5 Changes between 0.9.7e and 0.9.8 [xx XXX xxxx]
7 *) Compile clean with "-Wall -Wmissing-prototypes
8 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
9 DEBUG_SAFESTACK must also be set.
12 *) Change ./Configure so that certain algorithms can be disabled by default.
13 The new counterpiece to "no-xxx" is "enable-xxx".
15 The patented RC5 and MDC2 algorithms will now be disabled unless
16 "enable-rc5" and "enable-mdc2", respectively, are specified.
18 (IDEA remains enabled despite being patented. This is because IDEA
19 is frequently required for interoperability, and there is no license
20 fee for non-commercial use. As before, "no-idea" can be used to
21 avoid this algorithm.)
25 *) Add processing of proxy certificates (see RFC 3820). This work was
26 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
27 EGEE (Enabling Grids for E-science in Europe).
30 *) RC4 performance overhaul on modern architectures/implementations, such
31 as Intel P4, IA-64 and AMD64.
34 *) New utility extract-section.pl. This can be used specify an alternative
35 section number in a pod file instead of having to treat each file as
36 a separate case in Makefile. This can be done by adding two lines to the
39 =for comment openssl_section:XXX
41 The blank line is mandatory.
45 *) New arguments -certform, -keyform and -pass for s_client and s_server
46 to allow alternative format key and certificate files and passphrase
50 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
51 update associated structures and add various utility functions.
53 Add new policy related verify parameters, include policy checking in
54 standard verify code. Enhance 'smime' application with extra parameters
55 to support policy checking and print out.
58 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
59 Nehemiah processors. These extensions support AES encryption in hardware
60 as well as RNG (though RNG support is currently disabled).
61 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
63 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
66 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
67 [Andy Polyakov and a number of other people]
69 *) Improved PowerPC platform support. Most notably BIGNUM assembler
70 implementation contributed by IBM.
71 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
73 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
74 exponent rather than 'unsigned long'. There is a corresponding change to
75 the new 'rsa_keygen' element of the RSA_METHOD structure.
76 [Jelte Jansen, Geoff Thorpe]
78 *) Functionality for creating the initial serial number file is now
79 moved from CA.pl to the 'ca' utility with a new option -create_serial.
81 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
82 number file to 1, which is bound to cause problems. To avoid
83 the problems while respecting compatibility between different 0.9.7
84 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
85 CA.pl for serial number initialization. With the new release 0.9.8,
86 we can fix the problem directly in the 'ca' utility.)
89 *) Reduced header interdepencies by declaring more opaque objects in
90 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
91 give fewer recursive includes, which could break lazy source code - so
92 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
93 developers should define this symbol when building and using openssl to
94 ensure they track the recommended behaviour, interfaces, [etc], but
95 backwards-compatible behaviour prevails when this isn't defined.
98 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
101 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
102 This will generate a random key of the appropriate length based on the
103 cipher context. The EVP_CIPHER can provide its own random key generation
104 routine to support keys of a specific form. This is used in the des and
105 3des routines to generate a key of the correct parity. Update S/MIME
106 code to use new functions and hence generate correct parity DES keys.
107 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
108 valid (weak or incorrect parity).
111 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
112 as looking them up. This is useful when the verified structure may contain
113 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
114 present unless the new PKCS7_NO_CRL flag is asserted.
117 *) Extend ASN1 oid configuration module. It now additionally accepts the
120 shortName = some long name, 1.2.3.4
123 *) Reimplemented the BN_CTX implementation. There is now no more static
124 limitation on the number of variables it can handle nor the depth of the
125 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
126 information can now expand as required, and rather than having a single
127 static array of bignums, BN_CTX now uses a linked-list of such arrays
128 allowing it to expand on demand whilst maintaining the usefulness of
132 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
133 to allow all RSA operations to function using a single BN_CTX.
136 *) Preliminary support for certificate policy evaluation and checking. This
137 is initially intended to pass the tests outlined in "Conformance Testing
138 of Relying Party Client Certificate Path Processing Logic" v1.07.
141 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
142 remained unused and not that useful. A variety of other little bignum
143 tweaks and fixes have also been made continuing on from the audit (see
147 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
148 associated ASN1, EVP and SSL functions and old ASN1 macros.
151 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
152 and this should never fail. So the return value from the use of
153 BN_set_word() (which can fail due to needless expansion) is now deprecated;
154 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
157 *) BN_CTX_get() should return zero-valued bignums, providing the same
158 initialised value as BN_new().
159 [Geoff Thorpe, suggested by Ulf Möller]
161 *) Support for inhibitAnyPolicy certificate extension.
164 *) An audit of the BIGNUM code is underway, for which debugging code is
165 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
166 is considered valid when processing BIGNUMs, and causes execution to
167 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
168 further steps are taken to deliberately pollute unused data in BIGNUM
169 structures to try and expose faulty code further on. For now, openssl will
170 (in its default mode of operation) continue to tolerate the inconsistent
171 forms that it has tolerated in the past, but authors and packagers should
172 consider trying openssl and their own applications when compiled with
173 these debugging symbols defined. It will help highlight potential bugs in
174 their own code, and will improve the test coverage for OpenSSL itself. At
175 some point, these tighter rules will become openssl's default to improve
176 maintainability, though the assert()s and other overheads will remain only
177 in debugging configurations. See bn.h for more details.
178 [Geoff Thorpe, Nils Larsch, Ulf Möller]
180 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
181 that can only be obtained through BN_CTX_new() (which implicitly
182 initialises it). The presence of this function only made it possible
183 to overwrite an existing structure (and cause memory leaks).
186 *) Because of the callback-based approach for implementing LHASH as a
187 template type, lh_insert() adds opaque objects to hash-tables and
188 lh_doall() or lh_doall_arg() are typically used with a destructor callback
189 to clean up those corresponding objects before destroying the hash table
190 (and losing the object pointers). So some over-zealous constifications in
191 LHASH have been relaxed so that lh_insert() does not take (nor store) the
192 objects as "const" and the lh_doall[_arg] callback wrappers are not
193 prototyped to have "const" restrictions on the object pointers they are
194 given (and so aren't required to cast them away any more).
197 *) The tmdiff.h API was so ugly and minimal that our own timing utility
198 (speed) prefers to use its own implementation. The two implementations
199 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
200 its object type properly exposed (MS_TM) instead of casting to/from "char
201 *". This may still change yet if someone realises MS_TM and "ms_time_***"
202 aren't necessarily the greatest nomenclatures - but this is what was used
203 internally to the implementation so I've used that for now.
206 *) Ensure that deprecated functions do not get compiled when
207 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
208 the self-tests were still using deprecated key-generation functions so
209 these have been updated also.
212 *) Reorganise PKCS#7 code to separate the digest location functionality
213 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
214 New function PKCS7_set_digest() to set the digest type for PKCS#7
215 digestedData type. Add additional code to correctly generate the
216 digestedData type and add support for this type in PKCS7 initialization
220 *) New function PKCS7_set0_type_other() this initializes a PKCS7
221 structure of type "other".
224 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
225 sure the loop does correctly stop and breaking ("division by zero")
226 modulus operations are not performed. The (pre-generated) prime
227 table crypto/bn/bn_prime.h was already correct, but it could not be
228 re-generated on some platforms because of the "division by zero"
229 situation in the script.
230 [Ralf S. Engelschall]
232 *) Update support for ECC-based TLS ciphersuites according to
233 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
234 SHA-1 now is only used for "small" curves (where the
235 representation of a field element takes up to 24 bytes); for
236 larger curves, the field element resulting from ECDH is directly
237 used as premaster secret.
238 [Douglas Stebila (Sun Microsystems Laboratories)]
240 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
241 curve secp160r1 to the tests.
242 [Douglas Stebila (Sun Microsystems Laboratories)]
244 *) Add the possibility to load symbols globally with DSO.
245 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
247 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
248 control of the error stack.
251 *) Add support for STORE in ENGINE.
254 *) Add the STORE type. The intention is to provide a common interface
255 to certificate and key stores, be they simple file-based stores, or
256 HSM-type store, or LDAP stores, or...
257 NOTE: The code is currently UNTESTED and isn't really used anywhere.
260 *) Add a generic structure called OPENSSL_ITEM. This can be used to
261 pass a list of arguments to any function as well as provide a way
262 for a function to pass data back to the caller.
265 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
266 works like BUF_strdup() but can be used to duplicate a portion of
267 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
271 *) Add the function sk_find_ex() which works like sk_find(), but will
272 return an index to an element even if an exact match couldn't be
273 found. The index is guaranteed to point at the element where the
274 searched-for key would be inserted to preserve sorting order.
277 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
278 takes an extra flags argument for optional functionality. Currently,
279 the following flags are defined:
281 OBJ_BSEARCH_VALUE_ON_NOMATCH
282 This one gets OBJ_bsearch_ex() to return a pointer to the first
283 element where the comparing function returns a negative or zero
286 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
287 This one gets OBJ_bsearch_ex() to return a pointer to the first
288 element where the comparing function returns zero. This is useful
289 if there are more than one element where the comparing function
293 *) Make it possible to create self-signed certificates with 'openssl ca'
294 in such a way that the self-signed certificate becomes part of the
295 CA database and uses the same mechanisms for serial number generation
296 as all other certificate signing. The new flag '-selfsign' enables
297 this functionality. Adapt CA.sh and CA.pl.in.
300 *) Add functionality to check the public key of a certificate request
301 against a given private. This is useful to check that a certificate
302 request can be signed by that key (self-signing).
305 *) Make it possible to have multiple active certificates with the same
306 subject in the CA index file. This is done only if the keyword
307 'unique_subject' is set to 'no' in the main CA section (default
308 if 'CA_default') of the configuration file. The value is saved
309 with the database itself in a separate index attribute file,
310 named like the index file with '.attr' appended to the name.
313 *) Generate muti valued AVAs using '+' notation in config files for
317 *) Support for nameConstraints certificate extension.
320 *) Support for policyConstraints certificate extension.
323 *) Support for policyMappings certificate extension.
326 *) Make sure the default DSA_METHOD implementation only uses its
327 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
328 and change its own handlers to be NULL so as to remove unnecessary
329 indirection. This lets alternative implementations fallback to the
330 default implementation more easily.
333 *) Support for directoryName in GeneralName related extensions
337 *) Make it possible to link applications using Makefile.shared.
338 Make that possible even when linking against static libraries!
341 *) Support for single pass processing for S/MIME signing. This now
342 means that S/MIME signing can be done from a pipe, in addition
343 cleartext signing (multipart/signed type) is effectively streaming
344 and the signed data does not need to be all held in memory.
346 This is done with a new flag PKCS7_STREAM. When this flag is set
347 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
348 is done after the data is output (and digests calculated) in
352 *) Add full support for -rpath/-R, both in shared libraries and
353 applications, at least on the platforms where it's known how
357 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
358 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
359 will now compute a table of multiples of the generator that
360 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
361 faster (notably in the case of a single point multiplication,
363 [Nils Larsch, Bodo Moeller]
365 *) IPv6 support for certificate extensions. The various extensions
366 which use the IP:a.b.c.d can now take IPv6 addresses using the
367 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
371 *) Added an ENGINE that implements RSA by performing private key
372 exponentiations with the GMP library. The conversions to and from
373 GMP's mpz_t format aren't optimised nor are any montgomery forms
374 cached, and on x86 it appears OpenSSL's own performance has caught up.
375 However there are likely to be other architectures where GMP could
376 provide a boost. This ENGINE is not built in by default, but it can be
377 specified at Configure time and should be accompanied by the necessary
378 linker additions, eg;
379 ./config -DOPENSSL_USE_GMP -lgmp
382 *) "openssl engine" will not display ENGINE/DSO load failure errors when
383 testing availability of engines with "-t" - the old behaviour is
384 produced by increasing the feature's verbosity with "-tt".
387 *) ECDSA routines: under certain error conditions uninitialized BN objects
388 could be freed. Solution: make sure initialization is performed early
389 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
393 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
394 and DH_METHOD (eg. by ENGINE implementations) to override the normal
395 software implementations. For DSA and DH, parameter generation can
396 also be overriden by providing the appropriate method callbacks.
399 *) Change the "progress" mechanism used in key-generation and
400 primality testing to functions that take a new BN_GENCB pointer in
401 place of callback/argument pairs. The new API functions have "_ex"
402 postfixes and the older functions are reimplemented as wrappers for
403 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
404 declarations of the old functions to help (graceful) attempts to
405 migrate to the new functions. Also, the new key-generation API
406 functions operate on a caller-supplied key-structure and return
407 success/failure rather than returning a key or NULL - this is to
408 help make "keygen" another member function of RSA_METHOD etc.
410 Example for using the new callback interface:
412 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
416 BN_GENCB_set(&my_cb, my_callback, my_arg);
418 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
419 /* For the meaning of a, b in calls to my_callback(), see the
420 * documentation of the function that calls the callback.
421 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
422 * my_callback should return 1 if it wants BN_is_prime_ex()
423 * to continue, or 0 to stop.
428 *) Change the ZLIB compression method to be stateful, and make it
429 available to TLS with the number defined in
430 draft-ietf-tls-compression-04.txt.
433 *) Add the ASN.1 structures and functions for CertificatePair, which
434 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
436 CertificatePair ::= SEQUENCE {
437 forward [0] Certificate OPTIONAL,
438 reverse [1] Certificate OPTIONAL,
439 -- at least one of the pair shall be present -- }
441 Also implement the PEM functions to read and write certificate
442 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
444 This needed to be defined, mostly for the sake of the LDAP
445 attribute crossCertificatePair, but may prove useful elsewhere as
449 *) Make it possible to inhibit symlinking of shared libraries in
450 Makefile.shared, for Cygwin's sake.
453 *) Extend the BIGNUM API by creating new macros that behave like
456 void BN_set_sign(BIGNUM *a, int neg);
457 int BN_get_sign(const BIGNUM *a);
459 and avoid the need to access 'a->neg' directly in applications.
460 [Nils Larsch <nla@trustcenter.de>]
462 *) Implement fast modular reduction for pseudo-Mersenne primes
463 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
464 EC_GROUP_new_curve_GFp() will now automatically use this
466 [Nils Larsch <nla@trustcenter.de>]
468 *) Add new lock type (CRYPTO_LOCK_BN).
471 *) Change the ENGINE framework to automatically load engines
472 dynamically from specific directories unless they could be
473 found to already be built in or loaded. Move all the
474 current engines except for the cryptodev one to a new
476 The engines in engines/ are built as shared libraries if
477 the "shared" options was given to ./Configure or ./config.
478 Otherwise, they are inserted in libcrypto.a.
479 /usr/local/ssl/engines is the default directory for dynamic
480 engines, but that can be overriden at configure time through
481 the usual use of --prefix and/or --openssldir, and at run
482 time with the environment variable OPENSSL_ENGINES.
483 [Geoff Thorpe and Richard Levitte]
485 *) Add Makefile.shared, a helper makefile to build shared
486 libraries. Addapt Makefile.org.
489 *) Add version info to Win32 DLLs.
490 [Peter 'Luna' Runestig" <peter@runestig.com>]
492 *) Add new 'medium level' PKCS#12 API. Certificates and keys
493 can be added using this API to created arbitrary PKCS#12
494 files while avoiding the low level API.
496 New options to PKCS12_create(), key or cert can be NULL and
497 will then be omitted from the output file. The encryption
498 algorithm NIDs can be set to -1 for no encryption, the mac
499 iteration count can be set to 0 to omit the mac.
501 Enhance pkcs12 utility by making the -nokeys and -nocerts
502 options work when creating a PKCS#12 file. New option -nomac
503 to omit the mac, NONE can be set for an encryption algorithm.
504 New code is modified to use the enhanced PKCS12_create()
505 instead of the low level API.
508 *) Extend ASN1 encoder to support indefinite length constructed
509 encoding. This can output sequences tags and octet strings in
510 this form. Modify pk7_asn1.c to support indefinite length
511 encoding. This is experimental and needs additional code to
512 be useful, such as an ASN1 bio and some enhanced streaming
515 Extend template encode functionality so that tagging is passed
516 down to the template encoder.
519 *) Let 'openssl req' fail if an argument to '-newkey' is not
520 recognized instead of using RSA as a default.
523 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
524 As these are not official, they are not included in "ALL";
525 the "ECCdraft" ciphersuite group alias can be used to select them.
526 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
528 *) Add ECDH engine support.
529 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
531 *) Add ECDH in new directory crypto/ecdh/.
532 [Douglas Stebila (Sun Microsystems Laboratories)]
534 *) Let BN_rand_range() abort with an error after 100 iterations
535 without success (which indicates a broken PRNG).
538 *) Change BN_mod_sqrt() so that it verifies that the input value
539 is really the square of the return value. (Previously,
540 BN_mod_sqrt would show GIGO behaviour.)
543 *) Add named elliptic curves over binary fields from X9.62, SECG,
544 and WAP/WTLS; add OIDs that were still missing.
546 [Sheueling Chang Shantz and Douglas Stebila
547 (Sun Microsystems Laboratories)]
549 *) Extend the EC library for elliptic curves over binary fields
550 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
553 EC_GF2m_simple_method
557 EC_GROUP_new_curve_GF2m
558 EC_GROUP_set_curve_GF2m
559 EC_GROUP_get_curve_GF2m
560 EC_POINT_set_affine_coordinates_GF2m
561 EC_POINT_get_affine_coordinates_GF2m
562 EC_POINT_set_compressed_coordinates_GF2m
564 Point compression for binary fields is disabled by default for
565 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
568 As binary polynomials are represented as BIGNUMs, various members
569 of the EC_GROUP and EC_POINT data structures can be shared
570 between the implementations for prime fields and binary fields;
571 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
572 are essentially identical to their ..._GFp counterparts.
573 (For simplicity, the '..._GFp' prefix has been dropped from
574 various internal method names.)
576 An internal 'field_div' method (similar to 'field_mul' and
577 'field_sqr') has been added; this is used only for binary fields.
579 [Sheueling Chang Shantz and Douglas Stebila
580 (Sun Microsystems Laboratories)]
582 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
583 through methods ('mul', 'precompute_mult').
585 The generic implementations (now internally called 'ec_wNAF_mul'
586 and 'ec_wNAF_precomputed_mult') remain the default if these
587 methods are undefined.
589 [Sheueling Chang Shantz and Douglas Stebila
590 (Sun Microsystems Laboratories)]
592 *) New function EC_GROUP_get_degree, which is defined through
593 EC_METHOD. For curves over prime fields, this returns the bit
594 length of the modulus.
596 [Sheueling Chang Shantz and Douglas Stebila
597 (Sun Microsystems Laboratories)]
599 *) New functions EC_GROUP_dup, EC_POINT_dup.
600 (These simply call ..._new and ..._copy).
602 [Sheueling Chang Shantz and Douglas Stebila
603 (Sun Microsystems Laboratories)]
605 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
606 Polynomials are represented as BIGNUMs (where the sign bit is not
607 used) in the following functions [macros]:
610 BN_GF2m_sub [= BN_GF2m_add]
611 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
612 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
613 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
615 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
616 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
617 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
618 BN_GF2m_cmp [= BN_ucmp]
620 (Note that only the 'mod' functions are actually for fields GF(2^m).
621 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
623 For some functions, an the irreducible polynomial defining a
624 field can be given as an 'unsigned int[]' with strictly
625 decreasing elements giving the indices of those bits that are set;
626 i.e., p[] represents the polynomial
627 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
629 p[0] > p[1] > ... > p[k] = 0.
630 This applies to the following functions:
635 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
636 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
639 BN_GF2m_mod_solve_quad_arr
643 Conversion can be performed by the following functions:
648 bntest.c has additional tests for binary polynomial arithmetic.
650 Two implementations for BN_GF2m_mod_div() are available.
651 The default algorithm simply uses BN_GF2m_mod_inv() and
652 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
653 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
654 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
656 [Sheueling Chang Shantz and Douglas Stebila
657 (Sun Microsystems Laboratories)]
659 *) Add new error code 'ERR_R_DISABLED' that can be used when some
660 functionality is disabled at compile-time.
661 [Douglas Stebila <douglas.stebila@sun.com>]
663 *) Change default behaviour of 'openssl asn1parse' so that more
664 information is visible when viewing, e.g., a certificate:
666 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
667 mode the content of non-printable OCTET STRINGs is output in a
668 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
669 avoid the appearance of a printable string.
670 [Nils Larsch <nla@trustcenter.de>]
672 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
674 EC_GROUP_set_asn1_flag()
675 EC_GROUP_get_asn1_flag()
676 EC_GROUP_set_point_conversion_form()
677 EC_GROUP_get_point_conversion_form()
678 These control ASN1 encoding details:
679 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
680 has been set to OPENSSL_EC_NAMED_CURVE.
681 - Points are encoded in uncompressed form by default; options for
682 asn1_for are as for point2oct, namely
683 POINT_CONVERSION_COMPRESSED
684 POINT_CONVERSION_UNCOMPRESSED
685 POINT_CONVERSION_HYBRID
687 Also add 'seed' and 'seed_len' members to EC_GROUP with access
691 EC_GROUP_get_seed_len()
692 This is used only for ASN1 purposes (so far).
693 [Nils Larsch <nla@trustcenter.de>]
695 *) Add 'field_type' member to EC_METHOD, which holds the NID
696 of the appropriate field type OID. The new function
697 EC_METHOD_get_field_type() returns this value.
698 [Nils Larsch <nla@trustcenter.de>]
705 providing useful interfaces to EC_POINT_point2oct() and
706 EC_POINT_oct2point().
707 [Nils Larsch <nla@trustcenter.de>]
709 *) Change internals of the EC library so that the functions
710 EC_GROUP_set_generator()
711 EC_GROUP_get_generator()
713 EC_GROUP_get_cofactor()
714 are implemented directly in crypto/ec/ec_lib.c and not dispatched
715 to methods, which would lead to unnecessary code duplication when
716 adding different types of curves.
717 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
719 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
720 arithmetic, and such that modified wNAFs are generated
721 (which avoid length expansion in many cases).
724 *) Add a function EC_GROUP_check_discriminant() (defined via
725 EC_METHOD) that verifies that the curve discriminant is non-zero.
727 Add a function EC_GROUP_check() that makes some sanity tests
728 on a EC_GROUP, its generator and order. This includes
729 EC_GROUP_check_discriminant().
730 [Nils Larsch <nla@trustcenter.de>]
732 *) Add ECDSA in new directory crypto/ecdsa/.
734 Add applications 'openssl ecparam' and 'openssl ecdsa'
735 (these are based on 'openssl dsaparam' and 'openssl dsa').
737 ECDSA support is also included in various other files across the
738 library. Most notably,
739 - 'openssl req' now has a '-newkey ecdsa:file' option;
740 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
741 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
742 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
743 them suitable for ECDSA where domain parameters must be
744 extracted before the specific public key;
745 - ECDSA engine support has been added.
746 [Nils Larsch <nla@trustcenter.de>]
748 *) Include some named elliptic curves, and add OIDs from X9.62,
749 SECG, and WAP/WTLS. Each curve can be obtained from the new
751 EC_GROUP_new_by_nid(),
752 and the list of available named curves can be obtained with
753 EC_get_builtin_curves().
754 Also add a 'curve_name' member to EC_GROUP objects, which can be
758 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
760 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
761 was actually never needed) and in BN_mul(). The removal in BN_mul()
762 required a small change in bn_mul_part_recursive() and the addition
763 of the functions bn_cmp_part_words(), bn_sub_part_words() and
764 bn_add_part_words(), which do the same thing as bn_cmp_words(),
765 bn_sub_words() and bn_add_words() except they take arrays with
769 Changes between 0.9.7f and 0.9.7g [XX xxx xxxx]
771 *) Undo Cygwin change.
774 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
776 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
777 server and client random values. Previously
778 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
779 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
781 This change has negligible security impact because:
783 1. Server and client random values still have 24 bytes of pseudo random
786 2. Server and client random values are sent in the clear in the initial
789 3. The master secret is derived using the premaster secret (48 bytes in
790 size for static RSA ciphersuites) as well as client server and random
793 The OpenSSL team would like to thank the UK NISCC for bringing this issue
796 [Stephen Henson, reported by UK NISCC]
798 *) Use Windows randomness collection on Cygwin.
801 *) Fix hang in EGD/PRNGD query when communication socket is closed
802 prematurely by EGD/PRNGD.
803 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
805 *) Prompt for pass phrases when appropriate for PKCS12 input format.
808 *) Back-port of selected performance improvements from development
809 branch, as well as improved support for PowerPC platforms.
812 *) Add lots of checks for memory allocation failure, error codes to indicate
813 failure and freeing up memory if a failure occurs.
814 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
816 *) Add new -passin argument to dgst.
819 *) Perform some character comparisons of different types in X509_NAME_cmp:
820 this is needed for some certificates that reencode DNs into UTF8Strings
821 (in violation of RFC3280) and can't or wont issue name rollover
825 *) Make an explicit check during certificate validation to see that
826 the CA setting in each certificate on the chain is correct. As a
827 side effect always do the following basic checks on extensions,
828 not just when there's an associated purpose to the check:
830 - if there is an unhandled critical extension (unless the user
831 has chosen to ignore this fault)
832 - if the path length has been exceeded (if one is set at all)
833 - that certain extensions fit the associated purpose (if one has
837 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
839 *) Avoid a race condition when CRLs are checked in a multi threaded
840 environment. This would happen due to the reordering of the revoked
841 entries during signature checking and serial number lookup. Now the
842 encoding is cached and the serial number sort performed under a lock.
843 Add new STACK function sk_is_sorted().
846 *) Add Delta CRL to the extension code.
849 *) Various fixes to s3_pkt.c so alerts are sent properly.
850 [David Holmes <d.holmes@f5.com>]
852 *) Reduce the chances of duplicate issuer name and serial numbers (in
853 violation of RFC3280) using the OpenSSL certificate creation utilities.
854 This is done by creating a random 64 bit value for the initial serial
855 number when a serial number file is created or when a self signed
856 certificate is created using 'openssl req -x509'. The initial serial
857 number file is created using 'openssl x509 -next_serial' in CA.pl
858 rather than being initialized to 1.
861 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
863 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
864 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
865 [Joe Orton, Steve Henson]
867 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
869 [Joe Orton, Steve Henson]
871 *) Make it possible to have multiple active certificates with the same
872 subject in the CA index file. This is done only if the keyword
873 'unique_subject' is set to 'no' in the main CA section (default
874 if 'CA_default') of the configuration file. The value is saved
875 with the database itself in a separate index attribute file,
876 named like the index file with '.attr' appended to the name.
