verify routines to perform initial checks (just CA validity) and
2296 to obtain the certificate chain. Then additional checks will be
2297 performed on the chain. Currently the root CA is checked to see
2298 if it is explicitly trusted for OCSP signing. This is used to set
2299 a root CA as a global signing root: that is any certificate that
2300 chains to that CA is an acceptable OCSP signing certificate.
2303 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2304 extensions from a separate configuration file.
2305 As when reading extensions from the main configuration file,
2306 the '-extensions ...' option may be used for specifying the
2308 [Massimiliano Pala <madwolf@comune.modena.it>]
2310 *) New OCSP utility. Allows OCSP requests to be generated or
2311 read. The request can be sent to a responder and the output
2312 parsed, outputed or printed in text form. Not complete yet:
2313 still needs to check the OCSP response validity.
2316 *) New subcommands for 'openssl ca':
2317 'openssl ca -status <serial>' prints the status of the cert with
2318 the given serial number (according to the index file).
2319 'openssl ca -updatedb' updates the expiry status of certificates
2321 [Massimiliano Pala <madwolf@comune.modena.it>]
2323 *) New '-newreq-nodes' command option to CA.pl. This is like
2324 '-newreq', but calls 'openssl req' with the '-nodes' option
2325 so that the resulting key is not encrypted.
2326 [Damien Miller <djm@mindrot.org>]
2328 *) New configuration for the GNU Hurd.
2329 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2331 *) Initial code to implement OCSP basic response verify. This
2332 is currently incomplete. Currently just finds the signer's
2333 certificate and verifies the signature on the response.
2336 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2337 value of OPENSSLDIR. This is available via the new '-d' option
2338 to 'openssl version', and is also included in 'openssl version -a'.
2341 *) Allowing defining memory allocation callbacks that will be given
2342 file name and line number information in additional arguments
2343 (a const char* and an int). The basic functionality remains, as
2344 well as the original possibility to just replace malloc(),
2345 realloc() and free() by functions that do not know about these
2346 additional arguments. To register and find out the current
2347 settings for extended allocation functions, the following
2348 functions are provided:
2350 CRYPTO_set_mem_ex_functions
2351 CRYPTO_set_locked_mem_ex_functions
2352 CRYPTO_get_mem_ex_functions
2353 CRYPTO_get_locked_mem_ex_functions
2355 These work the same way as CRYPTO_set_mem_functions and friends.
2356 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2357 extended allocation function is enabled.
2358 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2359 a conventional allocation function is enabled.
2360 [Richard Levitte, Bodo Moeller]
2362 *) Finish off removing the remaining LHASH function pointer casts.
2363 There should no longer be any prototype-casting required when using
2364 the LHASH abstraction, and any casts that remain are "bugs". See
2365 the callback types and macros at the head of lhash.h for details
2366 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2369 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2370 If /dev/[u]random devices are not available or do not return enough
2371 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2373 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2374 /etc/entropy will be queried once each in this sequence, quering stops
2375 when enough entropy was collected without querying more sockets.
2378 *) Change the Unix RAND_poll() variant to be able to poll several
2379 random devices, as specified by DEVRANDOM, until a sufficient amount
2380 of data has been collected. We spend at most 10 ms on each file
2381 (select timeout) and read in non-blocking mode. DEVRANDOM now
2382 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2383 (previously it was just the string "/dev/urandom"), so on typical
2384 platforms the 10 ms delay will never occur.
2385 Also separate out the Unix variant to its own file, rand_unix.c.
2386 For VMS, there's a currently-empty rand_vms.c.
2389 *) Move OCSP client related routines to ocsp_cl.c. These
2390 provide utility functions which an application needing
2391 to issue a request to an OCSP responder and analyse the
2392 response will typically need: as opposed to those which an
2393 OCSP responder itself would need which will be added later.
2395 OCSP_request_sign() signs an OCSP request with an API similar
2396 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2397 response. OCSP_response_get1_basic() extracts basic response
2398 from response. OCSP_resp_find_status(): finds and extracts status
2399 information from an OCSP_CERTID structure (which will be created
2400 when the request structure is built). These are built from lower
2401 level functions which work on OCSP_SINGLERESP structures but
2402 wont normally be used unless the application wishes to examine
2403 extensions in the OCSP response for example.
2405 Replace nonce routines with a pair of functions.
2406 OCSP_request_add1_nonce() adds a nonce value and optionally
2407 generates a random value. OCSP_check_nonce() checks the
2408 validity of the nonce in an OCSP response.
2411 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2412 This doesn't copy the supplied OCSP_CERTID and avoids the
2413 need to free up the newly created id. Change return type
2414 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2415 This can then be used to add extensions to the request.
2416 Deleted OCSP_request_new(), since most of its functionality
2417 is now in OCSP_REQUEST_new() (and the case insensitive name
2418 clash) apart from the ability to set the request name which
2419 will be added elsewhere.
2422 *) Update OCSP API. Remove obsolete extensions argument from
2423 various functions. Extensions are now handled using the new
2424 OCSP extension code. New simple OCSP HTTP function which
2425 can be used to send requests and parse the response.
2428 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2429 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2430 uses the special reorder version of SET OF to sort the attributes
2431 and reorder them to match the encoded order. This resolves a long
2432 standing problem: a verify on a PKCS7 structure just after signing
2433 it used to fail because the attribute order did not match the
2434 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2435 it uses the received order. This is necessary to tolerate some broken
2436 software that does not order SET OF. This is handled by encoding
2437 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2438 to produce the required SET OF.
2441 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2442 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2443 files to get correct declarations of the ASN.1 item variables.
2446 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2447 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2448 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2449 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2450 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2451 ASN1_ITEM and no wrapper functions.
2454 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2455 replace the old function pointer based I/O routines. Change most of
2456 the *_d2i_bio() and *_d2i_fp() functions to use these.
2459 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2460 lines, recognice more "algorithms" that can be deselected, and make
2461 it complain about algorithm deselection that isn't recognised.
2464 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2465 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2466 to use new functions. Add NO_ASN1_OLD which can be set to remove
2467 some old style ASN1 functions: this can be used to determine if old
2468 code will still work when these eventually go away.
2471 *) New extension functions for OCSP structures, these follow the
2472 same conventions as certificates and CRLs.
2475 *) New function X509V3_add1_i2d(). This automatically encodes and
2476 adds an extension. Its behaviour can be customised with various
2477 flags to append, replace or delete. Various wrappers added for
2478 certifcates and CRLs.
2481 *) Fix to avoid calling the underlying ASN1 print routine when
2482 an extension cannot be parsed. Correct a typo in the
2483 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2486 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2487 entries for variables.
2490 *) Add functionality to apps/openssl.c for detecting locking
2491 problems: As the program is single-threaded, all we have
2492 to do is register a locking callback using an array for
2493 storing which locks are currently held by the program.
2496 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2497 SSL_get_ex_data_X509_STORE_idx(), which is used in
2498 ssl_verify_cert_chain() and thus can be called at any time
2499 during TLS/SSL handshakes so that thread-safety is essential.
2500 Unfortunately, the ex_data design is not at all suited
2501 for multi-threaded use, so it probably should be abolished.
2504 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2505 [Broadcom, tweaked and integrated by Geoff Thorpe]
2507 *) Move common extension printing code to new function
2508 X509V3_print_extensions(). Reorganise OCSP print routines and
2509 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2512 *) New function X509_signature_print() to remove duplication in some
2516 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2517 set (this was treated exactly the same as SET OF previously). This
2518 is used to reorder the STACK representing the structure to match the
2519 encoding. This will be used to get round a problem where a PKCS7
2520 structure which was signed could not be verified because the STACK
2521 order did not reflect the encoded order.
2524 *) Reimplement the OCSP ASN1 module using the new code.
2527 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2528 for its ASN1 operations. The old style function pointers still exist
2529 for now but they will eventually go away.
2532 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2533 completely replaces the old ASN1 functionality with a table driven
2534 encoder and decoder which interprets an ASN1_ITEM structure describing
2535 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2536 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2537 has also been converted to the new form.
2540 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2541 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2542 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2543 for negative moduli.
2546 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2547 of not touching the result's sign bit.
2550 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2554 *) Changed the LHASH code to use prototypes for callbacks, and created
2555 macros to declare and implement thin (optionally static) functions
2556 that provide type-safety and avoid function pointer casting for the
2557 type-specific callbacks.
2560 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2562 [Veers Staats <staatsvr@asc.hpc.mil>,
2563 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2565 *) Reformat the FAQ so the different questions and answers can be divided
2566 in sections depending on the subject.
2569 *) Have the zlib compression code load ZLIB.DLL dynamically under
2573 *) New function BN_mod_sqrt for computing square roots modulo a prime
2574 (using the probabilistic Tonelli-Shanks algorithm unless
2575 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2576 be handled deterministically).
2577 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2579 *) Make BN_mod_inverse faster by explicitly handling small quotients
2580 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2581 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2584 *) New function BN_kronecker.
2587 *) Fix BN_gcd so that it works on negative inputs; the result is
2588 positive unless both parameters are zero.
2589 Previously something reasonably close to an infinite loop was
2590 possible because numbers could be growing instead of shrinking
2591 in the implementation of Euclid's algorithm.
2594 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2595 sign of the number in question.
2597 Fix BN_is_word(a,w) to work correctly for w == 0.
2599 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2600 because its test if the absolute value of 'a' equals 'w'.
2601 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2602 it exists mostly for use in the implementations of BN_is_zero(),
2603 BN_is_one(), and BN_is_word().
2606 *) New function BN_swap.
2609 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2610 the exponentiation functions are more likely to produce reasonable
2611 results on negative inputs.
2614 *) Change BN_mod_mul so that the result is always non-negative.
2615 Previously, it could be negative if one of the factors was negative;
2616 I don't think anyone really wanted that behaviour.
2619 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2620 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2621 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2622 and add new functions:
2631 BN_mod_lshift1_quick
2635 These functions always generate non-negative results.
2637 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2638 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2640 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2641 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2642 be reduced modulo m.
2643 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2646 The following entry accidentily appeared in the CHANGES file
2647 distributed with OpenSSL 0.9.7. The modifications described in
2648 it do *not* apply to OpenSSL 0.9.7.
2650 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2651 was actually never needed) and in BN_mul(). The removal in BN_mul()
2652 required a small change in bn_mul_part_recursive() and the addition
2653 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2654 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2655 bn_sub_words() and bn_add_words() except they take arrays with
2660 *) In 'openssl passwd', verify passwords read from the terminal
2661 unless the '-salt' option is used (which usually means that
2662 verification would just waste user's time since the resulting
2663 hash is going to be compared with some given password hash)
2664 or the new '-noverify' option is used.
2666 This is an incompatible change, but it does not affect
2667 non-interactive use of 'openssl passwd' (passwords on the command
2668 line, '-stdin' option, '-in ...' option) and thus should not
2672 *) Remove all references to RSAref, since there's no more need for it.
2675 *) Make DSO load along a path given through an environment variable
2676 (SHLIB_PATH) with shl_load().
2679 *) Constify the ENGINE code as a result of BIGNUM constification.
2680 Also constify the RSA code and most things related to it. In a
2681 few places, most notable in the depth of the ASN.1 code, ugly
2682 casts back to non-const were required (to be solved at a later
2686 *) Make it so the openssl application has all engines loaded by default.
2689 *) Constify the BIGNUM routines a little more.
2692 *) Add the following functions:
2694 ENGINE_load_cswift()
2696 ENGINE_load_atalla()
2698 ENGINE_load_builtin_engines()
2700 That way, an application can itself choose if external engines that
2701 are built-in in OpenSSL shall ever be used or not. The benefit is
2702 that applications won't have to be linked with libdl or other dso
2703 libraries unless it's really needed.
2705 Changed 'openssl engine' to load all engines on demand.
2706 Changed the engine header files to avoid the duplication of some
2707 declarations (they differed!).
2710 *) 'openssl engine' can now list capabilities.
2713 *) Better error reporting in 'openssl engine'.
2716 *) Never call load_dh_param(NULL) in s_server.
2719 *) Add engine application. It can currently list engines by name and
2720 identity, and test if they are actually available.
2723 *) Improve RPM specification file by forcing symbolic linking and making
2724 sure the installed documentation is also owned by root.root.
2725 [Damien Miller <djm@mindrot.org>]
2727 *) Give the OpenSSL applications more possibilities to make use of
2728 keys (public as well as private) handled by engines.
2731 *) Add OCSP code that comes from CertCo.
2734 *) Add VMS support for the Rijndael code.
2737 *) Added untested support for Nuron crypto accelerator.
2740 *) Add support for external cryptographic devices. This code was
2741 previously distributed separately as the "engine" branch.
2742 [Geoff Thorpe, Richard Levitte]
2744 *) Rework the filename-translation in the DSO code. It is now possible to
2745 have far greater control over how a "name" is turned into a filename
2746 depending on the operating environment and any oddities about the
2747 different shared library filenames on each system.
2750 *) Support threads on FreeBSD-elf in Configure.
2753 *) Fix for SHA1 assembly problem with MASM: it produces
2754 warnings about corrupt line number information when assembling
2755 with debugging information. This is caused by the overlapping
2757 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2760 NCONF_get_number() has no error checking at all. As a replacement,
2761 NCONF_get_number_e() is defined (_e for "error checking") and is
2762 promoted strongly. The old NCONF_get_number is kept around for
2763 binary backward compatibility.
2764 Make it possible for methods to load from something other than a BIO,
2765 by providing a function pointer that is given a name instead of a BIO.
2766 For example, this could be used to load configuration data from an
2770 *) Fix for non blocking accept BIOs. Added new I/O special reason
2771 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2772 with non blocking I/O was not possible because no retry code was
2773 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2777 *) Added the beginnings of Rijndael support.
2780 *) Fix for bug in DirectoryString mask setting. Add support for
2781 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2782 to allow certificate printing to more controllable, additional
2783 'certopt' option to 'x509' to allow new printing options to be
2787 *) Clean old EAY MD5 hack from e_os.h.
2790 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2792 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2793 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2794 [Joe Orton, Steve Henson]
2796 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2798 *) Fix additional bug revealed by the NISCC test suite:
2800 Stop bug triggering large recursion when presented with
2801 certain ASN.1 tags (CAN-2003-0851)
2804 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2806 *) Fix various bugs revealed by running the NISCC test suite:
2808 Stop out of bounds reads in the ASN1 code when presented with
2809 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2811 If verify callback ignores invalid public key errors don't try to check
2812 certificate signature with the NULL public key.
2816 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2817 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2821 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2822 extra data after the compression methods not only for TLS 1.0
2823 but also for SSL 3.0 (as required by the specification).
2824 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2826 *) Change X509_certificate_type() to mark the key as exported/exportable
2827 when it's 512 *bits* long, not 512 bytes.
2830 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2832 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2833 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2834 a protocol version number mismatch like a decryption error
2835 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2838 *) Turn on RSA blinding by default in the default implementation
2839 to avoid a timing attack. Applications that don't want it can call
2840 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2841 They would be ill-advised to do so in most cases.
2842 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2844 *) Change RSA blinding code so that it works when the PRNG is not
2845 seeded (in this case, the secret RSA exponent is abused as
2846 an unpredictable seed -- if it is not unpredictable, there
2847 is no point in blinding anyway). Make RSA blinding thread-safe
2848 by remembering the creator's thread ID in rsa->blinding and
2849 having all other threads use local one-time blinding factors
2850 (this requires more computation than sharing rsa->blinding, but
2851 avoids excessive locking; and if an RSA object is not shared
2852 between threads, blinding will still be very fast).
2855 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2857 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2858 via timing by performing a MAC computation even if incorrrect
2859 block cipher padding has been found. This is a countermeasure
2860 against active attacks where the attacker has to distinguish
2861 between bad padding and a MAC verification error. (CAN-2003-0078)
2863 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2864 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2865 Martin Vuagnoux (EPFL, Ilion)]
2867 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2869 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2870 memory from it's contents. This is done with a counter that will
2871 place alternating values in each byte. This can be used to solve
2872 two issues: 1) the removal of calls to memset() by highly optimizing
2873 compilers, and 2) cleansing with other values than 0, since those can
2874 be read through on certain media, for example a swap space on disk.
2877 *) Bugfix: client side session caching did not work with external caching,
2878 because the session->cipher setting was not restored when reloading
2879 from the external cache. This problem was masked, when
2880 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2881 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2884 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2885 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2886 [Zeev Lieber <zeev-l@yahoo.com>]
2888 *) Undo an undocumented change introduced in 0.9.6e which caused
2889 repeated calls to OpenSSL_add_all_ciphers() and
2890 OpenSSL_add_all_digests() to be ignored, even after calling
2894 *) Change the default configuration reader to deal with last line not
2895 being properly terminated.
2898 *) Change X509_NAME_cmp() so it applies the special rules on handling
2899 DN values that are of type PrintableString, as well as RDNs of type
2900 emailAddress where the value has the type ia5String.
2901 [stefank@valicert.com via Richard Levitte]
2903 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
2904 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
2905 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
2906 the bitwise-OR of the two for use by the majority of applications
2907 wanting this behaviour, and update the docs. The documented
2908 behaviour and actual behaviour were inconsistent and had been
2909 changing anyway, so this is more a bug-fix than a behavioural
2911 [Geoff Thorpe, diagnosed by Nadav Har'El]
2913 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
2914 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
2917 *) Fix initialization code race conditions in
2918 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
2919 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
2920 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
2921 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
2922 ssl2_get_cipher_by_char(),
2923 ssl3_get_cipher_by_char().
