5 Changes between 0.9.8c and 0.9.8d [xx XXX xxxx]
7 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
8 match only those. Before that, "AES256-SHA" would be interpreted
9 as a pattern and match "AES128-SHA" too since we currently only
10 have a single AES bit in the ciphersuite description bitmap.
11 That change, however, also applied to ciphersuite strings such as
12 "RC4-MD5" that intentionally matched multiple ciphersuites --
13 namely, SSL 2.0 ciphersuites in addition to the more common ones
16 So we change the selection algorithm again: Naming an explicit
17 ciphersuite selects this one ciphersuite, and any other similar
18 ciphersuite (same bitmap) from *other* protocol versions.
19 Thus, "RC4-MD5" again will properly select both the SSL 2.0
20 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
22 Since SSL 2.0 does not have any ciphersuites for which the
23 128/256 bit distinction would be relevant, this works for now.
24 The proper fix will be to use different bits for AES128 and
25 AES256, which would have avoided the problems from the beginning;
26 however, bits are scarce, so we can only do this in a new release
27 (not just a patchlevel) when we can change the SSL_CIPHER
28 definition to split the single 'unsigned long mask' bitmap into
29 multiple values to extend the available space.
33 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
35 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
36 (CVE-2006-4339) [Ben Laurie and Google Security Team]
38 *) Add AES IGE and biIGE modes.
41 *) Change the Unix randomness entropy gathering to use poll() when
42 possible instead of select(), since the latter has some
43 undesirable limitations.
44 [Darryl Miles via Richard Levitte and Bodo Moeller]
46 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
47 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
48 cannot be implicitly activated as part of, e.g., the "AES" alias.
49 However, please upgrade to OpenSSL 0.9.9[-dev] for
50 non-experimental use of the ECC ciphersuites to get TLS extension
51 support, which is required for curve and point format negotiation
52 to avoid potential handshake problems.
55 *) Disable rogue ciphersuites:
57 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
58 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
59 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
61 The latter two were purportedly from
62 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
65 Also deactive the remaining ciphersuites from
66 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
67 unofficial, and the ID has long expired.
70 *) Fix RSA blinding Heisenbug (problems sometimes occured on
71 dual-core machines) and other potential thread-safety issues.
74 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
75 versions), which is now available for royalty-free use
76 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
77 Also, add Camellia TLS ciphersuites from RFC 4132.
79 To minimize changes between patchlevels in the OpenSSL 0.9.8
80 series, Camellia remains excluded from compilation unless OpenSSL
81 is configured with 'enable-camellia'.
84 *) Disable the padding bug check when compression is in use. The padding
85 bug check assumes the first packet is of even length, this is not
86 necessarily true if compresssion is enabled and can result in false
87 positives causing handshake failure. The actual bug test is ancient
88 code so it is hoped that implementations will either have fixed it by
89 now or any which still have the bug do not support compression.
92 Changes between 0.9.8a and 0.9.8b [04 May 2006]
94 *) When applying a cipher rule check to see if string match is an explicit
95 cipher suite and only match that one cipher suite if it is.
98 *) Link in manifests for VC++ if needed.
99 [Austin Ziegler <halostatue@gmail.com>]
101 *) Update support for ECC-based TLS ciphersuites according to
102 draft-ietf-tls-ecc-12.txt with proposed changes (but without
103 TLS extensions, which are supported starting with the 0.9.9
104 branch, not in the OpenSSL 0.9.8 branch).
107 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
108 opaque EVP_CIPHER_CTX handling.
111 *) Fixes and enhancements to zlib compression code. We now only use
112 "zlib1.dll" and use the default __cdecl calling convention on Win32
113 to conform with the standards mentioned here:
114 http://www.zlib.net/DLL_FAQ.txt
115 Static zlib linking now works on Windows and the new --with-zlib-include
116 --with-zlib-lib options to Configure can be used to supply the location
117 of the headers and library. Gracefully handle case where zlib library
121 *) Several fixes and enhancements to the OID generation code. The old code
122 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
123 handle numbers larger than ULONG_MAX, truncated printing and had a
124 non standard OBJ_obj2txt() behaviour.
127 *) Add support for building of engines under engine/ as shared libraries
128 under VC++ build system.
131 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
132 Hopefully, we will not see any false combination of paths any more.
135 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
137 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
138 (part of SSL_OP_ALL). This option used to disable the
139 countermeasure against man-in-the-middle protocol-version
140 rollback in the SSL 2.0 server implementation, which is a bad
141 idea. (CVE-2005-2969)
143 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
144 for Information Security, National Institute of Advanced Industrial
145 Science and Technology [AIST], Japan)]
147 *) Add two function to clear and return the verify parameter flags.
150 *) Keep cipherlists sorted in the source instead of sorting them at
151 runtime, thus removing the need for a lock.
154 *) Avoid some small subgroup attacks in Diffie-Hellman.
155 [Nick Mathewson and Ben Laurie]
157 *) Add functions for well-known primes.
160 *) Extended Windows CE support.
161 [Satoshi Nakamura and Andy Polyakov]
163 *) Initialize SSL_METHOD structures at compile time instead of during
164 runtime, thus removing the need for a lock.
167 *) Make PKCS7_decrypt() work even if no certificate is supplied by
168 attempting to decrypt each encrypted key in turn. Add support to
172 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
174 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
177 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
180 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
181 key into the same file any more.
184 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
187 *) Add -utf8 command line and config file option to 'ca'.
188 [Stefan <stf@udoma.org]
190 *) Removed the macro des_crypt(), as it seems to conflict with some
191 libraries. Use DES_crypt().
194 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
195 involves renaming the source and generated shared-libs for
196 both. The engines will accept the corrected or legacy ids
197 ('ncipher' and '4758_cca' respectively) when binding. NB,
198 this only applies when building 'shared'.
199 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
201 *) Add attribute functions to EVP_PKEY structure. Modify
202 PKCS12_create() to recognize a CSP name attribute and
203 use it. Make -CSP option work again in pkcs12 utility.
206 *) Add new functionality to the bn blinding code:
207 - automatic re-creation of the BN_BLINDING parameters after
208 a fixed number of uses (currently 32)
209 - add new function for parameter creation
210 - introduce flags to control the update behaviour of the
211 BN_BLINDING parameters
212 - hide BN_BLINDING structure
213 Add a second BN_BLINDING slot to the RSA structure to improve
214 performance when a single RSA object is shared among several
218 *) Add support for DTLS.
219 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
221 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
222 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
225 *) Remove buggy and incompletet DH cert support from
226 ssl/ssl_rsa.c and ssl/s3_both.c
229 *) Use SHA-1 instead of MD5 as the default digest algorithm for
230 the apps/openssl applications.
233 *) Compile clean with "-Wall -Wmissing-prototypes
234 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
235 DEBUG_SAFESTACK must also be set.
238 *) Change ./Configure so that certain algorithms can be disabled by default.
239 The new counterpiece to "no-xxx" is "enable-xxx".
241 The patented RC5 and MDC2 algorithms will now be disabled unless
242 "enable-rc5" and "enable-mdc2", respectively, are specified.
244 (IDEA remains enabled despite being patented. This is because IDEA
245 is frequently required for interoperability, and there is no license
246 fee for non-commercial use. As before, "no-idea" can be used to
247 avoid this algorithm.)
251 *) Add processing of proxy certificates (see RFC 3820). This work was
252 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
253 EGEE (Enabling Grids for E-science in Europe).
256 *) RC4 performance overhaul on modern architectures/implementations, such
257 as Intel P4, IA-64 and AMD64.
260 *) New utility extract-section.pl. This can be used specify an alternative
261 section number in a pod file instead of having to treat each file as
262 a separate case in Makefile. This can be done by adding two lines to the
265 =for comment openssl_section:XXX
267 The blank line is mandatory.
271 *) New arguments -certform, -keyform and -pass for s_client and s_server
272 to allow alternative format key and certificate files and passphrase
276 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
277 update associated structures and add various utility functions.
279 Add new policy related verify parameters, include policy checking in
280 standard verify code. Enhance 'smime' application with extra parameters
281 to support policy checking and print out.
284 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
285 Nehemiah processors. These extensions support AES encryption in hardware
286 as well as RNG (though RNG support is currently disabled).
287 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
289 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
292 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
293 [Andy Polyakov and a number of other people]
295 *) Improved PowerPC platform support. Most notably BIGNUM assembler
296 implementation contributed by IBM.
297 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
299 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
300 exponent rather than 'unsigned long'. There is a corresponding change to
301 the new 'rsa_keygen' element of the RSA_METHOD structure.
302 [Jelte Jansen, Geoff Thorpe]
304 *) Functionality for creating the initial serial number file is now
305 moved from CA.pl to the 'ca' utility with a new option -create_serial.
307 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
308 number file to 1, which is bound to cause problems. To avoid
309 the problems while respecting compatibility between different 0.9.7
310 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
311 CA.pl for serial number initialization. With the new release 0.9.8,
312 we can fix the problem directly in the 'ca' utility.)
315 *) Reduced header interdepencies by declaring more opaque objects in
316 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
317 give fewer recursive includes, which could break lazy source code - so
318 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
319 developers should define this symbol when building and using openssl to
320 ensure they track the recommended behaviour, interfaces, [etc], but
321 backwards-compatible behaviour prevails when this isn't defined.
324 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
327 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
328 This will generate a random key of the appropriate length based on the
329 cipher context. The EVP_CIPHER can provide its own random key generation
330 routine to support keys of a specific form. This is used in the des and
331 3des routines to generate a key of the correct parity. Update S/MIME
332 code to use new functions and hence generate correct parity DES keys.
333 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
334 valid (weak or incorrect parity).
337 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
338 as looking them up. This is useful when the verified structure may contain
339 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
340 present unless the new PKCS7_NO_CRL flag is asserted.
343 *) Extend ASN1 oid configuration module. It now additionally accepts the
346 shortName = some long name, 1.2.3.4
349 *) Reimplemented the BN_CTX implementation. There is now no more static
350 limitation on the number of variables it can handle nor the depth of the
351 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
352 information can now expand as required, and rather than having a single
353 static array of bignums, BN_CTX now uses a linked-list of such arrays
354 allowing it to expand on demand whilst maintaining the usefulness of
358 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
359 to allow all RSA operations to function using a single BN_CTX.
362 *) Preliminary support for certificate policy evaluation and checking. This
363 is initially intended to pass the tests outlined in "Conformance Testing
364 of Relying Party Client Certificate Path Processing Logic" v1.07.
367 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
368 remained unused and not that useful. A variety of other little bignum
369 tweaks and fixes have also been made continuing on from the audit (see
373 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
374 associated ASN1, EVP and SSL functions and old ASN1 macros.
377 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
378 and this should never fail. So the return value from the use of
379 BN_set_word() (which can fail due to needless expansion) is now deprecated;
380 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
383 *) BN_CTX_get() should return zero-valued bignums, providing the same
384 initialised value as BN_new().
385 [Geoff Thorpe, suggested by Ulf Möller]
387 *) Support for inhibitAnyPolicy certificate extension.
390 *) An audit of the BIGNUM code is underway, for which debugging code is
391 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
392 is considered valid when processing BIGNUMs, and causes execution to
393 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
394 further steps are taken to deliberately pollute unused data in BIGNUM
395 structures to try and expose faulty code further on. For now, openssl will
396 (in its default mode of operation) continue to tolerate the inconsistent
397 forms that it has tolerated in the past, but authors and packagers should
398 consider trying openssl and their own applications when compiled with
399 these debugging symbols defined. It will help highlight potential bugs in
400 their own code, and will improve the test coverage for OpenSSL itself. At
401 some point, these tighter rules will become openssl's default to improve
402 maintainability, though the assert()s and other overheads will remain only
403 in debugging configurations. See bn.h for more details.
404 [Geoff Thorpe, Nils Larsch, Ulf Möller]
406 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
407 that can only be obtained through BN_CTX_new() (which implicitly
408 initialises it). The presence of this function only made it possible
409 to overwrite an existing structure (and cause memory leaks).
412 *) Because of the callback-based approach for implementing LHASH as a
413 template type, lh_insert() adds opaque objects to hash-tables and
414 lh_doall() or lh_doall_arg() are typically used with a destructor callback
415 to clean up those corresponding objects before destroying the hash table
416 (and losing the object pointers). So some over-zealous constifications in
417 LHASH have been relaxed so that lh_insert() does not take (nor store) the
418 objects as "const" and the lh_doall[_arg] callback wrappers are not
419 prototyped to have "const" restrictions on the object pointers they are
420 given (and so aren't required to cast them away any more).
423 *) The tmdiff.h API was so ugly and minimal that our own timing utility
424 (speed) prefers to use its own implementation. The two implementations
425 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
426 its object type properly exposed (MS_TM) instead of casting to/from "char
427 *". This may still change yet if someone realises MS_TM and "ms_time_***"
428 aren't necessarily the greatest nomenclatures - but this is what was used
429 internally to the implementation so I've used that for now.
432 *) Ensure that deprecated functions do not get compiled when
433 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
434 the self-tests were still using deprecated key-generation functions so
435 these have been updated also.
438 *) Reorganise PKCS#7 code to separate the digest location functionality
439 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
440 New function PKCS7_set_digest() to set the digest type for PKCS#7
441 digestedData type. Add additional code to correctly generate the
442 digestedData type and add support for this type in PKCS7 initialization
446 *) New function PKCS7_set0_type_other() this initializes a PKCS7
447 structure of type "other".
450 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
451 sure the loop does correctly stop and breaking ("division by zero")
452 modulus operations are not performed. The (pre-generated) prime
453 table crypto/bn/bn_prime.h was already correct, but it could not be
454 re-generated on some platforms because of the "division by zero"
455 situation in the script.
456 [Ralf S. Engelschall]
458 *) Update support for ECC-based TLS ciphersuites according to
459 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
460 SHA-1 now is only used for "small" curves (where the
461 representation of a field element takes up to 24 bytes); for
462 larger curves, the field element resulting from ECDH is directly
463 used as premaster secret.
464 [Douglas Stebila (Sun Microsystems Laboratories)]
466 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
467 curve secp160r1 to the tests.
468 [Douglas Stebila (Sun Microsystems Laboratories)]
470 *) Add the possibility to load symbols globally with DSO.
471 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
473 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
474 control of the error stack.
477 *) Add support for STORE in ENGINE.
480 *) Add the STORE type. The intention is to provide a common interface
481 to certificate and key stores, be they simple file-based stores, or
482 HSM-type store, or LDAP stores, or...
483 NOTE: The code is currently UNTESTED and isn't really used anywhere.
486 *) Add a generic structure called OPENSSL_ITEM. This can be used to
487 pass a list of arguments to any function as well as provide a way
488 for a function to pass data back to the caller.
491 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
492 works like BUF_strdup() but can be used to duplicate a portion of
493 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
497 *) Add the function sk_find_ex() which works like sk_find(), but will
498 return an index to an element even if an exact match couldn't be
499 found. The index is guaranteed to point at the element where the
500 searched-for key would be inserted to preserve sorting order.
503 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
504 takes an extra flags argument for optional functionality. Currently,
505 the following flags are defined:
507 OBJ_BSEARCH_VALUE_ON_NOMATCH
508 This one gets OBJ_bsearch_ex() to return a pointer to the first
509 element where the comparing function returns a negative or zero
512 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
513 This one gets OBJ_bsearch_ex() to return a pointer to the first
514 element where the comparing function returns zero. This is useful
515 if there are more than one element where the comparing function
519 *) Make it possible to create self-signed certificates with 'openssl ca'
520 in such a way that the self-signed certificate becomes part of the
521 CA database and uses the same mechanisms for serial number generation
522 as all other certificate signing. The new flag '-selfsign' enables
523 this functionality. Adapt CA.sh and CA.pl.in.
526 *) Add functionality to check the public key of a certificate request
527 against a given private. This is useful to check that a certificate
528 request can be signed by that key (self-signing).
531 *) Make it possible to have multiple active certificates with the same
532 subject in the CA index file. This is done only if the keyword
533 'unique_subject' is set to 'no' in the main CA section (default
534 if 'CA_default') of the configuration file. The value is saved
535 with the database itself in a separate index attribute file,
536 named like the index file with '.attr' appended to the name.
539 *) Generate muti valued AVAs using '+' notation in config files for
543 *) Support for nameConstraints certificate extension.
546 *) Support for policyConstraints certificate extension.
549 *) Support for policyMappings certificate extension.
552 *) Make sure the default DSA_METHOD implementation only uses its
553 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
554 and change its own handlers to be NULL so as to remove unnecessary
555 indirection. This lets alternative implementations fallback to the
556 default implementation more easily.
559 *) Support for directoryName in GeneralName related extensions
563 *) Make it possible to link applications using Makefile.shared.
564 Make that possible even when linking against static libraries!
567 *) Support for single pass processing for S/MIME signing. This now
568 means that S/MIME signing can be done from a pipe, in addition
569 cleartext signing (multipart/signed type) is effectively streaming
570 and the signed data does not need to be all held in memory.
572 This is done with a new flag PKCS7_STREAM. When this flag is set
573 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
574 is done after the data is output (and digests calculated) in
578 *) Add full support for -rpath/-R, both in shared libraries and
579 applications, at least on the platforms where it's known how
583 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
584 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
585 will now compute a table of multiples of the generator that
586 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
587 faster (notably in the case of a single point multiplication,
589 [Nils Larsch, Bodo Moeller]
591 *) IPv6 support for certificate extensions. The various extensions
592 which use the IP:a.b.c.d can now take IPv6 addresses using the
593 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
597 *) Added an ENGINE that implements RSA by performing private key
598 exponentiations with the GMP library. The conversions to and from
599 GMP's mpz_t format aren't optimised nor are any montgomery forms
600 cached, and on x86 it appears OpenSSL's own performance has caught up.
601 However there are likely to be other architectures where GMP could
602 provide a boost. This ENGINE is not built in by default, but it can be
603 specified at Configure time and should be accompanied by the necessary
604 linker additions, eg;
605 ./config -DOPENSSL_USE_GMP -lgmp
608 *) "openssl engine" will not display ENGINE/DSO load failure errors when
609 testing availability of engines with "-t" - the old behaviour is
610 produced by increasing the feature's verbosity with "-tt".
613 *) ECDSA routines: under certain error conditions uninitialized BN objects
614 could be freed. Solution: make sure initialization is performed early
615 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
619 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
620 and DH_METHOD (eg. by ENGINE implementations) to override the normal
621 software implementations. For DSA and DH, parameter generation can
622 also be overriden by providing the appropriate method callbacks.
625 *) Change the "progress" mechanism used in key-generation and
626 primality testing to functions that take a new BN_GENCB pointer in
627 place of callback/argument pairs. The new API functions have "_ex"
628 postfixes and the older functions are reimplemented as wrappers for
629 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
630 declarations of the old functions to help (graceful) attempts to
631 migrate to the new functions. Also, the new key-generation API
632 functions operate on a caller-supplied key-structure and return
633 success/failure rather than returning a key or NULL - this is to
634 help make "keygen" another member function of RSA_METHOD etc.
636 Example for using the new callback interface:
638 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
642 BN_GENCB_set(&my_cb, my_callback, my_arg);
644 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
645 /* For the meaning of a, b in calls to my_callback(), see the
646 * documentation of the function that calls the callback.
647 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
648 * my_callback should return 1 if it wants BN_is_prime_ex()
649 * to continue, or 0 to stop.
654 *) Change the ZLIB compression method to be stateful, and make it
655 available to TLS with the number defined in
656 draft-ietf-tls-compression-04.txt.
659 *) Add the ASN.1 structures and functions for CertificatePair, which
660 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
662 CertificatePair ::= SEQUENCE {
663 forward [0] Certificate OPTIONAL,
664 reverse [1] Certificate OPTIONAL,
665 -- at least one of the pair shall be present -- }
667 Also implement the PEM functions to read and write certificate
668 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
670 This needed to be defined, mostly for the sake of the LDAP
671 attribute crossCertificatePair, but may prove useful elsewhere as
675 *) Make it possible to inhibit symlinking of shared libraries in
676 Makefile.shared, for Cygwin's sake.
679 *) Extend the BIGNUM API by creating a function
680 void BN_set_negative(BIGNUM *a, int neg);
681 and a macro that behave like
682 int BN_is_negative(const BIGNUM *a);
684 to avoid the need to access 'a->neg' directly in applications.
687 *) Implement fast modular reduction for pseudo-Mersenne primes
688 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
689 EC_GROUP_new_curve_GFp() will now automatically use this
691 [Nils Larsch <nla@trustcenter.de>]
693 *) Add new lock type (CRYPTO_LOCK_BN).
696 *) Change the ENGINE framework to automatically load engines
697 dynamically from specific directories unless they could be
698 found to already be built in or loaded. Move all the
699 current engines except for the cryptodev one to a new
701 The engines in engines/ are built as shared libraries if
702 the "shared" options was given to ./Configure or ./config.
703 Otherwise, they are inserted in libcrypto.a.
704 /usr/local/ssl/engines is the default directory for dynamic
705 engines, but that can be overriden at configure time through
706 the usual use of --prefix and/or --openssldir, and at run
707 time with the environment variable OPENSSL_ENGINES.
708 [Geoff Thorpe and Richard Levitte]
710 *) Add Makefile.shared, a helper makefile to build shared
711 libraries. Addapt Makefile.org.
714 *) Add version info to Win32 DLLs.
715 [Peter 'Luna' Runestig" <peter@runestig.com>]
717 *) Add new 'medium level' PKCS#12 API. Certificates and keys
718 can be added using this API to created arbitrary PKCS#12
719 files while avoiding the low level API.
721 New options to PKCS12_create(), key or cert can be NULL and
722 will then be omitted from the output file. The encryption
723 algorithm NIDs can be set to -1 for no encryption, the mac
724 iteration count can be set to 0 to omit the mac.
726 Enhance pkcs12 utility by making the -nokeys and -nocerts
727 options work when creating a PKCS#12 file. New option -nomac
728 to omit the mac, NONE can be set for an encryption algorithm.
729 New code is modified to use the enhanced PKCS12_create()
730 instead of the low level API.
733 *) Extend ASN1 encoder to support indefinite length constructed
734 encoding. This can output sequences tags and octet strings in
735 this form. Modify pk7_asn1.c to support indefinite length
736 encoding. This is experimental and needs additional code to
737 be useful, such as an ASN1 bio and some enhanced streaming
740 Extend template encode functionality so that tagging is passed
741 down to the template encoder.
744 *) Let 'openssl req' fail if an argument to '-newkey' is not
745 recognized instead of using RSA as a default.
748 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
749 As these are not official, they are not included in "ALL";
750 the "ECCdraft" ciphersuite group alias can be used to select them.
751 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
753 *) Add ECDH engine support.
754 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
756 *) Add ECDH in new directory crypto/ecdh/.
757 [Douglas Stebila (Sun Microsystems Laboratories)]
759 *) Let BN_rand_range() abort with an error after 100 iterations
760 without success (which indicates a broken PRNG).
763 *) Change BN_mod_sqrt() so that it verifies that the input value
764 is really the square of the return value. (Previously,
765 BN_mod_sqrt would show GIGO behaviour.)
768 *) Add named elliptic curves over binary fields from X9.62, SECG,
769 and WAP/WTLS; add OIDs that were still missing.
771 [Sheueling Chang Shantz and Douglas Stebila
772 (Sun Microsystems Laboratories)]
774 *) Extend the EC library for elliptic curves over binary fields
775 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
778 EC_GF2m_simple_method
782 EC_GROUP_new_curve_GF2m
783 EC_GROUP_set_curve_GF2m
784 EC_GROUP_get_curve_GF2m
785 EC_POINT_set_affine_coordinates_GF2m
786 EC_POINT_get_affine_coordinates_GF2m
787 EC_POINT_set_compressed_coordinates_GF2m
789 Point compression for binary fields is disabled by default for
790 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
793 As binary polynomials are represented as BIGNUMs, various members
794 of the EC_GROUP and EC_POINT data structures can be shared
795 between the implementations for prime fields and binary fields;
796 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
797 are essentially identical to their ..._GFp counterparts.
798 (For simplicity, the '..._GFp' prefix has been dropped from
799 various internal method names.)
801 An internal 'field_div' method (similar to 'field_mul' and
802 'field_sqr') has been added; this is used only for binary fields.
804 [Sheueling Chang Shantz and Douglas Stebila
805 (Sun Microsystems Laboratories)]
807 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
808 through methods ('mul', 'precompute_mult').
810 The generic implementations (now internally called 'ec_wNAF_mul'
811 and 'ec_wNAF_precomputed_mult') remain the default if these
812 methods are undefined.
814 [Sheueling Chang Shantz and Douglas Stebila
815 (Sun Microsystems Laboratories)]
817 *) New function EC_GROUP_get_degree, which is defined through
818 EC_METHOD. For curves over prime fields, this returns the bit
819 length of the modulus.
821 [Sheueling Chang Shantz and Douglas Stebila
822 (Sun Microsystems Laboratories)]
824 *) New functions EC_GROUP_dup, EC_POINT_dup.
825 (These simply call ..._new and ..._copy).
827 [Sheueling Chang Shantz and Douglas Stebila
828 (Sun Microsystems Laboratories)]
830 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
831 Polynomials are represented as BIGNUMs (where the sign bit is not
832 used) in the following functions [macros]:
835 BN_GF2m_sub [= BN_GF2m_add]
836 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
837 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
838 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
840 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
841 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
842 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
843 BN_GF2m_cmp [= BN_ucmp]
845 (Note that only the 'mod' functions are actually for fields GF(2^m).
846 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
848 For some functions, an the irreducible polynomial defining a
849 field can be given as an 'unsigned int[]' with strictly
850 decreasing elements giving the indices of those bits that are set;
851 i.e., p[] represents the polynomial
852 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
854 p[0] > p[1] > ... > p[k] = 0.
855 This applies to the following functions:
860 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
861 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
864 BN_GF2m_mod_solve_quad_arr
868 Conversion can be performed by the following functions:
873 bntest.c has additional tests for binary polynomial arithmetic.
875 Two implementations for BN_GF2m_mod_div() are available.
876 The default algorithm simply uses BN_GF2m_mod_inv() and
877 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
878 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
879 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
881 [Sheueling Chang Shantz and Douglas Stebila
882 (Sun Microsystems Laboratories)]
884 *) Add new error code 'ERR_R_DISABLED' that can be used when some
885 functionality is disabled at compile-time.
886 [Douglas Stebila <douglas.stebila@sun.com>]
888 *) Change default behaviour of 'openssl asn1parse' so that more
889 information is visible when viewing, e.g., a certificate:
891 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
892 mode the content of non-printable OCTET STRINGs is output in a
893 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
894 avoid the appearance of a printable string.
895 [Nils Larsch <nla@trustcenter.de>]
897 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
899 EC_GROUP_set_asn1_flag()
900 EC_GROUP_get_asn1_flag()
901 EC_GROUP_set_point_conversion_form()
902 EC_GROUP_get_point_conversion_form()
903 These control ASN1 encoding details:
904 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
905 has been set to OPENSSL_EC_NAMED_CURVE.