879 *) X509 verify fixes. Disable broken certificate workarounds when
880 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
881 keyUsage extension present. Don't accept CRLs with unhandled critical
882 extensions: since verify currently doesn't process CRL extensions this
883 rejects a CRL with *any* critical extensions. Add new verify error codes
887 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
888 A clarification of RFC2560 will require the use of OCTET STRINGs and
889 some implementations cannot handle the current raw format. Since OpenSSL
890 copies and compares OCSP nonces as opaque blobs without any attempt at
891 parsing them this should not create any compatibility issues.
894 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
895 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
896 this HMAC (and other) operations are several times slower than OpenSSL
900 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
901 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
903 *) Use the correct content when signing type "other".
906 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
908 *) Fix various bugs revealed by running the NISCC test suite:
910 Stop out of bounds reads in the ASN1 code when presented with
911 invalid tags (CAN-2003-0543 and CAN-2003-0544).
913 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
915 If verify callback ignores invalid public key errors don't try to check
916 certificate signature with the NULL public key.
920 *) New -ignore_err option in ocsp application to stop the server
921 exiting on the first error in a request.
924 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
925 if the server requested one: as stated in TLS 1.0 and SSL 3.0
929 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
930 extra data after the compression methods not only for TLS 1.0
931 but also for SSL 3.0 (as required by the specification).
932 [Bodo Moeller; problem pointed out by Matthias Loepfe]
934 *) Change X509_certificate_type() to mark the key as exported/exportable
935 when it's 512 *bits* long, not 512 bytes.
938 *) Change AES_cbc_encrypt() so it outputs exact multiple of
939 blocks during encryption.
942 *) Various fixes to base64 BIO and non blocking I/O. On write
943 flushes were not handled properly if the BIO retried. On read
944 data was not being buffered properly and had various logic bugs.
945 This also affects blocking I/O when the data being decoded is a
949 *) Various S/MIME bugfixes and compatibility changes:
950 output correct application/pkcs7 MIME type if
951 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
952 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
953 of files as .eml work). Correctly handle very long lines in MIME
957 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
959 *) Countermeasure against the Klima-Pokorny-Rosa extension of
960 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
961 a protocol version number mismatch like a decryption error
962 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
965 *) Turn on RSA blinding by default in the default implementation
966 to avoid a timing attack. Applications that don't want it can call
967 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
968 They would be ill-advised to do so in most cases.
969 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
971 *) Change RSA blinding code so that it works when the PRNG is not
972 seeded (in this case, the secret RSA exponent is abused as
973 an unpredictable seed -- if it is not unpredictable, there
974 is no point in blinding anyway). Make RSA blinding thread-safe
975 by remembering the creator's thread ID in rsa->blinding and
976 having all other threads use local one-time blinding factors
977 (this requires more computation than sharing rsa->blinding, but
978 avoids excessive locking; and if an RSA object is not shared
979 between threads, blinding will still be very fast).
982 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
983 ENGINE as defaults for all supported algorithms irrespective of
984 the 'flags' parameter. 'flags' is now honoured, so applications
985 should make sure they are passing it correctly.
988 *) Target "mingw" now allows native Windows code to be generated in
989 the Cygwin environment as well as with the MinGW compiler.
992 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
994 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
995 via timing by performing a MAC computation even if incorrrect
996 block cipher padding has been found. This is a countermeasure
997 against active attacks where the attacker has to distinguish
998 between bad padding and a MAC verification error. (CAN-2003-0078)
1000 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1001 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1002 Martin Vuagnoux (EPFL, Ilion)]
1004 *) Make the no-err option work as intended. The intention with no-err
1005 is not to have the whole error stack handling routines removed from
1006 libcrypto, it's only intended to remove all the function name and
1007 reason texts, thereby removing some of the footprint that may not
1008 be interesting if those errors aren't displayed anyway.
1010 NOTE: it's still possible for any application or module to have it's
1011 own set of error texts inserted. The routines are there, just not
1012 used by default when no-err is given.
1015 *) Add support for FreeBSD on IA64.
1016 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1018 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1019 Kerberos function mit_des_cbc_cksum(). Before this change,
1020 the value returned by DES_cbc_cksum() was like the one from
1021 mit_des_cbc_cksum(), except the bytes were swapped.
1022 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1024 *) Allow an application to disable the automatic SSL chain building.
1025 Before this a rather primitive chain build was always performed in
1026 ssl3_output_cert_chain(): an application had no way to send the
1027 correct chain if the automatic operation produced an incorrect result.
1029 Now the chain builder is disabled if either:
1031 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1033 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1035 The reasoning behind this is that an application would not want the
1036 auto chain building to take place if extra chain certificates are
1037 present and it might also want a means of sending no additional
1038 certificates (for example the chain has two certificates and the
1042 *) Add the possibility to build without the ENGINE framework.
1043 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1045 *) Under Win32 gmtime() can return NULL: check return value in
1046 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1049 *) DSA routines: under certain error conditions uninitialized BN objects
1050 could be freed. Solution: make sure initialization is performed early
1051 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1052 Nils Larsch <nla@trustcenter.de> via PR#459)
1055 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1056 checked on reconnect on the client side, therefore session resumption
1057 could still fail with a "ssl session id is different" error. This
1058 behaviour is masked when SSL_OP_ALL is used due to
1059 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1060 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1061 followup to PR #377.
1064 *) IA-32 assembler support enhancements: unified ELF targets, support
1065 for SCO/Caldera platforms, fix for Cygwin shared build.
1068 *) Add support for FreeBSD on sparc64. As a consequence, support for
1069 FreeBSD on non-x86 processors is separate from x86 processors on
1070 the config script, much like the NetBSD support.
1071 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1073 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1075 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1078 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1079 code (06) was taken as the first octet of the session ID and the last
1080 octet was ignored consequently. As a result SSLv2 client side session
1081 caching could not have worked due to the session ID mismatch between
1083 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1087 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1088 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1092 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1093 seems that in spite of existing for more than a year, many application
1094 author have done nothing to provide the necessary callbacks, which
1095 means that this particular engine will not work properly anywhere.
1096 This is a very unfortunate situation which forces us, in the name
1097 of usability, to give the hw_ncipher.c a static lock, which is part
1099 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1100 appear in 0.9.8 or later. We EXPECT application authors to have
1101 dealt properly with this when 0.9.8 is released (unless we actually
1102 make such changes in the libcrypto locking code that changes will
1103 have to be made anyway).
1106 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1107 octets have been read, EOF or an error occurs. Without this change
1108 some truncated ASN1 structures will not produce an error.
1111 *) Disable Heimdal support, since it hasn't been fully implemented.
1112 Still give the possibility to force the use of Heimdal, but with
1113 warnings and a request that patches get sent to openssl-dev.
1116 *) Add the VC-CE target, introduce the WINCE sysname, and add
1117 INSTALL.WCE and appropriate conditionals to make it build.
1118 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1120 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1121 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1122 edit numbers of the version.
1123 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1125 *) Introduce safe string copy and catenation functions
1126 (BUF_strlcpy() and BUF_strlcat()).
1127 [Ben Laurie (CHATS) and Richard Levitte]
1129 *) Avoid using fixed-size buffers for one-line DNs.
1130 [Ben Laurie (CHATS)]
1132 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1133 resizing buffers containing secrets, and use where appropriate.
1134 [Ben Laurie (CHATS)]
1136 *) Avoid using fixed size buffers for configuration file location.
1137 [Ben Laurie (CHATS)]
1139 *) Avoid filename truncation for various CA files.
1140 [Ben Laurie (CHATS)]
1142 *) Use sizeof in preference to magic numbers.
1143 [Ben Laurie (CHATS)]
1145 *) Avoid filename truncation in cert requests.
1146 [Ben Laurie (CHATS)]
1148 *) Add assertions to check for (supposedly impossible) buffer
1150 [Ben Laurie (CHATS)]
1152 *) Don't cache truncated DNS entries in the local cache (this could
1153 potentially lead to a spoofing attack).
1154 [Ben Laurie (CHATS)]
1156 *) Fix various buffers to be large enough for hex/decimal
1157 representations in a platform independent manner.
1158 [Ben Laurie (CHATS)]
1160 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1161 resizing buffers containing secrets, and use where appropriate.
1162 [Ben Laurie (CHATS)]
1164 *) Add BIO_indent() to avoid much slightly worrying code to do
1166 [Ben Laurie (CHATS)]
1168 *) Convert sprintf()/BIO_puts() to BIO_printf().
1169 [Ben Laurie (CHATS)]
1171 *) buffer_gets() could terminate with the buffer only half
1173 [Ben Laurie (CHATS)]
1175 *) Add assertions to prevent user-supplied crypto functions from
1176 overflowing internal buffers by having large block sizes, etc.
1177 [Ben Laurie (CHATS)]
1179 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1181 [Ben Laurie (CHATS)]
1183 *) Eliminate unused copy of key in RC4.
1184 [Ben Laurie (CHATS)]
1186 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1187 [Ben Laurie (CHATS)]
1189 *) Fix off-by-one error in EGD path.
1190 [Ben Laurie (CHATS)]
1192 *) If RANDFILE path is too long, ignore instead of truncating.
1193 [Ben Laurie (CHATS)]
1195 *) Eliminate unused and incorrectly sized X.509 structure
1197 [Ben Laurie (CHATS)]
1199 *) Eliminate unused and dangerous function knumber().
1200 [Ben Laurie (CHATS)]
1202 *) Eliminate unused and dangerous structure, KSSL_ERR.
1203 [Ben Laurie (CHATS)]
1205 *) Protect against overlong session ID context length in an encoded
1206 session object. Since these are local, this does not appear to be
1208 [Ben Laurie (CHATS)]
1210 *) Change from security patch (see 0.9.6e below) that did not affect
1211 the 0.9.6 release series:
1213 Remote buffer overflow in SSL3 protocol - an attacker could
1214 supply an oversized master key in Kerberos-enabled versions.
1216 [Ben Laurie (CHATS)]
1218 *) Change the SSL kerb5 codes to match RFC 2712.
1221 *) Make -nameopt work fully for req and add -reqopt switch.
1222 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1224 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1225 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1227 *) Make sure tests can be performed even if the corresponding algorithms
1228 have been removed entirely. This was also the last step to make
1229 OpenSSL compilable with DJGPP under all reasonable conditions.
1230 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1232 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1233 to allow version independent disabling of normally unselected ciphers,
1234 which may be activated as a side-effect of selecting a single cipher.
1236 (E.g., cipher list string "RSA" enables ciphersuites that are left
1237 out of "ALL" because they do not provide symmetric encryption.
1238 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1239 [Lutz Jaenicke, Bodo Moeller]
1241 *) Add appropriate support for separate platform-dependent build
1242 directories. The recommended way to make a platform-dependent
1243 build directory is the following (tested on Linux), maybe with
1246 # Place yourself outside of the OpenSSL source tree. In
1247 # this example, the environment variable OPENSSL_SOURCE
1248 # is assumed to contain the absolute OpenSSL source directory.
1249 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1250 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1251 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1252 mkdir -p `dirname $F`
1253 ln -s $OPENSSL_SOURCE/$F $F
1256 To be absolutely sure not to disturb the source tree, a "make clean"
1257 is a good thing. If it isn't successfull, don't worry about it,
1258 it probably means the source directory is very clean.
1261 *) Make sure any ENGINE control commands make local copies of string
1262 pointers passed to them whenever necessary. Otherwise it is possible
1263 the caller may have overwritten (or deallocated) the original string
1264 data when a later ENGINE operation tries to use the stored values.
1265 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1267 *) Improve diagnostics in file reading and command-line digests.
1268 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1270 *) Add AES modes CFB and OFB to the object database. Correct an
1271 error in AES-CFB decryption.
1274 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1275 allows existing EVP_CIPHER_CTX structures to be reused after
1276 calling EVP_*Final(). This behaviour is used by encryption
1277 BIOs and some applications. This has the side effect that
1278 applications must explicitly clean up cipher contexts with
1279 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1282 *) Check the values of dna and dnb in bn_mul_recursive before calling
1283 bn_mul_comba (a non zero value means the a or b arrays do not contain
1284 n2 elements) and fallback to bn_mul_normal if either is not zero.
1287 *) Fix escaping of non-ASCII characters when using the -subj option
1288 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1291 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1292 form for "surname", serialNumber has no short form.
1293 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1294 therefore remove "mail" short name for "internet 7".
1295 The OID for unique identifiers in X509 certificates is
1296 x500UniqueIdentifier, not uniqueIdentifier.
1297 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1300 *) Add an "init" command to the ENGINE config module and auto initialize
1301 ENGINEs. Without any "init" command the ENGINE will be initialized
1302 after all ctrl commands have been executed on it. If init=1 the
1303 ENGINE is initailized at that point (ctrls before that point are run
1304 on the uninitialized ENGINE and after on the initialized one). If
1305 init=0 then the ENGINE will not be iniatialized at all.
1308 *) Fix the 'app_verify_callback' interface so that the user-defined
1309 argument is actually passed to the callback: In the
1310 SSL_CTX_set_cert_verify_callback() prototype, the callback
1311 declaration has been changed from
1314 int (*cb)(X509_STORE_CTX *,void *);
1315 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1316 i=s->ctx->app_verify_callback(&ctx)
1317 has been changed into
1318 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1320 To update applications using SSL_CTX_set_cert_verify_callback(),
1321 a dummy argument can be added to their callback functions.
1322 [D. K. Smetters <smetters@parc.xerox.com>]
1324 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1325 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1327 *) Add and OPENSSL_LOAD_CONF define which will cause
1328 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1329 This allows older applications to transparently support certain
1330 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1331 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1332 load the config file and OPENSSL_add_all_algorithms_conf() which will
1333 always load it have also been added.
1336 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1337 Adjust NIDs and EVP layer.
1338 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1340 *) Config modules support in openssl utility.
1342 Most commands now load modules from the config file,
1343 though in a few (such as version) this isn't done
1344 because it couldn't be used for anything.
1346 In the case of ca and req the config file used is
1347 the same as the utility itself: that is the -config
1348 command line option can be used to specify an
1352 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1353 use "openssl_conf" if filename is NULL use default openssl config file.
1356 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1357 config section name. Add a new flag to tolerate a missing config file
1358 and move code to CONF_modules_load_file().
1361 *) Support for crypto accelerator cards from Accelerated Encryption
1362 Processing, www.aep.ie. (Use engine 'aep')
1363 The support was copied from 0.9.6c [engine] and adapted/corrected
1364 to work with the new engine framework.
1365 [AEP Inc. and Richard Levitte]
1367 *) Support for SureWare crypto accelerator cards from Baltimore
1368 Technologies. (Use engine 'sureware')
1369 The support was copied from 0.9.6c [engine] and adapted
1370 to work with the new engine framework.
1373 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1374 make the newer ENGINE framework commands for the CHIL engine work.
1375 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1377 *) Make it possible to produce shared libraries on ReliantUNIX.
1378 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1380 *) Add the configuration target debug-linux-ppro.
1381 Make 'openssl rsa' use the general key loading routines
1382 implemented in apps.c, and make those routines able to
1383 handle the key format FORMAT_NETSCAPE and the variant
1385 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1387 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1388 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1390 *) Add -keyform to rsautl, and document -engine.
1391 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1393 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1394 BIO_R_NO_SUCH_FILE error code rather than the generic
1395 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1398 *) Add new functions
1400 ERR_peek_last_error_line
1401 ERR_peek_last_error_line_data.
1402 These are similar to
1405 ERR_peek_error_line_data,
1406 but report on the latest error recorded rather than the first one
1407 still in the error queue.
1408 [Ben Laurie, Bodo Moeller]
1410 *) default_algorithms option in ENGINE config module. This allows things
1412 default_algorithms = ALL
1413 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1416 *) Prelminary ENGINE config module.
1419 *) New experimental application configuration code.
1422 *) Change the AES code to follow the same name structure as all other
1423 symmetric ciphers, and behave the same way. Move everything to
1424 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1425 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1427 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1428 [Ben Laurie and Theo de Raadt]
1430 *) Add option to output public keys in req command.
1431 [Massimiliano Pala madwolf@openca.org]
1433 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1434 (up to about 10% better than before for P-192 and P-224).
1437 *) New functions/macros
1439 SSL_CTX_set_msg_callback(ctx, cb)
1440 SSL_CTX_set_msg_callback_arg(ctx, arg)
1441 SSL_set_msg_callback(ssl, cb)
1442 SSL_set_msg_callback_arg(ssl, arg)
1444 to request calling a callback function
1446 void cb(int write_p, int version, int content_type,
1447 const void *buf, size_t len, SSL *ssl, void *arg)
1449 whenever a protocol message has been completely received
1450 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1451 protocol version according to which the SSL library interprets
1452 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1453 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1454 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1455 specification (change_cipher_spec(20), alert(21), handshake(22)).
1456 'buf' and 'len' point to the actual message, 'ssl' to the
1457 SSL object, and 'arg' is the application-defined value set by
1458 SSL[_CTX]_set_msg_callback_arg().
1460 'openssl s_client' and 'openssl s_server' have new '-msg' options
1461 to enable a callback that displays all protocol messages.
1464 *) Change the shared library support so shared libraries are built as
1465 soon as the corresponding static library is finished, and thereby get
1466 openssl and the test programs linked against the shared library.
1467 This still only happens when the keyword "shard" has been given to
1468 the configuration scripts.
1470 NOTE: shared library support is still an experimental thing, and
1471 backward binary compatibility is still not guaranteed.
1472 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1474 *) Add support for Subject Information Access extension.
1475 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1477 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1478 additional bytes when new memory had to be allocated, not just
1479 when reusing an existing buffer.
1482 *) New command line and configuration option 'utf8' for the req command.
1483 This allows field values to be specified as UTF8 strings.
1486 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1487 runs for the former and machine-readable output for the latter.
1490 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1491 of the e-mail address in the DN (i.e., it will go into a certificate
1492 extension only). The new configuration file option 'email_in_dn = no'
1493 has the same effect.
1494 [Massimiliano Pala madwolf@openca.org]
1496 *) Change all functions with names starting with des_ to be starting
1497 with DES_ instead. Add wrappers that are compatible with libdes,
1498 but are named _ossl_old_des_*. Finally, add macros that map the
1499 des_* symbols to the corresponding _ossl_old_des_* if libdes
1500 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1501 desired, the des_* symbols will be mapped to DES_*, with one
1504 Since we provide two compatibility mappings, the user needs to
1505 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1506 compatibility is desired. The default (i.e., when that macro
1507 isn't defined) is OpenSSL 0.9.6c compatibility.
1509 There are also macros that enable and disable the support of old
1510 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1511 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1512 are defined, the default will apply: to support the old des routines.
1514 In either case, one must include openssl/des.h to get the correct
1515 definitions. Do not try to just include openssl/des_old.h, that
1518 NOTE: This is a major break of an old API into a new one. Software
1519 authors are encouraged to switch to the DES_ style functions. Some
1520 time in the future, des_old.h and the libdes compatibility functions
1521 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1522 default), and then completely removed.
1525 *) Test for certificates which contain unsupported critical extensions.
1526 If such a certificate is found during a verify operation it is
1527 rejected by default: this behaviour can be overridden by either
1528 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1529 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1530 X509_supported_extension() has also been added which returns 1 if a
1531 particular extension is supported.
1534 *) Modify the behaviour of EVP cipher functions in similar way to digests
1535 to retain compatibility with existing code.
1538 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1539 compatibility with existing code. In particular the 'ctx' parameter does
1540 not have to be to be initialized before the call to EVP_DigestInit() and
1541 it is tidied up after a call to EVP_DigestFinal(). New function
1542 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1543 EVP_MD_CTX_copy() changed to not require the destination to be
1544 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1545 requires the destination to be valid.
1547 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1548 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1551 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1552 so that complete 'Handshake' protocol structures are kept in memory
1553 instead of overwriting 'msg_type' and 'length' with 'body' data.
1556 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1557 [Massimo Santin via Richard Levitte]
1559 *) Major restructuring to the underlying ENGINE code. This includes
1560 reduction of linker bloat, separation of pure "ENGINE" manipulation
1561 (initialisation, etc) from functionality dealing with implementations
1562 of specific crypto iterfaces. This change also introduces integrated
1563 support for symmetric ciphers and digest implementations - so ENGINEs
1564 can now accelerate these by providing EVP_CIPHER and EVP_MD
1565 implementations of their own. This is detailed in crypto/engine/README
1566 as it couldn't be adequately described here. However, there are a few
1567 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1568 were changed in the original introduction of ENGINE code have now
1569 reverted back - the hooking from this code to ENGINE is now a good
1570 deal more passive and at run-time, operations deal directly with
1571 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1572 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1573 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1574 they were not being used by the framework as there is no concept of a
1575 BIGNUM_METHOD and they could not be generalised to the new
1576 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1577 ENGINE_cpy() has been removed as it cannot be consistently defined in
1581 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1584 *) Change mkdef.pl to sort symbols that get the same entry number,
1585 and make sure the automatically generated functions ERR_load_*
1586 become part of libeay.num as well.
1589 *) New function SSL_renegotiate_pending(). This returns true once
1590 renegotiation has been requested (either SSL_renegotiate() call
1591 or HelloRequest/ClientHello receveived from the peer) and becomes
1592 false once a handshake has been completed.
1593 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1594 sends a HelloRequest, but does not ensure that a handshake takes
1595 place. SSL_renegotiate_pending() is useful for checking if the
1596 client has followed the request.)
1599 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1600 By default, clients may request session resumption even during
1601 renegotiation (if session ID contexts permit); with this option,
1602 session resumption is possible only in the first handshake.
1604 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1605 more bits available for options that should not be part of
1606 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1609 *) Add some demos for certificate and certificate request creation.
1612 *) Make maximum certificate chain size accepted from the peer application
1613 settable (SSL*_get/set_max_cert_list()), as proposed by
1614 "Douglas E. Engert" <deengert@anl.gov>.
1617 *) Add support for shared libraries for Unixware-7
1618 (Boyd Lynn Gerber <gerberb@zenez.com>).
1621 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1622 be done prior to destruction. Use this to unload error strings from
1623 ENGINEs that load their own error strings. NB: This adds two new API
1624 functions to "get" and "set" this destroy handler in an ENGINE.
1627 *) Alter all existing ENGINE implementations (except "openssl" and
1628 "openbsd") to dynamically instantiate their own error strings. This
1629 makes them more flexible to be built both as statically-linked ENGINEs
1630 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1631 Also, add stub code to each that makes building them as self-contained
1632 shared-libraries easier (see README.ENGINE).
1635 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1636 implementations into applications that are completely implemented in
1637 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1638 commands that can be used to configure what shared-library to load and
1639 to control aspects of the way it is handled. Also, made an update to
1640 the README.ENGINE file that brings its information up-to-date and
1641 provides some information and instructions on the "dynamic" ENGINE
1642 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1645 *) Make it possible to unload ranges of ERR strings with a new
1646 "ERR_unload_strings" function.
1649 *) Add a copy() function to EVP_MD.
1652 *) Make EVP_MD routines take a context pointer instead of just the
1653 md_data void pointer.
1656 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1657 that the digest can only process a single chunk of data
1658 (typically because it is provided by a piece of
1659 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1660 is only going to provide a single chunk of data, and hence the
1661 framework needn't accumulate the data for oneshot drivers.
1664 *) As with "ERR", make it possible to replace the underlying "ex_data"
1665 functions. This change also alters the storage and management of global
1666 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1667 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1668 index counters. The API functions that use this state have been changed
1669 to take a "class_index" rather than pointers to the class's local STACK
1670 and counter, and there is now an API function to dynamically create new
1671 classes. This centralisation allows us to (a) plug a lot of the
1672 thread-safety problems that existed, and (b) makes it possible to clean
1673 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1674 such data would previously have always leaked in application code and
1675 workarounds were in place to make the memory debugging turn a blind eye
1676 to it. Application code that doesn't use this new function will still
1677 leak as before, but their memory debugging output will announce it now
1678 rather than letting it slide.
1680 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1681 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1682 has a return value to indicate success or failure.
1685 *) Make it possible to replace the underlying "ERR" functions such that the
1686 global state (2 LHASH tables and 2 locks) is only used by the "default"
1687 implementation. This change also adds two functions to "get" and "set"
1688 the implementation prior to it being automatically set the first time
1689 any other ERR function takes place. Ie. an application can call "get",
1690 pass the return value to a module it has just loaded, and that module
1691 can call its own "set" function using that value. This means the
1692 module's "ERR" operations will use (and modify) the error state in the
1693 application and not in its own statically linked copy of OpenSSL code.
1696 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1697 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1698 the operation, and provides a more encapsulated way for external code
1699 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1700 to use these functions rather than manually incrementing the counts.
1702 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1705 *) Add EVP test program.
1708 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1711 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1712 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1713 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1714 These allow a CRL to be built without having to access X509_CRL fields
1715 directly. Modify 'ca' application to use new functions.
1718 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1719 bug workarounds. Rollback attack detection is a security feature.
1720 The problem will only arise on OpenSSL servers when TLSv1 is not
1721 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1722 Software authors not wanting to support TLSv1 will have special reasons
1723 for their choice and can explicitly enable this option.
1724 [Bodo Moeller, Lutz Jaenicke]
1726 *) Rationalise EVP so it can be extended: don't include a union of
1727 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1728 (similar to those existing for EVP_CIPHER_CTX).
1733 EVP_MD_CTX_init(&md); /* new function call */
1734 EVP_DigestInit(&md, EVP_sha1());
1735 EVP_DigestUpdate(&md, in, len);
1736 EVP_DigestFinal(&md, out, NULL);
1737 EVP_MD_CTX_cleanup(&md); /* new function call */
1741 *) Make DES key schedule conform to the usual scheme, as well as
1742 correcting its structure. This means that calls to DES functions
1743 now have to pass a pointer to a des_key_schedule instead of a
1744 plain des_key_schedule (which was actually always a pointer
1747 des_key_schedule ks;
1749 des_set_key_checked(..., &ks);
1750 des_ncbc_encrypt(..., &ks, ...);
1752 (Note that a later change renames 'des_...' into 'DES_...'.)
1755 *) Initial reduction of linker bloat: the use of some functions, such as
1756 PEM causes large amounts of unused functions to be linked in due to
1757 poor organisation. For example pem_all.c contains every PEM function
1758 which has a knock on effect of linking in large amounts of (unused)
1759 ASN1 code. Grouping together similar functions and splitting unrelated
1760 functions prevents this.