2924 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
2926 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
2927 the cached sessions are flushed, as the remove_cb() might use ex_data
2928 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
2929 (see [openssl.org #212]).
2930 [Geoff Thorpe, Lutz Jaenicke]
2932 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
2933 length, instead of the encoding length to d2i_ASN1_OBJECT.
2936 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
2938 *) [In 0.9.6g-engine release:]
2939 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
2940 [Lynn Gazis <lgazis@rainbow.com>]
2942 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
2944 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
2945 and get fix the header length calculation.
2946 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
2947 Alon Kantor <alonk@checkpoint.com> (and others),
2950 *) Use proper error handling instead of 'assertions' in buffer
2951 overflow checks added in 0.9.6e. This prevents DoS (the
2952 assertions could call abort()).
2953 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
2955 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
2957 *) Add various sanity checks to asn1_get_length() to reject
2958 the ASN1 length bytes if they exceed sizeof(long), will appear
2959 negative or the content length exceeds the length of the
2961 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2963 *) Fix cipher selection routines: ciphers without encryption had no flags
2964 for the cipher strength set and where therefore not handled correctly
2965 by the selection routines (PR #130).
2968 *) Fix EVP_dsa_sha macro.
2972 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
2973 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
2974 that was added in OpenSSL 0.9.6d.
2976 As the countermeasure turned out to be incompatible with some
2977 broken SSL implementations, the new option is part of SSL_OP_ALL.
2978 SSL_OP_ALL is usually employed when compatibility with weird SSL
2979 implementations is desired (e.g. '-bugs' option to 's_client' and
2980 's_server'), so the new option is automatically set in many
2984 *) Changes in security patch:
2986 Changes marked "(CHATS)" were sponsored by the Defense Advanced
2987 Research Projects Agency (DARPA) and Air Force Research Laboratory,
2988 Air Force Materiel Command, USAF, under agreement number
2991 *) Add various sanity checks to asn1_get_length() to reject
2992 the ASN1 length bytes if they exceed sizeof(long), will appear
2993 negative or the content length exceeds the length of the
2994 supplied buffer. (CAN-2002-0659)
2995 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
2997 *) Assertions for various potential buffer overflows, not known to
2999 [Ben Laurie (CHATS)]
3001 *) Various temporary buffers to hold ASCII versions of integers were
3002 too small for 64 bit platforms. (CAN-2002-0655)
3003 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3005 *) Remote buffer overflow in SSL3 protocol - an attacker could
3006 supply an oversized session ID to a client. (CAN-2002-0656)
3007 [Ben Laurie (CHATS)]
3009 *) Remote buffer overflow in SSL2 protocol - an attacker could
3010 supply an oversized client master key. (CAN-2002-0656)
3011 [Ben Laurie (CHATS)]
3013 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3015 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3016 encoded as NULL) with id-dsa-with-sha1.
3017 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3019 *) Check various X509_...() return values in apps/req.c.
3020 [Nils Larsch <nla@trustcenter.de>]
3022 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3023 an end-of-file condition would erronously be flagged, when the CRLF
3024 was just at the end of a processed block. The bug was discovered when
3025 processing data through a buffering memory BIO handing the data to a
3026 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3027 <ptsekov@syntrex.com> and Nedelcho Stanev.
3030 *) Implement a countermeasure against a vulnerability recently found
3031 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3032 before application data chunks to avoid the use of known IVs
3033 with data potentially chosen by the attacker.
3036 *) Fix length checks in ssl3_get_client_hello().
3039 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3040 to prevent ssl3_read_internal() from incorrectly assuming that
3041 ssl3_read_bytes() found application data while handshake
3042 processing was enabled when in fact s->s3->in_read_app_data was
3043 merely automatically cleared during the initial handshake.
3044 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3046 *) Fix object definitions for Private and Enterprise: they were not
3047 recognized in their shortname (=lowercase) representation. Extend
3048 obj_dat.pl to issue an error when using undefined keywords instead
3049 of silently ignoring the problem (Svenning Sorensen
3050 <sss@sss.dnsalias.net>).
3053 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3054 generators, i.e. generators other than 2 and 5. (Previously, the
3055 code did not properly initialise the 'add' and 'rem' values to
3056 BN_generate_prime().)
3058 In the new general case, we do not insist that 'generator' is
3059 actually a primitive root: This requirement is rather pointless;
3060 a generator of the order-q subgroup is just as good, if not
3064 *) Map new X509 verification errors to alerts. Discovered and submitted by
3065 Tom Wu <tom@arcot.com>.
3068 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3069 returning non-zero before the data has been completely received
3070 when using non-blocking I/O.
3071 [Bodo Moeller; problem pointed out by John Hughes]
3073 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3074 [Ben Laurie, Lutz Jaenicke]
3076 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3077 Yoram Zahavi <YoramZ@gilian.com>).
3080 *) Add information about CygWin 1.3 and on, and preserve proper
3081 configuration for the versions before that.
3082 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3084 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3085 check whether we deal with a copy of a session and do not delete from
3086 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3087 <izhar@checkpoint.com>.
3090 *) Do not store session data into the internal session cache, if it
3091 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3092 flag is set). Proposed by Aslam <aslam@funk.com>.
3095 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3099 *) [In 0.9.6d-engine release:]
3100 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3101 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3103 *) Add the configuration target linux-s390x.
3104 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3106 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3107 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3108 variable as an indication that a ClientHello message has been
3109 received. As the flag value will be lost between multiple
3110 invocations of ssl3_accept when using non-blocking I/O, the
3111 function may not be aware that a handshake has actually taken
3112 place, thus preventing a new session from being added to the
3115 To avoid this problem, we now set s->new_session to 2 instead of
3116 using a local variable.
3117 [Lutz Jaenicke, Bodo Moeller]
3119 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3120 if the SSL_R_LENGTH_MISMATCH error is detected.
3121 [Geoff Thorpe, Bodo Moeller]
3123 *) New 'shared_ldflag' column in Configure platform table.
3126 *) Fix EVP_CIPHER_mode macro.
3127 ["Dan S. Camper" <dan@bti.net>]
3129 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3130 type, we must throw them away by setting rr->length to 0.
3131 [D P Chang <dpc@qualys.com>]
3133 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3135 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3136 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3137 worked incorrectly for those cases where range = 10..._2 and
3138 3*range is two bits longer than range.)
3141 *) Only add signing time to PKCS7 structures if it is not already
3145 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3146 OBJ_ld_ce should be OBJ_id_ce.
3147 Also some ip-pda OIDs in crypto/objects/objects.txt were
3148 incorrect (cf. RFC 3039).
3149 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3151 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3152 returns early because it has nothing to do.
3153 [Andy Schneider <andy.schneider@bjss.co.uk>]
3155 *) [In 0.9.6c-engine release:]
3156 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3157 [Andy Schneider <andy.schneider@bjss.co.uk>]
3159 *) [In 0.9.6c-engine release:]
3160 Add support for Cryptographic Appliance's keyserver technology.
3161 (Use engine 'keyclient')
3162 [Cryptographic Appliances and Geoff Thorpe]
3164 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3165 is called via tools/c89.sh because arguments have to be
3166 rearranged (all '-L' options must appear before the first object
3168 [Richard Shapiro <rshapiro@abinitio.com>]
3170 *) [In 0.9.6c-engine release:]
3171 Add support for Broadcom crypto accelerator cards, backported
3173 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3175 *) [In 0.9.6c-engine release:]
3176 Add support for SureWare crypto accelerator cards from
3177 Baltimore Technologies. (Use engine 'sureware')
3178 [Baltimore Technologies and Mark Cox]
3180 *) [In 0.9.6c-engine release:]
3181 Add support for crypto accelerator cards from Accelerated
3182 Encryption Processing, www.aep.ie. (Use engine 'aep')
3183 [AEP Inc. and Mark Cox]
3185 *) Add a configuration entry for gcc on UnixWare.
3186 [Gary Benson <gbenson@redhat.com>]
3188 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3189 messages are stored in a single piece (fixed-length part and
3190 variable-length part combined) and fix various bugs found on the way.
3193 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3194 instead. BIO_gethostbyname() does not know what timeouts are
3195 appropriate, so entries would stay in cache even when they have
3197 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3199 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3200 faced with a pathologically small ClientHello fragment that does
3201 not contain client_version: Instead of aborting with an error,
3202 simply choose the highest available protocol version (i.e.,
3203 TLS 1.0 unless it is disabled). In practice, ClientHello
3204 messages are never sent like this, but this change gives us
3205 strictly correct behaviour at least for TLS.
3208 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3209 never resets s->method to s->ctx->method when called from within
3210 one of the SSL handshake functions.
3211 [Bodo Moeller; problem pointed out by Niko Baric]
3213 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3214 (sent using the client's version number) if client_version is
3215 smaller than the protocol version in use. Also change
3216 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3217 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3218 the client will at least see that alert.
3221 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3225 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3226 client receives HelloRequest while in a handshake.
3227 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3229 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3230 should end in 'break', not 'goto end' which circuments various
3231 cleanups done in state SSL_ST_OK. But session related stuff
3232 must be disabled for SSL_ST_OK in the case that we just sent a
3235 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3236 before just sending a HelloRequest.
3237 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3239 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3240 reveal whether illegal block cipher padding was found or a MAC
3241 verification error occured. (Neither SSLerr() codes nor alerts
3242 are directly visible to potential attackers, but the information
3243 may leak via logfiles.)
3245 Similar changes are not required for the SSL 2.0 implementation
3246 because the number of padding bytes is sent in clear for SSL 2.0,
3247 and the extra bytes are just ignored. However ssl/s2_pkt.c
3248 failed to verify that the purported number of padding bytes is in
3252 *) Add OpenUNIX-8 support including shared libraries
3253 (Boyd Lynn Gerber <gerberb@zenez.com>).
3256 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3257 'wristwatch attack' using huge encoding parameters (cf.
3258 James H. Manger's CRYPTO 2001 paper). Note that the
3259 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3260 encoding parameters and hence was not vulnerable.
3263 *) BN_sqr() bug fix.
3264 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3266 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3267 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3268 followed by modular reduction.
3269 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3271 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3272 equivalent based on BN_pseudo_rand() instead of BN_rand().
3275 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3276 This function was broken, as the check for a new client hello message
3277 to handle SGC did not allow these large messages.
3278 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3281 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3284 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3285 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3288 *) Rework the configuration and shared library support for Tru64 Unix.
3289 The configuration part makes use of modern compiler features and
3290 still retains old compiler behavior for those that run older versions
3291 of the OS. The shared library support part includes a variant that
3292 uses the RPATH feature, and is available through the special
3293 configuration target "alpha-cc-rpath", which will never be selected
3295 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3297 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3298 with the same message size as in ssl3_get_certificate_request().
3299 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3300 messages might inadvertently be reject as too long.
3301 [Petr Lampa <lampa@fee.vutbr.cz>]
3303 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3306 *) Modified SSL library such that the verify_callback that has been set
3307 specificly for an SSL object with SSL_set_verify() is actually being
3308 used. Before the change, a verify_callback set with this function was
3309 ignored and the verify_callback() set in the SSL_CTX at the time of
3310 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3311 to allow the necessary settings.
3314 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3315 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3316 done automatically (in contradiction to the requirements of the C
3317 standard). This made problems when used from OpenSSH.
3320 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3321 dh->length and always used
3323 BN_rand_range(priv_key, dh->p).
3325 BN_rand_range() is not necessary for Diffie-Hellman, and this
3326 specific range makes Diffie-Hellman unnecessarily inefficient if
3327 dh->length (recommended exponent length) is much smaller than the
3328 length of dh->p. We could use BN_rand_range() if the order of
3329 the subgroup was stored in the DH structure, but we only have
3334 BN_rand(priv_key, l, ...)
3336 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3342 RSA_eay_public_encrypt
3343 RSA_eay_private_decrypt
3344 RSA_eay_private_encrypt (signing)
3345 RSA_eay_public_decrypt (signature verification)
3347 (default implementations for RSA_public_encrypt,
3348 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3349 always reject numbers >= n.
3352 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3353 to synchronize access to 'locking_thread'. This is necessary on
3354 systems where access to 'locking_thread' (an 'unsigned long'
3355 variable) is not atomic.
3358 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3359 *before* setting the 'crypto_lock_rand' flag. The previous code had
3360 a race condition if 0 is a valid thread ID.
3361 [Travis Vitek <vitek@roguewave.com>]
3363 *) Add support for shared libraries under Irix.
3364 [Albert Chin-A-Young <china@thewrittenword.com>]
3366 *) Add configuration option to build on Linux on both big-endian and
3368 [Ralf Baechle <ralf@uni-koblenz.de>]
3370 *) Add the possibility to create shared libraries on HP-UX.
3373 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3375 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3376 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3377 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3378 PRNG state recovery was possible based on the output of
3379 one PRNG request appropriately sized to gain knowledge on
3380 'md' followed by enough consecutive 1-byte PRNG requests
3381 to traverse all of 'state'.
3383 1. When updating 'md_local' (the current thread's copy of 'md')
3384 during PRNG output generation, hash all of the previous
3385 'md_local' value, not just the half used for PRNG output.
3387 2. Make the number of bytes from 'state' included into the hash
3388 independent from the number of PRNG bytes requested.
3390 The first measure alone would be sufficient to avoid
3391 Markku-Juhani's attack. (Actually it had never occurred
3392 to me that the half of 'md_local' used for chaining was the
3393 half from which PRNG output bytes were taken -- I had always
3394 assumed that the secret half would be used.) The second
3395 measure makes sure that additional data from 'state' is never
3396 mixed into 'md_local' in small portions; this heuristically
3397 further strengthens the PRNG.
3400 *) Fix crypto/bn/asm/mips3.s.
3403 *) When only the key is given to "enc", the IV is undefined. Print out
3404 an error message in this case.
3407 *) Handle special case when X509_NAME is empty in X509 printing routines.
3410 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3411 positive and less than q.
3414 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3415 used: it isn't thread safe and the add_lock_callback should handle
3417 [Paul Rose <Paul.Rose@bridge.com>]
3419 *) Verify that incoming data obeys the block size in
3420 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3424 [Ulf Möller, Bodo Möller]
3426 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3427 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3428 when fixing the server behaviour for backwards-compatible 'client
3429 hello' messages. (Note that the attack is impractical against
3430 SSL 3.0 and TLS 1.0 anyway because length and version checking
3431 means that the probability of guessing a valid ciphertext is
3432 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3435 Before 0.9.5, the countermeasure (hide the error by generating a
3436 random 'decryption result') did not work properly because
3437 ERR_clear_error() was missing, meaning that SSL_get_error() would
3438 detect the supposedly ignored error.
3440 Both problems are now fixed.
3443 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3444 (previously it was 1024).
3447 *) Fix for compatibility mode trust settings: ignore trust settings
3448 unless some valid trust or reject settings are present.
3451 *) Fix for blowfish EVP: its a variable length cipher.
3454 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3455 parameters in DSA public key structures and return an error in the
3456 DSA routines if parameters are absent.
3459 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3460 in the current directory if neither $RANDFILE nor $HOME was set.
3461 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3462 caused some confusion to Windows users who haven't defined $HOME.
3463 Thus RAND_file_name() is changed again: e_os.h can define a
3464 DEFAULT_HOME, which will be used if $HOME is not set.
3465 For Windows, we use "C:"; on other platforms, we still require
3466 environment variables.
3468 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3469 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3470 having multiple threads call RAND_poll() concurrently.
3473 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3474 combination of a flag and a thread ID variable.
3475 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3476 flag), *other* threads can enter ssleay_add_bytes without obeying
3477 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3478 that they do not hold after the first thread unsets add_do_not_lock).
3481 *) Change bctest again: '-x' expressions are not available in all
3485 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3487 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3488 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3490 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3491 the default extension for executables, if any. Also, make the perl
3492 scripts that use symlink() to test if it really exists and use "cp"
3493 if it doesn't. All this made OpenSSL compilable and installable in
3497 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3498 If SEQUENCE is length is indefinite just set c->slen to the total
3499 amount of data available.
3500 [Steve Henson, reported by shige@FreeBSD.org]
3501 [This change does not apply to 0.9.7.]
3503 *) Change bctest to avoid here-documents inside command substitution
3504 (workaround for FreeBSD /bin/sh bug).
3505 For compatibility with Ultrix, avoid shell functions (introduced
3506 in the bctest version that searches along $PATH).
3509 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3510 with des_encrypt() defined on some operating systems, like Solaris
3514 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3515 On the Importance of Eliminating Errors in Cryptographic
3516 Computations, J. Cryptology 14 (2001) 2, 101-119,
3517 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3520 *) MIPS assembler BIGNUM division bug fix.
3523 *) Disabled incorrect Alpha assembler code.
3526 *) Fix PKCS#7 decode routines so they correctly update the length
3527 after reading an EOC for the EXPLICIT tag.
3529 [This change does not apply to 0.9.7.]
3531 *) Fix bug in PKCS#12 key generation routines. This was triggered
3532 if a 3DES key was generated with a 0 initial byte. Include
3533 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3534 (but broken) behaviour.
3537 *) Enhance bctest to search for a working bc along $PATH and print
3539 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3541 *) Fix memory leaks in err.c: free err_data string if necessary;
3542 don't write to the wrong index in ERR_set_error_data.
3545 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3549 *) Replace rdtsc with _emit statements for VC++ version 5.
3550 [Jeremy Cooper <jeremy@baymoo.org>]
3552 *) Make it possible to reuse SSLv2 sessions.