906 - Points are encoded in uncompressed form by default; options for
907 asn1_for are as for point2oct, namely
908 POINT_CONVERSION_COMPRESSED
909 POINT_CONVERSION_UNCOMPRESSED
910 POINT_CONVERSION_HYBRID
912 Also add 'seed' and 'seed_len' members to EC_GROUP with access
916 EC_GROUP_get_seed_len()
917 This is used only for ASN1 purposes (so far).
918 [Nils Larsch <nla@trustcenter.de>]
920 *) Add 'field_type' member to EC_METHOD, which holds the NID
921 of the appropriate field type OID. The new function
922 EC_METHOD_get_field_type() returns this value.
923 [Nils Larsch <nla@trustcenter.de>]
930 providing useful interfaces to EC_POINT_point2oct() and
931 EC_POINT_oct2point().
932 [Nils Larsch <nla@trustcenter.de>]
934 *) Change internals of the EC library so that the functions
935 EC_GROUP_set_generator()
936 EC_GROUP_get_generator()
938 EC_GROUP_get_cofactor()
939 are implemented directly in crypto/ec/ec_lib.c and not dispatched
940 to methods, which would lead to unnecessary code duplication when
941 adding different types of curves.
942 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
944 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
945 arithmetic, and such that modified wNAFs are generated
946 (which avoid length expansion in many cases).
949 *) Add a function EC_GROUP_check_discriminant() (defined via
950 EC_METHOD) that verifies that the curve discriminant is non-zero.
952 Add a function EC_GROUP_check() that makes some sanity tests
953 on a EC_GROUP, its generator and order. This includes
954 EC_GROUP_check_discriminant().
955 [Nils Larsch <nla@trustcenter.de>]
957 *) Add ECDSA in new directory crypto/ecdsa/.
959 Add applications 'openssl ecparam' and 'openssl ecdsa'
960 (these are based on 'openssl dsaparam' and 'openssl dsa').
962 ECDSA support is also included in various other files across the
963 library. Most notably,
964 - 'openssl req' now has a '-newkey ecdsa:file' option;
965 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
966 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
967 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
968 them suitable for ECDSA where domain parameters must be
969 extracted before the specific public key;
970 - ECDSA engine support has been added.
971 [Nils Larsch <nla@trustcenter.de>]
973 *) Include some named elliptic curves, and add OIDs from X9.62,
974 SECG, and WAP/WTLS. Each curve can be obtained from the new
976 EC_GROUP_new_by_curve_name(),
977 and the list of available named curves can be obtained with
978 EC_get_builtin_curves().
979 Also add a 'curve_name' member to EC_GROUP objects, which can be
981 EC_GROUP_set_curve_name()
982 EC_GROUP_get_curve_name()
983 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
985 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
986 was actually never needed) and in BN_mul(). The removal in BN_mul()
987 required a small change in bn_mul_part_recursive() and the addition
988 of the functions bn_cmp_part_words(), bn_sub_part_words() and
989 bn_add_part_words(), which do the same thing as bn_cmp_words(),
990 bn_sub_words() and bn_add_words() except they take arrays with
994 Changes between 0.9.7j and 0.9.7k [xx XXX xxxx]
996 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
997 (CVE-2006-4339) [Ben Laurie and Google Security Team]
999 *) Change the Unix randomness entropy gathering to use poll() when
1000 possible instead of select(), since the latter has some
1001 undesirable limitations.
1002 [Darryl Miles via Richard Levitte and Bodo Moeller]
1004 *) Disable rogue ciphersuites:
1006 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1007 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1008 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1010 The latter two were purportedly from
1011 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1014 Also deactive the remaining ciphersuites from
1015 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1016 unofficial, and the ID has long expired.
1019 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1020 dual-core machines) and other potential thread-safety issues.
1023 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1025 *) Adapt fipsld and the build system to link against the validated FIPS
1026 module in FIPS mode.
1029 *) Fixes for VC++ 2005 build under Windows.
1032 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1033 from a Windows bash shell such as MSYS. It is autodetected from the
1034 "config" script when run from a VC++ environment. Modify standard VC++
1035 build to use fipscanister.o from the GNU make build.
1038 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1040 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1041 The value now differs depending on if you build for FIPS or not.
1042 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1043 safely run with a non-FIPSed libcrypto, as it may crash because of
1044 the difference induced by this change.
1047 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1049 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1050 (part of SSL_OP_ALL). This option used to disable the
1051 countermeasure against man-in-the-middle protocol-version
1052 rollback in the SSL 2.0 server implementation, which is a bad
1053 idea. (CVE-2005-2969)
1055 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1056 for Information Security, National Institute of Advanced Industrial
1057 Science and Technology [AIST], Japan)]
1059 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1060 mainly for FIPS compliance and not fully integrated at this stage.
1063 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1064 the exponentiation using a fixed-length exponent. (Otherwise,
1065 the information leaked through timing could expose the secret key
1066 after many signatures; cf. Bleichenbacher's attack on DSA with
1070 *) Make a new fixed-window mod_exp implementation the default for
1071 RSA, DSA, and DH private-key operations so that the sequence of
1072 squares and multiplies and the memory access pattern are
1073 independent of the particular secret key. This will mitigate
1074 cache-timing and potential related attacks.
1076 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1077 and this is automatically used by BN_mod_exp_mont() if the new flag
1078 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1079 will use this BN flag for private exponents unless the flag
1080 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1081 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1083 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1085 *) Change the client implementation for SSLv23_method() and
1086 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1087 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1088 (Previously, the SSL 2.0 backwards compatible Client Hello
1089 message format would be used even with SSL_OP_NO_SSLv2.)
1092 *) Add support for smime-type MIME parameter in S/MIME messages which some
1096 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1097 a threadsafe manner. Modify rsa code to use new function and add calls
1098 to dsa and dh code (which had race conditions before).
1101 *) Include the fixed error library code in the C error file definitions
1102 instead of fixing them up at runtime. This keeps the error code
1103 structures constant.
1106 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1108 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1111 *) Fixes for newer kerberos headers. NB: the casts are needed because
1112 the 'length' field is signed on one version and unsigned on another
1113 with no (?) obvious way to tell the difference, without these VC++
1114 complains. Also the "definition" of FAR (blank) is no longer included
1115 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1116 some needed definitions.
1119 *) Undo Cygwin change.
1122 *) Added support for proxy certificates according to RFC 3820.
1123 Because they may be a security thread to unaware applications,
1124 they must be explicitely allowed in run-time. See
1125 docs/HOWTO/proxy_certificates.txt for further information.
1128 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1130 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1131 server and client random values. Previously
1132 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1133 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1135 This change has negligible security impact because:
1137 1. Server and client random values still have 24 bytes of pseudo random
1140 2. Server and client random values are sent in the clear in the initial
1143 3. The master secret is derived using the premaster secret (48 bytes in
1144 size for static RSA ciphersuites) as well as client server and random
1147 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1150 [Stephen Henson, reported by UK NISCC]
1152 *) Use Windows randomness collection on Cygwin.
1155 *) Fix hang in EGD/PRNGD query when communication socket is closed
1156 prematurely by EGD/PRNGD.
1157 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1159 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1162 *) Back-port of selected performance improvements from development
1163 branch, as well as improved support for PowerPC platforms.
1166 *) Add lots of checks for memory allocation failure, error codes to indicate
1167 failure and freeing up memory if a failure occurs.
1168 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1170 *) Add new -passin argument to dgst.
1173 *) Perform some character comparisons of different types in X509_NAME_cmp:
1174 this is needed for some certificates that reencode DNs into UTF8Strings
1175 (in violation of RFC3280) and can't or wont issue name rollover
1179 *) Make an explicit check during certificate validation to see that
1180 the CA setting in each certificate on the chain is correct. As a
1181 side effect always do the following basic checks on extensions,
1182 not just when there's an associated purpose to the check:
1184 - if there is an unhandled critical extension (unless the user
1185 has chosen to ignore this fault)
1186 - if the path length has been exceeded (if one is set at all)
1187 - that certain extensions fit the associated purpose (if one has
1191 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1193 *) Avoid a race condition when CRLs are checked in a multi threaded
1194 environment. This would happen due to the reordering of the revoked
1195 entries during signature checking and serial number lookup. Now the
1196 encoding is cached and the serial number sort performed under a lock.
1197 Add new STACK function sk_is_sorted().
1200 *) Add Delta CRL to the extension code.
1203 *) Various fixes to s3_pkt.c so alerts are sent properly.
1204 [David Holmes <d.holmes@f5.com>]
1206 *) Reduce the chances of duplicate issuer name and serial numbers (in
1207 violation of RFC3280) using the OpenSSL certificate creation utilities.
1208 This is done by creating a random 64 bit value for the initial serial
1209 number when a serial number file is created or when a self signed
1210 certificate is created using 'openssl req -x509'. The initial serial
1211 number file is created using 'openssl x509 -next_serial' in CA.pl
1212 rather than being initialized to 1.
1215 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1217 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1218 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1219 [Joe Orton, Steve Henson]
1221 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1223 [Joe Orton, Steve Henson]
1225 *) Make it possible to have multiple active certificates with the same
1226 subject in the CA index file. This is done only if the keyword
1227 'unique_subject' is set to 'no' in the main CA section (default
1228 if 'CA_default') of the configuration file. The value is saved
1229 with the database itself in a separate index attribute file,
1230 named like the index file with '.attr' appended to the name.
1233 *) X509 verify fixes. Disable broken certificate workarounds when
1234 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1235 keyUsage extension present. Don't accept CRLs with unhandled critical
1236 extensions: since verify currently doesn't process CRL extensions this
1237 rejects a CRL with *any* critical extensions. Add new verify error codes
1241 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1242 A clarification of RFC2560 will require the use of OCTET STRINGs and
1243 some implementations cannot handle the current raw format. Since OpenSSL
1244 copies and compares OCSP nonces as opaque blobs without any attempt at
1245 parsing them this should not create any compatibility issues.
1248 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1249 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1250 this HMAC (and other) operations are several times slower than OpenSSL
1254 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1255 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1257 *) Use the correct content when signing type "other".
1260 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1262 *) Fix various bugs revealed by running the NISCC test suite:
1264 Stop out of bounds reads in the ASN1 code when presented with
1265 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1267 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1269 If verify callback ignores invalid public key errors don't try to check
1270 certificate signature with the NULL public key.
1274 *) New -ignore_err option in ocsp application to stop the server
1275 exiting on the first error in a request.
1278 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1279 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1283 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1284 extra data after the compression methods not only for TLS 1.0
1285 but also for SSL 3.0 (as required by the specification).
1286 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1288 *) Change X509_certificate_type() to mark the key as exported/exportable
1289 when it's 512 *bits* long, not 512 bytes.
1292 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1293 blocks during encryption.
1296 *) Various fixes to base64 BIO and non blocking I/O. On write
1297 flushes were not handled properly if the BIO retried. On read
1298 data was not being buffered properly and had various logic bugs.
1299 This also affects blocking I/O when the data being decoded is a
1303 *) Various S/MIME bugfixes and compatibility changes:
1304 output correct application/pkcs7 MIME type if
1305 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1306 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1307 of files as .eml work). Correctly handle very long lines in MIME
1311 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1313 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1314 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1315 a protocol version number mismatch like a decryption error
1316 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1319 *) Turn on RSA blinding by default in the default implementation
1320 to avoid a timing attack. Applications that don't want it can call
1321 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1322 They would be ill-advised to do so in most cases.
1323 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1325 *) Change RSA blinding code so that it works when the PRNG is not
1326 seeded (in this case, the secret RSA exponent is abused as
1327 an unpredictable seed -- if it is not unpredictable, there
1328 is no point in blinding anyway). Make RSA blinding thread-safe
1329 by remembering the creator's thread ID in rsa->blinding and
1330 having all other threads use local one-time blinding factors
1331 (this requires more computation than sharing rsa->blinding, but
1332 avoids excessive locking; and if an RSA object is not shared
1333 between threads, blinding will still be very fast).
1336 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1337 ENGINE as defaults for all supported algorithms irrespective of
1338 the 'flags' parameter. 'flags' is now honoured, so applications
1339 should make sure they are passing it correctly.
1342 *) Target "mingw" now allows native Windows code to be generated in
1343 the Cygwin environment as well as with the MinGW compiler.
1346 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1348 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1349 via timing by performing a MAC computation even if incorrrect
1350 block cipher padding has been found. This is a countermeasure
1351 against active attacks where the attacker has to distinguish
1352 between bad padding and a MAC verification error. (CVE-2003-0078)
1354 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1355 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1356 Martin Vuagnoux (EPFL, Ilion)]
1358 *) Make the no-err option work as intended. The intention with no-err
1359 is not to have the whole error stack handling routines removed from
1360 libcrypto, it's only intended to remove all the function name and
1361 reason texts, thereby removing some of the footprint that may not
1362 be interesting if those errors aren't displayed anyway.
1364 NOTE: it's still possible for any application or module to have it's
1365 own set of error texts inserted. The routines are there, just not
1366 used by default when no-err is given.
1369 *) Add support for FreeBSD on IA64.
1370 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1372 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1373 Kerberos function mit_des_cbc_cksum(). Before this change,
1374 the value returned by DES_cbc_cksum() was like the one from
1375 mit_des_cbc_cksum(), except the bytes were swapped.
1376 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1378 *) Allow an application to disable the automatic SSL chain building.
1379 Before this a rather primitive chain build was always performed in
1380 ssl3_output_cert_chain(): an application had no way to send the
1381 correct chain if the automatic operation produced an incorrect result.
1383 Now the chain builder is disabled if either:
1385 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1387 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1389 The reasoning behind this is that an application would not want the
1390 auto chain building to take place if extra chain certificates are
1391 present and it might also want a means of sending no additional
1392 certificates (for example the chain has two certificates and the
1396 *) Add the possibility to build without the ENGINE framework.
1397 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1399 *) Under Win32 gmtime() can return NULL: check return value in
1400 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1403 *) DSA routines: under certain error conditions uninitialized BN objects
1404 could be freed. Solution: make sure initialization is performed early
1405 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1406 Nils Larsch <nla@trustcenter.de> via PR#459)
1409 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1410 checked on reconnect on the client side, therefore session resumption
1411 could still fail with a "ssl session id is different" error. This
1412 behaviour is masked when SSL_OP_ALL is used due to
1413 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1414 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1415 followup to PR #377.
1418 *) IA-32 assembler support enhancements: unified ELF targets, support
1419 for SCO/Caldera platforms, fix for Cygwin shared build.
1422 *) Add support for FreeBSD on sparc64. As a consequence, support for
1423 FreeBSD on non-x86 processors is separate from x86 processors on
1424 the config script, much like the NetBSD support.
1425 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1427 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1429 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1432 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1433 code (06) was taken as the first octet of the session ID and the last
1434 octet was ignored consequently. As a result SSLv2 client side session
1435 caching could not have worked due to the session ID mismatch between
1437 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1441 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1442 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1446 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1447 seems that in spite of existing for more than a year, many application
1448 author have done nothing to provide the necessary callbacks, which
1449 means that this particular engine will not work properly anywhere.
1450 This is a very unfortunate situation which forces us, in the name
1451 of usability, to give the hw_ncipher.c a static lock, which is part
1453 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1454 appear in 0.9.8 or later. We EXPECT application authors to have
1455 dealt properly with this when 0.9.8 is released (unless we actually
1456 make such changes in the libcrypto locking code that changes will
1457 have to be made anyway).
1460 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1461 octets have been read, EOF or an error occurs. Without this change
1462 some truncated ASN1 structures will not produce an error.
1465 *) Disable Heimdal support, since it hasn't been fully implemented.
1466 Still give the possibility to force the use of Heimdal, but with
1467 warnings and a request that patches get sent to openssl-dev.
1470 *) Add the VC-CE target, introduce the WINCE sysname, and add
1471 INSTALL.WCE and appropriate conditionals to make it build.
1472 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1474 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1475 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1476 edit numbers of the version.
1477 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1479 *) Introduce safe string copy and catenation functions
1480 (BUF_strlcpy() and BUF_strlcat()).
1481 [Ben Laurie (CHATS) and Richard Levitte]
1483 *) Avoid using fixed-size buffers for one-line DNs.
1484 [Ben Laurie (CHATS)]
1486 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1487 resizing buffers containing secrets, and use where appropriate.
1488 [Ben Laurie (CHATS)]
1490 *) Avoid using fixed size buffers for configuration file location.
1491 [Ben Laurie (CHATS)]
1493 *) Avoid filename truncation for various CA files.
1494 [Ben Laurie (CHATS)]
1496 *) Use sizeof in preference to magic numbers.
1497 [Ben Laurie (CHATS)]
1499 *) Avoid filename truncation in cert requests.
1500 [Ben Laurie (CHATS)]
1502 *) Add assertions to check for (supposedly impossible) buffer
1504 [Ben Laurie (CHATS)]
1506 *) Don't cache truncated DNS entries in the local cache (this could
1507 potentially lead to a spoofing attack).
1508 [Ben Laurie (CHATS)]
1510 *) Fix various buffers to be large enough for hex/decimal
1511 representations in a platform independent manner.
1512 [Ben Laurie (CHATS)]
1514 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1515 resizing buffers containing secrets, and use where appropriate.
1516 [Ben Laurie (CHATS)]
1518 *) Add BIO_indent() to avoid much slightly worrying code to do
1520 [Ben Laurie (CHATS)]
1522 *) Convert sprintf()/BIO_puts() to BIO_printf().
1523 [Ben Laurie (CHATS)]
1525 *) buffer_gets() could terminate with the buffer only half
1527 [Ben Laurie (CHATS)]
1529 *) Add assertions to prevent user-supplied crypto functions from
1530 overflowing internal buffers by having large block sizes, etc.
1531 [Ben Laurie (CHATS)]
1533 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1535 [Ben Laurie (CHATS)]
1537 *) Eliminate unused copy of key in RC4.
1538 [Ben Laurie (CHATS)]
1540 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1541 [Ben Laurie (CHATS)]
1543 *) Fix off-by-one error in EGD path.
1544 [Ben Laurie (CHATS)]
1546 *) If RANDFILE path is too long, ignore instead of truncating.
1547 [Ben Laurie (CHATS)]
1549 *) Eliminate unused and incorrectly sized X.509 structure
1551 [Ben Laurie (CHATS)]
1553 *) Eliminate unused and dangerous function knumber().
1554 [Ben Laurie (CHATS)]
1556 *) Eliminate unused and dangerous structure, KSSL_ERR.
1557 [Ben Laurie (CHATS)]
1559 *) Protect against overlong session ID context length in an encoded
1560 session object. Since these are local, this does not appear to be
1562 [Ben Laurie (CHATS)]
1564 *) Change from security patch (see 0.9.6e below) that did not affect
1565 the 0.9.6 release series:
1567 Remote buffer overflow in SSL3 protocol - an attacker could
1568 supply an oversized master key in Kerberos-enabled versions.
1570 [Ben Laurie (CHATS)]
1572 *) Change the SSL kerb5 codes to match RFC 2712.
1575 *) Make -nameopt work fully for req and add -reqopt switch.
1576 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1578 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1579 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1581 *) Make sure tests can be performed even if the corresponding algorithms
1582 have been removed entirely. This was also the last step to make
1583 OpenSSL compilable with DJGPP under all reasonable conditions.
1584 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1586 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1587 to allow version independent disabling of normally unselected ciphers,
1588 which may be activated as a side-effect of selecting a single cipher.
1590 (E.g., cipher list string "RSA" enables ciphersuites that are left
1591 out of "ALL" because they do not provide symmetric encryption.
1592 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1593 [Lutz Jaenicke, Bodo Moeller]
1595 *) Add appropriate support for separate platform-dependent build
1596 directories. The recommended way to make a platform-dependent
1597 build directory is the following (tested on Linux), maybe with
1600 # Place yourself outside of the OpenSSL source tree. In
1601 # this example, the environment variable OPENSSL_SOURCE
1602 # is assumed to contain the absolute OpenSSL source directory.
1603 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1604 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1605 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1606 mkdir -p `dirname $F`
1607 ln -s $OPENSSL_SOURCE/$F $F
1610 To be absolutely sure not to disturb the source tree, a "make clean"
1611 is a good thing. If it isn't successfull, don't worry about it,
1612 it probably means the source directory is very clean.
1615 *) Make sure any ENGINE control commands make local copies of string
1616 pointers passed to them whenever necessary. Otherwise it is possible
1617 the caller may have overwritten (or deallocated) the original string
1618 data when a later ENGINE operation tries to use the stored values.
1619 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1621 *) Improve diagnostics in file reading and command-line digests.
1622 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1624 *) Add AES modes CFB and OFB to the object database. Correct an
1625 error in AES-CFB decryption.
1628 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1629 allows existing EVP_CIPHER_CTX structures to be reused after
1630 calling EVP_*Final(). This behaviour is used by encryption
1631 BIOs and some applications. This has the side effect that
1632 applications must explicitly clean up cipher contexts with
1633 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1636 *) Check the values of dna and dnb in bn_mul_recursive before calling
1637 bn_mul_comba (a non zero value means the a or b arrays do not contain
1638 n2 elements) and fallback to bn_mul_normal if either is not zero.
1641 *) Fix escaping of non-ASCII characters when using the -subj option
1642 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1645 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1646 form for "surname", serialNumber has no short form.
1647 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1648 therefore remove "mail" short name for "internet 7".
1649 The OID for unique identifiers in X509 certificates is
1650 x500UniqueIdentifier, not uniqueIdentifier.
1651 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1654 *) Add an "init" command to the ENGINE config module and auto initialize
1655 ENGINEs. Without any "init" command the ENGINE will be initialized
1656 after all ctrl commands have been executed on it. If init=1 the
1657 ENGINE is initailized at that point (ctrls before that point are run
1658 on the uninitialized ENGINE and after on the initialized one). If
1659 init=0 then the ENGINE will not be iniatialized at all.
1662 *) Fix the 'app_verify_callback' interface so that the user-defined
1663 argument is actually passed to the callback: In the
1664 SSL_CTX_set_cert_verify_callback() prototype, the callback
1665 declaration has been changed from
1668 int (*cb)(X509_STORE_CTX *,void *);
1669 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1670 i=s->ctx->app_verify_callback(&ctx)
1671 has been changed into
1672 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1674 To update applications using SSL_CTX_set_cert_verify_callback(),
1675 a dummy argument can be added to their callback functions.
1676 [D. K. Smetters <smetters@parc.xerox.com>]
1678 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1679 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1681 *) Add and OPENSSL_LOAD_CONF define which will cause
1682 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1683 This allows older applications to transparently support certain
1684 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1685 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1686 load the config file and OPENSSL_add_all_algorithms_conf() which will
1687 always load it have also been added.
1690 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1691 Adjust NIDs and EVP layer.
1692 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1694 *) Config modules support in openssl utility.
1696 Most commands now load modules from the config file,
1697 though in a few (such as version) this isn't done
1698 because it couldn't be used for anything.
1700 In the case of ca and req the config file used is
1701 the same as the utility itself: that is the -config
1702 command line option can be used to specify an
1706 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1707 use "openssl_conf" if filename is NULL use default openssl config file.
1710 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1711 config section name. Add a new flag to tolerate a missing config file
1712 and move code to CONF_modules_load_file().
1715 *) Support for crypto accelerator cards from Accelerated Encryption
1716 Processing, www.aep.ie. (Use engine 'aep')
1717 The support was copied from 0.9.6c [engine] and adapted/corrected
1718 to work with the new engine framework.
1719 [AEP Inc. and Richard Levitte]
1721 *) Support for SureWare crypto accelerator cards from Baltimore
1722 Technologies. (Use engine 'sureware')
1723 The support was copied from 0.9.6c [engine] and adapted
1724 to work with the new engine framework.
1727 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1728 make the newer ENGINE framework commands for the CHIL engine work.
1729 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1731 *) Make it possible to produce shared libraries on ReliantUNIX.
1732 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1734 *) Add the configuration target debug-linux-ppro.
1735 Make 'openssl rsa' use the general key loading routines
1736 implemented in apps.c, and make those routines able to
1737 handle the key format FORMAT_NETSCAPE and the variant
1739 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1741 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1742 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1744 *) Add -keyform to rsautl, and document -engine.
1745 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1747 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1748 BIO_R_NO_SUCH_FILE error code rather than the generic
1749 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1752 *) Add new functions
1754 ERR_peek_last_error_line
1755 ERR_peek_last_error_line_data.
1756 These are similar to
1759 ERR_peek_error_line_data,
1760 but report on the latest error recorded rather than the first one
1761 still in the error queue.
1762 [Ben Laurie, Bodo Moeller]
1764 *) default_algorithms option in ENGINE config module. This allows things
1766 default_algorithms = ALL
1767 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1770 *) Prelminary ENGINE config module.
1773 *) New experimental application configuration code.
1776 *) Change the AES code to follow the same name structure as all other
1777 symmetric ciphers, and behave the same way. Move everything to
1778 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1779 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1781 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1782 [Ben Laurie and Theo de Raadt]
1784 *) Add option to output public keys in req command.
1785 [Massimiliano Pala madwolf@openca.org]
1787 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1788 (up to about 10% better than before for P-192 and P-224).
1791 *) New functions/macros
1793 SSL_CTX_set_msg_callback(ctx, cb)
1794 SSL_CTX_set_msg_callback_arg(ctx, arg)
1795 SSL_set_msg_callback(ssl, cb)
1796 SSL_set_msg_callback_arg(ssl, arg)
1798 to request calling a callback function
1800 void cb(int write_p, int version, int content_type,
1801 const void *buf, size_t len, SSL *ssl, void *arg)
1803 whenever a protocol message has been completely received
1804 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1805 protocol version according to which the SSL library interprets
1806 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1807 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1808 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1809 specification (change_cipher_spec(20), alert(21), handshake(22)).
1810 'buf' and 'len' point to the actual message, 'ssl' to the
1811 SSL object, and 'arg' is the application-defined value set by
1812 SSL[_CTX]_set_msg_callback_arg().
1814 'openssl s_client' and 'openssl s_server' have new '-msg' options
1815 to enable a callback that displays all protocol messages.
1818 *) Change the shared library support so shared libraries are built as
1819 soon as the corresponding static library is finished, and thereby get
1820 openssl and the test programs linked against the shared library.
1821 This still only happens when the keyword "shard" has been given to
1822 the configuration scripts.
1824 NOTE: shared library support is still an experimental thing, and
1825 backward binary compatibility is still not guaranteed.
1826 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1828 *) Add support for Subject Information Access extension.
1829 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1831 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1832 additional bytes when new memory had to be allocated, not just
1833 when reusing an existing buffer.
1836 *) New command line and configuration option 'utf8' for the req command.
1837 This allows field values to be specified as UTF8 strings.
1840 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1841 runs for the former and machine-readable output for the latter.
1844 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1845 of the e-mail address in the DN (i.e., it will go into a certificate
1846 extension only). The new configuration file option 'email_in_dn = no'
1847 has the same effect.
1848 [Massimiliano Pala madwolf@openca.org]
1850 *) Change all functions with names starting with des_ to be starting
1851 with DES_ instead. Add wrappers that are compatible with libdes,
1852 but are named _ossl_old_des_*. Finally, add macros that map the
1853 des_* symbols to the corresponding _ossl_old_des_* if libdes
1854 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1855 desired, the des_* symbols will be mapped to DES_*, with one
1858 Since we provide two compatibility mappings, the user needs to
1859 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1860 compatibility is desired. The default (i.e., when that macro
1861 isn't defined) is OpenSSL 0.9.6c compatibility.
1863 There are also macros that enable and disable the support of old
1864 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1865 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1866 are defined, the default will apply: to support the old des routines.
1868 In either case, one must include openssl/des.h to get the correct
1869 definitions. Do not try to just include openssl/des_old.h, that
1872 NOTE: This is a major break of an old API into a new one. Software
1873 authors are encouraged to switch to the DES_ style functions. Some
1874 time in the future, des_old.h and the libdes compatibility functions
1875 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1876 default), and then completely removed.
1879 *) Test for certificates which contain unsupported critical extensions.