1763 *) Cleanup of EVP macros.
1766 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1767 correct _ecb suffix.
1770 *) Add initial OCSP responder support to ocsp application. The
1771 revocation information is handled using the text based index
1772 use by the ca application. The responder can either handle
1773 requests generated internally, supplied in files (for example
1774 via a CGI script) or using an internal minimal server.
1777 *) Add configuration choices to get zlib compression for TLS.
1780 *) Changes to Kerberos SSL for RFC 2712 compliance:
1781 1. Implemented real KerberosWrapper, instead of just using
1782 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1783 2. Implemented optional authenticator field of KerberosWrapper.
1785 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1786 and authenticator structs; see crypto/krb5/.
1788 Generalized Kerberos calls to support multiple Kerberos libraries.
1789 [Vern Staats <staatsvr@asc.hpc.mil>,
1790 Jeffrey Altman <jaltman@columbia.edu>
1791 via Richard Levitte]
1793 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1794 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1795 values for each of the key sizes rather than having just
1796 parameters (and 'speed' generating keys each time).
1799 *) Speed up EVP routines.
1802 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1803 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1804 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1805 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1807 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1808 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1809 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1812 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1814 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1817 *) Added the OS2-EMX target.
1818 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1820 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1821 to support NCONF routines in extension code. New function CONF_set_nconf()
1822 to allow functions which take an NCONF to also handle the old LHASH
1823 structure: this means that the old CONF compatible routines can be
1824 retained (in particular wrt extensions) without having to duplicate the
1825 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1828 *) Enhance the general user interface with mechanisms for inner control
1829 and with possibilities to have yes/no kind of prompts.
1832 *) Change all calls to low level digest routines in the library and
1833 applications to use EVP. Add missing calls to HMAC_cleanup() and
1834 don't assume HMAC_CTX can be copied using memcpy().
1835 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1837 *) Add the possibility to control engines through control names but with
1838 arbitrary arguments instead of just a string.
1839 Change the key loaders to take a UI_METHOD instead of a callback
1840 function pointer. NOTE: this breaks binary compatibility with earlier
1841 versions of OpenSSL [engine].
1842 Adapt the nCipher code for these new conditions and add a card insertion
1846 *) Enhance the general user interface with mechanisms to better support
1847 dialog box interfaces, application-defined prompts, the possibility
1848 to use defaults (for example default passwords from somewhere else)
1849 and interrupts/cancellations.
1852 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1853 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1856 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1857 tidy up some unnecessarily weird code in 'sk_new()').
1858 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1860 *) Change the key loading routines for ENGINEs to use the same kind
1861 callback (pem_password_cb) as all other routines that need this
1865 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1866 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1867 than this minimum value is recommended.
1870 *) New random seeder for OpenVMS, using the system process statistics
1871 that are easily reachable.
1874 *) Windows apparently can't transparently handle global
1875 variables defined in DLLs. Initialisations such as:
1877 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1879 wont compile. This is used by the any applications that need to
1880 declare their own ASN1 modules. This was fixed by adding the option
1881 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1882 needed for static libraries under Win32.
1885 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1886 setting of purpose and trust fields. New X509_STORE trust and
1887 purpose functions and tidy up setting in other SSL functions.
1890 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1891 structure. These are inherited by X509_STORE_CTX when it is
1892 initialised. This allows various defaults to be set in the
1893 X509_STORE structure (such as flags for CRL checking and custom
1894 purpose or trust settings) for functions which only use X509_STORE_CTX
1895 internally such as S/MIME.
1897 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1898 trust settings if they are not set in X509_STORE. This allows X509_STORE
1899 purposes and trust (in S/MIME for example) to override any set by default.
1901 Add command line options for CRL checking to smime, s_client and s_server
1905 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1906 are set then the CRL is looked up in the X509_STORE structure and
1907 its validity and signature checked, then if the certificate is found
1908 in the CRL the verify fails with a revoked error.
1910 Various new CRL related callbacks added to X509_STORE_CTX structure.
1912 Command line options added to 'verify' application to support this.
1914 This needs some additional work, such as being able to handle multiple
1915 CRLs with different times, extension based lookup (rather than just
1916 by subject name) and ultimately more complete V2 CRL extension
1920 *) Add a general user interface API (crypto/ui/). This is designed
1921 to replace things like des_read_password and friends (backward
1922 compatibility functions using this new API are provided).
1923 The purpose is to remove prompting functions from the DES code
1924 section as well as provide for prompting through dialog boxes in
1925 a window system and the like.
1928 *) Add "ex_data" support to ENGINE so implementations can add state at a
1929 per-structure level rather than having to store it globally.
1932 *) Make it possible for ENGINE structures to be copied when retrieved by
1933 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1934 This causes the "original" ENGINE structure to act like a template,
1935 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1936 operational state can be localised to each ENGINE structure, despite the
1937 fact they all share the same "methods". New ENGINE structures returned in
1938 this case have no functional references and the return value is the single
1939 structural reference. This matches the single structural reference returned
1940 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1944 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1945 needs to match any other type at all we need to manually clear the
1949 *) Changes to the "openssl engine" utility to include;
1950 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1951 about an ENGINE's available control commands.
1952 - executing control commands from command line arguments using the
1953 '-pre' and '-post' switches. '-post' is only used if '-t' is
1954 specified and the ENGINE is successfully initialised. The syntax for
1955 the individual commands are colon-separated, for example;
1956 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1959 *) New dynamic control command support for ENGINEs. ENGINEs can now
1960 declare their own commands (numbers), names (strings), descriptions,
1961 and input types for run-time discovery by calling applications. A
1962 subset of these commands are implicitly classed as "executable"
1963 depending on their input type, and only these can be invoked through
1964 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1965 can be based on user input, config files, etc). The distinction is
1966 that "executable" commands cannot return anything other than a boolean
1967 result and can only support numeric or string input, whereas some
1968 discoverable commands may only be for direct use through
1969 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1970 pointers, or other custom uses. The "executable" commands are to
1971 support parameterisations of ENGINE behaviour that can be
1972 unambiguously defined by ENGINEs and used consistently across any
1973 OpenSSL-based application. Commands have been added to all the
1974 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1975 control over shared-library paths without source code alterations.
1978 *) Changed all ENGINE implementations to dynamically allocate their
1979 ENGINEs rather than declaring them statically. Apart from this being
1980 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1981 this also allows the implementations to compile without using the
1982 internal engine_int.h header.
1985 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1986 'const' value. Any code that should be able to modify a RAND_METHOD
1987 should already have non-const pointers to it (ie. they should only
1988 modify their own ones).
1991 *) Made a variety of little tweaks to the ENGINE code.
1992 - "atalla" and "ubsec" string definitions were moved from header files
1993 to C code. "nuron" string definitions were placed in variables
1994 rather than hard-coded - allowing parameterisation of these values
1995 later on via ctrl() commands.
1996 - Removed unused "#if 0"'d code.
1997 - Fixed engine list iteration code so it uses ENGINE_free() to release
1998 structural references.
1999 - Constified the RAND_METHOD element of ENGINE structures.
2000 - Constified various get/set functions as appropriate and added
2001 missing functions (including a catch-all ENGINE_cpy that duplicates
2002 all ENGINE values onto a new ENGINE except reference counts/state).
2003 - Removed NULL parameter checks in get/set functions. Setting a method
2004 or function to NULL is a way of cancelling out a previously set
2005 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2006 and doesn't justify the extra error symbols and code.
2007 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2008 flags from engine_int.h to engine.h.
2009 - Changed prototypes for ENGINE handler functions (init(), finish(),
2010 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2013 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2014 to the algorithm using long division. The binary algorithm can be
2015 used only if the modulus is odd. On 32-bit systems, it is faster
2016 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2017 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2018 up to 450 bits. In 64-bit environments, the binary algorithm
2019 appears to be advantageous for much longer moduli; here we use it
2020 for moduli up to 2048 bits.
2023 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2024 could not support the combine flag in choice fields.
2027 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2028 extensions from a certificate request to the certificate.
2031 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2032 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2033 file: this allows the display of the certificate about to be
2034 signed to be customised, to allow certain fields to be included
2035 or excluded and extension details. The old system didn't display
2036 multicharacter strings properly, omitted fields not in the policy
2037 and couldn't display additional details such as extensions.
2040 *) Function EC_POINTs_mul for multiple scalar multiplication
2041 of an arbitrary number of elliptic curve points
2042 \sum scalars[i]*points[i],
2043 optionally including the generator defined for the EC_GROUP:
2044 scalar*generator + \sum scalars[i]*points[i].
2046 EC_POINT_mul is a simple wrapper function for the typical case
2047 that the point list has just one item (besides the optional
2051 *) First EC_METHODs for curves over GF(p):
2053 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2054 operations and provides various method functions that can also
2055 operate with faster implementations of modular arithmetic.
2057 EC_GFp_mont_method() reuses most functions that are part of
2058 EC_GFp_simple_method, but uses Montgomery arithmetic.
2060 [Bodo Moeller; point addition and point doubling
2061 implementation directly derived from source code provided by
2062 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2064 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2065 crypto/ec/ec_lib.c):
2067 Curves are EC_GROUP objects (with an optional group generator)
2068 based on EC_METHODs that are built into the library.
2070 Points are EC_POINT objects based on EC_GROUP objects.
2072 Most of the framework would be able to handle curves over arbitrary
2073 finite fields, but as there are no obvious types for fields other
2074 than GF(p), some functions are limited to that for now.
2077 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2078 that the file contains a complete HTTP response.
2081 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2082 change the def and num file printf format specifier from "%-40sXXX"
2083 to "%-39s XXX". The latter will always guarantee a space after the
2084 field while the former will cause them to run together if the field
2085 is 40 of more characters long.
2088 *) Constify the cipher and digest 'method' functions and structures
2089 and modify related functions to take constant EVP_MD and EVP_CIPHER
2093 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2094 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2097 *) Modify EVP_Digest*() routines so they now return values. Although the
2098 internal software routines can never fail additional hardware versions
2102 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2104 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2105 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2108 ERR_R_NESTED_ASN1_ERROR
2110 ERR_R_MISSING_ASN1_EOS
2111 were 4 .. 9, conflicting with
2112 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2114 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2115 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2117 Add new error code 'ERR_R_INTERNAL_ERROR'.
2120 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2124 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2125 sets the subject name for a new request or supersedes the
2126 subject name in a given request. Formats that can be parsed are
2127 'CN=Some Name, OU=myOU, C=IT'
2129 'CN=Some Name/OU=myOU/C=IT'.
2131 Add options '-batch' and '-verbose' to 'openssl req'.
2132 [Massimiliano Pala <madwolf@hackmasters.net>]
2134 *) Introduce the possibility to access global variables through
2135 functions on platform were that's the best way to handle exporting
2136 global variables in shared libraries. To enable this functionality,
2137 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2138 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2139 is normally done by Configure or something similar).
2141 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2142 in the source file (foo.c) like this:
2144 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2145 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2147 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2148 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2150 OPENSSL_DECLARE_GLOBAL(int,foo);
2151 #define foo OPENSSL_GLOBAL_REF(foo)
2152 OPENSSL_DECLARE_GLOBAL(double,bar);
2153 #define bar OPENSSL_GLOBAL_REF(bar)
2155 The #defines are very important, and therefore so is including the
2156 header file everywhere where the defined globals are used.
2158 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2159 of ASN.1 items, but that structure is a bit different.
2161 The largest change is in util/mkdef.pl which has been enhanced with
2162 better and easier to understand logic to choose which symbols should
2163 go into the Windows .def files as well as a number of fixes and code
2164 cleanup (among others, algorithm keywords are now sorted
2165 lexicographically to avoid constant rewrites).
2168 *) In BN_div() keep a copy of the sign of 'num' before writing the
2169 result to 'rm' because if rm==num the value will be overwritten
2170 and produce the wrong result if 'num' is negative: this caused
2171 problems with BN_mod() and BN_nnmod().
2174 *) Function OCSP_request_verify(). This checks the signature on an
2175 OCSP request and verifies the signer certificate. The signer
2176 certificate is just checked for a generic purpose and OCSP request
2180 *) Add OCSP_check_validity() function to check the validity of OCSP
2181 responses. OCSP responses are prepared in real time and may only
2182 be a few seconds old. Simply checking that the current time lies
2183 between thisUpdate and nextUpdate max reject otherwise valid responses
2184 caused by either OCSP responder or client clock inaccuracy. Instead
2185 we allow thisUpdate and nextUpdate to fall within a certain period of
2186 the current time. The age of the response can also optionally be
2187 checked. Two new options -validity_period and -status_age added to
2191 *) If signature or public key algorithm is unrecognized print out its
2192 OID rather that just UNKNOWN.
2195 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2196 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2197 ID to be generated from the issuer certificate alone which can then be
2198 passed to OCSP_id_issuer_cmp().
2201 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2202 ASN1 modules to export functions returning ASN1_ITEM pointers
2203 instead of the ASN1_ITEM structures themselves. This adds several
2204 new macros which allow the underlying ASN1 function/structure to
2205 be accessed transparently. As a result code should not use ASN1_ITEM
2206 references directly (such as &X509_it) but instead use the relevant
2207 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2208 use of the new ASN1 code on platforms where exporting structures
2209 is problematical (for example in shared libraries) but exporting
2210 functions returning pointers to structures is not.
2213 *) Add support for overriding the generation of SSL/TLS session IDs.
2214 These callbacks can be registered either in an SSL_CTX or per SSL.
2215 The purpose of this is to allow applications to control, if they wish,
2216 the arbitrary values chosen for use as session IDs, particularly as it
2217 can be useful for session caching in multiple-server environments. A
2218 command-line switch for testing this (and any client code that wishes
2219 to use such a feature) has been added to "s_server".
2220 [Geoff Thorpe, Lutz Jaenicke]
2222 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2223 of the form '#if defined(...) || defined(...) || ...' and
2224 '#if !defined(...) && !defined(...) && ...'. This also avoids
2225 the growing number of special cases it was previously handling.
2228 *) Make all configuration macros available for application by making
2229 sure they are available in opensslconf.h, by giving them names starting
2230 with "OPENSSL_" to avoid conflicts with other packages and by making
2231 sure e_os2.h will cover all platform-specific cases together with
2233 Additionally, it is now possible to define configuration/platform-
2234 specific names (called "system identities"). In the C code, these
2235 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2236 macro with the name beginning with "OPENSSL_SYS_", which is determined
2237 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2241 *) New option -set_serial to 'req' and 'x509' this allows the serial
2242 number to use to be specified on the command line. Previously self
2243 signed certificates were hard coded with serial number 0 and the
2244 CA options of 'x509' had to use a serial number in a file which was
2248 *) New options to 'ca' utility to support V2 CRL entry extensions.
2249 Currently CRL reason, invalidity date and hold instruction are
2250 supported. Add new CRL extensions to V3 code and some new objects.
2253 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2254 disable standard block padding (aka PKCS#5 padding) in the EVP
2255 API, which was previously mandatory. This means that the data is
2256 not padded in any way and so the total length much be a multiple
2257 of the block size, otherwise an error occurs.
2260 *) Initial (incomplete) OCSP SSL support.
2263 *) New function OCSP_parse_url(). This splits up a URL into its host,
2264 port and path components: primarily to parse OCSP URLs. New -url
2265 option to ocsp utility.
2268 *) New nonce behavior. The return value of OCSP_check_nonce() now
2269 reflects the various checks performed. Applications can decide
2270 whether to tolerate certain situations such as an absent nonce
2271 in a response when one was present in a request: the ocsp application
2272 just prints out a warning. New function OCSP_add1_basic_nonce()
2273 this is to allow responders to include a nonce in a response even if
2274 the request is nonce-less.
2277 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2278 skipped when using openssl x509 multiple times on a single input file,
2279 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2282 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2283 set string type: to handle setting ASN1_TIME structures. Fix ca
2284 utility to correctly initialize revocation date of CRLs.
2287 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2288 the clients preferred ciphersuites and rather use its own preferences.
2289 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2290 Internet Explorer by ensuring unchanged hash method during stepup.
2291 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2294 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2295 to aes and add a new 'exist' option to print out symbols that don't
2299 *) Additional options to ocsp utility to allow flags to be set and
2300 additional certificates supplied.
2303 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2304 OCSP client a number of certificate to only verify the response
2308 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2309 handle the new API. Currently only ECB, CBC modes supported. Add new
2312 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2313 Encryption Standard (AES) Ciphersuites for Transport Layer
2314 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2315 not enabled by default and were not part of the "ALL" ciphersuite
2316 alias because they were not yet official; they could be
2317 explicitly requested by specifying the "AESdraft" ciphersuite
2318 group alias. In the final release of OpenSSL 0.9.7, the group
2319 alias is called "AES" and is part of "ALL".)
2320 [Ben Laurie, Steve Henson, Bodo Moeller]
2322 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2323 request to response.
2326 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2327 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2328 extract information from a certificate request. OCSP_response_create()
2329 creates a response and optionally adds a basic response structure.
2330 OCSP_basic_add1_status() adds a complete single response to a basic
2331 response and returns the OCSP_SINGLERESP structure just added (to allow
2332 extensions to be included for example). OCSP_basic_add1_cert() adds a
2333 certificate to a basic response and OCSP_basic_sign() signs a basic
2334 response with various flags. New helper functions ASN1_TIME_check()
2335 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2336 (converts ASN1_TIME to GeneralizedTime).
2339 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2340 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2341 structure from a certificate. X509_pubkey_digest() digests the public_key
2342 contents: this is used in various key identifiers.
2345 *) Make sk_sort() tolerate a NULL argument.
2346 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2348 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2349 passed by the function are trusted implicitly. If any of them signed the
2350 response then it is assumed to be valid and is not verified.
2353 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2354 to data. This was previously part of the PKCS7 ASN1 code. This
2355 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2356 [Steve Henson, reported by Kenneth R. Robinette
2357 <support@securenetterm.com>]
2359 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2360 routines: without these tracing memory leaks is very painful.
2361 Fix leaks in PKCS12 and PKCS7 routines.
2364 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2365 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2366 effectively meant GeneralizedTime would never be used. Now it
2367 is initialised to -1 but X509_time_adj() now has to check the value
2368 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2369 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2370 [Steve Henson, reported by Kenneth R. Robinette
2371 <support@securenetterm.com>]
2373 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2374 result in a zero length in the ASN1_INTEGER structure which was
2375 not consistent with the structure when d2i_ASN1_INTEGER() was used
2376 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2377 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2378 where it did not print out a minus for negative ASN1_INTEGER.
2381 *) Add summary printout to ocsp utility. The various functions which
2382 convert status values to strings have been renamed to:
2383 OCSP_response_status_str(), OCSP_cert_status_str() and
2384 OCSP_crl_reason_str() and are no longer static. New options
2385 to verify nonce values and to disable verification. OCSP response
2386 printout format cleaned up.
2389 *) Add additional OCSP certificate checks. These are those specified
2390 in RFC2560. This consists of two separate checks: the CA of the
2391 certificate being checked must either be the OCSP signer certificate
2392 or the issuer of the OCSP signer certificate. In the latter case the
2393 OCSP signer certificate must contain the OCSP signing extended key
2394 usage. This check is performed by attempting to match the OCSP
2395 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2396 in the OCSP_CERTID structures of the response.
2399 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2400 and related routines. This uses the standard OpenSSL certificate
2401 verify routines to perform initial checks (just CA validity) and
2402 to obtain the certificate chain. Then additional checks will be
2403 performed on the chain. Currently the root CA is checked to see
2404 if it is explicitly trusted for OCSP signing. This is used to set
2405 a root CA as a global signing root: that is any certificate that
2406 chains to that CA is an acceptable OCSP signing certificate.
2409 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2410 extensions from a separate configuration file.
2411 As when reading extensions from the main configuration file,
2412 the '-extensions ...' option may be used for specifying the
2414 [Massimiliano Pala <madwolf@comune.modena.it>]
2416 *) New OCSP utility. Allows OCSP requests to be generated or
2417 read. The request can be sent to a responder and the output
2418 parsed, outputed or printed in text form. Not complete yet:
2419 still needs to check the OCSP response validity.
2422 *) New subcommands for 'openssl ca':
2423 'openssl ca -status <serial>' prints the status of the cert with
2424 the given serial number (according to the index file).
2425 'openssl ca -updatedb' updates the expiry status of certificates
2427 [Massimiliano Pala <madwolf@comune.modena.it>]
2429 *) New '-newreq-nodes' command option to CA.pl. This is like
2430 '-newreq', but calls 'openssl req' with the '-nodes' option
2431 so that the resulting key is not encrypted.
2432 [Damien Miller <djm@mindrot.org>]
2434 *) New configuration for the GNU Hurd.
2435 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2437 *) Initial code to implement OCSP basic response verify. This
2438 is currently incomplete. Currently just finds the signer's
2439 certificate and verifies the signature on the response.
2442 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2443 value of OPENSSLDIR. This is available via the new '-d' option
2444 to 'openssl version', and is also included in 'openssl version -a'.
2447 *) Allowing defining memory allocation callbacks that will be given
2448 file name and line number information in additional arguments
2449 (a const char* and an int). The basic functionality remains, as
2450 well as the original possibility to just replace malloc(),
2451 realloc() and free() by functions that do not know about these
2452 additional arguments. To register and find out the current
2453 settings for extended allocation functions, the following
2454 functions are provided:
2456 CRYPTO_set_mem_ex_functions
2457 CRYPTO_set_locked_mem_ex_functions
2458 CRYPTO_get_mem_ex_functions
2459 CRYPTO_get_locked_mem_ex_functions
2461 These work the same way as CRYPTO_set_mem_functions and friends.
2462 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2463 extended allocation function is enabled.
2464 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2465 a conventional allocation function is enabled.
2466 [Richard Levitte, Bodo Moeller]
2468 *) Finish off removing the remaining LHASH function pointer casts.
2469 There should no longer be any prototype-casting required when using
2470 the LHASH abstraction, and any casts that remain are "bugs". See
2471 the callback types and macros at the head of lhash.h for details
2472 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2475 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2476 If /dev/[u]random devices are not available or do not return enough
2477 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2479 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2480 /etc/entropy will be queried once each in this sequence, quering stops
2481 when enough entropy was collected without querying more sockets.
2484 *) Change the Unix RAND_poll() variant to be able to poll several
2485 random devices, as specified by DEVRANDOM, until a sufficient amount
2486 of data has been collected. We spend at most 10 ms on each file
2487 (select timeout) and read in non-blocking mode. DEVRANDOM now
2488 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2489 (previously it was just the string "/dev/urandom"), so on typical
2490 platforms the 10 ms delay will never occur.
2491 Also separate out the Unix variant to its own file, rand_unix.c.
2492 For VMS, there's a currently-empty rand_vms.c.
2495 *) Move OCSP client related routines to ocsp_cl.c. These
2496 provide utility functions which an application needing
2497 to issue a request to an OCSP responder and analyse the
2498 response will typically need: as opposed to those which an
2499 OCSP responder itself would need which will be added later.
2501 OCSP_request_sign() signs an OCSP request with an API similar
2502 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2503 response. OCSP_response_get1_basic() extracts basic response
2504 from response. OCSP_resp_find_status(): finds and extracts status
2505 information from an OCSP_CERTID structure (which will be created
2506 when the request structure is built). These are built from lower
2507 level functions which work on OCSP_SINGLERESP structures but
2508 wont normally be used unless the application wishes to examine
2509 extensions in the OCSP response for example.
2511 Replace nonce routines with a pair of functions.
2512 OCSP_request_add1_nonce() adds a nonce value and optionally
2513 generates a random value. OCSP_check_nonce() checks the
2514 validity of the nonce in an OCSP response.
2517 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2518 This doesn't copy the supplied OCSP_CERTID and avoids the
2519 need to free up the newly created id. Change return type
2520 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2521 This can then be used to add extensions to the request.
2522 Deleted OCSP_request_new(), since most of its functionality
2523 is now in OCSP_REQUEST_new() (and the case insensitive name
2524 clash) apart from the ability to set the request name which
2525 will be added elsewhere.
2528 *) Update OCSP API. Remove obsolete extensions argument from
2529 various functions. Extensions are now handled using the new
2530 OCSP extension code. New simple OCSP HTTP function which
2531 can be used to send requests and parse the response.
2534 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2535 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2536 uses the special reorder version of SET OF to sort the attributes
2537 and reorder them to match the encoded order. This resolves a long
2538 standing problem: a verify on a PKCS7 structure just after signing
2539 it used to fail because the attribute order did not match the
2540 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2541 it uses the received order. This is necessary to tolerate some broken
2542 software that does not order SET OF. This is handled by encoding
2543 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2544 to produce the required SET OF.
2547 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2548 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2549 files to get correct declarations of the ASN.1 item variables.
2552 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2553 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2554 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2555 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2556 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2557 ASN1_ITEM and no wrapper functions.
2560 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2561 replace the old function pointer based I/O routines. Change most of
2562 the *_d2i_bio() and *_d2i_fp() functions to use these.
2565 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2566 lines, recognice more "algorithms" that can be deselected, and make
2567 it complain about algorithm deselection that isn't recognised.
2570 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2571 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2572 to use new functions. Add NO_ASN1_OLD which can be set to remove
2573 some old style ASN1 functions: this can be used to determine if old
2574 code will still work when these eventually go away.
2577 *) New extension functions for OCSP structures, these follow the
2578 same conventions as certificates and CRLs.
2581 *) New function X509V3_add1_i2d(). This automatically encodes and
2582 adds an extension. Its behaviour can be customised with various
2583 flags to append, replace or delete. Various wrappers added for
2584 certifcates and CRLs.
2587 *) Fix to avoid calling the underlying ASN1 print routine when
2588 an extension cannot be parsed. Correct a typo in the
2589 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2592 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2593 entries for variables.
2596 *) Add functionality to apps/openssl.c for detecting locking
2597 problems: As the program is single-threaded, all we have
2598 to do is register a locking callback using an array for
2599 storing which locks are currently held by the program.
2602 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2603 SSL_get_ex_data_X509_STORE_idx(), which is used in
2604 ssl_verify_cert_chain() and thus can be called at any time
2605 during TLS/SSL handshakes so that thread-safety is essential.
2606 Unfortunately, the ex_data design is not at all suited
2607 for multi-threaded use, so it probably should be abolished.