3555 *) In copy_email() check for >= 0 as a return value for
3556 X509_NAME_get_index_by_NID() since 0 is a valid index.
3557 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3559 *) Avoid coredump with unsupported or invalid public keys by checking if
3560 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3561 PKCS7_verify() fails with non detached data.
3564 *) Don't use getenv in library functions when run as setuid/setgid.
3565 New function OPENSSL_issetugid().
3568 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3569 due to incorrect handling of multi-threading:
3571 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3573 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3575 3. Count how many times MemCheck_off() has been called so that
3576 nested use can be treated correctly. This also avoids
3577 inband-signalling in the previous code (which relied on the
3578 assumption that thread ID 0 is impossible).
3581 *) Add "-rand" option also to s_client and s_server.
3584 *) Fix CPU detection on Irix 6.x.
3585 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3586 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3588 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3591 [This change does not apply to 0.9.7.]
3593 *) Use the cached encoding of an X509_NAME structure rather than
3594 copying it. This is apparently the reason for the libsafe "errors"
3595 but the code is actually correct.
3598 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3599 Bleichenbacher's DSA attack.
3600 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3601 to be set and top=0 forces the highest bit to be set; top=-1 is new
3602 and leaves the highest bit random.
3603 [Ulf Moeller, Bodo Moeller]
3605 *) In the NCONF_...-based implementations for CONF_... queries
3606 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3607 a temporary CONF structure with the data component set to NULL
3608 (which gives segmentation faults in lh_retrieve).
3609 Instead, use NULL for the CONF pointer in CONF_get_string and
3610 CONF_get_number (which may use environment variables) and directly
3611 return NULL from CONF_get_section.
3614 *) Fix potential buffer overrun for EBCDIC.
3617 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3618 keyUsage if basicConstraints absent for a CA.
3621 *) Make SMIME_write_PKCS7() write mail header values with a format that
3622 is more generally accepted (no spaces before the semicolon), since
3623 some programs can't parse those values properly otherwise. Also make
3624 sure BIO's that break lines after each write do not create invalid
3628 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3629 macros previously used would not encode an empty SEQUENCE OF
3630 and break the signature.
3632 [This change does not apply to 0.9.7.]
3634 *) Zero the premaster secret after deriving the master secret in
3638 *) Add some EVP_add_digest_alias registrations (as found in
3639 OpenSSL_add_all_digests()) to SSL_library_init()
3640 aka OpenSSL_add_ssl_algorithms(). This provides improved
3641 compatibility with peers using X.509 certificates
3642 with unconventional AlgorithmIdentifier OIDs.
3645 *) Fix for Irix with NO_ASM.
3646 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3648 *) ./config script fixes.
3649 [Ulf Moeller, Richard Levitte]
3651 *) Fix 'openssl passwd -1'.
3654 *) Change PKCS12_key_gen_asc() so it can cope with non null
3655 terminated strings whose length is passed in the passlen
3656 parameter, for example from PEM callbacks. This was done
3657 by adding an extra length parameter to asc2uni().
3658 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3660 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3661 call failed, free the DSA structure.
3664 *) Fix to uni2asc() to cope with zero length Unicode strings.
3665 These are present in some PKCS#12 files.
3668 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3669 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3670 when writing a 32767 byte record.
3671 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3673 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3674 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3676 (RSA objects have a reference count access to which is protected
3677 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3678 so they are meant to be shared between threads.)
3679 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3680 "Reddie, Steven" <Steven.Reddie@ca.com>]
3682 *) Fix a deadlock in CRYPTO_mem_leaks().
3685 *) Use better test patterns in bntest.
3688 *) rand_win.c fix for Borland C.
3691 *) BN_rshift bugfix for n == 0.
3694 *) Add a 'bctest' script that checks for some known 'bc' bugs
3695 so that 'make test' does not abort just because 'bc' is broken.
3698 *) Store verify_result within SSL_SESSION also for client side to
3699 avoid potential security hole. (Re-used sessions on the client side
3700 always resulted in verify_result==X509_V_OK, not using the original
3701 result of the server certificate verification.)
3704 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3705 SSL3_RT_APPLICATION_DATA, return 0.
3706 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3710 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3711 releases, have been re-implemented by renaming the previous
3712 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3713 and ssl3_read_internal, respectively, and adding 'peek' parameters
3714 to them. The new ssl[23]_{read,peek} functions are calls to
3715 ssl[23]_read_internal with the 'peek' flag set appropriately.
3716 A 'peek' parameter has also been added to ssl3_read_bytes, which
3717 does the actual work for ssl3_read_internal.
3720 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3721 the method-specific "init()" handler. Also clean up ex_data after
3722 calling the method-specific "finish()" handler. Previously, this was
3723 happening the other way round.
3726 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3727 The previous value, 12, was not always sufficient for BN_mod_exp().
3730 *) Make sure that shared libraries get the internal name engine with
3731 the full version number and not just 0. This should mark the
3732 shared libraries as not backward compatible. Of course, this should
3733 be changed again when we can guarantee backward binary compatibility.
3736 *) Fix typo in get_cert_by_subject() in by_dir.c
3737 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3739 *) Rework the system to generate shared libraries:
3741 - Make note of the expected extension for the shared libraries and
3742 if there is a need for symbolic links from for example libcrypto.so.0
3743 to libcrypto.so.0.9.7. There is extended info in Configure for
3746 - Make as few rebuilds of the shared libraries as possible.
3748 - Still avoid linking the OpenSSL programs with the shared libraries.
3750 - When installing, install the shared libraries separately from the
3754 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3756 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3757 and not in SSL_clear because the latter is also used by the
3758 accept/connect functions; previously, the settings made by
3759 SSL_set_read_ahead would be lost during the handshake.
3760 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3762 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3763 Previously, it would create entries for disableed algorithms no
3767 *) Added several new manual pages for SSL_* function.
3770 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3772 *) In ssl23_get_client_hello, generate an error message when faced
3773 with an initial SSL 3.0/TLS record that is too small to contain the
3774 first two bytes of the ClientHello message, i.e. client_version.
3775 (Note that this is a pathologic case that probably has never happened
3776 in real life.) The previous approach was to use the version number
3777 from the record header as a substitute; but our protocol choice
3778 should not depend on that one because it is not authenticated
3779 by the Finished messages.
3782 *) More robust randomness gathering functions for Windows.
3783 [Jeffrey Altman <jaltman@columbia.edu>]
3785 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3786 not set then we don't setup the error code for issuer check errors
3787 to avoid possibly overwriting other errors which the callback does
3788 handle. If an application does set the flag then we assume it knows
3789 what it is doing and can handle the new informational codes
3793 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3794 a general "ANY" type, as such it should be able to decode anything
3795 including tagged types. However it didn't check the class so it would
3796 wrongly interpret tagged types in the same way as their universal
3797 counterpart and unknown types were just rejected. Changed so that the
3798 tagged and unknown types are handled in the same way as a SEQUENCE:
3799 that is the encoding is stored intact. There is also a new type
3800 "V_ASN1_OTHER" which is used when the class is not universal, in this
3801 case we have no idea what the actual type is so we just lump them all
3805 *) On VMS, stdout may very well lead to a file that is written to
3806 in a record-oriented fashion. That means that every write() will
3807 write a separate record, which will be read separately by the
3808 programs trying to read from it. This can be very confusing.
3810 The solution is to put a BIO filter in the way that will buffer
3811 text until a linefeed is reached, and then write everything a
3812 line at a time, so every record written will be an actual line,
3813 not chunks of lines and not (usually doesn't happen, but I've
3814 seen it once) several lines in one record. BIO_f_linebuffer() is
3817 Currently, it's a VMS-only method, because that's where it has
3818 been tested well enough.
3821 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3822 it can return incorrect results.
3823 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3824 but it was in 0.9.6-beta[12].)
3827 *) Disable the check for content being present when verifying detached
3828 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3829 include zero length content when signing messages.
3832 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3833 BIO_ctrl (for BIO pairs).
3836 *) Add DSO method for VMS.
3839 *) Bug fix: Montgomery multiplication could produce results with the
3843 *) Add RPM specification openssl.spec and modify it to build three
3844 packages. The default package contains applications, application
3845 documentation and run-time libraries. The devel package contains
3846 include files, static libraries and function documentation. The
3847 doc package contains the contents of the doc directory. The original
3848 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3851 *) Add a large number of documentation files for many SSL routines.
3852 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3854 *) Add a configuration entry for Sony News 4.
3855 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3857 *) Don't set the two most significant bits to one when generating a
3858 random number < q in the DSA library.
3861 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3862 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3863 the underlying transport is blocking) if a handshake took place.
3864 (The default behaviour is needed by applications such as s_client
3865 and s_server that use select() to determine when to use SSL_read;
3866 but for applications that know in advance when to expect data, it
3867 just makes things more complicated.)
3870 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3874 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3875 work better on such systems.
3876 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3878 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3879 Update PKCS12_parse() so it copies the friendlyName and the
3880 keyid to the certificates aux info.
3883 *) Fix bug in PKCS7_verify() which caused an infinite loop
3884 if there was more than one signature.
3885 [Sven Uszpelkat <su@celocom.de>]
3887 *) Major change in util/mkdef.pl to include extra information
3888 about each symbol, as well as presentig variables as well
3889 as functions. This change means that there's n more need
3890 to rebuild the .num files when some algorithms are excluded.
3893 *) Allow the verify time to be set by an application,
3894 rather than always using the current time.
3897 *) Phase 2 verify code reorganisation. The certificate
3898 verify code now looks up an issuer certificate by a
3899 number of criteria: subject name, authority key id
3900 and key usage. It also verifies self signed certificates
3901 by the same criteria. The main comparison function is
3902 X509_check_issued() which performs these checks.
3904 Lot of changes were necessary in order to support this
3905 without completely rewriting the lookup code.
3907 Authority and subject key identifier are now cached.
3909 The LHASH 'certs' is X509_STORE has now been replaced
3910 by a STACK_OF(X509_OBJECT). This is mainly because an
3911 LHASH can't store or retrieve multiple objects with
3912 the same hash value.
3914 As a result various functions (which were all internal
3915 use only) have changed to handle the new X509_STORE
3916 structure. This will break anything that messed round
3917 with X509_STORE internally.
3919 The functions X509_STORE_add_cert() now checks for an
3920 exact match, rather than just subject name.
3922 The X509_STORE API doesn't directly support the retrieval
3923 of multiple certificates matching a given criteria, however
3924 this can be worked round by performing a lookup first
3925 (which will fill the cache with candidate certificates)
3926 and then examining the cache for matches. This is probably
3927 the best we can do without throwing out X509_LOOKUP
3928 entirely (maybe later...).
3930 The X509_VERIFY_CTX structure has been enhanced considerably.
3932 All certificate lookup operations now go via a get_issuer()
3933 callback. Although this currently uses an X509_STORE it
3934 can be replaced by custom lookups. This is a simple way
3935 to bypass the X509_STORE hackery necessary to make this
3936 work and makes it possible to use more efficient techniques
3937 in future. A very simple version which uses a simple
3938 STACK for its trusted certificate store is also provided
3939 using X509_STORE_CTX_trusted_stack().
3941 The verify_cb() and verify() callbacks now have equivalents
3942 in the X509_STORE_CTX structure.
3944 X509_STORE_CTX also has a 'flags' field which can be used
3945 to customise the verify behaviour.
3948 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
3949 excludes S/MIME capabilities.
3952 *) When a certificate request is read in keep a copy of the
3953 original encoding of the signed data and use it when outputing
3954 again. Signatures then use the original encoding rather than
3955 a decoded, encoded version which may cause problems if the
3956 request is improperly encoded.
3959 *) For consistency with other BIO_puts implementations, call
3960 buffer_write(b, ...) directly in buffer_puts instead of calling
3963 In BIO_puts, increment b->num_write as in BIO_write.
3964 [Peter.Sylvester@EdelWeb.fr]
3966 *) Fix BN_mul_word for the case where the word is 0. (We have to use
3967 BN_zero, we may not return a BIGNUM with an array consisting of
3971 *) Avoid calling abort() from within the library when problems are
3972 detected, except if preprocessor symbols have been defined
3973 (such as REF_CHECK, BN_DEBUG etc.).
3976 *) New openssl application 'rsautl'. This utility can be
3977 used for low level RSA operations. DER public key
3978 BIO/fp routines also added.
3981 *) New Configure entry and patches for compiling on QNX 4.
3982 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
3984 *) A demo state-machine implementation was sponsored by
3985 Nuron (http://www.nuron.com/) and is now available in
3986 demos/state_machine.
3989 *) New options added to the 'dgst' utility for signature
3990 generation and verification.
3993 *) Unrecognized PKCS#7 content types are now handled via a
3994 catch all ASN1_TYPE structure. This allows unsupported
3995 types to be stored as a "blob" and an application can
3996 encode and decode it manually.
3999 *) Fix various signed/unsigned issues to make a_strex.c
4001 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4003 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4004 length if passed a buffer. ASN1_INTEGER_to_BN failed
4005 if passed a NULL BN and its argument was negative.
4006 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4008 *) Modification to PKCS#7 encoding routines to output definite
4009 length encoding. Since currently the whole structures are in
4010 memory there's not real point in using indefinite length
4011 constructed encoding. However if OpenSSL is compiled with
4012 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4015 *) Added BIO_vprintf() and BIO_vsnprintf().
4018 *) Added more prefixes to parse for in the the strings written
4019 through a logging bio, to cover all the levels that are available
4020 through syslog. The prefixes are now:
4022 PANIC, EMERG, EMR => LOG_EMERG
4023 ALERT, ALR => LOG_ALERT
4024 CRIT, CRI => LOG_CRIT
4025 ERROR, ERR => LOG_ERR
4026 WARNING, WARN, WAR => LOG_WARNING
4027 NOTICE, NOTE, NOT => LOG_NOTICE
4028 INFO, INF => LOG_INFO
4029 DEBUG, DBG => LOG_DEBUG
4031 and as before, if none of those prefixes are present at the
4032 beginning of the string, LOG_ERR is chosen.
4034 On Win32, the LOG_* levels are mapped according to this:
4036 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4037 LOG_WARNING => EVENTLOG_WARNING_TYPE
4038 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4042 *) Made it possible to reconfigure with just the configuration
4043 argument "reconf" or "reconfigure". The command line arguments
4044 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4045 and are retrieved from there when reconfiguring.
4049 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4051 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4054 *) The obj_dat.pl script was messing up the sorting of object
4055 names. The reason was that it compared the quoted version
4056 of strings as a result "OCSP" > "OCSP Signing" because
4057 " > SPACE. Changed script to store unquoted versions of
4058 names and add quotes on output. It was also omitting some
4059 names from the lookup table if they were given a default
4060 value (that is if SN is missing it is given the same
4061 value as LN and vice versa), these are now added on the
4062 grounds that if an object has a name we should be able to
4063 look it up. Finally added warning output when duplicate
4064 short or long names are found.
4067 *) Changes needed for Tandem NSK.
4068 [Scott Uroff <scott@xypro.com>]
4070 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4071 RSA_padding_check_SSLv23(), special padding was never detected
4072 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4073 version rollback attacks was not effective.
4075 In s23_clnt.c, don't use special rollback-attack detection padding
4076 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4077 client; similarly, in s23_srvr.c, don't do the rollback check if
4078 SSL 2.0 is the only protocol enabled in the server.
4081 *) Make it possible to get hexdumps of unprintable data with 'openssl
4082 asn1parse'. By implication, the functions ASN1_parse_dump() and
4083 BIO_dump_indent() are added.
4086 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4087 these print out strings and name structures based on various
4088 flags including RFC2253 support and proper handling of
4089 multibyte characters. Added options to the 'x509' utility
4090 to allow the various flags to be set.
4093 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4094 Also change the functions X509_cmp_current_time() and
4095 X509_gmtime_adj() work with an ASN1_TIME structure,
4096 this will enable certificates using GeneralizedTime in validity
4097 dates to be checked.
4100 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4101 negative public key encodings) on by default,
4102 NO_NEG_PUBKEY_BUG can be set to disable it.
4105 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4106 content octets. An i2c_ASN1_OBJECT is unnecessary because
4107 the encoding can be trivially obtained from the structure.
4110 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4111 not read locks (CRYPTO_r_[un]lock).
4114 *) A first attempt at creating official support for shared
4115 libraries through configuration. I've kept it so the
4116 default is static libraries only, and the OpenSSL programs
4117 are always statically linked for now, but there are
4118 preparations for dynamic linking in place.
4119 This has been tested on Linux and Tru64.
4122 *) Randomness polling function for Win9x, as described in:
4123 Peter Gutmann, Software Generation of Practically Strong
4127 *) Fix so PRNG is seeded in req if using an already existing
4131 *) New options to smime application. -inform and -outform
4132 allow alternative formats for the S/MIME message including
4133 PEM and DER. The -content option allows the content to be
4134 specified separately. This should allow things like Netscape
4135 form signing output easier to verify.
4138 *) Fix the ASN1 encoding of tags using the 'long form'.
4141 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4142 STRING types. These convert content octets to and from the
4143 underlying type. The actual tag and length octets are
4144 already assumed to have been read in and checked. These
4145 are needed because all other string types have virtually
4146 identical handling apart from the tag. By having versions
4147 of the ASN1 functions that just operate on content octets
4148 IMPLICIT tagging can be handled properly. It also allows
4149 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4150 and ASN1_INTEGER are identical apart from the tag.
4153 *) Change the handling of OID objects as follows:
4155 - New object identifiers are inserted in objects.txt, following
4156 the syntax given in objects.README.
