5 Changes between 0.9.7h and 0.9.8 [xx XXX xxxx]
7 *) Add attribute functions to EVP_PKEY structure. Modify
8 PKCS12_create() to recognize a CSP name attribute and
9 use it. Make -CSP option work again in pkcs12 utility.
12 *) Add new functionality to the bn blinding code:
13 - automatic re-creation of the BN_BLINDING parameters after
14 a fixed number of uses (currently 32)
15 - add new function for parameter creation
16 - introduce flags to control the update behaviour of the
17 BN_BLINDING parameters
18 - hide BN_BLINDING structure
19 Add a second BN_BLINDING slot to the RSA structure to improve
20 performance when a single RSA object is shared among several
24 *) Add support for DTLS.
25 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
27 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
28 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
31 *) Remove buggy and incompletet DH cert support from
32 ssl/ssl_rsa.c and ssl/s3_both.c
35 *) Use SHA-1 instead of MD5 as the default digest algorithm for
36 the apps/openssl applications.
39 *) Compile clean with "-Wall -Wmissing-prototypes
40 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
41 DEBUG_SAFESTACK must also be set.
44 *) Change ./Configure so that certain algorithms can be disabled by default.
45 The new counterpiece to "no-xxx" is "enable-xxx".
47 The patented RC5 and MDC2 algorithms will now be disabled unless
48 "enable-rc5" and "enable-mdc2", respectively, are specified.
50 (IDEA remains enabled despite being patented. This is because IDEA
51 is frequently required for interoperability, and there is no license
52 fee for non-commercial use. As before, "no-idea" can be used to
53 avoid this algorithm.)
57 *) Add processing of proxy certificates (see RFC 3820). This work was
58 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
59 EGEE (Enabling Grids for E-science in Europe).
62 *) RC4 performance overhaul on modern architectures/implementations, such
63 as Intel P4, IA-64 and AMD64.
66 *) New utility extract-section.pl. This can be used specify an alternative
67 section number in a pod file instead of having to treat each file as
68 a separate case in Makefile. This can be done by adding two lines to the
71 =for comment openssl_section:XXX
73 The blank line is mandatory.
77 *) New arguments -certform, -keyform and -pass for s_client and s_server
78 to allow alternative format key and certificate files and passphrase
82 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
83 update associated structures and add various utility functions.
85 Add new policy related verify parameters, include policy checking in
86 standard verify code. Enhance 'smime' application with extra parameters
87 to support policy checking and print out.
90 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
91 Nehemiah processors. These extensions support AES encryption in hardware
92 as well as RNG (though RNG support is currently disabled).
93 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
95 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
98 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
99 [Andy Polyakov and a number of other people]
101 *) Improved PowerPC platform support. Most notably BIGNUM assembler
102 implementation contributed by IBM.
103 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
105 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
106 exponent rather than 'unsigned long'. There is a corresponding change to
107 the new 'rsa_keygen' element of the RSA_METHOD structure.
108 [Jelte Jansen, Geoff Thorpe]
110 *) Functionality for creating the initial serial number file is now
111 moved from CA.pl to the 'ca' utility with a new option -create_serial.
113 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
114 number file to 1, which is bound to cause problems. To avoid
115 the problems while respecting compatibility between different 0.9.7
116 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
117 CA.pl for serial number initialization. With the new release 0.9.8,
118 we can fix the problem directly in the 'ca' utility.)
121 *) Reduced header interdepencies by declaring more opaque objects in
122 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
123 give fewer recursive includes, which could break lazy source code - so
124 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
125 developers should define this symbol when building and using openssl to
126 ensure they track the recommended behaviour, interfaces, [etc], but
127 backwards-compatible behaviour prevails when this isn't defined.
130 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
133 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
134 This will generate a random key of the appropriate length based on the
135 cipher context. The EVP_CIPHER can provide its own random key generation
136 routine to support keys of a specific form. This is used in the des and
137 3des routines to generate a key of the correct parity. Update S/MIME
138 code to use new functions and hence generate correct parity DES keys.
139 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
140 valid (weak or incorrect parity).
143 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
144 as looking them up. This is useful when the verified structure may contain
145 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
146 present unless the new PKCS7_NO_CRL flag is asserted.
149 *) Extend ASN1 oid configuration module. It now additionally accepts the
152 shortName = some long name, 1.2.3.4
155 *) Reimplemented the BN_CTX implementation. There is now no more static
156 limitation on the number of variables it can handle nor the depth of the
157 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
158 information can now expand as required, and rather than having a single
159 static array of bignums, BN_CTX now uses a linked-list of such arrays
160 allowing it to expand on demand whilst maintaining the usefulness of
164 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
165 to allow all RSA operations to function using a single BN_CTX.
168 *) Preliminary support for certificate policy evaluation and checking. This
169 is initially intended to pass the tests outlined in "Conformance Testing
170 of Relying Party Client Certificate Path Processing Logic" v1.07.
173 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
174 remained unused and not that useful. A variety of other little bignum
175 tweaks and fixes have also been made continuing on from the audit (see
179 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
180 associated ASN1, EVP and SSL functions and old ASN1 macros.
183 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
184 and this should never fail. So the return value from the use of
185 BN_set_word() (which can fail due to needless expansion) is now deprecated;
186 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
189 *) BN_CTX_get() should return zero-valued bignums, providing the same
190 initialised value as BN_new().
191 [Geoff Thorpe, suggested by Ulf Möller]
193 *) Support for inhibitAnyPolicy certificate extension.
196 *) An audit of the BIGNUM code is underway, for which debugging code is
197 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
198 is considered valid when processing BIGNUMs, and causes execution to
199 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
200 further steps are taken to deliberately pollute unused data in BIGNUM
201 structures to try and expose faulty code further on. For now, openssl will
202 (in its default mode of operation) continue to tolerate the inconsistent
203 forms that it has tolerated in the past, but authors and packagers should
204 consider trying openssl and their own applications when compiled with
205 these debugging symbols defined. It will help highlight potential bugs in
206 their own code, and will improve the test coverage for OpenSSL itself. At
207 some point, these tighter rules will become openssl's default to improve
208 maintainability, though the assert()s and other overheads will remain only
209 in debugging configurations. See bn.h for more details.
210 [Geoff Thorpe, Nils Larsch, Ulf Möller]
212 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
213 that can only be obtained through BN_CTX_new() (which implicitly
214 initialises it). The presence of this function only made it possible
215 to overwrite an existing structure (and cause memory leaks).
218 *) Because of the callback-based approach for implementing LHASH as a
219 template type, lh_insert() adds opaque objects to hash-tables and
220 lh_doall() or lh_doall_arg() are typically used with a destructor callback
221 to clean up those corresponding objects before destroying the hash table
222 (and losing the object pointers). So some over-zealous constifications in
223 LHASH have been relaxed so that lh_insert() does not take (nor store) the
224 objects as "const" and the lh_doall[_arg] callback wrappers are not
225 prototyped to have "const" restrictions on the object pointers they are
226 given (and so aren't required to cast them away any more).
229 *) The tmdiff.h API was so ugly and minimal that our own timing utility
230 (speed) prefers to use its own implementation. The two implementations
231 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
232 its object type properly exposed (MS_TM) instead of casting to/from "char
233 *". This may still change yet if someone realises MS_TM and "ms_time_***"
234 aren't necessarily the greatest nomenclatures - but this is what was used
235 internally to the implementation so I've used that for now.
238 *) Ensure that deprecated functions do not get compiled when
239 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
240 the self-tests were still using deprecated key-generation functions so
241 these have been updated also.
244 *) Reorganise PKCS#7 code to separate the digest location functionality
245 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
246 New function PKCS7_set_digest() to set the digest type for PKCS#7
247 digestedData type. Add additional code to correctly generate the
248 digestedData type and add support for this type in PKCS7 initialization
252 *) New function PKCS7_set0_type_other() this initializes a PKCS7
253 structure of type "other".
256 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
257 sure the loop does correctly stop and breaking ("division by zero")
258 modulus operations are not performed. The (pre-generated) prime
259 table crypto/bn/bn_prime.h was already correct, but it could not be
260 re-generated on some platforms because of the "division by zero"
261 situation in the script.
262 [Ralf S. Engelschall]
264 *) Update support for ECC-based TLS ciphersuites according to
265 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
266 SHA-1 now is only used for "small" curves (where the
267 representation of a field element takes up to 24 bytes); for
268 larger curves, the field element resulting from ECDH is directly
269 used as premaster secret.
270 [Douglas Stebila (Sun Microsystems Laboratories)]
272 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
273 curve secp160r1 to the tests.
274 [Douglas Stebila (Sun Microsystems Laboratories)]
276 *) Add the possibility to load symbols globally with DSO.
277 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
279 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
280 control of the error stack.
283 *) Add support for STORE in ENGINE.
286 *) Add the STORE type. The intention is to provide a common interface
287 to certificate and key stores, be they simple file-based stores, or
288 HSM-type store, or LDAP stores, or...
289 NOTE: The code is currently UNTESTED and isn't really used anywhere.
292 *) Add a generic structure called OPENSSL_ITEM. This can be used to
293 pass a list of arguments to any function as well as provide a way
294 for a function to pass data back to the caller.
297 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
298 works like BUF_strdup() but can be used to duplicate a portion of
299 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
303 *) Add the function sk_find_ex() which works like sk_find(), but will
304 return an index to an element even if an exact match couldn't be
305 found. The index is guaranteed to point at the element where the
306 searched-for key would be inserted to preserve sorting order.
309 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
310 takes an extra flags argument for optional functionality. Currently,
311 the following flags are defined:
313 OBJ_BSEARCH_VALUE_ON_NOMATCH
314 This one gets OBJ_bsearch_ex() to return a pointer to the first
315 element where the comparing function returns a negative or zero
318 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
319 This one gets OBJ_bsearch_ex() to return a pointer to the first
320 element where the comparing function returns zero. This is useful
321 if there are more than one element where the comparing function
325 *) Make it possible to create self-signed certificates with 'openssl ca'
326 in such a way that the self-signed certificate becomes part of the
327 CA database and uses the same mechanisms for serial number generation
328 as all other certificate signing. The new flag '-selfsign' enables
329 this functionality. Adapt CA.sh and CA.pl.in.
332 *) Add functionality to check the public key of a certificate request
333 against a given private. This is useful to check that a certificate
334 request can be signed by that key (self-signing).
337 *) Make it possible to have multiple active certificates with the same
338 subject in the CA index file. This is done only if the keyword
339 'unique_subject' is set to 'no' in the main CA section (default
340 if 'CA_default') of the configuration file. The value is saved
341 with the database itself in a separate index attribute file,
342 named like the index file with '.attr' appended to the name.
345 *) Generate muti valued AVAs using '+' notation in config files for
349 *) Support for nameConstraints certificate extension.
352 *) Support for policyConstraints certificate extension.
355 *) Support for policyMappings certificate extension.
358 *) Make sure the default DSA_METHOD implementation only uses its
359 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
360 and change its own handlers to be NULL so as to remove unnecessary
361 indirection. This lets alternative implementations fallback to the
362 default implementation more easily.
365 *) Support for directoryName in GeneralName related extensions
369 *) Make it possible to link applications using Makefile.shared.
370 Make that possible even when linking against static libraries!
373 *) Support for single pass processing for S/MIME signing. This now
374 means that S/MIME signing can be done from a pipe, in addition
375 cleartext signing (multipart/signed type) is effectively streaming
376 and the signed data does not need to be all held in memory.
378 This is done with a new flag PKCS7_STREAM. When this flag is set
379 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
380 is done after the data is output (and digests calculated) in
384 *) Add full support for -rpath/-R, both in shared libraries and
385 applications, at least on the platforms where it's known how
389 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
390 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
391 will now compute a table of multiples of the generator that
392 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
393 faster (notably in the case of a single point multiplication,
395 [Nils Larsch, Bodo Moeller]
397 *) IPv6 support for certificate extensions. The various extensions
398 which use the IP:a.b.c.d can now take IPv6 addresses using the
399 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
403 *) Added an ENGINE that implements RSA by performing private key
404 exponentiations with the GMP library. The conversions to and from
405 GMP's mpz_t format aren't optimised nor are any montgomery forms
406 cached, and on x86 it appears OpenSSL's own performance has caught up.
407 However there are likely to be other architectures where GMP could
408 provide a boost. This ENGINE is not built in by default, but it can be
409 specified at Configure time and should be accompanied by the necessary
410 linker additions, eg;
411 ./config -DOPENSSL_USE_GMP -lgmp
414 *) "openssl engine" will not display ENGINE/DSO load failure errors when
415 testing availability of engines with "-t" - the old behaviour is
416 produced by increasing the feature's verbosity with "-tt".
419 *) ECDSA routines: under certain error conditions uninitialized BN objects
420 could be freed. Solution: make sure initialization is performed early
421 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
425 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
426 and DH_METHOD (eg. by ENGINE implementations) to override the normal
427 software implementations. For DSA and DH, parameter generation can
428 also be overriden by providing the appropriate method callbacks.
431 *) Change the "progress" mechanism used in key-generation and
432 primality testing to functions that take a new BN_GENCB pointer in
433 place of callback/argument pairs. The new API functions have "_ex"
434 postfixes and the older functions are reimplemented as wrappers for
435 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
436 declarations of the old functions to help (graceful) attempts to
437 migrate to the new functions. Also, the new key-generation API
438 functions operate on a caller-supplied key-structure and return
439 success/failure rather than returning a key or NULL - this is to
440 help make "keygen" another member function of RSA_METHOD etc.
442 Example for using the new callback interface:
444 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
448 BN_GENCB_set(&my_cb, my_callback, my_arg);
450 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
451 /* For the meaning of a, b in calls to my_callback(), see the
452 * documentation of the function that calls the callback.
453 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
454 * my_callback should return 1 if it wants BN_is_prime_ex()
455 * to continue, or 0 to stop.
460 *) Change the ZLIB compression method to be stateful, and make it
461 available to TLS with the number defined in
462 draft-ietf-tls-compression-04.txt.
465 *) Add the ASN.1 structures and functions for CertificatePair, which
466 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
468 CertificatePair ::= SEQUENCE {
469 forward [0] Certificate OPTIONAL,
470 reverse [1] Certificate OPTIONAL,
471 -- at least one of the pair shall be present -- }
473 Also implement the PEM functions to read and write certificate
474 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
476 This needed to be defined, mostly for the sake of the LDAP
477 attribute crossCertificatePair, but may prove useful elsewhere as
481 *) Make it possible to inhibit symlinking of shared libraries in
482 Makefile.shared, for Cygwin's sake.
485 *) Extend the BIGNUM API by creating a function
486 void BN_set_negative(BIGNUM *a, int neg);
487 and a macro that behave like
488 int BN_is_negative(const BIGNUM *a);
490 to avoid the need to access 'a->neg' directly in applications.
493 *) Implement fast modular reduction for pseudo-Mersenne primes
494 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
495 EC_GROUP_new_curve_GFp() will now automatically use this
497 [Nils Larsch <nla@trustcenter.de>]
499 *) Add new lock type (CRYPTO_LOCK_BN).
502 *) Change the ENGINE framework to automatically load engines
503 dynamically from specific directories unless they could be
504 found to already be built in or loaded. Move all the
505 current engines except for the cryptodev one to a new
507 The engines in engines/ are built as shared libraries if
508 the "shared" options was given to ./Configure or ./config.
509 Otherwise, they are inserted in libcrypto.a.
510 /usr/local/ssl/engines is the default directory for dynamic
511 engines, but that can be overriden at configure time through
512 the usual use of --prefix and/or --openssldir, and at run
513 time with the environment variable OPENSSL_ENGINES.
514 [Geoff Thorpe and Richard Levitte]
516 *) Add Makefile.shared, a helper makefile to build shared
517 libraries. Addapt Makefile.org.
520 *) Add version info to Win32 DLLs.
521 [Peter 'Luna' Runestig" <peter@runestig.com>]
523 *) Add new 'medium level' PKCS#12 API. Certificates and keys
524 can be added using this API to created arbitrary PKCS#12
525 files while avoiding the low level API.
527 New options to PKCS12_create(), key or cert can be NULL and
528 will then be omitted from the output file. The encryption
529 algorithm NIDs can be set to -1 for no encryption, the mac
530 iteration count can be set to 0 to omit the mac.
532 Enhance pkcs12 utility by making the -nokeys and -nocerts
533 options work when creating a PKCS#12 file. New option -nomac
534 to omit the mac, NONE can be set for an encryption algorithm.
535 New code is modified to use the enhanced PKCS12_create()
536 instead of the low level API.
539 *) Extend ASN1 encoder to support indefinite length constructed
540 encoding. This can output sequences tags and octet strings in
541 this form. Modify pk7_asn1.c to support indefinite length
542 encoding. This is experimental and needs additional code to
543 be useful, such as an ASN1 bio and some enhanced streaming
546 Extend template encode functionality so that tagging is passed
547 down to the template encoder.
550 *) Let 'openssl req' fail if an argument to '-newkey' is not
551 recognized instead of using RSA as a default.
554 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
555 As these are not official, they are not included in "ALL";
556 the "ECCdraft" ciphersuite group alias can be used to select them.
557 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
559 *) Add ECDH engine support.
560 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
562 *) Add ECDH in new directory crypto/ecdh/.
563 [Douglas Stebila (Sun Microsystems Laboratories)]
565 *) Let BN_rand_range() abort with an error after 100 iterations
566 without success (which indicates a broken PRNG).
569 *) Change BN_mod_sqrt() so that it verifies that the input value
570 is really the square of the return value. (Previously,
571 BN_mod_sqrt would show GIGO behaviour.)
574 *) Add named elliptic curves over binary fields from X9.62, SECG,
575 and WAP/WTLS; add OIDs that were still missing.
577 [Sheueling Chang Shantz and Douglas Stebila
578 (Sun Microsystems Laboratories)]
580 *) Extend the EC library for elliptic curves over binary fields
581 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
584 EC_GF2m_simple_method
588 EC_GROUP_new_curve_GF2m
589 EC_GROUP_set_curve_GF2m
590 EC_GROUP_get_curve_GF2m
591 EC_POINT_set_affine_coordinates_GF2m
592 EC_POINT_get_affine_coordinates_GF2m
593 EC_POINT_set_compressed_coordinates_GF2m
595 Point compression for binary fields is disabled by default for
596 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
599 As binary polynomials are represented as BIGNUMs, various members
600 of the EC_GROUP and EC_POINT data structures can be shared
601 between the implementations for prime fields and binary fields;
602 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
603 are essentially identical to their ..._GFp counterparts.
604 (For simplicity, the '..._GFp' prefix has been dropped from
605 various internal method names.)
607 An internal 'field_div' method (similar to 'field_mul' and
608 'field_sqr') has been added; this is used only for binary fields.
610 [Sheueling Chang Shantz and Douglas Stebila
611 (Sun Microsystems Laboratories)]
613 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
614 through methods ('mul', 'precompute_mult').
616 The generic implementations (now internally called 'ec_wNAF_mul'
617 and 'ec_wNAF_precomputed_mult') remain the default if these
618 methods are undefined.
620 [Sheueling Chang Shantz and Douglas Stebila
621 (Sun Microsystems Laboratories)]
623 *) New function EC_GROUP_get_degree, which is defined through
624 EC_METHOD. For curves over prime fields, this returns the bit
625 length of the modulus.
627 [Sheueling Chang Shantz and Douglas Stebila
628 (Sun Microsystems Laboratories)]
630 *) New functions EC_GROUP_dup, EC_POINT_dup.
631 (These simply call ..._new and ..._copy).
633 [Sheueling Chang Shantz and Douglas Stebila
634 (Sun Microsystems Laboratories)]
636 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
637 Polynomials are represented as BIGNUMs (where the sign bit is not
638 used) in the following functions [macros]:
641 BN_GF2m_sub [= BN_GF2m_add]
642 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
643 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
644 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
646 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
647 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
648 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
649 BN_GF2m_cmp [= BN_ucmp]
651 (Note that only the 'mod' functions are actually for fields GF(2^m).
652 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
654 For some functions, an the irreducible polynomial defining a
655 field can be given as an 'unsigned int[]' with strictly
656 decreasing elements giving the indices of those bits that are set;
657 i.e., p[] represents the polynomial
658 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
660 p[0] > p[1] > ... > p[k] = 0.
661 This applies to the following functions:
666 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
667 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
670 BN_GF2m_mod_solve_quad_arr
674 Conversion can be performed by the following functions:
679 bntest.c has additional tests for binary polynomial arithmetic.
681 Two implementations for BN_GF2m_mod_div() are available.
682 The default algorithm simply uses BN_GF2m_mod_inv() and
683 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
684 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
685 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
687 [Sheueling Chang Shantz and Douglas Stebila
688 (Sun Microsystems Laboratories)]
690 *) Add new error code 'ERR_R_DISABLED' that can be used when some
691 functionality is disabled at compile-time.
692 [Douglas Stebila <douglas.stebila@sun.com>]
694 *) Change default behaviour of 'openssl asn1parse' so that more
695 information is visible when viewing, e.g., a certificate:
697 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
698 mode the content of non-printable OCTET STRINGs is output in a
699 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
700 avoid the appearance of a printable string.
701 [Nils Larsch <nla@trustcenter.de>]
703 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
705 EC_GROUP_set_asn1_flag()
706 EC_GROUP_get_asn1_flag()
707 EC_GROUP_set_point_conversion_form()
708 EC_GROUP_get_point_conversion_form()
709 These control ASN1 encoding details:
710 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
711 has been set to OPENSSL_EC_NAMED_CURVE.
712 - Points are encoded in uncompressed form by default; options for
713 asn1_for are as for point2oct, namely
714 POINT_CONVERSION_COMPRESSED
715 POINT_CONVERSION_UNCOMPRESSED
716 POINT_CONVERSION_HYBRID
718 Also add 'seed' and 'seed_len' members to EC_GROUP with access
722 EC_GROUP_get_seed_len()
723 This is used only for ASN1 purposes (so far).
724 [Nils Larsch <nla@trustcenter.de>]
726 *) Add 'field_type' member to EC_METHOD, which holds the NID
727 of the appropriate field type OID. The new function
728 EC_METHOD_get_field_type() returns this value.
729 [Nils Larsch <nla@trustcenter.de>]
736 providing useful interfaces to EC_POINT_point2oct() and
737 EC_POINT_oct2point().
738 [Nils Larsch <nla@trustcenter.de>]
740 *) Change internals of the EC library so that the functions
741 EC_GROUP_set_generator()
742 EC_GROUP_get_generator()
744 EC_GROUP_get_cofactor()
745 are implemented directly in crypto/ec/ec_lib.c and not dispatched
746 to methods, which would lead to unnecessary code duplication when
747 adding different types of curves.
748 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
750 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
751 arithmetic, and such that modified wNAFs are generated
752 (which avoid length expansion in many cases).
755 *) Add a function EC_GROUP_check_discriminant() (defined via
756 EC_METHOD) that verifies that the curve discriminant is non-zero.
758 Add a function EC_GROUP_check() that makes some sanity tests
759 on a EC_GROUP, its generator and order. This includes
760 EC_GROUP_check_discriminant().
761 [Nils Larsch <nla@trustcenter.de>]
763 *) Add ECDSA in new directory crypto/ecdsa/.
765 Add applications 'openssl ecparam' and 'openssl ecdsa'
766 (these are based on 'openssl dsaparam' and 'openssl dsa').
768 ECDSA support is also included in various other files across the
769 library. Most notably,
770 - 'openssl req' now has a '-newkey ecdsa:file' option;
771 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
772 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
773 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
774 them suitable for ECDSA where domain parameters must be
775 extracted before the specific public key;
776 - ECDSA engine support has been added.
777 [Nils Larsch <nla@trustcenter.de>]
779 *) Include some named elliptic curves, and add OIDs from X9.62,
780 SECG, and WAP/WTLS. Each curve can be obtained from the new
782 EC_GROUP_new_by_curve_name(),
783 and the list of available named curves can be obtained with
784 EC_get_builtin_curves().
785 Also add a 'curve_name' member to EC_GROUP objects, which can be
787 EC_GROUP_set_curve_name()
788 EC_GROUP_get_curve_name()
789 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
791 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
792 was actually never needed) and in BN_mul(). The removal in BN_mul()
793 required a small change in bn_mul_part_recursive() and the addition
794 of the functions bn_cmp_part_words(), bn_sub_part_words() and
795 bn_add_part_words(), which do the same thing as bn_cmp_words(),
796 bn_sub_words() and bn_add_words() except they take arrays with
800 Changes between 0.9.7g and 0.9.7h [XX xxx XXXX]
802 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
803 the exponentiation using a fixed-length exponent. (Otherwise,
804 the information leaked through timing could expose the secret key
805 after many signatures; cf. Bleichenbacher's attack on DSA with
809 *) Make a new fixed-window mod_exp implementation the default for
810 RSA, DSA, and DH private-key operations so that the sequence of
811 squares and multiplies and the memory access pattern are
812 independent of the particular secret key. This will mitigate
813 cache-timing and potential related attacks.
815 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
816 and this is automatically used by BN_mod_exp_mont() if the new flag
817 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
818 will use this BN flag for private exponents unless the flag
819 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
820 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
822 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
824 *) Change the client implementation for SSLv23_method() and
825 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
826 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
827 (Previously, the SSL 2.0 backwards compatible Client Hello
828 message format would be used even with SSL_OP_NO_SSLv2.)
831 *) Add support for smime-type MIME parameter in S/MIME messages which some
835 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
836 a threadsafe manner. Modify rsa code to use new function and add calls
837 to dsa and dh code (which had race conditions before).
840 *) Include the fixed error library code in the C error file definitions
841 instead of fixing them up at runtime. This keeps the error code
845 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
847 *) Fixes for newer kerberos headers. NB: the casts are needed because
848 the 'length' field is signed on one version and unsigned on another
849 with no (?) obvious way to tell the difference, without these VC++
850 complains. Also the "definition" of FAR (blank) is no longer included
851 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
852 some needed definitions.
855 *) Undo Cygwin change.
858 *) Added support for proxy certificates according to RFC 3820.
859 Because they may be a security thread to unaware applications,
860 they must be explicitely allowed in run-time. See
861 docs/HOWTO/proxy_certificates.txt for further information.
864 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
866 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
867 server and client random values. Previously
868 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
869 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
871 This change has negligible security impact because:
873 1. Server and client random values still have 24 bytes of pseudo random
876 2. Server and client random values are sent in the clear in the initial
879 3. The master secret is derived using the premaster secret (48 bytes in
880 size for static RSA ciphersuites) as well as client server and random
883 The OpenSSL team would like to thank the UK NISCC for bringing this issue
886 [Stephen Henson, reported by UK NISCC]
888 *) Use Windows randomness collection on Cygwin.
891 *) Fix hang in EGD/PRNGD query when communication socket is closed
892 prematurely by EGD/PRNGD.
893 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
895 *) Prompt for pass phrases when appropriate for PKCS12 input format.
898 *) Back-port of selected performance improvements from development
899 branch, as well as improved support for PowerPC platforms.