1880 If such a certificate is found during a verify operation it is
1881 rejected by default: this behaviour can be overridden by either
1882 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1883 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1884 X509_supported_extension() has also been added which returns 1 if a
1885 particular extension is supported.
1888 *) Modify the behaviour of EVP cipher functions in similar way to digests
1889 to retain compatibility with existing code.
1892 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1893 compatibility with existing code. In particular the 'ctx' parameter does
1894 not have to be to be initialized before the call to EVP_DigestInit() and
1895 it is tidied up after a call to EVP_DigestFinal(). New function
1896 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1897 EVP_MD_CTX_copy() changed to not require the destination to be
1898 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1899 requires the destination to be valid.
1901 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1902 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1905 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1906 so that complete 'Handshake' protocol structures are kept in memory
1907 instead of overwriting 'msg_type' and 'length' with 'body' data.
1910 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1911 [Massimo Santin via Richard Levitte]
1913 *) Major restructuring to the underlying ENGINE code. This includes
1914 reduction of linker bloat, separation of pure "ENGINE" manipulation
1915 (initialisation, etc) from functionality dealing with implementations
1916 of specific crypto iterfaces. This change also introduces integrated
1917 support for symmetric ciphers and digest implementations - so ENGINEs
1918 can now accelerate these by providing EVP_CIPHER and EVP_MD
1919 implementations of their own. This is detailed in crypto/engine/README
1920 as it couldn't be adequately described here. However, there are a few
1921 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1922 were changed in the original introduction of ENGINE code have now
1923 reverted back - the hooking from this code to ENGINE is now a good
1924 deal more passive and at run-time, operations deal directly with
1925 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1926 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1927 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1928 they were not being used by the framework as there is no concept of a
1929 BIGNUM_METHOD and they could not be generalised to the new
1930 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1931 ENGINE_cpy() has been removed as it cannot be consistently defined in
1935 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1938 *) Change mkdef.pl to sort symbols that get the same entry number,
1939 and make sure the automatically generated functions ERR_load_*
1940 become part of libeay.num as well.
1943 *) New function SSL_renegotiate_pending(). This returns true once
1944 renegotiation has been requested (either SSL_renegotiate() call
1945 or HelloRequest/ClientHello receveived from the peer) and becomes
1946 false once a handshake has been completed.
1947 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1948 sends a HelloRequest, but does not ensure that a handshake takes
1949 place. SSL_renegotiate_pending() is useful for checking if the
1950 client has followed the request.)
1953 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1954 By default, clients may request session resumption even during
1955 renegotiation (if session ID contexts permit); with this option,
1956 session resumption is possible only in the first handshake.
1958 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1959 more bits available for options that should not be part of
1960 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1963 *) Add some demos for certificate and certificate request creation.
1966 *) Make maximum certificate chain size accepted from the peer application
1967 settable (SSL*_get/set_max_cert_list()), as proposed by
1968 "Douglas E. Engert" <deengert@anl.gov>.
1971 *) Add support for shared libraries for Unixware-7
1972 (Boyd Lynn Gerber <gerberb@zenez.com>).
1975 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1976 be done prior to destruction. Use this to unload error strings from
1977 ENGINEs that load their own error strings. NB: This adds two new API
1978 functions to "get" and "set" this destroy handler in an ENGINE.
1981 *) Alter all existing ENGINE implementations (except "openssl" and
1982 "openbsd") to dynamically instantiate their own error strings. This
1983 makes them more flexible to be built both as statically-linked ENGINEs
1984 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1985 Also, add stub code to each that makes building them as self-contained
1986 shared-libraries easier (see README.ENGINE).
1989 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1990 implementations into applications that are completely implemented in
1991 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1992 commands that can be used to configure what shared-library to load and
1993 to control aspects of the way it is handled. Also, made an update to
1994 the README.ENGINE file that brings its information up-to-date and
1995 provides some information and instructions on the "dynamic" ENGINE
1996 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1999 *) Make it possible to unload ranges of ERR strings with a new
2000 "ERR_unload_strings" function.
2003 *) Add a copy() function to EVP_MD.
2006 *) Make EVP_MD routines take a context pointer instead of just the
2007 md_data void pointer.
2010 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2011 that the digest can only process a single chunk of data
2012 (typically because it is provided by a piece of
2013 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2014 is only going to provide a single chunk of data, and hence the
2015 framework needn't accumulate the data for oneshot drivers.
2018 *) As with "ERR", make it possible to replace the underlying "ex_data"
2019 functions. This change also alters the storage and management of global
2020 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2021 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2022 index counters. The API functions that use this state have been changed
2023 to take a "class_index" rather than pointers to the class's local STACK
2024 and counter, and there is now an API function to dynamically create new
2025 classes. This centralisation allows us to (a) plug a lot of the
2026 thread-safety problems that existed, and (b) makes it possible to clean
2027 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2028 such data would previously have always leaked in application code and
2029 workarounds were in place to make the memory debugging turn a blind eye
2030 to it. Application code that doesn't use this new function will still
2031 leak as before, but their memory debugging output will announce it now
2032 rather than letting it slide.
2034 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2035 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2036 has a return value to indicate success or failure.
2039 *) Make it possible to replace the underlying "ERR" functions such that the
2040 global state (2 LHASH tables and 2 locks) is only used by the "default"
2041 implementation. This change also adds two functions to "get" and "set"
2042 the implementation prior to it being automatically set the first time
2043 any other ERR function takes place. Ie. an application can call "get",
2044 pass the return value to a module it has just loaded, and that module
2045 can call its own "set" function using that value. This means the
2046 module's "ERR" operations will use (and modify) the error state in the
2047 application and not in its own statically linked copy of OpenSSL code.
2050 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2051 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2052 the operation, and provides a more encapsulated way for external code
2053 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2054 to use these functions rather than manually incrementing the counts.
2056 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2059 *) Add EVP test program.
2062 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2065 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2066 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2067 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2068 These allow a CRL to be built without having to access X509_CRL fields
2069 directly. Modify 'ca' application to use new functions.
2072 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2073 bug workarounds. Rollback attack detection is a security feature.
2074 The problem will only arise on OpenSSL servers when TLSv1 is not
2075 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2076 Software authors not wanting to support TLSv1 will have special reasons
2077 for their choice and can explicitly enable this option.
2078 [Bodo Moeller, Lutz Jaenicke]
2080 *) Rationalise EVP so it can be extended: don't include a union of
2081 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2082 (similar to those existing for EVP_CIPHER_CTX).
2087 EVP_MD_CTX_init(&md); /* new function call */
2088 EVP_DigestInit(&md, EVP_sha1());
2089 EVP_DigestUpdate(&md, in, len);
2090 EVP_DigestFinal(&md, out, NULL);
2091 EVP_MD_CTX_cleanup(&md); /* new function call */
2095 *) Make DES key schedule conform to the usual scheme, as well as
2096 correcting its structure. This means that calls to DES functions
2097 now have to pass a pointer to a des_key_schedule instead of a
2098 plain des_key_schedule (which was actually always a pointer
2101 des_key_schedule ks;
2103 des_set_key_checked(..., &ks);
2104 des_ncbc_encrypt(..., &ks, ...);
2106 (Note that a later change renames 'des_...' into 'DES_...'.)
2109 *) Initial reduction of linker bloat: the use of some functions, such as
2110 PEM causes large amounts of unused functions to be linked in due to
2111 poor organisation. For example pem_all.c contains every PEM function
2112 which has a knock on effect of linking in large amounts of (unused)
2113 ASN1 code. Grouping together similar functions and splitting unrelated
2114 functions prevents this.
2117 *) Cleanup of EVP macros.
2120 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2121 correct _ecb suffix.
2124 *) Add initial OCSP responder support to ocsp application. The
2125 revocation information is handled using the text based index
2126 use by the ca application. The responder can either handle
2127 requests generated internally, supplied in files (for example
2128 via a CGI script) or using an internal minimal server.
2131 *) Add configuration choices to get zlib compression for TLS.
2134 *) Changes to Kerberos SSL for RFC 2712 compliance:
2135 1. Implemented real KerberosWrapper, instead of just using
2136 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2137 2. Implemented optional authenticator field of KerberosWrapper.
2139 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2140 and authenticator structs; see crypto/krb5/.
2142 Generalized Kerberos calls to support multiple Kerberos libraries.
2143 [Vern Staats <staatsvr@asc.hpc.mil>,
2144 Jeffrey Altman <jaltman@columbia.edu>
2145 via Richard Levitte]
2147 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2148 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2149 values for each of the key sizes rather than having just
2150 parameters (and 'speed' generating keys each time).
2153 *) Speed up EVP routines.
2156 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2157 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2158 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2159 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2161 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2162 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2163 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2166 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2168 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2171 *) Added the OS2-EMX target.
2172 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2174 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2175 to support NCONF routines in extension code. New function CONF_set_nconf()
2176 to allow functions which take an NCONF to also handle the old LHASH
2177 structure: this means that the old CONF compatible routines can be
2178 retained (in particular wrt extensions) without having to duplicate the
2179 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2182 *) Enhance the general user interface with mechanisms for inner control
2183 and with possibilities to have yes/no kind of prompts.
2186 *) Change all calls to low level digest routines in the library and
2187 applications to use EVP. Add missing calls to HMAC_cleanup() and
2188 don't assume HMAC_CTX can be copied using memcpy().
2189 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2191 *) Add the possibility to control engines through control names but with
2192 arbitrary arguments instead of just a string.
2193 Change the key loaders to take a UI_METHOD instead of a callback
2194 function pointer. NOTE: this breaks binary compatibility with earlier
2195 versions of OpenSSL [engine].
2196 Adapt the nCipher code for these new conditions and add a card insertion
2200 *) Enhance the general user interface with mechanisms to better support
2201 dialog box interfaces, application-defined prompts, the possibility
2202 to use defaults (for example default passwords from somewhere else)
2203 and interrupts/cancellations.
2206 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2207 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2210 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2211 tidy up some unnecessarily weird code in 'sk_new()').
2212 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2214 *) Change the key loading routines for ENGINEs to use the same kind
2215 callback (pem_password_cb) as all other routines that need this
2219 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2220 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2221 than this minimum value is recommended.
2224 *) New random seeder for OpenVMS, using the system process statistics
2225 that are easily reachable.
2228 *) Windows apparently can't transparently handle global
2229 variables defined in DLLs. Initialisations such as:
2231 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2233 wont compile. This is used by the any applications that need to
2234 declare their own ASN1 modules. This was fixed by adding the option
2235 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2236 needed for static libraries under Win32.
2239 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2240 setting of purpose and trust fields. New X509_STORE trust and
2241 purpose functions and tidy up setting in other SSL functions.
2244 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2245 structure. These are inherited by X509_STORE_CTX when it is
2246 initialised. This allows various defaults to be set in the
2247 X509_STORE structure (such as flags for CRL checking and custom
2248 purpose or trust settings) for functions which only use X509_STORE_CTX
2249 internally such as S/MIME.
2251 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2252 trust settings if they are not set in X509_STORE. This allows X509_STORE
2253 purposes and trust (in S/MIME for example) to override any set by default.
2255 Add command line options for CRL checking to smime, s_client and s_server
2259 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2260 are set then the CRL is looked up in the X509_STORE structure and
2261 its validity and signature checked, then if the certificate is found
2262 in the CRL the verify fails with a revoked error.
2264 Various new CRL related callbacks added to X509_STORE_CTX structure.
2266 Command line options added to 'verify' application to support this.
2268 This needs some additional work, such as being able to handle multiple
2269 CRLs with different times, extension based lookup (rather than just
2270 by subject name) and ultimately more complete V2 CRL extension
2274 *) Add a general user interface API (crypto/ui/). This is designed
2275 to replace things like des_read_password and friends (backward
2276 compatibility functions using this new API are provided).
2277 The purpose is to remove prompting functions from the DES code
2278 section as well as provide for prompting through dialog boxes in
2279 a window system and the like.
2282 *) Add "ex_data" support to ENGINE so implementations can add state at a
2283 per-structure level rather than having to store it globally.
2286 *) Make it possible for ENGINE structures to be copied when retrieved by
2287 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2288 This causes the "original" ENGINE structure to act like a template,
2289 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2290 operational state can be localised to each ENGINE structure, despite the
2291 fact they all share the same "methods". New ENGINE structures returned in
2292 this case have no functional references and the return value is the single
2293 structural reference. This matches the single structural reference returned
2294 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2298 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2299 needs to match any other type at all we need to manually clear the
2303 *) Changes to the "openssl engine" utility to include;
2304 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2305 about an ENGINE's available control commands.
2306 - executing control commands from command line arguments using the
2307 '-pre' and '-post' switches. '-post' is only used if '-t' is
2308 specified and the ENGINE is successfully initialised. The syntax for
2309 the individual commands are colon-separated, for example;
2310 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2313 *) New dynamic control command support for ENGINEs. ENGINEs can now
2314 declare their own commands (numbers), names (strings), descriptions,
2315 and input types for run-time discovery by calling applications. A
2316 subset of these commands are implicitly classed as "executable"
2317 depending on their input type, and only these can be invoked through
2318 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2319 can be based on user input, config files, etc). The distinction is
2320 that "executable" commands cannot return anything other than a boolean
2321 result and can only support numeric or string input, whereas some
2322 discoverable commands may only be for direct use through
2323 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2324 pointers, or other custom uses. The "executable" commands are to
2325 support parameterisations of ENGINE behaviour that can be
2326 unambiguously defined by ENGINEs and used consistently across any
2327 OpenSSL-based application. Commands have been added to all the
2328 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2329 control over shared-library paths without source code alterations.
2332 *) Changed all ENGINE implementations to dynamically allocate their
2333 ENGINEs rather than declaring them statically. Apart from this being
2334 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2335 this also allows the implementations to compile without using the
2336 internal engine_int.h header.
2339 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2340 'const' value. Any code that should be able to modify a RAND_METHOD
2341 should already have non-const pointers to it (ie. they should only
2342 modify their own ones).
2345 *) Made a variety of little tweaks to the ENGINE code.
2346 - "atalla" and "ubsec" string definitions were moved from header files
2347 to C code. "nuron" string definitions were placed in variables
2348 rather than hard-coded - allowing parameterisation of these values
2349 later on via ctrl() commands.
2350 - Removed unused "#if 0"'d code.
2351 - Fixed engine list iteration code so it uses ENGINE_free() to release
2352 structural references.
2353 - Constified the RAND_METHOD element of ENGINE structures.
2354 - Constified various get/set functions as appropriate and added
2355 missing functions (including a catch-all ENGINE_cpy that duplicates
2356 all ENGINE values onto a new ENGINE except reference counts/state).
2357 - Removed NULL parameter checks in get/set functions. Setting a method
2358 or function to NULL is a way of cancelling out a previously set
2359 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2360 and doesn't justify the extra error symbols and code.
2361 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2362 flags from engine_int.h to engine.h.
2363 - Changed prototypes for ENGINE handler functions (init(), finish(),
2364 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2367 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2368 to the algorithm using long division. The binary algorithm can be
2369 used only if the modulus is odd. On 32-bit systems, it is faster
2370 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2371 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2372 up to 450 bits. In 64-bit environments, the binary algorithm
2373 appears to be advantageous for much longer moduli; here we use it
2374 for moduli up to 2048 bits.
2377 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2378 could not support the combine flag in choice fields.
2381 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2382 extensions from a certificate request to the certificate.
2385 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2386 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2387 file: this allows the display of the certificate about to be
2388 signed to be customised, to allow certain fields to be included
2389 or excluded and extension details. The old system didn't display
2390 multicharacter strings properly, omitted fields not in the policy
2391 and couldn't display additional details such as extensions.
2394 *) Function EC_POINTs_mul for multiple scalar multiplication
2395 of an arbitrary number of elliptic curve points
2396 \sum scalars[i]*points[i],
2397 optionally including the generator defined for the EC_GROUP:
2398 scalar*generator + \sum scalars[i]*points[i].
2400 EC_POINT_mul is a simple wrapper function for the typical case
2401 that the point list has just one item (besides the optional
2405 *) First EC_METHODs for curves over GF(p):
2407 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2408 operations and provides various method functions that can also
2409 operate with faster implementations of modular arithmetic.
2411 EC_GFp_mont_method() reuses most functions that are part of
2412 EC_GFp_simple_method, but uses Montgomery arithmetic.
2414 [Bodo Moeller; point addition and point doubling
2415 implementation directly derived from source code provided by
2416 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2418 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2419 crypto/ec/ec_lib.c):
2421 Curves are EC_GROUP objects (with an optional group generator)
2422 based on EC_METHODs that are built into the library.
2424 Points are EC_POINT objects based on EC_GROUP objects.
2426 Most of the framework would be able to handle curves over arbitrary
2427 finite fields, but as there are no obvious types for fields other
2428 than GF(p), some functions are limited to that for now.
2431 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2432 that the file contains a complete HTTP response.
2435 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2436 change the def and num file printf format specifier from "%-40sXXX"
2437 to "%-39s XXX". The latter will always guarantee a space after the
2438 field while the former will cause them to run together if the field
2439 is 40 of more characters long.
2442 *) Constify the cipher and digest 'method' functions and structures
2443 and modify related functions to take constant EVP_MD and EVP_CIPHER
2447 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2448 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2451 *) Modify EVP_Digest*() routines so they now return values. Although the
2452 internal software routines can never fail additional hardware versions
2456 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2458 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2459 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2462 ERR_R_NESTED_ASN1_ERROR
2464 ERR_R_MISSING_ASN1_EOS
2465 were 4 .. 9, conflicting with
2466 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2468 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2469 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2471 Add new error code 'ERR_R_INTERNAL_ERROR'.
2474 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2478 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2479 sets the subject name for a new request or supersedes the
2480 subject name in a given request. Formats that can be parsed are
2481 'CN=Some Name, OU=myOU, C=IT'
2483 'CN=Some Name/OU=myOU/C=IT'.
2485 Add options '-batch' and '-verbose' to 'openssl req'.
2486 [Massimiliano Pala <madwolf@hackmasters.net>]
2488 *) Introduce the possibility to access global variables through
2489 functions on platform were that's the best way to handle exporting
2490 global variables in shared libraries. To enable this functionality,
2491 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2492 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2493 is normally done by Configure or something similar).
2495 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2496 in the source file (foo.c) like this:
2498 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2499 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2501 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2502 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2504 OPENSSL_DECLARE_GLOBAL(int,foo);
2505 #define foo OPENSSL_GLOBAL_REF(foo)
2506 OPENSSL_DECLARE_GLOBAL(double,bar);
2507 #define bar OPENSSL_GLOBAL_REF(bar)
2509 The #defines are very important, and therefore so is including the
2510 header file everywhere where the defined globals are used.
2512 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2513 of ASN.1 items, but that structure is a bit different.
2515 The largest change is in util/mkdef.pl which has been enhanced with
2516 better and easier to understand logic to choose which symbols should
2517 go into the Windows .def files as well as a number of fixes and code
2518 cleanup (among others, algorithm keywords are now sorted
2519 lexicographically to avoid constant rewrites).
2522 *) In BN_div() keep a copy of the sign of 'num' before writing the
2523 result to 'rm' because if rm==num the value will be overwritten
2524 and produce the wrong result if 'num' is negative: this caused
2525 problems with BN_mod() and BN_nnmod().
2528 *) Function OCSP_request_verify(). This checks the signature on an
2529 OCSP request and verifies the signer certificate. The signer
2530 certificate is just checked for a generic purpose and OCSP request
2534 *) Add OCSP_check_validity() function to check the validity of OCSP
2535 responses. OCSP responses are prepared in real time and may only
2536 be a few seconds old. Simply checking that the current time lies
2537 between thisUpdate and nextUpdate max reject otherwise valid responses
2538 caused by either OCSP responder or client clock inaccuracy. Instead
2539 we allow thisUpdate and nextUpdate to fall within a certain period of
2540 the current time. The age of the response can also optionally be
2541 checked. Two new options -validity_period and -status_age added to
2545 *) If signature or public key algorithm is unrecognized print out its
2546 OID rather that just UNKNOWN.
2549 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2550 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2551 ID to be generated from the issuer certificate alone which can then be
2552 passed to OCSP_id_issuer_cmp().
2555 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2556 ASN1 modules to export functions returning ASN1_ITEM pointers
2557 instead of the ASN1_ITEM structures themselves. This adds several
2558 new macros which allow the underlying ASN1 function/structure to
2559 be accessed transparently. As a result code should not use ASN1_ITEM
2560 references directly (such as &X509_it) but instead use the relevant
2561 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2562 use of the new ASN1 code on platforms where exporting structures
2563 is problematical (for example in shared libraries) but exporting
2564 functions returning pointers to structures is not.
2567 *) Add support for overriding the generation of SSL/TLS session IDs.
2568 These callbacks can be registered either in an SSL_CTX or per SSL.
2569 The purpose of this is to allow applications to control, if they wish,
2570 the arbitrary values chosen for use as session IDs, particularly as it
2571 can be useful for session caching in multiple-server environments. A
2572 command-line switch for testing this (and any client code that wishes
2573 to use such a feature) has been added to "s_server".
2574 [Geoff Thorpe, Lutz Jaenicke]
2576 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2577 of the form '#if defined(...) || defined(...) || ...' and
2578 '#if !defined(...) && !defined(...) && ...'. This also avoids
2579 the growing number of special cases it was previously handling.
2582 *) Make all configuration macros available for application by making
2583 sure they are available in opensslconf.h, by giving them names starting
2584 with "OPENSSL_" to avoid conflicts with other packages and by making
2585 sure e_os2.h will cover all platform-specific cases together with
2587 Additionally, it is now possible to define configuration/platform-
2588 specific names (called "system identities"). In the C code, these
2589 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2590 macro with the name beginning with "OPENSSL_SYS_", which is determined
2591 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2595 *) New option -set_serial to 'req' and 'x509' this allows the serial
2596 number to use to be specified on the command line. Previously self
2597 signed certificates were hard coded with serial number 0 and the
2598 CA options of 'x509' had to use a serial number in a file which was
2602 *) New options to 'ca' utility to support V2 CRL entry extensions.
2603 Currently CRL reason, invalidity date and hold instruction are
2604 supported. Add new CRL extensions to V3 code and some new objects.
2607 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2608 disable standard block padding (aka PKCS#5 padding) in the EVP
2609 API, which was previously mandatory. This means that the data is
2610 not padded in any way and so the total length much be a multiple
2611 of the block size, otherwise an error occurs.
2614 *) Initial (incomplete) OCSP SSL support.
2617 *) New function OCSP_parse_url(). This splits up a URL into its host,
2618 port and path components: primarily to parse OCSP URLs. New -url
2619 option to ocsp utility.
2622 *) New nonce behavior. The return value of OCSP_check_nonce() now
2623 reflects the various checks performed. Applications can decide
2624 whether to tolerate certain situations such as an absent nonce
2625 in a response when one was present in a request: the ocsp application
2626 just prints out a warning. New function OCSP_add1_basic_nonce()
2627 this is to allow responders to include a nonce in a response even if
2628 the request is nonce-less.
2631 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2632 skipped when using openssl x509 multiple times on a single input file,
2633 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2636 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2637 set string type: to handle setting ASN1_TIME structures. Fix ca
2638 utility to correctly initialize revocation date of CRLs.
2641 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2642 the clients preferred ciphersuites and rather use its own preferences.
2643 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2644 Internet Explorer by ensuring unchanged hash method during stepup.
2645 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2648 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2649 to aes and add a new 'exist' option to print out symbols that don't
2653 *) Additional options to ocsp utility to allow flags to be set and
2654 additional certificates supplied.
2657 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2658 OCSP client a number of certificate to only verify the response
2662 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2663 handle the new API. Currently only ECB, CBC modes supported. Add new
2666 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2667 Encryption Standard (AES) Ciphersuites for Transport Layer
2668 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2669 not enabled by default and were not part of the "ALL" ciphersuite
2670 alias because they were not yet official; they could be
2671 explicitly requested by specifying the "AESdraft" ciphersuite
2672 group alias. In the final release of OpenSSL 0.9.7, the group
2673 alias is called "AES" and is part of "ALL".)
2674 [Ben Laurie, Steve Henson, Bodo Moeller]
2676 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2677 request to response.
2680 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2681 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2682 extract information from a certificate request. OCSP_response_create()
2683 creates a response and optionally adds a basic response structure.
2684 OCSP_basic_add1_status() adds a complete single response to a basic
2685 response and returns the OCSP_SINGLERESP structure just added (to allow
2686 extensions to be included for example). OCSP_basic_add1_cert() adds a
2687 certificate to a basic response and OCSP_basic_sign() signs a basic
2688 response with various flags. New helper functions ASN1_TIME_check()
2689 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2690 (converts ASN1_TIME to GeneralizedTime).
2693 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2694 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2695 structure from a certificate. X509_pubkey_digest() digests the public_key
2696 contents: this is used in various key identifiers.
2699 *) Make sk_sort() tolerate a NULL argument.
2700 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2702 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2703 passed by the function are trusted implicitly. If any of them signed the
2704 response then it is assumed to be valid and is not verified.
2707 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2708 to data. This was previously part of the PKCS7 ASN1 code. This
2709 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2710 [Steve Henson, reported by Kenneth R. Robinette
2711 <support@securenetterm.com>]
2713 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2714 routines: without these tracing memory leaks is very painful.
2715 Fix leaks in PKCS12 and PKCS7 routines.
2718 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2719 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2720 effectively meant GeneralizedTime would never be used. Now it
2721 is initialised to -1 but X509_time_adj() now has to check the value
2722 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2723 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2724 [Steve Henson, reported by Kenneth R. Robinette
2725 <support@securenetterm.com>]
2727 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2728 result in a zero length in the ASN1_INTEGER structure which was
2729 not consistent with the structure when d2i_ASN1_INTEGER() was used
2730 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2731 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2732 where it did not print out a minus for negative ASN1_INTEGER.
2735 *) Add summary printout to ocsp utility. The various functions which
2736 convert status values to strings have been renamed to:
2737 OCSP_response_status_str(), OCSP_cert_status_str() and
2738 OCSP_crl_reason_str() and are no longer static. New options
2739 to verify nonce values and to disable verification. OCSP response
2740 printout format cleaned up.
2743 *) Add additional OCSP certificate checks. These are those specified
2744 in RFC2560. This consists of two separate checks: the CA of the
2745 certificate being checked must either be the OCSP signer certificate
2746 or the issuer of the OCSP signer certificate. In the latter case the
2747 OCSP signer certificate must contain the OCSP signing extended key
2748 usage. This check is performed by attempting to match the OCSP
2749 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2750 in the OCSP_CERTID structures of the response.
2753 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2754 and related routines. This uses the standard OpenSSL certificate
2755 verify routines to perform initial checks (just CA validity) and
2756 to obtain the certificate chain. Then additional checks will be
2757 performed on the chain. Currently the root CA is checked to see
2758 if it is explicitly trusted for OCSP signing. This is used to set
2759 a root CA as a global signing root: that is any certificate that
2760 chains to that CA is an acceptable OCSP signing certificate.
2763 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2764 extensions from a separate configuration file.
2765 As when reading extensions from the main configuration file,
2766 the '-extensions ...' option may be used for specifying the
2768 [Massimiliano Pala <madwolf@comune.modena.it>]
2770 *) New OCSP utility. Allows OCSP requests to be generated or
2771 read. The request can be sent to a responder and the output
2772 parsed, outputed or printed in text form. Not complete yet:
2773 still needs to check the OCSP response validity.
2776 *) New subcommands for 'openssl ca':
2777 'openssl ca -status <serial>' prints the status of the cert with
2778 the given serial number (according to the index file).