4157 - objects.pl is used to process obj_mac.num and create a new
4159 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4162 This is currently kind of a hack, and the perl code in objects.pl
4163 isn't very elegant, but it works as I intended. The simplest way
4164 to check that it worked correctly is to look in obj_dat.h and
4165 check the array nid_objs and make sure the objects haven't moved
4166 around (this is important!). Additions are OK, as well as
4167 consistent name changes.
4170 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4173 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4174 The given file adds to whatever has already been seeded into the
4175 random pool through the RANDFILE configuration file option or
4176 environment variable, or the default random state file.
4179 *) mkstack.pl now sorts each macro group into lexical order.
4180 Previously the output order depended on the order the files
4181 appeared in the directory, resulting in needless rewriting
4185 *) Patches to make OpenSSL compile under Win32 again. Mostly
4186 work arounds for the VC++ problem that it treats func() as
4187 func(void). Also stripped out the parts of mkdef.pl that
4188 added extra typesafe functions: these no longer exist.
4191 *) Reorganisation of the stack code. The macros are now all
4192 collected in safestack.h . Each macro is defined in terms of
4193 a "stack macro" of the form SKM_<name>(type, a, b). The
4194 DEBUG_SAFESTACK is now handled in terms of function casts,
4195 this has the advantage of retaining type safety without the
4196 use of additional functions. If DEBUG_SAFESTACK is not defined
4197 then the non typesafe macros are used instead. Also modified the
4198 mkstack.pl script to handle the new form. Needs testing to see
4199 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4200 the default if no major problems. Similar behaviour for ASN1_SET_OF
4201 and PKCS12_STACK_OF.
4204 *) When some versions of IIS use the 'NET' form of private key the
4205 key derivation algorithm is different. Normally MD5(password) is
4206 used as a 128 bit RC4 key. In the modified case
4207 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4208 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4209 as the old Netscape_RSA functions except they have an additional
4210 'sgckey' parameter which uses the modified algorithm. Also added
4211 an -sgckey command line option to the rsa utility. Thanks to
4212 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4213 algorithm to openssl-dev.
4216 *) The evp_local.h macros were using 'c.##kname' which resulted in
4217 invalid expansion on some systems (SCO 5.0.5 for example).
4218 Corrected to 'c.kname'.
4219 [Phillip Porch <root@theporch.com>]
4221 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4222 a STACK of email addresses from a certificate or request, these look
4223 in the subject name and the subject alternative name extensions and
4224 omit any duplicate addresses.
4227 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4228 This makes DSA verification about 2 % faster.
4231 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4232 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4233 plus overhead for 1024 bit moduli).
4234 This makes exponentiations about 0.5 % faster for 1024 bit
4235 exponents (as measured by "openssl speed rsa2048").
4238 *) Rename memory handling macros to avoid conflicts with other
4240 Malloc => OPENSSL_malloc
4241 Malloc_locked => OPENSSL_malloc_locked
4242 Realloc => OPENSSL_realloc
4243 Free => OPENSSL_free
4246 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4247 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4250 *) CygWin32 support.
4251 [John Jarvie <jjarvie@newsguy.com>]
4253 *) The type-safe stack code has been rejigged. It is now only compiled
4254 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4255 by default all type-specific stack functions are "#define"d back to
4256 standard stack functions. This results in more streamlined output
4257 but retains the type-safety checking possibilities of the original
4261 *) The STACK code has been cleaned up, and certain type declarations
4262 that didn't make a lot of sense have been brought in line. This has
4263 also involved a cleanup of sorts in safestack.h to more correctly
4264 map type-safe stack functions onto their plain stack counterparts.
4265 This work has also resulted in a variety of "const"ifications of
4266 lots of the code, especially "_cmp" operations which should normally
4267 be prototyped with "const" parameters anyway.
4270 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4271 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4272 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4273 where all of 'md' is used each time the PRNG is used, but 'state'
4274 is used only indexed by a cyclic counter. As entropy may not be
4275 well distributed from the beginning, 'md' is important as a
4276 chaining variable. However, the output function chains only half
4277 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4278 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4279 in all of 'state' being rewritten, with the new values depending
4280 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4283 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4284 the handshake is continued after ssl_verify_cert_chain();
4285 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4286 can lead to 'unexplainable' connection aborts later.
4287 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4289 *) Major EVP API cipher revision.
4290 Add hooks for extra EVP features. This allows various cipher
4291 parameters to be set in the EVP interface. Support added for variable
4292 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4293 setting of RC2 and RC5 parameters.
4295 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4298 Remove lots of duplicated code from the EVP library. For example *every*
4299 cipher init() function handles the 'iv' in the same way according to the
4300 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4301 for CFB and OFB modes they zero ctx->num.
4303 New functionality allows removal of S/MIME code RC2 hack.
4305 Most of the routines have the same form and so can be declared in terms
4308 By shifting this to the top level EVP_CipherInit() it can be removed from
4309 all individual ciphers. If the cipher wants to handle IVs or keys
4310 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4313 Change lots of functions like EVP_EncryptUpdate() to now return a
4314 value: although software versions of the algorithms cannot fail
4315 any installed hardware versions can.
4318 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4319 this option is set, tolerate broken clients that send the negotiated
4320 protocol version number instead of the requested protocol version
4324 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4325 i.e. non-zero for export ciphersuites, zero otherwise.
4326 Previous versions had this flag inverted, inconsistent with
4327 rsa_tmp_cb (..._TMP_RSA_CB).
4328 [Bodo Moeller; problem reported by Amit Chopra]
4330 *) Add missing DSA library text string. Work around for some IIS
4331 key files with invalid SEQUENCE encoding.
4334 *) Add a document (doc/standards.txt) that list all kinds of standards
4335 and so on that are implemented in OpenSSL.
4338 *) Enhance c_rehash script. Old version would mishandle certificates
4339 with the same subject name hash and wouldn't handle CRLs at all.
4340 Added -fingerprint option to crl utility, to support new c_rehash
4344 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4347 *) Fix for SSL server purpose checking. Server checking was
4348 rejecting certificates which had extended key usage present
4349 but no ssl client purpose.
4350 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4352 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4353 is a little unclear about how a blank password is handled.
4354 Since the password in encoded as a BMPString with terminating
4355 double NULL a zero length password would end up as just the
4356 double NULL. However no password at all is different and is
4357 handled differently in the PKCS#12 key generation code. NS
4358 treats a blank password as zero length. MSIE treats it as no
4359 password on export: but it will try both on import. We now do
4360 the same: PKCS12_parse() tries zero length and no password if
4361 the password is set to "" or NULL (NULL is now a valid password:
4362 it wasn't before) as does the pkcs12 application.
4365 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4366 perror when PEM_read_bio_X509_REQ fails, the error message must
4367 be obtained from the error queue.
4370 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4371 it in ERR_remove_state if appropriate, and change ERR_get_state
4372 accordingly to avoid race conditions (this is necessary because
4373 thread_hash is no longer constant once set).
4376 *) Bugfix for linux-elf makefile.one.
4379 *) RSA_get_default_method() will now cause a default
4380 RSA_METHOD to be chosen if one doesn't exist already.
4381 Previously this was only set during a call to RSA_new()
4382 or RSA_new_method(NULL) meaning it was possible for
4383 RSA_get_default_method() to return NULL.
4386 *) Added native name translation to the existing DSO code
4387 that will convert (if the flag to do so is set) filenames
4388 that are sufficiently small and have no path information
4389 into a canonical native form. Eg. "blah" converted to
4390 "libblah.so" or "blah.dll" etc.
4393 *) New function ERR_error_string_n(e, buf, len) which is like
4394 ERR_error_string(e, buf), but writes at most 'len' bytes
4395 including the 0 terminator. For ERR_error_string_n, 'buf'
4397 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4399 *) CONF library reworked to become more general. A new CONF
4400 configuration file reader "class" is implemented as well as a
4401 new functions (NCONF_*, for "New CONF") to handle it. The now
4402 old CONF_* functions are still there, but are reimplemented to
4403 work in terms of the new functions. Also, a set of functions
4404 to handle the internal storage of the configuration data is
4405 provided to make it easier to write new configuration file
4406 reader "classes" (I can definitely see something reading a
4407 configuration file in XML format, for example), called _CONF_*,
4408 or "the configuration storage API"...
4410 The new configuration file reading functions are:
4412 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4413 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4415 NCONF_default, NCONF_WIN32
4417 NCONF_dump_fp, NCONF_dump_bio
4419 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4420 NCONF_new creates a new CONF object. This works in the same way
4421 as other interfaces in OpenSSL, like the BIO interface.
4422 NCONF_dump_* dump the internal storage of the configuration file,
4423 which is useful for debugging. All other functions take the same
4424 arguments as the old CONF_* functions wth the exception of the
4425 first that must be a `CONF *' instead of a `LHASH *'.
4427 To make it easer to use the new classes with the old CONF_* functions,
4428 the function CONF_set_default_method is provided.
4431 *) Add '-tls1' option to 'openssl ciphers', which was already
4432 mentioned in the documentation but had not been implemented.
4433 (This option is not yet really useful because even the additional
4434 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4437 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4438 OpenSSL-based applications) load shared libraries and bind to
4439 them in a portable way.
4440 [Geoff Thorpe, with contributions from Richard Levitte]
4442 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4444 *) Make sure _lrotl and _lrotr are only used with MSVC.
4446 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4447 (the default implementation of RAND_status).
4449 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4450 to '-clrext' (= clear extensions), as intended and documented.
4451 [Bodo Moeller; inconsistency pointed out by Michael Attili
4454 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4455 was larger than the MD block size.
4456 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4458 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4459 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4460 using the passed key: if the passed key was a private key the result
4461 of X509_print(), for example, would be to print out all the private key
4465 *) des_quad_cksum() byte order bug fix.
4466 [Ulf Möller, using the problem description in krb4-0.9.7, where
4467 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4469 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4471 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4473 *) For easily testing in shell scripts whether some command
4474 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4475 returns with exit code 0 iff no command of the given name is available.
4476 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4477 the output goes to stdout and nothing is printed to stderr.
4478 Additional arguments are always ignored.
4480 Since for each cipher there is a command of the same name,
4481 the 'no-cipher' compilation switches can be tested this way.
4483 ('openssl no-XXX' is not able to detect pseudo-commands such
4484 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4487 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4490 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4491 is set; it will be thrown away anyway because each handshake creates
4493 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4494 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4495 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4496 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4499 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4500 'Q' and 'R' lose their special meanings (quit/renegotiate).
4501 This is part of what -quiet does; unlike -quiet, -ign_eof
4502 does not suppress any output.
4505 *) Add compatibility options to the purpose and trust code. The
4506 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4507 accepts a certificate or CA, this was the previous behaviour,
4508 with all the associated security issues.
4510 X509_TRUST_COMPAT is the old trust behaviour: only and
4511 automatically trust self signed roots in certificate store. A
4512 new trust setting X509_TRUST_DEFAULT is used to specify that
4513 a purpose has no associated trust setting and it should instead
4514 use the value in the default purpose.
4517 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4518 and fix a memory leak.
4521 *) In util/mkerr.pl (which implements 'make errors'), preserve
4522 reason strings from the previous version of the .c file, as
4523 the default to have only downcase letters (and digits) in
4524 automatically generated reasons codes is not always appropriate.
4527 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4528 using strerror. Previously, ERR_reason_error_string() returned
4529 library names as reason strings for SYSerr; but SYSerr is a special
4530 case where small numbers are errno values, not library numbers.
4533 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4534 converts DSA parameters into DH parameters. (When creating parameters,
4535 DSA_generate_parameters is used.)
4538 *) Include 'length' (recommended exponent length) in C code generated
4539 by 'openssl dhparam -C'.
4542 *) The second argument to set_label in perlasm was already being used
4543 so couldn't be used as a "file scope" flag. Moved to third argument
4547 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4548 instead of RAND_bytes for encryption IVs and salts.
4551 *) Include RAND_status() into RAND_METHOD instead of implementing
4552 it only for md_rand.c Otherwise replacing the PRNG by calling
4553 RAND_set_rand_method would be impossible.
4556 *) Don't let DSA_generate_key() enter an infinite loop if the random
4557 number generation fails.
4560 *) New 'rand' application for creating pseudo-random output.
4563 *) Added configuration support for Linux/IA64
4564 [Rolf Haberrecker <rolf@suse.de>]
4566 *) Assembler module support for Mingw32.
4569 *) Shared library support for HPUX (in shlib/).
4570 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4572 *) Shared library support for Solaris gcc.
4573 [Lutz Behnke <behnke@trustcenter.de>]
4575 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4577 *) PKCS7_encrypt() was adding text MIME headers twice because they
4578 were added manually and by SMIME_crlf_copy().
4581 *) In bntest.c don't call BN_rand with zero bits argument.
4582 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4584 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4585 case was implemented. This caused BN_div_recp() to fail occasionally.
4588 *) Add an optional second argument to the set_label() in the perl
4589 assembly language builder. If this argument exists and is set
4590 to 1 it signals that the assembler should use a symbol whose
4591 scope is the entire file, not just the current function. This
4592 is needed with MASM which uses the format label:: for this scope.
4593 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4595 *) Change the ASN1 types so they are typedefs by default. Before
4596 almost all types were #define'd to ASN1_STRING which was causing
4597 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4601 *) Change names of new functions to the new get1/get0 naming
4602 convention: After 'get1', the caller owns a reference count
4603 and has to call ..._free; 'get0' returns a pointer to some
4604 data structure without incrementing reference counters.
4605 (Some of the existing 'get' functions increment a reference
4606 counter, some don't.)
4607 Similarly, 'set1' and 'add1' functions increase reference
4608 counters or duplicate objects.
4611 *) Allow for the possibility of temp RSA key generation failure:
4612 the code used to assume it always worked and crashed on failure.
4615 *) Fix potential buffer overrun problem in BIO_printf().
4616 [Ulf Möller, using public domain code by Patrick Powell; problem
4617 pointed out by David Sacerdote <das33@cornell.edu>]
4619 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4620 RAND_egd() and RAND_status(). In the command line application,
4621 the EGD socket can be specified like a seed file using RANDFILE
4625 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4626 Some CAs (e.g. Verisign) distribute certificates in this form.
4629 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4630 list to exclude them. This means that no special compilation option
4631 is needed to use anonymous DH: it just needs to be included in the
4635 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4636 EVP_MD_type. The old functionality is available in a new macro called
4637 EVP_MD_md(). Change code that uses it and update docs.
4640 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4641 where the 'void *' argument is replaced by a function pointer argument.
4642 Previously 'void *' was abused to point to functions, which works on
4643 many platforms, but is not correct. As these functions are usually
4644 called by macros defined in OpenSSL header files, most source code
4645 should work without changes.
4648 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4649 sections with information on -D... compiler switches used for
4650 compiling the library so that applications can see them. To enable
4651 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4652 must be defined. E.g.,
4653 #define OPENSSL_ALGORITHM_DEFINES
4654 #include <openssl/opensslconf.h>
4655 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4656 [Richard Levitte, Ulf and Bodo Möller]
4658 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4662 *) Change the 'other' type in certificate aux info to a STACK_OF
4663 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4664 the required ASN1 format: arbitrary types determined by an OID.
4667 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4668 argument to 'req'. This is not because the function is newer or
4669 better than others it just uses the work 'NEW' in the certificate
4670 request header lines. Some software needs this.
4673 *) Reorganise password command line arguments: now passwords can be
4674 obtained from various sources. Delete the PEM_cb function and make
4675 it the default behaviour: i.e. if the callback is NULL and the
4676 usrdata argument is not NULL interpret it as a null terminated pass
4677 phrase. If usrdata and the callback are NULL then the pass phrase
4678 is prompted for as usual.
4681 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4682 the support is automatically enabled. The resulting binaries will
4683 autodetect the card and use it if present.
4684 [Ben Laurie and Compaq Inc.]
4686 *) Work around for Netscape hang bug. This sends certificate request
4687 and server done in one record. Since this is perfectly legal in the
4688 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4689 the bugs/SSLv3 entry for more info.
4692 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4695 *) Add -rand argument to smime and pkcs12 applications and read/write
4699 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4702 *) Add command line password options to the remaining applications.
4705 *) Bug fix for BN_div_recp() for numerators with an even number of
4709 *) More tests in bntest.c, and changed test_bn output.
4712 *) ./config recognizes MacOS X now.
4715 *) Bug fix for BN_div() when the first words of num and divsor are
4716 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4719 *) Add support for various broken PKCS#8 formats, and command line
4720 options to produce them.
4723 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4724 get temporary BIGNUMs from a BN_CTX.
4727 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4731 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4732 include a #define from the old name to the new. The original intent
4733 was that statically linked binaries could for example just call
4734 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4735 link with digests. This never worked becayse SSLeay_add_all_digests()
4736 and SSLeay_add_all_ciphers() were in the same source file so calling
4737 one would link with the other. They are now in separate source files.
4740 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4743 *) Use a less unusual form of the Miller-Rabin primality test (it used
4744 a binary algorithm for exponentiation integrated into the Miller-Rabin
4745 loop, our standard modexp algorithms are faster).
4748 *) Support for the EBCDIC character set completed.
4749 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4751 *) Source code cleanups: use const where appropriate, eliminate casts,
4752 use void * instead of char * in lhash.
4755 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4756 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4757 this the server could overwrite ephemeral keys that the client
4761 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4762 using 50 iterations of the Rabin-Miller test.
4764 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4765 iterations of the Rabin-Miller test as required by the appendix
4766 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4767 As BN_is_prime_fasttest includes trial division, DSA parameter
4768 generation becomes much faster.