2610 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2611 [Broadcom, tweaked and integrated by Geoff Thorpe]
2613 *) Move common extension printing code to new function
2614 X509V3_print_extensions(). Reorganise OCSP print routines and
2615 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2618 *) New function X509_signature_print() to remove duplication in some
2622 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2623 set (this was treated exactly the same as SET OF previously). This
2624 is used to reorder the STACK representing the structure to match the
2625 encoding. This will be used to get round a problem where a PKCS7
2626 structure which was signed could not be verified because the STACK
2627 order did not reflect the encoded order.
2630 *) Reimplement the OCSP ASN1 module using the new code.
2633 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2634 for its ASN1 operations. The old style function pointers still exist
2635 for now but they will eventually go away.
2638 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2639 completely replaces the old ASN1 functionality with a table driven
2640 encoder and decoder which interprets an ASN1_ITEM structure describing
2641 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2642 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2643 has also been converted to the new form.
2646 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2647 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2648 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2649 for negative moduli.
2652 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2653 of not touching the result's sign bit.
2656 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2660 *) Changed the LHASH code to use prototypes for callbacks, and created
2661 macros to declare and implement thin (optionally static) functions
2662 that provide type-safety and avoid function pointer casting for the
2663 type-specific callbacks.
2666 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2668 [Veers Staats <staatsvr@asc.hpc.mil>,
2669 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2671 *) Reformat the FAQ so the different questions and answers can be divided
2672 in sections depending on the subject.
2675 *) Have the zlib compression code load ZLIB.DLL dynamically under
2679 *) New function BN_mod_sqrt for computing square roots modulo a prime
2680 (using the probabilistic Tonelli-Shanks algorithm unless
2681 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2682 be handled deterministically).
2683 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2685 *) Make BN_mod_inverse faster by explicitly handling small quotients
2686 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2687 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2690 *) New function BN_kronecker.
2693 *) Fix BN_gcd so that it works on negative inputs; the result is
2694 positive unless both parameters are zero.
2695 Previously something reasonably close to an infinite loop was
2696 possible because numbers could be growing instead of shrinking
2697 in the implementation of Euclid's algorithm.
2700 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2701 sign of the number in question.
2703 Fix BN_is_word(a,w) to work correctly for w == 0.
2705 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2706 because its test if the absolute value of 'a' equals 'w'.
2707 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2708 it exists mostly for use in the implementations of BN_is_zero(),
2709 BN_is_one(), and BN_is_word().
2712 *) New function BN_swap.
2715 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2716 the exponentiation functions are more likely to produce reasonable
2717 results on negative inputs.
2720 *) Change BN_mod_mul so that the result is always non-negative.
2721 Previously, it could be negative if one of the factors was negative;
2722 I don't think anyone really wanted that behaviour.
2725 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2726 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2727 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2728 and add new functions:
2737 BN_mod_lshift1_quick
2741 These functions always generate non-negative results.
2743 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2744 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2746 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2747 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2748 be reduced modulo m.
2749 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2752 The following entry accidentily appeared in the CHANGES file
2753 distributed with OpenSSL 0.9.7. The modifications described in
2754 it do *not* apply to OpenSSL 0.9.7.
2756 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2757 was actually never needed) and in BN_mul(). The removal in BN_mul()
2758 required a small change in bn_mul_part_recursive() and the addition
2759 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2760 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2761 bn_sub_words() and bn_add_words() except they take arrays with
2766 *) In 'openssl passwd', verify passwords read from the terminal
2767 unless the '-salt' option is used (which usually means that
2768 verification would just waste user's time since the resulting
2769 hash is going to be compared with some given password hash)
2770 or the new '-noverify' option is used.
2772 This is an incompatible change, but it does not affect
2773 non-interactive use of 'openssl passwd' (passwords on the command
2774 line, '-stdin' option, '-in ...' option) and thus should not
2778 *) Remove all references to RSAref, since there's no more need for it.
2781 *) Make DSO load along a path given through an environment variable
2782 (SHLIB_PATH) with shl_load().
2785 *) Constify the ENGINE code as a result of BIGNUM constification.
2786 Also constify the RSA code and most things related to it. In a
2787 few places, most notable in the depth of the ASN.1 code, ugly
2788 casts back to non-const were required (to be solved at a later
2792 *) Make it so the openssl application has all engines loaded by default.
2795 *) Constify the BIGNUM routines a little more.
2798 *) Add the following functions:
2800 ENGINE_load_cswift()
2802 ENGINE_load_atalla()
2804 ENGINE_load_builtin_engines()
2806 That way, an application can itself choose if external engines that
2807 are built-in in OpenSSL shall ever be used or not. The benefit is
2808 that applications won't have to be linked with libdl or other dso
2809 libraries unless it's really needed.
2811 Changed 'openssl engine' to load all engines on demand.
2812 Changed the engine header files to avoid the duplication of some
2813 declarations (they differed!).
2816 *) 'openssl engine' can now list capabilities.
2819 *) Better error reporting in 'openssl engine'.
2822 *) Never call load_dh_param(NULL) in s_server.
2825 *) Add engine application. It can currently list engines by name and
2826 identity, and test if they are actually available.
2829 *) Improve RPM specification file by forcing symbolic linking and making
2830 sure the installed documentation is also owned by root.root.
2831 [Damien Miller <djm@mindrot.org>]
2833 *) Give the OpenSSL applications more possibilities to make use of
2834 keys (public as well as private) handled by engines.
2837 *) Add OCSP code that comes from CertCo.
2840 *) Add VMS support for the Rijndael code.
2843 *) Added untested support for Nuron crypto accelerator.
2846 *) Add support for external cryptographic devices. This code was
2847 previously distributed separately as the "engine" branch.
2848 [Geoff Thorpe, Richard Levitte]
2850 *) Rework the filename-translation in the DSO code. It is now possible to
2851 have far greater control over how a "name" is turned into a filename
2852 depending on the operating environment and any oddities about the
2853 different shared library filenames on each system.
2856 *) Support threads on FreeBSD-elf in Configure.
2859 *) Fix for SHA1 assembly problem with MASM: it produces
2860 warnings about corrupt line number information when assembling
2861 with debugging information. This is caused by the overlapping
2863 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2866 NCONF_get_number() has no error checking at all. As a replacement,
2867 NCONF_get_number_e() is defined (_e for "error checking") and is
2868 promoted strongly. The old NCONF_get_number is kept around for
2869 binary backward compatibility.
2870 Make it possible for methods to load from something other than a BIO,
2871 by providing a function pointer that is given a name instead of a BIO.
2872 For example, this could be used to load configuration data from an
2876 *) Fix for non blocking accept BIOs. Added new I/O special reason
2877 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2878 with non blocking I/O was not possible because no retry code was
2879 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2883 *) Added the beginnings of Rijndael support.
2886 *) Fix for bug in DirectoryString mask setting. Add support for
2887 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2888 to allow certificate printing to more controllable, additional
2889 'certopt' option to 'x509' to allow new printing options to be
2893 *) Clean old EAY MD5 hack from e_os.h.
2896 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2898 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2899 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2900 [Joe Orton, Steve Henson]
2902 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2904 *) Fix additional bug revealed by the NISCC test suite:
2906 Stop bug triggering large recursion when presented with
2907 certain ASN.1 tags (CAN-2003-0851)
2910 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
2912 *) Fix various bugs revealed by running the NISCC test suite:
2914 Stop out of bounds reads in the ASN1 code when presented with
2915 invalid tags (CAN-2003-0543 and CAN-2003-0544).
2917 If verify callback ignores invalid public key errors don't try to check
2918 certificate signature with the NULL public key.
2922 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2923 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2927 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2928 extra data after the compression methods not only for TLS 1.0
2929 but also for SSL 3.0 (as required by the specification).
2930 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2932 *) Change X509_certificate_type() to mark the key as exported/exportable
2933 when it's 512 *bits* long, not 512 bytes.
2936 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
2938 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2939 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2940 a protocol version number mismatch like a decryption error
2941 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2944 *) Turn on RSA blinding by default in the default implementation
2945 to avoid a timing attack. Applications that don't want it can call
2946 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2947 They would be ill-advised to do so in most cases.
2948 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2950 *) Change RSA blinding code so that it works when the PRNG is not
2951 seeded (in this case, the secret RSA exponent is abused as
2952 an unpredictable seed -- if it is not unpredictable, there
2953 is no point in blinding anyway). Make RSA blinding thread-safe
2954 by remembering the creator's thread ID in rsa->blinding and
2955 having all other threads use local one-time blinding factors
2956 (this requires more computation than sharing rsa->blinding, but
2957 avoids excessive locking; and if an RSA object is not shared
2958 between threads, blinding will still be very fast).
2961 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
2963 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2964 via timing by performing a MAC computation even if incorrrect
2965 block cipher padding has been found. This is a countermeasure
2966 against active attacks where the attacker has to distinguish
2967 between bad padding and a MAC verification error. (CAN-2003-0078)
2969 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2970 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2971 Martin Vuagnoux (EPFL, Ilion)]
2973 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
2975 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
2976 memory from it's contents. This is done with a counter that will
2977 place alternating values in each byte. This can be used to solve
2978 two issues: 1) the removal of calls to memset() by highly optimizing
2979 compilers, and 2) cleansing with other values than 0, since those can
2980 be read through on certain media, for example a swap space on disk.
2983 *) Bugfix: client side session caching did not work with external caching,
2984 because the session->cipher setting was not restored when reloading
2985 from the external cache. This problem was masked, when
2986 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
2987 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
2990 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
2991 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
2992 [Zeev Lieber <zeev-l@yahoo.com>]
2994 *) Undo an undocumented change introduced in 0.9.6e which caused
2995 repeated calls to OpenSSL_add_all_ciphers() and
2996 OpenSSL_add_all_digests() to be ignored, even after calling
3000 *) Change the default configuration reader to deal with last line not
3001 being properly terminated.
3004 *) Change X509_NAME_cmp() so it applies the special rules on handling
3005 DN values that are of type PrintableString, as well as RDNs of type
3006 emailAddress where the value has the type ia5String.
3007 [stefank@valicert.com via Richard Levitte]
3009 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3010 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3011 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3012 the bitwise-OR of the two for use by the majority of applications
3013 wanting this behaviour, and update the docs. The documented
3014 behaviour and actual behaviour were inconsistent and had been
3015 changing anyway, so this is more a bug-fix than a behavioural
3017 [Geoff Thorpe, diagnosed by Nadav Har'El]
3019 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3020 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3023 *) Fix initialization code race conditions in
3024 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3025 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3026 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3027 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3028 ssl2_get_cipher_by_char(),
3029 ssl3_get_cipher_by_char().
3030 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3032 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3033 the cached sessions are flushed, as the remove_cb() might use ex_data
3034 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3035 (see [openssl.org #212]).
3036 [Geoff Thorpe, Lutz Jaenicke]
3038 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3039 length, instead of the encoding length to d2i_ASN1_OBJECT.
3042 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3044 *) [In 0.9.6g-engine release:]
3045 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3046 [Lynn Gazis <lgazis@rainbow.com>]
3048 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3050 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3051 and get fix the header length calculation.
3052 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3053 Alon Kantor <alonk@checkpoint.com> (and others),
3056 *) Use proper error handling instead of 'assertions' in buffer
3057 overflow checks added in 0.9.6e. This prevents DoS (the
3058 assertions could call abort()).
3059 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3061 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3063 *) Add various sanity checks to asn1_get_length() to reject
3064 the ASN1 length bytes if they exceed sizeof(long), will appear
3065 negative or the content length exceeds the length of the
3067 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3069 *) Fix cipher selection routines: ciphers without encryption had no flags
3070 for the cipher strength set and where therefore not handled correctly
3071 by the selection routines (PR #130).
3074 *) Fix EVP_dsa_sha macro.
3078 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3079 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3080 that was added in OpenSSL 0.9.6d.
3082 As the countermeasure turned out to be incompatible with some
3083 broken SSL implementations, the new option is part of SSL_OP_ALL.
3084 SSL_OP_ALL is usually employed when compatibility with weird SSL
3085 implementations is desired (e.g. '-bugs' option to 's_client' and
3086 's_server'), so the new option is automatically set in many
3090 *) Changes in security patch:
3092 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3093 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3094 Air Force Materiel Command, USAF, under agreement number
3097 *) Add various sanity checks to asn1_get_length() to reject
3098 the ASN1 length bytes if they exceed sizeof(long), will appear
3099 negative or the content length exceeds the length of the
3100 supplied buffer. (CAN-2002-0659)
3101 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3103 *) Assertions for various potential buffer overflows, not known to
3105 [Ben Laurie (CHATS)]
3107 *) Various temporary buffers to hold ASCII versions of integers were
3108 too small for 64 bit platforms. (CAN-2002-0655)
3109 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3111 *) Remote buffer overflow in SSL3 protocol - an attacker could
3112 supply an oversized session ID to a client. (CAN-2002-0656)
3113 [Ben Laurie (CHATS)]
3115 *) Remote buffer overflow in SSL2 protocol - an attacker could
3116 supply an oversized client master key. (CAN-2002-0656)
3117 [Ben Laurie (CHATS)]
3119 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3121 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3122 encoded as NULL) with id-dsa-with-sha1.
3123 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3125 *) Check various X509_...() return values in apps/req.c.
3126 [Nils Larsch <nla@trustcenter.de>]
3128 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3129 an end-of-file condition would erronously be flagged, when the CRLF
3130 was just at the end of a processed block. The bug was discovered when
3131 processing data through a buffering memory BIO handing the data to a
3132 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3133 <ptsekov@syntrex.com> and Nedelcho Stanev.
3136 *) Implement a countermeasure against a vulnerability recently found
3137 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3138 before application data chunks to avoid the use of known IVs
3139 with data potentially chosen by the attacker.
3142 *) Fix length checks in ssl3_get_client_hello().
3145 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3146 to prevent ssl3_read_internal() from incorrectly assuming that
3147 ssl3_read_bytes() found application data while handshake
3148 processing was enabled when in fact s->s3->in_read_app_data was
3149 merely automatically cleared during the initial handshake.
3150 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3152 *) Fix object definitions for Private and Enterprise: they were not
3153 recognized in their shortname (=lowercase) representation. Extend
3154 obj_dat.pl to issue an error when using undefined keywords instead
3155 of silently ignoring the problem (Svenning Sorensen
3156 <sss@sss.dnsalias.net>).
3159 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3160 generators, i.e. generators other than 2 and 5. (Previously, the
3161 code did not properly initialise the 'add' and 'rem' values to
3162 BN_generate_prime().)
3164 In the new general case, we do not insist that 'generator' is
3165 actually a primitive root: This requirement is rather pointless;
3166 a generator of the order-q subgroup is just as good, if not
3170 *) Map new X509 verification errors to alerts. Discovered and submitted by
3171 Tom Wu <tom@arcot.com>.
3174 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3175 returning non-zero before the data has been completely received
3176 when using non-blocking I/O.
3177 [Bodo Moeller; problem pointed out by John Hughes]
3179 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3180 [Ben Laurie, Lutz Jaenicke]
3182 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3183 Yoram Zahavi <YoramZ@gilian.com>).
3186 *) Add information about CygWin 1.3 and on, and preserve proper
3187 configuration for the versions before that.
3188 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3190 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3191 check whether we deal with a copy of a session and do not delete from
3192 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3193 <izhar@checkpoint.com>.
3196 *) Do not store session data into the internal session cache, if it
3197 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3198 flag is set). Proposed by Aslam <aslam@funk.com>.
3201 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3205 *) [In 0.9.6d-engine release:]
3206 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3207 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3209 *) Add the configuration target linux-s390x.
3210 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3212 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3213 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3214 variable as an indication that a ClientHello message has been
3215 received. As the flag value will be lost between multiple
3216 invocations of ssl3_accept when using non-blocking I/O, the
3217 function may not be aware that a handshake has actually taken
3218 place, thus preventing a new session from being added to the
3221 To avoid this problem, we now set s->new_session to 2 instead of
3222 using a local variable.
3223 [Lutz Jaenicke, Bodo Moeller]
3225 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3226 if the SSL_R_LENGTH_MISMATCH error is detected.
3227 [Geoff Thorpe, Bodo Moeller]
3229 *) New 'shared_ldflag' column in Configure platform table.
3232 *) Fix EVP_CIPHER_mode macro.
3233 ["Dan S. Camper" <dan@bti.net>]
3235 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3236 type, we must throw them away by setting rr->length to 0.
3237 [D P Chang <dpc@qualys.com>]
3239 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3241 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3242 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3243 worked incorrectly for those cases where range = 10..._2 and
3244 3*range is two bits longer than range.)
3247 *) Only add signing time to PKCS7 structures if it is not already
3251 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3252 OBJ_ld_ce should be OBJ_id_ce.
3253 Also some ip-pda OIDs in crypto/objects/objects.txt were
3254 incorrect (cf. RFC 3039).
3255 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3257 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3258 returns early because it has nothing to do.
3259 [Andy Schneider <andy.schneider@bjss.co.uk>]
3261 *) [In 0.9.6c-engine release:]
3262 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3263 [Andy Schneider <andy.schneider@bjss.co.uk>]
3265 *) [In 0.9.6c-engine release:]
3266 Add support for Cryptographic Appliance's keyserver technology.
3267 (Use engine 'keyclient')
3268 [Cryptographic Appliances and Geoff Thorpe]
3270 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3271 is called via tools/c89.sh because arguments have to be
3272 rearranged (all '-L' options must appear before the first object
3274 [Richard Shapiro <rshapiro@abinitio.com>]
3276 *) [In 0.9.6c-engine release:]
3277 Add support for Broadcom crypto accelerator cards, backported
3279 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3281 *) [In 0.9.6c-engine release:]
3282 Add support for SureWare crypto accelerator cards from
3283 Baltimore Technologies. (Use engine 'sureware')
3284 [Baltimore Technologies and Mark Cox]
3286 *) [In 0.9.6c-engine release:]
3287 Add support for crypto accelerator cards from Accelerated
3288 Encryption Processing, www.aep.ie. (Use engine 'aep')
3289 [AEP Inc. and Mark Cox]
3291 *) Add a configuration entry for gcc on UnixWare.
3292 [Gary Benson <gbenson@redhat.com>]
3294 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3295 messages are stored in a single piece (fixed-length part and
3296 variable-length part combined) and fix various bugs found on the way.
3299 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3300 instead. BIO_gethostbyname() does not know what timeouts are
3301 appropriate, so entries would stay in cache even when they have
3303 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3305 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3306 faced with a pathologically small ClientHello fragment that does
3307 not contain client_version: Instead of aborting with an error,
3308 simply choose the highest available protocol version (i.e.,
3309 TLS 1.0 unless it is disabled). In practice, ClientHello
3310 messages are never sent like this, but this change gives us
3311 strictly correct behaviour at least for TLS.
3314 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3315 never resets s->method to s->ctx->method when called from within
3316 one of the SSL handshake functions.
3317 [Bodo Moeller; problem pointed out by Niko Baric]
3319 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3320 (sent using the client's version number) if client_version is
3321 smaller than the protocol version in use. Also change
3322 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3323 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3324 the client will at least see that alert.
3327 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3331 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3332 client receives HelloRequest while in a handshake.
3333 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3335 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3336 should end in 'break', not 'goto end' which circuments various
3337 cleanups done in state SSL_ST_OK. But session related stuff
3338 must be disabled for SSL_ST_OK in the case that we just sent a
3341 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3342 before just sending a HelloRequest.
3343 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3345 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3346 reveal whether illegal block cipher padding was found or a MAC
3347 verification error occured. (Neither SSLerr() codes nor alerts
3348 are directly visible to potential attackers, but the information
3349 may leak via logfiles.)
3351 Similar changes are not required for the SSL 2.0 implementation
3352 because the number of padding bytes is sent in clear for SSL 2.0,
3353 and the extra bytes are just ignored. However ssl/s2_pkt.c
3354 failed to verify that the purported number of padding bytes is in
3358 *) Add OpenUNIX-8 support including shared libraries
3359 (Boyd Lynn Gerber <gerberb@zenez.com>).
3362 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3363 'wristwatch attack' using huge encoding parameters (cf.
3364 James H. Manger's CRYPTO 2001 paper). Note that the
3365 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3366 encoding parameters and hence was not vulnerable.
3369 *) BN_sqr() bug fix.
3370 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3372 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3373 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3374 followed by modular reduction.
3375 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3377 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3378 equivalent based on BN_pseudo_rand() instead of BN_rand().
3381 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3382 This function was broken, as the check for a new client hello message
3383 to handle SGC did not allow these large messages.
3384 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3387 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3390 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3391 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3394 *) Rework the configuration and shared library support for Tru64 Unix.
3395 The configuration part makes use of modern compiler features and
3396 still retains old compiler behavior for those that run older versions
3397 of the OS. The shared library support part includes a variant that
3398 uses the RPATH feature, and is available through the special
3399 configuration target "alpha-cc-rpath", which will never be selected
3401 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3403 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3404 with the same message size as in ssl3_get_certificate_request().
3405 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3406 messages might inadvertently be reject as too long.
3407 [Petr Lampa <lampa@fee.vutbr.cz>]
3409 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3412 *) Modified SSL library such that the verify_callback that has been set
3413 specificly for an SSL object with SSL_set_verify() is actually being
3414 used. Before the change, a verify_callback set with this function was
3415 ignored and the verify_callback() set in the SSL_CTX at the time of
3416 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3417 to allow the necessary settings.
3420 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3421 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3422 done automatically (in contradiction to the requirements of the C
3423 standard). This made problems when used from OpenSSH.
3426 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3427 dh->length and always used
3429 BN_rand_range(priv_key, dh->p).
3431 BN_rand_range() is not necessary for Diffie-Hellman, and this
3432 specific range makes Diffie-Hellman unnecessarily inefficient if
3433 dh->length (recommended exponent length) is much smaller than the
3434 length of dh->p. We could use BN_rand_range() if the order of
3435 the subgroup was stored in the DH structure, but we only have
3440 BN_rand(priv_key, l, ...)
3442 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3448 RSA_eay_public_encrypt
3449 RSA_eay_private_decrypt
3450 RSA_eay_private_encrypt (signing)
3451 RSA_eay_public_decrypt (signature verification)
3453 (default implementations for RSA_public_encrypt,
3454 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3455 always reject numbers >= n.
3458 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3459 to synchronize access to 'locking_thread'. This is necessary on
3460 systems where access to 'locking_thread' (an 'unsigned long'
3461 variable) is not atomic.
3464 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3465 *before* setting the 'crypto_lock_rand' flag. The previous code had
3466 a race condition if 0 is a valid thread ID.
3467 [Travis Vitek <vitek@roguewave.com>]
3469 *) Add support for shared libraries under Irix.
3470 [Albert Chin-A-Young <china@thewrittenword.com>]
3472 *) Add configuration option to build on Linux on both big-endian and
3474 [Ralf Baechle <ralf@uni-koblenz.de>]
3476 *) Add the possibility to create shared libraries on HP-UX.
3479 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3481 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3482 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3483 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3484 PRNG state recovery was possible based on the output of
3485 one PRNG request appropriately sized to gain knowledge on
3486 'md' followed by enough consecutive 1-byte PRNG requests
3487 to traverse all of 'state'.
3489 1. When updating 'md_local' (the current thread's copy of 'md')
3490 during PRNG output generation, hash all of the previous
3491 'md_local' value, not just the half used for PRNG output.
3493 2. Make the number of bytes from 'state' included into the hash
3494 independent from the number of PRNG bytes requested.
3496 The first measure alone would be sufficient to avoid
3497 Markku-Juhani's attack. (Actually it had never occurred
3498 to me that the half of 'md_local' used for chaining was the
3499 half from which PRNG output bytes were taken -- I had always
3500 assumed that the secret half would be used.) The second
3501 measure makes sure that additional data from 'state' is never
3502 mixed into 'md_local' in small portions; this heuristically
3503 further strengthens the PRNG.
3506 *) Fix crypto/bn/asm/mips3.s.
3509 *) When only the key is given to "enc", the IV is undefined. Print out
3510 an error message in this case.
3513 *) Handle special case when X509_NAME is empty in X509 printing routines.
3516 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3517 positive and less than q.
3520 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3521 used: it isn't thread safe and the add_lock_callback should handle
3523 [Paul Rose <Paul.Rose@bridge.com>]
3525 *) Verify that incoming data obeys the block size in
3526 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3530 [Ulf Möller, Bodo Möller]
3532 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3533 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3534 when fixing the server behaviour for backwards-compatible 'client
3535 hello' messages. (Note that the attack is impractical against
3536 SSL 3.0 and TLS 1.0 anyway because length and version checking
3537 means that the probability of guessing a valid ciphertext is
3538 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3541 Before 0.9.5, the countermeasure (hide the error by generating a
3542 random 'decryption result') did not work properly because
3543 ERR_clear_error() was missing, meaning that SSL_get_error() would
3544 detect the supposedly ignored error.
3546 Both problems are now fixed.
3549 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3550 (previously it was 1024).
3553 *) Fix for compatibility mode trust settings: ignore trust settings
3554 unless some valid trust or reject settings are present.
3557 *) Fix for blowfish EVP: its a variable length cipher.
3560 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3561 parameters in DSA public key structures and return an error in the
3562 DSA routines if parameters are absent.
3565 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3566 in the current directory if neither $RANDFILE nor $HOME was set.
3567 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3568 caused some confusion to Windows users who haven't defined $HOME.
3569 Thus RAND_file_name() is changed again: e_os.h can define a
3570 DEFAULT_HOME, which will be used if $HOME is not set.
3571 For Windows, we use "C:"; on other platforms, we still require
3572 environment variables.
3574 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3575 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3576 having multiple threads call RAND_poll() concurrently.
3579 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3580 combination of a flag and a thread ID variable.
3581 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3582 flag), *other* threads can enter ssleay_add_bytes without obeying
3583 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3584 that they do not hold after the first thread unsets add_do_not_lock).
3587 *) Change bctest again: '-x' expressions are not available in all
3591 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3593 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3594 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3596 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3597 the default extension for executables, if any. Also, make the perl
3598 scripts that use symlink() to test if it really exists and use "cp"
3599 if it doesn't. All this made OpenSSL compilable and installable in
3603 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3604 If SEQUENCE is length is indefinite just set c->slen to the total
3605 amount of data available.
3606 [Steve Henson, reported by shige@FreeBSD.org]
3607 [This change does not apply to 0.9.7.]