2779 'openssl ca -updatedb' updates the expiry status of certificates
2781 [Massimiliano Pala <madwolf@comune.modena.it>]
2783 *) New '-newreq-nodes' command option to CA.pl. This is like
2784 '-newreq', but calls 'openssl req' with the '-nodes' option
2785 so that the resulting key is not encrypted.
2786 [Damien Miller <djm@mindrot.org>]
2788 *) New configuration for the GNU Hurd.
2789 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2791 *) Initial code to implement OCSP basic response verify. This
2792 is currently incomplete. Currently just finds the signer's
2793 certificate and verifies the signature on the response.
2796 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2797 value of OPENSSLDIR. This is available via the new '-d' option
2798 to 'openssl version', and is also included in 'openssl version -a'.
2801 *) Allowing defining memory allocation callbacks that will be given
2802 file name and line number information in additional arguments
2803 (a const char* and an int). The basic functionality remains, as
2804 well as the original possibility to just replace malloc(),
2805 realloc() and free() by functions that do not know about these
2806 additional arguments. To register and find out the current
2807 settings for extended allocation functions, the following
2808 functions are provided:
2810 CRYPTO_set_mem_ex_functions
2811 CRYPTO_set_locked_mem_ex_functions
2812 CRYPTO_get_mem_ex_functions
2813 CRYPTO_get_locked_mem_ex_functions
2815 These work the same way as CRYPTO_set_mem_functions and friends.
2816 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2817 extended allocation function is enabled.
2818 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2819 a conventional allocation function is enabled.
2820 [Richard Levitte, Bodo Moeller]
2822 *) Finish off removing the remaining LHASH function pointer casts.
2823 There should no longer be any prototype-casting required when using
2824 the LHASH abstraction, and any casts that remain are "bugs". See
2825 the callback types and macros at the head of lhash.h for details
2826 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2829 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2830 If /dev/[u]random devices are not available or do not return enough
2831 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2833 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2834 /etc/entropy will be queried once each in this sequence, quering stops
2835 when enough entropy was collected without querying more sockets.
2838 *) Change the Unix RAND_poll() variant to be able to poll several
2839 random devices, as specified by DEVRANDOM, until a sufficient amount
2840 of data has been collected. We spend at most 10 ms on each file
2841 (select timeout) and read in non-blocking mode. DEVRANDOM now
2842 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2843 (previously it was just the string "/dev/urandom"), so on typical
2844 platforms the 10 ms delay will never occur.
2845 Also separate out the Unix variant to its own file, rand_unix.c.
2846 For VMS, there's a currently-empty rand_vms.c.
2849 *) Move OCSP client related routines to ocsp_cl.c. These
2850 provide utility functions which an application needing
2851 to issue a request to an OCSP responder and analyse the
2852 response will typically need: as opposed to those which an
2853 OCSP responder itself would need which will be added later.
2855 OCSP_request_sign() signs an OCSP request with an API similar
2856 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2857 response. OCSP_response_get1_basic() extracts basic response
2858 from response. OCSP_resp_find_status(): finds and extracts status
2859 information from an OCSP_CERTID structure (which will be created
2860 when the request structure is built). These are built from lower
2861 level functions which work on OCSP_SINGLERESP structures but
2862 wont normally be used unless the application wishes to examine
2863 extensions in the OCSP response for example.
2865 Replace nonce routines with a pair of functions.
2866 OCSP_request_add1_nonce() adds a nonce value and optionally
2867 generates a random value. OCSP_check_nonce() checks the
2868 validity of the nonce in an OCSP response.
2871 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2872 This doesn't copy the supplied OCSP_CERTID and avoids the
2873 need to free up the newly created id. Change return type
2874 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2875 This can then be used to add extensions to the request.
2876 Deleted OCSP_request_new(), since most of its functionality
2877 is now in OCSP_REQUEST_new() (and the case insensitive name
2878 clash) apart from the ability to set the request name which
2879 will be added elsewhere.
2882 *) Update OCSP API. Remove obsolete extensions argument from
2883 various functions. Extensions are now handled using the new
2884 OCSP extension code. New simple OCSP HTTP function which
2885 can be used to send requests and parse the response.
2888 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2889 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2890 uses the special reorder version of SET OF to sort the attributes
2891 and reorder them to match the encoded order. This resolves a long
2892 standing problem: a verify on a PKCS7 structure just after signing
2893 it used to fail because the attribute order did not match the
2894 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2895 it uses the received order. This is necessary to tolerate some broken
2896 software that does not order SET OF. This is handled by encoding
2897 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2898 to produce the required SET OF.
2901 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2902 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2903 files to get correct declarations of the ASN.1 item variables.
2906 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2907 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2908 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2909 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2910 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2911 ASN1_ITEM and no wrapper functions.
2914 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2915 replace the old function pointer based I/O routines. Change most of
2916 the *_d2i_bio() and *_d2i_fp() functions to use these.
2919 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2920 lines, recognice more "algorithms" that can be deselected, and make
2921 it complain about algorithm deselection that isn't recognised.
2924 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2925 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2926 to use new functions. Add NO_ASN1_OLD which can be set to remove
2927 some old style ASN1 functions: this can be used to determine if old
2928 code will still work when these eventually go away.
2931 *) New extension functions for OCSP structures, these follow the
2932 same conventions as certificates and CRLs.
2935 *) New function X509V3_add1_i2d(). This automatically encodes and
2936 adds an extension. Its behaviour can be customised with various
2937 flags to append, replace or delete. Various wrappers added for
2938 certifcates and CRLs.
2941 *) Fix to avoid calling the underlying ASN1 print routine when
2942 an extension cannot be parsed. Correct a typo in the
2943 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2946 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2947 entries for variables.
2950 *) Add functionality to apps/openssl.c for detecting locking
2951 problems: As the program is single-threaded, all we have
2952 to do is register a locking callback using an array for
2953 storing which locks are currently held by the program.
2956 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2957 SSL_get_ex_data_X509_STORE_idx(), which is used in
2958 ssl_verify_cert_chain() and thus can be called at any time
2959 during TLS/SSL handshakes so that thread-safety is essential.
2960 Unfortunately, the ex_data design is not at all suited
2961 for multi-threaded use, so it probably should be abolished.
2964 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2965 [Broadcom, tweaked and integrated by Geoff Thorpe]
2967 *) Move common extension printing code to new function
2968 X509V3_print_extensions(). Reorganise OCSP print routines and
2969 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2972 *) New function X509_signature_print() to remove duplication in some
2976 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2977 set (this was treated exactly the same as SET OF previously). This
2978 is used to reorder the STACK representing the structure to match the
2979 encoding. This will be used to get round a problem where a PKCS7
2980 structure which was signed could not be verified because the STACK
2981 order did not reflect the encoded order.
2984 *) Reimplement the OCSP ASN1 module using the new code.
2987 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2988 for its ASN1 operations. The old style function pointers still exist
2989 for now but they will eventually go away.
2992 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2993 completely replaces the old ASN1 functionality with a table driven
2994 encoder and decoder which interprets an ASN1_ITEM structure describing
2995 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2996 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2997 has also been converted to the new form.
3000 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3001 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3002 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3003 for negative moduli.
3006 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3007 of not touching the result's sign bit.
3010 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3014 *) Changed the LHASH code to use prototypes for callbacks, and created
3015 macros to declare and implement thin (optionally static) functions
3016 that provide type-safety and avoid function pointer casting for the
3017 type-specific callbacks.
3020 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3022 [Veers Staats <staatsvr@asc.hpc.mil>,
3023 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3025 *) Reformat the FAQ so the different questions and answers can be divided
3026 in sections depending on the subject.
3029 *) Have the zlib compression code load ZLIB.DLL dynamically under
3033 *) New function BN_mod_sqrt for computing square roots modulo a prime
3034 (using the probabilistic Tonelli-Shanks algorithm unless
3035 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3036 be handled deterministically).
3037 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3039 *) Make BN_mod_inverse faster by explicitly handling small quotients
3040 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3041 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3044 *) New function BN_kronecker.
3047 *) Fix BN_gcd so that it works on negative inputs; the result is
3048 positive unless both parameters are zero.
3049 Previously something reasonably close to an infinite loop was
3050 possible because numbers could be growing instead of shrinking
3051 in the implementation of Euclid's algorithm.
3054 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3055 sign of the number in question.
3057 Fix BN_is_word(a,w) to work correctly for w == 0.
3059 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3060 because its test if the absolute value of 'a' equals 'w'.
3061 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3062 it exists mostly for use in the implementations of BN_is_zero(),
3063 BN_is_one(), and BN_is_word().
3066 *) New function BN_swap.
3069 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3070 the exponentiation functions are more likely to produce reasonable
3071 results on negative inputs.
3074 *) Change BN_mod_mul so that the result is always non-negative.
3075 Previously, it could be negative if one of the factors was negative;
3076 I don't think anyone really wanted that behaviour.
3079 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3080 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3081 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3082 and add new functions:
3091 BN_mod_lshift1_quick
3095 These functions always generate non-negative results.
3097 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3098 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3100 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3101 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3102 be reduced modulo m.
3103 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3106 The following entry accidentily appeared in the CHANGES file
3107 distributed with OpenSSL 0.9.7. The modifications described in
3108 it do *not* apply to OpenSSL 0.9.7.
3110 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3111 was actually never needed) and in BN_mul(). The removal in BN_mul()
3112 required a small change in bn_mul_part_recursive() and the addition
3113 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3114 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3115 bn_sub_words() and bn_add_words() except they take arrays with
3120 *) In 'openssl passwd', verify passwords read from the terminal
3121 unless the '-salt' option is used (which usually means that
3122 verification would just waste user's time since the resulting
3123 hash is going to be compared with some given password hash)
3124 or the new '-noverify' option is used.
3126 This is an incompatible change, but it does not affect
3127 non-interactive use of 'openssl passwd' (passwords on the command
3128 line, '-stdin' option, '-in ...' option) and thus should not
3132 *) Remove all references to RSAref, since there's no more need for it.
3135 *) Make DSO load along a path given through an environment variable
3136 (SHLIB_PATH) with shl_load().
3139 *) Constify the ENGINE code as a result of BIGNUM constification.
3140 Also constify the RSA code and most things related to it. In a
3141 few places, most notable in the depth of the ASN.1 code, ugly
3142 casts back to non-const were required (to be solved at a later
3146 *) Make it so the openssl application has all engines loaded by default.
3149 *) Constify the BIGNUM routines a little more.
3152 *) Add the following functions:
3154 ENGINE_load_cswift()
3156 ENGINE_load_atalla()
3158 ENGINE_load_builtin_engines()
3160 That way, an application can itself choose if external engines that
3161 are built-in in OpenSSL shall ever be used or not. The benefit is
3162 that applications won't have to be linked with libdl or other dso
3163 libraries unless it's really needed.
3165 Changed 'openssl engine' to load all engines on demand.
3166 Changed the engine header files to avoid the duplication of some
3167 declarations (they differed!).
3170 *) 'openssl engine' can now list capabilities.
3173 *) Better error reporting in 'openssl engine'.
3176 *) Never call load_dh_param(NULL) in s_server.
3179 *) Add engine application. It can currently list engines by name and
3180 identity, and test if they are actually available.
3183 *) Improve RPM specification file by forcing symbolic linking and making
3184 sure the installed documentation is also owned by root.root.
3185 [Damien Miller <djm@mindrot.org>]
3187 *) Give the OpenSSL applications more possibilities to make use of
3188 keys (public as well as private) handled by engines.
3191 *) Add OCSP code that comes from CertCo.
3194 *) Add VMS support for the Rijndael code.
3197 *) Added untested support for Nuron crypto accelerator.
3200 *) Add support for external cryptographic devices. This code was
3201 previously distributed separately as the "engine" branch.
3202 [Geoff Thorpe, Richard Levitte]
3204 *) Rework the filename-translation in the DSO code. It is now possible to
3205 have far greater control over how a "name" is turned into a filename
3206 depending on the operating environment and any oddities about the
3207 different shared library filenames on each system.
3210 *) Support threads on FreeBSD-elf in Configure.
3213 *) Fix for SHA1 assembly problem with MASM: it produces
3214 warnings about corrupt line number information when assembling
3215 with debugging information. This is caused by the overlapping
3217 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3220 NCONF_get_number() has no error checking at all. As a replacement,
3221 NCONF_get_number_e() is defined (_e for "error checking") and is
3222 promoted strongly. The old NCONF_get_number is kept around for
3223 binary backward compatibility.
3224 Make it possible for methods to load from something other than a BIO,
3225 by providing a function pointer that is given a name instead of a BIO.
3226 For example, this could be used to load configuration data from an
3230 *) Fix for non blocking accept BIOs. Added new I/O special reason
3231 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3232 with non blocking I/O was not possible because no retry code was
3233 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3237 *) Added the beginnings of Rijndael support.
3240 *) Fix for bug in DirectoryString mask setting. Add support for
3241 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3242 to allow certificate printing to more controllable, additional
3243 'certopt' option to 'x509' to allow new printing options to be
3247 *) Clean old EAY MD5 hack from e_os.h.
3250 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3252 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3253 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3254 [Joe Orton, Steve Henson]
3256 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3258 *) Fix additional bug revealed by the NISCC test suite:
3260 Stop bug triggering large recursion when presented with
3261 certain ASN.1 tags (CVE-2003-0851)
3264 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3266 *) Fix various bugs revealed by running the NISCC test suite:
3268 Stop out of bounds reads in the ASN1 code when presented with
3269 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3271 If verify callback ignores invalid public key errors don't try to check
3272 certificate signature with the NULL public key.
3276 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3277 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3281 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3282 extra data after the compression methods not only for TLS 1.0
3283 but also for SSL 3.0 (as required by the specification).
3284 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3286 *) Change X509_certificate_type() to mark the key as exported/exportable
3287 when it's 512 *bits* long, not 512 bytes.
3290 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3292 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3293 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3294 a protocol version number mismatch like a decryption error
3295 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3298 *) Turn on RSA blinding by default in the default implementation
3299 to avoid a timing attack. Applications that don't want it can call
3300 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3301 They would be ill-advised to do so in most cases.
3302 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3304 *) Change RSA blinding code so that it works when the PRNG is not
3305 seeded (in this case, the secret RSA exponent is abused as
3306 an unpredictable seed -- if it is not unpredictable, there
3307 is no point in blinding anyway). Make RSA blinding thread-safe
3308 by remembering the creator's thread ID in rsa->blinding and
3309 having all other threads use local one-time blinding factors
3310 (this requires more computation than sharing rsa->blinding, but
3311 avoids excessive locking; and if an RSA object is not shared
3312 between threads, blinding will still be very fast).
3315 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3317 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3318 via timing by performing a MAC computation even if incorrrect
3319 block cipher padding has been found. This is a countermeasure
3320 against active attacks where the attacker has to distinguish
3321 between bad padding and a MAC verification error. (CVE-2003-0078)
3323 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3324 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3325 Martin Vuagnoux (EPFL, Ilion)]
3327 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3329 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3330 memory from it's contents. This is done with a counter that will
3331 place alternating values in each byte. This can be used to solve
3332 two issues: 1) the removal of calls to memset() by highly optimizing
3333 compilers, and 2) cleansing with other values than 0, since those can
3334 be read through on certain media, for example a swap space on disk.
3337 *) Bugfix: client side session caching did not work with external caching,
3338 because the session->cipher setting was not restored when reloading
3339 from the external cache. This problem was masked, when
3340 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3341 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3344 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3345 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3346 [Zeev Lieber <zeev-l@yahoo.com>]
3348 *) Undo an undocumented change introduced in 0.9.6e which caused
3349 repeated calls to OpenSSL_add_all_ciphers() and
3350 OpenSSL_add_all_digests() to be ignored, even after calling
3354 *) Change the default configuration reader to deal with last line not
3355 being properly terminated.
3358 *) Change X509_NAME_cmp() so it applies the special rules on handling
3359 DN values that are of type PrintableString, as well as RDNs of type
3360 emailAddress where the value has the type ia5String.
3361 [stefank@valicert.com via Richard Levitte]
3363 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3364 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3365 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3366 the bitwise-OR of the two for use by the majority of applications
3367 wanting this behaviour, and update the docs. The documented
3368 behaviour and actual behaviour were inconsistent and had been
3369 changing anyway, so this is more a bug-fix than a behavioural
3371 [Geoff Thorpe, diagnosed by Nadav Har'El]
3373 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3374 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3377 *) Fix initialization code race conditions in
3378 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3379 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3380 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3381 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3382 ssl2_get_cipher_by_char(),
3383 ssl3_get_cipher_by_char().
3384 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3386 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3387 the cached sessions are flushed, as the remove_cb() might use ex_data
3388 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3389 (see [openssl.org #212]).
3390 [Geoff Thorpe, Lutz Jaenicke]
3392 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3393 length, instead of the encoding length to d2i_ASN1_OBJECT.
3396 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3398 *) [In 0.9.6g-engine release:]
3399 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3400 [Lynn Gazis <lgazis@rainbow.com>]
3402 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3404 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3405 and get fix the header length calculation.
3406 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3407 Alon Kantor <alonk@checkpoint.com> (and others),
3410 *) Use proper error handling instead of 'assertions' in buffer
3411 overflow checks added in 0.9.6e. This prevents DoS (the
3412 assertions could call abort()).
3413 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3415 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3417 *) Add various sanity checks to asn1_get_length() to reject
3418 the ASN1 length bytes if they exceed sizeof(long), will appear
3419 negative or the content length exceeds the length of the
3421 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3423 *) Fix cipher selection routines: ciphers without encryption had no flags
3424 for the cipher strength set and where therefore not handled correctly
3425 by the selection routines (PR #130).
3428 *) Fix EVP_dsa_sha macro.
3432 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3433 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3434 that was added in OpenSSL 0.9.6d.
3436 As the countermeasure turned out to be incompatible with some
3437 broken SSL implementations, the new option is part of SSL_OP_ALL.
3438 SSL_OP_ALL is usually employed when compatibility with weird SSL
3439 implementations is desired (e.g. '-bugs' option to 's_client' and
3440 's_server'), so the new option is automatically set in many
3444 *) Changes in security patch:
3446 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3447 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3448 Air Force Materiel Command, USAF, under agreement number
3451 *) Add various sanity checks to asn1_get_length() to reject
3452 the ASN1 length bytes if they exceed sizeof(long), will appear
3453 negative or the content length exceeds the length of the
3454 supplied buffer. (CVE-2002-0659)
3455 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3457 *) Assertions for various potential buffer overflows, not known to
3459 [Ben Laurie (CHATS)]
3461 *) Various temporary buffers to hold ASCII versions of integers were
3462 too small for 64 bit platforms. (CVE-2002-0655)
3463 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3465 *) Remote buffer overflow in SSL3 protocol - an attacker could
3466 supply an oversized session ID to a client. (CVE-2002-0656)
3467 [Ben Laurie (CHATS)]
3469 *) Remote buffer overflow in SSL2 protocol - an attacker could
3470 supply an oversized client master key. (CVE-2002-0656)
3471 [Ben Laurie (CHATS)]
3473 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3475 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3476 encoded as NULL) with id-dsa-with-sha1.
3477 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3479 *) Check various X509_...() return values in apps/req.c.
3480 [Nils Larsch <nla@trustcenter.de>]
3482 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3483 an end-of-file condition would erronously be flagged, when the CRLF
3484 was just at the end of a processed block. The bug was discovered when
3485 processing data through a buffering memory BIO handing the data to a
3486 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3487 <ptsekov@syntrex.com> and Nedelcho Stanev.
3490 *) Implement a countermeasure against a vulnerability recently found
3491 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3492 before application data chunks to avoid the use of known IVs
3493 with data potentially chosen by the attacker.
3496 *) Fix length checks in ssl3_get_client_hello().
3499 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3500 to prevent ssl3_read_internal() from incorrectly assuming that
3501 ssl3_read_bytes() found application data while handshake
3502 processing was enabled when in fact s->s3->in_read_app_data was
3503 merely automatically cleared during the initial handshake.
3504 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3506 *) Fix object definitions for Private and Enterprise: they were not
3507 recognized in their shortname (=lowercase) representation. Extend
3508 obj_dat.pl to issue an error when using undefined keywords instead
3509 of silently ignoring the problem (Svenning Sorensen
3510 <sss@sss.dnsalias.net>).
3513 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3514 generators, i.e. generators other than 2 and 5. (Previously, the
3515 code did not properly initialise the 'add' and 'rem' values to
3516 BN_generate_prime().)
3518 In the new general case, we do not insist that 'generator' is
3519 actually a primitive root: This requirement is rather pointless;
3520 a generator of the order-q subgroup is just as good, if not
3524 *) Map new X509 verification errors to alerts. Discovered and submitted by
3525 Tom Wu <tom@arcot.com>.
3528 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3529 returning non-zero before the data has been completely received
3530 when using non-blocking I/O.
3531 [Bodo Moeller; problem pointed out by John Hughes]
3533 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3534 [Ben Laurie, Lutz Jaenicke]
3536 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3537 Yoram Zahavi <YoramZ@gilian.com>).
3540 *) Add information about CygWin 1.3 and on, and preserve proper
3541 configuration for the versions before that.
3542 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3544 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3545 check whether we deal with a copy of a session and do not delete from
3546 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3547 <izhar@checkpoint.com>.
3550 *) Do not store session data into the internal session cache, if it
3551 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3552 flag is set). Proposed by Aslam <aslam@funk.com>.
3555 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3559 *) [In 0.9.6d-engine release:]
3560 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3561 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3563 *) Add the configuration target linux-s390x.
3564 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3566 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3567 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3568 variable as an indication that a ClientHello message has been
3569 received. As the flag value will be lost between multiple
3570 invocations of ssl3_accept when using non-blocking I/O, the
3571 function may not be aware that a handshake has actually taken
3572 place, thus preventing a new session from being added to the
3575 To avoid this problem, we now set s->new_session to 2 instead of
3576 using a local variable.
3577 [Lutz Jaenicke, Bodo Moeller]
3579 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3580 if the SSL_R_LENGTH_MISMATCH error is detected.
3581 [Geoff Thorpe, Bodo Moeller]
3583 *) New 'shared_ldflag' column in Configure platform table.
3586 *) Fix EVP_CIPHER_mode macro.
3587 ["Dan S. Camper" <dan@bti.net>]
3589 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3590 type, we must throw them away by setting rr->length to 0.
3591 [D P Chang <dpc@qualys.com>]
3593 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3595 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3596 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3597 worked incorrectly for those cases where range = 10..._2 and
3598 3*range is two bits longer than range.)
3601 *) Only add signing time to PKCS7 structures if it is not already
3605 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3606 OBJ_ld_ce should be OBJ_id_ce.
3607 Also some ip-pda OIDs in crypto/objects/objects.txt were
3608 incorrect (cf. RFC 3039).
3609 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3611 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3612 returns early because it has nothing to do.
3613 [Andy Schneider <andy.schneider@bjss.co.uk>]
3615 *) [In 0.9.6c-engine release:]
3616 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3617 [Andy Schneider <andy.schneider@bjss.co.uk>]
3619 *) [In 0.9.6c-engine release:]
3620 Add support for Cryptographic Appliance's keyserver technology.
3621 (Use engine 'keyclient')
3622 [Cryptographic Appliances and Geoff Thorpe]
3624 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3625 is called via tools/c89.sh because arguments have to be
3626 rearranged (all '-L' options must appear before the first object
3628 [Richard Shapiro <rshapiro@abinitio.com>]
3630 *) [In 0.9.6c-engine release:]
3631 Add support for Broadcom crypto accelerator cards, backported
3633 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3635 *) [In 0.9.6c-engine release:]
3636 Add support for SureWare crypto accelerator cards from
3637 Baltimore Technologies. (Use engine 'sureware')
3638 [Baltimore Technologies and Mark Cox]
3640 *) [In 0.9.6c-engine release:]
3641 Add support for crypto accelerator cards from Accelerated
3642 Encryption Processing, www.aep.ie. (Use engine 'aep')
3643 [AEP Inc. and Mark Cox]
3645 *) Add a configuration entry for gcc on UnixWare.
3646 [Gary Benson <gbenson@redhat.com>]
3648 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3649 messages are stored in a single piece (fixed-length part and
3650 variable-length part combined) and fix various bugs found on the way.
3653 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3654 instead. BIO_gethostbyname() does not know what timeouts are
3655 appropriate, so entries would stay in cache even when they have
3657 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3659 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3660 faced with a pathologically small ClientHello fragment that does
3661 not contain client_version: Instead of aborting with an error,
3662 simply choose the highest available protocol version (i.e.,
3663 TLS 1.0 unless it is disabled). In practice, ClientHello
3664 messages are never sent like this, but this change gives us
3665 strictly correct behaviour at least for TLS.
3668 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3669 never resets s->method to s->ctx->method when called from within
3670 one of the SSL handshake functions.
3671 [Bodo Moeller; problem pointed out by Niko Baric]
3673 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3674 (sent using the client's version number) if client_version is
3675 smaller than the protocol version in use. Also change
3676 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3677 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3678 the client will at least see that alert.
3681 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3685 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3686 client receives HelloRequest while in a handshake.
3687 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3689 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3690 should end in 'break', not 'goto end' which circuments various
3691 cleanups done in state SSL_ST_OK. But session related stuff
3692 must be disabled for SSL_ST_OK in the case that we just sent a
3695 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3696 before just sending a HelloRequest.
3697 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3699 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3700 reveal whether illegal block cipher padding was found or a MAC
3701 verification error occured. (Neither SSLerr() codes nor alerts
3702 are directly visible to potential attackers, but the information
3703 may leak via logfiles.)
3705 Similar changes are not required for the SSL 2.0 implementation
3706 because the number of padding bytes is sent in clear for SSL 2.0,
3707 and the extra bytes are just ignored. However ssl/s2_pkt.c
3708 failed to verify that the purported number of padding bytes is in
3712 *) Add OpenUNIX-8 support including shared libraries
3713 (Boyd Lynn Gerber <gerberb@zenez.com>).
3716 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3717 'wristwatch attack' using huge encoding parameters (cf.
3718 James H. Manger's CRYPTO 2001 paper). Note that the
3719 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3720 encoding parameters and hence was not vulnerable.
3723 *) BN_sqr() bug fix.
3724 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3726 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3727 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3728 followed by modular reduction.
3729 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3731 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3732 equivalent based on BN_pseudo_rand() instead of BN_rand().
3735 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3736 This function was broken, as the check for a new client hello message
3737 to handle SGC did not allow these large messages.
3738 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3741 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3744 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3745 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3748 *) Rework the configuration and shared library support for Tru64 Unix.
3749 The configuration part makes use of modern compiler features and
3750 still retains old compiler behavior for those that run older versions
3751 of the OS. The shared library support part includes a variant that
3752 uses the RPATH feature, and is available through the special
3753 configuration target "alpha-cc-rpath", which will never be selected
3755 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3757 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3758 with the same message size as in ssl3_get_certificate_request().
3759 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3760 messages might inadvertently be reject as too long.
3761 [Petr Lampa <lampa@fee.vutbr.cz>]
3763 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3766 *) Modified SSL library such that the verify_callback that has been set
3767 specificly for an SSL object with SSL_set_verify() is actually being
3768 used. Before the change, a verify_callback set with this function was
3769 ignored and the verify_callback() set in the SSL_CTX at the time of
3770 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3771 to allow the necessary settings.
3774 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3775 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3776 done automatically (in contradiction to the requirements of the C
3777 standard). This made problems when used from OpenSSH.