4770 This implies a change for the callback functions in DSA_is_prime
4771 and DSA_generate_parameters: The callback function is called once
4772 for each positive witness in the Rabin-Miller test, not just
4773 occasionally in the inner loop; and the parameters to the
4774 callback function now provide an iteration count for the outer
4775 loop rather than for the current invocation of the inner loop.
4776 DSA_generate_parameters additionally can call the callback
4777 function with an 'iteration count' of -1, meaning that a
4778 candidate has passed the trial division test (when q is generated
4779 from an application-provided seed, trial division is skipped).
4782 *) New function BN_is_prime_fasttest that optionally does trial
4783 division before starting the Rabin-Miller test and has
4784 an additional BN_CTX * argument (whereas BN_is_prime always
4785 has to allocate at least one BN_CTX).
4786 'callback(1, -1, cb_arg)' is called when a number has passed the
4787 trial division stage.
4790 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4794 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4797 *) New function BN_pseudo_rand().
4800 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4801 bignum version of BN_from_montgomery() with the working code from
4802 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4806 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4807 made it impossible to use the same SSL_SESSION data structure in
4808 SSL2 clients in multiple threads.
4811 *) The return value of RAND_load_file() no longer counts bytes obtained
4812 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4813 to seed the PRNG (previously an explicit byte count was required).
4814 [Ulf Möller, Bodo Möller]
4816 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4817 used (char *) instead of (void *) and had casts all over the place.
4820 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4823 *) Retain source code compatibility for BN_prime_checks macro:
4824 BN_is_prime(..., BN_prime_checks, ...) now uses
4825 BN_prime_checks_for_size to determine the appropriate number of
4826 Rabin-Miller iterations.
4829 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4830 DH_CHECK_P_NOT_SAFE_PRIME.
4831 (Check if this is true? OpenPGP calls them "strong".)
4834 *) Merge the functionality of "dh" and "gendh" programs into a new program
4835 "dhparam". The old programs are retained for now but will handle DH keys
4836 (instead of parameters) in future.
4839 *) Make the ciphers, s_server and s_client programs check the return values
4840 when a new cipher list is set.
4843 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4844 ciphers. Before when the 56bit ciphers were enabled the sorting was
4847 The syntax for the cipher sorting has been extended to support sorting by
4848 cipher-strength (using the strength_bits hard coded in the tables).
4849 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4851 Fix a bug in the cipher-command parser: when supplying a cipher command
4852 string with an "undefined" symbol (neither command nor alphanumeric
4853 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4854 an error is flagged.
4856 Due to the strength-sorting extension, the code of the
4857 ssl_create_cipher_list() function was completely rearranged. I hope that
4858 the readability was also increased :-)
4859 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4861 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4862 for the first serial number and places 2 in the serial number file. This
4863 avoids problems when the root CA is created with serial number zero and
4864 the first user certificate has the same issuer name and serial number
4868 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4869 the new code. Add documentation for this stuff.
4872 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4873 X509_*() to X509at_*() on the grounds that they don't handle X509
4874 structures and behave in an analagous way to the X509v3 functions:
4875 they shouldn't be called directly but wrapper functions should be used
4878 So we also now have some wrapper functions that call the X509at functions
4879 when passed certificate requests. (TO DO: similar things can be done with
4880 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4881 things. Some of these need some d2i or i2d and print functionality
4882 because they handle more complex structures.)
4885 *) Add missing #ifndefs that caused missing symbols when building libssl
4886 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4887 NO_RSA in ssl/s2*.c.
4888 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4890 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4891 has a return value which indicates the quality of the random data
4892 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4893 error queue. New function RAND_pseudo_bytes() generates output that is
4894 guaranteed to be unique but not unpredictable. RAND_add is like
4895 RAND_seed, but takes an extra argument for an entropy estimate
4896 (RAND_seed always assumes full entropy).
4899 *) Do more iterations of Rabin-Miller probable prime test (specifically,
4900 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
4901 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
4902 in crypto/bn/bn_prime.c for the complete table). This guarantees a
4903 false-positive rate of at most 2^-80 for random input.
4906 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
4909 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
4910 in the 0.9.5 release), this returns the chain
4911 from an X509_CTX structure with a dup of the stack and all
4912 the X509 reference counts upped: so the stack will exist
4913 after X509_CTX_cleanup() has been called. Modify pkcs12.c
4916 Also make SSL_SESSION_print() print out the verify return
4920 *) Add manpage for the pkcs12 command. Also change the default
4921 behaviour so MAC iteration counts are used unless the new
4922 -nomaciter option is used. This improves file security and
4923 only older versions of MSIE (4.0 for example) need it.
4926 *) Honor the no-xxx Configure options when creating .DEF files.
4929 *) Add PKCS#10 attributes to field table: challengePassword,
4930 unstructuredName and unstructuredAddress. These are taken from
4931 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
4932 international characters are used.
4934 More changes to X509_ATTRIBUTE code: allow the setting of types
4935 based on strings. Remove the 'loc' parameter when adding
4936 attributes because these will be a SET OF encoding which is sorted
4940 *) Initial changes to the 'req' utility to allow request generation
4941 automation. This will allow an application to just generate a template
4942 file containing all the field values and have req construct the
4945 Initial support for X509_ATTRIBUTE handling. Stacks of these are
4946 used all over the place including certificate requests and PKCS#7
4947 structures. They are currently handled manually where necessary with
4948 some primitive wrappers for PKCS#7. The new functions behave in a
4949 manner analogous to the X509 extension functions: they allow
4950 attributes to be looked up by NID and added.
4952 Later something similar to the X509V3 code would be desirable to
4953 automatically handle the encoding, decoding and printing of the
4954 more complex types. The string types like challengePassword can
4955 be handled by the string table functions.
4957 Also modified the multi byte string table handling. Now there is
4958 a 'global mask' which masks out certain types. The table itself
4959 can use the flag STABLE_NO_MASK to ignore the mask setting: this
4960 is useful when for example there is only one permissible type
4961 (as in countryName) and using the mask might result in no valid
4965 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
4966 SSL_get_peer_finished to allow applications to obtain the latest
4967 Finished messages sent to the peer or expected from the peer,
4968 respectively. (SSL_get_peer_finished is usually the Finished message
4969 actually received from the peer, otherwise the protocol will be aborted.)
4971 As the Finished message are message digests of the complete handshake
4972 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
4973 be used for external authentication procedures when the authentication
4974 provided by SSL/TLS is not desired or is not enough.
4977 *) Enhanced support for Alpha Linux is added. Now ./config checks if
4978 the host supports BWX extension and if Compaq C is present on the
4979 $PATH. Just exploiting of the BWX extension results in 20-30%
4980 performance kick for some algorithms, e.g. DES and RC4 to mention
4981 a couple. Compaq C in turn generates ~20% faster code for MD5 and
4985 *) Add support for MS "fast SGC". This is arguably a violation of the
4986 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
4987 weak crypto and after checking the certificate is SGC a second one
4988 with strong crypto. MS SGC stops the first handshake after receiving
4989 the server certificate message and sends a second client hello. Since
4990 a server will typically do all the time consuming operations before
4991 expecting any further messages from the client (server key exchange
4992 is the most expensive) there is little difference between the two.
4994 To get OpenSSL to support MS SGC we have to permit a second client
4995 hello message after we have sent server done. In addition we have to
4996 reset the MAC if we do get this second client hello.
4999 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5000 if a DER encoded private key is RSA or DSA traditional format. Changed
5001 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5002 format DER encoded private key. Newer code should use PKCS#8 format which
5003 has the key type encoded in the ASN1 structure. Added DER private key
5004 support to pkcs8 application.
5007 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5008 ciphersuites has been selected (as required by the SSL 3/TLS 1
5009 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5010 is set, we interpret this as a request to violate the specification
5011 (the worst that can happen is a handshake failure, and 'correct'
5012 behaviour would result in a handshake failure anyway).
5015 *) In SSL_CTX_add_session, take into account that there might be multiple
5016 SSL_SESSION structures with the same session ID (e.g. when two threads
5017 concurrently obtain them from an external cache).
5018 The internal cache can handle only one SSL_SESSION with a given ID,
5019 so if there's a conflict, we now throw out the old one to achieve
5023 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5024 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5025 some routines that use cipher OIDs: some ciphers do not have OIDs
5026 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5030 *) Simplify the trust setting structure and code. Now we just have
5031 two sequences of OIDs for trusted and rejected settings. These will
5032 typically have values the same as the extended key usage extension
5033 and any application specific purposes.
5035 The trust checking code now has a default behaviour: it will just
5036 check for an object with the same NID as the passed id. Functions can
5037 be provided to override either the default behaviour or the behaviour
5038 for a given id. SSL client, server and email already have functions
5039 in place for compatibility: they check the NID and also return "trusted"
5040 if the certificate is self signed.
5043 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5044 traditional format into an EVP_PKEY structure.
5047 *) Add a password callback function PEM_cb() which either prompts for
5048 a password if usr_data is NULL or otherwise assumes it is a null
5049 terminated password. Allow passwords to be passed on command line
5050 environment or config files in a few more utilities.
5053 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5054 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5055 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5056 Update documentation.
5059 *) Support for ASN1 "NULL" type. This could be handled before by using
5060 ASN1_TYPE but there wasn't any function that would try to read a NULL
5061 and produce an error if it couldn't. For compatibility we also have
5062 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5063 don't allocate anything because they don't need to.
5066 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5068 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5070 *) Rebuild of the memory allocation routines used by OpenSSL code and
5071 possibly others as well. The purpose is to make an interface that
5072 provide hooks so anyone can build a separate set of allocation and
5073 deallocation routines to be used by OpenSSL, for example memory
5074 pool implementations, or something else, which was previously hard
5075 since Malloc(), Realloc() and Free() were defined as macros having
5076 the values malloc, realloc and free, respectively (except for Win32
5077 compilations). The same is provided for memory debugging code.
5078 OpenSSL already comes with functionality to find memory leaks, but
5079 this gives people a chance to debug other memory problems.
5081 With these changes, a new set of functions and macros have appeared:
5083 CRYPTO_set_mem_debug_functions() [F]
5084 CRYPTO_get_mem_debug_functions() [F]
5085 CRYPTO_dbg_set_options() [F]
5086 CRYPTO_dbg_get_options() [F]
5087 CRYPTO_malloc_debug_init() [M]
5089 The memory debug functions are NULL by default, unless the library
5090 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5091 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5092 gives the standard debugging functions that come with OpenSSL) or
5093 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5094 provided by the library user) must be used. When the standard
5095 debugging functions are used, CRYPTO_dbg_set_options can be used to
5096 request additional information:
5097 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5098 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5100 Also, things like CRYPTO_set_mem_functions will always give the
5101 expected result (the new set of functions is used for allocation
5102 and deallocation) at all times, regardless of platform and compiler
5105 To finish it up, some functions that were never use in any other
5106 way than through macros have a new API and new semantic:
5109 CRYPTO_dbg_realloc()
5112 All macros of value have retained their old syntax.
5113 [Richard Levitte and Bodo Moeller]
5115 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5116 ordering of SMIMECapabilities wasn't in "strength order" and there
5117 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5121 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5122 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5123 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5125 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5126 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5127 functionality to handle multipart/signed properly) and a utility
5128 called 'smime' to call all this stuff. This is based on code I
5129 originally wrote for Celo who have kindly allowed it to be
5130 included in OpenSSL.
5133 *) Add variants des_set_key_checked and des_set_key_unchecked of
5134 des_set_key (aka des_key_sched). Global variable des_check_key
5135 decides which of these is called by des_set_key; this way
5136 des_check_key behaves as it always did, but applications and
5137 the library itself, which was buggy for des_check_key == 1,
5138 have a cleaner way to pick the version they need.
5141 *) New function PKCS12_newpass() which changes the password of a
5145 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5146 dynamic mix. In both cases the ids can be used as an index into the
5147 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5148 functions so they accept a list of the field values and the
5149 application doesn't need to directly manipulate the X509_TRUST
5153 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5157 *) Modify the way the V3 extension code looks up extensions. This now
5158 works in a similar way to the object code: we have some "standard"
5159 extensions in a static table which is searched with OBJ_bsearch()
5160 and the application can add dynamic ones if needed. The file
5161 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5162 updated whenever a new extension is added to the core code and kept
5163 in ext_nid order. There is a simple program 'tabtest.c' which checks
5164 this. New extensions are not added too often so this file can readily
5165 be maintained manually.
5167 There are two big advantages in doing things this way. The extensions
5168 can be looked up immediately and no longer need to be "added" using
5169 X509V3_add_standard_extensions(): this function now does nothing.
5170 [Side note: I get *lots* of email saying the extension code doesn't
5171 work because people forget to call this function]
5172 Also no dynamic allocation is done unless new extensions are added:
5173 so if we don't add custom extensions there is no need to call
5174 X509V3_EXT_cleanup().
5177 *) Modify enc utility's salting as follows: make salting the default. Add a
5178 magic header, so unsalted files fail gracefully instead of just decrypting
5179 to garbage. This is because not salting is a big security hole, so people
5180 should be discouraged from doing it.
5183 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5184 digest to be passed on the command line but it only used this
5185 parameter when signing a certificate. Modified so all relevant
5186 operations are affected by the digest parameter including the
5187 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5188 DSA key was used because it didn't fix the digest.
5191 *) Initial certificate chain verify code. Currently tests the untrusted
5192 certificates for consistency with the verify purpose (which is set
5193 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5195 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5196 this is because it will reject chains with invalid extensions whereas
5197 every previous version of OpenSSL and SSLeay made no checks at all.
5199 Trust code: checks the root CA for the relevant trust settings. Trust
5200 settings have an initial value consistent with the verify purpose: e.g.
5201 if the verify purpose is for SSL client use it expects the CA to be
5202 trusted for SSL client use. However the default value can be changed to
5203 permit custom trust settings: one example of this would be to only trust
5204 certificates from a specific "secure" set of CAs.
5206 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5207 which should be used for version portability: especially since the
5208 verify structure is likely to change more often now.
5210 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5211 to set them. If not set then assume SSL clients will verify SSL servers
5214 Two new options to the verify program: -untrusted allows a set of
5215 untrusted certificates to be passed in and -purpose which sets the
5216 intended purpose of the certificate. If a purpose is set then the
5217 new chain verify code is used to check extension consistency.
5220 *) Support for the authority information access extension.
5223 *) Modify RSA and DSA PEM read routines to transparently handle
5224 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5225 public keys in a format compatible with certificate
5226 SubjectPublicKeyInfo structures. Unfortunately there were already
5227 functions called *_PublicKey_* which used various odd formats so
5228 these are retained for compatibility: however the DSA variants were
5229 never in a public release so they have been deleted. Changed dsa/rsa
5230 utilities to handle the new format: note no releases ever handled public
5231 keys so we should be OK.
5233 The primary motivation for this change is to avoid the same fiasco
5234 that dogs private keys: there are several incompatible private key
5235 formats some of which are standard and some OpenSSL specific and
5236 require various evil hacks to allow partial transparent handling and
5237 even then it doesn't work with DER formats. Given the option anything
5238 other than PKCS#8 should be dumped: but the other formats have to
5239 stay in the name of compatibility.
5241 With public keys and the benefit of hindsight one standard format
5242 is used which works with EVP_PKEY, RSA or DSA structures: though
5243 it clearly returns an error if you try to read the wrong kind of key.
5245 Added a -pubkey option to the 'x509' utility to output the public key.
5246 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5247 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5248 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5249 that do the same as the EVP_PKEY_assign_*() except they up the
5250 reference count of the added key (they don't "swallow" the
5254 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5255 CRLs would fail if the file contained no certificates or no CRLs:
5256 added a new function to read in both types and return the number
5257 read: this means that if none are read it will be an error. The
5258 DER versions of the certificate and CRL reader would always fail
5259 because it isn't possible to mix certificates and CRLs in DER format
5260 without choking one or the other routine. Changed this to just read
5261 a certificate: this is the best we can do. Also modified the code
5262 in apps/verify.c to take notice of return codes: it was previously
5263 attempting to read in certificates from NULL pointers and ignoring
5264 any errors: this is one reason why the cert and CRL reader seemed
5265 to work. It doesn't check return codes from the default certificate
5266 routines: these may well fail if the certificates aren't installed.
5269 *) Code to support otherName option in GeneralName.
5272 *) First update to verify code. Change the verify utility
5273 so it warns if it is passed a self signed certificate:
5274 for consistency with the normal behaviour. X509_verify
5275 has been modified to it will now verify a self signed
5276 certificate if *exactly* the same certificate appears
5277 in the store: it was previously impossible to trust a
5278 single self signed certificate. This means that:
5279 openssl verify ss.pem
5280 now gives a warning about a self signed certificate but
5281 openssl verify -CAfile ss.pem ss.pem
5285 *) For servers, store verify_result in SSL_SESSION data structure
5286 (and add it to external session representation).
5287 This is needed when client certificate verifications fails,
5288 but an application-provided verification callback (set by
5289 SSL_CTX_set_cert_verify_callback) allows accepting the session
5290 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5291 but returns 1): When the session is reused, we have to set
5292 ssl->verify_result to the appropriate error code to avoid
5294 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5296 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5297 case in PKCS7_dataInit() where the signed PKCS7 structure
5298 didn't contain any existing data because it was being created.
5299 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5301 *) Add a salt to the key derivation routines in enc.c. This
5302 forms the first 8 bytes of the encrypted file. Also add a
5303 -S option to allow a salt to be input on the command line.