3609 *) Change bctest to avoid here-documents inside command substitution
3610 (workaround for FreeBSD /bin/sh bug).
3611 For compatibility with Ultrix, avoid shell functions (introduced
3612 in the bctest version that searches along $PATH).
3615 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3616 with des_encrypt() defined on some operating systems, like Solaris
3620 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3621 On the Importance of Eliminating Errors in Cryptographic
3622 Computations, J. Cryptology 14 (2001) 2, 101-119,
3623 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3626 *) MIPS assembler BIGNUM division bug fix.
3629 *) Disabled incorrect Alpha assembler code.
3632 *) Fix PKCS#7 decode routines so they correctly update the length
3633 after reading an EOC for the EXPLICIT tag.
3635 [This change does not apply to 0.9.7.]
3637 *) Fix bug in PKCS#12 key generation routines. This was triggered
3638 if a 3DES key was generated with a 0 initial byte. Include
3639 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3640 (but broken) behaviour.
3643 *) Enhance bctest to search for a working bc along $PATH and print
3645 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3647 *) Fix memory leaks in err.c: free err_data string if necessary;
3648 don't write to the wrong index in ERR_set_error_data.
3651 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3655 *) Replace rdtsc with _emit statements for VC++ version 5.
3656 [Jeremy Cooper <jeremy@baymoo.org>]
3658 *) Make it possible to reuse SSLv2 sessions.
3661 *) In copy_email() check for >= 0 as a return value for
3662 X509_NAME_get_index_by_NID() since 0 is a valid index.
3663 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3665 *) Avoid coredump with unsupported or invalid public keys by checking if
3666 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3667 PKCS7_verify() fails with non detached data.
3670 *) Don't use getenv in library functions when run as setuid/setgid.
3671 New function OPENSSL_issetugid().
3674 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3675 due to incorrect handling of multi-threading:
3677 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3679 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3681 3. Count how many times MemCheck_off() has been called so that
3682 nested use can be treated correctly. This also avoids
3683 inband-signalling in the previous code (which relied on the
3684 assumption that thread ID 0 is impossible).
3687 *) Add "-rand" option also to s_client and s_server.
3690 *) Fix CPU detection on Irix 6.x.
3691 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3692 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3694 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3697 [This change does not apply to 0.9.7.]
3699 *) Use the cached encoding of an X509_NAME structure rather than
3700 copying it. This is apparently the reason for the libsafe "errors"
3701 but the code is actually correct.
3704 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3705 Bleichenbacher's DSA attack.
3706 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3707 to be set and top=0 forces the highest bit to be set; top=-1 is new
3708 and leaves the highest bit random.
3709 [Ulf Moeller, Bodo Moeller]
3711 *) In the NCONF_...-based implementations for CONF_... queries
3712 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3713 a temporary CONF structure with the data component set to NULL
3714 (which gives segmentation faults in lh_retrieve).
3715 Instead, use NULL for the CONF pointer in CONF_get_string and
3716 CONF_get_number (which may use environment variables) and directly
3717 return NULL from CONF_get_section.
3720 *) Fix potential buffer overrun for EBCDIC.
3723 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3724 keyUsage if basicConstraints absent for a CA.
3727 *) Make SMIME_write_PKCS7() write mail header values with a format that
3728 is more generally accepted (no spaces before the semicolon), since
3729 some programs can't parse those values properly otherwise. Also make
3730 sure BIO's that break lines after each write do not create invalid
3734 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3735 macros previously used would not encode an empty SEQUENCE OF
3736 and break the signature.
3738 [This change does not apply to 0.9.7.]
3740 *) Zero the premaster secret after deriving the master secret in
3744 *) Add some EVP_add_digest_alias registrations (as found in
3745 OpenSSL_add_all_digests()) to SSL_library_init()
3746 aka OpenSSL_add_ssl_algorithms(). This provides improved
3747 compatibility with peers using X.509 certificates
3748 with unconventional AlgorithmIdentifier OIDs.
3751 *) Fix for Irix with NO_ASM.
3752 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3754 *) ./config script fixes.
3755 [Ulf Moeller, Richard Levitte]
3757 *) Fix 'openssl passwd -1'.
3760 *) Change PKCS12_key_gen_asc() so it can cope with non null
3761 terminated strings whose length is passed in the passlen
3762 parameter, for example from PEM callbacks. This was done
3763 by adding an extra length parameter to asc2uni().
3764 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3766 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3767 call failed, free the DSA structure.
3770 *) Fix to uni2asc() to cope with zero length Unicode strings.
3771 These are present in some PKCS#12 files.
3774 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3775 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3776 when writing a 32767 byte record.
3777 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3779 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3780 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3782 (RSA objects have a reference count access to which is protected
3783 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3784 so they are meant to be shared between threads.)
3785 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3786 "Reddie, Steven" <Steven.Reddie@ca.com>]
3788 *) Fix a deadlock in CRYPTO_mem_leaks().
3791 *) Use better test patterns in bntest.
3794 *) rand_win.c fix for Borland C.
3797 *) BN_rshift bugfix for n == 0.
3800 *) Add a 'bctest' script that checks for some known 'bc' bugs
3801 so that 'make test' does not abort just because 'bc' is broken.
3804 *) Store verify_result within SSL_SESSION also for client side to
3805 avoid potential security hole. (Re-used sessions on the client side
3806 always resulted in verify_result==X509_V_OK, not using the original
3807 result of the server certificate verification.)
3810 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3811 SSL3_RT_APPLICATION_DATA, return 0.
3812 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3816 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3817 releases, have been re-implemented by renaming the previous
3818 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3819 and ssl3_read_internal, respectively, and adding 'peek' parameters
3820 to them. The new ssl[23]_{read,peek} functions are calls to
3821 ssl[23]_read_internal with the 'peek' flag set appropriately.
3822 A 'peek' parameter has also been added to ssl3_read_bytes, which
3823 does the actual work for ssl3_read_internal.
3826 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3827 the method-specific "init()" handler. Also clean up ex_data after
3828 calling the method-specific "finish()" handler. Previously, this was
3829 happening the other way round.
3832 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3833 The previous value, 12, was not always sufficient for BN_mod_exp().
3836 *) Make sure that shared libraries get the internal name engine with
3837 the full version number and not just 0. This should mark the
3838 shared libraries as not backward compatible. Of course, this should
3839 be changed again when we can guarantee backward binary compatibility.
3842 *) Fix typo in get_cert_by_subject() in by_dir.c
3843 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3845 *) Rework the system to generate shared libraries:
3847 - Make note of the expected extension for the shared libraries and
3848 if there is a need for symbolic links from for example libcrypto.so.0
3849 to libcrypto.so.0.9.7. There is extended info in Configure for
3852 - Make as few rebuilds of the shared libraries as possible.
3854 - Still avoid linking the OpenSSL programs with the shared libraries.
3856 - When installing, install the shared libraries separately from the
3860 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3862 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3863 and not in SSL_clear because the latter is also used by the
3864 accept/connect functions; previously, the settings made by
3865 SSL_set_read_ahead would be lost during the handshake.
3866 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3868 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3869 Previously, it would create entries for disableed algorithms no
3873 *) Added several new manual pages for SSL_* function.
3876 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3878 *) In ssl23_get_client_hello, generate an error message when faced
3879 with an initial SSL 3.0/TLS record that is too small to contain the
3880 first two bytes of the ClientHello message, i.e. client_version.
3881 (Note that this is a pathologic case that probably has never happened
3882 in real life.) The previous approach was to use the version number
3883 from the record header as a substitute; but our protocol choice
3884 should not depend on that one because it is not authenticated
3885 by the Finished messages.
3888 *) More robust randomness gathering functions for Windows.
3889 [Jeffrey Altman <jaltman@columbia.edu>]
3891 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3892 not set then we don't setup the error code for issuer check errors
3893 to avoid possibly overwriting other errors which the callback does
3894 handle. If an application does set the flag then we assume it knows
3895 what it is doing and can handle the new informational codes
3899 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3900 a general "ANY" type, as such it should be able to decode anything
3901 including tagged types. However it didn't check the class so it would
3902 wrongly interpret tagged types in the same way as their universal
3903 counterpart and unknown types were just rejected. Changed so that the
3904 tagged and unknown types are handled in the same way as a SEQUENCE:
3905 that is the encoding is stored intact. There is also a new type
3906 "V_ASN1_OTHER" which is used when the class is not universal, in this
3907 case we have no idea what the actual type is so we just lump them all
3911 *) On VMS, stdout may very well lead to a file that is written to
3912 in a record-oriented fashion. That means that every write() will
3913 write a separate record, which will be read separately by the
3914 programs trying to read from it. This can be very confusing.
3916 The solution is to put a BIO filter in the way that will buffer
3917 text until a linefeed is reached, and then write everything a
3918 line at a time, so every record written will be an actual line,
3919 not chunks of lines and not (usually doesn't happen, but I've
3920 seen it once) several lines in one record. BIO_f_linebuffer() is
3923 Currently, it's a VMS-only method, because that's where it has
3924 been tested well enough.
3927 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
3928 it can return incorrect results.
3929 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
3930 but it was in 0.9.6-beta[12].)
3933 *) Disable the check for content being present when verifying detached
3934 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
3935 include zero length content when signing messages.
3938 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
3939 BIO_ctrl (for BIO pairs).
3942 *) Add DSO method for VMS.
3945 *) Bug fix: Montgomery multiplication could produce results with the
3949 *) Add RPM specification openssl.spec and modify it to build three
3950 packages. The default package contains applications, application
3951 documentation and run-time libraries. The devel package contains
3952 include files, static libraries and function documentation. The
3953 doc package contains the contents of the doc directory. The original
3954 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
3957 *) Add a large number of documentation files for many SSL routines.
3958 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3960 *) Add a configuration entry for Sony News 4.
3961 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
3963 *) Don't set the two most significant bits to one when generating a
3964 random number < q in the DSA library.
3967 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
3968 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
3969 the underlying transport is blocking) if a handshake took place.
3970 (The default behaviour is needed by applications such as s_client
3971 and s_server that use select() to determine when to use SSL_read;
3972 but for applications that know in advance when to expect data, it
3973 just makes things more complicated.)
3976 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
3980 *) Add a few more EBCDIC conditionals that make `req' and `x509'
3981 work better on such systems.
3982 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
3984 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
3985 Update PKCS12_parse() so it copies the friendlyName and the
3986 keyid to the certificates aux info.
3989 *) Fix bug in PKCS7_verify() which caused an infinite loop
3990 if there was more than one signature.
3991 [Sven Uszpelkat <su@celocom.de>]
3993 *) Major change in util/mkdef.pl to include extra information
3994 about each symbol, as well as presentig variables as well
3995 as functions. This change means that there's n more need
3996 to rebuild the .num files when some algorithms are excluded.
3999 *) Allow the verify time to be set by an application,
4000 rather than always using the current time.
4003 *) Phase 2 verify code reorganisation. The certificate
4004 verify code now looks up an issuer certificate by a
4005 number of criteria: subject name, authority key id
4006 and key usage. It also verifies self signed certificates
4007 by the same criteria. The main comparison function is
4008 X509_check_issued() which performs these checks.
4010 Lot of changes were necessary in order to support this
4011 without completely rewriting the lookup code.
4013 Authority and subject key identifier are now cached.
4015 The LHASH 'certs' is X509_STORE has now been replaced
4016 by a STACK_OF(X509_OBJECT). This is mainly because an
4017 LHASH can't store or retrieve multiple objects with
4018 the same hash value.
4020 As a result various functions (which were all internal
4021 use only) have changed to handle the new X509_STORE
4022 structure. This will break anything that messed round
4023 with X509_STORE internally.
4025 The functions X509_STORE_add_cert() now checks for an
4026 exact match, rather than just subject name.
4028 The X509_STORE API doesn't directly support the retrieval
4029 of multiple certificates matching a given criteria, however
4030 this can be worked round by performing a lookup first
4031 (which will fill the cache with candidate certificates)
4032 and then examining the cache for matches. This is probably
4033 the best we can do without throwing out X509_LOOKUP
4034 entirely (maybe later...).
4036 The X509_VERIFY_CTX structure has been enhanced considerably.
4038 All certificate lookup operations now go via a get_issuer()
4039 callback. Although this currently uses an X509_STORE it
4040 can be replaced by custom lookups. This is a simple way
4041 to bypass the X509_STORE hackery necessary to make this
4042 work and makes it possible to use more efficient techniques
4043 in future. A very simple version which uses a simple
4044 STACK for its trusted certificate store is also provided
4045 using X509_STORE_CTX_trusted_stack().
4047 The verify_cb() and verify() callbacks now have equivalents
4048 in the X509_STORE_CTX structure.
4050 X509_STORE_CTX also has a 'flags' field which can be used
4051 to customise the verify behaviour.
4054 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4055 excludes S/MIME capabilities.
4058 *) When a certificate request is read in keep a copy of the
4059 original encoding of the signed data and use it when outputing
4060 again. Signatures then use the original encoding rather than
4061 a decoded, encoded version which may cause problems if the
4062 request is improperly encoded.
4065 *) For consistency with other BIO_puts implementations, call
4066 buffer_write(b, ...) directly in buffer_puts instead of calling
4069 In BIO_puts, increment b->num_write as in BIO_write.
4070 [Peter.Sylvester@EdelWeb.fr]
4072 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4073 BN_zero, we may not return a BIGNUM with an array consisting of
4077 *) Avoid calling abort() from within the library when problems are
4078 detected, except if preprocessor symbols have been defined
4079 (such as REF_CHECK, BN_DEBUG etc.).
4082 *) New openssl application 'rsautl'. This utility can be
4083 used for low level RSA operations. DER public key
4084 BIO/fp routines also added.
4087 *) New Configure entry and patches for compiling on QNX 4.
4088 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4090 *) A demo state-machine implementation was sponsored by
4091 Nuron (http://www.nuron.com/) and is now available in
4092 demos/state_machine.
4095 *) New options added to the 'dgst' utility for signature
4096 generation and verification.
4099 *) Unrecognized PKCS#7 content types are now handled via a
4100 catch all ASN1_TYPE structure. This allows unsupported
4101 types to be stored as a "blob" and an application can
4102 encode and decode it manually.
4105 *) Fix various signed/unsigned issues to make a_strex.c
4107 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4109 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4110 length if passed a buffer. ASN1_INTEGER_to_BN failed
4111 if passed a NULL BN and its argument was negative.
4112 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4114 *) Modification to PKCS#7 encoding routines to output definite
4115 length encoding. Since currently the whole structures are in
4116 memory there's not real point in using indefinite length
4117 constructed encoding. However if OpenSSL is compiled with
4118 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4121 *) Added BIO_vprintf() and BIO_vsnprintf().
4124 *) Added more prefixes to parse for in the the strings written
4125 through a logging bio, to cover all the levels that are available
4126 through syslog. The prefixes are now:
4128 PANIC, EMERG, EMR => LOG_EMERG
4129 ALERT, ALR => LOG_ALERT
4130 CRIT, CRI => LOG_CRIT
4131 ERROR, ERR => LOG_ERR
4132 WARNING, WARN, WAR => LOG_WARNING
4133 NOTICE, NOTE, NOT => LOG_NOTICE
4134 INFO, INF => LOG_INFO
4135 DEBUG, DBG => LOG_DEBUG
4137 and as before, if none of those prefixes are present at the
4138 beginning of the string, LOG_ERR is chosen.
4140 On Win32, the LOG_* levels are mapped according to this:
4142 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4143 LOG_WARNING => EVENTLOG_WARNING_TYPE
4144 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4148 *) Made it possible to reconfigure with just the configuration
4149 argument "reconf" or "reconfigure". The command line arguments
4150 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4151 and are retrieved from there when reconfiguring.
4155 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4157 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4160 *) The obj_dat.pl script was messing up the sorting of object
4161 names. The reason was that it compared the quoted version
4162 of strings as a result "OCSP" > "OCSP Signing" because
4163 " > SPACE. Changed script to store unquoted versions of
4164 names and add quotes on output. It was also omitting some
4165 names from the lookup table if they were given a default
4166 value (that is if SN is missing it is given the same
4167 value as LN and vice versa), these are now added on the
4168 grounds that if an object has a name we should be able to
4169 look it up. Finally added warning output when duplicate
4170 short or long names are found.
4173 *) Changes needed for Tandem NSK.
4174 [Scott Uroff <scott@xypro.com>]
4176 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4177 RSA_padding_check_SSLv23(), special padding was never detected
4178 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4179 version rollback attacks was not effective.
4181 In s23_clnt.c, don't use special rollback-attack detection padding
4182 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4183 client; similarly, in s23_srvr.c, don't do the rollback check if
4184 SSL 2.0 is the only protocol enabled in the server.
4187 *) Make it possible to get hexdumps of unprintable data with 'openssl
4188 asn1parse'. By implication, the functions ASN1_parse_dump() and
4189 BIO_dump_indent() are added.
4192 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4193 these print out strings and name structures based on various
4194 flags including RFC2253 support and proper handling of
4195 multibyte characters. Added options to the 'x509' utility
4196 to allow the various flags to be set.
4199 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4200 Also change the functions X509_cmp_current_time() and
4201 X509_gmtime_adj() work with an ASN1_TIME structure,
4202 this will enable certificates using GeneralizedTime in validity
4203 dates to be checked.
4206 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4207 negative public key encodings) on by default,
4208 NO_NEG_PUBKEY_BUG can be set to disable it.
4211 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4212 content octets. An i2c_ASN1_OBJECT is unnecessary because
4213 the encoding can be trivially obtained from the structure.
4216 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4217 not read locks (CRYPTO_r_[un]lock).
4220 *) A first attempt at creating official support for shared
4221 libraries through configuration. I've kept it so the
4222 default is static libraries only, and the OpenSSL programs
4223 are always statically linked for now, but there are
4224 preparations for dynamic linking in place.
4225 This has been tested on Linux and Tru64.
4228 *) Randomness polling function for Win9x, as described in:
4229 Peter Gutmann, Software Generation of Practically Strong
4233 *) Fix so PRNG is seeded in req if using an already existing
4237 *) New options to smime application. -inform and -outform
4238 allow alternative formats for the S/MIME message including
4239 PEM and DER. The -content option allows the content to be
4240 specified separately. This should allow things like Netscape
4241 form signing output easier to verify.
4244 *) Fix the ASN1 encoding of tags using the 'long form'.
4247 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4248 STRING types. These convert content octets to and from the
4249 underlying type. The actual tag and length octets are
4250 already assumed to have been read in and checked. These
4251 are needed because all other string types have virtually
4252 identical handling apart from the tag. By having versions
4253 of the ASN1 functions that just operate on content octets
4254 IMPLICIT tagging can be handled properly. It also allows
4255 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4256 and ASN1_INTEGER are identical apart from the tag.
4259 *) Change the handling of OID objects as follows:
4261 - New object identifiers are inserted in objects.txt, following
4262 the syntax given in objects.README.
4263 - objects.pl is used to process obj_mac.num and create a new
4265 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4268 This is currently kind of a hack, and the perl code in objects.pl
4269 isn't very elegant, but it works as I intended. The simplest way
4270 to check that it worked correctly is to look in obj_dat.h and
4271 check the array nid_objs and make sure the objects haven't moved
4272 around (this is important!). Additions are OK, as well as
4273 consistent name changes.
4276 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4279 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4280 The given file adds to whatever has already been seeded into the
4281 random pool through the RANDFILE configuration file option or
4282 environment variable, or the default random state file.
4285 *) mkstack.pl now sorts each macro group into lexical order.
4286 Previously the output order depended on the order the files
4287 appeared in the directory, resulting in needless rewriting
4291 *) Patches to make OpenSSL compile under Win32 again. Mostly
4292 work arounds for the VC++ problem that it treats func() as
4293 func(void). Also stripped out the parts of mkdef.pl that
4294 added extra typesafe functions: these no longer exist.
4297 *) Reorganisation of the stack code. The macros are now all
4298 collected in safestack.h . Each macro is defined in terms of
4299 a "stack macro" of the form SKM_<name>(type, a, b). The
4300 DEBUG_SAFESTACK is now handled in terms of function casts,
4301 this has the advantage of retaining type safety without the
4302 use of additional functions. If DEBUG_SAFESTACK is not defined
4303 then the non typesafe macros are used instead. Also modified the
4304 mkstack.pl script to handle the new form. Needs testing to see
4305 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4306 the default if no major problems. Similar behaviour for ASN1_SET_OF
4307 and PKCS12_STACK_OF.
4310 *) When some versions of IIS use the 'NET' form of private key the
4311 key derivation algorithm is different. Normally MD5(password) is
4312 used as a 128 bit RC4 key. In the modified case
4313 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4314 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4315 as the old Netscape_RSA functions except they have an additional
4316 'sgckey' parameter which uses the modified algorithm. Also added
4317 an -sgckey command line option to the rsa utility. Thanks to
4318 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4319 algorithm to openssl-dev.
4322 *) The evp_local.h macros were using 'c.##kname' which resulted in
4323 invalid expansion on some systems (SCO 5.0.5 for example).
4324 Corrected to 'c.kname'.
4325 [Phillip Porch <root@theporch.com>]
4327 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4328 a STACK of email addresses from a certificate or request, these look
4329 in the subject name and the subject alternative name extensions and
4330 omit any duplicate addresses.
4333 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4334 This makes DSA verification about 2 % faster.
4337 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4338 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4339 plus overhead for 1024 bit moduli).
4340 This makes exponentiations about 0.5 % faster for 1024 bit
4341 exponents (as measured by "openssl speed rsa2048").
4344 *) Rename memory handling macros to avoid conflicts with other
4346 Malloc => OPENSSL_malloc
4347 Malloc_locked => OPENSSL_malloc_locked
4348 Realloc => OPENSSL_realloc
4349 Free => OPENSSL_free
4352 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4353 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4356 *) CygWin32 support.
4357 [John Jarvie <jjarvie@newsguy.com>]
4359 *) The type-safe stack code has been rejigged. It is now only compiled
4360 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4361 by default all type-specific stack functions are "#define"d back to
4362 standard stack functions. This results in more streamlined output
4363 but retains the type-safety checking possibilities of the original
4367 *) The STACK code has been cleaned up, and certain type declarations
4368 that didn't make a lot of sense have been brought in line. This has
4369 also involved a cleanup of sorts in safestack.h to more correctly
4370 map type-safe stack functions onto their plain stack counterparts.
4371 This work has also resulted in a variety of "const"ifications of
4372 lots of the code, especially "_cmp" operations which should normally
4373 be prototyped with "const" parameters anyway.
4376 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4377 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4378 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4379 where all of 'md' is used each time the PRNG is used, but 'state'
4380 is used only indexed by a cyclic counter. As entropy may not be
4381 well distributed from the beginning, 'md' is important as a
4382 chaining variable. However, the output function chains only half
4383 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4384 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4385 in all of 'state' being rewritten, with the new values depending
4386 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4389 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4390 the handshake is continued after ssl_verify_cert_chain();
4391 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4392 can lead to 'unexplainable' connection aborts later.
4393 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4395 *) Major EVP API cipher revision.
4396 Add hooks for extra EVP features. This allows various cipher
4397 parameters to be set in the EVP interface. Support added for variable
4398 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4399 setting of RC2 and RC5 parameters.
4401 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4404 Remove lots of duplicated code from the EVP library. For example *every*
4405 cipher init() function handles the 'iv' in the same way according to the
4406 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4407 for CFB and OFB modes they zero ctx->num.
4409 New functionality allows removal of S/MIME code RC2 hack.
4411 Most of the routines have the same form and so can be declared in terms
4414 By shifting this to the top level EVP_CipherInit() it can be removed from
4415 all individual ciphers. If the cipher wants to handle IVs or keys
4416 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4419 Change lots of functions like EVP_EncryptUpdate() to now return a
4420 value: although software versions of the algorithms cannot fail
4421 any installed hardware versions can.
4424 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4425 this option is set, tolerate broken clients that send the negotiated
4426 protocol version number instead of the requested protocol version
4430 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4431 i.e. non-zero for export ciphersuites, zero otherwise.
4432 Previous versions had this flag inverted, inconsistent with
4433 rsa_tmp_cb (..._TMP_RSA_CB).
4434 [Bodo Moeller; problem reported by Amit Chopra]
4436 *) Add missing DSA library text string. Work around for some IIS
4437 key files with invalid SEQUENCE encoding.
4440 *) Add a document (doc/standards.txt) that list all kinds of standards
4441 and so on that are implemented in OpenSSL.
4444 *) Enhance c_rehash script. Old version would mishandle certificates
4445 with the same subject name hash and wouldn't handle CRLs at all.
4446 Added -fingerprint option to crl utility, to support new c_rehash
4450 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4453 *) Fix for SSL server purpose checking. Server checking was
4454 rejecting certificates which had extended key usage present
4455 but no ssl client purpose.
4456 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4458 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4459 is a little unclear about how a blank password is handled.
4460 Since the password in encoded as a BMPString with terminating
4461 double NULL a zero length password would end up as just the
4462 double NULL. However no password at all is different and is
4463 handled differently in the PKCS#12 key generation code. NS
4464 treats a blank password as zero length. MSIE treats it as no
4465 password on export: but it will try both on import. We now do
4466 the same: PKCS12_parse() tries zero length and no password if
4467 the password is set to "" or NULL (NULL is now a valid password:
4468 it wasn't before) as does the pkcs12 application.
4471 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4472 perror when PEM_read_bio_X509_REQ fails, the error message must
4473 be obtained from the error queue.
4476 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4477 it in ERR_remove_state if appropriate, and change ERR_get_state
4478 accordingly to avoid race conditions (this is necessary because
4479 thread_hash is no longer constant once set).
4482 *) Bugfix for linux-elf makefile.one.
4485 *) RSA_get_default_method() will now cause a default
4486 RSA_METHOD to be chosen if one doesn't exist already.
4487 Previously this was only set during a call to RSA_new()
4488 or RSA_new_method(NULL) meaning it was possible for
4489 RSA_get_default_method() to return NULL.
4492 *) Added native name translation to the existing DSO code
4493 that will convert (if the flag to do so is set) filenames
4494 that are sufficiently small and have no path information
4495 into a canonical native form. Eg. "blah" converted to
4496 "libblah.so" or "blah.dll" etc.