3780 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3781 dh->length and always used
3783 BN_rand_range(priv_key, dh->p).
3785 BN_rand_range() is not necessary for Diffie-Hellman, and this
3786 specific range makes Diffie-Hellman unnecessarily inefficient if
3787 dh->length (recommended exponent length) is much smaller than the
3788 length of dh->p. We could use BN_rand_range() if the order of
3789 the subgroup was stored in the DH structure, but we only have
3794 BN_rand(priv_key, l, ...)
3796 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3802 RSA_eay_public_encrypt
3803 RSA_eay_private_decrypt
3804 RSA_eay_private_encrypt (signing)
3805 RSA_eay_public_decrypt (signature verification)
3807 (default implementations for RSA_public_encrypt,
3808 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3809 always reject numbers >= n.
3812 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3813 to synchronize access to 'locking_thread'. This is necessary on
3814 systems where access to 'locking_thread' (an 'unsigned long'
3815 variable) is not atomic.
3818 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3819 *before* setting the 'crypto_lock_rand' flag. The previous code had
3820 a race condition if 0 is a valid thread ID.
3821 [Travis Vitek <vitek@roguewave.com>]
3823 *) Add support for shared libraries under Irix.
3824 [Albert Chin-A-Young <china@thewrittenword.com>]
3826 *) Add configuration option to build on Linux on both big-endian and
3828 [Ralf Baechle <ralf@uni-koblenz.de>]
3830 *) Add the possibility to create shared libraries on HP-UX.
3833 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3835 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3836 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3837 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3838 PRNG state recovery was possible based on the output of
3839 one PRNG request appropriately sized to gain knowledge on
3840 'md' followed by enough consecutive 1-byte PRNG requests
3841 to traverse all of 'state'.
3843 1. When updating 'md_local' (the current thread's copy of 'md')
3844 during PRNG output generation, hash all of the previous
3845 'md_local' value, not just the half used for PRNG output.
3847 2. Make the number of bytes from 'state' included into the hash
3848 independent from the number of PRNG bytes requested.
3850 The first measure alone would be sufficient to avoid
3851 Markku-Juhani's attack. (Actually it had never occurred
3852 to me that the half of 'md_local' used for chaining was the
3853 half from which PRNG output bytes were taken -- I had always
3854 assumed that the secret half would be used.) The second
3855 measure makes sure that additional data from 'state' is never
3856 mixed into 'md_local' in small portions; this heuristically
3857 further strengthens the PRNG.
3860 *) Fix crypto/bn/asm/mips3.s.
3863 *) When only the key is given to "enc", the IV is undefined. Print out
3864 an error message in this case.
3867 *) Handle special case when X509_NAME is empty in X509 printing routines.
3870 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3871 positive and less than q.
3874 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3875 used: it isn't thread safe and the add_lock_callback should handle
3877 [Paul Rose <Paul.Rose@bridge.com>]
3879 *) Verify that incoming data obeys the block size in
3880 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3884 [Ulf Möller, Bodo Möller]
3886 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3887 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3888 when fixing the server behaviour for backwards-compatible 'client
3889 hello' messages. (Note that the attack is impractical against
3890 SSL 3.0 and TLS 1.0 anyway because length and version checking
3891 means that the probability of guessing a valid ciphertext is
3892 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3895 Before 0.9.5, the countermeasure (hide the error by generating a
3896 random 'decryption result') did not work properly because
3897 ERR_clear_error() was missing, meaning that SSL_get_error() would
3898 detect the supposedly ignored error.
3900 Both problems are now fixed.
3903 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3904 (previously it was 1024).
3907 *) Fix for compatibility mode trust settings: ignore trust settings
3908 unless some valid trust or reject settings are present.
3911 *) Fix for blowfish EVP: its a variable length cipher.
3914 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3915 parameters in DSA public key structures and return an error in the
3916 DSA routines if parameters are absent.
3919 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3920 in the current directory if neither $RANDFILE nor $HOME was set.
3921 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3922 caused some confusion to Windows users who haven't defined $HOME.
3923 Thus RAND_file_name() is changed again: e_os.h can define a
3924 DEFAULT_HOME, which will be used if $HOME is not set.
3925 For Windows, we use "C:"; on other platforms, we still require
3926 environment variables.
3928 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3929 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3930 having multiple threads call RAND_poll() concurrently.
3933 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3934 combination of a flag and a thread ID variable.
3935 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3936 flag), *other* threads can enter ssleay_add_bytes without obeying
3937 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3938 that they do not hold after the first thread unsets add_do_not_lock).
3941 *) Change bctest again: '-x' expressions are not available in all
3945 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3947 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3948 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3950 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3951 the default extension for executables, if any. Also, make the perl
3952 scripts that use symlink() to test if it really exists and use "cp"
3953 if it doesn't. All this made OpenSSL compilable and installable in
3957 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3958 If SEQUENCE is length is indefinite just set c->slen to the total
3959 amount of data available.
3960 [Steve Henson, reported by shige@FreeBSD.org]
3961 [This change does not apply to 0.9.7.]
3963 *) Change bctest to avoid here-documents inside command substitution
3964 (workaround for FreeBSD /bin/sh bug).
3965 For compatibility with Ultrix, avoid shell functions (introduced
3966 in the bctest version that searches along $PATH).
3969 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3970 with des_encrypt() defined on some operating systems, like Solaris
3974 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3975 On the Importance of Eliminating Errors in Cryptographic
3976 Computations, J. Cryptology 14 (2001) 2, 101-119,
3977 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3980 *) MIPS assembler BIGNUM division bug fix.
3983 *) Disabled incorrect Alpha assembler code.
3986 *) Fix PKCS#7 decode routines so they correctly update the length
3987 after reading an EOC for the EXPLICIT tag.
3989 [This change does not apply to 0.9.7.]
3991 *) Fix bug in PKCS#12 key generation routines. This was triggered
3992 if a 3DES key was generated with a 0 initial byte. Include
3993 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3994 (but broken) behaviour.
3997 *) Enhance bctest to search for a working bc along $PATH and print
3999 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4001 *) Fix memory leaks in err.c: free err_data string if necessary;
4002 don't write to the wrong index in ERR_set_error_data.
4005 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4009 *) Replace rdtsc with _emit statements for VC++ version 5.
4010 [Jeremy Cooper <jeremy@baymoo.org>]
4012 *) Make it possible to reuse SSLv2 sessions.
4015 *) In copy_email() check for >= 0 as a return value for
4016 X509_NAME_get_index_by_NID() since 0 is a valid index.
4017 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4019 *) Avoid coredump with unsupported or invalid public keys by checking if
4020 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4021 PKCS7_verify() fails with non detached data.
4024 *) Don't use getenv in library functions when run as setuid/setgid.
4025 New function OPENSSL_issetugid().
4028 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4029 due to incorrect handling of multi-threading:
4031 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4033 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4035 3. Count how many times MemCheck_off() has been called so that
4036 nested use can be treated correctly. This also avoids
4037 inband-signalling in the previous code (which relied on the
4038 assumption that thread ID 0 is impossible).
4041 *) Add "-rand" option also to s_client and s_server.
4044 *) Fix CPU detection on Irix 6.x.
4045 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4046 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4048 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4051 [This change does not apply to 0.9.7.]
4053 *) Use the cached encoding of an X509_NAME structure rather than
4054 copying it. This is apparently the reason for the libsafe "errors"
4055 but the code is actually correct.
4058 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4059 Bleichenbacher's DSA attack.
4060 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4061 to be set and top=0 forces the highest bit to be set; top=-1 is new
4062 and leaves the highest bit random.
4063 [Ulf Moeller, Bodo Moeller]
4065 *) In the NCONF_...-based implementations for CONF_... queries
4066 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4067 a temporary CONF structure with the data component set to NULL
4068 (which gives segmentation faults in lh_retrieve).
4069 Instead, use NULL for the CONF pointer in CONF_get_string and
4070 CONF_get_number (which may use environment variables) and directly
4071 return NULL from CONF_get_section.
4074 *) Fix potential buffer overrun for EBCDIC.
4077 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4078 keyUsage if basicConstraints absent for a CA.
4081 *) Make SMIME_write_PKCS7() write mail header values with a format that
4082 is more generally accepted (no spaces before the semicolon), since
4083 some programs can't parse those values properly otherwise. Also make
4084 sure BIO's that break lines after each write do not create invalid
4088 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4089 macros previously used would not encode an empty SEQUENCE OF
4090 and break the signature.
4092 [This change does not apply to 0.9.7.]
4094 *) Zero the premaster secret after deriving the master secret in
4098 *) Add some EVP_add_digest_alias registrations (as found in
4099 OpenSSL_add_all_digests()) to SSL_library_init()
4100 aka OpenSSL_add_ssl_algorithms(). This provides improved
4101 compatibility with peers using X.509 certificates
4102 with unconventional AlgorithmIdentifier OIDs.
4105 *) Fix for Irix with NO_ASM.
4106 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4108 *) ./config script fixes.
4109 [Ulf Moeller, Richard Levitte]
4111 *) Fix 'openssl passwd -1'.
4114 *) Change PKCS12_key_gen_asc() so it can cope with non null
4115 terminated strings whose length is passed in the passlen
4116 parameter, for example from PEM callbacks. This was done
4117 by adding an extra length parameter to asc2uni().
4118 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4120 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4121 call failed, free the DSA structure.
4124 *) Fix to uni2asc() to cope with zero length Unicode strings.
4125 These are present in some PKCS#12 files.
4128 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4129 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4130 when writing a 32767 byte record.
4131 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4133 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4134 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4136 (RSA objects have a reference count access to which is protected
4137 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4138 so they are meant to be shared between threads.)
4139 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4140 "Reddie, Steven" <Steven.Reddie@ca.com>]
4142 *) Fix a deadlock in CRYPTO_mem_leaks().
4145 *) Use better test patterns in bntest.
4148 *) rand_win.c fix for Borland C.
4151 *) BN_rshift bugfix for n == 0.
4154 *) Add a 'bctest' script that checks for some known 'bc' bugs
4155 so that 'make test' does not abort just because 'bc' is broken.
4158 *) Store verify_result within SSL_SESSION also for client side to
4159 avoid potential security hole. (Re-used sessions on the client side
4160 always resulted in verify_result==X509_V_OK, not using the original
4161 result of the server certificate verification.)
4164 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4165 SSL3_RT_APPLICATION_DATA, return 0.
4166 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4170 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4171 releases, have been re-implemented by renaming the previous
4172 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4173 and ssl3_read_internal, respectively, and adding 'peek' parameters
4174 to them. The new ssl[23]_{read,peek} functions are calls to
4175 ssl[23]_read_internal with the 'peek' flag set appropriately.
4176 A 'peek' parameter has also been added to ssl3_read_bytes, which
4177 does the actual work for ssl3_read_internal.
4180 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4181 the method-specific "init()" handler. Also clean up ex_data after
4182 calling the method-specific "finish()" handler. Previously, this was
4183 happening the other way round.
4186 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4187 The previous value, 12, was not always sufficient for BN_mod_exp().
4190 *) Make sure that shared libraries get the internal name engine with
4191 the full version number and not just 0. This should mark the
4192 shared libraries as not backward compatible. Of course, this should
4193 be changed again when we can guarantee backward binary compatibility.
4196 *) Fix typo in get_cert_by_subject() in by_dir.c
4197 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4199 *) Rework the system to generate shared libraries:
4201 - Make note of the expected extension for the shared libraries and
4202 if there is a need for symbolic links from for example libcrypto.so.0
4203 to libcrypto.so.0.9.7. There is extended info in Configure for
4206 - Make as few rebuilds of the shared libraries as possible.
4208 - Still avoid linking the OpenSSL programs with the shared libraries.
4210 - When installing, install the shared libraries separately from the
4214 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4216 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4217 and not in SSL_clear because the latter is also used by the
4218 accept/connect functions; previously, the settings made by
4219 SSL_set_read_ahead would be lost during the handshake.
4220 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4222 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4223 Previously, it would create entries for disableed algorithms no
4227 *) Added several new manual pages for SSL_* function.
4230 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4232 *) In ssl23_get_client_hello, generate an error message when faced
4233 with an initial SSL 3.0/TLS record that is too small to contain the
4234 first two bytes of the ClientHello message, i.e. client_version.
4235 (Note that this is a pathologic case that probably has never happened
4236 in real life.) The previous approach was to use the version number
4237 from the record header as a substitute; but our protocol choice
4238 should not depend on that one because it is not authenticated
4239 by the Finished messages.
4242 *) More robust randomness gathering functions for Windows.
4243 [Jeffrey Altman <jaltman@columbia.edu>]
4245 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4246 not set then we don't setup the error code for issuer check errors
4247 to avoid possibly overwriting other errors which the callback does
4248 handle. If an application does set the flag then we assume it knows
4249 what it is doing and can handle the new informational codes
4253 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4254 a general "ANY" type, as such it should be able to decode anything
4255 including tagged types. However it didn't check the class so it would
4256 wrongly interpret tagged types in the same way as their universal
4257 counterpart and unknown types were just rejected. Changed so that the
4258 tagged and unknown types are handled in the same way as a SEQUENCE:
4259 that is the encoding is stored intact. There is also a new type
4260 "V_ASN1_OTHER" which is used when the class is not universal, in this
4261 case we have no idea what the actual type is so we just lump them all
4265 *) On VMS, stdout may very well lead to a file that is written to
4266 in a record-oriented fashion. That means that every write() will
4267 write a separate record, which will be read separately by the
4268 programs trying to read from it. This can be very confusing.
4270 The solution is to put a BIO filter in the way that will buffer
4271 text until a linefeed is reached, and then write everything a
4272 line at a time, so every record written will be an actual line,
4273 not chunks of lines and not (usually doesn't happen, but I've
4274 seen it once) several lines in one record. BIO_f_linebuffer() is
4277 Currently, it's a VMS-only method, because that's where it has
4278 been tested well enough.
4281 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4282 it can return incorrect results.
4283 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4284 but it was in 0.9.6-beta[12].)
4287 *) Disable the check for content being present when verifying detached
4288 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4289 include zero length content when signing messages.
4292 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4293 BIO_ctrl (for BIO pairs).
4296 *) Add DSO method for VMS.
4299 *) Bug fix: Montgomery multiplication could produce results with the
4303 *) Add RPM specification openssl.spec and modify it to build three
4304 packages. The default package contains applications, application
4305 documentation and run-time libraries. The devel package contains
4306 include files, static libraries and function documentation. The
4307 doc package contains the contents of the doc directory. The original
4308 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4311 *) Add a large number of documentation files for many SSL routines.
4312 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4314 *) Add a configuration entry for Sony News 4.
4315 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4317 *) Don't set the two most significant bits to one when generating a
4318 random number < q in the DSA library.
4321 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4322 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4323 the underlying transport is blocking) if a handshake took place.
4324 (The default behaviour is needed by applications such as s_client
4325 and s_server that use select() to determine when to use SSL_read;
4326 but for applications that know in advance when to expect data, it
4327 just makes things more complicated.)
4330 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4334 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4335 work better on such systems.
4336 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4338 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4339 Update PKCS12_parse() so it copies the friendlyName and the
4340 keyid to the certificates aux info.
4343 *) Fix bug in PKCS7_verify() which caused an infinite loop
4344 if there was more than one signature.
4345 [Sven Uszpelkat <su@celocom.de>]
4347 *) Major change in util/mkdef.pl to include extra information
4348 about each symbol, as well as presentig variables as well
4349 as functions. This change means that there's n more need
4350 to rebuild the .num files when some algorithms are excluded.
4353 *) Allow the verify time to be set by an application,
4354 rather than always using the current time.
4357 *) Phase 2 verify code reorganisation. The certificate
4358 verify code now looks up an issuer certificate by a
4359 number of criteria: subject name, authority key id
4360 and key usage. It also verifies self signed certificates
4361 by the same criteria. The main comparison function is
4362 X509_check_issued() which performs these checks.
4364 Lot of changes were necessary in order to support this
4365 without completely rewriting the lookup code.
4367 Authority and subject key identifier are now cached.
4369 The LHASH 'certs' is X509_STORE has now been replaced
4370 by a STACK_OF(X509_OBJECT). This is mainly because an
4371 LHASH can't store or retrieve multiple objects with
4372 the same hash value.
4374 As a result various functions (which were all internal
4375 use only) have changed to handle the new X509_STORE
4376 structure. This will break anything that messed round
4377 with X509_STORE internally.
4379 The functions X509_STORE_add_cert() now checks for an
4380 exact match, rather than just subject name.
4382 The X509_STORE API doesn't directly support the retrieval
4383 of multiple certificates matching a given criteria, however
4384 this can be worked round by performing a lookup first
4385 (which will fill the cache with candidate certificates)
4386 and then examining the cache for matches. This is probably
4387 the best we can do without throwing out X509_LOOKUP
4388 entirely (maybe later...).
4390 The X509_VERIFY_CTX structure has been enhanced considerably.
4392 All certificate lookup operations now go via a get_issuer()
4393 callback. Although this currently uses an X509_STORE it
4394 can be replaced by custom lookups. This is a simple way
4395 to bypass the X509_STORE hackery necessary to make this
4396 work and makes it possible to use more efficient techniques
4397 in future. A very simple version which uses a simple
4398 STACK for its trusted certificate store is also provided
4399 using X509_STORE_CTX_trusted_stack().
4401 The verify_cb() and verify() callbacks now have equivalents
4402 in the X509_STORE_CTX structure.
4404 X509_STORE_CTX also has a 'flags' field which can be used
4405 to customise the verify behaviour.
4408 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4409 excludes S/MIME capabilities.
4412 *) When a certificate request is read in keep a copy of the
4413 original encoding of the signed data and use it when outputing
4414 again. Signatures then use the original encoding rather than
4415 a decoded, encoded version which may cause problems if the
4416 request is improperly encoded.
4419 *) For consistency with other BIO_puts implementations, call
4420 buffer_write(b, ...) directly in buffer_puts instead of calling
4423 In BIO_puts, increment b->num_write as in BIO_write.
4424 [Peter.Sylvester@EdelWeb.fr]
4426 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4427 BN_zero, we may not return a BIGNUM with an array consisting of
4431 *) Avoid calling abort() from within the library when problems are
4432 detected, except if preprocessor symbols have been defined
4433 (such as REF_CHECK, BN_DEBUG etc.).
4436 *) New openssl application 'rsautl'. This utility can be
4437 used for low level RSA operations. DER public key
4438 BIO/fp routines also added.
4441 *) New Configure entry and patches for compiling on QNX 4.
4442 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4444 *) A demo state-machine implementation was sponsored by
4445 Nuron (http://www.nuron.com/) and is now available in
4446 demos/state_machine.
4449 *) New options added to the 'dgst' utility for signature
4450 generation and verification.
4453 *) Unrecognized PKCS#7 content types are now handled via a
4454 catch all ASN1_TYPE structure. This allows unsupported
4455 types to be stored as a "blob" and an application can
4456 encode and decode it manually.
4459 *) Fix various signed/unsigned issues to make a_strex.c
4461 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4463 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4464 length if passed a buffer. ASN1_INTEGER_to_BN failed
4465 if passed a NULL BN and its argument was negative.
4466 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4468 *) Modification to PKCS#7 encoding routines to output definite
4469 length encoding. Since currently the whole structures are in
4470 memory there's not real point in using indefinite length
4471 constructed encoding. However if OpenSSL is compiled with
4472 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4475 *) Added BIO_vprintf() and BIO_vsnprintf().
4478 *) Added more prefixes to parse for in the the strings written
4479 through a logging bio, to cover all the levels that are available
4480 through syslog. The prefixes are now:
4482 PANIC, EMERG, EMR => LOG_EMERG
4483 ALERT, ALR => LOG_ALERT
4484 CRIT, CRI => LOG_CRIT
4485 ERROR, ERR => LOG_ERR
4486 WARNING, WARN, WAR => LOG_WARNING
4487 NOTICE, NOTE, NOT => LOG_NOTICE
4488 INFO, INF => LOG_INFO
4489 DEBUG, DBG => LOG_DEBUG
4491 and as before, if none of those prefixes are present at the
4492 beginning of the string, LOG_ERR is chosen.
4494 On Win32, the LOG_* levels are mapped according to this:
4496 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4497 LOG_WARNING => EVENTLOG_WARNING_TYPE
4498 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4502 *) Made it possible to reconfigure with just the configuration
4503 argument "reconf" or "reconfigure". The command line arguments
4504 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4505 and are retrieved from there when reconfiguring.
4509 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4511 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4514 *) The obj_dat.pl script was messing up the sorting of object
4515 names. The reason was that it compared the quoted version
4516 of strings as a result "OCSP" > "OCSP Signing" because
4517 " > SPACE. Changed script to store unquoted versions of
4518 names and add quotes on output. It was also omitting some
4519 names from the lookup table if they were given a default
4520 value (that is if SN is missing it is given the same
4521 value as LN and vice versa), these are now added on the
4522 grounds that if an object has a name we should be able to
4523 look it up. Finally added warning output when duplicate
4524 short or long names are found.
4527 *) Changes needed for Tandem NSK.
4528 [Scott Uroff <scott@xypro.com>]
4530 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4531 RSA_padding_check_SSLv23(), special padding was never detected
4532 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4533 version rollback attacks was not effective.
4535 In s23_clnt.c, don't use special rollback-attack detection padding
4536 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4537 client; similarly, in s23_srvr.c, don't do the rollback check if
4538 SSL 2.0 is the only protocol enabled in the server.
4541 *) Make it possible to get hexdumps of unprintable data with 'openssl
4542 asn1parse'. By implication, the functions ASN1_parse_dump() and
4543 BIO_dump_indent() are added.
4546 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4547 these print out strings and name structures based on various
4548 flags including RFC2253 support and proper handling of
4549 multibyte characters. Added options to the 'x509' utility
4550 to allow the various flags to be set.
4553 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4554 Also change the functions X509_cmp_current_time() and
4555 X509_gmtime_adj() work with an ASN1_TIME structure,
4556 this will enable certificates using GeneralizedTime in validity
4557 dates to be checked.
4560 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4561 negative public key encodings) on by default,
4562 NO_NEG_PUBKEY_BUG can be set to disable it.
4565 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4566 content octets. An i2c_ASN1_OBJECT is unnecessary because
4567 the encoding can be trivially obtained from the structure.
4570 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4571 not read locks (CRYPTO_r_[un]lock).
4574 *) A first attempt at creating official support for shared
4575 libraries through configuration. I've kept it so the
4576 default is static libraries only, and the OpenSSL programs
4577 are always statically linked for now, but there are
4578 preparations for dynamic linking in place.
4579 This has been tested on Linux and Tru64.
4582 *) Randomness polling function for Win9x, as described in:
4583 Peter Gutmann, Software Generation of Practically Strong
4587 *) Fix so PRNG is seeded in req if using an already existing
4591 *) New options to smime application. -inform and -outform
4592 allow alternative formats for the S/MIME message including
4593 PEM and DER. The -content option allows the content to be
4594 specified separately. This should allow things like Netscape
4595 form signing output easier to verify.
4598 *) Fix the ASN1 encoding of tags using the 'long form'.
4601 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4602 STRING types. These convert content octets to and from the
4603 underlying type. The actual tag and length octets are
4604 already assumed to have been read in and checked. These
4605 are needed because all other string types have virtually
4606 identical handling apart from the tag. By having versions
4607 of the ASN1 functions that just operate on content octets
4608 IMPLICIT tagging can be handled properly. It also allows
4609 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4610 and ASN1_INTEGER are identical apart from the tag.
4613 *) Change the handling of OID objects as follows:
4615 - New object identifiers are inserted in objects.txt, following
4616 the syntax given in objects.README.
4617 - objects.pl is used to process obj_mac.num and create a new
4619 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4622 This is currently kind of a hack, and the perl code in objects.pl
4623 isn't very elegant, but it works as I intended. The simplest way
4624 to check that it worked correctly is to look in obj_dat.h and
4625 check the array nid_objs and make sure the objects haven't moved
4626 around (this is important!). Additions are OK, as well as
4627 consistent name changes.
4630 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4633 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4634 The given file adds to whatever has already been seeded into the
4635 random pool through the RANDFILE configuration file option or
4636 environment variable, or the default random state file.
4639 *) mkstack.pl now sorts each macro group into lexical order.
4640 Previously the output order depended on the order the files
4641 appeared in the directory, resulting in needless rewriting
4645 *) Patches to make OpenSSL compile under Win32 again. Mostly
4646 work arounds for the VC++ problem that it treats func() as
4647 func(void). Also stripped out the parts of mkdef.pl that
4648 added extra typesafe functions: these no longer exist.
4651 *) Reorganisation of the stack code. The macros are now all
4652 collected in safestack.h . Each macro is defined in terms of
4653 a "stack macro" of the form SKM_<name>(type, a, b). The
4654 DEBUG_SAFESTACK is now handled in terms of function casts,
4655 this has the advantage of retaining type safety without the
4656 use of additional functions. If DEBUG_SAFESTACK is not defined
4657 then the non typesafe macros are used instead. Also modified the
4658 mkstack.pl script to handle the new form. Needs testing to see
4659 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4660 the default if no major problems. Similar behaviour for ASN1_SET_OF
4661 and PKCS12_STACK_OF.
4664 *) When some versions of IIS use the 'NET' form of private key the
4665 key derivation algorithm is different. Normally MD5(password) is
4666 used as a 128 bit RC4 key. In the modified case
4667 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4668 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4669 as the old Netscape_RSA functions except they have an additional
4670 'sgckey' parameter which uses the modified algorithm. Also added
4671 an -sgckey command line option to the rsa utility. Thanks to
4672 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4673 algorithm to openssl-dev.
4676 *) The evp_local.h macros were using 'c.##kname' which resulted in
4677 invalid expansion on some systems (SCO 5.0.5 for example).
4678 Corrected to 'c.kname'.
4679 [Phillip Porch <root@theporch.com>]
4681 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4682 a STACK of email addresses from a certificate or request, these look
4683 in the subject name and the subject alternative name extensions and
4684 omit any duplicate addresses.
4687 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4688 This makes DSA verification about 2 % faster.
4691 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4692 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4693 plus overhead for 1024 bit moduli).
4694 This makes exponentiations about 0.5 % faster for 1024 bit
4695 exponents (as measured by "openssl speed rsa2048").
4698 *) Rename memory handling macros to avoid conflicts with other
4700 Malloc => OPENSSL_malloc
4701 Malloc_locked => OPENSSL_malloc_locked
4702 Realloc => OPENSSL_realloc
4703 Free => OPENSSL_free
4706 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4707 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4710 *) CygWin32 support.
4711 [John Jarvie <jjarvie@newsguy.com>]
4713 *) The type-safe stack code has been rejigged. It is now only compiled
4714 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4715 by default all type-specific stack functions are "#define"d back to
4716 standard stack functions. This results in more streamlined output
4717 but retains the type-safety checking possibilities of the original
4721 *) The STACK code has been cleaned up, and certain type declarations
4722 that didn't make a lot of sense have been brought in line. This has
4723 also involved a cleanup of sorts in safestack.h to more correctly
4724 map type-safe stack functions onto their plain stack counterparts.
4725 This work has also resulted in a variety of "const"ifications of
4726 lots of the code, especially "_cmp" operations which should normally
4727 be prototyped with "const" parameters anyway.
4730 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4731 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4732 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4733 where all of 'md' is used each time the PRNG is used, but 'state'
4734 is used only indexed by a cyclic counter. As entropy may not be
4735 well distributed from the beginning, 'md' is important as a
4736 chaining variable. However, the output function chains only half
4737 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4738 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4739 in all of 'state' being rewritten, with the new values depending
4740 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4743 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4744 the handshake is continued after ssl_verify_cert_chain();
4745 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4746 can lead to 'unexplainable' connection aborts later.