5306 *) New function X509_cmp(). Oddly enough there wasn't a function
5307 to compare two certificates. We do this by working out the SHA1
5308 hash and comparing that. X509_cmp() will be needed by the trust
5312 *) SSL_get1_session() is like SSL_get_session(), but increments
5313 the reference count in the SSL_SESSION returned.
5314 [Geoff Thorpe <geoff@eu.c2.net>]
5316 *) Fix for 'req': it was adding a null to request attributes.
5317 Also change the X509_LOOKUP and X509_INFO code to handle
5318 certificate auxiliary information.
5321 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5325 *) Add the possibility to add extra information to the memory leak
5326 detecting output, to form tracebacks, showing from where each
5327 allocation was originated: CRYPTO_push_info("constant string") adds
5328 the string plus current file name and line number to a per-thread
5329 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5330 is like calling CYRPTO_pop_info() until the stack is empty.
5331 Also updated memory leak detection code to be multi-thread-safe.
5334 *) Add options -text and -noout to pkcs7 utility and delete the
5335 encryption options which never did anything. Update docs.
5338 *) Add options to some of the utilities to allow the pass phrase
5339 to be included on either the command line (not recommended on
5340 OSes like Unix) or read from the environment. Update the
5341 manpages and fix a few bugs.
5344 *) Add a few manpages for some of the openssl commands.
5347 *) Fix the -revoke option in ca. It was freeing up memory twice,
5348 leaking and not finding already revoked certificates.
5351 *) Extensive changes to support certificate auxiliary information.
5352 This involves the use of X509_CERT_AUX structure and X509_AUX
5353 functions. An X509_AUX function such as PEM_read_X509_AUX()
5354 can still read in a certificate file in the usual way but it
5355 will also read in any additional "auxiliary information". By
5356 doing things this way a fair degree of compatibility can be
5357 retained: existing certificates can have this information added
5358 using the new 'x509' options.
5360 Current auxiliary information includes an "alias" and some trust
5361 settings. The trust settings will ultimately be used in enhanced
5362 certificate chain verification routines: currently a certificate
5363 can only be trusted if it is self signed and then it is trusted
5367 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5368 The problem was that one of the replacement routines had not been working
5369 since SSLeay releases. For now the offending routine has been replaced
5370 with non-optimised assembler. Even so, this now gives around 95%
5371 performance improvement for 1024 bit RSA signs.
5374 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5375 handling. Most clients have the effective key size in bits equal to
5376 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5377 A few however don't do this and instead use the size of the decrypted key
5378 to determine the RC2 key length and the AlgorithmIdentifier to determine
5379 the effective key length. In this case the effective key length can still
5380 be 40 bits but the key length can be 168 bits for example. This is fixed
5381 by manually forcing an RC2 key into the EVP_PKEY structure because the
5382 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5383 the key length and effective key length are equal.
5386 *) Add a bunch of functions that should simplify the creation of
5387 X509_NAME structures. Now you should be able to do:
5388 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5389 and have it automatically work out the correct field type and fill in
5390 the structures. The more adventurous can try:
5391 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5392 and it will (hopefully) work out the correct multibyte encoding.
5395 *) Change the 'req' utility to use the new field handling and multibyte
5396 copy routines. Before the DN field creation was handled in an ad hoc
5397 way in req, ca, and x509 which was rather broken and didn't support
5398 BMPStrings or UTF8Strings. Since some software doesn't implement
5399 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5400 using the dirstring_type option. See the new comment in the default
5401 openssl.cnf for more info.
5404 *) Make crypto/rand/md_rand.c more robust:
5405 - Assure unique random numbers after fork().
5406 - Make sure that concurrent threads access the global counter and
5407 md serializably so that we never lose entropy in them
5408 or use exactly the same state in multiple threads.
5409 Access to the large state is not always serializable because
5410 the additional locking could be a performance killer, and
5411 md should be large enough anyway.
5414 *) New file apps/app_rand.c with commonly needed functionality
5415 for handling the random seed file.
5417 Use the random seed file in some applications that previously did not:
5419 dsaparam -genkey (which also ignored its '-rand' option),
5422 x509 (when signing).
5423 Except on systems with /dev/urandom, it is crucial to have a random
5424 seed file at least for key creation, DSA signing, and for DH exchanges;
5425 for RSA signatures we could do without one.
5427 gendh and gendsa (unlike genrsa) used to read only the first byte
5428 of each file listed in the '-rand' option. The function as previously
5429 found in genrsa is now in app_rand.c and is used by all programs
5430 that support '-rand'.
5433 *) In RAND_write_file, use mode 0600 for creating files;
5434 don't just chmod when it may be too late.
5437 *) Report an error from X509_STORE_load_locations
5438 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5441 *) New function ASN1_mbstring_copy() this copies a string in either
5442 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5443 into an ASN1_STRING type. A mask of permissible types is passed
5444 and it chooses the "minimal" type to use or an error if not type
5448 *) Add function equivalents to the various macros in asn1.h. The old
5449 macros are retained with an M_ prefix. Code inside the library can
5450 use the M_ macros. External code (including the openssl utility)
5451 should *NOT* in order to be "shared library friendly".
5454 *) Add various functions that can check a certificate's extensions
5455 to see if it usable for various purposes such as SSL client,
5456 server or S/MIME and CAs of these types. This is currently
5457 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5458 verification. Also added a -purpose flag to x509 utility to
5459 print out all the purposes.
5462 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5466 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5467 for, obtain and decode and extension and obtain its critical flag.
5468 This allows all the necessary extension code to be handled in a
5469 single function call.
5472 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5473 platforms. See crypto/rc4/rc4_enc.c for further details.
5476 *) New -noout option to asn1parse. This causes no output to be produced
5477 its main use is when combined with -strparse and -out to extract data
5478 from a file (which may not be in ASN.1 format).
5481 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5482 when producing the local key id.
5483 [Richard Levitte <levitte@stacken.kth.se>]
5485 *) New option -dhparam in s_server. This allows a DH parameter file to be
5486 stated explicitly. If it is not stated then it tries the first server
5487 certificate file. The previous behaviour hard coded the filename
5491 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5492 a public key to be input or output. For example:
5493 openssl rsa -in key.pem -pubout -out pubkey.pem
5494 Also added necessary DSA public key functions to handle this.
5497 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5498 in the message. This was handled by allowing
5499 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5500 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5502 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5503 to the end of the strings whereas this didn't. This would cause problems
5504 if strings read with d2i_ASN1_bytes() were later modified.
5505 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5507 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5508 data and it contains EOF it will end up returning an error. This is
5509 caused by input 46 bytes long. The cause is due to the way base64
5510 BIOs find the start of base64 encoded data. They do this by trying a
5511 trial decode on each line until they find one that works. When they
5512 do a flag is set and it starts again knowing it can pass all the
5513 data directly through the decoder. Unfortunately it doesn't reset
5514 the context it uses. This means that if EOF is reached an attempt
5515 is made to pass two EOFs through the context and this causes the
5516 resulting error. This can also cause other problems as well. As is
5517 usual with these problems it takes *ages* to find and the fix is
5518 trivial: move one line.
5519 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5521 *) Ugly workaround to get s_client and s_server working under Windows. The
5522 old code wouldn't work because it needed to select() on sockets and the
5523 tty (for keypresses and to see if data could be written). Win32 only
5524 supports select() on sockets so we select() with a 1s timeout on the
5525 sockets and then see if any characters are waiting to be read, if none
5526 are present then we retry, we also assume we can always write data to
5527 the tty. This isn't nice because the code then blocks until we've
5528 received a complete line of data and it is effectively polling the
5529 keyboard at 1s intervals: however it's quite a bit better than not
5530 working at all :-) A dedicated Windows application might handle this
5531 with an event loop for example.
5534 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5535 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5536 will be called when RSA_sign() and RSA_verify() are used. This is useful
5537 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5538 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5539 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5540 This necessitated the support of an extra signature type NID_md5_sha1
5541 for SSL signatures and modifications to the SSL library to use it instead
5542 of calling RSA_public_decrypt() and RSA_private_encrypt().
5545 *) Add new -verify -CAfile and -CApath options to the crl program, these
5546 will lookup a CRL issuers certificate and verify the signature in a
5547 similar way to the verify program. Tidy up the crl program so it
5548 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5549 less strict. It will now permit CRL extensions even if it is not
5550 a V2 CRL: this will allow it to tolerate some broken CRLs.
5553 *) Initialize all non-automatic variables each time one of the openssl
5554 sub-programs is started (this is necessary as they may be started
5555 multiple times from the "OpenSSL>" prompt).
5556 [Lennart Bang, Bodo Moeller]
5558 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5559 removing all other RSA functionality (this is what NO_RSA does). This
5560 is so (for example) those in the US can disable those operations covered
5561 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5565 *) Non-copying interface to BIO pairs.
5566 (still largely untested)
5569 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5570 ASCII string. This was handled independently in various places before.
5573 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5574 UTF8 strings a character at a time.
5577 *) Use client_version from client hello to select the protocol
5578 (s23_srvr.c) and for RSA client key exchange verification
5579 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5582 *) Add various utility functions to handle SPKACs, these were previously
5583 handled by poking round in the structure internals. Added new function
5584 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5585 print, verify and generate SPKACs. Based on an original idea from
5586 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5589 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5592 *) Allow the config file extension section to be overwritten on the
5593 command line. Based on an original idea from Massimiliano Pala
5594 <madwolf@comune.modena.it>. The new option is called -extensions
5595 and can be applied to ca, req and x509. Also -reqexts to override
5596 the request extensions in req and -crlexts to override the crl extensions
5600 *) Add new feature to the SPKAC handling in ca. Now you can include
5601 the same field multiple times by preceding it by "XXXX." for example:
5604 this is the same syntax as used in the req config file.
5607 *) Allow certificate extensions to be added to certificate requests. These
5608 are specified in a 'req_extensions' option of the req section of the
5609 config file. They can be printed out with the -text option to req but
5610 are otherwise ignored at present.
5613 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5614 data read consists of only the final block it would not decrypted because
5615 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5616 A misplaced 'break' also meant the decrypted final block might not be
5617 copied until the next read.
5620 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5621 a few extra parameters to the DH structure: these will be useful if
5622 for example we want the value of 'q' or implement X9.42 DH.
5625 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5626 provides hooks that allow the default DSA functions or functions on a
5627 "per key" basis to be replaced. This allows hardware acceleration and
5628 hardware key storage to be handled without major modification to the
5629 library. Also added low level modexp hooks and CRYPTO_EX structure and
5630 associated functions.
5633 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5634 as "read only": it can't be written to and the buffer it points to will
5635 not be freed. Reading from a read only BIO is much more efficient than
5636 a normal memory BIO. This was added because there are several times when
5637 an area of memory needs to be read from a BIO. The previous method was
5638 to create a memory BIO and write the data to it, this results in two
5639 copies of the data and an O(n^2) reading algorithm. There is a new
5640 function BIO_new_mem_buf() which creates a read only memory BIO from
5641 an area of memory. Also modified the PKCS#7 routines to use read only
5645 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5646 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5647 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5648 but a retry condition occured while trying to read the rest.
5651 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5652 NID_pkcs7_encrypted by default: this was wrong since this should almost
5653 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5654 the encrypted data type: this is a more sensible place to put it and it
5655 allows the PKCS#12 code to be tidied up that duplicated this
5659 *) Changed obj_dat.pl script so it takes its input and output files on
5660 the command line. This should avoid shell escape redirection problems
5664 *) Initial support for certificate extension requests, these are included
5665 in things like Xenroll certificate requests. Included functions to allow
5666 extensions to be obtained and added.
5669 *) -crlf option to s_client and s_server for sending newlines as
5670 CRLF (as required by many protocols).
5673 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5675 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5676 [Ralf S. Engelschall]
5678 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5679 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5681 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5685 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5686 DH parameters/keys (q is lost during that conversion, but the resulting
5687 DH parameters contain its length).
5689 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5690 much faster than DH_generate_parameters (which creates parameters
5691 where p = 2*q + 1), and also the smaller q makes DH computations
5692 much more efficient (160-bit exponentiation instead of 1024-bit
5693 exponentiation); so this provides a convenient way to support DHE
5694 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5695 utter importance to use
5696 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5698 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5699 when such DH parameters are used, because otherwise small subgroup
5700 attacks may become possible!
5703 *) Avoid memory leak in i2d_DHparams.
5706 *) Allow the -k option to be used more than once in the enc program:
5707 this allows the same encrypted message to be read by multiple recipients.
5710 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5711 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5712 it will always use the numerical form of the OID, even if it has a short
5716 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5717 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5718 otherwise bn_mod_exp was called. In the case of hardware keys for example
5719 no private key components need be present and it might store extra data
5720 in the RSA structure, which cannot be accessed from bn_mod_exp.
5721 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5722 private key operations.
5725 *) Added support for SPARC Linux.
5728 *) pem_password_cb function type incompatibly changed from
5729 typedef int pem_password_cb(char *buf, int size, int rwflag);
5731 ....(char *buf, int size, int rwflag, void *userdata);
5732 so that applications can pass data to their callbacks:
5733 The PEM[_ASN1]_{read,write}... functions and macros now take an
5734 additional void * argument, which is just handed through whenever
5735 the password callback is called.
5736 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5738 New function SSL_CTX_set_default_passwd_cb_userdata.
5740 Compatibility note: As many C implementations push function arguments
5741 onto the stack in reverse order, the new library version is likely to
5742 interoperate with programs that have been compiled with the old
5743 pem_password_cb definition (PEM_whatever takes some data that
5744 happens to be on the stack as its last argument, and the callback
5745 just ignores this garbage); but there is no guarantee whatsoever that
5748 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5749 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5750 problems not only on Windows, but also on some Unix platforms.
5751 To avoid problematic command lines, these definitions are now in an
5752 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5753 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5756 *) MIPS III/IV assembler module is reimplemented.
5759 *) More DES library cleanups: remove references to srand/rand and
5760 delete an unused file.
5763 *) Add support for the the free Netwide assembler (NASM) under Win32,
5764 since not many people have MASM (ml) and it can be hard to obtain.
5765 This is currently experimental but it seems to work OK and pass all
5766 the tests. Check out INSTALL.W32 for info.
5769 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5770 without temporary keys kept an extra copy of the server key,
5771 and connections with temporary keys did not free everything in case
5775 *) New function RSA_check_key and new openssl rsa option -check
5776 for verifying the consistency of RSA keys.
5777 [Ulf Moeller, Bodo Moeller]
5779 *) Various changes to make Win32 compile work:
5780 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5781 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5782 comparison" warnings.
5783 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5786 *) Add a debugging option to PKCS#5 v2 key generation function: when
5787 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5788 derived keys are printed to stderr.
5791 *) Copy the flags in ASN1_STRING_dup().
5792 [Roman E. Pavlov <pre@mo.msk.ru>]
5794 *) The x509 application mishandled signing requests containing DSA
5795 keys when the signing key was also DSA and the parameters didn't match.
5797 It was supposed to omit the parameters when they matched the signing key:
5798 the verifying software was then supposed to automatically use the CA's
5799 parameters if they were absent from the end user certificate.
5801 Omitting parameters is no longer recommended. The test was also
5802 the wrong way round! This was probably due to unusual behaviour in
5803 EVP_cmp_parameters() which returns 1 if the parameters match.
5804 This meant that parameters were omitted when they *didn't* match and
5805 the certificate was useless. Certificates signed with 'ca' didn't have
5807 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5809 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5810 The interface is as follows:
5811 Applications can use
5812 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5813 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5814 "off" is now the default.
5815 The library internally uses
5816 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5817 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5818 to disable memory-checking temporarily.
5820 Some inconsistent states that previously were possible (and were
5821 even the default) are now avoided.
5823 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5824 with each memory chunk allocated; this is occasionally more helpful
5825 than just having a counter.
5827 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5829 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5833 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5834 which largely parallels "options", but is for changing API behaviour,
5835 whereas "options" are about protocol behaviour.
5836 Initial "mode" flags are:
5838 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5839 a single record has been written.
5840 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5841 retries use the same buffer location.
5842 (But all of the contents must be
5846 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5849 *) Fix problems with no-hmac etc.
5850 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5852 *) New functions RSA_get_default_method(), RSA_set_method() and
5853 RSA_get_method(). These allows replacement of RSA_METHODs without having
5854 to mess around with the internals of an RSA structure.
5857 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5858 Also really enable memory leak checks in openssl.c and in some
5860 [Chad C. Mulligan, Bodo Moeller]
5862 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5863 up the length of negative integers. This has now been simplified to just
5864 store the length when it is first determined and use it later, rather
5865 than trying to keep track of where data is copied and updating it to
5867 [Steve Henson, reported by Brien Wheeler
5868 <bwheeler@authentica-security.com>]
5870 *) Add a new function PKCS7_signatureVerify. This allows the verification
5871 of a PKCS#7 signature but with the signing certificate passed to the
5872 function itself. This contrasts with PKCS7_dataVerify which assumes the
5873 certificate is present in the PKCS#7 structure. This isn't always the
5874 case: certificates can be omitted from a PKCS#7 structure and be
5875 distributed by "out of band" means (such as a certificate database).
5878 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5879 function prototypes in pem.h, also change util/mkdef.pl to add the
5880 necessary function names.
5883 *) mk1mf.pl (used by Windows builds) did not properly read the
5884 options set by Configure in the top level Makefile, and Configure
5885 was not even able to write more than one option correctly.
5886 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5889 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5890 file to be loaded from a BIO or FILE pointer. The BIO version will
5891 for example allow memory BIOs to contain config info.
5894 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
5895 Whoever hopes to achieve shared-library compatibility across versions
5896 must use this, not the compile-time macro.