902 *) Add lots of checks for memory allocation failure, error codes to indicate
903 failure and freeing up memory if a failure occurs.
904 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
906 *) Add new -passin argument to dgst.
909 *) Perform some character comparisons of different types in X509_NAME_cmp:
910 this is needed for some certificates that reencode DNs into UTF8Strings
911 (in violation of RFC3280) and can't or wont issue name rollover
915 *) Make an explicit check during certificate validation to see that
916 the CA setting in each certificate on the chain is correct. As a
917 side effect always do the following basic checks on extensions,
918 not just when there's an associated purpose to the check:
920 - if there is an unhandled critical extension (unless the user
921 has chosen to ignore this fault)
922 - if the path length has been exceeded (if one is set at all)
923 - that certain extensions fit the associated purpose (if one has
927 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
929 *) Avoid a race condition when CRLs are checked in a multi threaded
930 environment. This would happen due to the reordering of the revoked
931 entries during signature checking and serial number lookup. Now the
932 encoding is cached and the serial number sort performed under a lock.
933 Add new STACK function sk_is_sorted().
936 *) Add Delta CRL to the extension code.
939 *) Various fixes to s3_pkt.c so alerts are sent properly.
940 [David Holmes <d.holmes@f5.com>]
942 *) Reduce the chances of duplicate issuer name and serial numbers (in
943 violation of RFC3280) using the OpenSSL certificate creation utilities.
944 This is done by creating a random 64 bit value for the initial serial
945 number when a serial number file is created or when a self signed
946 certificate is created using 'openssl req -x509'. The initial serial
947 number file is created using 'openssl x509 -next_serial' in CA.pl
948 rather than being initialized to 1.
951 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
953 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
954 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
955 [Joe Orton, Steve Henson]
957 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
959 [Joe Orton, Steve Henson]
961 *) Make it possible to have multiple active certificates with the same
962 subject in the CA index file. This is done only if the keyword
963 'unique_subject' is set to 'no' in the main CA section (default
964 if 'CA_default') of the configuration file. The value is saved
965 with the database itself in a separate index attribute file,
966 named like the index file with '.attr' appended to the name.
969 *) X509 verify fixes. Disable broken certificate workarounds when
970 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
971 keyUsage extension present. Don't accept CRLs with unhandled critical
972 extensions: since verify currently doesn't process CRL extensions this
973 rejects a CRL with *any* critical extensions. Add new verify error codes
977 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
978 A clarification of RFC2560 will require the use of OCTET STRINGs and
979 some implementations cannot handle the current raw format. Since OpenSSL
980 copies and compares OCSP nonces as opaque blobs without any attempt at
981 parsing them this should not create any compatibility issues.
984 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
985 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
986 this HMAC (and other) operations are several times slower than OpenSSL
990 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
991 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
993 *) Use the correct content when signing type "other".
996 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
998 *) Fix various bugs revealed by running the NISCC test suite:
1000 Stop out of bounds reads in the ASN1 code when presented with
1001 invalid tags (CAN-2003-0543 and CAN-2003-0544).
1003 Free up ASN1_TYPE correctly if ANY type is invalid (CAN-2003-0545).
1005 If verify callback ignores invalid public key errors don't try to check
1006 certificate signature with the NULL public key.
1010 *) New -ignore_err option in ocsp application to stop the server
1011 exiting on the first error in a request.
1014 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1015 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1019 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1020 extra data after the compression methods not only for TLS 1.0
1021 but also for SSL 3.0 (as required by the specification).
1022 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1024 *) Change X509_certificate_type() to mark the key as exported/exportable
1025 when it's 512 *bits* long, not 512 bytes.
1028 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1029 blocks during encryption.
1032 *) Various fixes to base64 BIO and non blocking I/O. On write
1033 flushes were not handled properly if the BIO retried. On read
1034 data was not being buffered properly and had various logic bugs.
1035 This also affects blocking I/O when the data being decoded is a
1039 *) Various S/MIME bugfixes and compatibility changes:
1040 output correct application/pkcs7 MIME type if
1041 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1042 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1043 of files as .eml work). Correctly handle very long lines in MIME
1047 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1049 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1050 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1051 a protocol version number mismatch like a decryption error
1052 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1055 *) Turn on RSA blinding by default in the default implementation
1056 to avoid a timing attack. Applications that don't want it can call
1057 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1058 They would be ill-advised to do so in most cases.
1059 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1061 *) Change RSA blinding code so that it works when the PRNG is not
1062 seeded (in this case, the secret RSA exponent is abused as
1063 an unpredictable seed -- if it is not unpredictable, there
1064 is no point in blinding anyway). Make RSA blinding thread-safe
1065 by remembering the creator's thread ID in rsa->blinding and
1066 having all other threads use local one-time blinding factors
1067 (this requires more computation than sharing rsa->blinding, but
1068 avoids excessive locking; and if an RSA object is not shared
1069 between threads, blinding will still be very fast).
1072 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1073 ENGINE as defaults for all supported algorithms irrespective of
1074 the 'flags' parameter. 'flags' is now honoured, so applications
1075 should make sure they are passing it correctly.
1078 *) Target "mingw" now allows native Windows code to be generated in
1079 the Cygwin environment as well as with the MinGW compiler.
1082 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1084 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1085 via timing by performing a MAC computation even if incorrrect
1086 block cipher padding has been found. This is a countermeasure
1087 against active attacks where the attacker has to distinguish
1088 between bad padding and a MAC verification error. (CAN-2003-0078)
1090 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1091 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1092 Martin Vuagnoux (EPFL, Ilion)]
1094 *) Make the no-err option work as intended. The intention with no-err
1095 is not to have the whole error stack handling routines removed from
1096 libcrypto, it's only intended to remove all the function name and
1097 reason texts, thereby removing some of the footprint that may not
1098 be interesting if those errors aren't displayed anyway.
1100 NOTE: it's still possible for any application or module to have it's
1101 own set of error texts inserted. The routines are there, just not
1102 used by default when no-err is given.
1105 *) Add support for FreeBSD on IA64.
1106 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1108 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1109 Kerberos function mit_des_cbc_cksum(). Before this change,
1110 the value returned by DES_cbc_cksum() was like the one from
1111 mit_des_cbc_cksum(), except the bytes were swapped.
1112 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1114 *) Allow an application to disable the automatic SSL chain building.
1115 Before this a rather primitive chain build was always performed in
1116 ssl3_output_cert_chain(): an application had no way to send the
1117 correct chain if the automatic operation produced an incorrect result.
1119 Now the chain builder is disabled if either:
1121 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1123 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1125 The reasoning behind this is that an application would not want the
1126 auto chain building to take place if extra chain certificates are
1127 present and it might also want a means of sending no additional
1128 certificates (for example the chain has two certificates and the
1132 *) Add the possibility to build without the ENGINE framework.
1133 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1135 *) Under Win32 gmtime() can return NULL: check return value in
1136 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1139 *) DSA routines: under certain error conditions uninitialized BN objects
1140 could be freed. Solution: make sure initialization is performed early
1141 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1142 Nils Larsch <nla@trustcenter.de> via PR#459)
1145 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1146 checked on reconnect on the client side, therefore session resumption
1147 could still fail with a "ssl session id is different" error. This
1148 behaviour is masked when SSL_OP_ALL is used due to
1149 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1150 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1151 followup to PR #377.
1154 *) IA-32 assembler support enhancements: unified ELF targets, support
1155 for SCO/Caldera platforms, fix for Cygwin shared build.
1158 *) Add support for FreeBSD on sparc64. As a consequence, support for
1159 FreeBSD on non-x86 processors is separate from x86 processors on
1160 the config script, much like the NetBSD support.
1161 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1163 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1165 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1168 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1169 code (06) was taken as the first octet of the session ID and the last
1170 octet was ignored consequently. As a result SSLv2 client side session
1171 caching could not have worked due to the session ID mismatch between
1173 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1177 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1178 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1182 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1183 seems that in spite of existing for more than a year, many application
1184 author have done nothing to provide the necessary callbacks, which
1185 means that this particular engine will not work properly anywhere.
1186 This is a very unfortunate situation which forces us, in the name
1187 of usability, to give the hw_ncipher.c a static lock, which is part
1189 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1190 appear in 0.9.8 or later. We EXPECT application authors to have
1191 dealt properly with this when 0.9.8 is released (unless we actually
1192 make such changes in the libcrypto locking code that changes will
1193 have to be made anyway).
1196 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1197 octets have been read, EOF or an error occurs. Without this change
1198 some truncated ASN1 structures will not produce an error.
1201 *) Disable Heimdal support, since it hasn't been fully implemented.
1202 Still give the possibility to force the use of Heimdal, but with
1203 warnings and a request that patches get sent to openssl-dev.
1206 *) Add the VC-CE target, introduce the WINCE sysname, and add
1207 INSTALL.WCE and appropriate conditionals to make it build.
1208 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1210 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1211 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1212 edit numbers of the version.
1213 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1215 *) Introduce safe string copy and catenation functions
1216 (BUF_strlcpy() and BUF_strlcat()).
1217 [Ben Laurie (CHATS) and Richard Levitte]
1219 *) Avoid using fixed-size buffers for one-line DNs.
1220 [Ben Laurie (CHATS)]
1222 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1223 resizing buffers containing secrets, and use where appropriate.
1224 [Ben Laurie (CHATS)]
1226 *) Avoid using fixed size buffers for configuration file location.
1227 [Ben Laurie (CHATS)]
1229 *) Avoid filename truncation for various CA files.
1230 [Ben Laurie (CHATS)]
1232 *) Use sizeof in preference to magic numbers.
1233 [Ben Laurie (CHATS)]
1235 *) Avoid filename truncation in cert requests.
1236 [Ben Laurie (CHATS)]
1238 *) Add assertions to check for (supposedly impossible) buffer
1240 [Ben Laurie (CHATS)]
1242 *) Don't cache truncated DNS entries in the local cache (this could
1243 potentially lead to a spoofing attack).
1244 [Ben Laurie (CHATS)]
1246 *) Fix various buffers to be large enough for hex/decimal
1247 representations in a platform independent manner.
1248 [Ben Laurie (CHATS)]
1250 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1251 resizing buffers containing secrets, and use where appropriate.
1252 [Ben Laurie (CHATS)]
1254 *) Add BIO_indent() to avoid much slightly worrying code to do
1256 [Ben Laurie (CHATS)]
1258 *) Convert sprintf()/BIO_puts() to BIO_printf().
1259 [Ben Laurie (CHATS)]
1261 *) buffer_gets() could terminate with the buffer only half
1263 [Ben Laurie (CHATS)]
1265 *) Add assertions to prevent user-supplied crypto functions from
1266 overflowing internal buffers by having large block sizes, etc.
1267 [Ben Laurie (CHATS)]
1269 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1271 [Ben Laurie (CHATS)]
1273 *) Eliminate unused copy of key in RC4.
1274 [Ben Laurie (CHATS)]
1276 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1277 [Ben Laurie (CHATS)]
1279 *) Fix off-by-one error in EGD path.
1280 [Ben Laurie (CHATS)]
1282 *) If RANDFILE path is too long, ignore instead of truncating.
1283 [Ben Laurie (CHATS)]
1285 *) Eliminate unused and incorrectly sized X.509 structure
1287 [Ben Laurie (CHATS)]
1289 *) Eliminate unused and dangerous function knumber().
1290 [Ben Laurie (CHATS)]
1292 *) Eliminate unused and dangerous structure, KSSL_ERR.
1293 [Ben Laurie (CHATS)]
1295 *) Protect against overlong session ID context length in an encoded
1296 session object. Since these are local, this does not appear to be
1298 [Ben Laurie (CHATS)]
1300 *) Change from security patch (see 0.9.6e below) that did not affect
1301 the 0.9.6 release series:
1303 Remote buffer overflow in SSL3 protocol - an attacker could
1304 supply an oversized master key in Kerberos-enabled versions.
1306 [Ben Laurie (CHATS)]
1308 *) Change the SSL kerb5 codes to match RFC 2712.
1311 *) Make -nameopt work fully for req and add -reqopt switch.
1312 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1314 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1315 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1317 *) Make sure tests can be performed even if the corresponding algorithms
1318 have been removed entirely. This was also the last step to make
1319 OpenSSL compilable with DJGPP under all reasonable conditions.
1320 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1322 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1323 to allow version independent disabling of normally unselected ciphers,
1324 which may be activated as a side-effect of selecting a single cipher.
1326 (E.g., cipher list string "RSA" enables ciphersuites that are left
1327 out of "ALL" because they do not provide symmetric encryption.
1328 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1329 [Lutz Jaenicke, Bodo Moeller]
1331 *) Add appropriate support for separate platform-dependent build
1332 directories. The recommended way to make a platform-dependent
1333 build directory is the following (tested on Linux), maybe with
1336 # Place yourself outside of the OpenSSL source tree. In
1337 # this example, the environment variable OPENSSL_SOURCE
1338 # is assumed to contain the absolute OpenSSL source directory.
1339 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1340 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1341 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1342 mkdir -p `dirname $F`
1343 ln -s $OPENSSL_SOURCE/$F $F
1346 To be absolutely sure not to disturb the source tree, a "make clean"
1347 is a good thing. If it isn't successfull, don't worry about it,
1348 it probably means the source directory is very clean.
1351 *) Make sure any ENGINE control commands make local copies of string
1352 pointers passed to them whenever necessary. Otherwise it is possible
1353 the caller may have overwritten (or deallocated) the original string
1354 data when a later ENGINE operation tries to use the stored values.
1355 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1357 *) Improve diagnostics in file reading and command-line digests.
1358 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1360 *) Add AES modes CFB and OFB to the object database. Correct an
1361 error in AES-CFB decryption.
1364 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1365 allows existing EVP_CIPHER_CTX structures to be reused after
1366 calling EVP_*Final(). This behaviour is used by encryption
1367 BIOs and some applications. This has the side effect that
1368 applications must explicitly clean up cipher contexts with
1369 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1372 *) Check the values of dna and dnb in bn_mul_recursive before calling
1373 bn_mul_comba (a non zero value means the a or b arrays do not contain
1374 n2 elements) and fallback to bn_mul_normal if either is not zero.
1377 *) Fix escaping of non-ASCII characters when using the -subj option
1378 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1381 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1382 form for "surname", serialNumber has no short form.
1383 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1384 therefore remove "mail" short name for "internet 7".
1385 The OID for unique identifiers in X509 certificates is
1386 x500UniqueIdentifier, not uniqueIdentifier.
1387 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1390 *) Add an "init" command to the ENGINE config module and auto initialize
1391 ENGINEs. Without any "init" command the ENGINE will be initialized
1392 after all ctrl commands have been executed on it. If init=1 the
1393 ENGINE is initailized at that point (ctrls before that point are run
1394 on the uninitialized ENGINE and after on the initialized one). If
1395 init=0 then the ENGINE will not be iniatialized at all.
1398 *) Fix the 'app_verify_callback' interface so that the user-defined
1399 argument is actually passed to the callback: In the
1400 SSL_CTX_set_cert_verify_callback() prototype, the callback
1401 declaration has been changed from
1404 int (*cb)(X509_STORE_CTX *,void *);
1405 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1406 i=s->ctx->app_verify_callback(&ctx)
1407 has been changed into
1408 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1410 To update applications using SSL_CTX_set_cert_verify_callback(),
1411 a dummy argument can be added to their callback functions.
1412 [D. K. Smetters <smetters@parc.xerox.com>]
1414 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1415 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1417 *) Add and OPENSSL_LOAD_CONF define which will cause
1418 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1419 This allows older applications to transparently support certain
1420 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1421 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1422 load the config file and OPENSSL_add_all_algorithms_conf() which will
1423 always load it have also been added.
1426 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1427 Adjust NIDs and EVP layer.
1428 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1430 *) Config modules support in openssl utility.
1432 Most commands now load modules from the config file,
1433 though in a few (such as version) this isn't done
1434 because it couldn't be used for anything.
1436 In the case of ca and req the config file used is
1437 the same as the utility itself: that is the -config
1438 command line option can be used to specify an
1442 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1443 use "openssl_conf" if filename is NULL use default openssl config file.
1446 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1447 config section name. Add a new flag to tolerate a missing config file
1448 and move code to CONF_modules_load_file().
1451 *) Support for crypto accelerator cards from Accelerated Encryption
1452 Processing, www.aep.ie. (Use engine 'aep')
1453 The support was copied from 0.9.6c [engine] and adapted/corrected
1454 to work with the new engine framework.
1455 [AEP Inc. and Richard Levitte]
1457 *) Support for SureWare crypto accelerator cards from Baltimore
1458 Technologies. (Use engine 'sureware')
1459 The support was copied from 0.9.6c [engine] and adapted
1460 to work with the new engine framework.
1463 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1464 make the newer ENGINE framework commands for the CHIL engine work.
1465 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1467 *) Make it possible to produce shared libraries on ReliantUNIX.
1468 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1470 *) Add the configuration target debug-linux-ppro.
1471 Make 'openssl rsa' use the general key loading routines
1472 implemented in apps.c, and make those routines able to
1473 handle the key format FORMAT_NETSCAPE and the variant
1475 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1477 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1478 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1480 *) Add -keyform to rsautl, and document -engine.
1481 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1483 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1484 BIO_R_NO_SUCH_FILE error code rather than the generic
1485 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1488 *) Add new functions
1490 ERR_peek_last_error_line
1491 ERR_peek_last_error_line_data.
1492 These are similar to
1495 ERR_peek_error_line_data,
1496 but report on the latest error recorded rather than the first one
1497 still in the error queue.
1498 [Ben Laurie, Bodo Moeller]
1500 *) default_algorithms option in ENGINE config module. This allows things
1502 default_algorithms = ALL
1503 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1506 *) Prelminary ENGINE config module.
1509 *) New experimental application configuration code.
1512 *) Change the AES code to follow the same name structure as all other
1513 symmetric ciphers, and behave the same way. Move everything to
1514 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1515 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1517 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1518 [Ben Laurie and Theo de Raadt]
1520 *) Add option to output public keys in req command.
1521 [Massimiliano Pala madwolf@openca.org]
1523 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1524 (up to about 10% better than before for P-192 and P-224).
1527 *) New functions/macros
1529 SSL_CTX_set_msg_callback(ctx, cb)
1530 SSL_CTX_set_msg_callback_arg(ctx, arg)
1531 SSL_set_msg_callback(ssl, cb)
1532 SSL_set_msg_callback_arg(ssl, arg)
1534 to request calling a callback function
1536 void cb(int write_p, int version, int content_type,
1537 const void *buf, size_t len, SSL *ssl, void *arg)
1539 whenever a protocol message has been completely received
1540 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1541 protocol version according to which the SSL library interprets
1542 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1543 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1544 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1545 specification (change_cipher_spec(20), alert(21), handshake(22)).
1546 'buf' and 'len' point to the actual message, 'ssl' to the
1547 SSL object, and 'arg' is the application-defined value set by
1548 SSL[_CTX]_set_msg_callback_arg().
1550 'openssl s_client' and 'openssl s_server' have new '-msg' options
1551 to enable a callback that displays all protocol messages.
1554 *) Change the shared library support so shared libraries are built as
1555 soon as the corresponding static library is finished, and thereby get
1556 openssl and the test programs linked against the shared library.
1557 This still only happens when the keyword "shard" has been given to
1558 the configuration scripts.
1560 NOTE: shared library support is still an experimental thing, and
1561 backward binary compatibility is still not guaranteed.
1562 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1564 *) Add support for Subject Information Access extension.
1565 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1567 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1568 additional bytes when new memory had to be allocated, not just
1569 when reusing an existing buffer.
1572 *) New command line and configuration option 'utf8' for the req command.
1573 This allows field values to be specified as UTF8 strings.
1576 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1577 runs for the former and machine-readable output for the latter.
1580 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1581 of the e-mail address in the DN (i.e., it will go into a certificate
1582 extension only). The new configuration file option 'email_in_dn = no'
1583 has the same effect.
1584 [Massimiliano Pala madwolf@openca.org]
1586 *) Change all functions with names starting with des_ to be starting
1587 with DES_ instead. Add wrappers that are compatible with libdes,
1588 but are named _ossl_old_des_*. Finally, add macros that map the
1589 des_* symbols to the corresponding _ossl_old_des_* if libdes
1590 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1591 desired, the des_* symbols will be mapped to DES_*, with one
1594 Since we provide two compatibility mappings, the user needs to
1595 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1596 compatibility is desired. The default (i.e., when that macro
1597 isn't defined) is OpenSSL 0.9.6c compatibility.
1599 There are also macros that enable and disable the support of old
1600 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1601 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1602 are defined, the default will apply: to support the old des routines.
1604 In either case, one must include openssl/des.h to get the correct
1605 definitions. Do not try to just include openssl/des_old.h, that
1608 NOTE: This is a major break of an old API into a new one. Software
1609 authors are encouraged to switch to the DES_ style functions. Some
1610 time in the future, des_old.h and the libdes compatibility functions
1611 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1612 default), and then completely removed.
1615 *) Test for certificates which contain unsupported critical extensions.
1616 If such a certificate is found during a verify operation it is
1617 rejected by default: this behaviour can be overridden by either
1618 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1619 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1620 X509_supported_extension() has also been added which returns 1 if a
1621 particular extension is supported.
1624 *) Modify the behaviour of EVP cipher functions in similar way to digests
1625 to retain compatibility with existing code.
1628 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1629 compatibility with existing code. In particular the 'ctx' parameter does
1630 not have to be to be initialized before the call to EVP_DigestInit() and
1631 it is tidied up after a call to EVP_DigestFinal(). New function
1632 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1633 EVP_MD_CTX_copy() changed to not require the destination to be
1634 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1635 requires the destination to be valid.
1637 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1638 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1641 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1642 so that complete 'Handshake' protocol structures are kept in memory
1643 instead of overwriting 'msg_type' and 'length' with 'body' data.
1646 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1647 [Massimo Santin via Richard Levitte]
1649 *) Major restructuring to the underlying ENGINE code. This includes
1650 reduction of linker bloat, separation of pure "ENGINE" manipulation
1651 (initialisation, etc) from functionality dealing with implementations
1652 of specific crypto iterfaces. This change also introduces integrated
1653 support for symmetric ciphers and digest implementations - so ENGINEs
1654 can now accelerate these by providing EVP_CIPHER and EVP_MD
1655 implementations of their own. This is detailed in crypto/engine/README
1656 as it couldn't be adequately described here. However, there are a few
1657 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1658 were changed in the original introduction of ENGINE code have now
1659 reverted back - the hooking from this code to ENGINE is now a good
1660 deal more passive and at run-time, operations deal directly with
1661 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1662 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1663 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1664 they were not being used by the framework as there is no concept of a
1665 BIGNUM_METHOD and they could not be generalised to the new
1666 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1667 ENGINE_cpy() has been removed as it cannot be consistently defined in
1671 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1674 *) Change mkdef.pl to sort symbols that get the same entry number,
1675 and make sure the automatically generated functions ERR_load_*
1676 become part of libeay.num as well.
1679 *) New function SSL_renegotiate_pending(). This returns true once
1680 renegotiation has been requested (either SSL_renegotiate() call
1681 or HelloRequest/ClientHello receveived from the peer) and becomes
1682 false once a handshake has been completed.
1683 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1684 sends a HelloRequest, but does not ensure that a handshake takes
1685 place. SSL_renegotiate_pending() is useful for checking if the
1686 client has followed the request.)
1689 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1690 By default, clients may request session resumption even during
1691 renegotiation (if session ID contexts permit); with this option,
1692 session resumption is possible only in the first handshake.
1694 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1695 more bits available for options that should not be part of
1696 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1699 *) Add some demos for certificate and certificate request creation.
1702 *) Make maximum certificate chain size accepted from the peer application
1703 settable (SSL*_get/set_max_cert_list()), as proposed by
1704 "Douglas E. Engert" <deengert@anl.gov>.
1707 *) Add support for shared libraries for Unixware-7
1708 (Boyd Lynn Gerber <gerberb@zenez.com>).
1711 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1712 be done prior to destruction. Use this to unload error strings from
1713 ENGINEs that load their own error strings. NB: This adds two new API
1714 functions to "get" and "set" this destroy handler in an ENGINE.
1717 *) Alter all existing ENGINE implementations (except "openssl" and
1718 "openbsd") to dynamically instantiate their own error strings. This
1719 makes them more flexible to be built both as statically-linked ENGINEs
1720 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1721 Also, add stub code to each that makes building them as self-contained
1722 shared-libraries easier (see README.ENGINE).
1725 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1726 implementations into applications that are completely implemented in
1727 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1728 commands that can be used to configure what shared-library to load and
1729 to control aspects of the way it is handled. Also, made an update to
1730 the README.ENGINE file that brings its information up-to-date and
1731 provides some information and instructions on the "dynamic" ENGINE
1732 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1735 *) Make it possible to unload ranges of ERR strings with a new
1736 "ERR_unload_strings" function.
1739 *) Add a copy() function to EVP_MD.
1742 *) Make EVP_MD routines take a context pointer instead of just the
1743 md_data void pointer.
1746 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1747 that the digest can only process a single chunk of data
1748 (typically because it is provided by a piece of
1749 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1750 is only going to provide a single chunk of data, and hence the
1751 framework needn't accumulate the data for oneshot drivers.
1754 *) As with "ERR", make it possible to replace the underlying "ex_data"
1755 functions. This change also alters the storage and management of global
1756 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1757 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1758 index counters. The API functions that use this state have been changed
1759 to take a "class_index" rather than pointers to the class's local STACK
1760 and counter, and there is now an API function to dynamically create new
1761 classes. This centralisation allows us to (a) plug a lot of the
1762 thread-safety problems that existed, and (b) makes it possible to clean
1763 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
1764 such data would previously have always leaked in application code and
1765 workarounds were in place to make the memory debugging turn a blind eye
1766 to it. Application code that doesn't use this new function will still
1767 leak as before, but their memory debugging output will announce it now
1768 rather than letting it slide.
1770 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
1771 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
1772 has a return value to indicate success or failure.
1775 *) Make it possible to replace the underlying "ERR" functions such that the
1776 global state (2 LHASH tables and 2 locks) is only used by the "default"
1777 implementation. This change also adds two functions to "get" and "set"
1778 the implementation prior to it being automatically set the first time
1779 any other ERR function takes place. Ie. an application can call "get",
1780 pass the return value to a module it has just loaded, and that module
1781 can call its own "set" function using that value. This means the
1782 module's "ERR" operations will use (and modify) the error state in the
1783 application and not in its own statically linked copy of OpenSSL code.
1786 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
1787 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
1788 the operation, and provides a more encapsulated way for external code
1789 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
1790 to use these functions rather than manually incrementing the counts.
1792 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
1795 *) Add EVP test program.
1798 *) Add symmetric cipher support to ENGINE. Expect the API to change!
1801 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
1802 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
1803 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
1804 These allow a CRL to be built without having to access X509_CRL fields
1805 directly. Modify 'ca' application to use new functions.
1808 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
1809 bug workarounds. Rollback attack detection is a security feature.
1810 The problem will only arise on OpenSSL servers when TLSv1 is not
1811 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
1812 Software authors not wanting to support TLSv1 will have special reasons
1813 for their choice and can explicitly enable this option.
1814 [Bodo Moeller, Lutz Jaenicke]
1816 *) Rationalise EVP so it can be extended: don't include a union of
1817 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
1818 (similar to those existing for EVP_CIPHER_CTX).
1823 EVP_MD_CTX_init(&md); /* new function call */
1824 EVP_DigestInit(&md, EVP_sha1());
1825 EVP_DigestUpdate(&md, in, len);
1826 EVP_DigestFinal(&md, out, NULL);
1827 EVP_MD_CTX_cleanup(&md); /* new function call */
1831 *) Make DES key schedule conform to the usual scheme, as well as
1832 correcting its structure. This means that calls to DES functions
1833 now have to pass a pointer to a des_key_schedule instead of a
1834 plain des_key_schedule (which was actually always a pointer
1837 des_key_schedule ks;
1839 des_set_key_checked(..., &ks);
1840 des_ncbc_encrypt(..., &ks, ...);
1842 (Note that a later change renames 'des_...' into 'DES_...'.)
1845 *) Initial reduction of linker bloat: the use of some functions, such as
1846 PEM causes large amounts of unused functions to be linked in due to
1847 poor organisation. For example pem_all.c contains every PEM function
1848 which has a knock on effect of linking in large amounts of (unused)
1849 ASN1 code. Grouping together similar functions and splitting unrelated
1850 functions prevents this.
1853 *) Cleanup of EVP macros.
1856 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
1857 correct _ecb suffix.
1860 *) Add initial OCSP responder support to ocsp application. The
1861 revocation information is handled using the text based index
1862 use by the ca application. The responder can either handle
1863 requests generated internally, supplied in files (for example
1864 via a CGI script) or using an internal minimal server.
1867 *) Add configuration choices to get zlib compression for TLS.
1870 *) Changes to Kerberos SSL for RFC 2712 compliance:
1871 1. Implemented real KerberosWrapper, instead of just using
1872 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
1873 2. Implemented optional authenticator field of KerberosWrapper.
1875 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
1876 and authenticator structs; see crypto/krb5/.
1878 Generalized Kerberos calls to support multiple Kerberos libraries.
1879 [Vern Staats <staatsvr@asc.hpc.mil>,
1880 Jeffrey Altman <jaltman@columbia.edu>
1881 via Richard Levitte]
1883 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
1884 already does with RSA. testdsa.h now has 'priv_key/pub_key'
1885 values for each of the key sizes rather than having just
1886 parameters (and 'speed' generating keys each time).
1889 *) Speed up EVP routines.
1892 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
1893 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
1894 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
1895 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
1897 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
1898 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
1899 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
1902 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
1904 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
1907 *) Added the OS2-EMX target.
1908 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
1910 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
1911 to support NCONF routines in extension code. New function CONF_set_nconf()
1912 to allow functions which take an NCONF to also handle the old LHASH
1913 structure: this means that the old CONF compatible routines can be
1914 retained (in particular wrt extensions) without having to duplicate the
1915 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
1918 *) Enhance the general user interface with mechanisms for inner control
1919 and with possibilities to have yes/no kind of prompts.
1922 *) Change all calls to low level digest routines in the library and
1923 applications to use EVP. Add missing calls to HMAC_cleanup() and
1924 don't assume HMAC_CTX can be copied using memcpy().
1925 [Verdon Walker <VWalker@novell.com>, Steve Henson]
1927 *) Add the possibility to control engines through control names but with
1928 arbitrary arguments instead of just a string.
1929 Change the key loaders to take a UI_METHOD instead of a callback
1930 function pointer. NOTE: this breaks binary compatibility with earlier
1931 versions of OpenSSL [engine].
1932 Adapt the nCipher code for these new conditions and add a card insertion
1936 *) Enhance the general user interface with mechanisms to better support
1937 dialog box interfaces, application-defined prompts, the possibility
1938 to use defaults (for example default passwords from somewhere else)
1939 and interrupts/cancellations.
1942 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
1943 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
1946 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
1947 tidy up some unnecessarily weird code in 'sk_new()').
1948 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
1950 *) Change the key loading routines for ENGINEs to use the same kind
1951 callback (pem_password_cb) as all other routines that need this
1955 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
1956 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
1957 than this minimum value is recommended.
1960 *) New random seeder for OpenVMS, using the system process statistics
1961 that are easily reachable.
1964 *) Windows apparently can't transparently handle global
1965 variables defined in DLLs. Initialisations such as:
1967 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1969 wont compile. This is used by the any applications that need to
1970 declare their own ASN1 modules. This was fixed by adding the option
1971 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1972 needed for static libraries under Win32.
1975 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1976 setting of purpose and trust fields. New X509_STORE trust and
1977 purpose functions and tidy up setting in other SSL functions.
1980 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1981 structure. These are inherited by X509_STORE_CTX when it is
1982 initialised. This allows various defaults to be set in the
1983 X509_STORE structure (such as flags for CRL checking and custom
1984 purpose or trust settings) for functions which only use X509_STORE_CTX
1985 internally such as S/MIME.
1987 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1988 trust settings if they are not set in X509_STORE. This allows X509_STORE
1989 purposes and trust (in S/MIME for example) to override any set by default.
1991 Add command line options for CRL checking to smime, s_client and s_server
1995 *) Initial CRL based revocation checking. If the CRL checking flag(s)
1996 are set then the CRL is looked up in the X509_STORE structure and
1997 its validity and signature checked, then if the certificate is found
1998 in the CRL the verify fails with a revoked error.
2000 Various new CRL related callbacks added to X509_STORE_CTX structure.
2002 Command line options added to 'verify' application to support this.
2004 This needs some additional work, such as being able to handle multiple
2005 CRLs with different times, extension based lookup (rather than just
2006 by subject name) and ultimately more complete V2 CRL extension
2010 *) Add a general user interface API (crypto/ui/). This is designed
2011 to replace things like des_read_password and friends (backward
2012 compatibility functions using this new API are provided).
2013 The purpose is to remove prompting functions from the DES code
2014 section as well as provide for prompting through dialog boxes in
2015 a window system and the like.
2018 *) Add "ex_data" support to ENGINE so implementations can add state at a
2019 per-structure level rather than having to store it globally.
2022 *) Make it possible for ENGINE structures to be copied when retrieved by
2023 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2024 This causes the "original" ENGINE structure to act like a template,
2025 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2026 operational state can be localised to each ENGINE structure, despite the
2027 fact they all share the same "methods". New ENGINE structures returned in
2028 this case have no functional references and the return value is the single
2029 structural reference. This matches the single structural reference returned
2030 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2034 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2035 needs to match any other type at all we need to manually clear the
2039 *) Changes to the "openssl engine" utility to include;
2040 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2041 about an ENGINE's available control commands.
2042 - executing control commands from command line arguments using the
2043 '-pre' and '-post' switches. '-post' is only used if '-t' is
2044 specified and the ENGINE is successfully initialised. The syntax for
2045 the individual commands are colon-separated, for example;
2046 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2049 *) New dynamic control command support for ENGINEs. ENGINEs can now
2050 declare their own commands (numbers), names (strings), descriptions,
2051 and input types for run-time discovery by calling applications. A
2052 subset of these commands are implicitly classed as "executable"
2053 depending on their input type, and only these can be invoked through
2054 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2055 can be based on user input, config files, etc). The distinction is
2056 that "executable" commands cannot return anything other than a boolean
2057 result and can only support numeric or string input, whereas some
2058 discoverable commands may only be for direct use through
2059 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2060 pointers, or other custom uses. The "executable" commands are to
2061 support parameterisations of ENGINE behaviour that can be
2062 unambiguously defined by ENGINEs and used consistently across any
2063 OpenSSL-based application. Commands have been added to all the
2064 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2065 control over shared-library paths without source code alterations.
2068 *) Changed all ENGINE implementations to dynamically allocate their
2069 ENGINEs rather than declaring them statically. Apart from this being
2070 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2071 this also allows the implementations to compile without using the
2072 internal engine_int.h header.
2075 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2076 'const' value. Any code that should be able to modify a RAND_METHOD
2077 should already have non-const pointers to it (ie. they should only
2078 modify their own ones).
2081 *) Made a variety of little tweaks to the ENGINE code.
2082 - "atalla" and "ubsec" string definitions were moved from header files
2083 to C code. "nuron" string definitions were placed in variables
2084 rather than hard-coded - allowing parameterisation of these values
2085 later on via ctrl() commands.
2086 - Removed unused "#if 0"'d code.
2087 - Fixed engine list iteration code so it uses ENGINE_free() to release
2088 structural references.
2089 - Constified the RAND_METHOD element of ENGINE structures.
2090 - Constified various get/set functions as appropriate and added
2091 missing functions (including a catch-all ENGINE_cpy that duplicates
2092 all ENGINE values onto a new ENGINE except reference counts/state).
2093 - Removed NULL parameter checks in get/set functions. Setting a method
2094 or function to NULL is a way of cancelling out a previously set
2095 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2096 and doesn't justify the extra error symbols and code.
2097 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2098 flags from engine_int.h to engine.h.
2099 - Changed prototypes for ENGINE handler functions (init(), finish(),
2100 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2103 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2104 to the algorithm using long division. The binary algorithm can be
2105 used only if the modulus is odd. On 32-bit systems, it is faster
2106 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2107 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2108 up to 450 bits. In 64-bit environments, the binary algorithm
2109 appears to be advantageous for much longer moduli; here we use it
2110 for moduli up to 2048 bits.
2113 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2114 could not support the combine flag in choice fields.
2117 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2118 extensions from a certificate request to the certificate.
2121 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2122 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2123 file: this allows the display of the certificate about to be
2124 signed to be customised, to allow certain fields to be included
2125 or excluded and extension details. The old system didn't display
2126 multicharacter strings properly, omitted fields not in the policy
2127 and couldn't display additional details such as extensions.
2130 *) Function EC_POINTs_mul for multiple scalar multiplication
2131 of an arbitrary number of elliptic curve points
2132 \sum scalars[i]*points[i],
2133 optionally including the generator defined for the EC_GROUP:
2134 scalar*generator + \sum scalars[i]*points[i].
2136 EC_POINT_mul is a simple wrapper function for the typical case
2137 that the point list has just one item (besides the optional
2141 *) First EC_METHODs for curves over GF(p):
2143 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2144 operations and provides various method functions that can also
2145 operate with faster implementations of modular arithmetic.
2147 EC_GFp_mont_method() reuses most functions that are part of
2148 EC_GFp_simple_method, but uses Montgomery arithmetic.
2150 [Bodo Moeller; point addition and point doubling
2151 implementation directly derived from source code provided by
2152 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2154 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2155 crypto/ec/ec_lib.c):
2157 Curves are EC_GROUP objects (with an optional group generator)
2158 based on EC_METHODs that are built into the library.
2160 Points are EC_POINT objects based on EC_GROUP objects.
2162 Most of the framework would be able to handle curves over arbitrary
2163 finite fields, but as there are no obvious types for fields other
2164 than GF(p), some functions are limited to that for now.
2167 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2168 that the file contains a complete HTTP response.
2171 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2172 change the def and num file printf format specifier from "%-40sXXX"
2173 to "%-39s XXX". The latter will always guarantee a space after the
2174 field while the former will cause them to run together if the field
2175 is 40 of more characters long.
2178 *) Constify the cipher and digest 'method' functions and structures
2179 and modify related functions to take constant EVP_MD and EVP_CIPHER
2183 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2184 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2187 *) Modify EVP_Digest*() routines so they now return values. Although the
2188 internal software routines can never fail additional hardware versions
2192 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2194 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2195 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2198 ERR_R_NESTED_ASN1_ERROR
2200 ERR_R_MISSING_ASN1_EOS
2201 were 4 .. 9, conflicting with
2202 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2204 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2205 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2207 Add new error code 'ERR_R_INTERNAL_ERROR'.
2210 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2214 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2215 sets the subject name for a new request or supersedes the
2216 subject name in a given request. Formats that can be parsed are
2217 'CN=Some Name, OU=myOU, C=IT'
2219 'CN=Some Name/OU=myOU/C=IT'.
2221 Add options '-batch' and '-verbose' to 'openssl req'.
2222 [Massimiliano Pala <madwolf@hackmasters.net>]
2224 *) Introduce the possibility to access global variables through
2225 functions on platform were that's the best way to handle exporting
2226 global variables in shared libraries. To enable this functionality,
2227 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2228 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2229 is normally done by Configure or something similar).
2231 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2232 in the source file (foo.c) like this:
2234 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2235 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2237 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2238 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2240 OPENSSL_DECLARE_GLOBAL(int,foo);
2241 #define foo OPENSSL_GLOBAL_REF(foo)
2242 OPENSSL_DECLARE_GLOBAL(double,bar);
2243 #define bar OPENSSL_GLOBAL_REF(bar)
2245 The #defines are very important, and therefore so is including the
2246 header file everywhere where the defined globals are used.
2248 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2249 of ASN.1 items, but that structure is a bit different.
2251 The largest change is in util/mkdef.pl which has been enhanced with
2252 better and easier to understand logic to choose which symbols should
2253 go into the Windows .def files as well as a number of fixes and code
2254 cleanup (among others, algorithm keywords are now sorted
2255 lexicographically to avoid constant rewrites).
2258 *) In BN_div() keep a copy of the sign of 'num' before writing the
2259 result to 'rm' because if rm==num the value will be overwritten
2260 and produce the wrong result if 'num' is negative: this caused
2261 problems with BN_mod() and BN_nnmod().
2264 *) Function OCSP_request_verify(). This checks the signature on an
2265 OCSP request and verifies the signer certificate. The signer
2266 certificate is just checked for a generic purpose and OCSP request
2270 *) Add OCSP_check_validity() function to check the validity of OCSP
2271 responses. OCSP responses are prepared in real time and may only
2272 be a few seconds old. Simply checking that the current time lies
2273 between thisUpdate and nextUpdate max reject otherwise valid responses
2274 caused by either OCSP responder or client clock inaccuracy. Instead
2275 we allow thisUpdate and nextUpdate to fall within a certain period of
2276 the current time. The age of the response can also optionally be
2277 checked. Two new options -validity_period and -status_age added to
2281 *) If signature or public key algorithm is unrecognized print out its
2282 OID rather that just UNKNOWN.
2285 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2286 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2287 ID to be generated from the issuer certificate alone which can then be
2288 passed to OCSP_id_issuer_cmp().
2291 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2292 ASN1 modules to export functions returning ASN1_ITEM pointers
2293 instead of the ASN1_ITEM structures themselves. This adds several
2294 new macros which allow the underlying ASN1 function/structure to
2295 be accessed transparently. As a result code should not use ASN1_ITEM
2296 references directly (such as &X509_it) but instead use the relevant
2297 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2298 use of the new ASN1 code on platforms where exporting structures
2299 is problematical (for example in shared libraries) but exporting
2300 functions returning pointers to structures is not.
2303 *) Add support for overriding the generation of SSL/TLS session IDs.
2304 These callbacks can be registered either in an SSL_CTX or per SSL.
2305 The purpose of this is to allow applications to control, if they wish,
2306 the arbitrary values chosen for use as session IDs, particularly as it
2307 can be useful for session caching in multiple-server environments. A
2308 command-line switch for testing this (and any client code that wishes
2309 to use such a feature) has been added to "s_server".
2310 [Geoff Thorpe, Lutz Jaenicke]
2312 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2313 of the form '#if defined(...) || defined(...) || ...' and
2314 '#if !defined(...) && !defined(...) && ...'. This also avoids
2315 the growing number of special cases it was previously handling.
2318 *) Make all configuration macros available for application by making
2319 sure they are available in opensslconf.h, by giving them names starting
2320 with "OPENSSL_" to avoid conflicts with other packages and by making
2321 sure e_os2.h will cover all platform-specific cases together with
2323 Additionally, it is now possible to define configuration/platform-
2324 specific names (called "system identities"). In the C code, these
2325 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2326 macro with the name beginning with "OPENSSL_SYS_", which is determined
2327 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2331 *) New option -set_serial to 'req' and 'x509' this allows the serial
2332 number to use to be specified on the command line. Previously self
2333 signed certificates were hard coded with serial number 0 and the
2334 CA options of 'x509' had to use a serial number in a file which was
2338 *) New options to 'ca' utility to support V2 CRL entry extensions.
2339 Currently CRL reason, invalidity date and hold instruction are
2340 supported. Add new CRL extensions to V3 code and some new objects.
2343 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2344 disable standard block padding (aka PKCS#5 padding) in the EVP
2345 API, which was previously mandatory. This means that the data is
2346 not padded in any way and so the total length much be a multiple
2347 of the block size, otherwise an error occurs.
2350 *) Initial (incomplete) OCSP SSL support.
2353 *) New function OCSP_parse_url(). This splits up a URL into its host,
2354 port and path components: primarily to parse OCSP URLs. New -url
2355 option to ocsp utility.
2358 *) New nonce behavior. The return value of OCSP_check_nonce() now
2359 reflects the various checks performed. Applications can decide
2360 whether to tolerate certain situations such as an absent nonce
2361 in a response when one was present in a request: the ocsp application
2362 just prints out a warning. New function OCSP_add1_basic_nonce()
2363 this is to allow responders to include a nonce in a response even if
2364 the request is nonce-less.
2367 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2368 skipped when using openssl x509 multiple times on a single input file,
2369 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2372 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2373 set string type: to handle setting ASN1_TIME structures. Fix ca
2374 utility to correctly initialize revocation date of CRLs.
2377 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2378 the clients preferred ciphersuites and rather use its own preferences.
2379 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2380 Internet Explorer by ensuring unchanged hash method during stepup.
2381 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2384 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2385 to aes and add a new 'exist' option to print out symbols that don't
2389 *) Additional options to ocsp utility to allow flags to be set and
2390 additional certificates supplied.
2393 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2394 OCSP client a number of certificate to only verify the response
2398 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2399 handle the new API. Currently only ECB, CBC modes supported. Add new
2402 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2403 Encryption Standard (AES) Ciphersuites for Transport Layer
2404 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2405 not enabled by default and were not part of the "ALL" ciphersuite
2406 alias because they were not yet official; they could be
2407 explicitly requested by specifying the "AESdraft" ciphersuite
2408 group alias. In the final release of OpenSSL 0.9.7, the group
2409 alias is called "AES" and is part of "ALL".)
2410 [Ben Laurie, Steve Henson, Bodo Moeller]
2412 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2413 request to response.
2416 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2417 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2418 extract information from a certificate request. OCSP_response_create()
2419 creates a response and optionally adds a basic response structure.
2420 OCSP_basic_add1_status() adds a complete single response to a basic
2421 response and returns the OCSP_SINGLERESP structure just added (to allow
2422 extensions to be included for example). OCSP_basic_add1_cert() adds a
2423 certificate to a basic response and OCSP_basic_sign() signs a basic
2424 response with various flags. New helper functions ASN1_TIME_check()
2425 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2426 (converts ASN1_TIME to GeneralizedTime).
2429 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2430 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2431 structure from a certificate. X509_pubkey_digest() digests the public_key
2432 contents: this is used in various key identifiers.
2435 *) Make sk_sort() tolerate a NULL argument.
2436 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2438 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2439 passed by the function are trusted implicitly. If any of them signed the
2440 response then it is assumed to be valid and is not verified.
2443 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2444 to data. This was previously part of the PKCS7 ASN1 code. This
2445 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2446 [Steve Henson, reported by Kenneth R. Robinette
2447 <support@securenetterm.com>]
2449 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2450 routines: without these tracing memory leaks is very painful.
2451 Fix leaks in PKCS12 and PKCS7 routines.
2454 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2455 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2456 effectively meant GeneralizedTime would never be used. Now it
2457 is initialised to -1 but X509_time_adj() now has to check the value
2458 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2459 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2460 [Steve Henson, reported by Kenneth R. Robinette
2461 <support@securenetterm.com>]
2463 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2464 result in a zero length in the ASN1_INTEGER structure which was
2465 not consistent with the structure when d2i_ASN1_INTEGER() was used
2466 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2467 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2468 where it did not print out a minus for negative ASN1_INTEGER.
2471 *) Add summary printout to ocsp utility. The various functions which
2472 convert status values to strings have been renamed to:
2473 OCSP_response_status_str(), OCSP_cert_status_str() and
2474 OCSP_crl_reason_str() and are no longer static. New options
2475 to verify nonce values and to disable verification. OCSP response
2476 printout format cleaned up.
2479 *) Add additional OCSP certificate checks. These are those specified
2480 in RFC2560. This consists of two separate checks: the CA of the
2481 certificate being checked must either be the OCSP signer certificate
2482 or the issuer of the OCSP signer certificate. In the latter case the
2483 OCSP signer certificate must contain the OCSP signing extended key
2484 usage. This check is performed by attempting to match the OCSP
2485 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2486 in the OCSP_CERTID structures of the response.
2489 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2490 and related routines. This uses the standard OpenSSL certificate
2491 verify routines to perform initial checks (just CA validity) and
2492 to obtain the certificate chain. Then additional checks will be
2493 performed on the chain. Currently the root CA is checked to see
2494 if it is explicitly trusted for OCSP signing. This is used to set
2495 a root CA as a global signing root: that is any certificate that
2496 chains to that CA is an acceptable OCSP signing certificate.
2499 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2500 extensions from a separate configuration file.
2501 As when reading extensions from the main configuration file,
2502 the '-extensions ...' option may be used for specifying the
2504 [Massimiliano Pala <madwolf@comune.modena.it>]
2506 *) New OCSP utility. Allows OCSP requests to be generated or
2507 read. The request can be sent to a responder and the output
2508 parsed, outputed or printed in text form. Not complete yet:
2509 still needs to check the OCSP response validity.
2512 *) New subcommands for 'openssl ca':
2513 'openssl ca -status <serial>' prints the status of the cert with
2514 the given serial number (according to the index file).
2515 'openssl ca -updatedb' updates the expiry status of certificates
2517 [Massimiliano Pala <madwolf@comune.modena.it>]
2519 *) New '-newreq-nodes' command option to CA.pl. This is like
2520 '-newreq', but calls 'openssl req' with the '-nodes' option
2521 so that the resulting key is not encrypted.
2522 [Damien Miller <djm@mindrot.org>]
2524 *) New configuration for the GNU Hurd.
2525 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2527 *) Initial code to implement OCSP basic response verify. This
2528 is currently incomplete. Currently just finds the signer's
2529 certificate and verifies the signature on the response.
2532 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2533 value of OPENSSLDIR. This is available via the new '-d' option
2534 to 'openssl version', and is also included in 'openssl version -a'.
2537 *) Allowing defining memory allocation callbacks that will be given
2538 file name and line number information in additional arguments
2539 (a const char* and an int). The basic functionality remains, as
2540 well as the original possibility to just replace malloc(),
2541 realloc() and free() by functions that do not know about these
2542 additional arguments. To register and find out the current
2543 settings for extended allocation functions, the following
2544 functions are provided:
2546 CRYPTO_set_mem_ex_functions
2547 CRYPTO_set_locked_mem_ex_functions
2548 CRYPTO_get_mem_ex_functions
2549 CRYPTO_get_locked_mem_ex_functions
2551 These work the same way as CRYPTO_set_mem_functions and friends.
2552 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2553 extended allocation function is enabled.
2554 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2555 a conventional allocation function is enabled.
2556 [Richard Levitte, Bodo Moeller]
2558 *) Finish off removing the remaining LHASH function pointer casts.
2559 There should no longer be any prototype-casting required when using
2560 the LHASH abstraction, and any casts that remain are "bugs". See
2561 the callback types and macros at the head of lhash.h for details
2562 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2565 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2566 If /dev/[u]random devices are not available or do not return enough
2567 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2569 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2570 /etc/entropy will be queried once each in this sequence, quering stops
2571 when enough entropy was collected without querying more sockets.
2574 *) Change the Unix RAND_poll() variant to be able to poll several
2575 random devices, as specified by DEVRANDOM, until a sufficient amount
2576 of data has been collected. We spend at most 10 ms on each file
2577 (select timeout) and read in non-blocking mode. DEVRANDOM now
2578 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2579 (previously it was just the string "/dev/urandom"), so on typical
2580 platforms the 10 ms delay will never occur.
2581 Also separate out the Unix variant to its own file, rand_unix.c.
2582 For VMS, there's a currently-empty rand_vms.c.
2585 *) Move OCSP client related routines to ocsp_cl.c. These
2586 provide utility functions which an application needing
2587 to issue a request to an OCSP responder and analyse the
2588 response will typically need: as opposed to those which an
2589 OCSP responder itself would need which will be added later.
2591 OCSP_request_sign() signs an OCSP request with an API similar
2592 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2593 response. OCSP_response_get1_basic() extracts basic response
2594 from response. OCSP_resp_find_status(): finds and extracts status
2595 information from an OCSP_CERTID structure (which will be created
2596 when the request structure is built). These are built from lower
2597 level functions which work on OCSP_SINGLERESP structures but
2598 wont normally be used unless the application wishes to examine
2599 extensions in the OCSP response for example.
2601 Replace nonce routines with a pair of functions.
2602 OCSP_request_add1_nonce() adds a nonce value and optionally
2603 generates a random value. OCSP_check_nonce() checks the
2604 validity of the nonce in an OCSP response.
2607 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2608 This doesn't copy the supplied OCSP_CERTID and avoids the
2609 need to free up the newly created id. Change return type
2610 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2611 This can then be used to add extensions to the request.
2612 Deleted OCSP_request_new(), since most of its functionality
2613 is now in OCSP_REQUEST_new() (and the case insensitive name
2614 clash) apart from the ability to set the request name which
2615 will be added elsewhere.
2618 *) Update OCSP API. Remove obsolete extensions argument from
2619 various functions. Extensions are now handled using the new
2620 OCSP extension code. New simple OCSP HTTP function which
2621 can be used to send requests and parse the response.
2624 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2625 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2626 uses the special reorder version of SET OF to sort the attributes
2627 and reorder them to match the encoded order. This resolves a long
2628 standing problem: a verify on a PKCS7 structure just after signing
2629 it used to fail because the attribute order did not match the
2630 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2631 it uses the received order. This is necessary to tolerate some broken
2632 software that does not order SET OF. This is handled by encoding
2633 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2634 to produce the required SET OF.
2637 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2638 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2639 files to get correct declarations of the ASN.1 item variables.
2642 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2643 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2644 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2645 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2646 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2647 ASN1_ITEM and no wrapper functions.
2650 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2651 replace the old function pointer based I/O routines. Change most of
2652 the *_d2i_bio() and *_d2i_fp() functions to use these.
2655 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2656 lines, recognice more "algorithms" that can be deselected, and make
2657 it complain about algorithm deselection that isn't recognised.
2660 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2661 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2662 to use new functions. Add NO_ASN1_OLD which can be set to remove
2663 some old style ASN1 functions: this can be used to determine if old
2664 code will still work when these eventually go away.
2667 *) New extension functions for OCSP structures, these follow the
2668 same conventions as certificates and CRLs.
2671 *) New function X509V3_add1_i2d(). This automatically encodes and
2672 adds an extension. Its behaviour can be customised with various
2673 flags to append, replace or delete. Various wrappers added for
2674 certifcates and CRLs.
2677 *) Fix to avoid calling the underlying ASN1 print routine when
2678 an extension cannot be parsed. Correct a typo in the
2679 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2682 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2683 entries for variables.
2686 *) Add functionality to apps/openssl.c for detecting locking
2687 problems: As the program is single-threaded, all we have
2688 to do is register a locking callback using an array for
2689 storing which locks are currently held by the program.
2692 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2693 SSL_get_ex_data_X509_STORE_idx(), which is used in
2694 ssl_verify_cert_chain() and thus can be called at any time
2695 during TLS/SSL handshakes so that thread-safety is essential.
2696 Unfortunately, the ex_data design is not at all suited
2697 for multi-threaded use, so it probably should be abolished.
2700 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2701 [Broadcom, tweaked and integrated by Geoff Thorpe]
2703 *) Move common extension printing code to new function
2704 X509V3_print_extensions(). Reorganise OCSP print routines and
2705 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2708 *) New function X509_signature_print() to remove duplication in some
2712 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2713 set (this was treated exactly the same as SET OF previously). This
2714 is used to reorder the STACK representing the structure to match the
2715 encoding. This will be used to get round a problem where a PKCS7
2716 structure which was signed could not be verified because the STACK
2717 order did not reflect the encoded order.
2720 *) Reimplement the OCSP ASN1 module using the new code.
2723 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2724 for its ASN1 operations. The old style function pointers still exist
2725 for now but they will eventually go away.
2728 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2729 completely replaces the old ASN1 functionality with a table driven
2730 encoder and decoder which interprets an ASN1_ITEM structure describing
2731 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2732 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2733 has also been converted to the new form.
2736 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2737 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2738 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2739 for negative moduli.
2742 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2743 of not touching the result's sign bit.
2746 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2750 *) Changed the LHASH code to use prototypes for callbacks, and created
2751 macros to declare and implement thin (optionally static) functions
2752 that provide type-safety and avoid function pointer casting for the
2753 type-specific callbacks.
2756 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2758 [Veers Staats <staatsvr@asc.hpc.mil>,
2759 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2761 *) Reformat the FAQ so the different questions and answers can be divided
2762 in sections depending on the subject.
2765 *) Have the zlib compression code load ZLIB.DLL dynamically under
2769 *) New function BN_mod_sqrt for computing square roots modulo a prime
2770 (using the probabilistic Tonelli-Shanks algorithm unless
2771 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2772 be handled deterministically).
2773 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2775 *) Make BN_mod_inverse faster by explicitly handling small quotients
2776 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2777 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2780 *) New function BN_kronecker.
2783 *) Fix BN_gcd so that it works on negative inputs; the result is
2784 positive unless both parameters are zero.
2785 Previously something reasonably close to an infinite loop was
2786 possible because numbers could be growing instead of shrinking
2787 in the implementation of Euclid's algorithm.
2790 *) Fix BN_is_word() and BN_is_one() macros to take into account the
2791 sign of the number in question.
2793 Fix BN_is_word(a,w) to work correctly for w == 0.
2795 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2796 because its test if the absolute value of 'a' equals 'w'.
2797 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2798 it exists mostly for use in the implementations of BN_is_zero(),
2799 BN_is_one(), and BN_is_word().
2802 *) New function BN_swap.
2805 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2806 the exponentiation functions are more likely to produce reasonable
2807 results on negative inputs.
2810 *) Change BN_mod_mul so that the result is always non-negative.
2811 Previously, it could be negative if one of the factors was negative;
2812 I don't think anyone really wanted that behaviour.
2815 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2816 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2817 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2818 and add new functions:
2827 BN_mod_lshift1_quick
2831 These functions always generate non-negative results.
2833 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2834 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2836 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2837 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2838 be reduced modulo m.
2839 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2842 The following entry accidentily appeared in the CHANGES file
2843 distributed with OpenSSL 0.9.7. The modifications described in
2844 it do *not* apply to OpenSSL 0.9.7.
2846 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2847 was actually never needed) and in BN_mul(). The removal in BN_mul()
2848 required a small change in bn_mul_part_recursive() and the addition
2849 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2850 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2851 bn_sub_words() and bn_add_words() except they take arrays with
2856 *) In 'openssl passwd', verify passwords read from the terminal
2857 unless the '-salt' option is used (which usually means that
2858 verification would just waste user's time since the resulting
2859 hash is going to be compared with some given password hash)
2860 or the new '-noverify' option is used.
2862 This is an incompatible change, but it does not affect
2863 non-interactive use of 'openssl passwd' (passwords on the command
2864 line, '-stdin' option, '-in ...' option) and thus should not
2868 *) Remove all references to RSAref, since there's no more need for it.
2871 *) Make DSO load along a path given through an environment variable
2872 (SHLIB_PATH) with shl_load().
2875 *) Constify the ENGINE code as a result of BIGNUM constification.
2876 Also constify the RSA code and most things related to it. In a
2877 few places, most notable in the depth of the ASN.1 code, ugly
2878 casts back to non-const were required (to be solved at a later
2882 *) Make it so the openssl application has all engines loaded by default.
2885 *) Constify the BIGNUM routines a little more.
2888 *) Add the following functions:
2890 ENGINE_load_cswift()
2892 ENGINE_load_atalla()
2894 ENGINE_load_builtin_engines()
2896 That way, an application can itself choose if external engines that
2897 are built-in in OpenSSL shall ever be used or not. The benefit is
2898 that applications won't have to be linked with libdl or other dso
2899 libraries unless it's really needed.
2901 Changed 'openssl engine' to load all engines on demand.
2902 Changed the engine header files to avoid the duplication of some
2903 declarations (they differed!).
2906 *) 'openssl engine' can now list capabilities.
2909 *) Better error reporting in 'openssl engine'.
2912 *) Never call load_dh_param(NULL) in s_server.
2915 *) Add engine application. It can currently list engines by name and
2916 identity, and test if they are actually available.
2919 *) Improve RPM specification file by forcing symbolic linking and making
2920 sure the installed documentation is also owned by root.root.
2921 [Damien Miller <djm@mindrot.org>]
2923 *) Give the OpenSSL applications more possibilities to make use of
2924 keys (public as well as private) handled by engines.
2927 *) Add OCSP code that comes from CertCo.
2930 *) Add VMS support for the Rijndael code.
2933 *) Added untested support for Nuron crypto accelerator.
2936 *) Add support for external cryptographic devices. This code was
2937 previously distributed separately as the "engine" branch.
2938 [Geoff Thorpe, Richard Levitte]
2940 *) Rework the filename-translation in the DSO code. It is now possible to
2941 have far greater control over how a "name" is turned into a filename
2942 depending on the operating environment and any oddities about the
2943 different shared library filenames on each system.
2946 *) Support threads on FreeBSD-elf in Configure.
2949 *) Fix for SHA1 assembly problem with MASM: it produces
2950 warnings about corrupt line number information when assembling
2951 with debugging information. This is caused by the overlapping
2953 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2956 NCONF_get_number() has no error checking at all. As a replacement,
2957 NCONF_get_number_e() is defined (_e for "error checking") and is
2958 promoted strongly. The old NCONF_get_number is kept around for
2959 binary backward compatibility.
2960 Make it possible for methods to load from something other than a BIO,
2961 by providing a function pointer that is given a name instead of a BIO.
2962 For example, this could be used to load configuration data from an
2966 *) Fix for non blocking accept BIOs. Added new I/O special reason
2967 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2968 with non blocking I/O was not possible because no retry code was
2969 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2973 *) Added the beginnings of Rijndael support.
2976 *) Fix for bug in DirectoryString mask setting. Add support for
2977 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2978 to allow certificate printing to more controllable, additional
2979 'certopt' option to 'x509' to allow new printing options to be
2983 *) Clean old EAY MD5 hack from e_os.h.
2986 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
2988 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2989 by using the Codenomicon TLS Test Tool (CAN-2004-0079)
2990 [Joe Orton, Steve Henson]
2992 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
2994 *) Fix additional bug revealed by the NISCC test suite:
2996 Stop bug triggering large recursion when presented with
2997 certain ASN.1 tags (CAN-2003-0851)
3000 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3002 *) Fix various bugs revealed by running the NISCC test suite:
3004 Stop out of bounds reads in the ASN1 code when presented with
3005 invalid tags (CAN-2003-0543 and CAN-2003-0544).
3007 If verify callback ignores invalid public key errors don't try to check
3008 certificate signature with the NULL public key.
3012 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3013 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3017 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3018 extra data after the compression methods not only for TLS 1.0
3019 but also for SSL 3.0 (as required by the specification).
3020 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3022 *) Change X509_certificate_type() to mark the key as exported/exportable
3023 when it's 512 *bits* long, not 512 bytes.
3026 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3028 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3029 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3030 a protocol version number mismatch like a decryption error
3031 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3034 *) Turn on RSA blinding by default in the default implementation
3035 to avoid a timing attack. Applications that don't want it can call
3036 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3037 They would be ill-advised to do so in most cases.
3038 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3040 *) Change RSA blinding code so that it works when the PRNG is not
3041 seeded (in this case, the secret RSA exponent is abused as
3042 an unpredictable seed -- if it is not unpredictable, there
3043 is no point in blinding anyway). Make RSA blinding thread-safe
3044 by remembering the creator's thread ID in rsa->blinding and
3045 having all other threads use local one-time blinding factors
3046 (this requires more computation than sharing rsa->blinding, but
3047 avoids excessive locking; and if an RSA object is not shared
3048 between threads, blinding will still be very fast).
3051 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3053 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3054 via timing by performing a MAC computation even if incorrrect
3055 block cipher padding has been found. This is a countermeasure
3056 against active attacks where the attacker has to distinguish
3057 between bad padding and a MAC verification error. (CAN-2003-0078)
3059 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3060 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3061 Martin Vuagnoux (EPFL, Ilion)]
3063 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3065 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3066 memory from it's contents. This is done with a counter that will
3067 place alternating values in each byte. This can be used to solve
3068 two issues: 1) the removal of calls to memset() by highly optimizing
3069 compilers, and 2) cleansing with other values than 0, since those can
3070 be read through on certain media, for example a swap space on disk.
3073 *) Bugfix: client side session caching did not work with external caching,
3074 because the session->cipher setting was not restored when reloading
3075 from the external cache. This problem was masked, when
3076 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3077 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3080 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3081 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3082 [Zeev Lieber <zeev-l@yahoo.com>]
3084 *) Undo an undocumented change introduced in 0.9.6e which caused
3085 repeated calls to OpenSSL_add_all_ciphers() and
3086 OpenSSL_add_all_digests() to be ignored, even after calling
3090 *) Change the default configuration reader to deal with last line not
3091 being properly terminated.
3094 *) Change X509_NAME_cmp() so it applies the special rules on handling
3095 DN values that are of type PrintableString, as well as RDNs of type
3096 emailAddress where the value has the type ia5String.
3097 [stefank@valicert.com via Richard Levitte]
3099 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3100 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3101 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3102 the bitwise-OR of the two for use by the majority of applications
3103 wanting this behaviour, and update the docs. The documented
3104 behaviour and actual behaviour were inconsistent and had been
3105 changing anyway, so this is more a bug-fix than a behavioural
3107 [Geoff Thorpe, diagnosed by Nadav Har'El]
3109 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3110 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3113 *) Fix initialization code race conditions in
3114 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3115 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3116 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3117 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3118 ssl2_get_cipher_by_char(),
3119 ssl3_get_cipher_by_char().
3120 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3122 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3123 the cached sessions are flushed, as the remove_cb() might use ex_data
3124 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3125 (see [openssl.org #212]).
3126 [Geoff Thorpe, Lutz Jaenicke]
3128 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3129 length, instead of the encoding length to d2i_ASN1_OBJECT.
3132 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3134 *) [In 0.9.6g-engine release:]
3135 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3136 [Lynn Gazis <lgazis@rainbow.com>]
3138 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3140 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3141 and get fix the header length calculation.
3142 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3143 Alon Kantor <alonk@checkpoint.com> (and others),
3146 *) Use proper error handling instead of 'assertions' in buffer
3147 overflow checks added in 0.9.6e. This prevents DoS (the
3148 assertions could call abort()).
3149 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3151 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3153 *) Add various sanity checks to asn1_get_length() to reject
3154 the ASN1 length bytes if they exceed sizeof(long), will appear
3155 negative or the content length exceeds the length of the
3157 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3159 *) Fix cipher selection routines: ciphers without encryption had no flags
3160 for the cipher strength set and where therefore not handled correctly
3161 by the selection routines (PR #130).
3164 *) Fix EVP_dsa_sha macro.
3168 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3169 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3170 that was added in OpenSSL 0.9.6d.
3172 As the countermeasure turned out to be incompatible with some
3173 broken SSL implementations, the new option is part of SSL_OP_ALL.
3174 SSL_OP_ALL is usually employed when compatibility with weird SSL
3175 implementations is desired (e.g. '-bugs' option to 's_client' and
3176 's_server'), so the new option is automatically set in many
3180 *) Changes in security patch:
3182 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3183 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3184 Air Force Materiel Command, USAF, under agreement number
3187 *) Add various sanity checks to asn1_get_length() to reject
3188 the ASN1 length bytes if they exceed sizeof(long), will appear
3189 negative or the content length exceeds the length of the
3190 supplied buffer. (CAN-2002-0659)
3191 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3193 *) Assertions for various potential buffer overflows, not known to
3195 [Ben Laurie (CHATS)]
3197 *) Various temporary buffers to hold ASCII versions of integers were
3198 too small for 64 bit platforms. (CAN-2002-0655)
3199 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3201 *) Remote buffer overflow in SSL3 protocol - an attacker could
3202 supply an oversized session ID to a client. (CAN-2002-0656)
3203 [Ben Laurie (CHATS)]
3205 *) Remote buffer overflow in SSL2 protocol - an attacker could
3206 supply an oversized client master key. (CAN-2002-0656)
3207 [Ben Laurie (CHATS)]
3209 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3211 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3212 encoded as NULL) with id-dsa-with-sha1.
3213 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3215 *) Check various X509_...() return values in apps/req.c.
3216 [Nils Larsch <nla@trustcenter.de>]
3218 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3219 an end-of-file condition would erronously be flagged, when the CRLF
3220 was just at the end of a processed block. The bug was discovered when
3221 processing data through a buffering memory BIO handing the data to a
3222 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3223 <ptsekov@syntrex.com> and Nedelcho Stanev.
3226 *) Implement a countermeasure against a vulnerability recently found
3227 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3228 before application data chunks to avoid the use of known IVs
3229 with data potentially chosen by the attacker.
3232 *) Fix length checks in ssl3_get_client_hello().
3235 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3236 to prevent ssl3_read_internal() from incorrectly assuming that
3237 ssl3_read_bytes() found application data while handshake
3238 processing was enabled when in fact s->s3->in_read_app_data was
3239 merely automatically cleared during the initial handshake.
3240 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3242 *) Fix object definitions for Private and Enterprise: they were not
3243 recognized in their shortname (=lowercase) representation. Extend
3244 obj_dat.pl to issue an error when using undefined keywords instead
3245 of silently ignoring the problem (Svenning Sorensen
3246 <sss@sss.dnsalias.net>).
3249 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3250 generators, i.e. generators other than 2 and 5. (Previously, the
3251 code did not properly initialise the 'add' and 'rem' values to
3252 BN_generate_prime().)
3254 In the new general case, we do not insist that 'generator' is
3255 actually a primitive root: This requirement is rather pointless;
3256 a generator of the order-q subgroup is just as good, if not
3260 *) Map new X509 verification errors to alerts. Discovered and submitted by
3261 Tom Wu <tom@arcot.com>.
3264 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3265 returning non-zero before the data has been completely received
3266 when using non-blocking I/O.
3267 [Bodo Moeller; problem pointed out by John Hughes]
3269 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3270 [Ben Laurie, Lutz Jaenicke]
3272 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3273 Yoram Zahavi <YoramZ@gilian.com>).
3276 *) Add information about CygWin 1.3 and on, and preserve proper
3277 configuration for the versions before that.
3278 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3280 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3281 check whether we deal with a copy of a session and do not delete from
3282 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3283 <izhar@checkpoint.com>.
3286 *) Do not store session data into the internal session cache, if it
3287 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3288 flag is set). Proposed by Aslam <aslam@funk.com>.
3291 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3295 *) [In 0.9.6d-engine release:]
3296 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3297 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3299 *) Add the configuration target linux-s390x.
3300 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3302 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3303 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3304 variable as an indication that a ClientHello message has been
3305 received. As the flag value will be lost between multiple
3306 invocations of ssl3_accept when using non-blocking I/O, the
3307 function may not be aware that a handshake has actually taken
3308 place, thus preventing a new session from being added to the
3311 To avoid this problem, we now set s->new_session to 2 instead of
3312 using a local variable.
3313 [Lutz Jaenicke, Bodo Moeller]
3315 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3316 if the SSL_R_LENGTH_MISMATCH error is detected.
3317 [Geoff Thorpe, Bodo Moeller]
3319 *) New 'shared_ldflag' column in Configure platform table.
3322 *) Fix EVP_CIPHER_mode macro.
3323 ["Dan S. Camper" <dan@bti.net>]
3325 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3326 type, we must throw them away by setting rr->length to 0.
3327 [D P Chang <dpc@qualys.com>]
3329 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3331 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3332 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3333 worked incorrectly for those cases where range = 10..._2 and
3334 3*range is two bits longer than range.)
3337 *) Only add signing time to PKCS7 structures if it is not already
3341 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3342 OBJ_ld_ce should be OBJ_id_ce.
3343 Also some ip-pda OIDs in crypto/objects/objects.txt were
3344 incorrect (cf. RFC 3039).
3345 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3347 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3348 returns early because it has nothing to do.
3349 [Andy Schneider <andy.schneider@bjss.co.uk>]
3351 *) [In 0.9.6c-engine release:]
3352 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3353 [Andy Schneider <andy.schneider@bjss.co.uk>]
3355 *) [In 0.9.6c-engine release:]
3356 Add support for Cryptographic Appliance's keyserver technology.
3357 (Use engine 'keyclient')
3358 [Cryptographic Appliances and Geoff Thorpe]
3360 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3361 is called via tools/c89.sh because arguments have to be
3362 rearranged (all '-L' options must appear before the first object
3364 [Richard Shapiro <rshapiro@abinitio.com>]
3366 *) [In 0.9.6c-engine release:]
3367 Add support for Broadcom crypto accelerator cards, backported
3369 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3371 *) [In 0.9.6c-engine release:]
3372 Add support for SureWare crypto accelerator cards from
3373 Baltimore Technologies. (Use engine 'sureware')
3374 [Baltimore Technologies and Mark Cox]
3376 *) [In 0.9.6c-engine release:]
3377 Add support for crypto accelerator cards from Accelerated
3378 Encryption Processing, www.aep.ie. (Use engine 'aep')
3379 [AEP Inc. and Mark Cox]
3381 *) Add a configuration entry for gcc on UnixWare.
3382 [Gary Benson <gbenson@redhat.com>]
3384 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3385 messages are stored in a single piece (fixed-length part and
3386 variable-length part combined) and fix various bugs found on the way.
3389 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3390 instead. BIO_gethostbyname() does not know what timeouts are
3391 appropriate, so entries would stay in cache even when they have
3393 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3395 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3396 faced with a pathologically small ClientHello fragment that does
3397 not contain client_version: Instead of aborting with an error,
3398 simply choose the highest available protocol version (i.e.,
3399 TLS 1.0 unless it is disabled). In practice, ClientHello
3400 messages are never sent like this, but this change gives us
3401 strictly correct behaviour at least for TLS.
3404 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3405 never resets s->method to s->ctx->method when called from within
3406 one of the SSL handshake functions.
3407 [Bodo Moeller; problem pointed out by Niko Baric]
3409 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3410 (sent using the client's version number) if client_version is
3411 smaller than the protocol version in use. Also change
3412 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3413 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3414 the client will at least see that alert.
3417 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3421 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3422 client receives HelloRequest while in a handshake.
3423 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3425 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3426 should end in 'break', not 'goto end' which circuments various
3427 cleanups done in state SSL_ST_OK. But session related stuff
3428 must be disabled for SSL_ST_OK in the case that we just sent a
3431 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3432 before just sending a HelloRequest.
3433 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3435 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3436 reveal whether illegal block cipher padding was found or a MAC
3437 verification error occured. (Neither SSLerr() codes nor alerts
3438 are directly visible to potential attackers, but the information
3439 may leak via logfiles.)
3441 Similar changes are not required for the SSL 2.0 implementation
3442 because the number of padding bytes is sent in clear for SSL 2.0,
3443 and the extra bytes are just ignored. However ssl/s2_pkt.c
3444 failed to verify that the purported number of padding bytes is in
3448 *) Add OpenUNIX-8 support including shared libraries
3449 (Boyd Lynn Gerber <gerberb@zenez.com>).
3452 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3453 'wristwatch attack' using huge encoding parameters (cf.
3454 James H. Manger's CRYPTO 2001 paper). Note that the
3455 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3456 encoding parameters and hence was not vulnerable.
3459 *) BN_sqr() bug fix.
3460 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3462 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3463 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3464 followed by modular reduction.
3465 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3467 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3468 equivalent based on BN_pseudo_rand() instead of BN_rand().
3471 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3472 This function was broken, as the check for a new client hello message
3473 to handle SGC did not allow these large messages.
3474 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3477 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3480 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3481 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3484 *) Rework the configuration and shared library support for Tru64 Unix.
3485 The configuration part makes use of modern compiler features and
3486 still retains old compiler behavior for those that run older versions
3487 of the OS. The shared library support part includes a variant that
3488 uses the RPATH feature, and is available through the special
3489 configuration target "alpha-cc-rpath", which will never be selected
3491 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3493 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3494 with the same message size as in ssl3_get_certificate_request().
3495 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3496 messages might inadvertently be reject as too long.
3497 [Petr Lampa <lampa@fee.vutbr.cz>]
3499 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3502 *) Modified SSL library such that the verify_callback that has been set
3503 specificly for an SSL object with SSL_set_verify() is actually being
3504 used. Before the change, a verify_callback set with this function was
3505 ignored and the verify_callback() set in the SSL_CTX at the time of
3506 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3507 to allow the necessary settings.
3510 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3511 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3512 done automatically (in contradiction to the requirements of the C
3513 standard). This made problems when used from OpenSSH.
3516 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3517 dh->length and always used
3519 BN_rand_range(priv_key, dh->p).
3521 BN_rand_range() is not necessary for Diffie-Hellman, and this
3522 specific range makes Diffie-Hellman unnecessarily inefficient if
3523 dh->length (recommended exponent length) is much smaller than the
3524 length of dh->p. We could use BN_rand_range() if the order of
3525 the subgroup was stored in the DH structure, but we only have
3530 BN_rand(priv_key, l, ...)
3532 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3538 RSA_eay_public_encrypt
3539 RSA_eay_private_decrypt
3540 RSA_eay_private_encrypt (signing)
3541 RSA_eay_public_decrypt (signature verification)
3543 (default implementations for RSA_public_encrypt,
3544 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3545 always reject numbers >= n.
3548 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3549 to synchronize access to 'locking_thread'. This is necessary on
3550 systems where access to 'locking_thread' (an 'unsigned long'
3551 variable) is not atomic.
3554 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3555 *before* setting the 'crypto_lock_rand' flag. The previous code had
3556 a race condition if 0 is a valid thread ID.
3557 [Travis Vitek <vitek@roguewave.com>]
3559 *) Add support for shared libraries under Irix.
3560 [Albert Chin-A-Young <china@thewrittenword.com>]
3562 *) Add configuration option to build on Linux on both big-endian and
3564 [Ralf Baechle <ralf@uni-koblenz.de>]
3566 *) Add the possibility to create shared libraries on HP-UX.
3569 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3571 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3572 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3573 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3574 PRNG state recovery was possible based on the output of
3575 one PRNG request appropriately sized to gain knowledge on
3576 'md' followed by enough consecutive 1-byte PRNG requests
3577 to traverse all of 'state'.
3579 1. When updating 'md_local' (the current thread's copy of 'md')
3580 during PRNG output generation, hash all of the previous
3581 'md_local' value, not just the half used for PRNG output.
3583 2. Make the number of bytes from 'state' included into the hash
3584 independent from the number of PRNG bytes requested.
3586 The first measure alone would be sufficient to avoid
3587 Markku-Juhani's attack. (Actually it had never occurred
3588 to me that the half of 'md_local' used for chaining was the
3589 half from which PRNG output bytes were taken -- I had always
3590 assumed that the secret half would be used.) The second
3591 measure makes sure that additional data from 'state' is never
3592 mixed into 'md_local' in small portions; this heuristically
3593 further strengthens the PRNG.
3596 *) Fix crypto/bn/asm/mips3.s.
3599 *) When only the key is given to "enc", the IV is undefined. Print out
3600 an error message in this case.
3603 *) Handle special case when X509_NAME is empty in X509 printing routines.
3606 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3607 positive and less than q.
3610 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3611 used: it isn't thread safe and the add_lock_callback should handle
3613 [Paul Rose <Paul.Rose@bridge.com>]
3615 *) Verify that incoming data obeys the block size in
3616 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3620 [Ulf Möller, Bodo Möller]
3622 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3623 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3624 when fixing the server behaviour for backwards-compatible 'client
3625 hello' messages. (Note that the attack is impractical against
3626 SSL 3.0 and TLS 1.0 anyway because length and version checking
3627 means that the probability of guessing a valid ciphertext is
3628 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3631 Before 0.9.5, the countermeasure (hide the error by generating a
3632 random 'decryption result') did not work properly because
3633 ERR_clear_error() was missing, meaning that SSL_get_error() would
3634 detect the supposedly ignored error.
3636 Both problems are now fixed.
3639 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3640 (previously it was 1024).
3643 *) Fix for compatibility mode trust settings: ignore trust settings
3644 unless some valid trust or reject settings are present.
3647 *) Fix for blowfish EVP: its a variable length cipher.
3650 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3651 parameters in DSA public key structures and return an error in the
3652 DSA routines if parameters are absent.
3655 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3656 in the current directory if neither $RANDFILE nor $HOME was set.
3657 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3658 caused some confusion to Windows users who haven't defined $HOME.
3659 Thus RAND_file_name() is changed again: e_os.h can define a
3660 DEFAULT_HOME, which will be used if $HOME is not set.
3661 For Windows, we use "C:"; on other platforms, we still require
3662 environment variables.
3664 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3665 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3666 having multiple threads call RAND_poll() concurrently.
3669 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3670 combination of a flag and a thread ID variable.
3671 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3672 flag), *other* threads can enter ssleay_add_bytes without obeying
3673 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3674 that they do not hold after the first thread unsets add_do_not_lock).
3677 *) Change bctest again: '-x' expressions are not available in all
3681 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3683 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3684 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3686 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3687 the default extension for executables, if any. Also, make the perl
3688 scripts that use symlink() to test if it really exists and use "cp"
3689 if it doesn't. All this made OpenSSL compilable and installable in
3693 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3694 If SEQUENCE is length is indefinite just set c->slen to the total
3695 amount of data available.
3696 [Steve Henson, reported by shige@FreeBSD.org]
3697 [This change does not apply to 0.9.7.]
3699 *) Change bctest to avoid here-documents inside command substitution
3700 (workaround for FreeBSD /bin/sh bug).
3701 For compatibility with Ultrix, avoid shell functions (introduced
3702 in the bctest version that searches along $PATH).
3705 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3706 with des_encrypt() defined on some operating systems, like Solaris
3710 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3711 On the Importance of Eliminating Errors in Cryptographic
3712 Computations, J. Cryptology 14 (2001) 2, 101-119,
3713 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3716 *) MIPS assembler BIGNUM division bug fix.
3719 *) Disabled incorrect Alpha assembler code.
3722 *) Fix PKCS#7 decode routines so they correctly update the length
3723 after reading an EOC for the EXPLICIT tag.
3725 [This change does not apply to 0.9.7.]
3727 *) Fix bug in PKCS#12 key generation routines. This was triggered
3728 if a 3DES key was generated with a 0 initial byte. Include
3729 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3730 (but broken) behaviour.
3733 *) Enhance bctest to search for a working bc along $PATH and print
3735 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3737 *) Fix memory leaks in err.c: free err_data string if necessary;
3738 don't write to the wrong index in ERR_set_error_data.
3741 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3745 *) Replace rdtsc with _emit statements for VC++ version 5.
3746 [Jeremy Cooper <jeremy@baymoo.org>]
3748 *) Make it possible to reuse SSLv2 sessions.
3751 *) In copy_email() check for >= 0 as a return value for
3752 X509_NAME_get_index_by_NID() since 0 is a valid index.
3753 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3755 *) Avoid coredump with unsupported or invalid public keys by checking if
3756 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3757 PKCS7_verify() fails with non detached data.
3760 *) Don't use getenv in library functions when run as setuid/setgid.
3761 New function OPENSSL_issetugid().
3764 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
3765 due to incorrect handling of multi-threading:
3767 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
3769 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
3771 3. Count how many times MemCheck_off() has been called so that
3772 nested use can be treated correctly. This also avoids
3773 inband-signalling in the previous code (which relied on the
3774 assumption that thread ID 0 is impossible).
3777 *) Add "-rand" option also to s_client and s_server.
3780 *) Fix CPU detection on Irix 6.x.
3781 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
3782 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3784 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
3787 [This change does not apply to 0.9.7.]
3789 *) Use the cached encoding of an X509_NAME structure rather than
3790 copying it. This is apparently the reason for the libsafe "errors"
3791 but the code is actually correct.
3794 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
3795 Bleichenbacher's DSA attack.
3796 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
3797 to be set and top=0 forces the highest bit to be set; top=-1 is new
3798 and leaves the highest bit random.
3799 [Ulf Moeller, Bodo Moeller]
3801 *) In the NCONF_...-based implementations for CONF_... queries
3802 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
3803 a temporary CONF structure with the data component set to NULL
3804 (which gives segmentation faults in lh_retrieve).
3805 Instead, use NULL for the CONF pointer in CONF_get_string and
3806 CONF_get_number (which may use environment variables) and directly
3807 return NULL from CONF_get_section.
3810 *) Fix potential buffer overrun for EBCDIC.
3813 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
3814 keyUsage if basicConstraints absent for a CA.
3817 *) Make SMIME_write_PKCS7() write mail header values with a format that
3818 is more generally accepted (no spaces before the semicolon), since
3819 some programs can't parse those values properly otherwise. Also make
3820 sure BIO's that break lines after each write do not create invalid
3824 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
3825 macros previously used would not encode an empty SEQUENCE OF
3826 and break the signature.
3828 [This change does not apply to 0.9.7.]
3830 *) Zero the premaster secret after deriving the master secret in
3834 *) Add some EVP_add_digest_alias registrations (as found in
3835 OpenSSL_add_all_digests()) to SSL_library_init()
3836 aka OpenSSL_add_ssl_algorithms(). This provides improved
3837 compatibility with peers using X.509 certificates
3838 with unconventional AlgorithmIdentifier OIDs.
3841 *) Fix for Irix with NO_ASM.
3842 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
3844 *) ./config script fixes.
3845 [Ulf Moeller, Richard Levitte]
3847 *) Fix 'openssl passwd -1'.
3850 *) Change PKCS12_key_gen_asc() so it can cope with non null
3851 terminated strings whose length is passed in the passlen
3852 parameter, for example from PEM callbacks. This was done
3853 by adding an extra length parameter to asc2uni().
3854 [Steve Henson, reported by <oddissey@samsung.co.kr>]
3856 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
3857 call failed, free the DSA structure.
3860 *) Fix to uni2asc() to cope with zero length Unicode strings.
3861 These are present in some PKCS#12 files.
3864 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
3865 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
3866 when writing a 32767 byte record.
3867 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
3869 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
3870 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
3872 (RSA objects have a reference count access to which is protected
3873 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
3874 so they are meant to be shared between threads.)
3875 [Bodo Moeller, Geoff Thorpe; original patch submitted by
3876 "Reddie, Steven" <Steven.Reddie@ca.com>]
3878 *) Fix a deadlock in CRYPTO_mem_leaks().
3881 *) Use better test patterns in bntest.
3884 *) rand_win.c fix for Borland C.
3887 *) BN_rshift bugfix for n == 0.
3890 *) Add a 'bctest' script that checks for some known 'bc' bugs
3891 so that 'make test' does not abort just because 'bc' is broken.
3894 *) Store verify_result within SSL_SESSION also for client side to
3895 avoid potential security hole. (Re-used sessions on the client side
3896 always resulted in verify_result==X509_V_OK, not using the original
3897 result of the server certificate verification.)
3900 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
3901 SSL3_RT_APPLICATION_DATA, return 0.
3902 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
3906 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
3907 releases, have been re-implemented by renaming the previous
3908 implementations of ssl2_read and ssl3_read to ssl2_read_internal
3909 and ssl3_read_internal, respectively, and adding 'peek' parameters
3910 to them. The new ssl[23]_{read,peek} functions are calls to
3911 ssl[23]_read_internal with the 'peek' flag set appropriately.
3912 A 'peek' parameter has also been added to ssl3_read_bytes, which
3913 does the actual work for ssl3_read_internal.
3916 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
3917 the method-specific "init()" handler. Also clean up ex_data after
3918 calling the method-specific "finish()" handler. Previously, this was
3919 happening the other way round.
3922 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
3923 The previous value, 12, was not always sufficient for BN_mod_exp().
3926 *) Make sure that shared libraries get the internal name engine with
3927 the full version number and not just 0. This should mark the
3928 shared libraries as not backward compatible. Of course, this should
3929 be changed again when we can guarantee backward binary compatibility.
3932 *) Fix typo in get_cert_by_subject() in by_dir.c
3933 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
3935 *) Rework the system to generate shared libraries:
3937 - Make note of the expected extension for the shared libraries and
3938 if there is a need for symbolic links from for example libcrypto.so.0
3939 to libcrypto.so.0.9.7. There is extended info in Configure for
3942 - Make as few rebuilds of the shared libraries as possible.
3944 - Still avoid linking the OpenSSL programs with the shared libraries.
3946 - When installing, install the shared libraries separately from the
3950 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
3952 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
3953 and not in SSL_clear because the latter is also used by the
3954 accept/connect functions; previously, the settings made by
3955 SSL_set_read_ahead would be lost during the handshake.
3956 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
3958 *) Correct util/mkdef.pl to be selective about disabled algorithms.
3959 Previously, it would create entries for disableed algorithms no
3963 *) Added several new manual pages for SSL_* function.
3966 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
3968 *) In ssl23_get_client_hello, generate an error message when faced
3969 with an initial SSL 3.0/TLS record that is too small to contain the
3970 first two bytes of the ClientHello message, i.e. client_version.
3971 (Note that this is a pathologic case that probably has never happened
3972 in real life.) The previous approach was to use the version number
3973 from the record header as a substitute; but our protocol choice
3974 should not depend on that one because it is not authenticated
3975 by the Finished messages.
3978 *) More robust randomness gathering functions for Windows.
3979 [Jeffrey Altman <jaltman@columbia.edu>]
3981 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
3982 not set then we don't setup the error code for issuer check errors
3983 to avoid possibly overwriting other errors which the callback does
3984 handle. If an application does set the flag then we assume it knows
3985 what it is doing and can handle the new informational codes
3989 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
3990 a general "ANY" type, as such it should be able to decode anything
3991 including tagged types. However it didn't check the class so it would
3992 wrongly interpret tagged types in the same way as their universal
3993 counterpart and unknown types were just rejected. Changed so that the
3994 tagged and unknown types are handled in the same way as a SEQUENCE:
3995 that is the encoding is stored intact. There is also a new type
3996 "V_ASN1_OTHER" which is used when the class is not universal, in this
3997 case we have no idea what the actual type is so we just lump them all
4001 *) On VMS, stdout may very well lead to a file that is written to
4002 in a record-oriented fashion. That means that every write() will
4003 write a separate record, which will be read separately by the
4004 programs trying to read from it. This can be very confusing.
4006 The solution is to put a BIO filter in the way that will buffer
4007 text until a linefeed is reached, and then write everything a
4008 line at a time, so every record written will be an actual line,
4009 not chunks of lines and not (usually doesn't happen, but I've
4010 seen it once) several lines in one record. BIO_f_linebuffer() is
4013 Currently, it's a VMS-only method, because that's where it has
4014 been tested well enough.
4017 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4018 it can return incorrect results.
4019 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4020 but it was in 0.9.6-beta[12].)
4023 *) Disable the check for content being present when verifying detached
4024 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4025 include zero length content when signing messages.
4028 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4029 BIO_ctrl (for BIO pairs).
4032 *) Add DSO method for VMS.
4035 *) Bug fix: Montgomery multiplication could produce results with the
4039 *) Add RPM specification openssl.spec and modify it to build three
4040 packages. The default package contains applications, application
4041 documentation and run-time libraries. The devel package contains
4042 include files, static libraries and function documentation. The
4043 doc package contains the contents of the doc directory. The original
4044 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4047 *) Add a large number of documentation files for many SSL routines.
4048 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4050 *) Add a configuration entry for Sony News 4.
4051 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4053 *) Don't set the two most significant bits to one when generating a
4054 random number < q in the DSA library.
4057 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4058 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4059 the underlying transport is blocking) if a handshake took place.
4060 (The default behaviour is needed by applications such as s_client
4061 and s_server that use select() to determine when to use SSL_read;
4062 but for applications that know in advance when to expect data, it
4063 just makes things more complicated.)
4066 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4070 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4071 work better on such systems.
4072 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4074 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4075 Update PKCS12_parse() so it copies the friendlyName and the
4076 keyid to the certificates aux info.
4079 *) Fix bug in PKCS7_verify() which caused an infinite loop
4080 if there was more than one signature.
4081 [Sven Uszpelkat <su@celocom.de>]
4083 *) Major change in util/mkdef.pl to include extra information
4084 about each symbol, as well as presentig variables as well
4085 as functions. This change means that there's n more need
4086 to rebuild the .num files when some algorithms are excluded.
4089 *) Allow the verify time to be set by an application,
4090 rather than always using the current time.
4093 *) Phase 2 verify code reorganisation. The certificate
4094 verify code now looks up an issuer certificate by a
4095 number of criteria: subject name, authority key id
4096 and key usage. It also verifies self signed certificates
4097 by the same criteria. The main comparison function is
4098 X509_check_issued() which performs these checks.
4100 Lot of changes were necessary in order to support this
4101 without completely rewriting the lookup code.
4103 Authority and subject key identifier are now cached.
4105 The LHASH 'certs' is X509_STORE has now been replaced
4106 by a STACK_OF(X509_OBJECT). This is mainly because an
4107 LHASH can't store or retrieve multiple objects with
4108 the same hash value.
4110 As a result various functions (which were all internal
4111 use only) have changed to handle the new X509_STORE
4112 structure. This will break anything that messed round
4113 with X509_STORE internally.
4115 The functions X509_STORE_add_cert() now checks for an
4116 exact match, rather than just subject name.
4118 The X509_STORE API doesn't directly support the retrieval
4119 of multiple certificates matching a given criteria, however
4120 this can be worked round by performing a lookup first
4121 (which will fill the cache with candidate certificates)
4122 and then examining the cache for matches. This is probably
4123 the best we can do without throwing out X509_LOOKUP
4124 entirely (maybe later...).
4126 The X509_VERIFY_CTX structure has been enhanced considerably.
4128 All certificate lookup operations now go via a get_issuer()
4129 callback. Although this currently uses an X509_STORE it
4130 can be replaced by custom lookups. This is a simple way
4131 to bypass the X509_STORE hackery necessary to make this
4132 work and makes it possible to use more efficient techniques
4133 in future. A very simple version which uses a simple
4134 STACK for its trusted certificate store is also provided
4135 using X509_STORE_CTX_trusted_stack().
4137 The verify_cb() and verify() callbacks now have equivalents
4138 in the X509_STORE_CTX structure.
4140 X509_STORE_CTX also has a 'flags' field which can be used
4141 to customise the verify behaviour.
4144 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4145 excludes S/MIME capabilities.
4148 *) When a certificate request is read in keep a copy of the
4149 original encoding of the signed data and use it when outputing
4150 again. Signatures then use the original encoding rather than
4151 a decoded, encoded version which may cause problems if the
4152 request is improperly encoded.
4155 *) For consistency with other BIO_puts implementations, call
4156 buffer_write(b, ...) directly in buffer_puts instead of calling
4159 In BIO_puts, increment b->num_write as in BIO_write.
4160 [Peter.Sylvester@EdelWeb.fr]
4162 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4163 BN_zero, we may not return a BIGNUM with an array consisting of
4167 *) Avoid calling abort() from within the library when problems are
4168 detected, except if preprocessor symbols have been defined
4169 (such as REF_CHECK, BN_DEBUG etc.).
4172 *) New openssl application 'rsautl'. This utility can be
4173 used for low level RSA operations. DER public key
4174 BIO/fp routines also added.
4177 *) New Configure entry and patches for compiling on QNX 4.
4178 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4180 *) A demo state-machine implementation was sponsored by
4181 Nuron (http://www.nuron.com/) and is now available in
4182 demos/state_machine.
4185 *) New options added to the 'dgst' utility for signature
4186 generation and verification.
4189 *) Unrecognized PKCS#7 content types are now handled via a
4190 catch all ASN1_TYPE structure. This allows unsupported
4191 types to be stored as a "blob" and an application can
4192 encode and decode it manually.
4195 *) Fix various signed/unsigned issues to make a_strex.c
4197 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4199 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4200 length if passed a buffer. ASN1_INTEGER_to_BN failed
4201 if passed a NULL BN and its argument was negative.
4202 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4204 *) Modification to PKCS#7 encoding routines to output definite
4205 length encoding. Since currently the whole structures are in
4206 memory there's not real point in using indefinite length
4207 constructed encoding. However if OpenSSL is compiled with
4208 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4211 *) Added BIO_vprintf() and BIO_vsnprintf().
4214 *) Added more prefixes to parse for in the the strings written
4215 through a logging bio, to cover all the levels that are available
4216 through syslog. The prefixes are now:
4218 PANIC, EMERG, EMR => LOG_EMERG
4219 ALERT, ALR => LOG_ALERT
4220 CRIT, CRI => LOG_CRIT
4221 ERROR, ERR => LOG_ERR
4222 WARNING, WARN, WAR => LOG_WARNING
4223 NOTICE, NOTE, NOT => LOG_NOTICE
4224 INFO, INF => LOG_INFO
4225 DEBUG, DBG => LOG_DEBUG
4227 and as before, if none of those prefixes are present at the
4228 beginning of the string, LOG_ERR is chosen.
4230 On Win32, the LOG_* levels are mapped according to this:
4232 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4233 LOG_WARNING => EVENTLOG_WARNING_TYPE
4234 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4238 *) Made it possible to reconfigure with just the configuration
4239 argument "reconf" or "reconfigure". The command line arguments
4240 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4241 and are retrieved from there when reconfiguring.
4245 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4247 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4250 *) The obj_dat.pl script was messing up the sorting of object
4251 names. The reason was that it compared the quoted version
4252 of strings as a result "OCSP" > "OCSP Signing" because
4253 " > SPACE. Changed script to store unquoted versions of
4254 names and add quotes on output. It was also omitting some
4255 names from the lookup table if they were given a default
4256 value (that is if SN is missing it is given the same
4257 value as LN and vice versa), these are now added on the
4258 grounds that if an object has a name we should be able to
4259 look it up. Finally added warning output when duplicate
4260 short or long names are found.
4263 *) Changes needed for Tandem NSK.
4264 [Scott Uroff <scott@xypro.com>]
4266 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4267 RSA_padding_check_SSLv23(), special padding was never detected
4268 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4269 version rollback attacks was not effective.
4271 In s23_clnt.c, don't use special rollback-attack detection padding
4272 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4273 client; similarly, in s23_srvr.c, don't do the rollback check if
4274 SSL 2.0 is the only protocol enabled in the server.
4277 *) Make it possible to get hexdumps of unprintable data with 'openssl
4278 asn1parse'. By implication, the functions ASN1_parse_dump() and
4279 BIO_dump_indent() are added.
4282 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4283 these print out strings and name structures based on various
4284 flags including RFC2253 support and proper handling of
4285 multibyte characters. Added options to the 'x509' utility
4286 to allow the various flags to be set.
4289 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4290 Also change the functions X509_cmp_current_time() and
4291 X509_gmtime_adj() work with an ASN1_TIME structure,
4292 this will enable certificates using GeneralizedTime in validity
4293 dates to be checked.
4296 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4297 negative public key encodings) on by default,
4298 NO_NEG_PUBKEY_BUG can be set to disable it.
4301 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4302 content octets. An i2c_ASN1_OBJECT is unnecessary because
4303 the encoding can be trivially obtained from the structure.
4306 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4307 not read locks (CRYPTO_r_[un]lock).
4310 *) A first attempt at creating official support for shared
4311 libraries through configuration. I've kept it so the
4312 default is static libraries only, and the OpenSSL programs
4313 are always statically linked for now, but there are
4314 preparations for dynamic linking in place.
4315 This has been tested on Linux and Tru64.
4318 *) Randomness polling function for Win9x, as described in:
4319 Peter Gutmann, Software Generation of Practically Strong
4323 *) Fix so PRNG is seeded in req if using an already existing
4327 *) New options to smime application. -inform and -outform
4328 allow alternative formats for the S/MIME message including
4329 PEM and DER. The -content option allows the content to be
4330 specified separately. This should allow things like Netscape
4331 form signing output easier to verify.
4334 *) Fix the ASN1 encoding of tags using the 'long form'.
4337 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4338 STRING types. These convert content octets to and from the
4339 underlying type. The actual tag and length octets are
4340 already assumed to have been read in and checked. These
4341 are needed because all other string types have virtually
4342 identical handling apart from the tag. By having versions
4343 of the ASN1 functions that just operate on content octets
4344 IMPLICIT tagging can be handled properly. It also allows
4345 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4346 and ASN1_INTEGER are identical apart from the tag.
4349 *) Change the handling of OID objects as follows:
4351 - New object identifiers are inserted in objects.txt, following
4352 the syntax given in objects.README.
4353 - objects.pl is used to process obj_mac.num and create a new
4355 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4358 This is currently kind of a hack, and the perl code in objects.pl
4359 isn't very elegant, but it works as I intended. The simplest way
4360 to check that it worked correctly is to look in obj_dat.h and
4361 check the array nid_objs and make sure the objects haven't moved
4362 around (this is important!). Additions are OK, as well as
4363 consistent name changes.
4366 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4369 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4370 The given file adds to whatever has already been seeded into the
4371 random pool through the RANDFILE configuration file option or
4372 environment variable, or the default random state file.
4375 *) mkstack.pl now sorts each macro group into lexical order.
4376 Previously the output order depended on the order the files
4377 appeared in the directory, resulting in needless rewriting
4381 *) Patches to make OpenSSL compile under Win32 again. Mostly
4382 work arounds for the VC++ problem that it treats func() as
4383 func(void). Also stripped out the parts of mkdef.pl that
4384 added extra typesafe functions: these no longer exist.
4387 *) Reorganisation of the stack code. The macros are now all
4388 collected in safestack.h . Each macro is defined in terms of
4389 a "stack macro" of the form SKM_<name>(type, a, b). The
4390 DEBUG_SAFESTACK is now handled in terms of function casts,
4391 this has the advantage of retaining type safety without the
4392 use of additional functions. If DEBUG_SAFESTACK is not defined
4393 then the non typesafe macros are used instead. Also modified the
4394 mkstack.pl script to handle the new form. Needs testing to see
4395 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4396 the default if no major problems. Similar behaviour for ASN1_SET_OF
4397 and PKCS12_STACK_OF.
4400 *) When some versions of IIS use the 'NET' form of private key the
4401 key derivation algorithm is different. Normally MD5(password) is
4402 used as a 128 bit RC4 key. In the modified case
4403 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4404 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4405 as the old Netscape_RSA functions except they have an additional
4406 'sgckey' parameter which uses the modified algorithm. Also added
4407 an -sgckey command line option to the rsa utility. Thanks to
4408 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4409 algorithm to openssl-dev.
4412 *) The evp_local.h macros were using 'c.##kname' which resulted in
4413 invalid expansion on some systems (SCO 5.0.5 for example).
4414 Corrected to 'c.kname'.
4415 [Phillip Porch <root@theporch.com>]
4417 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4418 a STACK of email addresses from a certificate or request, these look
4419 in the subject name and the subject alternative name extensions and
4420 omit any duplicate addresses.
4423 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4424 This makes DSA verification about 2 % faster.
4427 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4428 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4429 plus overhead for 1024 bit moduli).
4430 This makes exponentiations about 0.5 % faster for 1024 bit
4431 exponents (as measured by "openssl speed rsa2048").
4434 *) Rename memory handling macros to avoid conflicts with other
4436 Malloc => OPENSSL_malloc
4437 Malloc_locked => OPENSSL_malloc_locked
4438 Realloc => OPENSSL_realloc
4439 Free => OPENSSL_free
4442 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4443 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4446 *) CygWin32 support.
4447 [John Jarvie <jjarvie@newsguy.com>]
4449 *) The type-safe stack code has been rejigged. It is now only compiled
4450 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4451 by default all type-specific stack functions are "#define"d back to
4452 standard stack functions. This results in more streamlined output
4453 but retains the type-safety checking possibilities of the original
4457 *) The STACK code has been cleaned up, and certain type declarations
4458 that didn't make a lot of sense have been brought in line. This has
4459 also involved a cleanup of sorts in safestack.h to more correctly
4460 map type-safe stack functions onto their plain stack counterparts.
4461 This work has also resulted in a variety of "const"ifications of
4462 lots of the code, especially "_cmp" operations which should normally
4463 be prototyped with "const" parameters anyway.
4466 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4467 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4468 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4469 where all of 'md' is used each time the PRNG is used, but 'state'
4470 is used only indexed by a cyclic counter. As entropy may not be
4471 well distributed from the beginning, 'md' is important as a
4472 chaining variable. However, the output function chains only half
4473 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4474 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4475 in all of 'state' being rewritten, with the new values depending
4476 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4479 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4480 the handshake is continued after ssl_verify_cert_chain();
4481 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4482 can lead to 'unexplainable' connection aborts later.
4483 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4485 *) Major EVP API cipher revision.
4486 Add hooks for extra EVP features. This allows various cipher
4487 parameters to be set in the EVP interface. Support added for variable
4488 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4489 setting of RC2 and RC5 parameters.
4491 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4494 Remove lots of duplicated code from the EVP library. For example *every*
4495 cipher init() function handles the 'iv' in the same way according to the
4496 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4497 for CFB and OFB modes they zero ctx->num.
4499 New functionality allows removal of S/MIME code RC2 hack.
4501 Most of the routines have the same form and so can be declared in terms
4504 By shifting this to the top level EVP_CipherInit() it can be removed from
4505 all individual ciphers. If the cipher wants to handle IVs or keys
4506 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4509 Change lots of functions like EVP_EncryptUpdate() to now return a
4510 value: although software versions of the algorithms cannot fail
4511 any installed hardware versions can.
4514 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4515 this option is set, tolerate broken clients that send the negotiated
4516 protocol version number instead of the requested protocol version
4520 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4521 i.e. non-zero for export ciphersuites, zero otherwise.
4522 Previous versions had this flag inverted, inconsistent with
4523 rsa_tmp_cb (..._TMP_RSA_CB).
4524 [Bodo Moeller; problem reported by Amit Chopra]
4526 *) Add missing DSA library text string. Work around for some IIS
4527 key files with invalid SEQUENCE encoding.
4530 *) Add a document (doc/standards.txt) that list all kinds of standards
4531 and so on that are implemented in OpenSSL.
4534 *) Enhance c_rehash script. Old version would mishandle certificates
4535 with the same subject name hash and wouldn't handle CRLs at all.
4536 Added -fingerprint option to crl utility, to support new c_rehash
4540 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4543 *) Fix for SSL server purpose checking. Server checking was
4544 rejecting certificates which had extended key usage present
4545 but no ssl client purpose.
4546 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4548 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4549 is a little unclear about how a blank password is handled.
4550 Since the password in encoded as a BMPString with terminating
4551 double NULL a zero length password would end up as just the
4552 double NULL. However no password at all is different and is
4553 handled differently in the PKCS#12 key generation code. NS
4554 treats a blank password as zero length. MSIE treats it as no
4555 password on export: but it will try both on import. We now do
4556 the same: PKCS12_parse() tries zero length and no password if
4557 the password is set to "" or NULL (NULL is now a valid password:
4558 it wasn't before) as does the pkcs12 application.
4561 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4562 perror when PEM_read_bio_X509_REQ fails, the error message must
4563 be obtained from the error queue.
4566 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4567 it in ERR_remove_state if appropriate, and change ERR_get_state
4568 accordingly to avoid race conditions (this is necessary because
4569 thread_hash is no longer constant once set).
4572 *) Bugfix for linux-elf makefile.one.
4575 *) RSA_get_default_method() will now cause a default
4576 RSA_METHOD to be chosen if one doesn't exist already.
4577 Previously this was only set during a call to RSA_new()
4578 or RSA_new_method(NULL) meaning it was possible for
4579 RSA_get_default_method() to return NULL.
4582 *) Added native name translation to the existing DSO code
4583 that will convert (if the flag to do so is set) filenames
4584 that are sufficiently small and have no path information
4585 into a canonical native form. Eg. "blah" converted to
4586 "libblah.so" or "blah.dll" etc.
4589 *) New function ERR_error_string_n(e, buf, len) which is like
4590 ERR_error_string(e, buf), but writes at most 'len' bytes
4591 including the 0 terminator. For ERR_error_string_n, 'buf'
4593 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4595 *) CONF library reworked to become more general. A new CONF
4596 configuration file reader "class" is implemented as well as a
4597 new functions (NCONF_*, for "New CONF") to handle it. The now
4598 old CONF_* functions are still there, but are reimplemented to
4599 work in terms of the new functions. Also, a set of functions
4600 to handle the internal storage of the configuration data is
4601 provided to make it easier to write new configuration file
4602 reader "classes" (I can definitely see something reading a
4603 configuration file in XML format, for example), called _CONF_*,
4604 or "the configuration storage API"...
4606 The new configuration file reading functions are:
4608 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4609 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4611 NCONF_default, NCONF_WIN32
4613 NCONF_dump_fp, NCONF_dump_bio
4615 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4616 NCONF_new creates a new CONF object. This works in the same way
4617 as other interfaces in OpenSSL, like the BIO interface.
4618 NCONF_dump_* dump the internal storage of the configuration file,
4619 which is useful for debugging. All other functions take the same
4620 arguments as the old CONF_* functions wth the exception of the
4621 first that must be a `CONF *' instead of a `LHASH *'.
4623 To make it easer to use the new classes with the old CONF_* functions,
4624 the function CONF_set_default_method is provided.
4627 *) Add '-tls1' option to 'openssl ciphers', which was already
4628 mentioned in the documentation but had not been implemented.
4629 (This option is not yet really useful because even the additional
4630 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4633 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4634 OpenSSL-based applications) load shared libraries and bind to
4635 them in a portable way.
4636 [Geoff Thorpe, with contributions from Richard Levitte]
4638 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4640 *) Make sure _lrotl and _lrotr are only used with MSVC.
4642 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4643 (the default implementation of RAND_status).
4645 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4646 to '-clrext' (= clear extensions), as intended and documented.
4647 [Bodo Moeller; inconsistency pointed out by Michael Attili
4650 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4651 was larger than the MD block size.
4652 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4654 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4655 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4656 using the passed key: if the passed key was a private key the result
4657 of X509_print(), for example, would be to print out all the private key
4661 *) des_quad_cksum() byte order bug fix.
4662 [Ulf Möller, using the problem description in krb4-0.9.7, where
4663 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4665 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4667 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4669 *) For easily testing in shell scripts whether some command
4670 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4671 returns with exit code 0 iff no command of the given name is available.
4672 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4673 the output goes to stdout and nothing is printed to stderr.
4674 Additional arguments are always ignored.
4676 Since for each cipher there is a command of the same name,
4677 the 'no-cipher' compilation switches can be tested this way.
4679 ('openssl no-XXX' is not able to detect pseudo-commands such
4680 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4683 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4686 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4687 is set; it will be thrown away anyway because each handshake creates
4689 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4690 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4691 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4692 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4695 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4696 'Q' and 'R' lose their special meanings (quit/renegotiate).
4697 This is part of what -quiet does; unlike -quiet, -ign_eof
4698 does not suppress any output.
4701 *) Add compatibility options to the purpose and trust code. The
4702 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4703 accepts a certificate or CA, this was the previous behaviour,
4704 with all the associated security issues.
4706 X509_TRUST_COMPAT is the old trust behaviour: only and
4707 automatically trust self signed roots in certificate store. A
4708 new trust setting X509_TRUST_DEFAULT is used to specify that
4709 a purpose has no associated trust setting and it should instead
4710 use the value in the default purpose.
4713 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4714 and fix a memory leak.
4717 *) In util/mkerr.pl (which implements 'make errors'), preserve
4718 reason strings from the previous version of the .c file, as
4719 the default to have only downcase letters (and digits) in
4720 automatically generated reasons codes is not always appropriate.
4723 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4724 using strerror. Previously, ERR_reason_error_string() returned
4725 library names as reason strings for SYSerr; but SYSerr is a special
4726 case where small numbers are errno values, not library numbers.
4729 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4730 converts DSA parameters into DH parameters. (When creating parameters,
4731 DSA_generate_parameters is used.)
4734 *) Include 'length' (recommended exponent length) in C code generated
4735 by 'openssl dhparam -C'.
4738 *) The second argument to set_label in perlasm was already being used
4739 so couldn't be used as a "file scope" flag. Moved to third argument
4743 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4744 instead of RAND_bytes for encryption IVs and salts.
4747 *) Include RAND_status() into RAND_METHOD instead of implementing
4748 it only for md_rand.c Otherwise replacing the PRNG by calling
4749 RAND_set_rand_method would be impossible.
4752 *) Don't let DSA_generate_key() enter an infinite loop if the random
4753 number generation fails.
4756 *) New 'rand' application for creating pseudo-random output.
4759 *) Added configuration support for Linux/IA64
4760 [Rolf Haberrecker <rolf@suse.de>]
4762 *) Assembler module support for Mingw32.
4765 *) Shared library support for HPUX (in shlib/).
4766 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
4768 *) Shared library support for Solaris gcc.
4769 [Lutz Behnke <behnke@trustcenter.de>]
4771 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
4773 *) PKCS7_encrypt() was adding text MIME headers twice because they
4774 were added manually and by SMIME_crlf_copy().
4777 *) In bntest.c don't call BN_rand with zero bits argument.
4778 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
4780 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
4781 case was implemented. This caused BN_div_recp() to fail occasionally.
4784 *) Add an optional second argument to the set_label() in the perl
4785 assembly language builder. If this argument exists and is set
4786 to 1 it signals that the assembler should use a symbol whose
4787 scope is the entire file, not just the current function. This
4788 is needed with MASM which uses the format label:: for this scope.
4789 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
4791 *) Change the ASN1 types so they are typedefs by default. Before
4792 almost all types were #define'd to ASN1_STRING which was causing
4793 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
4797 *) Change names of new functions to the new get1/get0 naming
4798 convention: After 'get1', the caller owns a reference count
4799 and has to call ..._free; 'get0' returns a pointer to some
4800 data structure without incrementing reference counters.
4801 (Some of the existing 'get' functions increment a reference
4802 counter, some don't.)
4803 Similarly, 'set1' and 'add1' functions increase reference
4804 counters or duplicate objects.
4807 *) Allow for the possibility of temp RSA key generation failure:
4808 the code used to assume it always worked and crashed on failure.
4811 *) Fix potential buffer overrun problem in BIO_printf().
4812 [Ulf Möller, using public domain code by Patrick Powell; problem
4813 pointed out by David Sacerdote <das33@cornell.edu>]
4815 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
4816 RAND_egd() and RAND_status(). In the command line application,
4817 the EGD socket can be specified like a seed file using RANDFILE
4821 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
4822 Some CAs (e.g. Verisign) distribute certificates in this form.
4825 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
4826 list to exclude them. This means that no special compilation option
4827 is needed to use anonymous DH: it just needs to be included in the
4831 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
4832 EVP_MD_type. The old functionality is available in a new macro called
4833 EVP_MD_md(). Change code that uses it and update docs.
4836 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
4837 where the 'void *' argument is replaced by a function pointer argument.
4838 Previously 'void *' was abused to point to functions, which works on
4839 many platforms, but is not correct. As these functions are usually
4840 called by macros defined in OpenSSL header files, most source code
4841 should work without changes.
4844 *) <openssl/opensslconf.h> (which is created by Configure) now contains
4845 sections with information on -D... compiler switches used for
4846 compiling the library so that applications can see them. To enable
4847 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
4848 must be defined. E.g.,
4849 #define OPENSSL_ALGORITHM_DEFINES
4850 #include <openssl/opensslconf.h>
4851 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
4852 [Richard Levitte, Ulf and Bodo Möller]
4854 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
4858 *) Change the 'other' type in certificate aux info to a STACK_OF
4859 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
4860 the required ASN1 format: arbitrary types determined by an OID.
4863 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
4864 argument to 'req'. This is not because the function is newer or
4865 better than others it just uses the work 'NEW' in the certificate
4866 request header lines. Some software needs this.
4869 *) Reorganise password command line arguments: now passwords can be
4870 obtained from various sources. Delete the PEM_cb function and make
4871 it the default behaviour: i.e. if the callback is NULL and the
4872 usrdata argument is not NULL interpret it as a null terminated pass
4873 phrase. If usrdata and the callback are NULL then the pass phrase
4874 is prompted for as usual.
4877 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
4878 the support is automatically enabled. The resulting binaries will
4879 autodetect the card and use it if present.
4880 [Ben Laurie and Compaq Inc.]
4882 *) Work around for Netscape hang bug. This sends certificate request
4883 and server done in one record. Since this is perfectly legal in the
4884 SSL/TLS protocol it isn't a "bug" option and is on by default. See
4885 the bugs/SSLv3 entry for more info.
4888 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
4891 *) Add -rand argument to smime and pkcs12 applications and read/write
4895 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
4898 *) Add command line password options to the remaining applications.
4901 *) Bug fix for BN_div_recp() for numerators with an even number of
4905 *) More tests in bntest.c, and changed test_bn output.
4908 *) ./config recognizes MacOS X now.
4911 *) Bug fix for BN_div() when the first words of num and divsor are
4912 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
4915 *) Add support for various broken PKCS#8 formats, and command line
4916 options to produce them.
4919 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
4920 get temporary BIGNUMs from a BN_CTX.
4923 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
4927 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
4928 include a #define from the old name to the new. The original intent
4929 was that statically linked binaries could for example just call
4930 SSLeay_add_all_ciphers() to just add ciphers to the table and not
4931 link with digests. This never worked becayse SSLeay_add_all_digests()
4932 and SSLeay_add_all_ciphers() were in the same source file so calling
4933 one would link with the other. They are now in separate source files.
4936 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
4939 *) Use a less unusual form of the Miller-Rabin primality test (it used
4940 a binary algorithm for exponentiation integrated into the Miller-Rabin
4941 loop, our standard modexp algorithms are faster).
4944 *) Support for the EBCDIC character set completed.
4945 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
4947 *) Source code cleanups: use const where appropriate, eliminate casts,
4948 use void * instead of char * in lhash.
4951 *) Bugfix: ssl3_send_server_key_exchange was not restartable
4952 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
4953 this the server could overwrite ephemeral keys that the client
4957 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
4958 using 50 iterations of the Rabin-Miller test.
4960 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
4961 iterations of the Rabin-Miller test as required by the appendix
4962 to FIPS PUB 186[-1]) instead of DSA_is_prime.
4963 As BN_is_prime_fasttest includes trial division, DSA parameter
4964 generation becomes much faster.
4966 This implies a change for the callback functions in DSA_is_prime
4967 and DSA_generate_parameters: The callback function is called once
4968 for each positive witness in the Rabin-Miller test, not just
4969 occasionally in the inner loop; and the parameters to the
4970 callback function now provide an iteration count for the outer
4971 loop rather than for the current invocation of the inner loop.
4972 DSA_generate_parameters additionally can call the callback
4973 function with an 'iteration count' of -1, meaning that a
4974 candidate has passed the trial division test (when q is generated
4975 from an application-provided seed, trial division is skipped).
4978 *) New function BN_is_prime_fasttest that optionally does trial
4979 division before starting the Rabin-Miller test and has
4980 an additional BN_CTX * argument (whereas BN_is_prime always
4981 has to allocate at least one BN_CTX).
4982 'callback(1, -1, cb_arg)' is called when a number has passed the
4983 trial division stage.
4986 *) Fix for bug in CRL encoding. The validity dates weren't being handled
4990 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
4993 *) New function BN_pseudo_rand().
4996 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
4997 bignum version of BN_from_montgomery() with the working code from
4998 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5002 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5003 made it impossible to use the same SSL_SESSION data structure in
5004 SSL2 clients in multiple threads.
5007 *) The return value of RAND_load_file() no longer counts bytes obtained
5008 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5009 to seed the PRNG (previously an explicit byte count was required).
5010 [Ulf Möller, Bodo Möller]
5012 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5013 used (char *) instead of (void *) and had casts all over the place.
5016 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5019 *) Retain source code compatibility for BN_prime_checks macro:
5020 BN_is_prime(..., BN_prime_checks, ...) now uses
5021 BN_prime_checks_for_size to determine the appropriate number of
5022 Rabin-Miller iterations.
5025 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5026 DH_CHECK_P_NOT_SAFE_PRIME.
5027 (Check if this is true? OpenPGP calls them "strong".)
5030 *) Merge the functionality of "dh" and "gendh" programs into a new program
5031 "dhparam". The old programs are retained for now but will handle DH keys
5032 (instead of parameters) in future.
5035 *) Make the ciphers, s_server and s_client programs check the return values
5036 when a new cipher list is set.
5039 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5040 ciphers. Before when the 56bit ciphers were enabled the sorting was
5043 The syntax for the cipher sorting has been extended to support sorting by
5044 cipher-strength (using the strength_bits hard coded in the tables).
5045 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5047 Fix a bug in the cipher-command parser: when supplying a cipher command
5048 string with an "undefined" symbol (neither command nor alphanumeric
5049 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5050 an error is flagged.
5052 Due to the strength-sorting extension, the code of the
5053 ssl_create_cipher_list() function was completely rearranged. I hope that
5054 the readability was also increased :-)
5055 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5057 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5058 for the first serial number and places 2 in the serial number file. This
5059 avoids problems when the root CA is created with serial number zero and
5060 the first user certificate has the same issuer name and serial number
5064 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5065 the new code. Add documentation for this stuff.
5068 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5069 X509_*() to X509at_*() on the grounds that they don't handle X509
5070 structures and behave in an analagous way to the X509v3 functions:
5071 they shouldn't be called directly but wrapper functions should be used
5074 So we also now have some wrapper functions that call the X509at functions
5075 when passed certificate requests. (TO DO: similar things can be done with
5076 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5077 things. Some of these need some d2i or i2d and print functionality
5078 because they handle more complex structures.)
5081 *) Add missing #ifndefs that caused missing symbols when building libssl
5082 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5083 NO_RSA in ssl/s2*.c.
5084 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5086 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5087 has a return value which indicates the quality of the random data
5088 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5089 error queue. New function RAND_pseudo_bytes() generates output that is
5090 guaranteed to be unique but not unpredictable. RAND_add is like
5091 RAND_seed, but takes an extra argument for an entropy estimate
5092 (RAND_seed always assumes full entropy).
5095 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5096 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5097 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5098 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5099 false-positive rate of at most 2^-80 for random input.
5102 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5105 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5106 in the 0.9.5 release), this returns the chain
5107 from an X509_CTX structure with a dup of the stack and all
5108 the X509 reference counts upped: so the stack will exist
5109 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5112 Also make SSL_SESSION_print() print out the verify return
5116 *) Add manpage for the pkcs12 command. Also change the default
5117 behaviour so MAC iteration counts are used unless the new
5118 -nomaciter option is used. This improves file security and
5119 only older versions of MSIE (4.0 for example) need it.
5122 *) Honor the no-xxx Configure options when creating .DEF files.
5125 *) Add PKCS#10 attributes to field table: challengePassword,
5126 unstructuredName and unstructuredAddress. These are taken from
5127 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5128 international characters are used.
5130 More changes to X509_ATTRIBUTE code: allow the setting of types
5131 based on strings. Remove the 'loc' parameter when adding
5132 attributes because these will be a SET OF encoding which is sorted
5136 *) Initial changes to the 'req' utility to allow request generation
5137 automation. This will allow an application to just generate a template
5138 file containing all the field values and have req construct the
5141 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5142 used all over the place including certificate requests and PKCS#7
5143 structures. They are currently handled manually where necessary with
5144 some primitive wrappers for PKCS#7. The new functions behave in a
5145 manner analogous to the X509 extension functions: they allow
5146 attributes to be looked up by NID and added.
5148 Later something similar to the X509V3 code would be desirable to
5149 automatically handle the encoding, decoding and printing of the
5150 more complex types. The string types like challengePassword can
5151 be handled by the string table functions.
5153 Also modified the multi byte string table handling. Now there is
5154 a 'global mask' which masks out certain types. The table itself
5155 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5156 is useful when for example there is only one permissible type
5157 (as in countryName) and using the mask might result in no valid
5161 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5162 SSL_get_peer_finished to allow applications to obtain the latest
5163 Finished messages sent to the peer or expected from the peer,
5164 respectively. (SSL_get_peer_finished is usually the Finished message
5165 actually received from the peer, otherwise the protocol will be aborted.)
5167 As the Finished message are message digests of the complete handshake
5168 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5169 be used for external authentication procedures when the authentication
5170 provided by SSL/TLS is not desired or is not enough.
5173 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5174 the host supports BWX extension and if Compaq C is present on the
5175 $PATH. Just exploiting of the BWX extension results in 20-30%
5176 performance kick for some algorithms, e.g. DES and RC4 to mention
5177 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5181 *) Add support for MS "fast SGC". This is arguably a violation of the
5182 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5183 weak crypto and after checking the certificate is SGC a second one
5184 with strong crypto. MS SGC stops the first handshake after receiving
5185 the server certificate message and sends a second client hello. Since
5186 a server will typically do all the time consuming operations before
5187 expecting any further messages from the client (server key exchange
5188 is the most expensive) there is little difference between the two.
5190 To get OpenSSL to support MS SGC we have to permit a second client
5191 hello message after we have sent server done. In addition we have to
5192 reset the MAC if we do get this second client hello.
5195 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5196 if a DER encoded private key is RSA or DSA traditional format. Changed
5197 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5198 format DER encoded private key. Newer code should use PKCS#8 format which
5199 has the key type encoded in the ASN1 structure. Added DER private key
5200 support to pkcs8 application.
5203 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5204 ciphersuites has been selected (as required by the SSL 3/TLS 1
5205 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5206 is set, we interpret this as a request to violate the specification
5207 (the worst that can happen is a handshake failure, and 'correct'
5208 behaviour would result in a handshake failure anyway).
5211 *) In SSL_CTX_add_session, take into account that there might be multiple
5212 SSL_SESSION structures with the same session ID (e.g. when two threads
5213 concurrently obtain them from an external cache).
5214 The internal cache can handle only one SSL_SESSION with a given ID,
5215 so if there's a conflict, we now throw out the old one to achieve
5219 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5220 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5221 some routines that use cipher OIDs: some ciphers do not have OIDs
5222 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5226 *) Simplify the trust setting structure and code. Now we just have
5227 two sequences of OIDs for trusted and rejected settings. These will
5228 typically have values the same as the extended key usage extension
5229 and any application specific purposes.
5231 The trust checking code now has a default behaviour: it will just
5232 check for an object with the same NID as the passed id. Functions can
5233 be provided to override either the default behaviour or the behaviour
5234 for a given id. SSL client, server and email already have functions
5235 in place for compatibility: they check the NID and also return "trusted"
5236 if the certificate is self signed.
5239 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5240 traditional format into an EVP_PKEY structure.
5243 *) Add a password callback function PEM_cb() which either prompts for
5244 a password if usr_data is NULL or otherwise assumes it is a null
5245 terminated password. Allow passwords to be passed on command line
5246 environment or config files in a few more utilities.
5249 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5250 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5251 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5252 Update documentation.
5255 *) Support for ASN1 "NULL" type. This could be handled before by using
5256 ASN1_TYPE but there wasn't any function that would try to read a NULL
5257 and produce an error if it couldn't. For compatibility we also have
5258 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5259 don't allocate anything because they don't need to.
5262 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5264 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5266 *) Rebuild of the memory allocation routines used by OpenSSL code and
5267 possibly others as well. The purpose is to make an interface that
5268 provide hooks so anyone can build a separate set of allocation and
5269 deallocation routines to be used by OpenSSL, for example memory
5270 pool implementations, or something else, which was previously hard
5271 since Malloc(), Realloc() and Free() were defined as macros having
5272 the values malloc, realloc and free, respectively (except for Win32
5273 compilations). The same is provided for memory debugging code.
5274 OpenSSL already comes with functionality to find memory leaks, but
5275 this gives people a chance to debug other memory problems.
5277 With these changes, a new set of functions and macros have appeared:
5279 CRYPTO_set_mem_debug_functions() [F]
5280 CRYPTO_get_mem_debug_functions() [F]
5281 CRYPTO_dbg_set_options() [F]
5282 CRYPTO_dbg_get_options() [F]
5283 CRYPTO_malloc_debug_init() [M]
5285 The memory debug functions are NULL by default, unless the library
5286 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5287 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5288 gives the standard debugging functions that come with OpenSSL) or
5289 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5290 provided by the library user) must be used. When the standard
5291 debugging functions are used, CRYPTO_dbg_set_options can be used to
5292 request additional information:
5293 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5294 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5296 Also, things like CRYPTO_set_mem_functions will always give the
5297 expected result (the new set of functions is used for allocation
5298 and deallocation) at all times, regardless of platform and compiler
5301 To finish it up, some functions that were never use in any other
5302 way than through macros have a new API and new semantic:
5305 CRYPTO_dbg_realloc()
5308 All macros of value have retained their old syntax.
5309 [Richard Levitte and Bodo Moeller]
5311 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5312 ordering of SMIMECapabilities wasn't in "strength order" and there
5313 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5317 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5318 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5319 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5321 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5322 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5323 functionality to handle multipart/signed properly) and a utility
5324 called 'smime' to call all this stuff. This is based on code I
5325 originally wrote for Celo who have kindly allowed it to be
5326 included in OpenSSL.
5329 *) Add variants des_set_key_checked and des_set_key_unchecked of
5330 des_set_key (aka des_key_sched). Global variable des_check_key
5331 decides which of these is called by des_set_key; this way
5332 des_check_key behaves as it always did, but applications and
5333 the library itself, which was buggy for des_check_key == 1,
5334 have a cleaner way to pick the version they need.
5337 *) New function PKCS12_newpass() which changes the password of a
5341 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5342 dynamic mix. In both cases the ids can be used as an index into the
5343 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5344 functions so they accept a list of the field values and the
5345 application doesn't need to directly manipulate the X509_TRUST
5349 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5353 *) Modify the way the V3 extension code looks up extensions. This now
5354 works in a similar way to the object code: we have some "standard"
5355 extensions in a static table which is searched with OBJ_bsearch()
5356 and the application can add dynamic ones if needed. The file
5357 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5358 updated whenever a new extension is added to the core code and kept
5359 in ext_nid order. There is a simple program 'tabtest.c' which checks
5360 this. New extensions are not added too often so this file can readily
5361 be maintained manually.
5363 There are two big advantages in doing things this way. The extensions
5364 can be looked up immediately and no longer need to be "added" using
5365 X509V3_add_standard_extensions(): this function now does nothing.
5366 [Side note: I get *lots* of email saying the extension code doesn't
5367 work because people forget to call this function]
5368 Also no dynamic allocation is done unless new extensions are added:
5369 so if we don't add custom extensions there is no need to call
5370 X509V3_EXT_cleanup().
5373 *) Modify enc utility's salting as follows: make salting the default. Add a
5374 magic header, so unsalted files fail gracefully instead of just decrypting
5375 to garbage. This is because not salting is a big security hole, so people
5376 should be discouraged from doing it.
5379 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5380 digest to be passed on the command line but it only used this
5381 parameter when signing a certificate. Modified so all relevant
5382 operations are affected by the digest parameter including the
5383 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5384 DSA key was used because it didn't fix the digest.
5387 *) Initial certificate chain verify code. Currently tests the untrusted
5388 certificates for consistency with the verify purpose (which is set
5389 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5391 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5392 this is because it will reject chains with invalid extensions whereas
5393 every previous version of OpenSSL and SSLeay made no checks at all.
5395 Trust code: checks the root CA for the relevant trust settings. Trust
5396 settings have an initial value consistent with the verify purpose: e.g.
5397 if the verify purpose is for SSL client use it expects the CA to be
5398 trusted for SSL client use. However the default value can be changed to
5399 permit custom trust settings: one example of this would be to only trust
5400 certificates from a specific "secure" set of CAs.
5402 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5403 which should be used for version portability: especially since the
5404 verify structure is likely to change more often now.
5406 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5407 to set them. If not set then assume SSL clients will verify SSL servers
5410 Two new options to the verify program: -untrusted allows a set of
5411 untrusted certificates to be passed in and -purpose which sets the
5412 intended purpose of the certificate. If a purpose is set then the
5413 new chain verify code is used to check extension consistency.
5416 *) Support for the authority information access extension.
5419 *) Modify RSA and DSA PEM read routines to transparently handle
5420 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5421 public keys in a format compatible with certificate
5422 SubjectPublicKeyInfo structures. Unfortunately there were already
5423 functions called *_PublicKey_* which used various odd formats so
5424 these are retained for compatibility: however the DSA variants were
5425 never in a public release so they have been deleted. Changed dsa/rsa
5426 utilities to handle the new format: note no releases ever handled public
5427 keys so we should be OK.
5429 The primary motivation for this change is to avoid the same fiasco
5430 that dogs private keys: there are several incompatible private key
5431 formats some of which are standard and some OpenSSL specific and
5432 require various evil hacks to allow partial transparent handling and
5433 even then it doesn't work with DER formats. Given the option anything
5434 other than PKCS#8 should be dumped: but the other formats have to
5435 stay in the name of compatibility.
5437 With public keys and the benefit of hindsight one standard format
5438 is used which works with EVP_PKEY, RSA or DSA structures: though
5439 it clearly returns an error if you try to read the wrong kind of key.
5441 Added a -pubkey option to the 'x509' utility to output the public key.
5442 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5443 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5444 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5445 that do the same as the EVP_PKEY_assign_*() except they up the
5446 reference count of the added key (they don't "swallow" the
5450 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5451 CRLs would fail if the file contained no certificates or no CRLs:
5452 added a new function to read in both types and return the number
5453 read: this means that if none are read it will be an error. The
5454 DER versions of the certificate and CRL reader would always fail
5455 because it isn't possible to mix certificates and CRLs in DER format
5456 without choking one or the other routine. Changed this to just read
5457 a certificate: this is the best we can do. Also modified the code
5458 in apps/verify.c to take notice of return codes: it was previously
5459 attempting to read in certificates from NULL pointers and ignoring
5460 any errors: this is one reason why the cert and CRL reader seemed
5461 to work. It doesn't check return codes from the default certificate
5462 routines: these may well fail if the certificates aren't installed.
5465 *) Code to support otherName option in GeneralName.
5468 *) First update to verify code. Change the verify utility
5469 so it warns if it is passed a self signed certificate:
5470 for consistency with the normal behaviour. X509_verify
5471 has been modified to it will now verify a self signed
5472 certificate if *exactly* the same certificate appears
5473 in the store: it was previously impossible to trust a
5474 single self signed certificate. This means that:
5475 openssl verify ss.pem
5476 now gives a warning about a self signed certificate but
5477 openssl verify -CAfile ss.pem ss.pem
5481 *) For servers, store verify_result in SSL_SESSION data structure
5482 (and add it to external session representation).
5483 This is needed when client certificate verifications fails,
5484 but an application-provided verification callback (set by
5485 SSL_CTX_set_cert_verify_callback) allows accepting the session
5486 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5487 but returns 1): When the session is reused, we have to set
5488 ssl->verify_result to the appropriate error code to avoid
5490 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5492 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5493 case in PKCS7_dataInit() where the signed PKCS7 structure
5494 didn't contain any existing data because it was being created.
5495 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5497 *) Add a salt to the key derivation routines in enc.c. This
5498 forms the first 8 bytes of the encrypted file. Also add a
5499 -S option to allow a salt to be input on the command line.
5502 *) New function X509_cmp(). Oddly enough there wasn't a function
5503 to compare two certificates. We do this by working out the SHA1
5504 hash and comparing that. X509_cmp() will be needed by the trust
5508 *) SSL_get1_session() is like SSL_get_session(), but increments
5509 the reference count in the SSL_SESSION returned.
5510 [Geoff Thorpe <geoff@eu.c2.net>]
5512 *) Fix for 'req': it was adding a null to request attributes.
5513 Also change the X509_LOOKUP and X509_INFO code to handle
5514 certificate auxiliary information.
5517 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5521 *) Add the possibility to add extra information to the memory leak
5522 detecting output, to form tracebacks, showing from where each
5523 allocation was originated: CRYPTO_push_info("constant string") adds
5524 the string plus current file name and line number to a per-thread
5525 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5526 is like calling CYRPTO_pop_info() until the stack is empty.
5527 Also updated memory leak detection code to be multi-thread-safe.
5530 *) Add options -text and -noout to pkcs7 utility and delete the
5531 encryption options which never did anything. Update docs.
5534 *) Add options to some of the utilities to allow the pass phrase
5535 to be included on either the command line (not recommended on
5536 OSes like Unix) or read from the environment. Update the
5537 manpages and fix a few bugs.
5540 *) Add a few manpages for some of the openssl commands.
5543 *) Fix the -revoke option in ca. It was freeing up memory twice,
5544 leaking and not finding already revoked certificates.
5547 *) Extensive changes to support certificate auxiliary information.
5548 This involves the use of X509_CERT_AUX structure and X509_AUX
5549 functions. An X509_AUX function such as PEM_read_X509_AUX()
5550 can still read in a certificate file in the usual way but it
5551 will also read in any additional "auxiliary information". By
5552 doing things this way a fair degree of compatibility can be
5553 retained: existing certificates can have this information added
5554 using the new 'x509' options.
5556 Current auxiliary information includes an "alias" and some trust
5557 settings. The trust settings will ultimately be used in enhanced
5558 certificate chain verification routines: currently a certificate
5559 can only be trusted if it is self signed and then it is trusted
5563 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5564 The problem was that one of the replacement routines had not been working
5565 since SSLeay releases. For now the offending routine has been replaced
5566 with non-optimised assembler. Even so, this now gives around 95%
5567 performance improvement for 1024 bit RSA signs.
5570 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5571 handling. Most clients have the effective key size in bits equal to
5572 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5573 A few however don't do this and instead use the size of the decrypted key
5574 to determine the RC2 key length and the AlgorithmIdentifier to determine
5575 the effective key length. In this case the effective key length can still
5576 be 40 bits but the key length can be 168 bits for example. This is fixed
5577 by manually forcing an RC2 key into the EVP_PKEY structure because the
5578 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5579 the key length and effective key length are equal.
5582 *) Add a bunch of functions that should simplify the creation of
5583 X509_NAME structures. Now you should be able to do:
5584 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5585 and have it automatically work out the correct field type and fill in
5586 the structures. The more adventurous can try:
5587 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5588 and it will (hopefully) work out the correct multibyte encoding.
5591 *) Change the 'req' utility to use the new field handling and multibyte
5592 copy routines. Before the DN field creation was handled in an ad hoc
5593 way in req, ca, and x509 which was rather broken and didn't support
5594 BMPStrings or UTF8Strings. Since some software doesn't implement
5595 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5596 using the dirstring_type option. See the new comment in the default
5597 openssl.cnf for more info.
5600 *) Make crypto/rand/md_rand.c more robust:
5601 - Assure unique random numbers after fork().
5602 - Make sure that concurrent threads access the global counter and
5603 md serializably so that we never lose entropy in them
5604 or use exactly the same state in multiple threads.
5605 Access to the large state is not always serializable because
5606 the additional locking could be a performance killer, and
5607 md should be large enough anyway.
5610 *) New file apps/app_rand.c with commonly needed functionality
5611 for handling the random seed file.
5613 Use the random seed file in some applications that previously did not:
5615 dsaparam -genkey (which also ignored its '-rand' option),
5618 x509 (when signing).
5619 Except on systems with /dev/urandom, it is crucial to have a random
5620 seed file at least for key creation, DSA signing, and for DH exchanges;
5621 for RSA signatures we could do without one.
5623 gendh and gendsa (unlike genrsa) used to read only the first byte
5624 of each file listed in the '-rand' option. The function as previously
5625 found in genrsa is now in app_rand.c and is used by all programs
5626 that support '-rand'.
5629 *) In RAND_write_file, use mode 0600 for creating files;
5630 don't just chmod when it may be too late.
5633 *) Report an error from X509_STORE_load_locations
5634 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5637 *) New function ASN1_mbstring_copy() this copies a string in either
5638 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5639 into an ASN1_STRING type. A mask of permissible types is passed
5640 and it chooses the "minimal" type to use or an error if not type
5644 *) Add function equivalents to the various macros in asn1.h. The old
5645 macros are retained with an M_ prefix. Code inside the library can
5646 use the M_ macros. External code (including the openssl utility)
5647 should *NOT* in order to be "shared library friendly".
5650 *) Add various functions that can check a certificate's extensions
5651 to see if it usable for various purposes such as SSL client,
5652 server or S/MIME and CAs of these types. This is currently
5653 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5654 verification. Also added a -purpose flag to x509 utility to
5655 print out all the purposes.
5658 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5662 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5663 for, obtain and decode and extension and obtain its critical flag.
5664 This allows all the necessary extension code to be handled in a
5665 single function call.
5668 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5669 platforms. See crypto/rc4/rc4_enc.c for further details.
5672 *) New -noout option to asn1parse. This causes no output to be produced
5673 its main use is when combined with -strparse and -out to extract data
5674 from a file (which may not be in ASN.1 format).
5677 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5678 when producing the local key id.
5679 [Richard Levitte <levitte@stacken.kth.se>]
5681 *) New option -dhparam in s_server. This allows a DH parameter file to be
5682 stated explicitly. If it is not stated then it tries the first server
5683 certificate file. The previous behaviour hard coded the filename
5687 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5688 a public key to be input or output. For example:
5689 openssl rsa -in key.pem -pubout -out pubkey.pem
5690 Also added necessary DSA public key functions to handle this.
5693 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5694 in the message. This was handled by allowing
5695 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5696 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5698 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5699 to the end of the strings whereas this didn't. This would cause problems
5700 if strings read with d2i_ASN1_bytes() were later modified.
5701 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5703 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5704 data and it contains EOF it will end up returning an error. This is
5705 caused by input 46 bytes long. The cause is due to the way base64
5706 BIOs find the start of base64 encoded data. They do this by trying a
5707 trial decode on each line until they find one that works. When they
5708 do a flag is set and it starts again knowing it can pass all the
5709 data directly through the decoder. Unfortunately it doesn't reset
5710 the context it uses. This means that if EOF is reached an attempt
5711 is made to pass two EOFs through the context and this causes the
5712 resulting error. This can also cause other problems as well. As is
5713 usual with these problems it takes *ages* to find and the fix is
5714 trivial: move one line.
5715 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5717 *) Ugly workaround to get s_client and s_server working under Windows. The
5718 old code wouldn't work because it needed to select() on sockets and the
5719 tty (for keypresses and to see if data could be written). Win32 only
5720 supports select() on sockets so we select() with a 1s timeout on the
5721 sockets and then see if any characters are waiting to be read, if none
5722 are present then we retry, we also assume we can always write data to
5723 the tty. This isn't nice because the code then blocks until we've
5724 received a complete line of data and it is effectively polling the
5725 keyboard at 1s intervals: however it's quite a bit better than not
5726 working at all :-) A dedicated Windows application might handle this
5727 with an event loop for example.
5730 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5731 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5732 will be called when RSA_sign() and RSA_verify() are used. This is useful
5733 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5734 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5735 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5736 This necessitated the support of an extra signature type NID_md5_sha1
5737 for SSL signatures and modifications to the SSL library to use it instead
5738 of calling RSA_public_decrypt() and RSA_private_encrypt().
5741 *) Add new -verify -CAfile and -CApath options to the crl program, these
5742 will lookup a CRL issuers certificate and verify the signature in a
5743 similar way to the verify program. Tidy up the crl program so it
5744 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5745 less strict. It will now permit CRL extensions even if it is not
5746 a V2 CRL: this will allow it to tolerate some broken CRLs.
5749 *) Initialize all non-automatic variables each time one of the openssl
5750 sub-programs is started (this is necessary as they may be started
5751 multiple times from the "OpenSSL>" prompt).
5752 [Lennart Bang, Bodo Moeller]
5754 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5755 removing all other RSA functionality (this is what NO_RSA does). This
5756 is so (for example) those in the US can disable those operations covered
5757 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5761 *) Non-copying interface to BIO pairs.
5762 (still largely untested)
5765 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
5766 ASCII string. This was handled independently in various places before.
5769 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
5770 UTF8 strings a character at a time.
5773 *) Use client_version from client hello to select the protocol
5774 (s23_srvr.c) and for RSA client key exchange verification
5775 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
5778 *) Add various utility functions to handle SPKACs, these were previously
5779 handled by poking round in the structure internals. Added new function
5780 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
5781 print, verify and generate SPKACs. Based on an original idea from
5782 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
5785 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
5788 *) Allow the config file extension section to be overwritten on the
5789 command line. Based on an original idea from Massimiliano Pala
5790 <madwolf@comune.modena.it>. The new option is called -extensions
5791 and can be applied to ca, req and x509. Also -reqexts to override
5792 the request extensions in req and -crlexts to override the crl extensions
5796 *) Add new feature to the SPKAC handling in ca. Now you can include
5797 the same field multiple times by preceding it by "XXXX." for example:
5800 this is the same syntax as used in the req config file.
5803 *) Allow certificate extensions to be added to certificate requests. These
5804 are specified in a 'req_extensions' option of the req section of the
5805 config file. They can be printed out with the -text option to req but
5806 are otherwise ignored at present.
5809 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
5810 data read consists of only the final block it would not decrypted because
5811 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
5812 A misplaced 'break' also meant the decrypted final block might not be
5813 copied until the next read.
5816 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
5817 a few extra parameters to the DH structure: these will be useful if
5818 for example we want the value of 'q' or implement X9.42 DH.
5821 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
5822 provides hooks that allow the default DSA functions or functions on a
5823 "per key" basis to be replaced. This allows hardware acceleration and
5824 hardware key storage to be handled without major modification to the
5825 library. Also added low level modexp hooks and CRYPTO_EX structure and
5826 associated functions.
5829 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
5830 as "read only": it can't be written to and the buffer it points to will
5831 not be freed. Reading from a read only BIO is much more efficient than
5832 a normal memory BIO. This was added because there are several times when
5833 an area of memory needs to be read from a BIO. The previous method was
5834 to create a memory BIO and write the data to it, this results in two
5835 copies of the data and an O(n^2) reading algorithm. There is a new
5836 function BIO_new_mem_buf() which creates a read only memory BIO from
5837 an area of memory. Also modified the PKCS#7 routines to use read only
5841 *) Bugfix: ssl23_get_client_hello did not work properly when called in
5842 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
5843 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
5844 but a retry condition occured while trying to read the rest.
5847 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
5848 NID_pkcs7_encrypted by default: this was wrong since this should almost
5849 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
5850 the encrypted data type: this is a more sensible place to put it and it
5851 allows the PKCS#12 code to be tidied up that duplicated this
5855 *) Changed obj_dat.pl script so it takes its input and output files on
5856 the command line. This should avoid shell escape redirection problems
5860 *) Initial support for certificate extension requests, these are included
5861 in things like Xenroll certificate requests. Included functions to allow
5862 extensions to be obtained and added.
5865 *) -crlf option to s_client and s_server for sending newlines as
5866 CRLF (as required by many protocols).
5869 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
5871 *) Install libRSAglue.a when OpenSSL is built with RSAref.
5872 [Ralf S. Engelschall]
5874 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
5875 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
5877 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
5881 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
5882 DH parameters/keys (q is lost during that conversion, but the resulting
5883 DH parameters contain its length).
5885 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
5886 much faster than DH_generate_parameters (which creates parameters
5887 where p = 2*q + 1), and also the smaller q makes DH computations
5888 much more efficient (160-bit exponentiation instead of 1024-bit
5889 exponentiation); so this provides a convenient way to support DHE
5890 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
5891 utter importance to use
5892 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5894 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
5895 when such DH parameters are used, because otherwise small subgroup
5896 attacks may become possible!
5899 *) Avoid memory leak in i2d_DHparams.
5902 *) Allow the -k option to be used more than once in the enc program:
5903 this allows the same encrypted message to be read by multiple recipients.
5906 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
5907 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
5908 it will always use the numerical form of the OID, even if it has a short
5912 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
5913 method only got called if p,q,dmp1,dmq1,iqmp components were present,
5914 otherwise bn_mod_exp was called. In the case of hardware keys for example
5915 no private key components need be present and it might store extra data
5916 in the RSA structure, which cannot be accessed from bn_mod_exp.
5917 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
5918 private key operations.
5921 *) Added support for SPARC Linux.
5924 *) pem_password_cb function type incompatibly changed from
5925 typedef int pem_password_cb(char *buf, int size, int rwflag);
5927 ....(char *buf, int size, int rwflag, void *userdata);
5928 so that applications can pass data to their callbacks:
5929 The PEM[_ASN1]_{read,write}... functions and macros now take an
5930 additional void * argument, which is just handed through whenever
5931 the password callback is called.
5932 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
5934 New function SSL_CTX_set_default_passwd_cb_userdata.
5936 Compatibility note: As many C implementations push function arguments
5937 onto the stack in reverse order, the new library version is likely to
5938 interoperate with programs that have been compiled with the old
5939 pem_password_cb definition (PEM_whatever takes some data that
5940 happens to be on the stack as its last argument, and the callback
5941 just ignores this garbage); but there is no guarantee whatsoever that
5944 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
5945 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
5946 problems not only on Windows, but also on some Unix platforms.
5947 To avoid problematic command lines, these definitions are now in an
5948 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
5949 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
5952 *) MIPS III/IV assembler module is reimplemented.
5955 *) More DES library cleanups: remove references to srand/rand and
5956 delete an unused file.
5959 *) Add support for the the free Netwide assembler (NASM) under Win32,
5960 since not many people have MASM (ml) and it can be hard to obtain.
5961 This is currently experimental but it seems to work OK and pass all
5962 the tests. Check out INSTALL.W32 for info.
5965 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
5966 without temporary keys kept an extra copy of the server key,
5967 and connections with temporary keys did not free everything in case
5971 *) New function RSA_check_key and new openssl rsa option -check
5972 for verifying the consistency of RSA keys.
5973 [Ulf Moeller, Bodo Moeller]
5975 *) Various changes to make Win32 compile work:
5976 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
5977 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
5978 comparison" warnings.
5979 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
5982 *) Add a debugging option to PKCS#5 v2 key generation function: when
5983 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
5984 derived keys are printed to stderr.
5987 *) Copy the flags in ASN1_STRING_dup().
5988 [Roman E. Pavlov <pre@mo.msk.ru>]
5990 *) The x509 application mishandled signing requests containing DSA
5991 keys when the signing key was also DSA and the parameters didn't match.
5993 It was supposed to omit the parameters when they matched the signing key:
5994 the verifying software was then supposed to automatically use the CA's
5995 parameters if they were absent from the end user certificate.
5997 Omitting parameters is no longer recommended. The test was also
5998 the wrong way round! This was probably due to unusual behaviour in
5999 EVP_cmp_parameters() which returns 1 if the parameters match.
6000 This meant that parameters were omitted when they *didn't* match and
6001 the certificate was useless. Certificates signed with 'ca' didn't have
6003 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6005 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6006 The interface is as follows:
6007 Applications can use
6008 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6009 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6010 "off" is now the default.
6011 The library internally uses
6012 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6013 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6014 to disable memory-checking temporarily.
6016 Some inconsistent states that previously were possible (and were
6017 even the default) are now avoided.
6019 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6020 with each memory chunk allocated; this is occasionally more helpful
6021 than just having a counter.
6023 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6025 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6029 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6030 which largely parallels "options", but is for changing API behaviour,
6031 whereas "options" are about protocol behaviour.
6032 Initial "mode" flags are:
6034 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6035 a single record has been written.
6036 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6037 retries use the same buffer location.
6038 (But all of the contents must be
6042 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6045 *) Fix problems with no-hmac etc.
6046 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6048 *) New functions RSA_get_default_method(), RSA_set_method() and
6049 RSA_get_method(). These allows replacement of RSA_METHODs without having
6050 to mess around with the internals of an RSA structure.
6053 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6054 Also really enable memory leak checks in openssl.c and in some
6056 [Chad C. Mulligan, Bodo Moeller]
6058 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6059 up the length of negative integers. This has now been simplified to just
6060 store the length when it is first determined and use it later, rather
6061 than trying to keep track of where data is copied and updating it to
6063 [Steve Henson, reported by Brien Wheeler
6064 <bwheeler@authentica-security.com>]
6066 *) Add a new function PKCS7_signatureVerify. This allows the verification
6067 of a PKCS#7 signature but with the signing certificate passed to the
6068 function itself. This contrasts with PKCS7_dataVerify which assumes the
6069 certificate is present in the PKCS#7 structure. This isn't always the
6070 case: certificates can be omitted from a PKCS#7 structure and be
6071 distributed by "out of band" means (such as a certificate database).
6074 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6075 function prototypes in pem.h, also change util/mkdef.pl to add the
6076 necessary function names.
6079 *) mk1mf.pl (used by Windows builds) did not properly read the
6080 options set by Configure in the top level Makefile, and Configure
6081 was not even able to write more than one option correctly.
6082 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6085 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6086 file to be loaded from a BIO or FILE pointer. The BIO version will
6087 for example allow memory BIOs to contain config info.
6090 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6091 Whoever hopes to achieve shared-library compatibility across versions
6092 must use this, not the compile-time macro.
6093 (Exercise 0.9.4: Which is the minimum library version required by
6095 Note: All this applies only to multi-threaded programs, others don't
6099 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6100 through a BIO pair triggered the default case, i.e.
6101 SSLerr(...,SSL_R_UNKNOWN_STATE).
6104 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6105 can use the SSL library even if none of the specific BIOs is
6109 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6110 for the encoded length.
6111 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6113 *) Add initial documentation of the X509V3 functions.
6116 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6117 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6118 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6119 secure PKCS#8 private key format with a high iteration count.
6122 *) Fix determination of Perl interpreter: A perl or perl5
6123 _directory_ in $PATH was also accepted as the interpreter.
6124 [Ralf S. Engelschall]
6126 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6127 wrong with it but it was very old and did things like calling
6128 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6132 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6133 to use the new extension code.
6136 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6137 with macros. This should make it easier to change their form, add extra
6138 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6142 *) Add to configuration table a new entry that can specify an alternative
6143 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6144 according to Mark Crispin <MRC@Panda.COM>.
6148 *) DES CBC did not update the IV. Weird.
6151 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6152 Changing the behaviour of the former might break existing programs --
6153 where IV updating is needed, des_ncbc_encrypt can be used.
6156 *) When bntest is run from "make test" it drives bc to check its
6157 calculations, as well as internally checking them. If an internal check
6158 fails, it needs to cause bc to give a non-zero result or make test carries
6159 on without noticing the failure. Fixed.
6162 *) DES library cleanups.
6165 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6166 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6167 ciphers. NOTE: although the key derivation function has been verified
6168 against some published test vectors it has not been extensively tested
6169 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6173 *) Instead of "mkdir -p", which is not fully portable, use new
6174 Perl script "util/mkdir-p.pl".
6177 *) Rewrite the way password based encryption (PBE) is handled. It used to
6178 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6179 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6180 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6181 the 'parameter' field of the AlgorithmIdentifier is passed to the
6182 underlying key generation function so it must do its own ASN1 parsing.
6183 This has also changed the EVP_PBE_CipherInit() function which now has a
6184 'parameter' argument instead of literal salt and iteration count values
6185 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6188 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6189 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6190 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6191 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6192 value was just used as a "magic string" and not used directly its
6193 value doesn't matter.
6196 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6200 *) "linux-sparc64" configuration (ultrapenguin).
6201 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6202 "linux-sparc" configuration.
6203 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6205 *) config now generates no-xxx options for missing ciphers.
6208 *) Support the EBCDIC character set (work in progress).
6209 File ebcdic.c not yet included because it has a different license.
6210 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6212 *) Support BS2000/OSD-POSIX.
6213 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6215 *) Make callbacks for key generation use void * instead of char *.
6218 *) Make S/MIME samples compile (not yet tested).
6221 *) Additional typesafe stacks.
6224 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6228 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6230 *) New configuration variant "sco5-gcc".
6232 *) Updated some demos.
6233 [Sean O Riordain, Wade Scholine]
6235 *) Add missing BIO_free at exit of pkcs12 application.
6238 *) Fix memory leak in conf.c.
6241 *) Updates for Win32 to assembler version of MD5.
6244 *) Set #! path to perl in apps/der_chop to where we found it
6245 instead of using a fixed path.
6248 *) SHA library changes for irix64-mips4-cc.
6251 *) Improvements for VMS support.
6255 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6257 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6258 This also avoids the problems with SC4.2 and unpatched SC5.
6259 [Andy Polyakov <appro@fy.chalmers.se>]
6261 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6262 These are required because of the typesafe stack would otherwise break
6263 existing code. If old code used a structure member which used to be STACK
6264 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6265 sk_num or sk_value it would produce an error because the num, data members
6266 are not present in STACK_OF. Now it just produces a warning. sk_set
6267 replaces the old method of assigning a value to sk_value
6268 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6269 that does this will no longer work (and should use sk_set instead) but
6270 this could be regarded as a "questionable" behaviour anyway.
6273 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6274 correctly handle encrypted S/MIME data.
6277 *) Change type of various DES function arguments from des_cblock
6278 (which means, in function argument declarations, pointer to char)
6279 to des_cblock * (meaning pointer to array with 8 char elements),
6280 which allows the compiler to do more typechecking; it was like
6281 that back in SSLeay, but with lots of ugly casts.
6283 Introduce new type const_des_cblock.
6286 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6287 problems: find RecipientInfo structure that matches recipient certificate
6288 and initialise the ASN1 structures properly based on passed cipher.
6291 *) Belatedly make the BN tests actually check the results.
6294 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6295 to and from BNs: it was completely broken. New compilation option
6296 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6297 key elements as negative integers.
6300 *) Reorganize and speed up MD5.
6301 [Andy Polyakov <appro@fy.chalmers.se>]
6304 [Richard Levitte <richard@levitte.org>]
6306 *) New option -out to asn1parse to allow the parsed structure to be
6307 output to a file. This is most useful when combined with the -strparse
6308 option to examine the output of things like OCTET STRINGS.
6311 *) Make SSL library a little more fool-proof by not requiring any longer
6312 that SSL_set_{accept,connect}_state be called before
6313 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6314 in many applications because usually everything *appeared* to work as
6315 intended anyway -- now it really works as intended).
6318 *) Move openssl.cnf out of lib/.
6321 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6322 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6323 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6324 [Ralf S. Engelschall]
6326 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6327 handle PKCS#7 enveloped data properly.
6328 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6330 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6331 copying pointers. The cert_st handling is changed by this in
6332 various ways (and thus what used to be known as ctx->default_cert
6333 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6334 any longer when s->cert does not give us what we need).
6335 ssl_cert_instantiate becomes obsolete by this change.
6336 As soon as we've got the new code right (possibly it already is?),
6337 we have solved a couple of bugs of the earlier code where s->cert
6338 was used as if it could not have been shared with other SSL structures.
6340 Note that using the SSL API in certain dirty ways now will result
6341 in different behaviour than observed with earlier library versions:
6342 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6343 does not influence s as it used to.
6345 In order to clean up things more thoroughly, inside SSL_SESSION
6346 we don't use CERT any longer, but a new structure SESS_CERT
6347 that holds per-session data (if available); currently, this is
6348 the peer's certificate chain and, for clients, the server's certificate
6349 and temporary key. CERT holds only those values that can have
6350 meaningful defaults in an SSL_CTX.
6353 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6354 from the internal representation. Various PKCS#7 fixes: remove some
6355 evil casts and set the enc_dig_alg field properly based on the signing
6359 *) Allow PKCS#12 password to be set from the command line or the
6360 environment. Let 'ca' get its config file name from the environment
6361 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6365 *) Allow certificate policies extension to use an IA5STRING for the
6366 organization field. This is contrary to the PKIX definition but
6367 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6371 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6372 without disallowing inline assembler and the like for non-pedantic builds.
6375 *) Support Borland C++ builder.
6376 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6381 *) SHA-1 cleanups and performance enhancements.
6382 [Andy Polyakov <appro@fy.chalmers.se>]
6384 *) Sparc v8plus assembler for the bignum library.
6385 [Andy Polyakov <appro@fy.chalmers.se>]
6387 *) Accept any -xxx and +xxx compiler options in Configure.
6390 *) Update HPUX configuration.
6393 *) Add missing sk_<type>_unshift() function to safestack.h
6394 [Ralf S. Engelschall]
6396 *) New function SSL_CTX_use_certificate_chain_file that sets the
6397 "extra_cert"s in addition to the certificate. (This makes sense
6398 only for "PEM" format files, as chains as a whole are not
6402 *) Support verify_depth from the SSL API.
6403 x509_vfy.c had what can be considered an off-by-one-error:
6404 Its depth (which was not part of the external interface)
6405 was actually counting the number of certificates in a chain;
6406 now it really counts the depth.
6409 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6410 instead of X509err, which often resulted in confusing error
6411 messages since the error codes are not globally unique
6412 (e.g. an alleged error in ssl3_accept when a certificate
6413 didn't match the private key).
6415 *) New function SSL_CTX_set_session_id_context that allows to set a default
6416 value (so that you don't need SSL_set_session_id_context for each
6417 connection using the SSL_CTX).
6420 *) OAEP decoding bug fix.
6423 *) Support INSTALL_PREFIX for package builders, as proposed by
6427 *) New Configure options "threads" and "no-threads". For systems
6428 where the proper compiler options are known (currently Solaris
6429 and Linux), "threads" is the default.
6432 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6435 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6436 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6437 such as /usr/local/bin.
6440 *) "make linux-shared" to build shared libraries.
6441 [Niels Poppe <niels@netbox.org>]
6443 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6446 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6447 extension adding in x509 utility.
6450 *) Remove NOPROTO sections and error code comments.
6453 *) Partial rewrite of the DEF file generator to now parse the ANSI
6457 *) New Configure options --prefix=DIR and --openssldir=DIR.
6460 *) Complete rewrite of the error code script(s). It is all now handled
6461 by one script at the top level which handles error code gathering,
6462 header rewriting and C source file generation. It should be much better
6463 than the old method: it now uses a modified version of Ulf's parser to
6464 read the ANSI prototypes in all header files (thus the old K&R definitions
6465 aren't needed for error creation any more) and do a better job of
6466 translating function codes into names. The old 'ASN1 error code imbedded
6467 in a comment' is no longer necessary and it doesn't use .err files which
6468 have now been deleted. Also the error code call doesn't have to appear all
6469 on one line (which resulted in some large lines...).
6472 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6475 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6476 0 (which usually indicates a closed connection), but continue reading.
6479 *) Fix some race conditions.
6482 *) Add support for CRL distribution points extension. Add Certificate
6483 Policies and CRL distribution points documentation.
6486 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6489 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6490 8 of keying material. Merlin has also confirmed interop with this fix
6491 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6492 [Merlin Hughes <merlin@baltimore.ie>]
6494 *) Fix lots of warnings.
6495 [Richard Levitte <levitte@stacken.kth.se>]
6497 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6498 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6499 [Richard Levitte <levitte@stacken.kth.se>]
6501 *) Fix problems with sizeof(long) == 8.
6502 [Andy Polyakov <appro@fy.chalmers.se>]
6504 *) Change functions to ANSI C.
6507 *) Fix typos in error codes.
6508 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6510 *) Remove defunct assembler files from Configure.
6513 *) SPARC v8 assembler BIGNUM implementation.
6514 [Andy Polyakov <appro@fy.chalmers.se>]
6516 *) Support for Certificate Policies extension: both print and set.
6517 Various additions to support the r2i method this uses.
6520 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6521 return a const string when you are expecting an allocated buffer.
6524 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6525 types DirectoryString and DisplayText.
6528 *) Add code to allow r2i extensions to access the configuration database,
6529 add an LHASH database driver and add several ctx helper functions.
6532 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6533 fail when they extended the size of a BIGNUM.
6536 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6537 support typesafe stack.
6540 *) Fix typo in SSL_[gs]et_options().
6541 [Nils Frostberg <nils@medcom.se>]
6543 *) Delete various functions and files that belonged to the (now obsolete)
6544 old X509V3 handling code.
6547 *) New Configure option "rsaref".
6550 *) Don't auto-generate pem.h.
6553 *) Introduce type-safe ASN.1 SETs.
6556 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6557 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6559 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6560 that links with OpenSSL (well at least cause lots of warnings), but fear
6561 not: the conversion is trivial, and it eliminates loads of evil casts. A
6562 few STACKed things have been converted already. Feel free to convert more.
6563 In the fullness of time, I'll do away with the STACK type altogether.
6566 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6567 specified in <certfile> by updating the entry in the index.txt file.
6568 This way one no longer has to edit the index.txt file manually for
6569 revoking a certificate. The -revoke option does the gory details now.
6570 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6572 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6573 `-text' option at all and this way the `-noout -text' combination was
6574 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6575 [Ralf S. Engelschall]
6577 *) Make sure a corresponding plain text error message exists for the
6578 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6579 verify callback function determined that a certificate was revoked.
6580 [Ralf S. Engelschall]
6582 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6583 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6584 all available cipers including rc5, which was forgotten until now.
6585 In order to let the testing shell script know which algorithms
6586 are available, a new (up to now undocumented) command
6587 "openssl list-cipher-commands" is used.
6590 *) Bugfix: s_client occasionally would sleep in select() when
6591 it should have checked SSL_pending() first.
6594 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6595 the raw DSA values prior to ASN.1 encoding.
6598 *) Tweaks to Configure
6599 [Niels Poppe <niels@netbox.org>]
6601 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6605 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6608 *) New config option to avoid instructions that are illegal on the 80386.
6609 The default code is faster, but requires at least a 486.
6612 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6613 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6614 same as SSL2_VERSION anyway.
6617 *) New "-showcerts" option for s_client.
6620 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6621 application. Various cleanups and fixes.
6624 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6625 modify error routines to work internally. Add error codes and PBE init
6626 to library startup routines.
6629 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6630 packing functions to asn1 and evp. Changed function names and error
6631 codes along the way.
6634 *) PKCS12 integration: and so it begins... First of several patches to
6635 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6636 objects to objects.h
6639 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6640 and display support for Thawte strong extranet extension.
6643 *) Add LinuxPPC support.
6644 [Jeff Dubrule <igor@pobox.org>]
6646 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6647 bn_div_words in alpha.s.
6648 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6650 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6651 OAEP isn't supported when OpenSSL is built with RSAref.
6652 [Ulf Moeller <ulf@fitug.de>]
6654 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6655 so they no longer are missing under -DNOPROTO.
6656 [Soren S. Jorvang <soren@t.dk>]
6659 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6661 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6662 doesn't work when the session is reused. Coming soon!
6665 *) Fix a security hole, that allows sessions to be reused in the wrong
6666 context thus bypassing client cert protection! All software that uses
6667 client certs and session caches in multiple contexts NEEDS PATCHING to
6668 allow session reuse! A fuller solution is in the works.
6669 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6671 *) Some more source tree cleanups (removed obsolete files
6672 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6673 permission on "config" script to be executable) and a fix for the INSTALL
6675 [Ulf Moeller <ulf@fitug.de>]
6677 *) Remove some legacy and erroneous uses of malloc, free instead of
6679 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6681 *) Make rsa_oaep_test return non-zero on error.
6682 [Ulf Moeller <ulf@fitug.de>]
6684 *) Add support for native Solaris shared libraries. Configure
6685 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6686 if someone would make that last step automatic.
6687 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6689 *) ctx_size was not built with the right compiler during "make links". Fixed.
6692 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6693 except NULL ciphers". This means the default cipher list will no longer
6694 enable NULL ciphers. They need to be specifically enabled e.g. with
6695 the string "DEFAULT:eNULL".
6698 *) Fix to RSA private encryption routines: if p < q then it would
6699 occasionally produce an invalid result. This will only happen with
6700 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6703 *) Be less restrictive and allow also `perl util/perlpath.pl
6704 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6705 because this way one can also use an interpreter named `perl5' (which is
6706 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6707 installed as `perl').
6708 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6710 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6711 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6713 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6714 advapi32.lib to Win32 build and change the pem test comparision
6715 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6716 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6717 and crypto/des/ede_cbcm_enc.c.
6720 *) DES quad checksum was broken on big-endian architectures. Fixed.
6723 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6724 Win32 test batch file so it (might) work again. The Win32 test batch file
6725 is horrible: I feel ill....
6728 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6729 in e_os.h. Audit of header files to check ANSI and non ANSI
6730 sections: 10 functions were absent from non ANSI section and not exported
6731 from Windows DLLs. Fixed up libeay.num for new functions.
6734 *) Make `openssl version' output lines consistent.
6735 [Ralf S. Engelschall]
6737 *) Fix Win32 symbol export lists for BIO functions: Added
6738 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6739 to ms/libeay{16,32}.def.
6740 [Ralf S. Engelschall]
6742 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6743 fine under Unix and passes some trivial tests I've now added. But the
6744 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6745 added to make sure no one expects that this stuff really works in the
6746 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6747 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6749 [Ralf S. Engelschall]
6751 *) Fix the generation of two part addresses in perl.
6752 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6754 *) Add config entry for Linux on MIPS.
6755 [John Tobey <jtobey@channel1.com>]
6757 *) Make links whenever Configure is run, unless we are on Windoze.
6760 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6761 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6765 *) Add a useful kludge to allow package maintainers to specify compiler and
6766 other platforms details on the command line without having to patch the
6767 Configure script everytime: One now can use ``perl Configure
6768 <id>:<details>'', i.e. platform ids are allowed to have details appended
6769 to them (seperated by colons). This is treated as there would be a static
6770 pre-configured entry in Configure's %table under key <id> with value
6771 <details> and ``perl Configure <id>'' is called. So, when you want to
6772 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
6773 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
6774 now, which overrides the FreeBSD-elf entry on-the-fly.
6775 [Ralf S. Engelschall]
6777 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
6780 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
6781 on the `perl Configure ...' command line. This way one can compile
6782 OpenSSL libraries with Position Independent Code (PIC) which is needed
6783 for linking it into DSOs.
6784 [Ralf S. Engelschall]
6786 *) Remarkably, export ciphers were totally broken and no-one had noticed!
6790 *) Cleaned up the LICENSE document: The official contact for any license
6791 questions now is the OpenSSL core team under openssl-core@openssl.org.
6792 And add a paragraph about the dual-license situation to make sure people
6793 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
6794 to the OpenSSL toolkit.
6795 [Ralf S. Engelschall]
6797 *) General source tree makefile cleanups: Made `making xxx in yyy...'
6798 display consistent in the source tree and replaced `/bin/rm' by `rm'.
6799 Additonally cleaned up the `make links' target: Remove unnecessary
6800 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
6801 to speed processing and no longer clutter the display with confusing
6802 stuff. Instead only the actually done links are displayed.
6803 [Ralf S. Engelschall]
6805 *) Permit null encryption ciphersuites, used for authentication only. It used
6806 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
6807 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
6811 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
6812 signed attributes when verifying signatures (this would break them),
6813 the detached data encoding was wrong and public keys obtained using
6814 X509_get_pubkey() weren't freed.
6817 *) Add text documentation for the BUFFER functions. Also added a work around
6818 to a Win95 console bug. This was triggered by the password read stuff: the
6819 last character typed gets carried over to the next fread(). If you were
6820 generating a new cert request using 'req' for example then the last
6821 character of the passphrase would be CR which would then enter the first
6825 *) Added the new `Includes OpenSSL Cryptography Software' button as
6826 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
6827 button and can be used by applications based on OpenSSL to show the
6828 relationship to the OpenSSL project.
6829 [Ralf S. Engelschall]
6831 *) Remove confusing variables in function signatures in files
6832 ssl/ssl_lib.c and ssl/ssl.h.
6833 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6835 *) Don't install bss_file.c under PREFIX/include/
6836 [Lennart Bong <lob@kulthea.stacken.kth.se>]
6838 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
6839 functions that return function pointers and has support for NT specific
6840 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
6841 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
6842 unsigned to signed types: this was killing the Win32 compile.
6845 *) Add new certificate file to stack functions,
6846 SSL_add_dir_cert_subjects_to_stack() and
6847 SSL_add_file_cert_subjects_to_stack(). These largely supplant
6848 SSL_load_client_CA_file(), and can be used to add multiple certs easily
6849 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
6850 This means that Apache-SSL and similar packages don't have to mess around
6851 to add as many CAs as they want to the preferred list.
6854 *) Experiment with doxygen documentation. Currently only partially applied to
6856 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
6857 openssl.doxy as the configuration file.
6860 *) Get rid of remaining C++-style comments which strict C compilers hate.
6861 [Ralf S. Engelschall, pointed out by Carlos Amengual]
6863 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
6864 compiled in by default: it has problems with large keys.
6867 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
6868 DH private keys and/or callback functions which directly correspond to
6869 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
6870 is needed for applications which have to configure certificates on a
6871 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
6873 For the RSA certificate situation is makes no difference, but
6874 for the DSA certificate situation this fixes the "no shared cipher"
6875 problem where the OpenSSL cipher selection procedure failed because the
6876 temporary keys were not overtaken from the context and the API provided
6877 no way to reconfigure them.
6878 The new functions now let applications reconfigure the stuff and they
6879 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
6880 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
6881 non-public-API function ssl_cert_instantiate() is used as a helper
6882 function and also to reduce code redundancy inside ssl_rsa.c.
6883 [Ralf S. Engelschall]
6885 *) Move s_server -dcert and -dkey options out of the undocumented feature
6886 area because they are useful for the DSA situation and should be
6887 recognized by the users.
6888 [Ralf S. Engelschall]
6890 *) Fix the cipher decision scheme for export ciphers: the export bits are
6891 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
6892 SSL_EXP_MASK. So, the original variable has to be used instead of the
6893 already masked variable.
6894 [Richard Levitte <levitte@stacken.kth.se>]
6896 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
6897 [Richard Levitte <levitte@stacken.kth.se>]
6899 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
6900 from `int' to `unsigned int' because it's a length and initialized by
6901 EVP_DigestFinal() which expects an `unsigned int *'.
6902 [Richard Levitte <levitte@stacken.kth.se>]
6904 *) Don't hard-code path to Perl interpreter on shebang line of Configure
6905 script. Instead use the usual Shell->Perl transition trick.
6906 [Ralf S. Engelschall]
6908 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
6909 (in addition to RSA certificates) to match the behaviour of `openssl dsa
6910 -noout -modulus' as it's already the case for `openssl rsa -noout
6911 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
6912 currently the public key is printed (a decision which was already done by
6913 `openssl dsa -modulus' in the past) which serves a similar purpose.
6914 Additionally the NO_RSA no longer completely removes the whole -modulus
6915 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
6917 [Ralf S. Engelschall]
6919 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
6920 BIO. See the source (crypto/evp/bio_ok.c) for more info.
6921 [Arne Ansper <arne@ats.cyber.ee>]
6923 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
6924 to be added. Now both 'req' and 'ca' can use new objects defined in the
6928 *) Add cool BIO that does syslog (or event log on NT).
6929 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
6931 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
6932 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
6933 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
6934 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
6937 *) Add preliminary config info for new extension code.
6940 *) Make RSA_NO_PADDING really use no padding.
6941 [Ulf Moeller <ulf@fitug.de>]
6943 *) Generate errors when private/public key check is done.
6946 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
6947 for some CRL extensions and new objects added.
6950 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
6951 key usage extension and fuller support for authority key id.
6954 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
6955 padding method for RSA, which is recommended for new applications in PKCS
6956 #1 v2.0 (RFC 2437, October 1998).
6957 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
6958 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
6959 against Bleichbacher's attack on RSA.
6960 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
6963 *) Updates to the new SSL compression code
6964 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6966 *) Fix so that the version number in the master secret, when passed
6967 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
6968 (because the server will not accept higher), that the version number
6969 is 0x03,0x01, not 0x03,0x00
6970 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
6972 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
6973 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
6974 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
6977 *) Support for RAW extensions where an arbitrary extension can be
6978 created by including its DER encoding. See apps/openssl.cnf for
6982 *) Make sure latest Perl versions don't interpret some generated C array
6983 code as Perl array code in the crypto/err/err_genc.pl script.
6984 [Lars Weber <3weber@informatik.uni-hamburg.de>]
6986 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
6987 not many people have the assembler. Various Win32 compilation fixes and
6988 update to the INSTALL.W32 file with (hopefully) more accurate Win32
6992 *) Modify configure script 'Configure' to automatically create crypto/date.h
6993 file under Win32 and also build pem.h from pem.org. New script
6994 util/mkfiles.pl to create the MINFO file on environments that can't do a
6995 'make files': perl util/mkfiles.pl >MINFO should work.
6998 *) Major rework of DES function declarations, in the pursuit of correctness
6999 and purity. As a result, many evil casts evaporated, and some weirdness,
7000 too. You may find this causes warnings in your code. Zapping your evil
7001 casts will probably fix them. Mostly.
7004 *) Fix for a typo in asn1.h. Bug fix to object creation script
7005 obj_dat.pl. It considered a zero in an object definition to mean
7006 "end of object": none of the objects in objects.h have any zeros
7007 so it wasn't spotted.
7008 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7010 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7011 Masking (CBCM). In the absence of test vectors, the best I have been able
7012 to do is check that the decrypt undoes the encrypt, so far. Send me test
7013 vectors if you have them.
7016 *) Correct calculation of key length for export ciphers (too much space was
7017 allocated for null ciphers). This has not been tested!
7020 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7021 message is now correct (it understands "crypto" and "ssl" on its
7022 command line). There is also now an "update" option. This will update
7023 the util/ssleay.num and util/libeay.num files with any new functions.
7025 perl util/mkdef.pl crypto ssl update
7026 it will update them.
7029 *) Overhauled the Perl interface (perl/*):
7030 - ported BN stuff to OpenSSL's different BN library
7031 - made the perl/ source tree CVS-aware
7032 - renamed the package from SSLeay to OpenSSL (the files still contain
7033 their history because I've copied them in the repository)
7034 - removed obsolete files (the test scripts will be replaced
7035 by better Test::Harness variants in the future)
7036 [Ralf S. Engelschall]
7038 *) First cut for a very conservative source tree cleanup:
7039 1. merge various obsolete readme texts into doc/ssleay.txt
7040 where we collect the old documents and readme texts.
7041 2. remove the first part of files where I'm already sure that we no
7042 longer need them because of three reasons: either they are just temporary
7043 files which were left by Eric or they are preserved original files where
7044 I've verified that the diff is also available in the CVS via "cvs diff
7045 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7046 the crypto/md/ stuff).
7047 [Ralf S. Engelschall]
7049 *) More extension code. Incomplete support for subject and issuer alt
7050 name, issuer and authority key id. Change the i2v function parameters
7051 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7052 what that's for :-) Fix to ASN1 macro which messed up
7053 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7056 *) Preliminary support for ENUMERATED type. This is largely copied from the
7060 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7061 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7063 *) Make sure `make rehash' target really finds the `openssl' program.
7064 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7066 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7067 like to hear about it if this slows down other processors.
7070 *) Add CygWin32 platform information to Configure script.
7071 [Alan Batie <batie@aahz.jf.intel.com>]
7073 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7074 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7076 *) New program nseq to manipulate netscape certificate sequences
7079 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7083 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7084 but the BN code had some problems that would cause failures when
7085 doing certificate verification and some other functions.
7086 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7088 *) Add ASN1 and PEM code to support netscape certificate sequences.
7091 *) Add ASN1 and PEM code to support netscape certificate sequences.
7094 *) Add several PKIX and private extended key usage OIDs.
7097 *) Modify the 'ca' program to handle the new extension code. Modify
7098 openssl.cnf for new extension format, add comments.
7101 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7102 and add a sample to openssl.cnf so req -x509 now adds appropriate
7106 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7107 error code, add initial support to X509_print() and x509 application.
7110 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7111 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7112 stuff is currently isolated and isn't even compiled yet.
7115 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7116 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7117 Removed the versions check from X509 routines when loading extensions:
7118 this allows certain broken certificates that don't set the version
7119 properly to be processed.
7122 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7123 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7124 can still be regenerated with "make depend".
7127 *) Spelling mistake in C version of CAST-128.
7128 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7130 *) Changes to the error generation code. The perl script err-code.pl
7131 now reads in the old error codes and retains the old numbers, only
7132 adding new ones if necessary. It also only changes the .err files if new
7133 codes are added. The makefiles have been modified to only insert errors
7134 when needed (to avoid needlessly modifying header files). This is done
7135 by only inserting errors if the .err file is newer than the auto generated
7136 C file. To rebuild all the error codes from scratch (the old behaviour)
7137 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7138 or delete all the .err files.
7141 *) CAST-128 was incorrectly implemented for short keys. The C version has
7142 been fixed, but is untested. The assembler versions are also fixed, but
7143 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7144 to regenerate it if needed.
7145 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7146 Hagino <itojun@kame.net>]
7148 *) File was opened incorrectly in randfile.c.
7149 [Ulf Möller <ulf@fitug.de>]
7151 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7152 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7153 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7154 al: it's just almost always a UTCTime. Note this patch adds new error
7155 codes so do a "make errors" if there are problems.
7158 *) Correct Linux 1 recognition in config.
7159 [Ulf Möller <ulf@fitug.de>]
7161 *) Remove pointless MD5 hash when using DSA keys in ca.
7162 [Anonymous <nobody@replay.com>]
7164 *) Generate an error if given an empty string as a cert directory. Also
7165 generate an error if handed NULL (previously returned 0 to indicate an
7166 error, but didn't set one).
7167 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7169 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7172 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7173 parameters. This was causing a warning which killed off the Win32 compile.
7176 *) Remove C++ style comments from crypto/bn/bn_local.h.
7177 [Neil Costigan <neil.costigan@celocom.com>]
7179 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7180 based on a text string, looking up short and long names and finally
7181 "dot" format. The "dot" format stuff didn't work. Added new function
7182 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7183 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7184 OID is not part of the table.
7187 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7188 X509_LOOKUP_by_alias().
7191 *) Sort openssl functions by name.
7194 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7195 encryption from sample DSA keys (in case anyone is interested the password
7199 *) Make _all_ *_free functions accept a NULL pointer.
7200 [Frans Heymans <fheymans@isaserver.be>]
7202 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7204 [Anonymous <nobody@replay.com>]
7206 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7207 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7209 *) Don't blow it for numeric -newkey arguments to apps/req.
7210 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7212 *) Temp key "for export" tests were wrong in s3_srvr.c.
7213 [Anonymous <nobody@replay.com>]
7215 *) Add prototype for temp key callback functions
7216 SSL_CTX_set_tmp_{rsa,dh}_callback().
7219 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7220 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7223 *) X509_name_add_entry() freed the wrong thing after an error.
7224 [Arne Ansper <arne@ats.cyber.ee>]
7226 *) rsa_eay.c would attempt to free a NULL context.
7227 [Arne Ansper <arne@ats.cyber.ee>]
7229 *) BIO_s_socket() had a broken should_retry() on Windoze.
7230 [Arne Ansper <arne@ats.cyber.ee>]
7232 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7233 [Arne Ansper <arne@ats.cyber.ee>]
7235 *) Make sure the already existing X509_STORE->depth variable is initialized
7236 in X509_STORE_new(), but document the fact that this variable is still
7237 unused in the certificate verification process.
7238 [Ralf S. Engelschall]
7240 *) Fix the various library and apps files to free up pkeys obtained from
7241 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7244 *) Fix reference counting in X509_PUBKEY_get(). This makes
7245 demos/maurice/example2.c work, amongst others, probably.
7246 [Steve Henson and Ben Laurie]
7248 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7249 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7250 are no longer created. This way we have a single and consistent command
7251 line interface `openssl <command>', similar to `cvs <command>'.
7252 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7254 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7255 BIT STRING wrapper always have zero unused bits.
7258 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7261 *) Make the top-level INSTALL documentation easier to understand.
7264 *) Makefiles updated to exit if an error occurs in a sub-directory
7265 make (including if user presses ^C) [Paul Sutton]
7267 *) Make Montgomery context stuff explicit in RSA data structure.
7270 *) Fix build order of pem and err to allow for generated pem.h.
7273 *) Fix renumbering bug in X509_NAME_delete_entry().
7276 *) Enhanced the err-ins.pl script so it makes the error library number
7277 global and can add a library name. This is needed for external ASN1 and
7278 other error libraries.
7281 *) Fixed sk_insert which never worked properly.
7284 *) Fix ASN1 macros so they can handle indefinite length construted
7285 EXPLICIT tags. Some non standard certificates use these: they can now
7289 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7290 into a single doc/ssleay.txt bundle. This way the information is still
7291 preserved but no longer messes up this directory. Now it's new room for
7292 the new set of documenation files.
7293 [Ralf S. Engelschall]
7295 *) SETs were incorrectly DER encoded. This was a major pain, because they
7296 shared code with SEQUENCEs, which aren't coded the same. This means that
7297 almost everything to do with SETs or SEQUENCEs has either changed name or
7298 number of arguments.
7299 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7301 *) Fix test data to work with the above.
7304 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7305 was already fixed by Eric for 0.9.1 it seems.
7306 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7308 *) Autodetect FreeBSD3.
7311 *) Fix various bugs in Configure. This affects the following platforms:
7315 unixware-2.0-pentium
7319 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7320 before they are needed.
7323 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7327 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7329 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7330 changed SSLeay to OpenSSL in version strings.
7331 [Ralf S. Engelschall]
7333 *) Some fixups to the top-level documents.
7336 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7337 because the symlink to include/ was missing.
7338 [Ralf S. Engelschall]
7340 *) Incorporated the popular no-RSA/DSA-only patches
7341 which allow to compile a RSA-free SSLeay.
7342 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7344 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7345 when "ssleay" is still not found.
7346 [Ralf S. Engelschall]
7348 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7349 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7351 *) Updated the README file.
7352 [Ralf S. Engelschall]
7354 *) Added various .cvsignore files in the CVS repository subdirs
7355 to make a "cvs update" really silent.
7356 [Ralf S. Engelschall]
7358 *) Recompiled the error-definition header files and added
7359 missing symbols to the Win32 linker tables.
7360 [Ralf S. Engelschall]
7362 *) Cleaned up the top-level documents;
7363 o new files: CHANGES and LICENSE
7364 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7365 o merged COPYRIGHT into LICENSE
7366 o removed obsolete TODO file
7367 o renamed MICROSOFT to INSTALL.W32
7368 [Ralf S. Engelschall]
7370 *) Removed dummy files from the 0.9.1b source tree:
7371 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7372 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7373 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7374 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7375 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7376 [Ralf S. Engelschall]
7378 *) Added various platform portability fixes.
7381 *) The Genesis of the OpenSSL rpject:
7382 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7383 Young and Tim J. Hudson created while they were working for C2Net until
7385 [The OpenSSL Project]
7388 Changes between 0.9.0b and 0.9.1b [not released]
7390 *) Updated a few CA certificates under certs/
7393 *) Changed some BIGNUM api stuff.
7396 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7397 DGUX x86, Linux Alpha, etc.
7400 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7401 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7405 *) Add -strparse option to asn1pars program which parses nested
7407 [Dr Stephen Henson <shenson@bigfoot.com>]
7409 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7412 *) DSA fix for "ca" program.
7415 *) Added "-genkey" option to "dsaparam" program.
7418 *) Added RIPE MD160 (rmd160) message digest.
7421 *) Added -a (all) option to "ssleay version" command.
7424 *) Added PLATFORM define which is the id given to Configure.
7427 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7430 *) Extended the ASN.1 parser routines.
7433 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7436 *) Added a BN_CTX to the BN library.
7439 *) Fixed the weak key values in DES library
7442 *) Changed API in EVP library for cipher aliases.
7445 *) Added support for RC2/64bit cipher.
7448 *) Converted the lhash library to the crypto/mem.c functions.
7451 *) Added more recognized ASN.1 object ids.
7454 *) Added more RSA padding checks for SSL/TLS.
7457 *) Added BIO proxy/filter functionality.
7460 *) Added extra_certs to SSL_CTX which can be used
7461 send extra CA certificates to the client in the CA cert chain sending
7462 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7465 *) Now Fortezza is denied in the authentication phase because
7466 this is key exchange mechanism is not supported by SSLeay at all.
7469 *) Additional PKCS1 checks.
7472 *) Support the string "TLSv1" for all TLS v1 ciphers.
7475 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7476 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7479 *) Fixed a few memory leaks.
7482 *) Fixed various code and comment typos.
7485 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7486 bytes sent in the client random.
7487 [Edward Bishop <ebishop@spyglass.com>]