4747 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4749 *) Major EVP API cipher revision.
4750 Add hooks for extra EVP features. This allows various cipher
4751 parameters to be set in the EVP interface. Support added for variable
4752 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4753 setting of RC2 and RC5 parameters.
4755 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4758 Remove lots of duplicated code from the EVP library. For example *every*
4759 cipher init() function handles the 'iv' in the same way according to the
4760 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4761 for CFB and OFB modes they zero ctx->num.
4763 New functionality allows removal of S/MIME code RC2 hack.
4765 Most of the routines have the same form and so can be declared in terms
4768 By shifting this to the top level EVP_CipherInit() it can be removed from
4769 all individual ciphers. If the cipher wants to handle IVs or keys
4770 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4773 Change lots of functions like EVP_EncryptUpdate() to now return a
4774 value: although software versions of the algorithms cannot fail
4775 any installed hardware versions can.
4778 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4779 this option is set, tolerate broken clients that send the negotiated
4780 protocol version number instead of the requested protocol version
4784 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4785 i.e. non-zero for export ciphersuites, zero otherwise.
4786 Previous versions had this flag inverted, inconsistent with
4787 rsa_tmp_cb (..._TMP_RSA_CB).
4788 [Bodo Moeller; problem reported by Amit Chopra]
4790 *) Add missing DSA library text string. Work around for some IIS
4791 key files with invalid SEQUENCE encoding.
4794 *) Add a document (doc/standards.txt) that list all kinds of standards
4795 and so on that are implemented in OpenSSL.
4798 *) Enhance c_rehash script. Old version would mishandle certificates
4799 with the same subject name hash and wouldn't handle CRLs at all.
4800 Added -fingerprint option to crl utility, to support new c_rehash
4804 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4807 *) Fix for SSL server purpose checking. Server checking was
4808 rejecting certificates which had extended key usage present
4809 but no ssl client purpose.
4810 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4812 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4813 is a little unclear about how a blank password is handled.
4814 Since the password in encoded as a BMPString with terminating
4815 double NULL a zero length password would end up as just the
4816 double NULL. However no password at all is different and is
4817 handled differently in the PKCS#12 key generation code. NS
4818 treats a blank password as zero length. MSIE treats it as no
4819 password on export: but it will try both on import. We now do
4820 the same: PKCS12_parse() tries zero length and no password if
4821 the password is set to "" or NULL (NULL is now a valid password:
4822 it wasn't before) as does the pkcs12 application.
4825 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4826 perror when PEM_read_bio_X509_REQ fails, the error message must
4827 be obtained from the error queue.
4830 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4831 it in ERR_remove_state if appropriate, and change ERR_get_state
4832 accordingly to avoid race conditions (this is necessary because
4833 thread_hash is no longer constant once set).
4836 *) Bugfix for linux-elf makefile.one.
4839 *) RSA_get_default_method() will now cause a default
4840 RSA_METHOD to be chosen if one doesn't exist already.
4841 Previously this was only set during a call to RSA_new()
4842 or RSA_new_method(NULL) meaning it was possible for
4843 RSA_get_default_method() to return NULL.
4846 *) Added native name translation to the existing DSO code
4847 that will convert (if the flag to do so is set) filenames
4848 that are sufficiently small and have no path information
4849 into a canonical native form. Eg. "blah" converted to
4850 "libblah.so" or "blah.dll" etc.
4853 *) New function ERR_error_string_n(e, buf, len) which is like
4854 ERR_error_string(e, buf), but writes at most 'len' bytes
4855 including the 0 terminator. For ERR_error_string_n, 'buf'
4857 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4859 *) CONF library reworked to become more general. A new CONF
4860 configuration file reader "class" is implemented as well as a
4861 new functions (NCONF_*, for "New CONF") to handle it. The now
4862 old CONF_* functions are still there, but are reimplemented to
4863 work in terms of the new functions. Also, a set of functions
4864 to handle the internal storage of the configuration data is
4865 provided to make it easier to write new configuration file
4866 reader "classes" (I can definitely see something reading a
4867 configuration file in XML format, for example), called _CONF_*,
4868 or "the configuration storage API"...
4870 The new configuration file reading functions are:
4872 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4873 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4875 NCONF_default, NCONF_WIN32
4877 NCONF_dump_fp, NCONF_dump_bio
4879 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4880 NCONF_new creates a new CONF object. This works in the same way
4881 as other interfaces in OpenSSL, like the BIO interface.
4882 NCONF_dump_* dump the internal storage of the configuration file,
4883 which is useful for debugging. All other functions take the same
4884 arguments as the old CONF_* functions wth the exception of the
4885 first that must be a `CONF *' instead of a `LHASH *'.
4887 To make it easer to use the new classes with the old CONF_* functions,
4888 the function CONF_set_default_method is provided.
4891 *) Add '-tls1' option to 'openssl ciphers', which was already
4892 mentioned in the documentation but had not been implemented.
4893 (This option is not yet really useful because even the additional
4894 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4897 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4898 OpenSSL-based applications) load shared libraries and bind to
4899 them in a portable way.
4900 [Geoff Thorpe, with contributions from Richard Levitte]
4902 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4904 *) Make sure _lrotl and _lrotr are only used with MSVC.
4906 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4907 (the default implementation of RAND_status).
4909 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4910 to '-clrext' (= clear extensions), as intended and documented.
4911 [Bodo Moeller; inconsistency pointed out by Michael Attili
4914 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4915 was larger than the MD block size.
4916 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4918 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4919 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4920 using the passed key: if the passed key was a private key the result
4921 of X509_print(), for example, would be to print out all the private key
4925 *) des_quad_cksum() byte order bug fix.
4926 [Ulf Möller, using the problem description in krb4-0.9.7, where
4927 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4929 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4931 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4933 *) For easily testing in shell scripts whether some command
4934 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4935 returns with exit code 0 iff no command of the given name is available.
4936 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4937 the output goes to stdout and nothing is printed to stderr.
4938 Additional arguments are always ignored.
4940 Since for each cipher there is a command of the same name,
4941 the 'no-cipher' compilation switches can be tested this way.
4943 ('openssl no-XXX' is not able to detect pseudo-commands such
4944 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4947 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4950 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4951 is set; it will be thrown away anyway because each handshake creates
4953 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4954 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4955 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4956 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4959 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4960 'Q' and 'R' lose their special meanings (quit/renegotiate).
4961 This is part of what -quiet does; unlike -quiet, -ign_eof
4962 does not suppress any output.
4965 *) Add compatibility options to the purpose and trust code. The
4966 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4967 accepts a certificate or CA, this was the previous behaviour,
4968 with all the associated security issues.
4970 X509_TRUST_COMPAT is the old trust behaviour: only and
4971 automatically trust self signed roots in certificate store. A
4972 new trust setting X509_TRUST_DEFAULT is used to specify that
4973 a purpose has no associated trust setting and it should instead
4974 use the value in the default purpose.
4977 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4978 and fix a memory leak.
4981 *) In util/mkerr.pl (which implements 'make errors'), preserve
4982 reason strings from the previous version of the .c file, as
4983 the default to have only downcase letters (and digits) in
4984 automatically generated reasons codes is not always appropriate.
4987 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4988 using strerror. Previously, ERR_reason_error_string() returned
4989 library names as reason strings for SYSerr; but SYSerr is a special
4990 case where small numbers are errno values, not library numbers.
4993 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4994 converts DSA parameters into DH parameters. (When creating parameters,
4995 DSA_generate_parameters is used.)
4998 *) Include 'length' (recommended exponent length) in C code generated
4999 by 'openssl dhparam -C'.
5002 *) The second argument to set_label in perlasm was already being used
5003 so couldn't be used as a "file scope" flag. Moved to third argument
5007 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5008 instead of RAND_bytes for encryption IVs and salts.
5011 *) Include RAND_status() into RAND_METHOD instead of implementing
5012 it only for md_rand.c Otherwise replacing the PRNG by calling
5013 RAND_set_rand_method would be impossible.
5016 *) Don't let DSA_generate_key() enter an infinite loop if the random
5017 number generation fails.
5020 *) New 'rand' application for creating pseudo-random output.
5023 *) Added configuration support for Linux/IA64
5024 [Rolf Haberrecker <rolf@suse.de>]
5026 *) Assembler module support for Mingw32.
5029 *) Shared library support for HPUX (in shlib/).
5030 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5032 *) Shared library support for Solaris gcc.
5033 [Lutz Behnke <behnke@trustcenter.de>]
5035 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5037 *) PKCS7_encrypt() was adding text MIME headers twice because they
5038 were added manually and by SMIME_crlf_copy().
5041 *) In bntest.c don't call BN_rand with zero bits argument.
5042 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5044 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5045 case was implemented. This caused BN_div_recp() to fail occasionally.
5048 *) Add an optional second argument to the set_label() in the perl
5049 assembly language builder. If this argument exists and is set
5050 to 1 it signals that the assembler should use a symbol whose
5051 scope is the entire file, not just the current function. This
5052 is needed with MASM which uses the format label:: for this scope.
5053 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5055 *) Change the ASN1 types so they are typedefs by default. Before
5056 almost all types were #define'd to ASN1_STRING which was causing
5057 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5061 *) Change names of new functions to the new get1/get0 naming
5062 convention: After 'get1', the caller owns a reference count
5063 and has to call ..._free; 'get0' returns a pointer to some
5064 data structure without incrementing reference counters.
5065 (Some of the existing 'get' functions increment a reference
5066 counter, some don't.)
5067 Similarly, 'set1' and 'add1' functions increase reference
5068 counters or duplicate objects.
5071 *) Allow for the possibility of temp RSA key generation failure:
5072 the code used to assume it always worked and crashed on failure.
5075 *) Fix potential buffer overrun problem in BIO_printf().
5076 [Ulf Möller, using public domain code by Patrick Powell; problem
5077 pointed out by David Sacerdote <das33@cornell.edu>]
5079 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5080 RAND_egd() and RAND_status(). In the command line application,
5081 the EGD socket can be specified like a seed file using RANDFILE
5085 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5086 Some CAs (e.g. Verisign) distribute certificates in this form.
5089 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5090 list to exclude them. This means that no special compilation option
5091 is needed to use anonymous DH: it just needs to be included in the
5095 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5096 EVP_MD_type. The old functionality is available in a new macro called
5097 EVP_MD_md(). Change code that uses it and update docs.
5100 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5101 where the 'void *' argument is replaced by a function pointer argument.
5102 Previously 'void *' was abused to point to functions, which works on
5103 many platforms, but is not correct. As these functions are usually
5104 called by macros defined in OpenSSL header files, most source code
5105 should work without changes.
5108 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5109 sections with information on -D... compiler switches used for
5110 compiling the library so that applications can see them. To enable
5111 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5112 must be defined. E.g.,
5113 #define OPENSSL_ALGORITHM_DEFINES
5114 #include <openssl/opensslconf.h>
5115 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5116 [Richard Levitte, Ulf and Bodo Möller]
5118 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5122 *) Change the 'other' type in certificate aux info to a STACK_OF
5123 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5124 the required ASN1 format: arbitrary types determined by an OID.
5127 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5128 argument to 'req'. This is not because the function is newer or
5129 better than others it just uses the work 'NEW' in the certificate
5130 request header lines. Some software needs this.
5133 *) Reorganise password command line arguments: now passwords can be
5134 obtained from various sources. Delete the PEM_cb function and make
5135 it the default behaviour: i.e. if the callback is NULL and the
5136 usrdata argument is not NULL interpret it as a null terminated pass
5137 phrase. If usrdata and the callback are NULL then the pass phrase
5138 is prompted for as usual.
5141 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5142 the support is automatically enabled. The resulting binaries will
5143 autodetect the card and use it if present.
5144 [Ben Laurie and Compaq Inc.]
5146 *) Work around for Netscape hang bug. This sends certificate request
5147 and server done in one record. Since this is perfectly legal in the
5148 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5149 the bugs/SSLv3 entry for more info.
5152 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5155 *) Add -rand argument to smime and pkcs12 applications and read/write
5159 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5162 *) Add command line password options to the remaining applications.
5165 *) Bug fix for BN_div_recp() for numerators with an even number of
5169 *) More tests in bntest.c, and changed test_bn output.
5172 *) ./config recognizes MacOS X now.
5175 *) Bug fix for BN_div() when the first words of num and divsor are
5176 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5179 *) Add support for various broken PKCS#8 formats, and command line
5180 options to produce them.
5183 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5184 get temporary BIGNUMs from a BN_CTX.
5187 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5191 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5192 include a #define from the old name to the new. The original intent
5193 was that statically linked binaries could for example just call
5194 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5195 link with digests. This never worked becayse SSLeay_add_all_digests()
5196 and SSLeay_add_all_ciphers() were in the same source file so calling
5197 one would link with the other. They are now in separate source files.
5200 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5203 *) Use a less unusual form of the Miller-Rabin primality test (it used
5204 a binary algorithm for exponentiation integrated into the Miller-Rabin
5205 loop, our standard modexp algorithms are faster).
5208 *) Support for the EBCDIC character set completed.
5209 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5211 *) Source code cleanups: use const where appropriate, eliminate casts,
5212 use void * instead of char * in lhash.
5215 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5216 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5217 this the server could overwrite ephemeral keys that the client
5221 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5222 using 50 iterations of the Rabin-Miller test.
5224 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5225 iterations of the Rabin-Miller test as required by the appendix
5226 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5227 As BN_is_prime_fasttest includes trial division, DSA parameter
5228 generation becomes much faster.
5230 This implies a change for the callback functions in DSA_is_prime
5231 and DSA_generate_parameters: The callback function is called once
5232 for each positive witness in the Rabin-Miller test, not just
5233 occasionally in the inner loop; and the parameters to the
5234 callback function now provide an iteration count for the outer
5235 loop rather than for the current invocation of the inner loop.
5236 DSA_generate_parameters additionally can call the callback
5237 function with an 'iteration count' of -1, meaning that a
5238 candidate has passed the trial division test (when q is generated
5239 from an application-provided seed, trial division is skipped).
5242 *) New function BN_is_prime_fasttest that optionally does trial
5243 division before starting the Rabin-Miller test and has
5244 an additional BN_CTX * argument (whereas BN_is_prime always
5245 has to allocate at least one BN_CTX).
5246 'callback(1, -1, cb_arg)' is called when a number has passed the
5247 trial division stage.
5250 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5254 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5257 *) New function BN_pseudo_rand().
5260 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5261 bignum version of BN_from_montgomery() with the working code from
5262 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5266 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5267 made it impossible to use the same SSL_SESSION data structure in
5268 SSL2 clients in multiple threads.
5271 *) The return value of RAND_load_file() no longer counts bytes obtained
5272 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5273 to seed the PRNG (previously an explicit byte count was required).
5274 [Ulf Möller, Bodo Möller]
5276 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5277 used (char *) instead of (void *) and had casts all over the place.
5280 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5283 *) Retain source code compatibility for BN_prime_checks macro:
5284 BN_is_prime(..., BN_prime_checks, ...) now uses
5285 BN_prime_checks_for_size to determine the appropriate number of
5286 Rabin-Miller iterations.
5289 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5290 DH_CHECK_P_NOT_SAFE_PRIME.
5291 (Check if this is true? OpenPGP calls them "strong".)
5294 *) Merge the functionality of "dh" and "gendh" programs into a new program
5295 "dhparam". The old programs are retained for now but will handle DH keys
5296 (instead of parameters) in future.
5299 *) Make the ciphers, s_server and s_client programs check the return values
5300 when a new cipher list is set.
5303 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5304 ciphers. Before when the 56bit ciphers were enabled the sorting was
5307 The syntax for the cipher sorting has been extended to support sorting by
5308 cipher-strength (using the strength_bits hard coded in the tables).
5309 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5311 Fix a bug in the cipher-command parser: when supplying a cipher command
5312 string with an "undefined" symbol (neither command nor alphanumeric
5313 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5314 an error is flagged.
5316 Due to the strength-sorting extension, the code of the
5317 ssl_create_cipher_list() function was completely rearranged. I hope that
5318 the readability was also increased :-)
5319 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5321 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5322 for the first serial number and places 2 in the serial number file. This
5323 avoids problems when the root CA is created with serial number zero and
5324 the first user certificate has the same issuer name and serial number
5328 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5329 the new code. Add documentation for this stuff.
5332 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5333 X509_*() to X509at_*() on the grounds that they don't handle X509
5334 structures and behave in an analagous way to the X509v3 functions:
5335 they shouldn't be called directly but wrapper functions should be used
5338 So we also now have some wrapper functions that call the X509at functions
5339 when passed certificate requests. (TO DO: similar things can be done with
5340 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5341 things. Some of these need some d2i or i2d and print functionality
5342 because they handle more complex structures.)
5345 *) Add missing #ifndefs that caused missing symbols when building libssl
5346 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5347 NO_RSA in ssl/s2*.c.
5348 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5350 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5351 has a return value which indicates the quality of the random data
5352 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5353 error queue. New function RAND_pseudo_bytes() generates output that is
5354 guaranteed to be unique but not unpredictable. RAND_add is like
5355 RAND_seed, but takes an extra argument for an entropy estimate
5356 (RAND_seed always assumes full entropy).
5359 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5360 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5361 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5362 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5363 false-positive rate of at most 2^-80 for random input.
5366 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5369 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5370 in the 0.9.5 release), this returns the chain
5371 from an X509_CTX structure with a dup of the stack and all
5372 the X509 reference counts upped: so the stack will exist
5373 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5376 Also make SSL_SESSION_print() print out the verify return
5380 *) Add manpage for the pkcs12 command. Also change the default
5381 behaviour so MAC iteration counts are used unless the new
5382 -nomaciter option is used. This improves file security and
5383 only older versions of MSIE (4.0 for example) need it.
5386 *) Honor the no-xxx Configure options when creating .DEF files.
5389 *) Add PKCS#10 attributes to field table: challengePassword,
5390 unstructuredName and unstructuredAddress. These are taken from
5391 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5392 international characters are used.
5394 More changes to X509_ATTRIBUTE code: allow the setting of types
5395 based on strings. Remove the 'loc' parameter when adding
5396 attributes because these will be a SET OF encoding which is sorted
5400 *) Initial changes to the 'req' utility to allow request generation
5401 automation. This will allow an application to just generate a template
5402 file containing all the field values and have req construct the
5405 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5406 used all over the place including certificate requests and PKCS#7
5407 structures. They are currently handled manually where necessary with
5408 some primitive wrappers for PKCS#7. The new functions behave in a
5409 manner analogous to the X509 extension functions: they allow
5410 attributes to be looked up by NID and added.
5412 Later something similar to the X509V3 code would be desirable to
5413 automatically handle the encoding, decoding and printing of the
5414 more complex types. The string types like challengePassword can
5415 be handled by the string table functions.
5417 Also modified the multi byte string table handling. Now there is
5418 a 'global mask' which masks out certain types. The table itself
5419 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5420 is useful when for example there is only one permissible type
5421 (as in countryName) and using the mask might result in no valid
5425 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5426 SSL_get_peer_finished to allow applications to obtain the latest
5427 Finished messages sent to the peer or expected from the peer,
5428 respectively. (SSL_get_peer_finished is usually the Finished message
5429 actually received from the peer, otherwise the protocol will be aborted.)
5431 As the Finished message are message digests of the complete handshake
5432 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5433 be used for external authentication procedures when the authentication
5434 provided by SSL/TLS is not desired or is not enough.
5437 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5438 the host supports BWX extension and if Compaq C is present on the
5439 $PATH. Just exploiting of the BWX extension results in 20-30%
5440 performance kick for some algorithms, e.g. DES and RC4 to mention
5441 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5445 *) Add support for MS "fast SGC". This is arguably a violation of the
5446 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5447 weak crypto and after checking the certificate is SGC a second one
5448 with strong crypto. MS SGC stops the first handshake after receiving
5449 the server certificate message and sends a second client hello. Since
5450 a server will typically do all the time consuming operations before
5451 expecting any further messages from the client (server key exchange
5452 is the most expensive) there is little difference between the two.
5454 To get OpenSSL to support MS SGC we have to permit a second client
5455 hello message after we have sent server done. In addition we have to
5456 reset the MAC if we do get this second client hello.
5459 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5460 if a DER encoded private key is RSA or DSA traditional format. Changed
5461 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5462 format DER encoded private key. Newer code should use PKCS#8 format which
5463 has the key type encoded in the ASN1 structure. Added DER private key
5464 support to pkcs8 application.
5467 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5468 ciphersuites has been selected (as required by the SSL 3/TLS 1
5469 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5470 is set, we interpret this as a request to violate the specification
5471 (the worst that can happen is a handshake failure, and 'correct'
5472 behaviour would result in a handshake failure anyway).
5475 *) In SSL_CTX_add_session, take into account that there might be multiple
5476 SSL_SESSION structures with the same session ID (e.g. when two threads
5477 concurrently obtain them from an external cache).
5478 The internal cache can handle only one SSL_SESSION with a given ID,
5479 so if there's a conflict, we now throw out the old one to achieve
5483 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5484 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5485 some routines that use cipher OIDs: some ciphers do not have OIDs
5486 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5490 *) Simplify the trust setting structure and code. Now we just have
5491 two sequences of OIDs for trusted and rejected settings. These will
5492 typically have values the same as the extended key usage extension
5493 and any application specific purposes.
5495 The trust checking code now has a default behaviour: it will just
5496 check for an object with the same NID as the passed id. Functions can
5497 be provided to override either the default behaviour or the behaviour
5498 for a given id. SSL client, server and email already have functions
5499 in place for compatibility: they check the NID and also return "trusted"
5500 if the certificate is self signed.
5503 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5504 traditional format into an EVP_PKEY structure.
5507 *) Add a password callback function PEM_cb() which either prompts for
5508 a password if usr_data is NULL or otherwise assumes it is a null
5509 terminated password. Allow passwords to be passed on command line
5510 environment or config files in a few more utilities.
5513 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5514 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5515 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5516 Update documentation.
5519 *) Support for ASN1 "NULL" type. This could be handled before by using
5520 ASN1_TYPE but there wasn't any function that would try to read a NULL
5521 and produce an error if it couldn't. For compatibility we also have
5522 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5523 don't allocate anything because they don't need to.
5526 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5528 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5530 *) Rebuild of the memory allocation routines used by OpenSSL code and
5531 possibly others as well. The purpose is to make an interface that
5532 provide hooks so anyone can build a separate set of allocation and
5533 deallocation routines to be used by OpenSSL, for example memory
5534 pool implementations, or something else, which was previously hard
5535 since Malloc(), Realloc() and Free() were defined as macros having
5536 the values malloc, realloc and free, respectively (except for Win32
5537 compilations). The same is provided for memory debugging code.
5538 OpenSSL already comes with functionality to find memory leaks, but
5539 this gives people a chance to debug other memory problems.
5541 With these changes, a new set of functions and macros have appeared:
5543 CRYPTO_set_mem_debug_functions() [F]
5544 CRYPTO_get_mem_debug_functions() [F]
5545 CRYPTO_dbg_set_options() [F]
5546 CRYPTO_dbg_get_options() [F]
5547 CRYPTO_malloc_debug_init() [M]
5549 The memory debug functions are NULL by default, unless the library
5550 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5551 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5552 gives the standard debugging functions that come with OpenSSL) or
5553 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5554 provided by the library user) must be used. When the standard
5555 debugging functions are used, CRYPTO_dbg_set_options can be used to
5556 request additional information:
5557 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5558 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5560 Also, things like CRYPTO_set_mem_functions will always give the
5561 expected result (the new set of functions is used for allocation
5562 and deallocation) at all times, regardless of platform and compiler
5565 To finish it up, some functions that were never use in any other
5566 way than through macros have a new API and new semantic:
5569 CRYPTO_dbg_realloc()
5572 All macros of value have retained their old syntax.
5573 [Richard Levitte and Bodo Moeller]
5575 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5576 ordering of SMIMECapabilities wasn't in "strength order" and there
5577 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5581 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5582 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5583 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5585 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5586 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5587 functionality to handle multipart/signed properly) and a utility
5588 called 'smime' to call all this stuff. This is based on code I
5589 originally wrote for Celo who have kindly allowed it to be
5590 included in OpenSSL.
5593 *) Add variants des_set_key_checked and des_set_key_unchecked of
5594 des_set_key (aka des_key_sched). Global variable des_check_key
5595 decides which of these is called by des_set_key; this way
5596 des_check_key behaves as it always did, but applications and
5597 the library itself, which was buggy for des_check_key == 1,
5598 have a cleaner way to pick the version they need.
5601 *) New function PKCS12_newpass() which changes the password of a
5605 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5606 dynamic mix. In both cases the ids can be used as an index into the
5607 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5608 functions so they accept a list of the field values and the
5609 application doesn't need to directly manipulate the X509_TRUST
5613 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5617 *) Modify the way the V3 extension code looks up extensions. This now
5618 works in a similar way to the object code: we have some "standard"
5619 extensions in a static table which is searched with OBJ_bsearch()
5620 and the application can add dynamic ones if needed. The file
5621 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5622 updated whenever a new extension is added to the core code and kept
5623 in ext_nid order. There is a simple program 'tabtest.c' which checks
5624 this. New extensions are not added too often so this file can readily
5625 be maintained manually.
5627 There are two big advantages in doing things this way. The extensions
5628 can be looked up immediately and no longer need to be "added" using
5629 X509V3_add_standard_extensions(): this function now does nothing.
5630 [Side note: I get *lots* of email saying the extension code doesn't
5631 work because people forget to call this function]
5632 Also no dynamic allocation is done unless new extensions are added:
5633 so if we don't add custom extensions there is no need to call
5634 X509V3_EXT_cleanup().
5637 *) Modify enc utility's salting as follows: make salting the default. Add a
5638 magic header, so unsalted files fail gracefully instead of just decrypting
5639 to garbage. This is because not salting is a big security hole, so people
5640 should be discouraged from doing it.
5643 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5644 digest to be passed on the command line but it only used this
5645 parameter when signing a certificate. Modified so all relevant
5646 operations are affected by the digest parameter including the
5647 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5648 DSA key was used because it didn't fix the digest.
5651 *) Initial certificate chain verify code. Currently tests the untrusted
5652 certificates for consistency with the verify purpose (which is set
5653 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5655 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5656 this is because it will reject chains with invalid extensions whereas
5657 every previous version of OpenSSL and SSLeay made no checks at all.
5659 Trust code: checks the root CA for the relevant trust settings. Trust
5660 settings have an initial value consistent with the verify purpose: e.g.
5661 if the verify purpose is for SSL client use it expects the CA to be
5662 trusted for SSL client use. However the default value can be changed to
5663 permit custom trust settings: one example of this would be to only trust
5664 certificates from a specific "secure" set of CAs.
5666 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5667 which should be used for version portability: especially since the
5668 verify structure is likely to change more often now.
5670 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5671 to set them. If not set then assume SSL clients will verify SSL servers
5674 Two new options to the verify program: -untrusted allows a set of
5675 untrusted certificates to be passed in and -purpose which sets the
5676 intended purpose of the certificate. If a purpose is set then the
5677 new chain verify code is used to check extension consistency.
5680 *) Support for the authority information access extension.
5683 *) Modify RSA and DSA PEM read routines to transparently handle
5684 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5685 public keys in a format compatible with certificate
5686 SubjectPublicKeyInfo structures. Unfortunately there were already
5687 functions called *_PublicKey_* which used various odd formats so
5688 these are retained for compatibility: however the DSA variants were
5689 never in a public release so they have been deleted. Changed dsa/rsa
5690 utilities to handle the new format: note no releases ever handled public
5691 keys so we should be OK.
5693 The primary motivation for this change is to avoid the same fiasco
5694 that dogs private keys: there are several incompatible private key
5695 formats some of which are standard and some OpenSSL specific and
5696 require various evil hacks to allow partial transparent handling and
5697 even then it doesn't work with DER formats. Given the option anything
5698 other than PKCS#8 should be dumped: but the other formats have to
5699 stay in the name of compatibility.
5701 With public keys and the benefit of hindsight one standard format
5702 is used which works with EVP_PKEY, RSA or DSA structures: though
5703 it clearly returns an error if you try to read the wrong kind of key.
5705 Added a -pubkey option to the 'x509' utility to output the public key.
5706 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5707 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5708 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5709 that do the same as the EVP_PKEY_assign_*() except they up the
5710 reference count of the added key (they don't "swallow" the
5714 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5715 CRLs would fail if the file contained no certificates or no CRLs:
5716 added a new function to read in both types and return the number
5717 read: this means that if none are read it will be an error. The
5718 DER versions of the certificate and CRL reader would always fail
5719 because it isn't possible to mix certificates and CRLs in DER format
5720 without choking one or the other routine. Changed this to just read
5721 a certificate: this is the best we can do. Also modified the code
5722 in apps/verify.c to take notice of return codes: it was previously
5723 attempting to read in certificates from NULL pointers and ignoring
5724 any errors: this is one reason why the cert and CRL reader seemed
5725 to work. It doesn't check return codes from the default certificate
5726 routines: these may well fail if the certificates aren't installed.
5729 *) Code to support otherName option in GeneralName.
5732 *) First update to verify code. Change the verify utility
5733 so it warns if it is passed a self signed certificate:
5734 for consistency with the normal behaviour. X509_verify
5735 has been modified to it will now verify a self signed
5736 certificate if *exactly* the same certificate appears
5737 in the store: it was previously impossible to trust a
5738 single self signed certificate. This means that:
5739 openssl verify ss.pem
5740 now gives a warning about a self signed certificate but
5741 openssl verify -CAfile ss.pem ss.pem
5745 *) For servers, store verify_result in SSL_SESSION data structure
5746 (and add it to external session representation).
5747 This is needed when client certificate verifications fails,
5748 but an application-provided verification callback (set by
5749 SSL_CTX_set_cert_verify_callback) allows accepting the session
5750 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5751 but returns 1): When the session is reused, we have to set
5752 ssl->verify_result to the appropriate error code to avoid
5754 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5756 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5757 case in PKCS7_dataInit() where the signed PKCS7 structure
5758 didn't contain any existing data because it was being created.
5759 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5761 *) Add a salt to the key derivation routines in enc.c. This
5762 forms the first 8 bytes of the encrypted file. Also add a
5763 -S option to allow a salt to be input on the command line.
5766 *) New function X509_cmp(). Oddly enough there wasn't a function
5767 to compare two certificates. We do this by working out the SHA1
5768 hash and comparing that. X509_cmp() will be needed by the trust
5772 *) SSL_get1_session() is like SSL_get_session(), but increments
5773 the reference count in the SSL_SESSION returned.
5774 [Geoff Thorpe <geoff@eu.c2.net>]
5776 *) Fix for 'req': it was adding a null to request attributes.
5777 Also change the X509_LOOKUP and X509_INFO code to handle
5778 certificate auxiliary information.
5781 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5785 *) Add the possibility to add extra information to the memory leak
5786 detecting output, to form tracebacks, showing from where each
5787 allocation was originated: CRYPTO_push_info("constant string") adds
5788 the string plus current file name and line number to a per-thread
5789 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5790 is like calling CYRPTO_pop_info() until the stack is empty.
5791 Also updated memory leak detection code to be multi-thread-safe.
5794 *) Add options -text and -noout to pkcs7 utility and delete the
5795 encryption options which never did anything. Update docs.
5798 *) Add options to some of the utilities to allow the pass phrase
5799 to be included on either the command line (not recommended on
5800 OSes like Unix) or read from the environment. Update the
5801 manpages and fix a few bugs.
5804 *) Add a few manpages for some of the openssl commands.
5807 *) Fix the -revoke option in ca. It was freeing up memory twice,
5808 leaking and not finding already revoked certificates.
5811 *) Extensive changes to support certificate auxiliary information.
5812 This involves the use of X509_CERT_AUX structure and X509_AUX
5813 functions. An X509_AUX function such as PEM_read_X509_AUX()
5814 can still read in a certificate file in the usual way but it
5815 will also read in any additional "auxiliary information". By
5816 doing things this way a fair degree of compatibility can be
5817 retained: existing certificates can have this information added
5818 using the new 'x509' options.
5820 Current auxiliary information includes an "alias" and some trust
5821 settings. The trust settings will ultimately be used in enhanced
5822 certificate chain verification routines: currently a certificate
5823 can only be trusted if it is self signed and then it is trusted
5827 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5828 The problem was that one of the replacement routines had not been working
5829 since SSLeay releases. For now the offending routine has been replaced
5830 with non-optimised assembler. Even so, this now gives around 95%
5831 performance improvement for 1024 bit RSA signs.
5834 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5835 handling. Most clients have the effective key size in bits equal to
5836 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5837 A few however don't do this and instead use the size of the decrypted key
5838 to determine the RC2 key length and the AlgorithmIdentifier to determine
5839 the effective key length. In this case the effective key length can still
5840 be 40 bits but the key length can be 168 bits for example. This is fixed
5841 by manually forcing an RC2 key into the EVP_PKEY structure because the
5842 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5843 the key length and effective key length are equal.
5846 *) Add a bunch of functions that should simplify the creation of
5847 X509_NAME structures. Now you should be able to do:
5848 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5849 and have it automatically work out the correct field type and fill in
5850 the structures. The more adventurous can try:
5851 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5852 and it will (hopefully) work out the correct multibyte encoding.
5855 *) Change the 'req' utility to use the new field handling and multibyte
5856 copy routines. Before the DN field creation was handled in an ad hoc
5857 way in req, ca, and x509 which was rather broken and didn't support
5858 BMPStrings or UTF8Strings. Since some software doesn't implement
5859 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5860 using the dirstring_type option. See the new comment in the default
5861 openssl.cnf for more info.
5864 *) Make crypto/rand/md_rand.c more robust:
5865 - Assure unique random numbers after fork().
5866 - Make sure that concurrent threads access the global counter and
5867 md serializably so that we never lose entropy in them
5868 or use exactly the same state in multiple threads.
5869 Access to the large state is not always serializable because
5870 the additional locking could be a performance killer, and
5871 md should be large enough anyway.
5874 *) New file apps/app_rand.c with commonly needed functionality
5875 for handling the random seed file.
5877 Use the random seed file in some applications that previously did not:
5879 dsaparam -genkey (which also ignored its '-rand' option),
5882 x509 (when signing).
5883 Except on systems with /dev/urandom, it is crucial to have a random
5884 seed file at least for key creation, DSA signing, and for DH exchanges;
5885 for RSA signatures we could do without one.
5887 gendh and gendsa (unlike genrsa) used to read only the first byte
5888 of each file listed in the '-rand' option. The function as previously
5889 found in genrsa is now in app_rand.c and is used by all programs
5890 that support '-rand'.
5893 *) In RAND_write_file, use mode 0600 for creating files;
5894 don't just chmod when it may be too late.
5897 *) Report an error from X509_STORE_load_locations
5898 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5901 *) New function ASN1_mbstring_copy() this copies a string in either
5902 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5903 into an ASN1_STRING type. A mask of permissible types is passed
5904 and it chooses the "minimal" type to use or an error if not type
5908 *) Add function equivalents to the various macros in asn1.h. The old
5909 macros are retained with an M_ prefix. Code inside the library can
5910 use the M_ macros. External code (including the openssl utility)
5911 should *NOT* in order to be "shared library friendly".
5914 *) Add various functions that can check a certificate's extensions
5915 to see if it usable for various purposes such as SSL client,
5916 server or S/MIME and CAs of these types. This is currently
5917 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5918 verification. Also added a -purpose flag to x509 utility to
5919 print out all the purposes.
5922 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5926 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5927 for, obtain and decode and extension and obtain its critical flag.
5928 This allows all the necessary extension code to be handled in a
5929 single function call.
5932 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5933 platforms. See crypto/rc4/rc4_enc.c for further details.
5936 *) New -noout option to asn1parse. This causes no output to be produced
5937 its main use is when combined with -strparse and -out to extract data
5938 from a file (which may not be in ASN.1 format).
5941 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5942 when producing the local key id.
5943 [Richard Levitte <levitte@stacken.kth.se>]
5945 *) New option -dhparam in s_server. This allows a DH parameter file to be
5946 stated explicitly. If it is not stated then it tries the first server
5947 certificate file. The previous behaviour hard coded the filename
5951 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5952 a public key to be input or output. For example:
5953 openssl rsa -in key.pem -pubout -out pubkey.pem
5954 Also added necessary DSA public key functions to handle this.
5957 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5958 in the message. This was handled by allowing
5959 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5960 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5962 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5963 to the end of the strings whereas this didn't. This would cause problems
5964 if strings read with d2i_ASN1_bytes() were later modified.
5965 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5967 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5968 data and it contains EOF it will end up returning an error. This is
5969 caused by input 46 bytes long. The cause is due to the way base64
5970 BIOs find the start of base64 encoded data. They do this by trying a
5971 trial decode on each line until they find one that works. When they
5972 do a flag is set and it starts again knowing it can pass all the
5973 data directly through the decoder. Unfortunately it doesn't reset
5974 the context it uses. This means that if EOF is reached an attempt
5975 is made to pass two EOFs through the context and this causes the
5976 resulting error. This can also cause other problems as well. As is
5977 usual with these problems it takes *ages* to find and the fix is
5978 trivial: move one line.
5979 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5981 *) Ugly workaround to get s_client and s_server working under Windows. The
5982 old code wouldn't work because it needed to select() on sockets and the
5983 tty (for keypresses and to see if data could be written). Win32 only
5984 supports select() on sockets so we select() with a 1s timeout on the
5985 sockets and then see if any characters are waiting to be read, if none
5986 are present then we retry, we also assume we can always write data to
5987 the tty. This isn't nice because the code then blocks until we've
5988 received a complete line of data and it is effectively polling the
5989 keyboard at 1s intervals: however it's quite a bit better than not
5990 working at all :-) A dedicated Windows application might handle this
5991 with an event loop for example.
5994 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5995 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5996 will be called when RSA_sign() and RSA_verify() are used. This is useful
5997 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5998 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5999 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6000 This necessitated the support of an extra signature type NID_md5_sha1
6001 for SSL signatures and modifications to the SSL library to use it instead
6002 of calling RSA_public_decrypt() and RSA_private_encrypt().
6005 *) Add new -verify -CAfile and -CApath options to the crl program, these
6006 will lookup a CRL issuers certificate and verify the signature in a
6007 similar way to the verify program. Tidy up the crl program so it
6008 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6009 less strict. It will now permit CRL extensions even if it is not
6010 a V2 CRL: this will allow it to tolerate some broken CRLs.
6013 *) Initialize all non-automatic variables each time one of the openssl
6014 sub-programs is started (this is necessary as they may be started
6015 multiple times from the "OpenSSL>" prompt).
6016 [Lennart Bang, Bodo Moeller]
6018 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6019 removing all other RSA functionality (this is what NO_RSA does). This
6020 is so (for example) those in the US can disable those operations covered
6021 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6025 *) Non-copying interface to BIO pairs.
6026 (still largely untested)
6029 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6030 ASCII string. This was handled independently in various places before.
6033 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6034 UTF8 strings a character at a time.
6037 *) Use client_version from client hello to select the protocol
6038 (s23_srvr.c) and for RSA client key exchange verification
6039 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6042 *) Add various utility functions to handle SPKACs, these were previously
6043 handled by poking round in the structure internals. Added new function
6044 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6045 print, verify and generate SPKACs. Based on an original idea from
6046 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6049 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6052 *) Allow the config file extension section to be overwritten on the
6053 command line. Based on an original idea from Massimiliano Pala
6054 <madwolf@comune.modena.it>. The new option is called -extensions
6055 and can be applied to ca, req and x509. Also -reqexts to override
6056 the request extensions in req and -crlexts to override the crl extensions
6060 *) Add new feature to the SPKAC handling in ca. Now you can include
6061 the same field multiple times by preceding it by "XXXX." for example:
6064 this is the same syntax as used in the req config file.
6067 *) Allow certificate extensions to be added to certificate requests. These
6068 are specified in a 'req_extensions' option of the req section of the
6069 config file. They can be printed out with the -text option to req but
6070 are otherwise ignored at present.
6073 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6074 data read consists of only the final block it would not decrypted because
6075 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6076 A misplaced 'break' also meant the decrypted final block might not be
6077 copied until the next read.
6080 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6081 a few extra parameters to the DH structure: these will be useful if
6082 for example we want the value of 'q' or implement X9.42 DH.
6085 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6086 provides hooks that allow the default DSA functions or functions on a
6087 "per key" basis to be replaced. This allows hardware acceleration and
6088 hardware key storage to be handled without major modification to the
6089 library. Also added low level modexp hooks and CRYPTO_EX structure and
6090 associated functions.
6093 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6094 as "read only": it can't be written to and the buffer it points to will
6095 not be freed. Reading from a read only BIO is much more efficient than
6096 a normal memory BIO. This was added because there are several times when
6097 an area of memory needs to be read from a BIO. The previous method was
6098 to create a memory BIO and write the data to it, this results in two
6099 copies of the data and an O(n^2) reading algorithm. There is a new
6100 function BIO_new_mem_buf() which creates a read only memory BIO from
6101 an area of memory. Also modified the PKCS#7 routines to use read only
6105 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6106 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6107 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6108 but a retry condition occured while trying to read the rest.
6111 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6112 NID_pkcs7_encrypted by default: this was wrong since this should almost
6113 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6114 the encrypted data type: this is a more sensible place to put it and it
6115 allows the PKCS#12 code to be tidied up that duplicated this
6119 *) Changed obj_dat.pl script so it takes its input and output files on
6120 the command line. This should avoid shell escape redirection problems
6124 *) Initial support for certificate extension requests, these are included
6125 in things like Xenroll certificate requests. Included functions to allow
6126 extensions to be obtained and added.
6129 *) -crlf option to s_client and s_server for sending newlines as
6130 CRLF (as required by many protocols).
6133 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6135 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6136 [Ralf S. Engelschall]
6138 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6139 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6141 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6145 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6146 DH parameters/keys (q is lost during that conversion, but the resulting
6147 DH parameters contain its length).
6149 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6150 much faster than DH_generate_parameters (which creates parameters
6151 where p = 2*q + 1), and also the smaller q makes DH computations
6152 much more efficient (160-bit exponentiation instead of 1024-bit
6153 exponentiation); so this provides a convenient way to support DHE
6154 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6155 utter importance to use
6156 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6158 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6159 when such DH parameters are used, because otherwise small subgroup
6160 attacks may become possible!
6163 *) Avoid memory leak in i2d_DHparams.
6166 *) Allow the -k option to be used more than once in the enc program:
6167 this allows the same encrypted message to be read by multiple recipients.
6170 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6171 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6172 it will always use the numerical form of the OID, even if it has a short
6176 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6177 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6178 otherwise bn_mod_exp was called. In the case of hardware keys for example
6179 no private key components need be present and it might store extra data
6180 in the RSA structure, which cannot be accessed from bn_mod_exp.
6181 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6182 private key operations.
6185 *) Added support for SPARC Linux.
6188 *) pem_password_cb function type incompatibly changed from
6189 typedef int pem_password_cb(char *buf, int size, int rwflag);
6191 ....(char *buf, int size, int rwflag, void *userdata);
6192 so that applications can pass data to their callbacks:
6193 The PEM[_ASN1]_{read,write}... functions and macros now take an
6194 additional void * argument, which is just handed through whenever
6195 the password callback is called.
6196 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6198 New function SSL_CTX_set_default_passwd_cb_userdata.
6200 Compatibility note: As many C implementations push function arguments
6201 onto the stack in reverse order, the new library version is likely to
6202 interoperate with programs that have been compiled with the old
6203 pem_password_cb definition (PEM_whatever takes some data that
6204 happens to be on the stack as its last argument, and the callback
6205 just ignores this garbage); but there is no guarantee whatsoever that
6208 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6209 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6210 problems not only on Windows, but also on some Unix platforms.
6211 To avoid problematic command lines, these definitions are now in an
6212 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6213 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6216 *) MIPS III/IV assembler module is reimplemented.
6219 *) More DES library cleanups: remove references to srand/rand and
6220 delete an unused file.
6223 *) Add support for the the free Netwide assembler (NASM) under Win32,
6224 since not many people have MASM (ml) and it can be hard to obtain.
6225 This is currently experimental but it seems to work OK and pass all
6226 the tests. Check out INSTALL.W32 for info.
6229 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6230 without temporary keys kept an extra copy of the server key,
6231 and connections with temporary keys did not free everything in case
6235 *) New function RSA_check_key and new openssl rsa option -check
6236 for verifying the consistency of RSA keys.
6237 [Ulf Moeller, Bodo Moeller]
6239 *) Various changes to make Win32 compile work:
6240 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6241 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6242 comparison" warnings.
6243 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6246 *) Add a debugging option to PKCS#5 v2 key generation function: when
6247 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6248 derived keys are printed to stderr.
6251 *) Copy the flags in ASN1_STRING_dup().
6252 [Roman E. Pavlov <pre@mo.msk.ru>]
6254 *) The x509 application mishandled signing requests containing DSA
6255 keys when the signing key was also DSA and the parameters didn't match.
6257 It was supposed to omit the parameters when they matched the signing key:
6258 the verifying software was then supposed to automatically use the CA's
6259 parameters if they were absent from the end user certificate.
6261 Omitting parameters is no longer recommended. The test was also
6262 the wrong way round! This was probably due to unusual behaviour in
6263 EVP_cmp_parameters() which returns 1 if the parameters match.
6264 This meant that parameters were omitted when they *didn't* match and
6265 the certificate was useless. Certificates signed with 'ca' didn't have
6267 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6269 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6270 The interface is as follows:
6271 Applications can use
6272 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6273 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6274 "off" is now the default.
6275 The library internally uses
6276 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6277 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6278 to disable memory-checking temporarily.
6280 Some inconsistent states that previously were possible (and were
6281 even the default) are now avoided.
6283 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6284 with each memory chunk allocated; this is occasionally more helpful
6285 than just having a counter.
6287 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6289 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6293 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6294 which largely parallels "options", but is for changing API behaviour,
6295 whereas "options" are about protocol behaviour.
6296 Initial "mode" flags are:
6298 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6299 a single record has been written.
6300 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6301 retries use the same buffer location.
6302 (But all of the contents must be
6306 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6309 *) Fix problems with no-hmac etc.
6310 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6312 *) New functions RSA_get_default_method(), RSA_set_method() and
6313 RSA_get_method(). These allows replacement of RSA_METHODs without having
6314 to mess around with the internals of an RSA structure.
6317 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6318 Also really enable memory leak checks in openssl.c and in some
6320 [Chad C. Mulligan, Bodo Moeller]
6322 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6323 up the length of negative integers. This has now been simplified to just
6324 store the length when it is first determined and use it later, rather
6325 than trying to keep track of where data is copied and updating it to
6327 [Steve Henson, reported by Brien Wheeler
6328 <bwheeler@authentica-security.com>]
6330 *) Add a new function PKCS7_signatureVerify. This allows the verification
6331 of a PKCS#7 signature but with the signing certificate passed to the
6332 function itself. This contrasts with PKCS7_dataVerify which assumes the
6333 certificate is present in the PKCS#7 structure. This isn't always the
6334 case: certificates can be omitted from a PKCS#7 structure and be
6335 distributed by "out of band" means (such as a certificate database).
6338 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6339 function prototypes in pem.h, also change util/mkdef.pl to add the
6340 necessary function names.
6343 *) mk1mf.pl (used by Windows builds) did not properly read the
6344 options set by Configure in the top level Makefile, and Configure
6345 was not even able to write more than one option correctly.
6346 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6349 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6350 file to be loaded from a BIO or FILE pointer. The BIO version will
6351 for example allow memory BIOs to contain config info.
6354 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6355 Whoever hopes to achieve shared-library compatibility across versions
6356 must use this, not the compile-time macro.
6357 (Exercise 0.9.4: Which is the minimum library version required by
6359 Note: All this applies only to multi-threaded programs, others don't
6363 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6364 through a BIO pair triggered the default case, i.e.
6365 SSLerr(...,SSL_R_UNKNOWN_STATE).
6368 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6369 can use the SSL library even if none of the specific BIOs is
6373 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6374 for the encoded length.
6375 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6377 *) Add initial documentation of the X509V3 functions.
6380 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6381 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6382 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6383 secure PKCS#8 private key format with a high iteration count.
6386 *) Fix determination of Perl interpreter: A perl or perl5
6387 _directory_ in $PATH was also accepted as the interpreter.
6388 [Ralf S. Engelschall]
6390 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6391 wrong with it but it was very old and did things like calling
6392 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6396 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6397 to use the new extension code.
6400 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6401 with macros. This should make it easier to change their form, add extra
6402 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6406 *) Add to configuration table a new entry that can specify an alternative
6407 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6408 according to Mark Crispin <MRC@Panda.COM>.
6412 *) DES CBC did not update the IV. Weird.
6415 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6416 Changing the behaviour of the former might break existing programs --
6417 where IV updating is needed, des_ncbc_encrypt can be used.
6420 *) When bntest is run from "make test" it drives bc to check its
6421 calculations, as well as internally checking them. If an internal check
6422 fails, it needs to cause bc to give a non-zero result or make test carries
6423 on without noticing the failure. Fixed.
6426 *) DES library cleanups.
6429 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6430 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6431 ciphers. NOTE: although the key derivation function has been verified
6432 against some published test vectors it has not been extensively tested
6433 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6437 *) Instead of "mkdir -p", which is not fully portable, use new
6438 Perl script "util/mkdir-p.pl".
6441 *) Rewrite the way password based encryption (PBE) is handled. It used to
6442 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6443 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6444 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6445 the 'parameter' field of the AlgorithmIdentifier is passed to the
6446 underlying key generation function so it must do its own ASN1 parsing.
6447 This has also changed the EVP_PBE_CipherInit() function which now has a
6448 'parameter' argument instead of literal salt and iteration count values
6449 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6452 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6453 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6454 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6455 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6456 value was just used as a "magic string" and not used directly its
6457 value doesn't matter.
6460 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6464 *) "linux-sparc64" configuration (ultrapenguin).
6465 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6466 "linux-sparc" configuration.
6467 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6469 *) config now generates no-xxx options for missing ciphers.
6472 *) Support the EBCDIC character set (work in progress).
6473 File ebcdic.c not yet included because it has a different license.
6474 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6476 *) Support BS2000/OSD-POSIX.
6477 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6479 *) Make callbacks for key generation use void * instead of char *.
6482 *) Make S/MIME samples compile (not yet tested).
6485 *) Additional typesafe stacks.
6488 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6492 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6494 *) New configuration variant "sco5-gcc".
6496 *) Updated some demos.
6497 [Sean O Riordain, Wade Scholine]
6499 *) Add missing BIO_free at exit of pkcs12 application.
6502 *) Fix memory leak in conf.c.
6505 *) Updates for Win32 to assembler version of MD5.
6508 *) Set #! path to perl in apps/der_chop to where we found it
6509 instead of using a fixed path.
6512 *) SHA library changes for irix64-mips4-cc.
6515 *) Improvements for VMS support.
6519 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6521 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6522 This also avoids the problems with SC4.2 and unpatched SC5.
6523 [Andy Polyakov <appro@fy.chalmers.se>]
6525 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6526 These are required because of the typesafe stack would otherwise break
6527 existing code. If old code used a structure member which used to be STACK
6528 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6529 sk_num or sk_value it would produce an error because the num, data members
6530 are not present in STACK_OF. Now it just produces a warning. sk_set
6531 replaces the old method of assigning a value to sk_value
6532 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6533 that does this will no longer work (and should use sk_set instead) but
6534 this could be regarded as a "questionable" behaviour anyway.
6537 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6538 correctly handle encrypted S/MIME data.
6541 *) Change type of various DES function arguments from des_cblock
6542 (which means, in function argument declarations, pointer to char)
6543 to des_cblock * (meaning pointer to array with 8 char elements),
6544 which allows the compiler to do more typechecking; it was like
6545 that back in SSLeay, but with lots of ugly casts.
6547 Introduce new type const_des_cblock.
6550 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6551 problems: find RecipientInfo structure that matches recipient certificate
6552 and initialise the ASN1 structures properly based on passed cipher.
6555 *) Belatedly make the BN tests actually check the results.
6558 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6559 to and from BNs: it was completely broken. New compilation option
6560 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6561 key elements as negative integers.
6564 *) Reorganize and speed up MD5.
6565 [Andy Polyakov <appro@fy.chalmers.se>]
6568 [Richard Levitte <richard@levitte.org>]
6570 *) New option -out to asn1parse to allow the parsed structure to be
6571 output to a file. This is most useful when combined with the -strparse
6572 option to examine the output of things like OCTET STRINGS.
6575 *) Make SSL library a little more fool-proof by not requiring any longer
6576 that SSL_set_{accept,connect}_state be called before
6577 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6578 in many applications because usually everything *appeared* to work as
6579 intended anyway -- now it really works as intended).
6582 *) Move openssl.cnf out of lib/.
6585 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6586 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6587 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6588 [Ralf S. Engelschall]
6590 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6591 handle PKCS#7 enveloped data properly.
6592 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6594 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6595 copying pointers. The cert_st handling is changed by this in
6596 various ways (and thus what used to be known as ctx->default_cert
6597 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6598 any longer when s->cert does not give us what we need).
6599 ssl_cert_instantiate becomes obsolete by this change.
6600 As soon as we've got the new code right (possibly it already is?),
6601 we have solved a couple of bugs of the earlier code where s->cert
6602 was used as if it could not have been shared with other SSL structures.
6604 Note that using the SSL API in certain dirty ways now will result
6605 in different behaviour than observed with earlier library versions:
6606 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6607 does not influence s as it used to.
6609 In order to clean up things more thoroughly, inside SSL_SESSION
6610 we don't use CERT any longer, but a new structure SESS_CERT
6611 that holds per-session data (if available); currently, this is
6612 the peer's certificate chain and, for clients, the server's certificate
6613 and temporary key. CERT holds only those values that can have
6614 meaningful defaults in an SSL_CTX.
6617 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6618 from the internal representation. Various PKCS#7 fixes: remove some
6619 evil casts and set the enc_dig_alg field properly based on the signing
6623 *) Allow PKCS#12 password to be set from the command line or the
6624 environment. Let 'ca' get its config file name from the environment
6625 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6629 *) Allow certificate policies extension to use an IA5STRING for the
6630 organization field. This is contrary to the PKIX definition but
6631 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6635 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6636 without disallowing inline assembler and the like for non-pedantic builds.
6639 *) Support Borland C++ builder.
6640 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6645 *) SHA-1 cleanups and performance enhancements.
6646 [Andy Polyakov <appro@fy.chalmers.se>]
6648 *) Sparc v8plus assembler for the bignum library.
6649 [Andy Polyakov <appro@fy.chalmers.se>]
6651 *) Accept any -xxx and +xxx compiler options in Configure.
6654 *) Update HPUX configuration.
6657 *) Add missing sk_<type>_unshift() function to safestack.h
6658 [Ralf S. Engelschall]
6660 *) New function SSL_CTX_use_certificate_chain_file that sets the
6661 "extra_cert"s in addition to the certificate. (This makes sense
6662 only for "PEM" format files, as chains as a whole are not
6666 *) Support verify_depth from the SSL API.
6667 x509_vfy.c had what can be considered an off-by-one-error:
6668 Its depth (which was not part of the external interface)
6669 was actually counting the number of certificates in a chain;
6670 now it really counts the depth.
6673 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6674 instead of X509err, which often resulted in confusing error
6675 messages since the error codes are not globally unique
6676 (e.g. an alleged error in ssl3_accept when a certificate
6677 didn't match the private key).
6679 *) New function SSL_CTX_set_session_id_context that allows to set a default
6680 value (so that you don't need SSL_set_session_id_context for each
6681 connection using the SSL_CTX).
6684 *) OAEP decoding bug fix.
6687 *) Support INSTALL_PREFIX for package builders, as proposed by
6691 *) New Configure options "threads" and "no-threads". For systems
6692 where the proper compiler options are known (currently Solaris
6693 and Linux), "threads" is the default.
6696 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6699 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6700 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6701 such as /usr/local/bin.
6704 *) "make linux-shared" to build shared libraries.
6705 [Niels Poppe <niels@netbox.org>]
6707 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6710 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6711 extension adding in x509 utility.
6714 *) Remove NOPROTO sections and error code comments.
6717 *) Partial rewrite of the DEF file generator to now parse the ANSI
6721 *) New Configure options --prefix=DIR and --openssldir=DIR.
6724 *) Complete rewrite of the error code script(s). It is all now handled
6725 by one script at the top level which handles error code gathering,
6726 header rewriting and C source file generation. It should be much better
6727 than the old method: it now uses a modified version of Ulf's parser to
6728 read the ANSI prototypes in all header files (thus the old K&R definitions
6729 aren't needed for error creation any more) and do a better job of
6730 translating function codes into names. The old 'ASN1 error code imbedded
6731 in a comment' is no longer necessary and it doesn't use .err files which
6732 have now been deleted. Also the error code call doesn't have to appear all
6733 on one line (which resulted in some large lines...).
6736 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6739 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6740 0 (which usually indicates a closed connection), but continue reading.
6743 *) Fix some race conditions.
6746 *) Add support for CRL distribution points extension. Add Certificate
6747 Policies and CRL distribution points documentation.
6750 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6753 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6754 8 of keying material. Merlin has also confirmed interop with this fix
6755 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6756 [Merlin Hughes <merlin@baltimore.ie>]
6758 *) Fix lots of warnings.
6759 [Richard Levitte <levitte@stacken.kth.se>]
6761 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6762 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6763 [Richard Levitte <levitte@stacken.kth.se>]
6765 *) Fix problems with sizeof(long) == 8.
6766 [Andy Polyakov <appro@fy.chalmers.se>]
6768 *) Change functions to ANSI C.
6771 *) Fix typos in error codes.
6772 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6774 *) Remove defunct assembler files from Configure.
6777 *) SPARC v8 assembler BIGNUM implementation.
6778 [Andy Polyakov <appro@fy.chalmers.se>]
6780 *) Support for Certificate Policies extension: both print and set.
6781 Various additions to support the r2i method this uses.
6784 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6785 return a const string when you are expecting an allocated buffer.
6788 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6789 types DirectoryString and DisplayText.
6792 *) Add code to allow r2i extensions to access the configuration database,
6793 add an LHASH database driver and add several ctx helper functions.
6796 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6797 fail when they extended the size of a BIGNUM.
6800 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6801 support typesafe stack.
6804 *) Fix typo in SSL_[gs]et_options().
6805 [Nils Frostberg <nils@medcom.se>]
6807 *) Delete various functions and files that belonged to the (now obsolete)
6808 old X509V3 handling code.
6811 *) New Configure option "rsaref".
6814 *) Don't auto-generate pem.h.
6817 *) Introduce type-safe ASN.1 SETs.
6820 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6821 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6823 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6824 that links with OpenSSL (well at least cause lots of warnings), but fear
6825 not: the conversion is trivial, and it eliminates loads of evil casts. A
6826 few STACKed things have been converted already. Feel free to convert more.
6827 In the fullness of time, I'll do away with the STACK type altogether.
6830 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6831 specified in <certfile> by updating the entry in the index.txt file.
6832 This way one no longer has to edit the index.txt file manually for
6833 revoking a certificate. The -revoke option does the gory details now.
6834 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6836 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6837 `-text' option at all and this way the `-noout -text' combination was
6838 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6839 [Ralf S. Engelschall]
6841 *) Make sure a corresponding plain text error message exists for the
6842 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6843 verify callback function determined that a certificate was revoked.
6844 [Ralf S. Engelschall]
6846 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6847 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6848 all available cipers including rc5, which was forgotten until now.
6849 In order to let the testing shell script know which algorithms
6850 are available, a new (up to now undocumented) command
6851 "openssl list-cipher-commands" is used.
6854 *) Bugfix: s_client occasionally would sleep in select() when
6855 it should have checked SSL_pending() first.
6858 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6859 the raw DSA values prior to ASN.1 encoding.
6862 *) Tweaks to Configure
6863 [Niels Poppe <niels@netbox.org>]
6865 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6869 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6872 *) New config option to avoid instructions that are illegal on the 80386.
6873 The default code is faster, but requires at least a 486.
6876 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6877 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6878 same as SSL2_VERSION anyway.
6881 *) New "-showcerts" option for s_client.
6884 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6885 application. Various cleanups and fixes.
6888 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6889 modify error routines to work internally. Add error codes and PBE init
6890 to library startup routines.
6893 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6894 packing functions to asn1 and evp. Changed function names and error
6895 codes along the way.
6898 *) PKCS12 integration: and so it begins... First of several patches to
6899 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6900 objects to objects.h
6903 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6904 and display support for Thawte strong extranet extension.
6907 *) Add LinuxPPC support.
6908 [Jeff Dubrule <igor@pobox.org>]
6910 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6911 bn_div_words in alpha.s.
6912 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6914 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6915 OAEP isn't supported when OpenSSL is built with RSAref.
6916 [Ulf Moeller <ulf@fitug.de>]
6918 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6919 so they no longer are missing under -DNOPROTO.
6920 [Soren S. Jorvang <soren@t.dk>]
6923 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6925 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6926 doesn't work when the session is reused. Coming soon!
6929 *) Fix a security hole, that allows sessions to be reused in the wrong
6930 context thus bypassing client cert protection! All software that uses
6931 client certs and session caches in multiple contexts NEEDS PATCHING to
6932 allow session reuse! A fuller solution is in the works.
6933 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6935 *) Some more source tree cleanups (removed obsolete files
6936 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6937 permission on "config" script to be executable) and a fix for the INSTALL
6939 [Ulf Moeller <ulf@fitug.de>]
6941 *) Remove some legacy and erroneous uses of malloc, free instead of
6943 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6945 *) Make rsa_oaep_test return non-zero on error.
6946 [Ulf Moeller <ulf@fitug.de>]
6948 *) Add support for native Solaris shared libraries. Configure
6949 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6950 if someone would make that last step automatic.
6951 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6953 *) ctx_size was not built with the right compiler during "make links". Fixed.
6956 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6957 except NULL ciphers". This means the default cipher list will no longer
6958 enable NULL ciphers. They need to be specifically enabled e.g. with
6959 the string "DEFAULT:eNULL".
6962 *) Fix to RSA private encryption routines: if p < q then it would
6963 occasionally produce an invalid result. This will only happen with
6964 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6967 *) Be less restrictive and allow also `perl util/perlpath.pl
6968 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6969 because this way one can also use an interpreter named `perl5' (which is
6970 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6971 installed as `perl').
6972 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6974 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6975 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6977 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6978 advapi32.lib to Win32 build and change the pem test comparision
6979 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6980 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6981 and crypto/des/ede_cbcm_enc.c.
6984 *) DES quad checksum was broken on big-endian architectures. Fixed.
6987 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6988 Win32 test batch file so it (might) work again. The Win32 test batch file
6989 is horrible: I feel ill....
6992 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6993 in e_os.h. Audit of header files to check ANSI and non ANSI
6994 sections: 10 functions were absent from non ANSI section and not exported
6995 from Windows DLLs. Fixed up libeay.num for new functions.
6998 *) Make `openssl version' output lines consistent.
6999 [Ralf S. Engelschall]
7001 *) Fix Win32 symbol export lists for BIO functions: Added
7002 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7003 to ms/libeay{16,32}.def.
7004 [Ralf S. Engelschall]
7006 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7007 fine under Unix and passes some trivial tests I've now added. But the
7008 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7009 added to make sure no one expects that this stuff really works in the
7010 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7011 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7013 [Ralf S. Engelschall]
7015 *) Fix the generation of two part addresses in perl.
7016 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7018 *) Add config entry for Linux on MIPS.
7019 [John Tobey <jtobey@channel1.com>]
7021 *) Make links whenever Configure is run, unless we are on Windoze.
7024 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7025 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7029 *) Add a useful kludge to allow package maintainers to specify compiler and
7030 other platforms details on the command line without having to patch the
7031 Configure script everytime: One now can use ``perl Configure
7032 <id>:<details>'', i.e. platform ids are allowed to have details appended
7033 to them (seperated by colons). This is treated as there would be a static
7034 pre-configured entry in Configure's %table under key <id> with value
7035 <details> and ``perl Configure <id>'' is called. So, when you want to
7036 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7037 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7038 now, which overrides the FreeBSD-elf entry on-the-fly.
7039 [Ralf S. Engelschall]
7041 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7044 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7045 on the `perl Configure ...' command line. This way one can compile
7046 OpenSSL libraries with Position Independent Code (PIC) which is needed
7047 for linking it into DSOs.
7048 [Ralf S. Engelschall]
7050 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7054 *) Cleaned up the LICENSE document: The official contact for any license
7055 questions now is the OpenSSL core team under openssl-core@openssl.org.
7056 And add a paragraph about the dual-license situation to make sure people
7057 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7058 to the OpenSSL toolkit.
7059 [Ralf S. Engelschall]
7061 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7062 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7063 Additonally cleaned up the `make links' target: Remove unnecessary
7064 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7065 to speed processing and no longer clutter the display with confusing
7066 stuff. Instead only the actually done links are displayed.
7067 [Ralf S. Engelschall]
7069 *) Permit null encryption ciphersuites, used for authentication only. It used
7070 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7071 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7075 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7076 signed attributes when verifying signatures (this would break them),
7077 the detached data encoding was wrong and public keys obtained using
7078 X509_get_pubkey() weren't freed.
7081 *) Add text documentation for the BUFFER functions. Also added a work around
7082 to a Win95 console bug. This was triggered by the password read stuff: the
7083 last character typed gets carried over to the next fread(). If you were
7084 generating a new cert request using 'req' for example then the last
7085 character of the passphrase would be CR which would then enter the first
7089 *) Added the new `Includes OpenSSL Cryptography Software' button as
7090 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7091 button and can be used by applications based on OpenSSL to show the
7092 relationship to the OpenSSL project.
7093 [Ralf S. Engelschall]
7095 *) Remove confusing variables in function signatures in files
7096 ssl/ssl_lib.c and ssl/ssl.h.
7097 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7099 *) Don't install bss_file.c under PREFIX/include/
7100 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7102 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7103 functions that return function pointers and has support for NT specific
7104 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7105 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7106 unsigned to signed types: this was killing the Win32 compile.
7109 *) Add new certificate file to stack functions,
7110 SSL_add_dir_cert_subjects_to_stack() and
7111 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7112 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7113 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7114 This means that Apache-SSL and similar packages don't have to mess around
7115 to add as many CAs as they want to the preferred list.
7118 *) Experiment with doxygen documentation. Currently only partially applied to
7120 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7121 openssl.doxy as the configuration file.
7124 *) Get rid of remaining C++-style comments which strict C compilers hate.
7125 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7127 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7128 compiled in by default: it has problems with large keys.
7131 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7132 DH private keys and/or callback functions which directly correspond to
7133 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7134 is needed for applications which have to configure certificates on a
7135 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7137 For the RSA certificate situation is makes no difference, but
7138 for the DSA certificate situation this fixes the "no shared cipher"
7139 problem where the OpenSSL cipher selection procedure failed because the
7140 temporary keys were not overtaken from the context and the API provided
7141 no way to reconfigure them.
7142 The new functions now let applications reconfigure the stuff and they
7143 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7144 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7145 non-public-API function ssl_cert_instantiate() is used as a helper
7146 function and also to reduce code redundancy inside ssl_rsa.c.
7147 [Ralf S. Engelschall]
7149 *) Move s_server -dcert and -dkey options out of the undocumented feature
7150 area because they are useful for the DSA situation and should be
7151 recognized by the users.
7152 [Ralf S. Engelschall]
7154 *) Fix the cipher decision scheme for export ciphers: the export bits are
7155 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7156 SSL_EXP_MASK. So, the original variable has to be used instead of the
7157 already masked variable.
7158 [Richard Levitte <levitte@stacken.kth.se>]
7160 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7161 [Richard Levitte <levitte@stacken.kth.se>]
7163 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7164 from `int' to `unsigned int' because it's a length and initialized by
7165 EVP_DigestFinal() which expects an `unsigned int *'.
7166 [Richard Levitte <levitte@stacken.kth.se>]
7168 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7169 script. Instead use the usual Shell->Perl transition trick.
7170 [Ralf S. Engelschall]
7172 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7173 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7174 -noout -modulus' as it's already the case for `openssl rsa -noout
7175 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7176 currently the public key is printed (a decision which was already done by
7177 `openssl dsa -modulus' in the past) which serves a similar purpose.
7178 Additionally the NO_RSA no longer completely removes the whole -modulus
7179 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7181 [Ralf S. Engelschall]
7183 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7184 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7185 [Arne Ansper <arne@ats.cyber.ee>]
7187 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7188 to be added. Now both 'req' and 'ca' can use new objects defined in the
7192 *) Add cool BIO that does syslog (or event log on NT).
7193 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7195 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7196 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7197 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7198 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7201 *) Add preliminary config info for new extension code.
7204 *) Make RSA_NO_PADDING really use no padding.
7205 [Ulf Moeller <ulf@fitug.de>]
7207 *) Generate errors when private/public key check is done.
7210 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7211 for some CRL extensions and new objects added.
7214 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7215 key usage extension and fuller support for authority key id.
7218 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7219 padding method for RSA, which is recommended for new applications in PKCS
7220 #1 v2.0 (RFC 2437, October 1998).
7221 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7222 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7223 against Bleichbacher's attack on RSA.
7224 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7227 *) Updates to the new SSL compression code
7228 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7230 *) Fix so that the version number in the master secret, when passed
7231 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7232 (because the server will not accept higher), that the version number
7233 is 0x03,0x01, not 0x03,0x00
7234 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7236 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7237 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7238 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7241 *) Support for RAW extensions where an arbitrary extension can be
7242 created by including its DER encoding. See apps/openssl.cnf for
7246 *) Make sure latest Perl versions don't interpret some generated C array
7247 code as Perl array code in the crypto/err/err_genc.pl script.
7248 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7250 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7251 not many people have the assembler. Various Win32 compilation fixes and
7252 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7256 *) Modify configure script 'Configure' to automatically create crypto/date.h
7257 file under Win32 and also build pem.h from pem.org. New script
7258 util/mkfiles.pl to create the MINFO file on environments that can't do a
7259 'make files': perl util/mkfiles.pl >MINFO should work.
7262 *) Major rework of DES function declarations, in the pursuit of correctness
7263 and purity. As a result, many evil casts evaporated, and some weirdness,
7264 too. You may find this causes warnings in your code. Zapping your evil
7265 casts will probably fix them. Mostly.
7268 *) Fix for a typo in asn1.h. Bug fix to object creation script
7269 obj_dat.pl. It considered a zero in an object definition to mean
7270 "end of object": none of the objects in objects.h have any zeros
7271 so it wasn't spotted.
7272 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7274 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7275 Masking (CBCM). In the absence of test vectors, the best I have been able
7276 to do is check that the decrypt undoes the encrypt, so far. Send me test
7277 vectors if you have them.
7280 *) Correct calculation of key length for export ciphers (too much space was
7281 allocated for null ciphers). This has not been tested!
7284 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7285 message is now correct (it understands "crypto" and "ssl" on its
7286 command line). There is also now an "update" option. This will update
7287 the util/ssleay.num and util/libeay.num files with any new functions.
7289 perl util/mkdef.pl crypto ssl update
7290 it will update them.
7293 *) Overhauled the Perl interface (perl/*):
7294 - ported BN stuff to OpenSSL's different BN library
7295 - made the perl/ source tree CVS-aware
7296 - renamed the package from SSLeay to OpenSSL (the files still contain
7297 their history because I've copied them in the repository)
7298 - removed obsolete files (the test scripts will be replaced
7299 by better Test::Harness variants in the future)
7300 [Ralf S. Engelschall]
7302 *) First cut for a very conservative source tree cleanup:
7303 1. merge various obsolete readme texts into doc/ssleay.txt
7304 where we collect the old documents and readme texts.
7305 2. remove the first part of files where I'm already sure that we no
7306 longer need them because of three reasons: either they are just temporary
7307 files which were left by Eric or they are preserved original files where
7308 I've verified that the diff is also available in the CVS via "cvs diff
7309 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7310 the crypto/md/ stuff).
7311 [Ralf S. Engelschall]
7313 *) More extension code. Incomplete support for subject and issuer alt
7314 name, issuer and authority key id. Change the i2v function parameters
7315 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7316 what that's for :-) Fix to ASN1 macro which messed up
7317 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7320 *) Preliminary support for ENUMERATED type. This is largely copied from the
7324 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7325 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7327 *) Make sure `make rehash' target really finds the `openssl' program.
7328 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7330 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7331 like to hear about it if this slows down other processors.
7334 *) Add CygWin32 platform information to Configure script.
7335 [Alan Batie <batie@aahz.jf.intel.com>]
7337 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7338 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7340 *) New program nseq to manipulate netscape certificate sequences
7343 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7347 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7348 but the BN code had some problems that would cause failures when
7349 doing certificate verification and some other functions.
7350 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7352 *) Add ASN1 and PEM code to support netscape certificate sequences.
7355 *) Add ASN1 and PEM code to support netscape certificate sequences.
7358 *) Add several PKIX and private extended key usage OIDs.
7361 *) Modify the 'ca' program to handle the new extension code. Modify
7362 openssl.cnf for new extension format, add comments.
7365 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7366 and add a sample to openssl.cnf so req -x509 now adds appropriate
7370 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7371 error code, add initial support to X509_print() and x509 application.
7374 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7375 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7376 stuff is currently isolated and isn't even compiled yet.
7379 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7380 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7381 Removed the versions check from X509 routines when loading extensions:
7382 this allows certain broken certificates that don't set the version
7383 properly to be processed.
7386 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7387 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7388 can still be regenerated with "make depend".
7391 *) Spelling mistake in C version of CAST-128.
7392 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7394 *) Changes to the error generation code. The perl script err-code.pl
7395 now reads in the old error codes and retains the old numbers, only
7396 adding new ones if necessary. It also only changes the .err files if new
7397 codes are added. The makefiles have been modified to only insert errors
7398 when needed (to avoid needlessly modifying header files). This is done
7399 by only inserting errors if the .err file is newer than the auto generated
7400 C file. To rebuild all the error codes from scratch (the old behaviour)
7401 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7402 or delete all the .err files.
7405 *) CAST-128 was incorrectly implemented for short keys. The C version has
7406 been fixed, but is untested. The assembler versions are also fixed, but
7407 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7408 to regenerate it if needed.
7409 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7410 Hagino <itojun@kame.net>]
7412 *) File was opened incorrectly in randfile.c.
7413 [Ulf Möller <ulf@fitug.de>]
7415 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7416 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7417 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7418 al: it's just almost always a UTCTime. Note this patch adds new error
7419 codes so do a "make errors" if there are problems.
7422 *) Correct Linux 1 recognition in config.
7423 [Ulf Möller <ulf@fitug.de>]
7425 *) Remove pointless MD5 hash when using DSA keys in ca.
7426 [Anonymous <nobody@replay.com>]
7428 *) Generate an error if given an empty string as a cert directory. Also
7429 generate an error if handed NULL (previously returned 0 to indicate an
7430 error, but didn't set one).
7431 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7433 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7436 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7437 parameters. This was causing a warning which killed off the Win32 compile.
7440 *) Remove C++ style comments from crypto/bn/bn_local.h.
7441 [Neil Costigan <neil.costigan@celocom.com>]
7443 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7444 based on a text string, looking up short and long names and finally
7445 "dot" format. The "dot" format stuff didn't work. Added new function
7446 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7447 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7448 OID is not part of the table.
7451 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7452 X509_LOOKUP_by_alias().
7455 *) Sort openssl functions by name.
7458 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7459 encryption from sample DSA keys (in case anyone is interested the password
7463 *) Make _all_ *_free functions accept a NULL pointer.
7464 [Frans Heymans <fheymans@isaserver.be>]
7466 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7468 [Anonymous <nobody@replay.com>]
7470 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7471 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7473 *) Don't blow it for numeric -newkey arguments to apps/req.
7474 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7476 *) Temp key "for export" tests were wrong in s3_srvr.c.
7477 [Anonymous <nobody@replay.com>]
7479 *) Add prototype for temp key callback functions
7480 SSL_CTX_set_tmp_{rsa,dh}_callback().
7483 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7484 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7487 *) X509_name_add_entry() freed the wrong thing after an error.
7488 [Arne Ansper <arne@ats.cyber.ee>]
7490 *) rsa_eay.c would attempt to free a NULL context.
7491 [Arne Ansper <arne@ats.cyber.ee>]
7493 *) BIO_s_socket() had a broken should_retry() on Windoze.
7494 [Arne Ansper <arne@ats.cyber.ee>]
7496 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7497 [Arne Ansper <arne@ats.cyber.ee>]
7499 *) Make sure the already existing X509_STORE->depth variable is initialized
7500 in X509_STORE_new(), but document the fact that this variable is still
7501 unused in the certificate verification process.
7502 [Ralf S. Engelschall]
7504 *) Fix the various library and apps files to free up pkeys obtained from
7505 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7508 *) Fix reference counting in X509_PUBKEY_get(). This makes
7509 demos/maurice/example2.c work, amongst others, probably.
7510 [Steve Henson and Ben Laurie]
7512 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7513 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7514 are no longer created. This way we have a single and consistent command
7515 line interface `openssl <command>', similar to `cvs <command>'.
7516 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7518 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7519 BIT STRING wrapper always have zero unused bits.
7522 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7525 *) Make the top-level INSTALL documentation easier to understand.
7528 *) Makefiles updated to exit if an error occurs in a sub-directory
7529 make (including if user presses ^C) [Paul Sutton]
7531 *) Make Montgomery context stuff explicit in RSA data structure.
7534 *) Fix build order of pem and err to allow for generated pem.h.
7537 *) Fix renumbering bug in X509_NAME_delete_entry().
7540 *) Enhanced the err-ins.pl script so it makes the error library number
7541 global and can add a library name. This is needed for external ASN1 and
7542 other error libraries.
7545 *) Fixed sk_insert which never worked properly.
7548 *) Fix ASN1 macros so they can handle indefinite length construted
7549 EXPLICIT tags. Some non standard certificates use these: they can now
7553 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7554 into a single doc/ssleay.txt bundle. This way the information is still
7555 preserved but no longer messes up this directory. Now it's new room for
7556 the new set of documenation files.
7557 [Ralf S. Engelschall]
7559 *) SETs were incorrectly DER encoded. This was a major pain, because they
7560 shared code with SEQUENCEs, which aren't coded the same. This means that
7561 almost everything to do with SETs or SEQUENCEs has either changed name or
7562 number of arguments.
7563 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7565 *) Fix test data to work with the above.
7568 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7569 was already fixed by Eric for 0.9.1 it seems.
7570 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7572 *) Autodetect FreeBSD3.
7575 *) Fix various bugs in Configure. This affects the following platforms:
7579 unixware-2.0-pentium
7583 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7584 before they are needed.
7587 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7591 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7593 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7594 changed SSLeay to OpenSSL in version strings.
7595 [Ralf S. Engelschall]
7597 *) Some fixups to the top-level documents.
7600 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7601 because the symlink to include/ was missing.
7602 [Ralf S. Engelschall]
7604 *) Incorporated the popular no-RSA/DSA-only patches
7605 which allow to compile a RSA-free SSLeay.
7606 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7608 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7609 when "ssleay" is still not found.
7610 [Ralf S. Engelschall]
7612 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7613 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7615 *) Updated the README file.
7616 [Ralf S. Engelschall]
7618 *) Added various .cvsignore files in the CVS repository subdirs
7619 to make a "cvs update" really silent.
7620 [Ralf S. Engelschall]
7622 *) Recompiled the error-definition header files and added
7623 missing symbols to the Win32 linker tables.
7624 [Ralf S. Engelschall]
7626 *) Cleaned up the top-level documents;
7627 o new files: CHANGES and LICENSE
7628 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7629 o merged COPYRIGHT into LICENSE
7630 o removed obsolete TODO file
7631 o renamed MICROSOFT to INSTALL.W32
7632 [Ralf S. Engelschall]
7634 *) Removed dummy files from the 0.9.1b source tree:
7635 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7636 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7637 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7638 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7639 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7640 [Ralf S. Engelschall]
7642 *) Added various platform portability fixes.
7645 *) The Genesis of the OpenSSL rpject:
7646 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7647 Young and Tim J. Hudson created while they were working for C2Net until
7649 [The OpenSSL Project]
7652 Changes between 0.9.0b and 0.9.1b [not released]
7654 *) Updated a few CA certificates under certs/
7657 *) Changed some BIGNUM api stuff.
7660 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7661 DGUX x86, Linux Alpha, etc.
7664 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7665 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7669 *) Add -strparse option to asn1pars program which parses nested
7671 [Dr Stephen Henson <shenson@bigfoot.com>]
7673 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7676 *) DSA fix for "ca" program.
7679 *) Added "-genkey" option to "dsaparam" program.
7682 *) Added RIPE MD160 (rmd160) message digest.
7685 *) Added -a (all) option to "ssleay version" command.
7688 *) Added PLATFORM define which is the id given to Configure.
7691 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7694 *) Extended the ASN.1 parser routines.
7697 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7700 *) Added a BN_CTX to the BN library.
7703 *) Fixed the weak key values in DES library
7706 *) Changed API in EVP library for cipher aliases.
7709 *) Added support for RC2/64bit cipher.
7712 *) Converted the lhash library to the crypto/mem.c functions.
7715 *) Added more recognized ASN.1 object ids.
7718 *) Added more RSA padding checks for SSL/TLS.
7721 *) Added BIO proxy/filter functionality.
7724 *) Added extra_certs to SSL_CTX which can be used
7725 send extra CA certificates to the client in the CA cert chain sending
7726 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7729 *) Now Fortezza is denied in the authentication phase because
7730 this is key exchange mechanism is not supported by SSLeay at all.
7733 *) Additional PKCS1 checks.
7736 *) Support the string "TLSv1" for all TLS v1 ciphers.
7739 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7740 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7743 *) Fixed a few memory leaks.
7746 *) Fixed various code and comment typos.
7749 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7750 bytes sent in the client random.
7751 [Edward Bishop <ebishop@spyglass.com>]