5897 (Exercise 0.9.4: Which is the minimum library version required by
5899 Note: All this applies only to multi-threaded programs, others don't
5903 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
5904 through a BIO pair triggered the default case, i.e.
5905 SSLerr(...,SSL_R_UNKNOWN_STATE).
5908 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
5909 can use the SSL library even if none of the specific BIOs is
5913 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
5914 for the encoded length.
5915 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
5917 *) Add initial documentation of the X509V3 functions.
5920 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
5921 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
5922 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
5923 secure PKCS#8 private key format with a high iteration count.
5926 *) Fix determination of Perl interpreter: A perl or perl5
5927 _directory_ in $PATH was also accepted as the interpreter.
5928 [Ralf S. Engelschall]
5930 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
5931 wrong with it but it was very old and did things like calling
5932 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
5936 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
5937 to use the new extension code.
5940 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
5941 with macros. This should make it easier to change their form, add extra
5942 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
5946 *) Add to configuration table a new entry that can specify an alternative
5947 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
5948 according to Mark Crispin <MRC@Panda.COM>.
5952 *) DES CBC did not update the IV. Weird.
5955 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
5956 Changing the behaviour of the former might break existing programs --
5957 where IV updating is needed, des_ncbc_encrypt can be used.
5960 *) When bntest is run from "make test" it drives bc to check its
5961 calculations, as well as internally checking them. If an internal check
5962 fails, it needs to cause bc to give a non-zero result or make test carries
5963 on without noticing the failure. Fixed.
5966 *) DES library cleanups.
5969 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
5970 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
5971 ciphers. NOTE: although the key derivation function has been verified
5972 against some published test vectors it has not been extensively tested
5973 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
5977 *) Instead of "mkdir -p", which is not fully portable, use new
5978 Perl script "util/mkdir-p.pl".
5981 *) Rewrite the way password based encryption (PBE) is handled. It used to
5982 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
5983 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
5984 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
5985 the 'parameter' field of the AlgorithmIdentifier is passed to the
5986 underlying key generation function so it must do its own ASN1 parsing.
5987 This has also changed the EVP_PBE_CipherInit() function which now has a
5988 'parameter' argument instead of literal salt and iteration count values
5989 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
5992 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
5993 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
5994 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
5995 KEY" because this clashed with PKCS#8 unencrypted string. Since this
5996 value was just used as a "magic string" and not used directly its
5997 value doesn't matter.
6000 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6004 *) "linux-sparc64" configuration (ultrapenguin).
6005 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6006 "linux-sparc" configuration.
6007 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6009 *) config now generates no-xxx options for missing ciphers.
6012 *) Support the EBCDIC character set (work in progress).
6013 File ebcdic.c not yet included because it has a different license.
6014 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6016 *) Support BS2000/OSD-POSIX.
6017 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6019 *) Make callbacks for key generation use void * instead of char *.
6022 *) Make S/MIME samples compile (not yet tested).
6025 *) Additional typesafe stacks.
6028 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6032 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6034 *) New configuration variant "sco5-gcc".
6036 *) Updated some demos.
6037 [Sean O Riordain, Wade Scholine]
6039 *) Add missing BIO_free at exit of pkcs12 application.
6042 *) Fix memory leak in conf.c.
6045 *) Updates for Win32 to assembler version of MD5.
6048 *) Set #! path to perl in apps/der_chop to where we found it
6049 instead of using a fixed path.
6052 *) SHA library changes for irix64-mips4-cc.
6055 *) Improvements for VMS support.
6059 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6061 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6062 This also avoids the problems with SC4.2 and unpatched SC5.
6063 [Andy Polyakov <appro@fy.chalmers.se>]
6065 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6066 These are required because of the typesafe stack would otherwise break
6067 existing code. If old code used a structure member which used to be STACK
6068 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6069 sk_num or sk_value it would produce an error because the num, data members
6070 are not present in STACK_OF. Now it just produces a warning. sk_set
6071 replaces the old method of assigning a value to sk_value
6072 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6073 that does this will no longer work (and should use sk_set instead) but
6074 this could be regarded as a "questionable" behaviour anyway.
6077 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6078 correctly handle encrypted S/MIME data.
6081 *) Change type of various DES function arguments from des_cblock
6082 (which means, in function argument declarations, pointer to char)
6083 to des_cblock * (meaning pointer to array with 8 char elements),
6084 which allows the compiler to do more typechecking; it was like
6085 that back in SSLeay, but with lots of ugly casts.
6087 Introduce new type const_des_cblock.
6090 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6091 problems: find RecipientInfo structure that matches recipient certificate
6092 and initialise the ASN1 structures properly based on passed cipher.
6095 *) Belatedly make the BN tests actually check the results.
6098 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6099 to and from BNs: it was completely broken. New compilation option
6100 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6101 key elements as negative integers.
6104 *) Reorganize and speed up MD5.
6105 [Andy Polyakov <appro@fy.chalmers.se>]
6108 [Richard Levitte <richard@levitte.org>]
6110 *) New option -out to asn1parse to allow the parsed structure to be
6111 output to a file. This is most useful when combined with the -strparse
6112 option to examine the output of things like OCTET STRINGS.
6115 *) Make SSL library a little more fool-proof by not requiring any longer
6116 that SSL_set_{accept,connect}_state be called before
6117 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6118 in many applications because usually everything *appeared* to work as
6119 intended anyway -- now it really works as intended).
6122 *) Move openssl.cnf out of lib/.
6125 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6126 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6127 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6128 [Ralf S. Engelschall]
6130 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6131 handle PKCS#7 enveloped data properly.
6132 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6134 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6135 copying pointers. The cert_st handling is changed by this in
6136 various ways (and thus what used to be known as ctx->default_cert
6137 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6138 any longer when s->cert does not give us what we need).
6139 ssl_cert_instantiate becomes obsolete by this change.
6140 As soon as we've got the new code right (possibly it already is?),
6141 we have solved a couple of bugs of the earlier code where s->cert
6142 was used as if it could not have been shared with other SSL structures.
6144 Note that using the SSL API in certain dirty ways now will result
6145 in different behaviour than observed with earlier library versions:
6146 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6147 does not influence s as it used to.
6149 In order to clean up things more thoroughly, inside SSL_SESSION
6150 we don't use CERT any longer, but a new structure SESS_CERT
6151 that holds per-session data (if available); currently, this is
6152 the peer's certificate chain and, for clients, the server's certificate
6153 and temporary key. CERT holds only those values that can have
6154 meaningful defaults in an SSL_CTX.
6157 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6158 from the internal representation. Various PKCS#7 fixes: remove some
6159 evil casts and set the enc_dig_alg field properly based on the signing
6163 *) Allow PKCS#12 password to be set from the command line or the
6164 environment. Let 'ca' get its config file name from the environment
6165 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6169 *) Allow certificate policies extension to use an IA5STRING for the
6170 organization field. This is contrary to the PKIX definition but
6171 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6175 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6176 without disallowing inline assembler and the like for non-pedantic builds.
6179 *) Support Borland C++ builder.
6180 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6185 *) SHA-1 cleanups and performance enhancements.
6186 [Andy Polyakov <appro@fy.chalmers.se>]
6188 *) Sparc v8plus assembler for the bignum library.
6189 [Andy Polyakov <appro@fy.chalmers.se>]
6191 *) Accept any -xxx and +xxx compiler options in Configure.
6194 *) Update HPUX configuration.
6197 *) Add missing sk_<type>_unshift() function to safestack.h
6198 [Ralf S. Engelschall]
6200 *) New function SSL_CTX_use_certificate_chain_file that sets the
6201 "extra_cert"s in addition to the certificate. (This makes sense
6202 only for "PEM" format files, as chains as a whole are not
6206 *) Support verify_depth from the SSL API.
6207 x509_vfy.c had what can be considered an off-by-one-error:
6208 Its depth (which was not part of the external interface)
6209 was actually counting the number of certificates in a chain;
6210 now it really counts the depth.
6213 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6214 instead of X509err, which often resulted in confusing error
6215 messages since the error codes are not globally unique
6216 (e.g. an alleged error in ssl3_accept when a certificate
6217 didn't match the private key).
6219 *) New function SSL_CTX_set_session_id_context that allows to set a default
6220 value (so that you don't need SSL_set_session_id_context for each
6221 connection using the SSL_CTX).
6224 *) OAEP decoding bug fix.
6227 *) Support INSTALL_PREFIX for package builders, as proposed by
6231 *) New Configure options "threads" and "no-threads". For systems
6232 where the proper compiler options are known (currently Solaris
6233 and Linux), "threads" is the default.
6236 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6239 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6240 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6241 such as /usr/local/bin.
6244 *) "make linux-shared" to build shared libraries.
6245 [Niels Poppe <niels@netbox.org>]
6247 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6250 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6251 extension adding in x509 utility.
6254 *) Remove NOPROTO sections and error code comments.
6257 *) Partial rewrite of the DEF file generator to now parse the ANSI
6261 *) New Configure options --prefix=DIR and --openssldir=DIR.
6264 *) Complete rewrite of the error code script(s). It is all now handled
6265 by one script at the top level which handles error code gathering,
6266 header rewriting and C source file generation. It should be much better
6267 than the old method: it now uses a modified version of Ulf's parser to
6268 read the ANSI prototypes in all header files (thus the old K&R definitions
6269 aren't needed for error creation any more) and do a better job of
6270 translating function codes into names. The old 'ASN1 error code imbedded
6271 in a comment' is no longer necessary and it doesn't use .err files which
6272 have now been deleted. Also the error code call doesn't have to appear all
6273 on one line (which resulted in some large lines...).
6276 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6279 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6280 0 (which usually indicates a closed connection), but continue reading.
6283 *) Fix some race conditions.
6286 *) Add support for CRL distribution points extension. Add Certificate
6287 Policies and CRL distribution points documentation.
6290 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6293 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6294 8 of keying material. Merlin has also confirmed interop with this fix
6295 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6296 [Merlin Hughes <merlin@baltimore.ie>]
6298 *) Fix lots of warnings.
6299 [Richard Levitte <levitte@stacken.kth.se>]
6301 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6302 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6303 [Richard Levitte <levitte@stacken.kth.se>]
6305 *) Fix problems with sizeof(long) == 8.
6306 [Andy Polyakov <appro@fy.chalmers.se>]
6308 *) Change functions to ANSI C.
6311 *) Fix typos in error codes.
6312 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6314 *) Remove defunct assembler files from Configure.
6317 *) SPARC v8 assembler BIGNUM implementation.
6318 [Andy Polyakov <appro@fy.chalmers.se>]
6320 *) Support for Certificate Policies extension: both print and set.
6321 Various additions to support the r2i method this uses.
6324 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6325 return a const string when you are expecting an allocated buffer.
6328 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6329 types DirectoryString and DisplayText.
6332 *) Add code to allow r2i extensions to access the configuration database,
6333 add an LHASH database driver and add several ctx helper functions.
6336 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6337 fail when they extended the size of a BIGNUM.
6340 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6341 support typesafe stack.
6344 *) Fix typo in SSL_[gs]et_options().
6345 [Nils Frostberg <nils@medcom.se>]
6347 *) Delete various functions and files that belonged to the (now obsolete)
6348 old X509V3 handling code.
6351 *) New Configure option "rsaref".
6354 *) Don't auto-generate pem.h.
6357 *) Introduce type-safe ASN.1 SETs.
6360 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6361 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6363 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6364 that links with OpenSSL (well at least cause lots of warnings), but fear
6365 not: the conversion is trivial, and it eliminates loads of evil casts. A
6366 few STACKed things have been converted already. Feel free to convert more.
6367 In the fullness of time, I'll do away with the STACK type altogether.
6370 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6371 specified in <certfile> by updating the entry in the index.txt file.
6372 This way one no longer has to edit the index.txt file manually for
6373 revoking a certificate. The -revoke option does the gory details now.
6374 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6376 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6377 `-text' option at all and this way the `-noout -text' combination was
6378 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6379 [Ralf S. Engelschall]
6381 *) Make sure a corresponding plain text error message exists for the
6382 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6383 verify callback function determined that a certificate was revoked.
6384 [Ralf S. Engelschall]
6386 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6387 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6388 all available cipers including rc5, which was forgotten until now.
6389 In order to let the testing shell script know which algorithms
6390 are available, a new (up to now undocumented) command
6391 "openssl list-cipher-commands" is used.
6394 *) Bugfix: s_client occasionally would sleep in select() when
6395 it should have checked SSL_pending() first.
6398 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6399 the raw DSA values prior to ASN.1 encoding.
6402 *) Tweaks to Configure
6403 [Niels Poppe <niels@netbox.org>]
6405 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6409 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6412 *) New config option to avoid instructions that are illegal on the 80386.
6413 The default code is faster, but requires at least a 486.
6416 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6417 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6418 same as SSL2_VERSION anyway.
6421 *) New "-showcerts" option for s_client.
6424 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6425 application. Various cleanups and fixes.
6428 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6429 modify error routines to work internally. Add error codes and PBE init
6430 to library startup routines.
6433 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6434 packing functions to asn1 and evp. Changed function names and error
6435 codes along the way.
6438 *) PKCS12 integration: and so it begins... First of several patches to
6439 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6440 objects to objects.h
6443 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6444 and display support for Thawte strong extranet extension.
6447 *) Add LinuxPPC support.
6448 [Jeff Dubrule <igor@pobox.org>]
6450 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6451 bn_div_words in alpha.s.
6452 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6454 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6455 OAEP isn't supported when OpenSSL is built with RSAref.
6456 [Ulf Moeller <ulf@fitug.de>]
6458 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6459 so they no longer are missing under -DNOPROTO.
6460 [Soren S. Jorvang <soren@t.dk>]
6463 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6465 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6466 doesn't work when the session is reused. Coming soon!
6469 *) Fix a security hole, that allows sessions to be reused in the wrong
6470 context thus bypassing client cert protection! All software that uses
6471 client certs and session caches in multiple contexts NEEDS PATCHING to
6472 allow session reuse! A fuller solution is in the works.
6473 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6475 *) Some more source tree cleanups (removed obsolete files
6476 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6477 permission on "config" script to be executable) and a fix for the INSTALL
6479 [Ulf Moeller <ulf@fitug.de>]
6481 *) Remove some legacy and erroneous uses of malloc, free instead of
6483 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6485 *) Make rsa_oaep_test return non-zero on error.
6486 [Ulf Moeller <ulf@fitug.de>]
6488 *) Add support for native Solaris shared libraries. Configure
6489 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6490 if someone would make that last step automatic.
6491 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6493 *) ctx_size was not built with the right compiler during "make links". Fixed.
6496 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6497 except NULL ciphers". This means the default cipher list will no longer
6498 enable NULL ciphers. They need to be specifically enabled e.g. with
6499 the string "DEFAULT:eNULL".
6502 *) Fix to RSA private encryption routines: if p < q then it would
6503 occasionally produce an invalid result. This will only happen with
6504 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6507 *) Be less restrictive and allow also `perl util/perlpath.pl
6508 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6509 because this way one can also use an interpreter named `perl5' (which is
6510 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6511 installed as `perl').
6512 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6514 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6515 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6517 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6518 advapi32.lib to Win32 build and change the pem test comparision
6519 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6520 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6521 and crypto/des/ede_cbcm_enc.c.
6524 *) DES quad checksum was broken on big-endian architectures. Fixed.
6527 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6528 Win32 test batch file so it (might) work again. The Win32 test batch file
6529 is horrible: I feel ill....
6532 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6533 in e_os.h. Audit of header files to check ANSI and non ANSI
6534 sections: 10 functions were absent from non ANSI section and not exported
6535 from Windows DLLs. Fixed up libeay.num for new functions.
6538 *) Make `openssl version' output lines consistent.
6539 [Ralf S. Engelschall]
6541 *) Fix Win32 symbol export lists for BIO functions: Added
6542 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6543 to ms/libeay{16,32}.def.
6544 [Ralf S. Engelschall]
6546 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6547 fine under Unix and passes some trivial tests I've now added. But the
6548 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6549 added to make sure no one expects that this stuff really works in the
6550 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6551 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6553 [Ralf S. Engelschall]
6555 *) Fix the generation of two part addresses in perl.
6556 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6558 *) Add config entry for Linux on MIPS.
6559 [John Tobey <jtobey@channel1.com>]
6561 *) Make links whenever Configure is run, unless we are on Windoze.
6564 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6565 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6569 *) Add a useful kludge to allow package maintainers to specify compiler and
6570 other platforms details on the command line without having to patch the
6571 Configure script everytime: One now can use ``perl Configure
6572 <id>:<details>'', i.e. platform ids are allowed to have details appended
6573 to them (seperated by colons). This is treated as there would be a static
6574 pre-configured entry in Configure's %table under key <id> with value
6575 <details> and ``perl Configure <id>'' is called. So, when you want to
6576 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6577 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6578 now, which overrides the FreeBSD-elf entry on-the-fly.
6579 [Ralf S. Engelschall]
6581 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6584 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6585 on the `perl Configure ...' command line. This way one can compile
6586 OpenSSL libraries with Position Independent Code (PIC) which is needed
6587 for linking it into DSOs.
6588 [Ralf S. Engelschall]
6590 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6594 *) Cleaned up the LICENSE document: The official contact for any license
6595 questions now is the OpenSSL core team under openssl-core@openssl.org.
6596 And add a paragraph about the dual-license situation to make sure people
6597 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6598 to the OpenSSL toolkit.