4499 *) New function ERR_error_string_n(e, buf, len) which is like
4500 ERR_error_string(e, buf), but writes at most 'len' bytes
4501 including the 0 terminator. For ERR_error_string_n, 'buf'
4503 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4505 *) CONF library reworked to become more general. A new CONF
4506 configuration file reader "class" is implemented as well as a
4507 new functions (NCONF_*, for "New CONF") to handle it. The now
4508 old CONF_* functions are still there, but are reimplemented to
4509 work in terms of the new functions. Also, a set of functions
4510 to handle the internal storage of the configuration data is
4511 provided to make it easier to write new configuration file
4512 reader "classes" (I can definitely see something reading a
4513 configuration file in XML format, for example), called _CONF_*,
4514 or "the configuration storage API"...
4516 The new configuration file reading functions are:
4518 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4519 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4521 NCONF_default, NCONF_WIN32
4523 NCONF_dump_fp, NCONF_dump_bio
4525 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4526 NCONF_new creates a new CONF object. This works in the same way
4527 as other interfaces in OpenSSL, like the BIO interface.
4528 NCONF_dump_* dump the internal storage of the configuration file,
4529 which is useful for debugging. All other functions take the same
4530 arguments as the old CONF_* functions wth the exception of the
4531 first that must be a `CONF *' instead of a `LHASH *'.
4533 To make it easer to use the new classes with the old CONF_* functions,
4534 the function CONF_set_default_method is provided.
4537 *) Add '-tls1' option to 'openssl ciphers', which was already
4538 mentioned in the documentation but had not been implemented.
4539 (This option is not yet really useful because even the additional
4540 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4543 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4544 OpenSSL-based applications) load shared libraries and bind to
4545 them in a portable way.
4546 [Geoff Thorpe, with contributions from Richard Levitte]
4548 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4550 *) Make sure _lrotl and _lrotr are only used with MSVC.
4552 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4553 (the default implementation of RAND_status).
4555 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4556 to '-clrext' (= clear extensions), as intended and documented.
4557 [Bodo Moeller; inconsistency pointed out by Michael Attili
4560 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4561 was larger than the MD block size.
4562 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4564 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4565 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4566 using the passed key: if the passed key was a private key the result
4567 of X509_print(), for example, would be to print out all the private key
4571 *) des_quad_cksum() byte order bug fix.
4572 [Ulf Möller, using the problem description in krb4-0.9.7, where
4573 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4575 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4577 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4579 *) For easily testing in shell scripts whether some command
4580 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4581 returns with exit code 0 iff no command of the given name is available.
4582 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4583 the output goes to stdout and nothing is printed to stderr.
4584 Additional arguments are always ignored.
4586 Since for each cipher there is a command of the same name,
4587 the 'no-cipher' compilation switches can be tested this way.
4589 ('openssl no-XXX' is not able to detect pseudo-commands such
4590 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4593 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4596 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4597 is set; it will be thrown away anyway because each handshake creates
4599 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4600 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4601 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4602 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4605 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4606 'Q' and 'R' lose their special meanings (quit/renegotiate).
4607 This is part of what -quiet does; unlike -quiet, -ign_eof
4608 does not suppress any output.
4611 *) Add compatibility options to the purpose and trust code. The
4612 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4613 accepts a certificate or CA, this was the previous behaviour,
4614 with all the associated security issues.
4616 X509_TRUST_COMPAT is the old trust behaviour: only and
4617 automatically trust self signed roots in certificate store. A
4618 new trust setting X509_TRUST_DEFAULT is used to specify that
4619 a purpose has no associated trust setting and it should instead
4620 use the value in the default purpose.
4623 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4624 and fix a memory leak.
4627 *) In util/mkerr.pl (which implements 'make errors'), preserve
4628 reason strings from the previous version of the .c file, as
4629 the default to have only downcase letters (and digits) in
4630 automatically generated reasons codes is not always appropriate.
4633 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4634 using strerror. Previously, ERR_reason_error_string() returned
4635 library names as reason strings for SYSerr; but SYSerr is a special
4636 case where small numbers are errno values, not library numbers.
4639 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4640 converts DSA parameters into DH parameters. (When creating parameters,
4641 DSA_generate_parameters is used.)
4644 *) Include 'length' (recommended exponent length) in C code generated
4645 by 'openssl dhparam -C'.
4648 *) The second argument to set_label in perlasm was already being used
4649 so couldn't be used as a "file scope" flag. Moved to third argument
4653 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4654 instead of RAND_bytes for encryption IVs and salts.
4657 *) Include RAND_status() into RAND_METHOD instead of implementing
4658 it only for md_rand.c Otherwise replacing the PRNG by calling
4659 RAND_set_rand_method would be impossible.
4662 *) Don't let DSA_generate_key() enter an infinite loop if the random
4663 number generation fails.
4666 *) New 'rand' application for creating pseudo-random output.
4669 *) Added configuration support for Linux/IA64
4670 [Rolf Haberrecker <rolf@suse.de>]
4672 *) Assembler module support for Mingw32.
4675 *) Shared library support for HPUX (in shlib/).
4676 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4678 *) Shared library support for Solaris gcc.
4679 [Lutz Behnke <behnke@trustcenter.de>]
4681 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4683 *) PKCS7_encrypt() was adding text MIME headers twice because they
4684 were added manually and by SMIME_crlf_copy().
4687 *) In bntest.c don't call BN_rand with zero bits argument.
4688 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4690 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4691 case was implemented. This caused BN_div_recp() to fail occasionally.
4694 *) Add an optional second argument to the set_label() in the perl
4695 assembly language builder. If this argument exists and is set
4696 to 1 it signals that the assembler should use a symbol whose
4697 scope is the entire file, not just the current function. This
4698 is needed with MASM which uses the format label:: for this scope.
4699 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4701 *) Change the ASN1 types so they are typedefs by default. Before
4702 almost all types were #define'd to ASN1_STRING which was causing
4703 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4707 *) Change names of new functions to the new get1/get0 naming
4708 convention: After 'get1', the caller owns a reference count
4709 and has to call ..._free; 'get0' returns a pointer to some
4710 data structure without incrementing reference counters.
4711 (Some of the existing 'get' functions increment a reference
4712 counter, some don't.)
4713 Similarly, 'set1' and 'add1' functions increase reference
4714 counters or duplicate objects.
4717 *) Allow for the possibility of temp RSA key generation failure:
4718 the code used to assume it always worked and crashed on failure.
4721 *) Fix potential buffer overrun problem in BIO_printf().
4722 [Ulf Möller, using public domain code by Patrick Powell; problem
4723 pointed out by David Sacerdote <das33@cornell.edu>]
4725 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4726 RAND_egd() and RAND_status(). In the command line application,
4727 the EGD socket can be specified like a seed file using RANDFILE
4731 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4732 Some CAs (e.g. Verisign) distribute certificates in this form.
4735 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4736 list to exclude them. This means that no special compilation option
4737 is needed to use anonymous DH: it just needs to be included in the
4741 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4742 EVP_MD_type. The old functionality is available in a new macro called
4743 EVP_MD_md(). Change code that uses it and update docs.
4746 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4747 where the 'void *' argument is replaced by a function pointer argument.
4748 Previously 'void *' was abused to point to functions, which works on
4749 many platforms, but is not correct. As these functions are usually
4750 called by macros defined in OpenSSL header files, most source code
4751 should work without changes.
4754 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4755 sections with information on -D... compiler switches used for
4756 compiling the library so that applications can see them. To enable
4757 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4758 must be defined. E.g.,
4759 #define OPENSSL_ALGORITHM_DEFINES
4760 #include <openssl/opensslconf.h>
4761 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4762 [Richard Levitte, Ulf and Bodo Möller]
4764 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4768 *) Change the 'other' type in certificate aux info to a STACK_OF
4769 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4770 the required ASN1 format: arbitrary types determined by an OID.
4773 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4774 argument to 'req'. This is not because the function is newer or
4775 better than others it just uses the work 'NEW' in the certificate
4776 request header lines. Some software needs this.
4779 *) Reorganise password command line arguments: now passwords can be
4780 obtained from various sources. Delete the PEM_cb function and make
4781 it the default behaviour: i.e. if the callback is NULL and the
4782 usrdata argument is not NULL interpret it as a null terminated pass
4783 phrase. If usrdata and the callback are NULL then the pass phrase
4784 is prompted for as usual.
4787 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4788 the support is automatically enabled. The resulting binaries will
4789 autodetect the card and use it if present.
4790 [Ben Laurie and Compaq Inc.]
4792 *) Work around for Netscape hang bug. This sends certificate request
4793 and server done in one record. Since this is perfectly legal in the
4794 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4795 the bugs/SSLv3 entry for more info.
4798 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4801 *) Add -rand argument to smime and pkcs12 applications and read/write
4805 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4808 *) Add command line password options to the remaining applications.
4811 *) Bug fix for BN_div_recp() for numerators with an even number of
4815 *) More tests in bntest.c, and changed test_bn output.
4818 *) ./config recognizes MacOS X now.
4821 *) Bug fix for BN_div() when the first words of num and divsor are
4822 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4825 *) Add support for various broken PKCS#8 formats, and command line
4826 options to produce them.
4829 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4830 get temporary BIGNUMs from a BN_CTX.
4833 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4837 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4838 include a #define from the old name to the new. The original intent
4839 was that statically linked binaries could for example just call
4840 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4841 link with digests. This never worked becayse SSLeay_add_all_digests()
4842 and SSLeay_add_all_ciphers() were in the same source file so calling
4843 one would link with the other. They are now in separate source files.
4846 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4849 *) Use a less unusual form of the Miller-Rabin primality test (it used
4850 a binary algorithm for exponentiation integrated into the Miller-Rabin
4851 loop, our standard modexp algorithms are faster).
4854 *) Support for the EBCDIC character set completed.
4855 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4857 *) Source code cleanups: use const where appropriate, eliminate casts,
4858 use void * instead of char * in lhash.
4861 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4862 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4863 this the server could overwrite ephemeral keys that the client
4867 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4868 using 50 iterations of the Rabin-Miller test.
4870 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4871 iterations of the Rabin-Miller test as required by the appendix
4872 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4873 As BN_is_prime_fasttest includes trial division, DSA parameter
4874 generation becomes much faster.
4876 This implies a change for the callback functions in DSA_is_prime
4877 and DSA_generate_parameters: The callback function is called once
4878 for each positive witness in the Rabin-Miller test, not just
4879 occasionally in the inner loop; and the parameters to the
4880 callback function now provide an iteration count for the outer
4881 loop rather than for the current invocation of the inner loop.
4882 DSA_generate_parameters additionally can call the callback
4883 function with an 'iteration count' of -1, meaning that a
4884 candidate has passed the trial division test (when q is generated
4885 from an application-provided seed, trial division is skipped).
4888 *) New function BN_is_prime_fasttest that optionally does trial
4889 division before starting the Rabin-Miller test and has
4890 an additional BN_CTX * argument (whereas BN_is_prime always
4891 has to allocate at least one BN_CTX).
4892 'callback(1, -1, cb_arg)' is called when a number has passed the
4893 trial division stage.
4896 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4900 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4903 *) New function BN_pseudo_rand().
4906 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4907 bignum version of BN_from_montgomery() with the working code from
4908 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
4912 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
4913 made it impossible to use the same SSL_SESSION data structure in
4914 SSL2 clients in multiple threads.
4917 *) The return value of RAND_load_file() no longer counts bytes obtained
4918 by stat(). RAND_load_file(..., -1) is new and uses the complete file
4919 to seed the PRNG (previously an explicit byte count was required).
4920 [Ulf Möller, Bodo Möller]
4922 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
4923 used (char *) instead of (void *) and had casts all over the place.
4926 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
4929 *) Retain source code compatibility for BN_prime_checks macro:
4930 BN_is_prime(..., BN_prime_checks, ...) now uses
4931 BN_prime_checks_for_size to determine the appropriate number of
4932 Rabin-Miller iterations.
4935 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
4936 DH_CHECK_P_NOT_SAFE_PRIME.
4937 (Check if this is true? OpenPGP calls them "strong".)
4940 *) Merge the functionality of "dh" and "gendh" programs into a new program
4941 "dhparam". The old programs are retained for now but will handle DH keys
4942 (instead of parameters) in future.
4945 *) Make the ciphers, s_server and s_client programs check the return values
4946 when a new cipher list is set.
4949 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
4950 ciphers. Before when the 56bit ciphers were enabled the sorting was
4953 The syntax for the cipher sorting has been extended to support sorting by
4954 cipher-strength (using the strength_bits hard coded in the tables).
4955 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
4957 Fix a bug in the cipher-command parser: when supplying a cipher command
4958 string with an "undefined" symbol (neither command nor alphanumeric
4959 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
4960 an error is flagged.
4962 Due to the strength-sorting extension, the code of the
4963 ssl_create_cipher_list() function was completely rearranged. I hope that
4964 the readability was also increased :-)
4965 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4967 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
4968 for the first serial number and places 2 in the serial number file. This
4969 avoids problems when the root CA is created with serial number zero and
4970 the first user certificate has the same issuer name and serial number
4974 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
4975 the new code. Add documentation for this stuff.
4978 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
4979 X509_*() to X509at_*() on the grounds that they don't handle X509
4980 structures and behave in an analagous way to the X509v3 functions:
4981 they shouldn't be called directly but wrapper functions should be used
4984 So we also now have some wrapper functions that call the X509at functions
4985 when passed certificate requests. (TO DO: similar things can be done with
4986 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
4987 things. Some of these need some d2i or i2d and print functionality
4988 because they handle more complex structures.)
4991 *) Add missing #ifndefs that caused missing symbols when building libssl
4992 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
4993 NO_RSA in ssl/s2*.c.
4994 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
4996 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
4997 has a return value which indicates the quality of the random data
4998 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
4999 error queue. New function RAND_pseudo_bytes() generates output that is
5000 guaranteed to be unique but not unpredictable. RAND_add is like
5001 RAND_seed, but takes an extra argument for an entropy estimate
5002 (RAND_seed always assumes full entropy).
5005 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5006 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5007 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5008 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5009 false-positive rate of at most 2^-80 for random input.
5012 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5015 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5016 in the 0.9.5 release), this returns the chain
5017 from an X509_CTX structure with a dup of the stack and all
5018 the X509 reference counts upped: so the stack will exist
5019 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5022 Also make SSL_SESSION_print() print out the verify return
5026 *) Add manpage for the pkcs12 command. Also change the default
5027 behaviour so MAC iteration counts are used unless the new
5028 -nomaciter option is used. This improves file security and
5029 only older versions of MSIE (4.0 for example) need it.
5032 *) Honor the no-xxx Configure options when creating .DEF files.
5035 *) Add PKCS#10 attributes to field table: challengePassword,
5036 unstructuredName and unstructuredAddress. These are taken from
5037 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5038 international characters are used.
5040 More changes to X509_ATTRIBUTE code: allow the setting of types
5041 based on strings. Remove the 'loc' parameter when adding
5042 attributes because these will be a SET OF encoding which is sorted
5046 *) Initial changes to the 'req' utility to allow request generation
5047 automation. This will allow an application to just generate a template
5048 file containing all the field values and have req construct the
5051 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5052 used all over the place including certificate requests and PKCS#7
5053 structures. They are currently handled manually where necessary with
5054 some primitive wrappers for PKCS#7. The new functions behave in a
5055 manner analogous to the X509 extension functions: they allow
5056 attributes to be looked up by NID and added.
5058 Later something similar to the X509V3 code would be desirable to
5059 automatically handle the encoding, decoding and printing of the
5060 more complex types. The string types like challengePassword can
5061 be handled by the string table functions.
5063 Also modified the multi byte string table handling. Now there is
5064 a 'global mask' which masks out certain types. The table itself
5065 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5066 is useful when for example there is only one permissible type
5067 (as in countryName) and using the mask might result in no valid
5071 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5072 SSL_get_peer_finished to allow applications to obtain the latest
5073 Finished messages sent to the peer or expected from the peer,
5074 respectively. (SSL_get_peer_finished is usually the Finished message
5075 actually received from the peer, otherwise the protocol will be aborted.)
5077 As the Finished message are message digests of the complete handshake
5078 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5079 be used for external authentication procedures when the authentication
5080 provided by SSL/TLS is not desired or is not enough.
5083 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5084 the host supports BWX extension and if Compaq C is present on the
5085 $PATH. Just exploiting of the BWX extension results in 20-30%
5086 performance kick for some algorithms, e.g. DES and RC4 to mention
5087 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5091 *) Add support for MS "fast SGC". This is arguably a violation of the
5092 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5093 weak crypto and after checking the certificate is SGC a second one
5094 with strong crypto. MS SGC stops the first handshake after receiving
5095 the server certificate message and sends a second client hello. Since
5096 a server will typically do all the time consuming operations before
5097 expecting any further messages from the client (server key exchange
5098 is the most expensive) there is little difference between the two.
5100 To get OpenSSL to support MS SGC we have to permit a second client
5101 hello message after we have sent server done. In addition we have to
5102 reset the MAC if we do get this second client hello.
5105 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5106 if a DER encoded private key is RSA or DSA traditional format. Changed
5107 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5108 format DER encoded private key. Newer code should use PKCS#8 format which
5109 has the key type encoded in the ASN1 structure. Added DER private key
5110 support to pkcs8 application.
5113 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5114 ciphersuites has been selected (as required by the SSL 3/TLS 1
5115 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5116 is set, we interpret this as a request to violate the specification
5117 (the worst that can happen is a handshake failure, and 'correct'
5118 behaviour would result in a handshake failure anyway).
5121 *) In SSL_CTX_add_session, take into account that there might be multiple
5122 SSL_SESSION structures with the same session ID (e.g. when two threads
5123 concurrently obtain them from an external cache).
5124 The internal cache can handle only one SSL_SESSION with a given ID,
5125 so if there's a conflict, we now throw out the old one to achieve
5129 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5130 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5131 some routines that use cipher OIDs: some ciphers do not have OIDs
5132 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5136 *) Simplify the trust setting structure and code. Now we just have
5137 two sequences of OIDs for trusted and rejected settings. These will
5138 typically have values the same as the extended key usage extension
5139 and any application specific purposes.
5141 The trust checking code now has a default behaviour: it will just
5142 check for an object with the same NID as the passed id. Functions can
5143 be provided to override either the default behaviour or the behaviour
5144 for a given id. SSL client, server and email already have functions
5145 in place for compatibility: they check the NID and also return "trusted"
5146 if the certificate is self signed.
5149 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5150 traditional format into an EVP_PKEY structure.
5153 *) Add a password callback function PEM_cb() which either prompts for
5154 a password if usr_data is NULL or otherwise assumes it is a null
5155 terminated password. Allow passwords to be passed on command line
5156 environment or config files in a few more utilities.
5159 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5160 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5161 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5162 Update documentation.
5165 *) Support for ASN1 "NULL" type. This could be handled before by using
5166 ASN1_TYPE but there wasn't any function that would try to read a NULL
5167 and produce an error if it couldn't. For compatibility we also have
5168 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5169 don't allocate anything because they don't need to.
5172 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5174 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5176 *) Rebuild of the memory allocation routines used by OpenSSL code and
5177 possibly others as well. The purpose is to make an interface that
5178 provide hooks so anyone can build a separate set of allocation and
5179 deallocation routines to be used by OpenSSL, for example memory
5180 pool implementations, or something else, which was previously hard
5181 since Malloc(), Realloc() and Free() were defined as macros having
5182 the values malloc, realloc and free, respectively (except for Win32
5183 compilations). The same is provided for memory debugging code.
5184 OpenSSL already comes with functionality to find memory leaks, but
5185 this gives people a chance to debug other memory problems.
5187 With these changes, a new set of functions and macros have appeared:
5189 CRYPTO_set_mem_debug_functions() [F]
5190 CRYPTO_get_mem_debug_functions() [F]
5191 CRYPTO_dbg_set_options() [F]
5192 CRYPTO_dbg_get_options() [F]
5193 CRYPTO_malloc_debug_init() [M]
5195 The memory debug functions are NULL by default, unless the library
5196 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5197 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5198 gives the standard debugging functions that come with OpenSSL) or
5199 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5200 provided by the library user) must be used. When the standard
5201 debugging functions are used, CRYPTO_dbg_set_options can be used to
5202 request additional information:
5203 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5204 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5206 Also, things like CRYPTO_set_mem_functions will always give the
5207 expected result (the new set of functions is used for allocation
5208 and deallocation) at all times, regardless of platform and compiler
5211 To finish it up, some functions that were never use in any other
5212 way than through macros have a new API and new semantic:
5215 CRYPTO_dbg_realloc()
5218 All macros of value have retained their old syntax.
5219 [Richard Levitte and Bodo Moeller]
5221 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5222 ordering of SMIMECapabilities wasn't in "strength order" and there
5223 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5227 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5228 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5229 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5231 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5232 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5233 functionality to handle multipart/signed properly) and a utility
5234 called 'smime' to call all this stuff. This is based on code I
5235 originally wrote for Celo who have kindly allowed it to be
5236 included in OpenSSL.
5239 *) Add variants des_set_key_checked and des_set_key_unchecked of
5240 des_set_key (aka des_key_sched). Global variable des_check_key
5241 decides which of these is called by des_set_key; this way
5242 des_check_key behaves as it always did, but applications and
5243 the library itself, which was buggy for des_check_key == 1,
5244 have a cleaner way to pick the version they need.
5247 *) New function PKCS12_newpass() which changes the password of a
5251 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5252 dynamic mix. In both cases the ids can be used as an index into the
5253 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5254 functions so they accept a list of the field values and the
5255 application doesn't need to directly manipulate the X509_TRUST
5259 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5263 *) Modify the way the V3 extension code looks up extensions. This now
5264 works in a similar way to the object code: we have some "standard"
5265 extensions in a static table which is searched with OBJ_bsearch()
5266 and the application can add dynamic ones if needed. The file
5267 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5268 updated whenever a new extension is added to the core code and kept
5269 in ext_nid order. There is a simple program 'tabtest.c' which checks
5270 this. New extensions are not added too often so this file can readily
5271 be maintained manually.
5273 There are two big advantages in doing things this way. The extensions
5274 can be looked up immediately and no longer need to be "added" using
5275 X509V3_add_standard_extensions(): this function now does nothing.
5276 [Side note: I get *lots* of email saying the extension code doesn't
5277 work because people forget to call this function]
5278 Also no dynamic allocation is done unless new extensions are added:
5279 so if we don't add custom extensions there is no need to call
5280 X509V3_EXT_cleanup().
5283 *) Modify enc utility's salting as follows: make salting the default. Add a
5284 magic header, so unsalted files fail gracefully instead of just decrypting
5285 to garbage. This is because not salting is a big security hole, so people
5286 should be discouraged from doing it.
5289 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5290 digest to be passed on the command line but it only used this
5291 parameter when signing a certificate. Modified so all relevant
5292 operations are affected by the digest parameter including the
5293 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5294 DSA key was used because it didn't fix the digest.
5297 *) Initial certificate chain verify code. Currently tests the untrusted
5298 certificates for consistency with the verify purpose (which is set
5299 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5301 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5302 this is because it will reject chains with invalid extensions whereas
5303 every previous version of OpenSSL and SSLeay made no checks at all.
5305 Trust code: checks the root CA for the relevant trust settings. Trust
5306 settings have an initial value consistent with the verify purpose: e.g.
5307 if the verify purpose is for SSL client use it expects the CA to be
5308 trusted for SSL client use. However the default value can be changed to
5309 permit custom trust settings: one example of this would be to only trust
5310 certificates from a specific "secure" set of CAs.
5312 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5313 which should be used for version portability: especially since the
5314 verify structure is likely to change more often now.
5316 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5317 to set them. If not set then assume SSL clients will verify SSL servers
5320 Two new options to the verify program: -untrusted allows a set of
5321 untrusted certificates to be passed in and -purpose which sets the
5322 intended purpose of the certificate. If a purpose is set then the
5323 new chain verify code is used to check extension consistency.
5326 *) Support for the authority information access extension.
5329 *) Modify RSA and DSA PEM read routines to transparently handle
5330 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5331 public keys in a format compatible with certificate
5332 SubjectPublicKeyInfo structures. Unfortunately there were already
5333 functions called *_PublicKey_* which used various odd formats so
5334 these are retained for compatibility: however the DSA variants were
5335 never in a public release so they have been deleted. Changed dsa/rsa
5336 utilities to handle the new format: note no releases ever handled public
5337 keys so we should be OK.
5339 The primary motivation for this change is to avoid the same fiasco
5340 that dogs private keys: there are several incompatible private key
5341 formats some of which are standard and some OpenSSL specific and
5342 require various evil hacks to allow partial transparent handling and
5343 even then it doesn't work with DER formats. Given the option anything
5344 other than PKCS#8 should be dumped: but the other formats have to
5345 stay in the name of compatibility.
5347 With public keys and the benefit of hindsight one standard format
5348 is used which works with EVP_PKEY, RSA or DSA structures: though
5349 it clearly returns an error if you try to read the wrong kind of key.
5351 Added a -pubkey option to the 'x509' utility to output the public key.
5352 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5353 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5354 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5355 that do the same as the EVP_PKEY_assign_*() except they up the
5356 reference count of the added key (they don't "swallow" the
5360 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5361 CRLs would fail if the file contained no certificates or no CRLs:
5362 added a new function to read in both types and return the number
5363 read: this means that if none are read it will be an error. The
5364 DER versions of the certificate and CRL reader would always fail
5365 because it isn't possible to mix certificates and CRLs in DER format
5366 without choking one or the other routine. Changed this to just read
5367 a certificate: this is the best we can do. Also modified the code
5368 in apps/verify.c to take notice of return codes: it was previously
5369 attempting to read in certificates from NULL pointers and ignoring
5370 any errors: this is one reason why the cert and CRL reader seemed
5371 to work. It doesn't check return codes from the default certificate
5372 routines: these may well fail if the certificates aren't installed.
5375 *) Code to support otherName option in GeneralName.
5378 *) First update to verify code. Change the verify utility
5379 so it warns if it is passed a self signed certificate:
5380 for consistency with the normal behaviour. X509_verify
5381 has been modified to it will now verify a self signed
5382 certificate if *exactly* the same certificate appears
5383 in the store: it was previously impossible to trust a
5384 single self signed certificate. This means that:
5385 openssl verify ss.pem
5386 now gives a warning about a self signed certificate but
5387 openssl verify -CAfile ss.pem ss.pem
5391 *) For servers, store verify_result in SSL_SESSION data structure
5392 (and add it to external session representation).
5393 This is needed when client certificate verifications fails,
5394 but an application-provided verification callback (set by
5395 SSL_CTX_set_cert_verify_callback) allows accepting the session
5396 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5397 but returns 1): When the session is reused, we have to set
5398 ssl->verify_result to the appropriate error code to avoid
5400 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5402 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5403 case in PKCS7_dataInit() where the signed PKCS7 structure
5404 didn't contain any existing data because it was being created.
5405 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5407 *) Add a salt to the key derivation routines in enc.c. This
5408 forms the first 8 bytes of the encrypted file. Also add a
5409 -S option to allow a salt to be input on the command line.
5412 *) New function X509_cmp(). Oddly enough there wasn't a function
5413 to compare two certificates. We do this by working out the SHA1
5414 hash and comparing that. X509_cmp() will be needed by the trust
5418 *) SSL_get1_session() is like SSL_get_session(), but increments
5419 the reference count in the SSL_SESSION returned.
5420 [Geoff Thorpe <geoff@eu.c2.net>]
5422 *) Fix for 'req': it was adding a null to request attributes.
5423 Also change the X509_LOOKUP and X509_INFO code to handle
5424 certificate auxiliary information.
5427 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5431 *) Add the possibility to add extra information to the memory leak
5432 detecting output, to form tracebacks, showing from where each
5433 allocation was originated: CRYPTO_push_info("constant string") adds
5434 the string plus current file name and line number to a per-thread
5435 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5436 is like calling CYRPTO_pop_info() until the stack is empty.
5437 Also updated memory leak detection code to be multi-thread-safe.
5440 *) Add options -text and -noout to pkcs7 utility and delete the
5441 encryption options which never did anything. Update docs.
5444 *) Add options to some of the utilities to allow the pass phrase
5445 to be included on either the command line (not recommended on
5446 OSes like Unix) or read from the environment. Update the
5447 manpages and fix a few bugs.
5450 *) Add a few manpages for some of the openssl commands.
5453 *) Fix the -revoke option in ca. It was freeing up memory twice,
5454 leaking and not finding already revoked certificates.
5457 *) Extensive changes to support certificate auxiliary information.
5458 This involves the use of X509_CERT_AUX structure and X509_AUX
5459 functions. An X509_AUX function such as PEM_read_X509_AUX()
5460 can still read in a certificate file in the usual way but it
5461 will also read in any additional "auxiliary information". By
5462 doing things this way a fair degree of compatibility can be
5463 retained: existing certificates can have this information added
5464 using the new 'x509' options.
5466 Current auxiliary information includes an "alias" and some trust
5467 settings. The trust settings will ultimately be used in enhanced
5468 certificate chain verification routines: currently a certificate
5469 can only be trusted if it is self signed and then it is trusted
5473 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5474 The problem was that one of the replacement routines had not been working
5475 since SSLeay releases. For now the offending routine has been replaced
5476 with non-optimised assembler. Even so, this now gives around 95%
5477 performance improvement for 1024 bit RSA signs.
5480 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5481 handling. Most clients have the effective key size in bits equal to
5482 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5483 A few however don't do this and instead use the size of the decrypted key
5484 to determine the RC2 key length and the AlgorithmIdentifier to determine
5485 the effective key length. In this case the effective key length can still
5486 be 40 bits but the key length can be 168 bits for example. This is fixed
5487 by manually forcing an RC2 key into the EVP_PKEY structure because the
5488 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5489 the key length and effective key length are equal.
5492 *) Add a bunch of functions that should simplify the creation of
5493 X509_NAME structures. Now you should be able to do:
5494 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5495 and have it automatically work out the correct field type and fill in
5496 the structures. The more adventurous can try:
5497 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5498 and it will (hopefully) work out the correct multibyte encoding.
5501 *) Change the 'req' utility to use the new field handling and multibyte
5502 copy routines. Before the DN field creation was handled in an ad hoc
5503 way in req, ca, and x509 which was rather broken and didn't support
5504 BMPStrings or UTF8Strings. Since some software doesn't implement
5505 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5506 using the dirstring_type option. See the new comment in the default
5507 openssl.cnf for more info.
5510 *) Make crypto/rand/md_rand.c more robust:
5511 - Assure unique random numbers after fork().
5512 - Make sure that concurrent threads access the global counter and
5513 md serializably so that we never lose entropy in them
5514 or use exactly the same state in multiple threads.
5515 Access to the large state is not always serializable because
5516 the additional locking could be a performance killer, and
5517 md should be large enough anyway.
5520 *) New file apps/app_rand.c with commonly needed functionality
5521 for handling the random seed file.
5523 Use the random seed file in some applications that previously did not:
5525 dsaparam -genkey (which also ignored its '-rand' option),
5528 x509 (when signing).
5529 Except on systems with /dev/urandom, it is crucial to have a random
5530 seed file at least for key creation, DSA signing, and for DH exchanges;
5531 for RSA signatures we could do without one.
5533 gendh and gendsa (unlike genrsa) used to read only the first byte
5534 of each file listed in the '-rand' option. The function as previously
5535 found in genrsa is now in app_rand.c and is used by all programs
5536 that support '-rand'.
5539 *) In RAND_write_file, use mode 0600 for creating files;
5540 don't just chmod when it may be too late.
5543 *) Report an error from X509_STORE_load_locations
5544 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5547 *) New function ASN1_mbstring_copy() this copies a string in either
5548 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5549 into an ASN1_STRING type. A mask of permissible types is passed
5550 and it chooses the "minimal" type to use or an error if not type
5554 *) Add function equivalents to the various macros in asn1.h. The old
5555 macros are retained with an M_ prefix. Code inside the library can
5556 use the M_ macros. External code (including the openssl utility)
5557 should *NOT* in order to be "shared library friendly".
5560 *) Add various functions that can check a certificate's extensions
5561 to see if it usable for various purposes such as SSL client,
5562 server or S/MIME and CAs of these types. This is currently
5563 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5564 verification. Also added a -purpose flag to x509 utility to
5565 print out all the purposes.
5568 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5572 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5573 for, obtain and decode and extension and obtain its critical flag.
5574 This allows all the necessary extension code to be handled in a
5575 single function call.
5578 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5579 platforms. See crypto/rc4/rc4_enc.c for further details.
5582 *) New -noout option to asn1parse. This causes no output to be produced
5583 its main use is when combined with -strparse and -out to extract data
5584 from a file (which may not be in ASN.1 format).
5587 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5588 when producing the local key id.
5589 [Richard Levitte <levitte@stacken.kth.se>]
5591 *) New option -dhparam in s_server. This allows a DH parameter file to be
5592 stated explicitly. If it is not stated then it tries the first server
5593 certificate file. The previous behaviour hard coded the filename
5597 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5598 a public key to be input or output. For example:
5599 openssl rsa -in key.pem -pubout -out pubkey.pem
5600 Also added necessary DSA public key functions to handle this.
5603 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5604 in the message. This was handled by allowing
5605 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5606 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5608 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5609 to the end of the strings whereas this didn't. This would cause problems
5610 if strings read with d2i_ASN1_bytes() were later modified.
5611 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5613 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5614 data and it contains EOF it will end up returning an error. This is
5615 caused by input 46 bytes long. The cause is due to the way base64
5616 BIOs find the start of base64 encoded data. They do this by trying a
5617 trial decode on each line until they find one that works. When they
5618 do a flag is set and it starts again knowing it can pass all the
5619 data directly through the decoder. Unfortunately it doesn't reset
5620 the context it uses. This means that if EOF is reached an attempt
5621 is made to pass two EOFs through the context and this causes the
5622 resulting error. This can also cause other problems as well. As is
5623 usual with these problems it takes *ages* to find and the fix is
5624 trivial: move one line.
5625 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5627 *) Ugly workaround to get s_client and s_server working under Windows. The
5628 old code wouldn't work because it needed to select() on sockets and the
5629 tty (for keypresses and to see if data could be written). Win32 only
5630 supports select() on sockets so we select() with a 1s timeout on the
5631 sockets and then see if any characters are waiting to be read, if none
5632 are present then we retry, we also assume we can always write data to
5633 the tty. This isn't nice because the code then blocks until we've
5634 received a complete line of data and it is effectively polling the
5635 keyboard at 1s intervals: however it's quite a bit better than not
5636 working at all :-) A dedicated Windows application might handle this
5637 with an event loop for example.
5640 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5641 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5642 will be called when RSA_sign() and RSA_verify() are used. This is useful
5643 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5644 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5645 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5646 This necessitated the support of an extra signature type NID_md5_sha1
5647 for SSL signatures and modifications to the SSL library to use it instead
5648 of calling RSA_public_decrypt() and RSA_private_encrypt().
5651 *) Add new -verify -CAfile and -CApath options to the crl program, these
5652 will lookup a CRL issuers certificate and verify the signature in a
5653 similar way to the verify program. Tidy up the crl program so it
5654 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5655 less strict. It will now permit CRL extensions even if it is not
5656 a V2 CRL: this will allow it to tolerate some broken CRLs.
5659 *) Initialize all non-automatic variables each time one of the openssl
5660 sub-programs is started (this is necessary as they may be started
5661 multiple times from the "OpenSSL>" prompt).
5662 [Lennart Bang, Bodo Moeller]
5664 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5665 removing all other RSA functionality (this is what NO_RSA does). This
5666 is so (for example) those in the US can disable those operations covered
5667 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5671 *) Non-copying interface to BIO pairs.
5672 (still largely untested)
5675 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5676 ASCII string. This was handled independently in various places before.
5679 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5680 UTF8 strings a character at a time.
5683 *) Use client_version from client hello to select the protocol
5684 (s23_srvr.c) and for RSA client key exchange verification
5685 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5688 *) Add various utility functions to handle SPKACs, these were previously
5689 handled by poking round in the structure internals. Added new function
5690 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5691 print, verify and generate SPKACs. Based on an original idea from
5692 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5695 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5698 *) Allow the config file extension section to be overwritten on the
5699 command line. Based on an original idea from Massimiliano Pala
5700 <madwolf@comune.modena.it>. The new option is called -extensions
5701 and can be applied to ca, req and x509. Also -reqexts to override
5702 the request extensions in req and -crlexts to override the crl extensions
5706 *) Add new feature to the SPKAC handling in ca. Now you can include
5707 the same field multiple times by preceding it by "XXXX." for example:
5710 this is the same syntax as used in the req config file.
5713 *) Allow certificate extensions to be added to certificate requests. These
5714 are specified in a 'req_extensions' option of the req section of the
5715 config file. They can be printed out with the -text option to req but
5716 are otherwise ignored at present.
5719 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5720 data read consists of only the final block it would not decrypted because
5721 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5722 A misplaced 'break' also meant the decrypted final block might not be
5723 copied until the next read.
5726 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5727 a few extra parameters to the DH structure: these will be useful if
5728 for example we want the value of 'q' or implement X9.42 DH.
5731 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5732 provides hooks that allow the default DSA functions or functions on a
5733 "per key" basis to be replaced. This allows hardware acceleration and
5734 hardware key storage to be handled without major modification to the
5735 library. Also added low level modexp hooks and CRYPTO_EX structure and
5736 associated functions.
5739 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5740 as "read only": it can't be written to and the buffer it points to will
5741 not be freed. Reading from a read only BIO is much more efficient than
5742 a normal memory BIO. This was added because there are several times when
5743 an area of memory needs to be read from a BIO. The previous method was
5744 to create a memory BIO and write the data to it, this results in two
5745 copies of the data and an O(n^2) reading algorithm. There is a new
5746 function BIO_new_mem_buf() which creates a read only memory BIO from
5747 an area of memory. Also modified the PKCS#7 routines to use read only
5751 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5752 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5753 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5754 but a retry condition occured while trying to read the rest.
5757 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5758 NID_pkcs7_encrypted by default: this was wrong since this should almost
5759 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5760 the encrypted data type: this is a more sensible place to put it and it
5761 allows the PKCS#12 code to be tidied up that duplicated this
5765 *) Changed obj_dat.pl script so it takes its input and output files on
5766 the command line. This should avoid shell escape redirection problems
5770 *) Initial support for certificate extension requests, these are included
5771 in things like Xenroll certificate requests. Included functions to allow
5772 extensions to be obtained and added.
5775 *) -crlf option to s_client and s_server for sending newlines as
5776 CRLF (as required by many protocols).
5779 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5781 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5782 [Ralf S. Engelschall]
5784 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5785 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5787 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5791 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5792 DH parameters/keys (q is lost during that conversion, but the resulting
5793 DH parameters contain its length).
5795 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5796 much faster than DH_generate_parameters (which creates parameters
5797 where p = 2*q + 1), and also the smaller q makes DH computations
5798 much more efficient (160-bit exponentiation instead of 1024-bit
5799 exponentiation); so this provides a convenient way to support DHE
5800 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5801 utter importance to use
5802 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5804 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5805 when such DH parameters are used, because otherwise small subgroup
5806 attacks may become possible!
5809 *) Avoid memory leak in i2d_DHparams.
5812 *) Allow the -k option to be used more than once in the enc program:
5813 this allows the same encrypted message to be read by multiple recipients.
5816 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5817 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5818 it will always use the numerical form of the OID, even if it has a short
5822 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5823 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5824 otherwise bn_mod_exp was called. In the case of hardware keys for example
5825 no private key components need be present and it might store extra data
5826 in the RSA structure, which cannot be accessed from bn_mod_exp.
5827 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5828 private key operations.
5831 *) Added support for SPARC Linux.
5834 *) pem_password_cb function type incompatibly changed from
5835 typedef int pem_password_cb(char *buf, int size, int rwflag);
5837 ....(char *buf, int size, int rwflag, void *userdata);
5838 so that applications can pass data to their callbacks:
5839 The PEM[_ASN1]_{read,write}... functions and macros now take an
5840 additional void * argument, which is just handed through whenever
5841 the password callback is called.
5842 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5844 New function SSL_CTX_set_default_passwd_cb_userdata.
5846 Compatibility note: As many C implementations push function arguments
5847 onto the stack in reverse order, the new library version is likely to
5848 interoperate with programs that have been compiled with the old
5849 pem_password_cb definition (PEM_whatever takes some data that
5850 happens to be on the stack as its last argument, and the callback
5851 just ignores this garbage); but there is no guarantee whatsoever that
5854 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5855 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5856 problems not only on Windows, but also on some Unix platforms.
5857 To avoid problematic command lines, these definitions are now in an
5858 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5859 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5862 *) MIPS III/IV assembler module is reimplemented.
5865 *) More DES library cleanups: remove references to srand/rand and
5866 delete an unused file.
5869 *) Add support for the the free Netwide assembler (NASM) under Win32,
5870 since not many people have MASM (ml) and it can be hard to obtain.
5871 This is currently experimental but it seems to work OK and pass all
5872 the tests. Check out INSTALL.W32 for info.
5875 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5876 without temporary keys kept an extra copy of the server key,
5877 and connections with temporary keys did not free everything in case
5881 *) New function RSA_check_key and new openssl rsa option -check
5882 for verifying the consistency of RSA keys.
5883 [Ulf Moeller, Bodo Moeller]
5885 *) Various changes to make Win32 compile work:
5886 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5887 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5888 comparison" warnings.
5889 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5892 *) Add a debugging option to PKCS#5 v2 key generation function: when
5893 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5894 derived keys are printed to stderr.
5897 *) Copy the flags in ASN1_STRING_dup().
5898 [Roman E. Pavlov <pre@mo.msk.ru>]
5900 *) The x509 application mishandled signing requests containing DSA
5901 keys when the signing key was also DSA and the parameters didn't match.
5903 It was supposed to omit the parameters when they matched the signing key:
5904 the verifying software was then supposed to automatically use the CA's
5905 parameters if they were absent from the end user certificate.
5907 Omitting parameters is no longer recommended. The test was also
5908 the wrong way round! This was probably due to unusual behaviour in
5909 EVP_cmp_parameters() which returns 1 if the parameters match.
5910 This meant that parameters were omitted when they *didn't* match and
5911 the certificate was useless. Certificates signed with 'ca' didn't have
5913 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
5915 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
5916 The interface is as follows:
5917 Applications can use
5918 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
5919 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
5920 "off" is now the default.
5921 The library internally uses
5922 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
5923 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
5924 to disable memory-checking temporarily.
5926 Some inconsistent states that previously were possible (and were
5927 even the default) are now avoided.
5929 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
5930 with each memory chunk allocated; this is occasionally more helpful
5931 than just having a counter.
5933 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
5935 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
5939 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
5940 which largely parallels "options", but is for changing API behaviour,
5941 whereas "options" are about protocol behaviour.
5942 Initial "mode" flags are:
5944 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
5945 a single record has been written.
5946 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
5947 retries use the same buffer location.
5948 (But all of the contents must be
5952 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
5955 *) Fix problems with no-hmac etc.
5956 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
5958 *) New functions RSA_get_default_method(), RSA_set_method() and
5959 RSA_get_method(). These allows replacement of RSA_METHODs without having
5960 to mess around with the internals of an RSA structure.
5963 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
5964 Also really enable memory leak checks in openssl.c and in some
5966 [Chad C. Mulligan, Bodo Moeller]
5968 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
5969 up the length of negative integers. This has now been simplified to just
5970 store the length when it is first determined and use it later, rather
5971 than trying to keep track of where data is copied and updating it to
5973 [Steve Henson, reported by Brien Wheeler
5974 <bwheeler@authentica-security.com>]
5976 *) Add a new function PKCS7_signatureVerify. This allows the verification
5977 of a PKCS#7 signature but with the signing certificate passed to the
5978 function itself. This contrasts with PKCS7_dataVerify which assumes the
5979 certificate is present in the PKCS#7 structure. This isn't always the
5980 case: certificates can be omitted from a PKCS#7 structure and be
5981 distributed by "out of band" means (such as a certificate database).
5984 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
5985 function prototypes in pem.h, also change util/mkdef.pl to add the
5986 necessary function names.
5989 *) mk1mf.pl (used by Windows builds) did not properly read the
5990 options set by Configure in the top level Makefile, and Configure
5991 was not even able to write more than one option correctly.
5992 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
5995 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
5996 file to be loaded from a BIO or FILE pointer. The BIO version will
5997 for example allow memory BIOs to contain config info.
6000 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6001 Whoever hopes to achieve shared-library compatibility across versions
6002 must use this, not the compile-time macro.
6003 (Exercise 0.9.4: Which is the minimum library version required by
6005 Note: All this applies only to multi-threaded programs, others don't
6009 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6010 through a BIO pair triggered the default case, i.e.
6011 SSLerr(...,SSL_R_UNKNOWN_STATE).
6014 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6015 can use the SSL library even if none of the specific BIOs is
6019 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6020 for the encoded length.
6021 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6023 *) Add initial documentation of the X509V3 functions.
6026 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6027 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6028 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6029 secure PKCS#8 private key format with a high iteration count.
6032 *) Fix determination of Perl interpreter: A perl or perl5
6033 _directory_ in $PATH was also accepted as the interpreter.
6034 [Ralf S. Engelschall]
6036 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6037 wrong with it but it was very old and did things like calling
6038 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6042 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6043 to use the new extension code.
6046 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6047 with macros. This should make it easier to change their form, add extra
6048 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6052 *) Add to configuration table a new entry that can specify an alternative
6053 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6054 according to Mark Crispin <MRC@Panda.COM>.
6058 *) DES CBC did not update the IV. Weird.
6061 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6062 Changing the behaviour of the former might break existing programs --
6063 where IV updating is needed, des_ncbc_encrypt can be used.
6066 *) When bntest is run from "make test" it drives bc to check its
6067 calculations, as well as internally checking them. If an internal check
6068 fails, it needs to cause bc to give a non-zero result or make test carries
6069 on without noticing the failure. Fixed.
6072 *) DES library cleanups.
6075 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6076 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6077 ciphers. NOTE: although the key derivation function has been verified
6078 against some published test vectors it has not been extensively tested
6079 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6083 *) Instead of "mkdir -p", which is not fully portable, use new
6084 Perl script "util/mkdir-p.pl".
6087 *) Rewrite the way password based encryption (PBE) is handled. It used to
6088 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6089 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6090 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6091 the 'parameter' field of the AlgorithmIdentifier is passed to the
6092 underlying key generation function so it must do its own ASN1 parsing.
6093 This has also changed the EVP_PBE_CipherInit() function which now has a
6094 'parameter' argument instead of literal salt and iteration count values
6095 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6098 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6099 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6100 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6101 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6102 value was just used as a "magic string" and not used directly its
6103 value doesn't matter.
6106 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6110 *) "linux-sparc64" configuration (ultrapenguin).
6111 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6112 "linux-sparc" configuration.
6113 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6115 *) config now generates no-xxx options for missing ciphers.
6118 *) Support the EBCDIC character set (work in progress).
6119 File ebcdic.c not yet included because it has a different license.
6120 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6122 *) Support BS2000/OSD-POSIX.
6123 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6125 *) Make callbacks for key generation use void * instead of char *.
6128 *) Make S/MIME samples compile (not yet tested).
6131 *) Additional typesafe stacks.
6134 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6138 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6140 *) New configuration variant "sco5-gcc".
6142 *) Updated some demos.
6143 [Sean O Riordain, Wade Scholine]
6145 *) Add missing BIO_free at exit of pkcs12 application.
6148 *) Fix memory leak in conf.c.
6151 *) Updates for Win32 to assembler version of MD5.
6154 *) Set #! path to perl in apps/der_chop to where we found it
6155 instead of using a fixed path.
6158 *) SHA library changes for irix64-mips4-cc.
6161 *) Improvements for VMS support.
6165 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6167 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6168 This also avoids the problems with SC4.2 and unpatched SC5.
6169 [Andy Polyakov <appro@fy.chalmers.se>]
6171 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6172 These are required because of the typesafe stack would otherwise break
6173 existing code. If old code used a structure member which used to be STACK
6174 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6175 sk_num or sk_value it would produce an error because the num, data members
6176 are not present in STACK_OF. Now it just produces a warning. sk_set
6177 replaces the old method of assigning a value to sk_value
6178 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6179 that does this will no longer work (and should use sk_set instead) but
6180 this could be regarded as a "questionable" behaviour anyway.
6183 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6184 correctly handle encrypted S/MIME data.
6187 *) Change type of various DES function arguments from des_cblock
6188 (which means, in function argument declarations, pointer to char)
6189 to des_cblock * (meaning pointer to array with 8 char elements),
6190 which allows the compiler to do more typechecking; it was like
6191 that back in SSLeay, but with lots of ugly casts.
6193 Introduce new type const_des_cblock.
6196 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6197 problems: find RecipientInfo structure that matches recipient certificate
6198 and initialise the ASN1 structures properly based on passed cipher.
6201 *) Belatedly make the BN tests actually check the results.
6204 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6205 to and from BNs: it was completely broken. New compilation option
6206 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6207 key elements as negative integers.
6210 *) Reorganize and speed up MD5.
6211 [Andy Polyakov <appro@fy.chalmers.se>]
6214 [Richard Levitte <richard@levitte.org>]
6216 *) New option -out to asn1parse to allow the parsed structure to be
6217 output to a file. This is most useful when combined with the -strparse
6218 option to examine the output of things like OCTET STRINGS.
6221 *) Make SSL library a little more fool-proof by not requiring any longer
6222 that SSL_set_{accept,connect}_state be called before
6223 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6224 in many applications because usually everything *appeared* to work as
6225 intended anyway -- now it really works as intended).
6228 *) Move openssl.cnf out of lib/.
6231 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6232 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6233 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6234 [Ralf S. Engelschall]
6236 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6237 handle PKCS#7 enveloped data properly.
6238 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6240 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6241 copying pointers. The cert_st handling is changed by this in
6242 various ways (and thus what used to be known as ctx->default_cert
6243 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6244 any longer when s->cert does not give us what we need).
6245 ssl_cert_instantiate becomes obsolete by this change.
6246 As soon as we've got the new code right (possibly it already is?),
6247 we have solved a couple of bugs of the earlier code where s->cert
6248 was used as if it could not have been shared with other SSL structures.
6250 Note that using the SSL API in certain dirty ways now will result
6251 in different behaviour than observed with earlier library versions:
6252 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6253 does not influence s as it used to.
6255 In order to clean up things more thoroughly, inside SSL_SESSION
6256 we don't use CERT any longer, but a new structure SESS_CERT
6257 that holds per-session data (if available); currently, this is
6258 the peer's certificate chain and, for clients, the server's certificate
6259 and temporary key. CERT holds only those values that can have
6260 meaningful defaults in an SSL_CTX.
6263 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6264 from the internal representation. Various PKCS#7 fixes: remove some
6265 evil casts and set the enc_dig_alg field properly based on the signing
6269 *) Allow PKCS#12 password to be set from the command line or the
6270 environment. Let 'ca' get its config file name from the environment
6271 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6275 *) Allow certificate policies extension to use an IA5STRING for the
6276 organization field. This is contrary to the PKIX definition but
6277 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6281 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6282 without disallowing inline assembler and the like for non-pedantic builds.
6285 *) Support Borland C++ builder.
6286 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6291 *) SHA-1 cleanups and performance enhancements.
6292 [Andy Polyakov <appro@fy.chalmers.se>]
6294 *) Sparc v8plus assembler for the bignum library.
6295 [Andy Polyakov <appro@fy.chalmers.se>]
6297 *) Accept any -xxx and +xxx compiler options in Configure.
6300 *) Update HPUX configuration.
6303 *) Add missing sk_<type>_unshift() function to safestack.h
6304 [Ralf S. Engelschall]
6306 *) New function SSL_CTX_use_certificate_chain_file that sets the
6307 "extra_cert"s in addition to the certificate. (This makes sense
6308 only for "PEM" format files, as chains as a whole are not
6312 *) Support verify_depth from the SSL API.
6313 x509_vfy.c had what can be considered an off-by-one-error:
6314 Its depth (which was not part of the external interface)
6315 was actually counting the number of certificates in a chain;
6316 now it really counts the depth.
6319 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6320 instead of X509err, which often resulted in confusing error
6321 messages since the error codes are not globally unique
6322 (e.g. an alleged error in ssl3_accept when a certificate
6323 didn't match the private key).
6325 *) New function SSL_CTX_set_session_id_context that allows to set a default
6326 value (so that you don't need SSL_set_session_id_context for each
6327 connection using the SSL_CTX).
6330 *) OAEP decoding bug fix.
6333 *) Support INSTALL_PREFIX for package builders, as proposed by
6337 *) New Configure options "threads" and "no-threads". For systems
6338 where the proper compiler options are known (currently Solaris
6339 and Linux), "threads" is the default.
6342 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6345 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6346 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6347 such as /usr/local/bin.
6350 *) "make linux-shared" to build shared libraries.
6351 [Niels Poppe <niels@netbox.org>]
6353 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6356 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6357 extension adding in x509 utility.
6360 *) Remove NOPROTO sections and error code comments.
6363 *) Partial rewrite of the DEF file generator to now parse the ANSI
6367 *) New Configure options --prefix=DIR and --openssldir=DIR.
6370 *) Complete rewrite of the error code script(s). It is all now handled
6371 by one script at the top level which handles error code gathering,
6372 header rewriting and C source file generation. It should be much better
6373 than the old method: it now uses a modified version of Ulf's parser to
6374 read the ANSI prototypes in all header files (thus the old K&R definitions
6375 aren't needed for error creation any more) and do a better job of
6376 translating function codes into names. The old 'ASN1 error code imbedded
6377 in a comment' is no longer necessary and it doesn't use .err files which
6378 have now been deleted. Also the error code call doesn't have to appear all
6379 on one line (which resulted in some large lines...).
6382 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6385 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6386 0 (which usually indicates a closed connection), but continue reading.
6389 *) Fix some race conditions.
6392 *) Add support for CRL distribution points extension. Add Certificate
6393 Policies and CRL distribution points documentation.
6396 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6399 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6400 8 of keying material. Merlin has also confirmed interop with this fix
6401 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6402 [Merlin Hughes <merlin@baltimore.ie>]
6404 *) Fix lots of warnings.
6405 [Richard Levitte <levitte@stacken.kth.se>]
6407 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6408 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6409 [Richard Levitte <levitte@stacken.kth.se>]
6411 *) Fix problems with sizeof(long) == 8.
6412 [Andy Polyakov <appro@fy.chalmers.se>]
6414 *) Change functions to ANSI C.
6417 *) Fix typos in error codes.
6418 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6420 *) Remove defunct assembler files from Configure.
6423 *) SPARC v8 assembler BIGNUM implementation.
6424 [Andy Polyakov <appro@fy.chalmers.se>]
6426 *) Support for Certificate Policies extension: both print and set.
6427 Various additions to support the r2i method this uses.
6430 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6431 return a const string when you are expecting an allocated buffer.
6434 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6435 types DirectoryString and DisplayText.
6438 *) Add code to allow r2i extensions to access the configuration database,
6439 add an LHASH database driver and add several ctx helper functions.
6442 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6443 fail when they extended the size of a BIGNUM.
6446 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6447 support typesafe stack.
6450 *) Fix typo in SSL_[gs]et_options().
6451 [Nils Frostberg <nils@medcom.se>]
6453 *) Delete various functions and files that belonged to the (now obsolete)
6454 old X509V3 handling code.
6457 *) New Configure option "rsaref".
6460 *) Don't auto-generate pem.h.
6463 *) Introduce type-safe ASN.1 SETs.
6466 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6467 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6469 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6470 that links with OpenSSL (well at least cause lots of warnings), but fear
6471 not: the conversion is trivial, and it eliminates loads of evil casts. A
6472 few STACKed things have been converted already. Feel free to convert more.
6473 In the fullness of time, I'll do away with the STACK type altogether.
6476 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6477 specified in <certfile> by updating the entry in the index.txt file.
6478 This way one no longer has to edit the index.txt file manually for
6479 revoking a certificate. The -revoke option does the gory details now.
6480 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6482 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6483 `-text' option at all and this way the `-noout -text' combination was
6484 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6485 [Ralf S. Engelschall]
6487 *) Make sure a corresponding plain text error message exists for the
6488 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6489 verify callback function determined that a certificate was revoked.
6490 [Ralf S. Engelschall]
6492 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6493 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6494 all available cipers including rc5, which was forgotten until now.
6495 In order to let the testing shell script know which algorithms
6496 are available, a new (up to now undocumented) command
6497 "openssl list-cipher-commands" is used.
6500 *) Bugfix: s_client occasionally would sleep in select() when
6501 it should have checked SSL_pending() first.
6504 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6505 the raw DSA values prior to ASN.1 encoding.
6508 *) Tweaks to Configure
6509 [Niels Poppe <niels@netbox.org>]
6511 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6515 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6518 *) New config option to avoid instructions that are illegal on the 80386.
6519 The default code is faster, but requires at least a 486.
6522 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6523 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6524 same as SSL2_VERSION anyway.
6527 *) New "-showcerts" option for s_client.
6530 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6531 application. Various cleanups and fixes.
6534 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6535 modify error routines to work internally. Add error codes and PBE init
6536 to library startup routines.
6539 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6540 packing functions to asn1 and evp. Changed function names and error
6541 codes along the way.
6544 *) PKCS12 integration: and so it begins... First of several patches to
6545 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6546 objects to objects.h
6549 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6550 and display support for Thawte strong extranet extension.
6553 *) Add LinuxPPC support.
6554 [Jeff Dubrule <igor@pobox.org>]
6556 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6557 bn_div_words in alpha.s.
6558 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6560 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6561 OAEP isn't supported when OpenSSL is built with RSAref.
6562 [Ulf Moeller <ulf@fitug.de>]
6564 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6565 so they no longer are missing under -DNOPROTO.
6566 [Soren S. Jorvang <soren@t.dk>]
6569 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6571 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6572 doesn't work when the session is reused. Coming soon!
6575 *) Fix a security hole, that allows sessions to be reused in the wrong
6576 context thus bypassing client cert protection! All software that uses
6577 client certs and session caches in multiple contexts NEEDS PATCHING to
6578 allow session reuse! A fuller solution is in the works.
6579 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6581 *) Some more source tree cleanups (removed obsolete files
6582 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6583 permission on "config" script to be executable) and a fix for the INSTALL
6585 [Ulf Moeller <ulf@fitug.de>]
6587 *) Remove some legacy and erroneous uses of malloc, free instead of
6589 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6591 *) Make rsa_oaep_test return non-zero on error.
6592 [Ulf Moeller <ulf@fitug.de>]
6594 *) Add support for native Solaris shared libraries. Configure
6595 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6596 if someone would make that last step automatic.
6597 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6599 *) ctx_size was not built with the right compiler during "make links". Fixed.
6602 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6603 except NULL ciphers". This means the default cipher list will no longer
6604 enable NULL ciphers. They need to be specifically enabled e.g. with
6605 the string "DEFAULT:eNULL".
6608 *) Fix to RSA private encryption routines: if p < q then it would
6609 occasionally produce an invalid result. This will only happen with
6610 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6613 *) Be less restrictive and allow also `perl util/perlpath.pl
6614 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6615 because this way one can also use an interpreter named `perl5' (which is
6616 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6617 installed as `perl').
6618 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6620 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6621 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6623 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6624 advapi32.lib to Win32 build and change the pem test comparision
6625 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6626 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6627 and crypto/des/ede_cbcm_enc.c.
6630 *) DES quad checksum was broken on big-endian architectures. Fixed.
6633 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6634 Win32 test batch file so it (might) work again. The Win32 test batch file
6635 is horrible: I feel ill....
6638 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6639 in e_os.h. Audit of header files to check ANSI and non ANSI
6640 sections: 10 functions were absent from non ANSI section and not exported
6641 from Windows DLLs. Fixed up libeay.num for new functions.
6644 *) Make `openssl version' output lines consistent.
6645 [Ralf S. Engelschall]
6647 *) Fix Win32 symbol export lists for BIO functions: Added
6648 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6649 to ms/libeay{16,32}.def.
6650 [Ralf S. Engelschall]
6652 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6653 fine under Unix and passes some trivial tests I've now added. But the
6654 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6655 added to make sure no one expects that this stuff really works in the
6656 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6657 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6659 [Ralf S. Engelschall]
6661 *) Fix the generation of two part addresses in perl.
6662 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6664 *) Add config entry for Linux on MIPS.
6665 [John Tobey <jtobey@channel1.com>]
6667 *) Make links whenever Configure is run, unless we are on Windoze.
6670 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6671 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6675 *) Add a useful kludge to allow package maintainers to specify compiler and
6676 other platforms details on the command line without having to patch the
6677 Configure script everytime: One now can use ``perl Configure
6678 <id>:<details>'', i.e. platform ids are allowed to have details appended
6679 to them (seperated by colons). This is treated as there would be a static
6680 pre-configured entry in Configure's %table under key <id> with value
6681 <details> and ``perl Configure <id>'' is called. So, when you want to
6682 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6683 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6684 now, which overrides the FreeBSD-elf entry on-the-fly.
6685 [Ralf S. Engelschall]
6687 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6690 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6691 on the `perl Configure ...' command line. This way one can compile
6692 OpenSSL libraries with Position Independent Code (PIC) which is needed
6693 for linking it into DSOs.
6694 [Ralf S. Engelschall]
6696 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6700 *) Cleaned up the LICENSE document: The official contact for any license
6701 questions now is the OpenSSL core team under openssl-core@openssl.org.
6702 And add a paragraph about the dual-license situation to make sure people
6703 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6704 to the OpenSSL toolkit.
6705 [Ralf S. Engelschall]
6707 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6708 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6709 Additonally cleaned up the `make links' target: Remove unnecessary
6710 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6711 to speed processing and no longer clutter the display with confusing
6712 stuff. Instead only the actually done links are displayed.
6713 [Ralf S. Engelschall]
6715 *) Permit null encryption ciphersuites, used for authentication only. It used
6716 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6717 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6721 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6722 signed attributes when verifying signatures (this would break them),
6723 the detached data encoding was wrong and public keys obtained using
6724 X509_get_pubkey() weren't freed.
6727 *) Add text documentation for the BUFFER functions. Also added a work around
6728 to a Win95 console bug. This was triggered by the password read stuff: the
6729 last character typed gets carried over to the next fread(). If you were
6730 generating a new cert request using 'req' for example then the last
6731 character of the passphrase would be CR which would then enter the first
6735 *) Added the new `Includes OpenSSL Cryptography Software' button as
6736 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6737 button and can be used by applications based on OpenSSL to show the
6738 relationship to the OpenSSL project.
6739 [Ralf S. Engelschall]
6741 *) Remove confusing variables in function signatures in files
6742 ssl/ssl_lib.c and ssl/ssl.h.
6743 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6745 *) Don't install bss_file.c under PREFIX/include/
6746 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6748 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6749 functions that return function pointers and has support for NT specific
6750 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6751 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6752 unsigned to signed types: this was killing the Win32 compile.
6755 *) Add new certificate file to stack functions,
6756 SSL_add_dir_cert_subjects_to_stack() and
6757 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6758 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6759 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6760 This means that Apache-SSL and similar packages don't have to mess around
6761 to add as many CAs as they want to the preferred list.
6764 *) Experiment with doxygen documentation. Currently only partially applied to
6766 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6767 openssl.doxy as the configuration file.
6770 *) Get rid of remaining C++-style comments which strict C compilers hate.
6771 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6773 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6774 compiled in by default: it has problems with large keys.
6777 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6778 DH private keys and/or callback functions which directly correspond to
6779 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6780 is needed for applications which have to configure certificates on a
6781 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6783 For the RSA certificate situation is makes no difference, but
6784 for the DSA certificate situation this fixes the "no shared cipher"
6785 problem where the OpenSSL cipher selection procedure failed because the
6786 temporary keys were not overtaken from the context and the API provided
6787 no way to reconfigure them.
6788 The new functions now let applications reconfigure the stuff and they
6789 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6790 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6791 non-public-API function ssl_cert_instantiate() is used as a helper
6792 function and also to reduce code redundancy inside ssl_rsa.c.
6793 [Ralf S. Engelschall]
6795 *) Move s_server -dcert and -dkey options out of the undocumented feature
6796 area because they are useful for the DSA situation and should be
6797 recognized by the users.
6798 [Ralf S. Engelschall]
6800 *) Fix the cipher decision scheme for export ciphers: the export bits are
6801 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6802 SSL_EXP_MASK. So, the original variable has to be used instead of the
6803 already masked variable.
6804 [Richard Levitte <levitte@stacken.kth.se>]
6806 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6807 [Richard Levitte <levitte@stacken.kth.se>]
6809 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6810 from `int' to `unsigned int' because it's a length and initialized by
6811 EVP_DigestFinal() which expects an `unsigned int *'.
6812 [Richard Levitte <levitte@stacken.kth.se>]
6814 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6815 script. Instead use the usual Shell->Perl transition trick.
6816 [Ralf S. Engelschall]
6818 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6819 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6820 -noout -modulus' as it's already the case for `openssl rsa -noout
6821 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6822 currently the public key is printed (a decision which was already done by
6823 `openssl dsa -modulus' in the past) which serves a similar purpose.
6824 Additionally the NO_RSA no longer completely removes the whole -modulus
6825 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6827 [Ralf S. Engelschall]
6829 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6830 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6831 [Arne Ansper <arne@ats.cyber.ee>]
6833 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6834 to be added. Now both 'req' and 'ca' can use new objects defined in the
6838 *) Add cool BIO that does syslog (or event log on NT).
6839 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6841 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6842 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6843 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6844 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6847 *) Add preliminary config info for new extension code.
6850 *) Make RSA_NO_PADDING really use no padding.
6851 [Ulf Moeller <ulf@fitug.de>]
6853 *) Generate errors when private/public key check is done.
6856 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6857 for some CRL extensions and new objects added.
6860 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6861 key usage extension and fuller support for authority key id.
6864 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6865 padding method for RSA, which is recommended for new applications in PKCS
6866 #1 v2.0 (RFC 2437, October 1998).
6867 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6868 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6869 against Bleichbacher's attack on RSA.
6870 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6873 *) Updates to the new SSL compression code
6874 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6876 *) Fix so that the version number in the master secret, when passed
6877 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6878 (because the server will not accept higher), that the version number
6879 is 0x03,0x01, not 0x03,0x00
6880 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6882 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6883 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6884 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6887 *) Support for RAW extensions where an arbitrary extension can be
6888 created by including its DER encoding. See apps/openssl.cnf for
6892 *) Make sure latest Perl versions don't interpret some generated C array
6893 code as Perl array code in the crypto/err/err_genc.pl script.
6894 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6896 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6897 not many people have the assembler. Various Win32 compilation fixes and
6898 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6902 *) Modify configure script 'Configure' to automatically create crypto/date.h
6903 file under Win32 and also build pem.h from pem.org. New script
6904 util/mkfiles.pl to create the MINFO file on environments that can't do a
6905 'make files': perl util/mkfiles.pl >MINFO should work.
6908 *) Major rework of DES function declarations, in the pursuit of correctness
6909 and purity. As a result, many evil casts evaporated, and some weirdness,
6910 too. You may find this causes warnings in your code. Zapping your evil
6911 casts will probably fix them. Mostly.
6914 *) Fix for a typo in asn1.h. Bug fix to object creation script
6915 obj_dat.pl. It considered a zero in an object definition to mean
6916 "end of object": none of the objects in objects.h have any zeros
6917 so it wasn't spotted.
6918 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
6920 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
6921 Masking (CBCM). In the absence of test vectors, the best I have been able
6922 to do is check that the decrypt undoes the encrypt, so far. Send me test
6923 vectors if you have them.
6926 *) Correct calculation of key length for export ciphers (too much space was
6927 allocated for null ciphers). This has not been tested!
6930 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
6931 message is now correct (it understands "crypto" and "ssl" on its
6932 command line). There is also now an "update" option. This will update
6933 the util/ssleay.num and util/libeay.num files with any new functions.
6935 perl util/mkdef.pl crypto ssl update
6936 it will update them.
6939 *) Overhauled the Perl interface (perl/*):
6940 - ported BN stuff to OpenSSL's different BN library
6941 - made the perl/ source tree CVS-aware
6942 - renamed the package from SSLeay to OpenSSL (the files still contain
6943 their history because I've copied them in the repository)
6944 - removed obsolete files (the test scripts will be replaced
6945 by better Test::Harness variants in the future)
6946 [Ralf S. Engelschall]
6948 *) First cut for a very conservative source tree cleanup:
6949 1. merge various obsolete readme texts into doc/ssleay.txt
6950 where we collect the old documents and readme texts.
6951 2. remove the first part of files where I'm already sure that we no
6952 longer need them because of three reasons: either they are just temporary
6953 files which were left by Eric or they are preserved original files where
6954 I've verified that the diff is also available in the CVS via "cvs diff
6955 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
6956 the crypto/md/ stuff).
6957 [Ralf S. Engelschall]
6959 *) More extension code. Incomplete support for subject and issuer alt
6960 name, issuer and authority key id. Change the i2v function parameters
6961 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
6962 what that's for :-) Fix to ASN1 macro which messed up
6963 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
6966 *) Preliminary support for ENUMERATED type. This is largely copied from the
6970 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
6971 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6973 *) Make sure `make rehash' target really finds the `openssl' program.
6974 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6976 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
6977 like to hear about it if this slows down other processors.
6980 *) Add CygWin32 platform information to Configure script.
6981 [Alan Batie <batie@aahz.jf.intel.com>]
6983 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
6984 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
6986 *) New program nseq to manipulate netscape certificate sequences
6989 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
6993 *) Fixes to BN code. Previously the default was to define BN_RECURSION
6994 but the BN code had some problems that would cause failures when
6995 doing certificate verification and some other functions.
6996 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6998 *) Add ASN1 and PEM code to support netscape certificate sequences.
7001 *) Add ASN1 and PEM code to support netscape certificate sequences.
7004 *) Add several PKIX and private extended key usage OIDs.
7007 *) Modify the 'ca' program to handle the new extension code. Modify
7008 openssl.cnf for new extension format, add comments.
7011 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7012 and add a sample to openssl.cnf so req -x509 now adds appropriate
7016 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7017 error code, add initial support to X509_print() and x509 application.
7020 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7021 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7022 stuff is currently isolated and isn't even compiled yet.
7025 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7026 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7027 Removed the versions check from X509 routines when loading extensions:
7028 this allows certain broken certificates that don't set the version
7029 properly to be processed.
7032 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7033 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7034 can still be regenerated with "make depend".
7037 *) Spelling mistake in C version of CAST-128.
7038 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7040 *) Changes to the error generation code. The perl script err-code.pl
7041 now reads in the old error codes and retains the old numbers, only
7042 adding new ones if necessary. It also only changes the .err files if new
7043 codes are added. The makefiles have been modified to only insert errors
7044 when needed (to avoid needlessly modifying header files). This is done
7045 by only inserting errors if the .err file is newer than the auto generated
7046 C file. To rebuild all the error codes from scratch (the old behaviour)
7047 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7048 or delete all the .err files.
7051 *) CAST-128 was incorrectly implemented for short keys. The C version has
7052 been fixed, but is untested. The assembler versions are also fixed, but
7053 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7054 to regenerate it if needed.
7055 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7056 Hagino <itojun@kame.net>]
7058 *) File was opened incorrectly in randfile.c.
7059 [Ulf Möller <ulf@fitug.de>]
7061 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7062 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7063 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7064 al: it's just almost always a UTCTime. Note this patch adds new error
7065 codes so do a "make errors" if there are problems.
7068 *) Correct Linux 1 recognition in config.
7069 [Ulf Möller <ulf@fitug.de>]
7071 *) Remove pointless MD5 hash when using DSA keys in ca.
7072 [Anonymous <nobody@replay.com>]
7074 *) Generate an error if given an empty string as a cert directory. Also
7075 generate an error if handed NULL (previously returned 0 to indicate an
7076 error, but didn't set one).
7077 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7079 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7082 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7083 parameters. This was causing a warning which killed off the Win32 compile.
7086 *) Remove C++ style comments from crypto/bn/bn_local.h.
7087 [Neil Costigan <neil.costigan@celocom.com>]
7089 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7090 based on a text string, looking up short and long names and finally
7091 "dot" format. The "dot" format stuff didn't work. Added new function
7092 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7093 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7094 OID is not part of the table.
7097 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7098 X509_LOOKUP_by_alias().
7101 *) Sort openssl functions by name.
7104 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7105 encryption from sample DSA keys (in case anyone is interested the password
7109 *) Make _all_ *_free functions accept a NULL pointer.
7110 [Frans Heymans <fheymans@isaserver.be>]
7112 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7114 [Anonymous <nobody@replay.com>]
7116 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7117 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7119 *) Don't blow it for numeric -newkey arguments to apps/req.
7120 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7122 *) Temp key "for export" tests were wrong in s3_srvr.c.
7123 [Anonymous <nobody@replay.com>]
7125 *) Add prototype for temp key callback functions
7126 SSL_CTX_set_tmp_{rsa,dh}_callback().
7129 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7130 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7133 *) X509_name_add_entry() freed the wrong thing after an error.
7134 [Arne Ansper <arne@ats.cyber.ee>]
7136 *) rsa_eay.c would attempt to free a NULL context.
7137 [Arne Ansper <arne@ats.cyber.ee>]
7139 *) BIO_s_socket() had a broken should_retry() on Windoze.
7140 [Arne Ansper <arne@ats.cyber.ee>]
7142 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7143 [Arne Ansper <arne@ats.cyber.ee>]
7145 *) Make sure the already existing X509_STORE->depth variable is initialized
7146 in X509_STORE_new(), but document the fact that this variable is still
7147 unused in the certificate verification process.
7148 [Ralf S. Engelschall]
7150 *) Fix the various library and apps files to free up pkeys obtained from
7151 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7154 *) Fix reference counting in X509_PUBKEY_get(). This makes
7155 demos/maurice/example2.c work, amongst others, probably.
7156 [Steve Henson and Ben Laurie]
7158 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7159 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7160 are no longer created. This way we have a single and consistent command
7161 line interface `openssl <command>', similar to `cvs <command>'.
7162 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7164 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7165 BIT STRING wrapper always have zero unused bits.
7168 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7171 *) Make the top-level INSTALL documentation easier to understand.
7174 *) Makefiles updated to exit if an error occurs in a sub-directory
7175 make (including if user presses ^C) [Paul Sutton]
7177 *) Make Montgomery context stuff explicit in RSA data structure.
7180 *) Fix build order of pem and err to allow for generated pem.h.
7183 *) Fix renumbering bug in X509_NAME_delete_entry().
7186 *) Enhanced the err-ins.pl script so it makes the error library number
7187 global and can add a library name. This is needed for external ASN1 and
7188 other error libraries.
7191 *) Fixed sk_insert which never worked properly.
7194 *) Fix ASN1 macros so they can handle indefinite length construted
7195 EXPLICIT tags. Some non standard certificates use these: they can now
7199 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7200 into a single doc/ssleay.txt bundle. This way the information is still
7201 preserved but no longer messes up this directory. Now it's new room for
7202 the new set of documenation files.
7203 [Ralf S. Engelschall]
7205 *) SETs were incorrectly DER encoded. This was a major pain, because they
7206 shared code with SEQUENCEs, which aren't coded the same. This means that
7207 almost everything to do with SETs or SEQUENCEs has either changed name or
7208 number of arguments.
7209 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7211 *) Fix test data to work with the above.
7214 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7215 was already fixed by Eric for 0.9.1 it seems.
7216 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7218 *) Autodetect FreeBSD3.
7221 *) Fix various bugs in Configure. This affects the following platforms:
7225 unixware-2.0-pentium
7229 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7230 before they are needed.
7233 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7237 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7239 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7240 changed SSLeay to OpenSSL in version strings.
7241 [Ralf S. Engelschall]
7243 *) Some fixups to the top-level documents.
7246 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7247 because the symlink to include/ was missing.
7248 [Ralf S. Engelschall]
7250 *) Incorporated the popular no-RSA/DSA-only patches
7251 which allow to compile a RSA-free SSLeay.
7252 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7254 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7255 when "ssleay" is still not found.
7256 [Ralf S. Engelschall]
7258 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7259 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7261 *) Updated the README file.
7262 [Ralf S. Engelschall]
7264 *) Added various .cvsignore files in the CVS repository subdirs
7265 to make a "cvs update" really silent.
7266 [Ralf S. Engelschall]
7268 *) Recompiled the error-definition header files and added
7269 missing symbols to the Win32 linker tables.
7270 [Ralf S. Engelschall]
7272 *) Cleaned up the top-level documents;
7273 o new files: CHANGES and LICENSE
7274 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7275 o merged COPYRIGHT into LICENSE
7276 o removed obsolete TODO file
7277 o renamed MICROSOFT to INSTALL.W32
7278 [Ralf S. Engelschall]
7280 *) Removed dummy files from the 0.9.1b source tree:
7281 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7282 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7283 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7284 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7285 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7286 [Ralf S. Engelschall]
7288 *) Added various platform portability fixes.
7291 *) The Genesis of the OpenSSL rpject:
7292 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7293 Young and Tim J. Hudson created while they were working for C2Net until
7295 [The OpenSSL Project]
7298 Changes between 0.9.0b and 0.9.1b [not released]
7300 *) Updated a few CA certificates under certs/
7303 *) Changed some BIGNUM api stuff.
7306 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7307 DGUX x86, Linux Alpha, etc.
7310 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7311 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7315 *) Add -strparse option to asn1pars program which parses nested
7317 [Dr Stephen Henson <shenson@bigfoot.com>]
7319 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7322 *) DSA fix for "ca" program.
7325 *) Added "-genkey" option to "dsaparam" program.
7328 *) Added RIPE MD160 (rmd160) message digest.
7331 *) Added -a (all) option to "ssleay version" command.
7334 *) Added PLATFORM define which is the id given to Configure.
7337 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7340 *) Extended the ASN.1 parser routines.
7343 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7346 *) Added a BN_CTX to the BN library.
7349 *) Fixed the weak key values in DES library
7352 *) Changed API in EVP library for cipher aliases.
7355 *) Added support for RC2/64bit cipher.
7358 *) Converted the lhash library to the crypto/mem.c functions.
7361 *) Added more recognized ASN.1 object ids.
7364 *) Added more RSA padding checks for SSL/TLS.
7367 *) Added BIO proxy/filter functionality.
7370 *) Added extra_certs to SSL_CTX which can be used
7371 send extra CA certificates to the client in the CA cert chain sending
7372 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7375 *) Now Fortezza is denied in the authentication phase because
7376 this is key exchange mechanism is not supported by SSLeay at all.
7379 *) Additional PKCS1 checks.
7382 *) Support the string "TLSv1" for all TLS v1 ciphers.
7385 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7386 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7389 *) Fixed a few memory leaks.
7392 *) Fixed various code and comment typos.
7395 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7396 bytes sent in the client random.
7397 [Edward Bishop <ebishop@spyglass.com>]