6599 [Ralf S. Engelschall]
6601 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6602 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6603 Additonally cleaned up the `make links' target: Remove unnecessary
6604 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6605 to speed processing and no longer clutter the display with confusing
6606 stuff. Instead only the actually done links are displayed.
6607 [Ralf S. Engelschall]
6609 *) Permit null encryption ciphersuites, used for authentication only. It used
6610 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6611 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6615 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6616 signed attributes when verifying signatures (this would break them),
6617 the detached data encoding was wrong and public keys obtained using
6618 X509_get_pubkey() weren't freed.
6621 *) Add text documentation for the BUFFER functions. Also added a work around
6622 to a Win95 console bug. This was triggered by the password read stuff: the
6623 last character typed gets carried over to the next fread(). If you were
6624 generating a new cert request using 'req' for example then the last
6625 character of the passphrase would be CR which would then enter the first
6629 *) Added the new `Includes OpenSSL Cryptography Software' button as
6630 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6631 button and can be used by applications based on OpenSSL to show the
6632 relationship to the OpenSSL project.
6633 [Ralf S. Engelschall]
6635 *) Remove confusing variables in function signatures in files
6636 ssl/ssl_lib.c and ssl/ssl.h.
6637 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6639 *) Don't install bss_file.c under PREFIX/include/
6640 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6642 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6643 functions that return function pointers and has support for NT specific
6644 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6645 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6646 unsigned to signed types: this was killing the Win32 compile.
6649 *) Add new certificate file to stack functions,
6650 SSL_add_dir_cert_subjects_to_stack() and
6651 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6652 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6653 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6654 This means that Apache-SSL and similar packages don't have to mess around
6655 to add as many CAs as they want to the preferred list.
6658 *) Experiment with doxygen documentation. Currently only partially applied to
6660 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6661 openssl.doxy as the configuration file.
6664 *) Get rid of remaining C++-style comments which strict C compilers hate.
6665 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6667 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6668 compiled in by default: it has problems with large keys.
6671 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6672 DH private keys and/or callback functions which directly correspond to
6673 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6674 is needed for applications which have to configure certificates on a
6675 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6677 For the RSA certificate situation is makes no difference, but
6678 for the DSA certificate situation this fixes the "no shared cipher"
6679 problem where the OpenSSL cipher selection procedure failed because the
6680 temporary keys were not overtaken from the context and the API provided
6681 no way to reconfigure them.
6682 The new functions now let applications reconfigure the stuff and they
6683 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6684 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6685 non-public-API function ssl_cert_instantiate() is used as a helper
6686 function and also to reduce code redundancy inside ssl_rsa.c.
6687 [Ralf S. Engelschall]
6689 *) Move s_server -dcert and -dkey options out of the undocumented feature
6690 area because they are useful for the DSA situation and should be
6691 recognized by the users.
6692 [Ralf S. Engelschall]
6694 *) Fix the cipher decision scheme for export ciphers: the export bits are
6695 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6696 SSL_EXP_MASK. So, the original variable has to be used instead of the
6697 already masked variable.
6698 [Richard Levitte <levitte@stacken.kth.se>]
6700 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6701 [Richard Levitte <levitte@stacken.kth.se>]
6703 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6704 from `int' to `unsigned int' because it's a length and initialized by
6705 EVP_DigestFinal() which expects an `unsigned int *'.
6706 [Richard Levitte <levitte@stacken.kth.se>]
6708 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6709 script. Instead use the usual Shell->Perl transition trick.
6710 [Ralf S. Engelschall]
6712 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6713 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6714 -noout -modulus' as it's already the case for `openssl rsa -noout
6715 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6716 currently the public key is printed (a decision which was already done by
6717 `openssl dsa -modulus' in the past) which serves a similar purpose.
6718 Additionally the NO_RSA no longer completely removes the whole -modulus
6719 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6721 [Ralf S. Engelschall]
6723 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6724 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6725 [Arne Ansper <arne@ats.cyber.ee>]
6727 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6728 to be added. Now both 'req' and 'ca' can use new objects defined in the
6732 *) Add cool BIO that does syslog (or event log on NT).
6733 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6735 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6736 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6737 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6738 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6741 *) Add preliminary config info for new extension code.
6744 *) Make RSA_NO_PADDING really use no padding.
6745 [Ulf Moeller <ulf@fitug.de>]
6747 *) Generate errors when private/public key check is done.
6750 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6751 for some CRL extensions and new objects added.
6754 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6755 key usage extension and fuller support for authority key id.
6758 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6759 padding method for RSA, which is recommended for new applications in PKCS
6760 #1 v2.0 (RFC 2437, October 1998).
6761 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6762 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6763 against Bleichbacher's attack on RSA.
6764 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6767 *) Updates to the new SSL compression code
6768 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6770 *) Fix so that the version number in the master secret, when passed
6771 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6772 (because the server will not accept higher), that the version number
6773 is 0x03,0x01, not 0x03,0x00
6774 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6776 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6777 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6778 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6781 *) Support for RAW extensions where an arbitrary extension can be
6782 created by including its DER encoding. See apps/openssl.cnf for
6786 *) Make sure latest Perl versions don't interpret some generated C array
6787 code as Perl array code in the crypto/err/err_genc.pl script.
6788 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6790 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6791 not many people have the assembler. Various Win32 compilation fixes and
6792 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6796 *) Modify configure script 'Configure' to automatically create crypto/date.h
6797 file under Win32 and also build pem.h from pem.org. New script
6798 util/mkfiles.pl to create the MINFO file on environments that can't do a
6799 'make files': perl util/mkfiles.pl >MINFO should work.
6802 *) Major rework of DES function declarations, in the pursuit of correctness
6803 and purity. As a result, many evil casts evaporated, and some weirdness,
6804 too. You may find this causes warnings in your code. Zapping your evil
6805 casts will probably fix them. Mostly.
6808 *) Fix for a typo in asn1.h. Bug fix to object creation script
6809 obj_dat.pl. It considered a zero in an object definition to mean
6810 "end of object": none of the objects in objects.h have any zeros
6811 so it wasn't spotted.
6812 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6814 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6815 Masking (CBCM). In the absence of test vectors, the best I have been able
6816 to do is check that the decrypt undoes the encrypt, so far. Send me test
6817 vectors if you have them.
6820 *) Correct calculation of key length for export ciphers (too much space was
6821 allocated for null ciphers). This has not been tested!
6824 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6825 message is now correct (it understands "crypto" and "ssl" on its
6826 command line). There is also now an "update" option. This will update
6827 the util/ssleay.num and util/libeay.num files with any new functions.
6829 perl util/mkdef.pl crypto ssl update
6830 it will update them.
6833 *) Overhauled the Perl interface (perl/*):
6834 - ported BN stuff to OpenSSL's different BN library
6835 - made the perl/ source tree CVS-aware
6836 - renamed the package from SSLeay to OpenSSL (the files still contain
6837 their history because I've copied them in the repository)
6838 - removed obsolete files (the test scripts will be replaced
6839 by better Test::Harness variants in the future)
6840 [Ralf S. Engelschall]
6842 *) First cut for a very conservative source tree cleanup:
6843 1. merge various obsolete readme texts into doc/ssleay.txt
6844 where we collect the old documents and readme texts.
6845 2. remove the first part of files where I'm already sure that we no
6846 longer need them because of three reasons: either they are just temporary
6847 files which were left by Eric or they are preserved original files where
6848 I've verified that the diff is also available in the CVS via "cvs diff
6849 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6850 the crypto/md/ stuff).
6851 [Ralf S. Engelschall]
6853 *) More extension code. Incomplete support for subject and issuer alt
6854 name, issuer and authority key id. Change the i2v function parameters
6855 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6856 what that's for :-) Fix to ASN1 macro which messed up
6857 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6860 *) Preliminary support for ENUMERATED type. This is largely copied from the
6864 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6865 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6867 *) Make sure `make rehash' target really finds the `openssl' program.
6868 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6870 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6871 like to hear about it if this slows down other processors.
6874 *) Add CygWin32 platform information to Configure script.
6875 [Alan Batie <batie@aahz.jf.intel.com>]
6877 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6878 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6880 *) New program nseq to manipulate netscape certificate sequences
6883 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6887 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6888 but the BN code had some problems that would cause failures when
6889 doing certificate verification and some other functions.
6890 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6892 *) Add ASN1 and PEM code to support netscape certificate sequences.
6895 *) Add ASN1 and PEM code to support netscape certificate sequences.
6898 *) Add several PKIX and private extended key usage OIDs.
6901 *) Modify the 'ca' program to handle the new extension code. Modify
6902 openssl.cnf for new extension format, add comments.
6905 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
6906 and add a sample to openssl.cnf so req -x509 now adds appropriate
6910 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
6911 error code, add initial support to X509_print() and x509 application.
6914 *) Takes a deep breath and start addding X509 V3 extension support code. Add
6915 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
6916 stuff is currently isolated and isn't even compiled yet.
6919 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
6920 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
6921 Removed the versions check from X509 routines when loading extensions:
6922 this allows certain broken certificates that don't set the version
6923 properly to be processed.
6926 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
6927 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
6928 can still be regenerated with "make depend".
6931 *) Spelling mistake in C version of CAST-128.
6932 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
6934 *) Changes to the error generation code. The perl script err-code.pl
6935 now reads in the old error codes and retains the old numbers, only
6936 adding new ones if necessary. It also only changes the .err files if new
6937 codes are added. The makefiles have been modified to only insert errors
6938 when needed (to avoid needlessly modifying header files). This is done
6939 by only inserting errors if the .err file is newer than the auto generated
6940 C file. To rebuild all the error codes from scratch (the old behaviour)
6941 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
6942 or delete all the .err files.
6945 *) CAST-128 was incorrectly implemented for short keys. The C version has
6946 been fixed, but is untested. The assembler versions are also fixed, but
6947 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
6948 to regenerate it if needed.
6949 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
6950 Hagino <itojun@kame.net>]
6952 *) File was opened incorrectly in randfile.c.
6953 [Ulf Möller <ulf@fitug.de>]
6955 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
6956 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
6957 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
6958 al: it's just almost always a UTCTime. Note this patch adds new error
6959 codes so do a "make errors" if there are problems.
6962 *) Correct Linux 1 recognition in config.
6963 [Ulf Möller <ulf@fitug.de>]
6965 *) Remove pointless MD5 hash when using DSA keys in ca.
6966 [Anonymous <nobody@replay.com>]
6968 *) Generate an error if given an empty string as a cert directory. Also
6969 generate an error if handed NULL (previously returned 0 to indicate an
6970 error, but didn't set one).
6971 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
6973 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
6976 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
6977 parameters. This was causing a warning which killed off the Win32 compile.
6980 *) Remove C++ style comments from crypto/bn/bn_local.h.
6981 [Neil Costigan <neil.costigan@celocom.com>]
6983 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
6984 based on a text string, looking up short and long names and finally
6985 "dot" format. The "dot" format stuff didn't work. Added new function
6986 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
6987 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
6988 OID is not part of the table.
6991 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
6992 X509_LOOKUP_by_alias().
6995 *) Sort openssl functions by name.
6998 *) Get the gendsa program working (hopefully) and add it to app list. Remove
6999 encryption from sample DSA keys (in case anyone is interested the password
7003 *) Make _all_ *_free functions accept a NULL pointer.
7004 [Frans Heymans <fheymans@isaserver.be>]
7006 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7008 [Anonymous <nobody@replay.com>]
7010 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7011 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7013 *) Don't blow it for numeric -newkey arguments to apps/req.
7014 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7016 *) Temp key "for export" tests were wrong in s3_srvr.c.
7017 [Anonymous <nobody@replay.com>]
7019 *) Add prototype for temp key callback functions
7020 SSL_CTX_set_tmp_{rsa,dh}_callback().
7023 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7024 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7027 *) X509_name_add_entry() freed the wrong thing after an error.
7028 [Arne Ansper <arne@ats.cyber.ee>]
7030 *) rsa_eay.c would attempt to free a NULL context.
7031 [Arne Ansper <arne@ats.cyber.ee>]
7033 *) BIO_s_socket() had a broken should_retry() on Windoze.
7034 [Arne Ansper <arne@ats.cyber.ee>]
7036 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7037 [Arne Ansper <arne@ats.cyber.ee>]
7039 *) Make sure the already existing X509_STORE->depth variable is initialized
7040 in X509_STORE_new(), but document the fact that this variable is still
7041 unused in the certificate verification process.
7042 [Ralf S. Engelschall]
7044 *) Fix the various library and apps files to free up pkeys obtained from
7045 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7048 *) Fix reference counting in X509_PUBKEY_get(). This makes
7049 demos/maurice/example2.c work, amongst others, probably.
7050 [Steve Henson and Ben Laurie]
7052 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7053 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7054 are no longer created. This way we have a single and consistent command
7055 line interface `openssl <command>', similar to `cvs <command>'.
7056 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7058 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7059 BIT STRING wrapper always have zero unused bits.
7062 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7065 *) Make the top-level INSTALL documentation easier to understand.
7068 *) Makefiles updated to exit if an error occurs in a sub-directory
7069 make (including if user presses ^C) [Paul Sutton]
7071 *) Make Montgomery context stuff explicit in RSA data structure.
7074 *) Fix build order of pem and err to allow for generated pem.h.
7077 *) Fix renumbering bug in X509_NAME_delete_entry().
7080 *) Enhanced the err-ins.pl script so it makes the error library number
7081 global and can add a library name. This is needed for external ASN1 and
7082 other error libraries.
7085 *) Fixed sk_insert which never worked properly.
7088 *) Fix ASN1 macros so they can handle indefinite length construted
7089 EXPLICIT tags. Some non standard certificates use these: they can now
7093 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7094 into a single doc/ssleay.txt bundle. This way the information is still
7095 preserved but no longer messes up this directory. Now it's new room for
7096 the new set of documenation files.
7097 [Ralf S. Engelschall]
7099 *) SETs were incorrectly DER encoded. This was a major pain, because they
7100 shared code with SEQUENCEs, which aren't coded the same. This means that
7101 almost everything to do with SETs or SEQUENCEs has either changed name or
7102 number of arguments.
7103 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7105 *) Fix test data to work with the above.
7108 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7109 was already fixed by Eric for 0.9.1 it seems.
7110 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7112 *) Autodetect FreeBSD3.
7115 *) Fix various bugs in Configure. This affects the following platforms:
7119 unixware-2.0-pentium
7123 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7124 before they are needed.
7127 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7131 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7133 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7134 changed SSLeay to OpenSSL in version strings.
7135 [Ralf S. Engelschall]
7137 *) Some fixups to the top-level documents.
7140 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7141 because the symlink to include/ was missing.
7142 [Ralf S. Engelschall]
7144 *) Incorporated the popular no-RSA/DSA-only patches
7145 which allow to compile a RSA-free SSLeay.
7146 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7148 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7149 when "ssleay" is still not found.
7150 [Ralf S. Engelschall]
7152 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7153 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7155 *) Updated the README file.
7156 [Ralf S. Engelschall]
7158 *) Added various .cvsignore files in the CVS repository subdirs
7159 to make a "cvs update" really silent.
7160 [Ralf S. Engelschall]
7162 *) Recompiled the error-definition header files and added
7163 missing symbols to the Win32 linker tables.
7164 [Ralf S. Engelschall]
7166 *) Cleaned up the top-level documents;
7167 o new files: CHANGES and LICENSE
7168 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7169 o merged COPYRIGHT into LICENSE
7170 o removed obsolete TODO file
7171 o renamed MICROSOFT to INSTALL.W32
7172 [Ralf S. Engelschall]
7174 *) Removed dummy files from the 0.9.1b source tree:
7175 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7176 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7177 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7178 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7179 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7180 [Ralf S. Engelschall]
7182 *) Added various platform portability fixes.
7185 *) The Genesis of the OpenSSL rpject:
7186 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7187 Young and Tim J. Hudson created while they were working for C2Net until
7189 [The OpenSSL Project]
7192 Changes between 0.9.0b and 0.9.1b [not released]
7194 *) Updated a few CA certificates under certs/
7197 *) Changed some BIGNUM api stuff.
7200 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7201 DGUX x86, Linux Alpha, etc.
7204 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7205 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7209 *) Add -strparse option to asn1pars program which parses nested
7211 [Dr Stephen Henson <shenson@bigfoot.com>]
7213 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7216 *) DSA fix for "ca" program.
7219 *) Added "-genkey" option to "dsaparam" program.
7222 *) Added RIPE MD160 (rmd160) message digest.
7225 *) Added -a (all) option to "ssleay version" command.
7228 *) Added PLATFORM define which is the id given to Configure.
7231 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7234 *) Extended the ASN.1 parser routines.
7237 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7240 *) Added a BN_CTX to the BN library.
7243 *) Fixed the weak key values in DES library
7246 *) Changed API in EVP library for cipher aliases.
7249 *) Added support for RC2/64bit cipher.
7252 *) Converted the lhash library to the crypto/mem.c functions.
7255 *) Added more recognized ASN.1 object ids.
7258 *) Added more RSA padding checks for SSL/TLS.
7261 *) Added BIO proxy/filter functionality.
7264 *) Added extra_certs to SSL_CTX which can be used
7265 send extra CA certificates to the client in the CA cert chain sending
7266 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7269 *) Now Fortezza is denied in the authentication phase because
7270 this is key exchange mechanism is not supported by SSLeay at all.
7273 *) Additional PKCS1 checks.
7276 *) Support the string "TLSv1" for all TLS v1 ciphers.
7279 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7280 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7283 *) Fixed a few memory leaks.
7286 *) Fixed various code and comment typos.
7289 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7290 bytes sent in the client random.
7291 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob