5 Changes between 0.9.8e and 0.9.8f-fips [xx XXX xxxx]
7 *) Modify VC++ build system to rename .text and .rdata segments in
8 FIPS sources to .fipst${a,b,c}, and $fipsr${a,b,c} and place them
9 in a static library fipscanister.lib. These are then combined by the
10 VC++ linker into a single segment in suffix order but without the
11 suffix (i.e. .fipst and .fipsr). This serves the same purpose as
12 fipscanister.o on other platforms but has the advantage that it can
13 be created using only standard VC++ utilities.
16 *) Modify WIN32 build system to forward references functions implemented
20 *) Move error library so that all lhash dependencies are in a separate
21 file. Include a simplified ERR_get_state() function for stand alone
22 FIPS applications. Include a initialization function OPENSSL_init()
23 to set all callbacks, automatically call OPENSSL_init() once when
24 a cipher or digest is added. This should mean that almost all applications
25 set the callbacks automatically. Exceptional cases can call OPENSSL_init()
28 #ifdef OPENSSL_HAVE_INIT
31 before starting any threads.
34 *) Collect common functions into header file "fips_utl.h".
37 *) Only enable dynamic lock functionality in CRYPTO_lock() when it is really
38 needed. Move some lock functionality into new file dyn_lck.c .
39 This further reduces FIPS dependencies allowing the complete removal
40 of STACK and OBJ_bsearch().
43 *) Reduce FIPS test program dependencies by providing stand alone
44 versions of some existing functions in libcrypto. Avoid use
45 of BIOs by converting to system stdio. Move some functions in FIPS
46 files: e.g. all use of BIO_printf().
49 *) Modify build of libcrypto in FIPS mode by using a perl
50 script "arx.pl" which calls the archiver specifically
51 excluding any FIPS dependencies in libcrypto.
54 *) Port OpenSSL 0.9.7 FIPS code to 0.9.8. Convert to new
55 Makefile form. Update Configure. Convert and update
56 FIPS source files. Update libcrypto, libssl and apps
57 with additional functionality from 0.9.7 FIPS code.
58 Update Windows build system.
61 Changes between 0.9.8e and 0.9.8f [23 Feb 2007]
63 *) In the SSL/TLS server implementation, be strict about session ID
64 context matching (which matters if an application uses a single
65 external cache for different purposes). Previously,
66 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
67 set. This did ensure strict client verification, but meant that,
68 with applications using a single external cache for quite
69 different requirements, clients could circumvent ciphersuite
70 restrictions for a given session ID context by starting a session
71 in a different context.
74 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
75 a ciphersuite string such as "DEFAULT:RSA" cannot enable
76 authentication-only ciphersuites.
79 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
81 *) Since AES128 and AES256 (and similarly Camellia128 and
82 Camellia256) share a single mask bit in the logic of
83 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
84 kludge to work properly if AES128 is available and AES256 isn't
85 (or if Camellia128 is available and Camellia256 isn't).
88 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
89 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
90 When a point or a seed is encoded in a BIT STRING, we need to
91 prevent the removal of trailing zero bits to get the proper DER
92 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
93 of a NamedBitList, for which trailing 0 bits need to be removed.)
96 *) Have SSL/TLS server implementation tolerate "mismatched" record
97 protocol version while receiving ClientHello even if the
98 ClientHello is fragmented. (The server can't insist on the
99 particular protocol version it has chosen before the ServerHello
100 message has informed the client about his choice.)
103 *) Add RFC 3779 support.
104 [Rob Austein for ARIN, Ben Laurie]
106 *) Load error codes if they are not already present instead of using a
107 static variable. This allows them to be cleanly unloaded and reloaded.
108 Improve header file function name parsing.
111 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
112 or CAPABILITY handshake as required by RFCs.
115 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
117 *) Introduce limits to prevent malicious keys being able to
118 cause a denial of service. (CVE-2006-2940)
119 [Steve Henson, Bodo Moeller]
121 *) Fix ASN.1 parsing of certain invalid structures that can result
122 in a denial of service. (CVE-2006-2937) [Steve Henson]
124 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
125 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
127 *) Fix SSL client code which could crash if connecting to a
128 malicious SSLv2 server. (CVE-2006-4343)
129 [Tavis Ormandy and Will Drewry, Google Security Team]
131 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
132 match only those. Before that, "AES256-SHA" would be interpreted
133 as a pattern and match "AES128-SHA" too (since AES128-SHA got
134 the same strength classification in 0.9.7h) as we currently only
135 have a single AES bit in the ciphersuite description bitmap.
136 That change, however, also applied to ciphersuite strings such as
137 "RC4-MD5" that intentionally matched multiple ciphersuites --
138 namely, SSL 2.0 ciphersuites in addition to the more common ones
139 from SSL 3.0/TLS 1.0.
141 So we change the selection algorithm again: Naming an explicit
142 ciphersuite selects this one ciphersuite, and any other similar
143 ciphersuite (same bitmap) from *other* protocol versions.
144 Thus, "RC4-MD5" again will properly select both the SSL 2.0
145 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
147 Since SSL 2.0 does not have any ciphersuites for which the
148 128/256 bit distinction would be relevant, this works for now.
149 The proper fix will be to use different bits for AES128 and
150 AES256, which would have avoided the problems from the beginning;
151 however, bits are scarce, so we can only do this in a new release
152 (not just a patchlevel) when we can change the SSL_CIPHER
153 definition to split the single 'unsigned long mask' bitmap into
154 multiple values to extend the available space.
158 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
160 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
161 (CVE-2006-4339) [Ben Laurie and Google Security Team]
163 *) Add AES IGE and biIGE modes.
166 *) Change the Unix randomness entropy gathering to use poll() when
167 possible instead of select(), since the latter has some
168 undesirable limitations.
169 [Darryl Miles via Richard Levitte and Bodo Moeller]
171 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
172 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
173 cannot be implicitly activated as part of, e.g., the "AES" alias.
174 However, please upgrade to OpenSSL 0.9.9[-dev] for
175 non-experimental use of the ECC ciphersuites to get TLS extension
176 support, which is required for curve and point format negotiation
177 to avoid potential handshake problems.
180 *) Disable rogue ciphersuites:
182 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
183 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
184 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
186 The latter two were purportedly from
187 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
190 Also deactivate the remaining ciphersuites from
191 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
192 unofficial, and the ID has long expired.
195 *) Fix RSA blinding Heisenbug (problems sometimes occured on
196 dual-core machines) and other potential thread-safety issues.
199 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
200 versions), which is now available for royalty-free use
201 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
202 Also, add Camellia TLS ciphersuites from RFC 4132.
204 To minimize changes between patchlevels in the OpenSSL 0.9.8
205 series, Camellia remains excluded from compilation unless OpenSSL
206 is configured with 'enable-camellia'.
209 *) Disable the padding bug check when compression is in use. The padding
210 bug check assumes the first packet is of even length, this is not
211 necessarily true if compresssion is enabled and can result in false
212 positives causing handshake failure. The actual bug test is ancient
213 code so it is hoped that implementations will either have fixed it by
214 now or any which still have the bug do not support compression.
217 Changes between 0.9.8a and 0.9.8b [04 May 2006]
219 *) When applying a cipher rule check to see if string match is an explicit
220 cipher suite and only match that one cipher suite if it is.
223 *) Link in manifests for VC++ if needed.
224 [Austin Ziegler <halostatue@gmail.com>]
226 *) Update support for ECC-based TLS ciphersuites according to
227 draft-ietf-tls-ecc-12.txt with proposed changes (but without
228 TLS extensions, which are supported starting with the 0.9.9
229 branch, not in the OpenSSL 0.9.8 branch).
232 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
233 opaque EVP_CIPHER_CTX handling.
236 *) Fixes and enhancements to zlib compression code. We now only use
237 "zlib1.dll" and use the default __cdecl calling convention on Win32
238 to conform with the standards mentioned here:
239 http://www.zlib.net/DLL_FAQ.txt
240 Static zlib linking now works on Windows and the new --with-zlib-include
241 --with-zlib-lib options to Configure can be used to supply the location
242 of the headers and library. Gracefully handle case where zlib library
246 *) Several fixes and enhancements to the OID generation code. The old code
247 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
248 handle numbers larger than ULONG_MAX, truncated printing and had a
249 non standard OBJ_obj2txt() behaviour.
252 *) Add support for building of engines under engine/ as shared libraries
253 under VC++ build system.
256 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
257 Hopefully, we will not see any false combination of paths any more.
260 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
262 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
263 (part of SSL_OP_ALL). This option used to disable the
264 countermeasure against man-in-the-middle protocol-version
265 rollback in the SSL 2.0 server implementation, which is a bad
266 idea. (CVE-2005-2969)
268 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
269 for Information Security, National Institute of Advanced Industrial
270 Science and Technology [AIST], Japan)]
272 *) Add two function to clear and return the verify parameter flags.
275 *) Keep cipherlists sorted in the source instead of sorting them at
276 runtime, thus removing the need for a lock.
279 *) Avoid some small subgroup attacks in Diffie-Hellman.
280 [Nick Mathewson and Ben Laurie]
282 *) Add functions for well-known primes.
285 *) Extended Windows CE support.
286 [Satoshi Nakamura and Andy Polyakov]
288 *) Initialize SSL_METHOD structures at compile time instead of during
289 runtime, thus removing the need for a lock.
292 *) Make PKCS7_decrypt() work even if no certificate is supplied by
293 attempting to decrypt each encrypted key in turn. Add support to
297 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
299 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
302 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
305 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
306 key into the same file any more.
309 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
312 *) Add -utf8 command line and config file option to 'ca'.
313 [Stefan <stf@udoma.org]
315 *) Removed the macro des_crypt(), as it seems to conflict with some
316 libraries. Use DES_crypt().
319 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
320 involves renaming the source and generated shared-libs for
321 both. The engines will accept the corrected or legacy ids
322 ('ncipher' and '4758_cca' respectively) when binding. NB,
323 this only applies when building 'shared'.
324 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
326 *) Add attribute functions to EVP_PKEY structure. Modify
327 PKCS12_create() to recognize a CSP name attribute and
328 use it. Make -CSP option work again in pkcs12 utility.
331 *) Add new functionality to the bn blinding code:
332 - automatic re-creation of the BN_BLINDING parameters after
333 a fixed number of uses (currently 32)
334 - add new function for parameter creation
335 - introduce flags to control the update behaviour of the
336 BN_BLINDING parameters
337 - hide BN_BLINDING structure
338 Add a second BN_BLINDING slot to the RSA structure to improve
339 performance when a single RSA object is shared among several
343 *) Add support for DTLS.
344 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
346 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
347 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
350 *) Remove buggy and incompletet DH cert support from
351 ssl/ssl_rsa.c and ssl/s3_both.c
354 *) Use SHA-1 instead of MD5 as the default digest algorithm for
355 the apps/openssl applications.
358 *) Compile clean with "-Wall -Wmissing-prototypes
359 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
360 DEBUG_SAFESTACK must also be set.
363 *) Change ./Configure so that certain algorithms can be disabled by default.
364 The new counterpiece to "no-xxx" is "enable-xxx".
366 The patented RC5 and MDC2 algorithms will now be disabled unless
367 "enable-rc5" and "enable-mdc2", respectively, are specified.
369 (IDEA remains enabled despite being patented. This is because IDEA
370 is frequently required for interoperability, and there is no license
371 fee for non-commercial use. As before, "no-idea" can be used to
372 avoid this algorithm.)
376 *) Add processing of proxy certificates (see RFC 3820). This work was
377 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
378 EGEE (Enabling Grids for E-science in Europe).
381 *) RC4 performance overhaul on modern architectures/implementations, such
382 as Intel P4, IA-64 and AMD64.
385 *) New utility extract-section.pl. This can be used specify an alternative
386 section number in a pod file instead of having to treat each file as
387 a separate case in Makefile. This can be done by adding two lines to the
390 =for comment openssl_section:XXX
392 The blank line is mandatory.
396 *) New arguments -certform, -keyform and -pass for s_client and s_server
397 to allow alternative format key and certificate files and passphrase
401 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
402 update associated structures and add various utility functions.
404 Add new policy related verify parameters, include policy checking in
405 standard verify code. Enhance 'smime' application with extra parameters
406 to support policy checking and print out.
409 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
410 Nehemiah processors. These extensions support AES encryption in hardware
411 as well as RNG (though RNG support is currently disabled).
412 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
414 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
417 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
418 [Andy Polyakov and a number of other people]
420 *) Improved PowerPC platform support. Most notably BIGNUM assembler
421 implementation contributed by IBM.
422 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
424 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
425 exponent rather than 'unsigned long'. There is a corresponding change to
426 the new 'rsa_keygen' element of the RSA_METHOD structure.
427 [Jelte Jansen, Geoff Thorpe]
429 *) Functionality for creating the initial serial number file is now
430 moved from CA.pl to the 'ca' utility with a new option -create_serial.
432 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
433 number file to 1, which is bound to cause problems. To avoid
434 the problems while respecting compatibility between different 0.9.7
435 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
436 CA.pl for serial number initialization. With the new release 0.9.8,
437 we can fix the problem directly in the 'ca' utility.)
440 *) Reduced header interdepencies by declaring more opaque objects in
441 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
442 give fewer recursive includes, which could break lazy source code - so
443 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
444 developers should define this symbol when building and using openssl to
445 ensure they track the recommended behaviour, interfaces, [etc], but
446 backwards-compatible behaviour prevails when this isn't defined.
449 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
452 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
453 This will generate a random key of the appropriate length based on the
454 cipher context. The EVP_CIPHER can provide its own random key generation
455 routine to support keys of a specific form. This is used in the des and
456 3des routines to generate a key of the correct parity. Update S/MIME
457 code to use new functions and hence generate correct parity DES keys.
458 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
459 valid (weak or incorrect parity).
462 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
463 as looking them up. This is useful when the verified structure may contain
464 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
465 present unless the new PKCS7_NO_CRL flag is asserted.
468 *) Extend ASN1 oid configuration module. It now additionally accepts the
471 shortName = some long name, 1.2.3.4
474 *) Reimplemented the BN_CTX implementation. There is now no more static
475 limitation on the number of variables it can handle nor the depth of the
476 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
477 information can now expand as required, and rather than having a single
478 static array of bignums, BN_CTX now uses a linked-list of such arrays
479 allowing it to expand on demand whilst maintaining the usefulness of
483 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
484 to allow all RSA operations to function using a single BN_CTX.
487 *) Preliminary support for certificate policy evaluation and checking. This
488 is initially intended to pass the tests outlined in "Conformance Testing
489 of Relying Party Client Certificate Path Processing Logic" v1.07.
492 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
493 remained unused and not that useful. A variety of other little bignum
494 tweaks and fixes have also been made continuing on from the audit (see
498 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
499 associated ASN1, EVP and SSL functions and old ASN1 macros.
502 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
503 and this should never fail. So the return value from the use of
504 BN_set_word() (which can fail due to needless expansion) is now deprecated;
505 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
508 *) BN_CTX_get() should return zero-valued bignums, providing the same
509 initialised value as BN_new().
510 [Geoff Thorpe, suggested by Ulf Möller]
512 *) Support for inhibitAnyPolicy certificate extension.
515 *) An audit of the BIGNUM code is underway, for which debugging code is
516 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
517 is considered valid when processing BIGNUMs, and causes execution to
518 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
519 further steps are taken to deliberately pollute unused data in BIGNUM
520 structures to try and expose faulty code further on. For now, openssl will
521 (in its default mode of operation) continue to tolerate the inconsistent
522 forms that it has tolerated in the past, but authors and packagers should
523 consider trying openssl and their own applications when compiled with
524 these debugging symbols defined. It will help highlight potential bugs in
525 their own code, and will improve the test coverage for OpenSSL itself. At
526 some point, these tighter rules will become openssl's default to improve
527 maintainability, though the assert()s and other overheads will remain only
528 in debugging configurations. See bn.h for more details.
529 [Geoff Thorpe, Nils Larsch, Ulf Möller]
531 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
532 that can only be obtained through BN_CTX_new() (which implicitly
533 initialises it). The presence of this function only made it possible
534 to overwrite an existing structure (and cause memory leaks).
537 *) Because of the callback-based approach for implementing LHASH as a
538 template type, lh_insert() adds opaque objects to hash-tables and
539 lh_doall() or lh_doall_arg() are typically used with a destructor callback
540 to clean up those corresponding objects before destroying the hash table
541 (and losing the object pointers). So some over-zealous constifications in
542 LHASH have been relaxed so that lh_insert() does not take (nor store) the
543 objects as "const" and the lh_doall[_arg] callback wrappers are not
544 prototyped to have "const" restrictions on the object pointers they are
545 given (and so aren't required to cast them away any more).
548 *) The tmdiff.h API was so ugly and minimal that our own timing utility
549 (speed) prefers to use its own implementation. The two implementations
550 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
551 its object type properly exposed (MS_TM) instead of casting to/from "char
552 *". This may still change yet if someone realises MS_TM and "ms_time_***"
553 aren't necessarily the greatest nomenclatures - but this is what was used
554 internally to the implementation so I've used that for now.
557 *) Ensure that deprecated functions do not get compiled when
558 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
559 the self-tests were still using deprecated key-generation functions so
560 these have been updated also.
563 *) Reorganise PKCS#7 code to separate the digest location functionality
564 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
565 New function PKCS7_set_digest() to set the digest type for PKCS#7
566 digestedData type. Add additional code to correctly generate the
567 digestedData type and add support for this type in PKCS7 initialization
571 *) New function PKCS7_set0_type_other() this initializes a PKCS7
572 structure of type "other".
575 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
576 sure the loop does correctly stop and breaking ("division by zero")
577 modulus operations are not performed. The (pre-generated) prime
578 table crypto/bn/bn_prime.h was already correct, but it could not be
579 re-generated on some platforms because of the "division by zero"
580 situation in the script.
581 [Ralf S. Engelschall]
583 *) Update support for ECC-based TLS ciphersuites according to
584 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
585 SHA-1 now is only used for "small" curves (where the
586 representation of a field element takes up to 24 bytes); for
587 larger curves, the field element resulting from ECDH is directly
588 used as premaster secret.
589 [Douglas Stebila (Sun Microsystems Laboratories)]
591 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
592 curve secp160r1 to the tests.
593 [Douglas Stebila (Sun Microsystems Laboratories)]
595 *) Add the possibility to load symbols globally with DSO.
596 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
598 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
599 control of the error stack.
602 *) Add support for STORE in ENGINE.
605 *) Add the STORE type. The intention is to provide a common interface
606 to certificate and key stores, be they simple file-based stores, or
607 HSM-type store, or LDAP stores, or...
608 NOTE: The code is currently UNTESTED and isn't really used anywhere.
611 *) Add a generic structure called OPENSSL_ITEM. This can be used to
612 pass a list of arguments to any function as well as provide a way
613 for a function to pass data back to the caller.
616 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
617 works like BUF_strdup() but can be used to duplicate a portion of
618 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
622 *) Add the function sk_find_ex() which works like sk_find(), but will
623 return an index to an element even if an exact match couldn't be
624 found. The index is guaranteed to point at the element where the
625 searched-for key would be inserted to preserve sorting order.
628 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
629 takes an extra flags argument for optional functionality. Currently,
630 the following flags are defined:
632 OBJ_BSEARCH_VALUE_ON_NOMATCH
633 This one gets OBJ_bsearch_ex() to return a pointer to the first
634 element where the comparing function returns a negative or zero
637 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
638 This one gets OBJ_bsearch_ex() to return a pointer to the first
639 element where the comparing function returns zero. This is useful
640 if there are more than one element where the comparing function
644 *) Make it possible to create self-signed certificates with 'openssl ca'
645 in such a way that the self-signed certificate becomes part of the
646 CA database and uses the same mechanisms for serial number generation
647 as all other certificate signing. The new flag '-selfsign' enables
648 this functionality. Adapt CA.sh and CA.pl.in.
651 *) Add functionality to check the public key of a certificate request
652 against a given private. This is useful to check that a certificate
653 request can be signed by that key (self-signing).
656 *) Make it possible to have multiple active certificates with the same
657 subject in the CA index file. This is done only if the keyword
658 'unique_subject' is set to 'no' in the main CA section (default
659 if 'CA_default') of the configuration file. The value is saved
660 with the database itself in a separate index attribute file,
661 named like the index file with '.attr' appended to the name.
664 *) Generate muti valued AVAs using '+' notation in config files for
668 *) Support for nameConstraints certificate extension.
671 *) Support for policyConstraints certificate extension.
674 *) Support for policyMappings certificate extension.
677 *) Make sure the default DSA_METHOD implementation only uses its
678 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
679 and change its own handlers to be NULL so as to remove unnecessary
680 indirection. This lets alternative implementations fallback to the
681 default implementation more easily.
684 *) Support for directoryName in GeneralName related extensions
688 *) Make it possible to link applications using Makefile.shared.
689 Make that possible even when linking against static libraries!
692 *) Support for single pass processing for S/MIME signing. This now
693 means that S/MIME signing can be done from a pipe, in addition
694 cleartext signing (multipart/signed type) is effectively streaming
695 and the signed data does not need to be all held in memory.
697 This is done with a new flag PKCS7_STREAM. When this flag is set
698 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
699 is done after the data is output (and digests calculated) in
703 *) Add full support for -rpath/-R, both in shared libraries and
704 applications, at least on the platforms where it's known how
708 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
709 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
710 will now compute a table of multiples of the generator that
711 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
712 faster (notably in the case of a single point multiplication,
714 [Nils Larsch, Bodo Moeller]
716 *) IPv6 support for certificate extensions. The various extensions
717 which use the IP:a.b.c.d can now take IPv6 addresses using the
718 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
722 *) Added an ENGINE that implements RSA by performing private key
723 exponentiations with the GMP library. The conversions to and from
724 GMP's mpz_t format aren't optimised nor are any montgomery forms
725 cached, and on x86 it appears OpenSSL's own performance has caught up.
726 However there are likely to be other architectures where GMP could
727 provide a boost. This ENGINE is not built in by default, but it can be
728 specified at Configure time and should be accompanied by the necessary
729 linker additions, eg;
730 ./config -DOPENSSL_USE_GMP -lgmp
733 *) "openssl engine" will not display ENGINE/DSO load failure errors when
734 testing availability of engines with "-t" - the old behaviour is
735 produced by increasing the feature's verbosity with "-tt".
738 *) ECDSA routines: under certain error conditions uninitialized BN objects
739 could be freed. Solution: make sure initialization is performed early
740 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
744 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
745 and DH_METHOD (eg. by ENGINE implementations) to override the normal
746 software implementations. For DSA and DH, parameter generation can
747 also be overriden by providing the appropriate method callbacks.
750 *) Change the "progress" mechanism used in key-generation and
751 primality testing to functions that take a new BN_GENCB pointer in
752 place of callback/argument pairs. The new API functions have "_ex"
753 postfixes and the older functions are reimplemented as wrappers for
754 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
755 declarations of the old functions to help (graceful) attempts to
756 migrate to the new functions. Also, the new key-generation API
757 functions operate on a caller-supplied key-structure and return
758 success/failure rather than returning a key or NULL - this is to
759 help make "keygen" another member function of RSA_METHOD etc.
761 Example for using the new callback interface:
763 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
767 BN_GENCB_set(&my_cb, my_callback, my_arg);
769 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
770 /* For the meaning of a, b in calls to my_callback(), see the
771 * documentation of the function that calls the callback.
772 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
773 * my_callback should return 1 if it wants BN_is_prime_ex()
774 * to continue, or 0 to stop.
779 *) Change the ZLIB compression method to be stateful, and make it
780 available to TLS with the number defined in
781 draft-ietf-tls-compression-04.txt.
784 *) Add the ASN.1 structures and functions for CertificatePair, which
785 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
787 CertificatePair ::= SEQUENCE {
788 forward [0] Certificate OPTIONAL,
789 reverse [1] Certificate OPTIONAL,
790 -- at least one of the pair shall be present -- }
792 Also implement the PEM functions to read and write certificate
793 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
795 This needed to be defined, mostly for the sake of the LDAP
796 attribute crossCertificatePair, but may prove useful elsewhere as
800 *) Make it possible to inhibit symlinking of shared libraries in
801 Makefile.shared, for Cygwin's sake.
804 *) Extend the BIGNUM API by creating a function
805 void BN_set_negative(BIGNUM *a, int neg);
806 and a macro that behave like
807 int BN_is_negative(const BIGNUM *a);
809 to avoid the need to access 'a->neg' directly in applications.
812 *) Implement fast modular reduction for pseudo-Mersenne primes
813 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
814 EC_GROUP_new_curve_GFp() will now automatically use this
816 [Nils Larsch <nla@trustcenter.de>]
818 *) Add new lock type (CRYPTO_LOCK_BN).
821 *) Change the ENGINE framework to automatically load engines
822 dynamically from specific directories unless they could be
823 found to already be built in or loaded. Move all the
824 current engines except for the cryptodev one to a new
826 The engines in engines/ are built as shared libraries if
827 the "shared" options was given to ./Configure or ./config.
828 Otherwise, they are inserted in libcrypto.a.
829 /usr/local/ssl/engines is the default directory for dynamic
830 engines, but that can be overriden at configure time through
831 the usual use of --prefix and/or --openssldir, and at run
832 time with the environment variable OPENSSL_ENGINES.
833 [Geoff Thorpe and Richard Levitte]
835 *) Add Makefile.shared, a helper makefile to build shared
836 libraries. Addapt Makefile.org.
839 *) Add version info to Win32 DLLs.
840 [Peter 'Luna' Runestig" <peter@runestig.com>]
842 *) Add new 'medium level' PKCS#12 API. Certificates and keys
843 can be added using this API to created arbitrary PKCS#12
844 files while avoiding the low level API.
846 New options to PKCS12_create(), key or cert can be NULL and
847 will then be omitted from the output file. The encryption
848 algorithm NIDs can be set to -1 for no encryption, the mac
849 iteration count can be set to 0 to omit the mac.
851 Enhance pkcs12 utility by making the -nokeys and -nocerts
852 options work when creating a PKCS#12 file. New option -nomac
853 to omit the mac, NONE can be set for an encryption algorithm.
854 New code is modified to use the enhanced PKCS12_create()
855 instead of the low level API.
858 *) Extend ASN1 encoder to support indefinite length constructed
859 encoding. This can output sequences tags and octet strings in
860 this form. Modify pk7_asn1.c to support indefinite length
861 encoding. This is experimental and needs additional code to
862 be useful, such as an ASN1 bio and some enhanced streaming
865 Extend template encode functionality so that tagging is passed
866 down to the template encoder.
869 *) Let 'openssl req' fail if an argument to '-newkey' is not
870 recognized instead of using RSA as a default.
873 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
874 As these are not official, they are not included in "ALL";
875 the "ECCdraft" ciphersuite group alias can be used to select them.
876 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
878 *) Add ECDH engine support.
879 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
881 *) Add ECDH in new directory crypto/ecdh/.
882 [Douglas Stebila (Sun Microsystems Laboratories)]
884 *) Let BN_rand_range() abort with an error after 100 iterations
885 without success (which indicates a broken PRNG).
888 *) Change BN_mod_sqrt() so that it verifies that the input value
889 is really the square of the return value. (Previously,
890 BN_mod_sqrt would show GIGO behaviour.)
893 *) Add named elliptic curves over binary fields from X9.62, SECG,
894 and WAP/WTLS; add OIDs that were still missing.
896 [Sheueling Chang Shantz and Douglas Stebila
897 (Sun Microsystems Laboratories)]
899 *) Extend the EC library for elliptic curves over binary fields
900 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
903 EC_GF2m_simple_method
907 EC_GROUP_new_curve_GF2m
908 EC_GROUP_set_curve_GF2m
909 EC_GROUP_get_curve_GF2m
910 EC_POINT_set_affine_coordinates_GF2m
911 EC_POINT_get_affine_coordinates_GF2m
912 EC_POINT_set_compressed_coordinates_GF2m
914 Point compression for binary fields is disabled by default for
915 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
918 As binary polynomials are represented as BIGNUMs, various members
919 of the EC_GROUP and EC_POINT data structures can be shared
920 between the implementations for prime fields and binary fields;
921 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
922 are essentially identical to their ..._GFp counterparts.
923 (For simplicity, the '..._GFp' prefix has been dropped from
924 various internal method names.)
926 An internal 'field_div' method (similar to 'field_mul' and
927 'field_sqr') has been added; this is used only for binary fields.
929 [Sheueling Chang Shantz and Douglas Stebila
930 (Sun Microsystems Laboratories)]
932 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
933 through methods ('mul', 'precompute_mult').
935 The generic implementations (now internally called 'ec_wNAF_mul'
936 and 'ec_wNAF_precomputed_mult') remain the default if these
937 methods are undefined.
939 [Sheueling Chang Shantz and Douglas Stebila
940 (Sun Microsystems Laboratories)]
942 *) New function EC_GROUP_get_degree, which is defined through
943 EC_METHOD. For curves over prime fields, this returns the bit
944 length of the modulus.
946 [Sheueling Chang Shantz and Douglas Stebila
947 (Sun Microsystems Laboratories)]
949 *) New functions EC_GROUP_dup, EC_POINT_dup.
950 (These simply call ..._new and ..._copy).
952 [Sheueling Chang Shantz and Douglas Stebila
953 (Sun Microsystems Laboratories)]
955 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
956 Polynomials are represented as BIGNUMs (where the sign bit is not
957 used) in the following functions [macros]:
960 BN_GF2m_sub [= BN_GF2m_add]
961 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
962 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
963 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
965 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
966 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
967 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
968 BN_GF2m_cmp [= BN_ucmp]
970 (Note that only the 'mod' functions are actually for fields GF(2^m).
971 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
973 For some functions, an the irreducible polynomial defining a
974 field can be given as an 'unsigned int[]' with strictly
975 decreasing elements giving the indices of those bits that are set;
976 i.e., p[] represents the polynomial
977 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
979 p[0] > p[1] > ... > p[k] = 0.
980 This applies to the following functions:
985 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
986 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
989 BN_GF2m_mod_solve_quad_arr
993 Conversion can be performed by the following functions:
998 bntest.c has additional tests for binary polynomial arithmetic.
1000 Two implementations for BN_GF2m_mod_div() are available.
1001 The default algorithm simply uses BN_GF2m_mod_inv() and
1002 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1003 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1004 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1006 [Sheueling Chang Shantz and Douglas Stebila
1007 (Sun Microsystems Laboratories)]
1009 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1010 functionality is disabled at compile-time.
1011 [Douglas Stebila <douglas.stebila@sun.com>]
1013 *) Change default behaviour of 'openssl asn1parse' so that more
1014 information is visible when viewing, e.g., a certificate:
1016 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1017 mode the content of non-printable OCTET STRINGs is output in a
1018 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1019 avoid the appearance of a printable string.
1020 [Nils Larsch <nla@trustcenter.de>]
1022 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1024 EC_GROUP_set_asn1_flag()
1025 EC_GROUP_get_asn1_flag()
1026 EC_GROUP_set_point_conversion_form()
1027 EC_GROUP_get_point_conversion_form()
1028 These control ASN1 encoding details:
1029 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1030 has been set to OPENSSL_EC_NAMED_CURVE.
1031 - Points are encoded in uncompressed form by default; options for
1032 asn1_for are as for point2oct, namely
1033 POINT_CONVERSION_COMPRESSED
1034 POINT_CONVERSION_UNCOMPRESSED
1035 POINT_CONVERSION_HYBRID
1037 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1040 EC_GROUP_get0_seed()
1041 EC_GROUP_get_seed_len()
1042 This is used only for ASN1 purposes (so far).
1043 [Nils Larsch <nla@trustcenter.de>]
1045 *) Add 'field_type' member to EC_METHOD, which holds the NID
1046 of the appropriate field type OID. The new function
1047 EC_METHOD_get_field_type() returns this value.
1048 [Nils Larsch <nla@trustcenter.de>]
1053 EC_POINT_point2hex()
1054 EC_POINT_hex2point()
1055 providing useful interfaces to EC_POINT_point2oct() and
1056 EC_POINT_oct2point().
1057 [Nils Larsch <nla@trustcenter.de>]
1059 *) Change internals of the EC library so that the functions
1060 EC_GROUP_set_generator()
1061 EC_GROUP_get_generator()
1062 EC_GROUP_get_order()
1063 EC_GROUP_get_cofactor()
1064 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1065 to methods, which would lead to unnecessary code duplication when
1066 adding different types of curves.
1067 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1069 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1070 arithmetic, and such that modified wNAFs are generated
1071 (which avoid length expansion in many cases).
1074 *) Add a function EC_GROUP_check_discriminant() (defined via
1075 EC_METHOD) that verifies that the curve discriminant is non-zero.
1077 Add a function EC_GROUP_check() that makes some sanity tests
1078 on a EC_GROUP, its generator and order. This includes
1079 EC_GROUP_check_discriminant().
1080 [Nils Larsch <nla@trustcenter.de>]
1082 *) Add ECDSA in new directory crypto/ecdsa/.
1084 Add applications 'openssl ecparam' and 'openssl ecdsa'
1085 (these are based on 'openssl dsaparam' and 'openssl dsa').
1087 ECDSA support is also included in various other files across the
1088 library. Most notably,
1089 - 'openssl req' now has a '-newkey ecdsa:file' option;
1090 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1091 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1092 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1093 them suitable for ECDSA where domain parameters must be
1094 extracted before the specific public key;
1095 - ECDSA engine support has been added.
1096 [Nils Larsch <nla@trustcenter.de>]
1098 *) Include some named elliptic curves, and add OIDs from X9.62,
1099 SECG, and WAP/WTLS. Each curve can be obtained from the new
1101 EC_GROUP_new_by_curve_name(),
1102 and the list of available named curves can be obtained with
1103 EC_get_builtin_curves().
1104 Also add a 'curve_name' member to EC_GROUP objects, which can be
1106 EC_GROUP_set_curve_name()
1107 EC_GROUP_get_curve_name()
1108 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1110 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1111 was actually never needed) and in BN_mul(). The removal in BN_mul()
1112 required a small change in bn_mul_part_recursive() and the addition
1113 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1114 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1115 bn_sub_words() and bn_add_words() except they take arrays with
1119 Changes between 0.9.7m and 0.9.7n [xx XXX xxxx]
1121 *) In the SSL/TLS server implementation, be strict about session ID
1122 context matching (which matters if an application uses a single
1123 external cache for different purposes). Previously,
1124 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1125 set. This did ensure strict client verification, but meant that,
1126 with applications using a single external cache for quite
1127 different requirements, clients could circumvent ciphersuite
1128 restrictions for a given session ID context by starting a session
1129 in a different context.
1132 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1134 *) Cleanse PEM buffers before freeing them since they may contain
1136 [Benjamin Bennett <ben@psc.edu>]
1138 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1139 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1140 authentication-only ciphersuites.
1143 *) Since AES128 and AES256 share a single mask bit in the logic of
1144 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1145 kludge to work properly if AES128 is available and AES256 isn't.
1148 *) Expand security boundary to match 1.1.1 module.
1151 *) Remove redundant features: hash file source, editing of test vectors
1152 modify fipsld to use external fips_premain.c signature.
1155 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1156 run algorithm test programs.
1159 *) Make algorithm test programs more tolerant of whitespace.
1162 *) Have SSL/TLS server implementation tolerate "mismatched" record
1163 protocol version while receiving ClientHello even if the
1164 ClientHello is fragmented. (The server can't insist on the
1165 particular protocol version it has chosen before the ServerHello
1166 message has informed the client about his choice.)
1169 *) Load error codes if they are not already present instead of using a
1170 static variable. This allows them to be cleanly unloaded and reloaded.
1173 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1175 *) Introduce limits to prevent malicious keys being able to
1176 cause a denial of service. (CVE-2006-2940)
1177 [Steve Henson, Bodo Moeller]
1179 *) Fix ASN.1 parsing of certain invalid structures that can result
1180 in a denial of service. (CVE-2006-2937) [Steve Henson]
1182 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1183 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1185 *) Fix SSL client code which could crash if connecting to a
1186 malicious SSLv2 server. (CVE-2006-4343)
1187 [Tavis Ormandy and Will Drewry, Google Security Team]
1189 *) Change ciphersuite string processing so that an explicit
1190 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1191 will no longer include "AES128-SHA"), and any other similar
1192 ciphersuite (same bitmap) from *other* protocol versions (so that
1193 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1194 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1195 changes from 0.9.8b and 0.9.8d.
1198 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1200 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1201 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1203 *) Change the Unix randomness entropy gathering to use poll() when
1204 possible instead of select(), since the latter has some
1205 undesirable limitations.
1206 [Darryl Miles via Richard Levitte and Bodo Moeller]
1208 *) Disable rogue ciphersuites:
1210 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1211 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1212 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1214 The latter two were purportedly from
1215 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1218 Also deactive the remaining ciphersuites from
1219 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1220 unofficial, and the ID has long expired.
1223 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1224 dual-core machines) and other potential thread-safety issues.
1227 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1229 *) Adapt fipsld and the build system to link against the validated FIPS
1230 module in FIPS mode.
1233 *) Fixes for VC++ 2005 build under Windows.
1236 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1237 from a Windows bash shell such as MSYS. It is autodetected from the
1238 "config" script when run from a VC++ environment. Modify standard VC++
1239 build to use fipscanister.o from the GNU make build.
1242 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1244 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1245 The value now differs depending on if you build for FIPS or not.
1246 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1247 safely run with a non-FIPSed libcrypto, as it may crash because of
1248 the difference induced by this change.
1251 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1253 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1254 (part of SSL_OP_ALL). This option used to disable the
1255 countermeasure against man-in-the-middle protocol-version
1256 rollback in the SSL 2.0 server implementation, which is a bad
1257 idea. (CVE-2005-2969)
1259 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1260 for Information Security, National Institute of Advanced Industrial
1261 Science and Technology [AIST], Japan)]
1263 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1264 mainly for FIPS compliance and not fully integrated at this stage.
1267 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1268 the exponentiation using a fixed-length exponent. (Otherwise,
1269 the information leaked through timing could expose the secret key
1270 after many signatures; cf. Bleichenbacher's attack on DSA with
1274 *) Make a new fixed-window mod_exp implementation the default for
1275 RSA, DSA, and DH private-key operations so that the sequence of
1276 squares and multiplies and the memory access pattern are
1277 independent of the particular secret key. This will mitigate
1278 cache-timing and potential related attacks.
1280 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1281 and this is automatically used by BN_mod_exp_mont() if the new flag
1282 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1283 will use this BN flag for private exponents unless the flag
1284 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1285 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1287 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1289 *) Change the client implementation for SSLv23_method() and
1290 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1291 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1292 (Previously, the SSL 2.0 backwards compatible Client Hello
1293 message format would be used even with SSL_OP_NO_SSLv2.)
1296 *) Add support for smime-type MIME parameter in S/MIME messages which some
1300 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1301 a threadsafe manner. Modify rsa code to use new function and add calls
1302 to dsa and dh code (which had race conditions before).
1305 *) Include the fixed error library code in the C error file definitions
1306 instead of fixing them up at runtime. This keeps the error code
1307 structures constant.
1310 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1312 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1315 *) Fixes for newer kerberos headers. NB: the casts are needed because
1316 the 'length' field is signed on one version and unsigned on another
1317 with no (?) obvious way to tell the difference, without these VC++
1318 complains. Also the "definition" of FAR (blank) is no longer included
1319 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1320 some needed definitions.
1323 *) Undo Cygwin change.
1326 *) Added support for proxy certificates according to RFC 3820.
1327 Because they may be a security thread to unaware applications,
1328 they must be explicitely allowed in run-time. See
1329 docs/HOWTO/proxy_certificates.txt for further information.
1332 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1334 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1335 server and client random values. Previously
1336 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1337 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1339 This change has negligible security impact because:
1341 1. Server and client random values still have 24 bytes of pseudo random
1344 2. Server and client random values are sent in the clear in the initial
1347 3. The master secret is derived using the premaster secret (48 bytes in
1348 size for static RSA ciphersuites) as well as client server and random
1351 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1354 [Stephen Henson, reported by UK NISCC]
1356 *) Use Windows randomness collection on Cygwin.
1359 *) Fix hang in EGD/PRNGD query when communication socket is closed
1360 prematurely by EGD/PRNGD.
1361 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1363 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1366 *) Back-port of selected performance improvements from development
1367 branch, as well as improved support for PowerPC platforms.
1370 *) Add lots of checks for memory allocation failure, error codes to indicate
1371 failure and freeing up memory if a failure occurs.
1372 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1374 *) Add new -passin argument to dgst.
1377 *) Perform some character comparisons of different types in X509_NAME_cmp:
1378 this is needed for some certificates that reencode DNs into UTF8Strings
1379 (in violation of RFC3280) and can't or wont issue name rollover
1383 *) Make an explicit check during certificate validation to see that
1384 the CA setting in each certificate on the chain is correct. As a
1385 side effect always do the following basic checks on extensions,
1386 not just when there's an associated purpose to the check:
1388 - if there is an unhandled critical extension (unless the user
1389 has chosen to ignore this fault)
1390 - if the path length has been exceeded (if one is set at all)
1391 - that certain extensions fit the associated purpose (if one has
1395 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1397 *) Avoid a race condition when CRLs are checked in a multi threaded
1398 environment. This would happen due to the reordering of the revoked
1399 entries during signature checking and serial number lookup. Now the
1400 encoding is cached and the serial number sort performed under a lock.
1401 Add new STACK function sk_is_sorted().
1404 *) Add Delta CRL to the extension code.
1407 *) Various fixes to s3_pkt.c so alerts are sent properly.
1408 [David Holmes <d.holmes@f5.com>]
1410 *) Reduce the chances of duplicate issuer name and serial numbers (in
1411 violation of RFC3280) using the OpenSSL certificate creation utilities.
1412 This is done by creating a random 64 bit value for the initial serial
1413 number when a serial number file is created or when a self signed
1414 certificate is created using 'openssl req -x509'. The initial serial
1415 number file is created using 'openssl x509 -next_serial' in CA.pl
1416 rather than being initialized to 1.
1419 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1421 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1422 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1423 [Joe Orton, Steve Henson]
1425 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1427 [Joe Orton, Steve Henson]
1429 *) Make it possible to have multiple active certificates with the same
1430 subject in the CA index file. This is done only if the keyword
1431 'unique_subject' is set to 'no' in the main CA section (default
1432 if 'CA_default') of the configuration file. The value is saved
1433 with the database itself in a separate index attribute file,
1434 named like the index file with '.attr' appended to the name.
1437 *) X509 verify fixes. Disable broken certificate workarounds when
1438 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1439 keyUsage extension present. Don't accept CRLs with unhandled critical
1440 extensions: since verify currently doesn't process CRL extensions this
1441 rejects a CRL with *any* critical extensions. Add new verify error codes
1445 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1446 A clarification of RFC2560 will require the use of OCTET STRINGs and
1447 some implementations cannot handle the current raw format. Since OpenSSL
1448 copies and compares OCSP nonces as opaque blobs without any attempt at
1449 parsing them this should not create any compatibility issues.
1452 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1453 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1454 this HMAC (and other) operations are several times slower than OpenSSL
1458 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1459 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1461 *) Use the correct content when signing type "other".
1464 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1466 *) Fix various bugs revealed by running the NISCC test suite:
1468 Stop out of bounds reads in the ASN1 code when presented with
1469 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1471 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1473 If verify callback ignores invalid public key errors don't try to check
1474 certificate signature with the NULL public key.
1478 *) New -ignore_err option in ocsp application to stop the server
1479 exiting on the first error in a request.
1482 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1483 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1487 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1488 extra data after the compression methods not only for TLS 1.0
1489 but also for SSL 3.0 (as required by the specification).
1490 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1492 *) Change X509_certificate_type() to mark the key as exported/exportable
1493 when it's 512 *bits* long, not 512 bytes.
1496 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1497 blocks during encryption.
1500 *) Various fixes to base64 BIO and non blocking I/O. On write
1501 flushes were not handled properly if the BIO retried. On read
1502 data was not being buffered properly and had various logic bugs.
1503 This also affects blocking I/O when the data being decoded is a
1507 *) Various S/MIME bugfixes and compatibility changes:
1508 output correct application/pkcs7 MIME type if
1509 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1510 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1511 of files as .eml work). Correctly handle very long lines in MIME
1515 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1517 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1518 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1519 a protocol version number mismatch like a decryption error
1520 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1523 *) Turn on RSA blinding by default in the default implementation
1524 to avoid a timing attack. Applications that don't want it can call
1525 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1526 They would be ill-advised to do so in most cases.
1527 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1529 *) Change RSA blinding code so that it works when the PRNG is not
1530 seeded (in this case, the secret RSA exponent is abused as
1531 an unpredictable seed -- if it is not unpredictable, there
1532 is no point in blinding anyway). Make RSA blinding thread-safe
1533 by remembering the creator's thread ID in rsa->blinding and
1534 having all other threads use local one-time blinding factors
1535 (this requires more computation than sharing rsa->blinding, but
1536 avoids excessive locking; and if an RSA object is not shared
1537 between threads, blinding will still be very fast).
1540 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1541 ENGINE as defaults for all supported algorithms irrespective of
1542 the 'flags' parameter. 'flags' is now honoured, so applications
1543 should make sure they are passing it correctly.
1546 *) Target "mingw" now allows native Windows code to be generated in
1547 the Cygwin environment as well as with the MinGW compiler.
1550 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1552 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1553 via timing by performing a MAC computation even if incorrrect
1554 block cipher padding has been found. This is a countermeasure
1555 against active attacks where the attacker has to distinguish
1556 between bad padding and a MAC verification error. (CVE-2003-0078)
1558 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1559 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1560 Martin Vuagnoux (EPFL, Ilion)]
1562 *) Make the no-err option work as intended. The intention with no-err
1563 is not to have the whole error stack handling routines removed from
1564 libcrypto, it's only intended to remove all the function name and
1565 reason texts, thereby removing some of the footprint that may not
1566 be interesting if those errors aren't displayed anyway.
1568 NOTE: it's still possible for any application or module to have it's
1569 own set of error texts inserted. The routines are there, just not
1570 used by default when no-err is given.
1573 *) Add support for FreeBSD on IA64.
1574 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1576 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1577 Kerberos function mit_des_cbc_cksum(). Before this change,
1578 the value returned by DES_cbc_cksum() was like the one from
1579 mit_des_cbc_cksum(), except the bytes were swapped.
1580 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1582 *) Allow an application to disable the automatic SSL chain building.
1583 Before this a rather primitive chain build was always performed in
1584 ssl3_output_cert_chain(): an application had no way to send the
1585 correct chain if the automatic operation produced an incorrect result.
1587 Now the chain builder is disabled if either:
1589 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1591 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1593 The reasoning behind this is that an application would not want the
1594 auto chain building to take place if extra chain certificates are
1595 present and it might also want a means of sending no additional
1596 certificates (for example the chain has two certificates and the
1600 *) Add the possibility to build without the ENGINE framework.
1601 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1603 *) Under Win32 gmtime() can return NULL: check return value in
1604 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1607 *) DSA routines: under certain error conditions uninitialized BN objects
1608 could be freed. Solution: make sure initialization is performed early
1609 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1610 Nils Larsch <nla@trustcenter.de> via PR#459)
1613 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1614 checked on reconnect on the client side, therefore session resumption
1615 could still fail with a "ssl session id is different" error. This
1616 behaviour is masked when SSL_OP_ALL is used due to
1617 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1618 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1619 followup to PR #377.
1622 *) IA-32 assembler support enhancements: unified ELF targets, support
1623 for SCO/Caldera platforms, fix for Cygwin shared build.
1626 *) Add support for FreeBSD on sparc64. As a consequence, support for
1627 FreeBSD on non-x86 processors is separate from x86 processors on
1628 the config script, much like the NetBSD support.
1629 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1631 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1633 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1636 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1637 code (06) was taken as the first octet of the session ID and the last
1638 octet was ignored consequently. As a result SSLv2 client side session
1639 caching could not have worked due to the session ID mismatch between
1641 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1645 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1646 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1650 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1651 seems that in spite of existing for more than a year, many application
1652 author have done nothing to provide the necessary callbacks, which
1653 means that this particular engine will not work properly anywhere.
1654 This is a very unfortunate situation which forces us, in the name
1655 of usability, to give the hw_ncipher.c a static lock, which is part
1657 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1658 appear in 0.9.8 or later. We EXPECT application authors to have
1659 dealt properly with this when 0.9.8 is released (unless we actually
1660 make such changes in the libcrypto locking code that changes will
1661 have to be made anyway).
1664 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1665 octets have been read, EOF or an error occurs. Without this change
1666 some truncated ASN1 structures will not produce an error.
1669 *) Disable Heimdal support, since it hasn't been fully implemented.
1670 Still give the possibility to force the use of Heimdal, but with
1671 warnings and a request that patches get sent to openssl-dev.
1674 *) Add the VC-CE target, introduce the WINCE sysname, and add
1675 INSTALL.WCE and appropriate conditionals to make it build.
1676 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1678 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1679 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1680 edit numbers of the version.
1681 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1683 *) Introduce safe string copy and catenation functions
1684 (BUF_strlcpy() and BUF_strlcat()).
1685 [Ben Laurie (CHATS) and Richard Levitte]
1687 *) Avoid using fixed-size buffers for one-line DNs.
1688 [Ben Laurie (CHATS)]
1690 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1691 resizing buffers containing secrets, and use where appropriate.
1692 [Ben Laurie (CHATS)]
1694 *) Avoid using fixed size buffers for configuration file location.
1695 [Ben Laurie (CHATS)]
1697 *) Avoid filename truncation for various CA files.
1698 [Ben Laurie (CHATS)]
1700 *) Use sizeof in preference to magic numbers.
1701 [Ben Laurie (CHATS)]
1703 *) Avoid filename truncation in cert requests.
1704 [Ben Laurie (CHATS)]
1706 *) Add assertions to check for (supposedly impossible) buffer
1708 [Ben Laurie (CHATS)]
1710 *) Don't cache truncated DNS entries in the local cache (this could
1711 potentially lead to a spoofing attack).
1712 [Ben Laurie (CHATS)]
1714 *) Fix various buffers to be large enough for hex/decimal
1715 representations in a platform independent manner.
1716 [Ben Laurie (CHATS)]
1718 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1719 resizing buffers containing secrets, and use where appropriate.
1720 [Ben Laurie (CHATS)]
1722 *) Add BIO_indent() to avoid much slightly worrying code to do
1724 [Ben Laurie (CHATS)]
1726 *) Convert sprintf()/BIO_puts() to BIO_printf().
1727 [Ben Laurie (CHATS)]
1729 *) buffer_gets() could terminate with the buffer only half
1731 [Ben Laurie (CHATS)]
1733 *) Add assertions to prevent user-supplied crypto functions from
1734 overflowing internal buffers by having large block sizes, etc.
1735 [Ben Laurie (CHATS)]
1737 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1739 [Ben Laurie (CHATS)]
1741 *) Eliminate unused copy of key in RC4.
1742 [Ben Laurie (CHATS)]
1744 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1745 [Ben Laurie (CHATS)]
1747 *) Fix off-by-one error in EGD path.
1748 [Ben Laurie (CHATS)]
1750 *) If RANDFILE path is too long, ignore instead of truncating.
1751 [Ben Laurie (CHATS)]
1753 *) Eliminate unused and incorrectly sized X.509 structure
1755 [Ben Laurie (CHATS)]
1757 *) Eliminate unused and dangerous function knumber().
1758 [Ben Laurie (CHATS)]
1760 *) Eliminate unused and dangerous structure, KSSL_ERR.
1761 [Ben Laurie (CHATS)]
1763 *) Protect against overlong session ID context length in an encoded
1764 session object. Since these are local, this does not appear to be
1766 [Ben Laurie (CHATS)]
1768 *) Change from security patch (see 0.9.6e below) that did not affect
1769 the 0.9.6 release series:
1771 Remote buffer overflow in SSL3 protocol - an attacker could
1772 supply an oversized master key in Kerberos-enabled versions.
1774 [Ben Laurie (CHATS)]
1776 *) Change the SSL kerb5 codes to match RFC 2712.
1779 *) Make -nameopt work fully for req and add -reqopt switch.
1780 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1782 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1783 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1785 *) Make sure tests can be performed even if the corresponding algorithms
1786 have been removed entirely. This was also the last step to make
1787 OpenSSL compilable with DJGPP under all reasonable conditions.
1788 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1790 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1791 to allow version independent disabling of normally unselected ciphers,
1792 which may be activated as a side-effect of selecting a single cipher.
1794 (E.g., cipher list string "RSA" enables ciphersuites that are left
1795 out of "ALL" because they do not provide symmetric encryption.
1796 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1797 [Lutz Jaenicke, Bodo Moeller]
1799 *) Add appropriate support for separate platform-dependent build
1800 directories. The recommended way to make a platform-dependent
1801 build directory is the following (tested on Linux), maybe with
1804 # Place yourself outside of the OpenSSL source tree. In
1805 # this example, the environment variable OPENSSL_SOURCE
1806 # is assumed to contain the absolute OpenSSL source directory.
1807 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1808 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1809 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1810 mkdir -p `dirname $F`
1811 ln -s $OPENSSL_SOURCE/$F $F
1814 To be absolutely sure not to disturb the source tree, a "make clean"
1815 is a good thing. If it isn't successfull, don't worry about it,
1816 it probably means the source directory is very clean.
1819 *) Make sure any ENGINE control commands make local copies of string
1820 pointers passed to them whenever necessary. Otherwise it is possible
1821 the caller may have overwritten (or deallocated) the original string
1822 data when a later ENGINE operation tries to use the stored values.
1823 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1825 *) Improve diagnostics in file reading and command-line digests.
1826 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1828 *) Add AES modes CFB and OFB to the object database. Correct an
1829 error in AES-CFB decryption.
1832 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1833 allows existing EVP_CIPHER_CTX structures to be reused after
1834 calling EVP_*Final(). This behaviour is used by encryption
1835 BIOs and some applications. This has the side effect that
1836 applications must explicitly clean up cipher contexts with
1837 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1840 *) Check the values of dna and dnb in bn_mul_recursive before calling
1841 bn_mul_comba (a non zero value means the a or b arrays do not contain
1842 n2 elements) and fallback to bn_mul_normal if either is not zero.
1845 *) Fix escaping of non-ASCII characters when using the -subj option
1846 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1849 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1850 form for "surname", serialNumber has no short form.
1851 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1852 therefore remove "mail" short name for "internet 7".
1853 The OID for unique identifiers in X509 certificates is
1854 x500UniqueIdentifier, not uniqueIdentifier.
1855 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1858 *) Add an "init" command to the ENGINE config module and auto initialize
1859 ENGINEs. Without any "init" command the ENGINE will be initialized
1860 after all ctrl commands have been executed on it. If init=1 the
1861 ENGINE is initailized at that point (ctrls before that point are run
1862 on the uninitialized ENGINE and after on the initialized one). If
1863 init=0 then the ENGINE will not be iniatialized at all.
1866 *) Fix the 'app_verify_callback' interface so that the user-defined
1867 argument is actually passed to the callback: In the
1868 SSL_CTX_set_cert_verify_callback() prototype, the callback
1869 declaration has been changed from
1872 int (*cb)(X509_STORE_CTX *,void *);
1873 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1874 i=s->ctx->app_verify_callback(&ctx)
1875 has been changed into
1876 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1878 To update applications using SSL_CTX_set_cert_verify_callback(),
1879 a dummy argument can be added to their callback functions.
1880 [D. K. Smetters <smetters@parc.xerox.com>]
1882 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1883 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1885 *) Add and OPENSSL_LOAD_CONF define which will cause
1886 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1887 This allows older applications to transparently support certain
1888 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1889 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1890 load the config file and OPENSSL_add_all_algorithms_conf() which will
1891 always load it have also been added.
1894 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1895 Adjust NIDs and EVP layer.
1896 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1898 *) Config modules support in openssl utility.
1900 Most commands now load modules from the config file,
1901 though in a few (such as version) this isn't done
1902 because it couldn't be used for anything.
1904 In the case of ca and req the config file used is
1905 the same as the utility itself: that is the -config
1906 command line option can be used to specify an
1910 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1911 use "openssl_conf" if filename is NULL use default openssl config file.
1914 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1915 config section name. Add a new flag to tolerate a missing config file
1916 and move code to CONF_modules_load_file().
1919 *) Support for crypto accelerator cards from Accelerated Encryption
1920 Processing, www.aep.ie. (Use engine 'aep')
1921 The support was copied from 0.9.6c [engine] and adapted/corrected
1922 to work with the new engine framework.
1923 [AEP Inc. and Richard Levitte]
1925 *) Support for SureWare crypto accelerator cards from Baltimore
1926 Technologies. (Use engine 'sureware')
1927 The support was copied from 0.9.6c [engine] and adapted
1928 to work with the new engine framework.
1931 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1932 make the newer ENGINE framework commands for the CHIL engine work.
1933 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1935 *) Make it possible to produce shared libraries on ReliantUNIX.
1936 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1938 *) Add the configuration target debug-linux-ppro.
1939 Make 'openssl rsa' use the general key loading routines
1940 implemented in apps.c, and make those routines able to
1941 handle the key format FORMAT_NETSCAPE and the variant
1943 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1945 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1946 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1948 *) Add -keyform to rsautl, and document -engine.
1949 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1951 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1952 BIO_R_NO_SUCH_FILE error code rather than the generic
1953 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1956 *) Add new functions
1958 ERR_peek_last_error_line
1959 ERR_peek_last_error_line_data.
1960 These are similar to
1963 ERR_peek_error_line_data,
1964 but report on the latest error recorded rather than the first one
1965 still in the error queue.
1966 [Ben Laurie, Bodo Moeller]
1968 *) default_algorithms option in ENGINE config module. This allows things
1970 default_algorithms = ALL
1971 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1974 *) Prelminary ENGINE config module.
1977 *) New experimental application configuration code.
1980 *) Change the AES code to follow the same name structure as all other
1981 symmetric ciphers, and behave the same way. Move everything to
1982 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1983 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1985 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1986 [Ben Laurie and Theo de Raadt]
1988 *) Add option to output public keys in req command.
1989 [Massimiliano Pala madwolf@openca.org]
1991 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1992 (up to about 10% better than before for P-192 and P-224).
1995 *) New functions/macros
1997 SSL_CTX_set_msg_callback(ctx, cb)
1998 SSL_CTX_set_msg_callback_arg(ctx, arg)
1999 SSL_set_msg_callback(ssl, cb)
2000 SSL_set_msg_callback_arg(ssl, arg)
2002 to request calling a callback function
2004 void cb(int write_p, int version, int content_type,
2005 const void *buf, size_t len, SSL *ssl, void *arg)
2007 whenever a protocol message has been completely received
2008 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2009 protocol version according to which the SSL library interprets
2010 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2011 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2012 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2013 specification (change_cipher_spec(20), alert(21), handshake(22)).
2014 'buf' and 'len' point to the actual message, 'ssl' to the
2015 SSL object, and 'arg' is the application-defined value set by
2016 SSL[_CTX]_set_msg_callback_arg().
2018 'openssl s_client' and 'openssl s_server' have new '-msg' options
2019 to enable a callback that displays all protocol messages.
2022 *) Change the shared library support so shared libraries are built as
2023 soon as the corresponding static library is finished, and thereby get
2024 openssl and the test programs linked against the shared library.
2025 This still only happens when the keyword "shard" has been given to
2026 the configuration scripts.
2028 NOTE: shared library support is still an experimental thing, and
2029 backward binary compatibility is still not guaranteed.
2030 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2032 *) Add support for Subject Information Access extension.
2033 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2035 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2036 additional bytes when new memory had to be allocated, not just
2037 when reusing an existing buffer.
2040 *) New command line and configuration option 'utf8' for the req command.
2041 This allows field values to be specified as UTF8 strings.
2044 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2045 runs for the former and machine-readable output for the latter.
2048 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2049 of the e-mail address in the DN (i.e., it will go into a certificate
2050 extension only). The new configuration file option 'email_in_dn = no'
2051 has the same effect.
2052 [Massimiliano Pala madwolf@openca.org]
2054 *) Change all functions with names starting with des_ to be starting
2055 with DES_ instead. Add wrappers that are compatible with libdes,
2056 but are named _ossl_old_des_*. Finally, add macros that map the
2057 des_* symbols to the corresponding _ossl_old_des_* if libdes
2058 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2059 desired, the des_* symbols will be mapped to DES_*, with one
2062 Since we provide two compatibility mappings, the user needs to
2063 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2064 compatibility is desired. The default (i.e., when that macro
2065 isn't defined) is OpenSSL 0.9.6c compatibility.
2067 There are also macros that enable and disable the support of old
2068 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2069 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2070 are defined, the default will apply: to support the old des routines.
2072 In either case, one must include openssl/des.h to get the correct
2073 definitions. Do not try to just include openssl/des_old.h, that
2076 NOTE: This is a major break of an old API into a new one. Software
2077 authors are encouraged to switch to the DES_ style functions. Some
2078 time in the future, des_old.h and the libdes compatibility functions
2079 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2080 default), and then completely removed.
2083 *) Test for certificates which contain unsupported critical extensions.
2084 If such a certificate is found during a verify operation it is
2085 rejected by default: this behaviour can be overridden by either
2086 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2087 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2088 X509_supported_extension() has also been added which returns 1 if a
2089 particular extension is supported.
2092 *) Modify the behaviour of EVP cipher functions in similar way to digests
2093 to retain compatibility with existing code.
2096 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2097 compatibility with existing code. In particular the 'ctx' parameter does
2098 not have to be to be initialized before the call to EVP_DigestInit() and
2099 it is tidied up after a call to EVP_DigestFinal(). New function
2100 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2101 EVP_MD_CTX_copy() changed to not require the destination to be
2102 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2103 requires the destination to be valid.
2105 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2106 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2109 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2110 so that complete 'Handshake' protocol structures are kept in memory
2111 instead of overwriting 'msg_type' and 'length' with 'body' data.
2114 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2115 [Massimo Santin via Richard Levitte]
2117 *) Major restructuring to the underlying ENGINE code. This includes
2118 reduction of linker bloat, separation of pure "ENGINE" manipulation
2119 (initialisation, etc) from functionality dealing with implementations
2120 of specific crypto iterfaces. This change also introduces integrated
2121 support for symmetric ciphers and digest implementations - so ENGINEs
2122 can now accelerate these by providing EVP_CIPHER and EVP_MD
2123 implementations of their own. This is detailed in crypto/engine/README
2124 as it couldn't be adequately described here. However, there are a few
2125 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2126 were changed in the original introduction of ENGINE code have now
2127 reverted back - the hooking from this code to ENGINE is now a good
2128 deal more passive and at run-time, operations deal directly with
2129 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2130 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2131 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2132 they were not being used by the framework as there is no concept of a
2133 BIGNUM_METHOD and they could not be generalised to the new
2134 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2135 ENGINE_cpy() has been removed as it cannot be consistently defined in
2139 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2142 *) Change mkdef.pl to sort symbols that get the same entry number,
2143 and make sure the automatically generated functions ERR_load_*
2144 become part of libeay.num as well.
2147 *) New function SSL_renegotiate_pending(). This returns true once
2148 renegotiation has been requested (either SSL_renegotiate() call
2149 or HelloRequest/ClientHello receveived from the peer) and becomes
2150 false once a handshake has been completed.
2151 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2152 sends a HelloRequest, but does not ensure that a handshake takes
2153 place. SSL_renegotiate_pending() is useful for checking if the
2154 client has followed the request.)
2157 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2158 By default, clients may request session resumption even during
2159 renegotiation (if session ID contexts permit); with this option,
2160 session resumption is possible only in the first handshake.
2162 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2163 more bits available for options that should not be part of
2164 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2167 *) Add some demos for certificate and certificate request creation.
2170 *) Make maximum certificate chain size accepted from the peer application
2171 settable (SSL*_get/set_max_cert_list()), as proposed by
2172 "Douglas E. Engert" <deengert@anl.gov>.
2175 *) Add support for shared libraries for Unixware-7
2176 (Boyd Lynn Gerber <gerberb@zenez.com>).
2179 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2180 be done prior to destruction. Use this to unload error strings from
2181 ENGINEs that load their own error strings. NB: This adds two new API
2182 functions to "get" and "set" this destroy handler in an ENGINE.
2185 *) Alter all existing ENGINE implementations (except "openssl" and
2186 "openbsd") to dynamically instantiate their own error strings. This
2187 makes them more flexible to be built both as statically-linked ENGINEs
2188 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2189 Also, add stub code to each that makes building them as self-contained
2190 shared-libraries easier (see README.ENGINE).
2193 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2194 implementations into applications that are completely implemented in
2195 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2196 commands that can be used to configure what shared-library to load and
2197 to control aspects of the way it is handled. Also, made an update to
2198 the README.ENGINE file that brings its information up-to-date and
2199 provides some information and instructions on the "dynamic" ENGINE
2200 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2203 *) Make it possible to unload ranges of ERR strings with a new
2204 "ERR_unload_strings" function.
2207 *) Add a copy() function to EVP_MD.
2210 *) Make EVP_MD routines take a context pointer instead of just the
2211 md_data void pointer.
2214 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2215 that the digest can only process a single chunk of data
2216 (typically because it is provided by a piece of
2217 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2218 is only going to provide a single chunk of data, and hence the
2219 framework needn't accumulate the data for oneshot drivers.
2222 *) As with "ERR", make it possible to replace the underlying "ex_data"
2223 functions. This change also alters the storage and management of global
2224 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2225 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2226 index counters. The API functions that use this state have been changed
2227 to take a "class_index" rather than pointers to the class's local STACK
2228 and counter, and there is now an API function to dynamically create new
2229 classes. This centralisation allows us to (a) plug a lot of the
2230 thread-safety problems that existed, and (b) makes it possible to clean
2231 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2232 such data would previously have always leaked in application code and
2233 workarounds were in place to make the memory debugging turn a blind eye
2234 to it. Application code that doesn't use this new function will still
2235 leak as before, but their memory debugging output will announce it now
2236 rather than letting it slide.
2238 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2239 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2240 has a return value to indicate success or failure.
2243 *) Make it possible to replace the underlying "ERR" functions such that the
2244 global state (2 LHASH tables and 2 locks) is only used by the "default"
2245 implementation. This change also adds two functions to "get" and "set"
2246 the implementation prior to it being automatically set the first time
2247 any other ERR function takes place. Ie. an application can call "get",
2248 pass the return value to a module it has just loaded, and that module
2249 can call its own "set" function using that value. This means the
2250 module's "ERR" operations will use (and modify) the error state in the
2251 application and not in its own statically linked copy of OpenSSL code.
2254 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2255 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2256 the operation, and provides a more encapsulated way for external code
2257 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2258 to use these functions rather than manually incrementing the counts.
2260 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2263 *) Add EVP test program.
2266 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2269 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2270 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2271 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2272 These allow a CRL to be built without having to access X509_CRL fields
2273 directly. Modify 'ca' application to use new functions.
2276 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2277 bug workarounds. Rollback attack detection is a security feature.
2278 The problem will only arise on OpenSSL servers when TLSv1 is not
2279 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2280 Software authors not wanting to support TLSv1 will have special reasons
2281 for their choice and can explicitly enable this option.
2282 [Bodo Moeller, Lutz Jaenicke]
2284 *) Rationalise EVP so it can be extended: don't include a union of
2285 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2286 (similar to those existing for EVP_CIPHER_CTX).
2291 EVP_MD_CTX_init(&md); /* new function call */
2292 EVP_DigestInit(&md, EVP_sha1());
2293 EVP_DigestUpdate(&md, in, len);
2294 EVP_DigestFinal(&md, out, NULL);
2295 EVP_MD_CTX_cleanup(&md); /* new function call */
2299 *) Make DES key schedule conform to the usual scheme, as well as
2300 correcting its structure. This means that calls to DES functions
2301 now have to pass a pointer to a des_key_schedule instead of a
2302 plain des_key_schedule (which was actually always a pointer
2305 des_key_schedule ks;
2307 des_set_key_checked(..., &ks);
2308 des_ncbc_encrypt(..., &ks, ...);
2310 (Note that a later change renames 'des_...' into 'DES_...'.)
2313 *) Initial reduction of linker bloat: the use of some functions, such as
2314 PEM causes large amounts of unused functions to be linked in due to
2315 poor organisation. For example pem_all.c contains every PEM function
2316 which has a knock on effect of linking in large amounts of (unused)
2317 ASN1 code. Grouping together similar functions and splitting unrelated
2318 functions prevents this.
2321 *) Cleanup of EVP macros.
2324 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2325 correct _ecb suffix.
2328 *) Add initial OCSP responder support to ocsp application. The
2329 revocation information is handled using the text based index
2330 use by the ca application. The responder can either handle
2331 requests generated internally, supplied in files (for example
2332 via a CGI script) or using an internal minimal server.
2335 *) Add configuration choices to get zlib compression for TLS.
2338 *) Changes to Kerberos SSL for RFC 2712 compliance:
2339 1. Implemented real KerberosWrapper, instead of just using
2340 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2341 2. Implemented optional authenticator field of KerberosWrapper.
2343 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2344 and authenticator structs; see crypto/krb5/.
2346 Generalized Kerberos calls to support multiple Kerberos libraries.
2347 [Vern Staats <staatsvr@asc.hpc.mil>,
2348 Jeffrey Altman <jaltman@columbia.edu>
2349 via Richard Levitte]
2351 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2352 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2353 values for each of the key sizes rather than having just
2354 parameters (and 'speed' generating keys each time).
2357 *) Speed up EVP routines.
2360 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2361 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2362 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2363 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2365 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2366 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2367 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2370 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2372 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2375 *) Added the OS2-EMX target.
2376 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2378 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2379 to support NCONF routines in extension code. New function CONF_set_nconf()
2380 to allow functions which take an NCONF to also handle the old LHASH
2381 structure: this means that the old CONF compatible routines can be
2382 retained (in particular wrt extensions) without having to duplicate the
2383 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2386 *) Enhance the general user interface with mechanisms for inner control
2387 and with possibilities to have yes/no kind of prompts.
2390 *) Change all calls to low level digest routines in the library and
2391 applications to use EVP. Add missing calls to HMAC_cleanup() and
2392 don't assume HMAC_CTX can be copied using memcpy().
2393 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2395 *) Add the possibility to control engines through control names but with
2396 arbitrary arguments instead of just a string.
2397 Change the key loaders to take a UI_METHOD instead of a callback
2398 function pointer. NOTE: this breaks binary compatibility with earlier
2399 versions of OpenSSL [engine].
2400 Adapt the nCipher code for these new conditions and add a card insertion
2404 *) Enhance the general user interface with mechanisms to better support
2405 dialog box interfaces, application-defined prompts, the possibility
2406 to use defaults (for example default passwords from somewhere else)
2407 and interrupts/cancellations.
2410 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2411 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2414 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2415 tidy up some unnecessarily weird code in 'sk_new()').
2416 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2418 *) Change the key loading routines for ENGINEs to use the same kind
2419 callback (pem_password_cb) as all other routines that need this
2423 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2424 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2425 than this minimum value is recommended.
2428 *) New random seeder for OpenVMS, using the system process statistics
2429 that are easily reachable.
2432 *) Windows apparently can't transparently handle global
2433 variables defined in DLLs. Initialisations such as:
2435 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2437 wont compile. This is used by the any applications that need to
2438 declare their own ASN1 modules. This was fixed by adding the option
2439 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2440 needed for static libraries under Win32.
2443 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2444 setting of purpose and trust fields. New X509_STORE trust and
2445 purpose functions and tidy up setting in other SSL functions.
2448 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2449 structure. These are inherited by X509_STORE_CTX when it is
2450 initialised. This allows various defaults to be set in the
2451 X509_STORE structure (such as flags for CRL checking and custom
2452 purpose or trust settings) for functions which only use X509_STORE_CTX
2453 internally such as S/MIME.
2455 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2456 trust settings if they are not set in X509_STORE. This allows X509_STORE
2457 purposes and trust (in S/MIME for example) to override any set by default.
2459 Add command line options for CRL checking to smime, s_client and s_server
2463 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2464 are set then the CRL is looked up in the X509_STORE structure and
2465 its validity and signature checked, then if the certificate is found
2466 in the CRL the verify fails with a revoked error.
2468 Various new CRL related callbacks added to X509_STORE_CTX structure.
2470 Command line options added to 'verify' application to support this.
2472 This needs some additional work, such as being able to handle multiple
2473 CRLs with different times, extension based lookup (rather than just
2474 by subject name) and ultimately more complete V2 CRL extension
2478 *) Add a general user interface API (crypto/ui/). This is designed
2479 to replace things like des_read_password and friends (backward
2480 compatibility functions using this new API are provided).
2481 The purpose is to remove prompting functions from the DES code
2482 section as well as provide for prompting through dialog boxes in
2483 a window system and the like.
2486 *) Add "ex_data" support to ENGINE so implementations can add state at a
2487 per-structure level rather than having to store it globally.
2490 *) Make it possible for ENGINE structures to be copied when retrieved by
2491 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2492 This causes the "original" ENGINE structure to act like a template,
2493 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2494 operational state can be localised to each ENGINE structure, despite the
2495 fact they all share the same "methods". New ENGINE structures returned in
2496 this case have no functional references and the return value is the single
2497 structural reference. This matches the single structural reference returned
2498 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2502 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2503 needs to match any other type at all we need to manually clear the
2507 *) Changes to the "openssl engine" utility to include;
2508 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2509 about an ENGINE's available control commands.
2510 - executing control commands from command line arguments using the
2511 '-pre' and '-post' switches. '-post' is only used if '-t' is
2512 specified and the ENGINE is successfully initialised. The syntax for
2513 the individual commands are colon-separated, for example;
2514 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2517 *) New dynamic control command support for ENGINEs. ENGINEs can now
2518 declare their own commands (numbers), names (strings), descriptions,
2519 and input types for run-time discovery by calling applications. A
2520 subset of these commands are implicitly classed as "executable"
2521 depending on their input type, and only these can be invoked through
2522 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2523 can be based on user input, config files, etc). The distinction is
2524 that "executable" commands cannot return anything other than a boolean
2525 result and can only support numeric or string input, whereas some
2526 discoverable commands may only be for direct use through
2527 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2528 pointers, or other custom uses. The "executable" commands are to
2529 support parameterisations of ENGINE behaviour that can be
2530 unambiguously defined by ENGINEs and used consistently across any
2531 OpenSSL-based application. Commands have been added to all the
2532 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2533 control over shared-library paths without source code alterations.
2536 *) Changed all ENGINE implementations to dynamically allocate their
2537 ENGINEs rather than declaring them statically. Apart from this being
2538 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2539 this also allows the implementations to compile without using the
2540 internal engine_int.h header.
2543 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2544 'const' value. Any code that should be able to modify a RAND_METHOD
2545 should already have non-const pointers to it (ie. they should only
2546 modify their own ones).
2549 *) Made a variety of little tweaks to the ENGINE code.
2550 - "atalla" and "ubsec" string definitions were moved from header files
2551 to C code. "nuron" string definitions were placed in variables
2552 rather than hard-coded - allowing parameterisation of these values
2553 later on via ctrl() commands.
2554 - Removed unused "#if 0"'d code.
2555 - Fixed engine list iteration code so it uses ENGINE_free() to release
2556 structural references.
2557 - Constified the RAND_METHOD element of ENGINE structures.
2558 - Constified various get/set functions as appropriate and added
2559 missing functions (including a catch-all ENGINE_cpy that duplicates
2560 all ENGINE values onto a new ENGINE except reference counts/state).
2561 - Removed NULL parameter checks in get/set functions. Setting a method
2562 or function to NULL is a way of cancelling out a previously set
2563 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2564 and doesn't justify the extra error symbols and code.
2565 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2566 flags from engine_int.h to engine.h.
2567 - Changed prototypes for ENGINE handler functions (init(), finish(),
2568 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2571 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2572 to the algorithm using long division. The binary algorithm can be
2573 used only if the modulus is odd. On 32-bit systems, it is faster
2574 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2575 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2576 up to 450 bits. In 64-bit environments, the binary algorithm
2577 appears to be advantageous for much longer moduli; here we use it
2578 for moduli up to 2048 bits.
2581 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2582 could not support the combine flag in choice fields.
2585 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2586 extensions from a certificate request to the certificate.
2589 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2590 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2591 file: this allows the display of the certificate about to be
2592 signed to be customised, to allow certain fields to be included
2593 or excluded and extension details. The old system didn't display
2594 multicharacter strings properly, omitted fields not in the policy
2595 and couldn't display additional details such as extensions.
2598 *) Function EC_POINTs_mul for multiple scalar multiplication
2599 of an arbitrary number of elliptic curve points
2600 \sum scalars[i]*points[i],
2601 optionally including the generator defined for the EC_GROUP:
2602 scalar*generator + \sum scalars[i]*points[i].
2604 EC_POINT_mul is a simple wrapper function for the typical case
2605 that the point list has just one item (besides the optional
2609 *) First EC_METHODs for curves over GF(p):
2611 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2612 operations and provides various method functions that can also
2613 operate with faster implementations of modular arithmetic.
2615 EC_GFp_mont_method() reuses most functions that are part of
2616 EC_GFp_simple_method, but uses Montgomery arithmetic.
2618 [Bodo Moeller; point addition and point doubling
2619 implementation directly derived from source code provided by
2620 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2622 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2623 crypto/ec/ec_lib.c):
2625 Curves are EC_GROUP objects (with an optional group generator)
2626 based on EC_METHODs that are built into the library.
2628 Points are EC_POINT objects based on EC_GROUP objects.
2630 Most of the framework would be able to handle curves over arbitrary
2631 finite fields, but as there are no obvious types for fields other
2632 than GF(p), some functions are limited to that for now.
2635 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2636 that the file contains a complete HTTP response.
2639 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2640 change the def and num file printf format specifier from "%-40sXXX"
2641 to "%-39s XXX". The latter will always guarantee a space after the
2642 field while the former will cause them to run together if the field
2643 is 40 of more characters long.
2646 *) Constify the cipher and digest 'method' functions and structures
2647 and modify related functions to take constant EVP_MD and EVP_CIPHER
2651 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2652 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2655 *) Modify EVP_Digest*() routines so they now return values. Although the
2656 internal software routines can never fail additional hardware versions
2660 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2662 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2663 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2666 ERR_R_NESTED_ASN1_ERROR
2668 ERR_R_MISSING_ASN1_EOS
2669 were 4 .. 9, conflicting with
2670 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2672 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2673 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2675 Add new error code 'ERR_R_INTERNAL_ERROR'.
2678 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2682 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2683 sets the subject name for a new request or supersedes the
2684 subject name in a given request. Formats that can be parsed are
2685 'CN=Some Name, OU=myOU, C=IT'
2687 'CN=Some Name/OU=myOU/C=IT'.
2689 Add options '-batch' and '-verbose' to 'openssl req'.
2690 [Massimiliano Pala <madwolf@hackmasters.net>]
2692 *) Introduce the possibility to access global variables through
2693 functions on platform were that's the best way to handle exporting
2694 global variables in shared libraries. To enable this functionality,
2695 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2696 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2697 is normally done by Configure or something similar).
2699 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2700 in the source file (foo.c) like this:
2702 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2703 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2705 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2706 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2708 OPENSSL_DECLARE_GLOBAL(int,foo);
2709 #define foo OPENSSL_GLOBAL_REF(foo)
2710 OPENSSL_DECLARE_GLOBAL(double,bar);
2711 #define bar OPENSSL_GLOBAL_REF(bar)
2713 The #defines are very important, and therefore so is including the
2714 header file everywhere where the defined globals are used.
2716 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2717 of ASN.1 items, but that structure is a bit different.
2719 The largest change is in util/mkdef.pl which has been enhanced with
2720 better and easier to understand logic to choose which symbols should
2721 go into the Windows .def files as well as a number of fixes and code
2722 cleanup (among others, algorithm keywords are now sorted
2723 lexicographically to avoid constant rewrites).
2726 *) In BN_div() keep a copy of the sign of 'num' before writing the
2727 result to 'rm' because if rm==num the value will be overwritten
2728 and produce the wrong result if 'num' is negative: this caused
2729 problems with BN_mod() and BN_nnmod().
2732 *) Function OCSP_request_verify(). This checks the signature on an
2733 OCSP request and verifies the signer certificate. The signer
2734 certificate is just checked for a generic purpose and OCSP request
2738 *) Add OCSP_check_validity() function to check the validity of OCSP
2739 responses. OCSP responses are prepared in real time and may only
2740 be a few seconds old. Simply checking that the current time lies
2741 between thisUpdate and nextUpdate max reject otherwise valid responses
2742 caused by either OCSP responder or client clock inaccuracy. Instead
2743 we allow thisUpdate and nextUpdate to fall within a certain period of
2744 the current time. The age of the response can also optionally be
2745 checked. Two new options -validity_period and -status_age added to
2749 *) If signature or public key algorithm is unrecognized print out its
2750 OID rather that just UNKNOWN.
2753 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2754 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2755 ID to be generated from the issuer certificate alone which can then be
2756 passed to OCSP_id_issuer_cmp().
2759 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2760 ASN1 modules to export functions returning ASN1_ITEM pointers
2761 instead of the ASN1_ITEM structures themselves. This adds several
2762 new macros which allow the underlying ASN1 function/structure to
2763 be accessed transparently. As a result code should not use ASN1_ITEM
2764 references directly (such as &X509_it) but instead use the relevant
2765 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2766 use of the new ASN1 code on platforms where exporting structures
2767 is problematical (for example in shared libraries) but exporting
2768 functions returning pointers to structures is not.
2771 *) Add support for overriding the generation of SSL/TLS session IDs.
2772 These callbacks can be registered either in an SSL_CTX or per SSL.
2773 The purpose of this is to allow applications to control, if they wish,
2774 the arbitrary values chosen for use as session IDs, particularly as it
2775 can be useful for session caching in multiple-server environments. A
2776 command-line switch for testing this (and any client code that wishes
2777 to use such a feature) has been added to "s_server".
2778 [Geoff Thorpe, Lutz Jaenicke]
2780 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2781 of the form '#if defined(...) || defined(...) || ...' and
2782 '#if !defined(...) && !defined(...) && ...'. This also avoids
2783 the growing number of special cases it was previously handling.
2786 *) Make all configuration macros available for application by making
2787 sure they are available in opensslconf.h, by giving them names starting
2788 with "OPENSSL_" to avoid conflicts with other packages and by making
2789 sure e_os2.h will cover all platform-specific cases together with
2791 Additionally, it is now possible to define configuration/platform-
2792 specific names (called "system identities"). In the C code, these
2793 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2794 macro with the name beginning with "OPENSSL_SYS_", which is determined
2795 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2799 *) New option -set_serial to 'req' and 'x509' this allows the serial
2800 number to use to be specified on the command line. Previously self
2801 signed certificates were hard coded with serial number 0 and the
2802 CA options of 'x509' had to use a serial number in a file which was
2806 *) New options to 'ca' utility to support V2 CRL entry extensions.
2807 Currently CRL reason, invalidity date and hold instruction are
2808 supported. Add new CRL extensions to V3 code and some new objects.
2811 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2812 disable standard block padding (aka PKCS#5 padding) in the EVP
2813 API, which was previously mandatory. This means that the data is
2814 not padded in any way and so the total length much be a multiple
2815 of the block size, otherwise an error occurs.
2818 *) Initial (incomplete) OCSP SSL support.
2821 *) New function OCSP_parse_url(). This splits up a URL into its host,
2822 port and path components: primarily to parse OCSP URLs. New -url
2823 option to ocsp utility.
2826 *) New nonce behavior. The return value of OCSP_check_nonce() now
2827 reflects the various checks performed. Applications can decide
2828 whether to tolerate certain situations such as an absent nonce
2829 in a response when one was present in a request: the ocsp application
2830 just prints out a warning. New function OCSP_add1_basic_nonce()
2831 this is to allow responders to include a nonce in a response even if
2832 the request is nonce-less.
2835 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2836 skipped when using openssl x509 multiple times on a single input file,
2837 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2840 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2841 set string type: to handle setting ASN1_TIME structures. Fix ca
2842 utility to correctly initialize revocation date of CRLs.
2845 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2846 the clients preferred ciphersuites and rather use its own preferences.
2847 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2848 Internet Explorer by ensuring unchanged hash method during stepup.
2849 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2852 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2853 to aes and add a new 'exist' option to print out symbols that don't
2857 *) Additional options to ocsp utility to allow flags to be set and
2858 additional certificates supplied.
2861 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2862 OCSP client a number of certificate to only verify the response
2866 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2867 handle the new API. Currently only ECB, CBC modes supported. Add new
2870 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2871 Encryption Standard (AES) Ciphersuites for Transport Layer
2872 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2873 not enabled by default and were not part of the "ALL" ciphersuite
2874 alias because they were not yet official; they could be
2875 explicitly requested by specifying the "AESdraft" ciphersuite
2876 group alias. In the final release of OpenSSL 0.9.7, the group
2877 alias is called "AES" and is part of "ALL".)
2878 [Ben Laurie, Steve Henson, Bodo Moeller]
2880 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2881 request to response.
2884 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2885 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2886 extract information from a certificate request. OCSP_response_create()
2887 creates a response and optionally adds a basic response structure.
2888 OCSP_basic_add1_status() adds a complete single response to a basic
2889 response and returns the OCSP_SINGLERESP structure just added (to allow
2890 extensions to be included for example). OCSP_basic_add1_cert() adds a
2891 certificate to a basic response and OCSP_basic_sign() signs a basic
2892 response with various flags. New helper functions ASN1_TIME_check()
2893 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2894 (converts ASN1_TIME to GeneralizedTime).
2897 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2898 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2899 structure from a certificate. X509_pubkey_digest() digests the public_key
2900 contents: this is used in various key identifiers.
2903 *) Make sk_sort() tolerate a NULL argument.
2904 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2906 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2907 passed by the function are trusted implicitly. If any of them signed the
2908 response then it is assumed to be valid and is not verified.
2911 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2912 to data. This was previously part of the PKCS7 ASN1 code. This
2913 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2914 [Steve Henson, reported by Kenneth R. Robinette
2915 <support@securenetterm.com>]
2917 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2918 routines: without these tracing memory leaks is very painful.
2919 Fix leaks in PKCS12 and PKCS7 routines.
2922 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2923 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2924 effectively meant GeneralizedTime would never be used. Now it
2925 is initialised to -1 but X509_time_adj() now has to check the value
2926 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2927 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2928 [Steve Henson, reported by Kenneth R. Robinette
2929 <support@securenetterm.com>]
2931 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2932 result in a zero length in the ASN1_INTEGER structure which was
2933 not consistent with the structure when d2i_ASN1_INTEGER() was used
2934 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2935 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2936 where it did not print out a minus for negative ASN1_INTEGER.
2939 *) Add summary printout to ocsp utility. The various functions which
2940 convert status values to strings have been renamed to:
2941 OCSP_response_status_str(), OCSP_cert_status_str() and
2942 OCSP_crl_reason_str() and are no longer static. New options
2943 to verify nonce values and to disable verification. OCSP response
2944 printout format cleaned up.
2947 *) Add additional OCSP certificate checks. These are those specified
2948 in RFC2560. This consists of two separate checks: the CA of the
2949 certificate being checked must either be the OCSP signer certificate
2950 or the issuer of the OCSP signer certificate. In the latter case the
2951 OCSP signer certificate must contain the OCSP signing extended key
2952 usage. This check is performed by attempting to match the OCSP
2953 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2954 in the OCSP_CERTID structures of the response.
2957 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2958 and related routines. This uses the standard OpenSSL certificate
2959 verify routines to perform initial checks (just CA validity) and
2960 to obtain the certificate chain. Then additional checks will be
2961 performed on the chain. Currently the root CA is checked to see
2962 if it is explicitly trusted for OCSP signing. This is used to set
2963 a root CA as a global signing root: that is any certificate that
2964 chains to that CA is an acceptable OCSP signing certificate.
2967 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2968 extensions from a separate configuration file.
2969 As when reading extensions from the main configuration file,
2970 the '-extensions ...' option may be used for specifying the
2972 [Massimiliano Pala <madwolf@comune.modena.it>]
2974 *) New OCSP utility. Allows OCSP requests to be generated or
2975 read. The request can be sent to a responder and the output
2976 parsed, outputed or printed in text form. Not complete yet:
2977 still needs to check the OCSP response validity.
2980 *) New subcommands for 'openssl ca':
2981 'openssl ca -status <serial>' prints the status of the cert with
2982 the given serial number (according to the index file).
2983 'openssl ca -updatedb' updates the expiry status of certificates
2985 [Massimiliano Pala <madwolf@comune.modena.it>]
2987 *) New '-newreq-nodes' command option to CA.pl. This is like
2988 '-newreq', but calls 'openssl req' with the '-nodes' option
2989 so that the resulting key is not encrypted.
2990 [Damien Miller <djm@mindrot.org>]
2992 *) New configuration for the GNU Hurd.
2993 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2995 *) Initial code to implement OCSP basic response verify. This
2996 is currently incomplete. Currently just finds the signer's
2997 certificate and verifies the signature on the response.
3000 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3001 value of OPENSSLDIR. This is available via the new '-d' option
3002 to 'openssl version', and is also included in 'openssl version -a'.
3005 *) Allowing defining memory allocation callbacks that will be given
3006 file name and line number information in additional arguments
3007 (a const char* and an int). The basic functionality remains, as
3008 well as the original possibility to just replace malloc(),
3009 realloc() and free() by functions that do not know about these
3010 additional arguments. To register and find out the current
3011 settings for extended allocation functions, the following
3012 functions are provided:
3014 CRYPTO_set_mem_ex_functions
3015 CRYPTO_set_locked_mem_ex_functions
3016 CRYPTO_get_mem_ex_functions
3017 CRYPTO_get_locked_mem_ex_functions
3019 These work the same way as CRYPTO_set_mem_functions and friends.
3020 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3021 extended allocation function is enabled.
3022 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3023 a conventional allocation function is enabled.
3024 [Richard Levitte, Bodo Moeller]
3026 *) Finish off removing the remaining LHASH function pointer casts.
3027 There should no longer be any prototype-casting required when using
3028 the LHASH abstraction, and any casts that remain are "bugs". See
3029 the callback types and macros at the head of lhash.h for details
3030 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3033 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3034 If /dev/[u]random devices are not available or do not return enough
3035 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3037 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3038 /etc/entropy will be queried once each in this sequence, quering stops
3039 when enough entropy was collected without querying more sockets.
3042 *) Change the Unix RAND_poll() variant to be able to poll several
3043 random devices, as specified by DEVRANDOM, until a sufficient amount
3044 of data has been collected. We spend at most 10 ms on each file
3045 (select timeout) and read in non-blocking mode. DEVRANDOM now
3046 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3047 (previously it was just the string "/dev/urandom"), so on typical
3048 platforms the 10 ms delay will never occur.
3049 Also separate out the Unix variant to its own file, rand_unix.c.
3050 For VMS, there's a currently-empty rand_vms.c.
3053 *) Move OCSP client related routines to ocsp_cl.c. These
3054 provide utility functions which an application needing
3055 to issue a request to an OCSP responder and analyse the
3056 response will typically need: as opposed to those which an
3057 OCSP responder itself would need which will be added later.
3059 OCSP_request_sign() signs an OCSP request with an API similar
3060 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3061 response. OCSP_response_get1_basic() extracts basic response
3062 from response. OCSP_resp_find_status(): finds and extracts status
3063 information from an OCSP_CERTID structure (which will be created
3064 when the request structure is built). These are built from lower
3065 level functions which work on OCSP_SINGLERESP structures but
3066 wont normally be used unless the application wishes to examine
3067 extensions in the OCSP response for example.
3069 Replace nonce routines with a pair of functions.
3070 OCSP_request_add1_nonce() adds a nonce value and optionally
3071 generates a random value. OCSP_check_nonce() checks the
3072 validity of the nonce in an OCSP response.
3075 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3076 This doesn't copy the supplied OCSP_CERTID and avoids the
3077 need to free up the newly created id. Change return type
3078 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3079 This can then be used to add extensions to the request.
3080 Deleted OCSP_request_new(), since most of its functionality
3081 is now in OCSP_REQUEST_new() (and the case insensitive name
3082 clash) apart from the ability to set the request name which
3083 will be added elsewhere.
3086 *) Update OCSP API. Remove obsolete extensions argument from
3087 various functions. Extensions are now handled using the new
3088 OCSP extension code. New simple OCSP HTTP function which
3089 can be used to send requests and parse the response.
3092 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3093 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3094 uses the special reorder version of SET OF to sort the attributes
3095 and reorder them to match the encoded order. This resolves a long
3096 standing problem: a verify on a PKCS7 structure just after signing
3097 it used to fail because the attribute order did not match the
3098 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3099 it uses the received order. This is necessary to tolerate some broken
3100 software that does not order SET OF. This is handled by encoding
3101 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3102 to produce the required SET OF.
3105 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3106 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3107 files to get correct declarations of the ASN.1 item variables.
3110 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3111 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3112 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3113 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3114 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3115 ASN1_ITEM and no wrapper functions.
3118 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3119 replace the old function pointer based I/O routines. Change most of
3120 the *_d2i_bio() and *_d2i_fp() functions to use these.
3123 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3124 lines, recognice more "algorithms" that can be deselected, and make
3125 it complain about algorithm deselection that isn't recognised.
3128 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3129 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3130 to use new functions. Add NO_ASN1_OLD which can be set to remove
3131 some old style ASN1 functions: this can be used to determine if old
3132 code will still work when these eventually go away.
3135 *) New extension functions for OCSP structures, these follow the
3136 same conventions as certificates and CRLs.
3139 *) New function X509V3_add1_i2d(). This automatically encodes and
3140 adds an extension. Its behaviour can be customised with various
3141 flags to append, replace or delete. Various wrappers added for
3142 certifcates and CRLs.
3145 *) Fix to avoid calling the underlying ASN1 print routine when
3146 an extension cannot be parsed. Correct a typo in the
3147 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3150 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3151 entries for variables.
3154 *) Add functionality to apps/openssl.c for detecting locking
3155 problems: As the program is single-threaded, all we have
3156 to do is register a locking callback using an array for
3157 storing which locks are currently held by the program.
3160 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3161 SSL_get_ex_data_X509_STORE_idx(), which is used in
3162 ssl_verify_cert_chain() and thus can be called at any time
3163 during TLS/SSL handshakes so that thread-safety is essential.
3164 Unfortunately, the ex_data design is not at all suited
3165 for multi-threaded use, so it probably should be abolished.
3168 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3169 [Broadcom, tweaked and integrated by Geoff Thorpe]
3171 *) Move common extension printing code to new function
3172 X509V3_print_extensions(). Reorganise OCSP print routines and
3173 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3176 *) New function X509_signature_print() to remove duplication in some
3180 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3181 set (this was treated exactly the same as SET OF previously). This
3182 is used to reorder the STACK representing the structure to match the
3183 encoding. This will be used to get round a problem where a PKCS7
3184 structure which was signed could not be verified because the STACK
3185 order did not reflect the encoded order.
3188 *) Reimplement the OCSP ASN1 module using the new code.
3191 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3192 for its ASN1 operations. The old style function pointers still exist
3193 for now but they will eventually go away.
3196 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3197 completely replaces the old ASN1 functionality with a table driven
3198 encoder and decoder which interprets an ASN1_ITEM structure describing
3199 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3200 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3201 has also been converted to the new form.
3204 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3205 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3206 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3207 for negative moduli.
3210 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3211 of not touching the result's sign bit.
3214 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3218 *) Changed the LHASH code to use prototypes for callbacks, and created
3219 macros to declare and implement thin (optionally static) functions
3220 that provide type-safety and avoid function pointer casting for the
3221 type-specific callbacks.
3224 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3226 [Veers Staats <staatsvr@asc.hpc.mil>,
3227 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3229 *) Reformat the FAQ so the different questions and answers can be divided
3230 in sections depending on the subject.
3233 *) Have the zlib compression code load ZLIB.DLL dynamically under
3237 *) New function BN_mod_sqrt for computing square roots modulo a prime
3238 (using the probabilistic Tonelli-Shanks algorithm unless
3239 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3240 be handled deterministically).
3241 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3243 *) Make BN_mod_inverse faster by explicitly handling small quotients
3244 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3245 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3248 *) New function BN_kronecker.
3251 *) Fix BN_gcd so that it works on negative inputs; the result is
3252 positive unless both parameters are zero.
3253 Previously something reasonably close to an infinite loop was
3254 possible because numbers could be growing instead of shrinking
3255 in the implementation of Euclid's algorithm.
3258 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3259 sign of the number in question.
3261 Fix BN_is_word(a,w) to work correctly for w == 0.
3263 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3264 because its test if the absolute value of 'a' equals 'w'.
3265 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3266 it exists mostly for use in the implementations of BN_is_zero(),
3267 BN_is_one(), and BN_is_word().
3270 *) New function BN_swap.
3273 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3274 the exponentiation functions are more likely to produce reasonable
3275 results on negative inputs.
3278 *) Change BN_mod_mul so that the result is always non-negative.
3279 Previously, it could be negative if one of the factors was negative;
3280 I don't think anyone really wanted that behaviour.
3283 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3284 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3285 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3286 and add new functions:
3295 BN_mod_lshift1_quick
3299 These functions always generate non-negative results.
3301 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3302 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3304 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3305 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3306 be reduced modulo m.
3307 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3310 The following entry accidentily appeared in the CHANGES file
3311 distributed with OpenSSL 0.9.7. The modifications described in
3312 it do *not* apply to OpenSSL 0.9.7.
3314 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3315 was actually never needed) and in BN_mul(). The removal in BN_mul()
3316 required a small change in bn_mul_part_recursive() and the addition
3317 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3318 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3319 bn_sub_words() and bn_add_words() except they take arrays with
3324 *) In 'openssl passwd', verify passwords read from the terminal
3325 unless the '-salt' option is used (which usually means that
3326 verification would just waste user's time since the resulting
3327 hash is going to be compared with some given password hash)
3328 or the new '-noverify' option is used.
3330 This is an incompatible change, but it does not affect
3331 non-interactive use of 'openssl passwd' (passwords on the command
3332 line, '-stdin' option, '-in ...' option) and thus should not
3336 *) Remove all references to RSAref, since there's no more need for it.
3339 *) Make DSO load along a path given through an environment variable
3340 (SHLIB_PATH) with shl_load().
3343 *) Constify the ENGINE code as a result of BIGNUM constification.
3344 Also constify the RSA code and most things related to it. In a
3345 few places, most notable in the depth of the ASN.1 code, ugly
3346 casts back to non-const were required (to be solved at a later
3350 *) Make it so the openssl application has all engines loaded by default.
3353 *) Constify the BIGNUM routines a little more.
3356 *) Add the following functions:
3358 ENGINE_load_cswift()
3360 ENGINE_load_atalla()
3362 ENGINE_load_builtin_engines()
3364 That way, an application can itself choose if external engines that
3365 are built-in in OpenSSL shall ever be used or not. The benefit is
3366 that applications won't have to be linked with libdl or other dso
3367 libraries unless it's really needed.
3369 Changed 'openssl engine' to load all engines on demand.
3370 Changed the engine header files to avoid the duplication of some
3371 declarations (they differed!).
3374 *) 'openssl engine' can now list capabilities.
3377 *) Better error reporting in 'openssl engine'.
3380 *) Never call load_dh_param(NULL) in s_server.
3383 *) Add engine application. It can currently list engines by name and
3384 identity, and test if they are actually available.
3387 *) Improve RPM specification file by forcing symbolic linking and making
3388 sure the installed documentation is also owned by root.root.
3389 [Damien Miller <djm@mindrot.org>]
3391 *) Give the OpenSSL applications more possibilities to make use of
3392 keys (public as well as private) handled by engines.
3395 *) Add OCSP code that comes from CertCo.
3398 *) Add VMS support for the Rijndael code.
3401 *) Added untested support for Nuron crypto accelerator.
3404 *) Add support for external cryptographic devices. This code was
3405 previously distributed separately as the "engine" branch.
3406 [Geoff Thorpe, Richard Levitte]
3408 *) Rework the filename-translation in the DSO code. It is now possible to
3409 have far greater control over how a "name" is turned into a filename
3410 depending on the operating environment and any oddities about the
3411 different shared library filenames on each system.
3414 *) Support threads on FreeBSD-elf in Configure.
3417 *) Fix for SHA1 assembly problem with MASM: it produces
3418 warnings about corrupt line number information when assembling
3419 with debugging information. This is caused by the overlapping
3421 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3424 NCONF_get_number() has no error checking at all. As a replacement,
3425 NCONF_get_number_e() is defined (_e for "error checking") and is
3426 promoted strongly. The old NCONF_get_number is kept around for
3427 binary backward compatibility.
3428 Make it possible for methods to load from something other than a BIO,
3429 by providing a function pointer that is given a name instead of a BIO.
3430 For example, this could be used to load configuration data from an
3434 *) Fix for non blocking accept BIOs. Added new I/O special reason
3435 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3436 with non blocking I/O was not possible because no retry code was
3437 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3441 *) Added the beginnings of Rijndael support.
3444 *) Fix for bug in DirectoryString mask setting. Add support for
3445 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3446 to allow certificate printing to more controllable, additional
3447 'certopt' option to 'x509' to allow new printing options to be
3451 *) Clean old EAY MD5 hack from e_os.h.
3454 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3456 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3457 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3458 [Joe Orton, Steve Henson]
3460 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3462 *) Fix additional bug revealed by the NISCC test suite:
3464 Stop bug triggering large recursion when presented with
3465 certain ASN.1 tags (CVE-2003-0851)
3468 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3470 *) Fix various bugs revealed by running the NISCC test suite:
3472 Stop out of bounds reads in the ASN1 code when presented with
3473 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3475 If verify callback ignores invalid public key errors don't try to check
3476 certificate signature with the NULL public key.
3480 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3481 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3485 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3486 extra data after the compression methods not only for TLS 1.0
3487 but also for SSL 3.0 (as required by the specification).
3488 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3490 *) Change X509_certificate_type() to mark the key as exported/exportable
3491 when it's 512 *bits* long, not 512 bytes.
3494 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3496 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3497 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3498 a protocol version number mismatch like a decryption error
3499 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3502 *) Turn on RSA blinding by default in the default implementation
3503 to avoid a timing attack. Applications that don't want it can call
3504 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3505 They would be ill-advised to do so in most cases.
3506 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3508 *) Change RSA blinding code so that it works when the PRNG is not
3509 seeded (in this case, the secret RSA exponent is abused as
3510 an unpredictable seed -- if it is not unpredictable, there
3511 is no point in blinding anyway). Make RSA blinding thread-safe
3512 by remembering the creator's thread ID in rsa->blinding and
3513 having all other threads use local one-time blinding factors
3514 (this requires more computation than sharing rsa->blinding, but
3515 avoids excessive locking; and if an RSA object is not shared
3516 between threads, blinding will still be very fast).
3519 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3521 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3522 via timing by performing a MAC computation even if incorrrect
3523 block cipher padding has been found. This is a countermeasure
3524 against active attacks where the attacker has to distinguish
3525 between bad padding and a MAC verification error. (CVE-2003-0078)
3527 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3528 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3529 Martin Vuagnoux (EPFL, Ilion)]
3531 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3533 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3534 memory from it's contents. This is done with a counter that will
3535 place alternating values in each byte. This can be used to solve
3536 two issues: 1) the removal of calls to memset() by highly optimizing
3537 compilers, and 2) cleansing with other values than 0, since those can
3538 be read through on certain media, for example a swap space on disk.
3541 *) Bugfix: client side session caching did not work with external caching,
3542 because the session->cipher setting was not restored when reloading
3543 from the external cache. This problem was masked, when
3544 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3545 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3548 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3549 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3550 [Zeev Lieber <zeev-l@yahoo.com>]
3552 *) Undo an undocumented change introduced in 0.9.6e which caused
3553 repeated calls to OpenSSL_add_all_ciphers() and
3554 OpenSSL_add_all_digests() to be ignored, even after calling
3558 *) Change the default configuration reader to deal with last line not
3559 being properly terminated.
3562 *) Change X509_NAME_cmp() so it applies the special rules on handling
3563 DN values that are of type PrintableString, as well as RDNs of type
3564 emailAddress where the value has the type ia5String.
3565 [stefank@valicert.com via Richard Levitte]
3567 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3568 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3569 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3570 the bitwise-OR of the two for use by the majority of applications
3571 wanting this behaviour, and update the docs. The documented
3572 behaviour and actual behaviour were inconsistent and had been
3573 changing anyway, so this is more a bug-fix than a behavioural
3575 [Geoff Thorpe, diagnosed by Nadav Har'El]
3577 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3578 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3581 *) Fix initialization code race conditions in
3582 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3583 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3584 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3585 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3586 ssl2_get_cipher_by_char(),
3587 ssl3_get_cipher_by_char().
3588 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3590 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3591 the cached sessions are flushed, as the remove_cb() might use ex_data
3592 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3593 (see [openssl.org #212]).
3594 [Geoff Thorpe, Lutz Jaenicke]
3596 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3597 length, instead of the encoding length to d2i_ASN1_OBJECT.
3600 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3602 *) [In 0.9.6g-engine release:]
3603 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3604 [Lynn Gazis <lgazis@rainbow.com>]
3606 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3608 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3609 and get fix the header length calculation.
3610 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3611 Alon Kantor <alonk@checkpoint.com> (and others),
3614 *) Use proper error handling instead of 'assertions' in buffer
3615 overflow checks added in 0.9.6e. This prevents DoS (the
3616 assertions could call abort()).
3617 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3619 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3621 *) Add various sanity checks to asn1_get_length() to reject
3622 the ASN1 length bytes if they exceed sizeof(long), will appear
3623 negative or the content length exceeds the length of the
3625 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3627 *) Fix cipher selection routines: ciphers without encryption had no flags
3628 for the cipher strength set and where therefore not handled correctly
3629 by the selection routines (PR #130).
3632 *) Fix EVP_dsa_sha macro.
3636 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3637 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3638 that was added in OpenSSL 0.9.6d.
3640 As the countermeasure turned out to be incompatible with some
3641 broken SSL implementations, the new option is part of SSL_OP_ALL.
3642 SSL_OP_ALL is usually employed when compatibility with weird SSL
3643 implementations is desired (e.g. '-bugs' option to 's_client' and
3644 's_server'), so the new option is automatically set in many
3648 *) Changes in security patch:
3650 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3651 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3652 Air Force Materiel Command, USAF, under agreement number
3655 *) Add various sanity checks to asn1_get_length() to reject
3656 the ASN1 length bytes if they exceed sizeof(long), will appear
3657 negative or the content length exceeds the length of the
3658 supplied buffer. (CVE-2002-0659)
3659 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3661 *) Assertions for various potential buffer overflows, not known to
3663 [Ben Laurie (CHATS)]
3665 *) Various temporary buffers to hold ASCII versions of integers were
3666 too small for 64 bit platforms. (CVE-2002-0655)
3667 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3669 *) Remote buffer overflow in SSL3 protocol - an attacker could
3670 supply an oversized session ID to a client. (CVE-2002-0656)
3671 [Ben Laurie (CHATS)]
3673 *) Remote buffer overflow in SSL2 protocol - an attacker could
3674 supply an oversized client master key. (CVE-2002-0656)
3675 [Ben Laurie (CHATS)]
3677 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3679 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3680 encoded as NULL) with id-dsa-with-sha1.
3681 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3683 *) Check various X509_...() return values in apps/req.c.
3684 [Nils Larsch <nla@trustcenter.de>]
3686 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3687 an end-of-file condition would erronously be flagged, when the CRLF
3688 was just at the end of a processed block. The bug was discovered when
3689 processing data through a buffering memory BIO handing the data to a
3690 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3691 <ptsekov@syntrex.com> and Nedelcho Stanev.
3694 *) Implement a countermeasure against a vulnerability recently found
3695 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3696 before application data chunks to avoid the use of known IVs
3697 with data potentially chosen by the attacker.
3700 *) Fix length checks in ssl3_get_client_hello().
3703 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3704 to prevent ssl3_read_internal() from incorrectly assuming that
3705 ssl3_read_bytes() found application data while handshake
3706 processing was enabled when in fact s->s3->in_read_app_data was
3707 merely automatically cleared during the initial handshake.
3708 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3710 *) Fix object definitions for Private and Enterprise: they were not
3711 recognized in their shortname (=lowercase) representation. Extend
3712 obj_dat.pl to issue an error when using undefined keywords instead
3713 of silently ignoring the problem (Svenning Sorensen
3714 <sss@sss.dnsalias.net>).
3717 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3718 generators, i.e. generators other than 2 and 5. (Previously, the
3719 code did not properly initialise the 'add' and 'rem' values to
3720 BN_generate_prime().)
3722 In the new general case, we do not insist that 'generator' is
3723 actually a primitive root: This requirement is rather pointless;
3724 a generator of the order-q subgroup is just as good, if not
3728 *) Map new X509 verification errors to alerts. Discovered and submitted by
3729 Tom Wu <tom@arcot.com>.
3732 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3733 returning non-zero before the data has been completely received
3734 when using non-blocking I/O.
3735 [Bodo Moeller; problem pointed out by John Hughes]
3737 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3738 [Ben Laurie, Lutz Jaenicke]
3740 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3741 Yoram Zahavi <YoramZ@gilian.com>).
3744 *) Add information about CygWin 1.3 and on, and preserve proper
3745 configuration for the versions before that.
3746 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3748 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3749 check whether we deal with a copy of a session and do not delete from
3750 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3751 <izhar@checkpoint.com>.
3754 *) Do not store session data into the internal session cache, if it
3755 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3756 flag is set). Proposed by Aslam <aslam@funk.com>.
3759 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3763 *) [In 0.9.6d-engine release:]
3764 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3765 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3767 *) Add the configuration target linux-s390x.
3768 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3770 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3771 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3772 variable as an indication that a ClientHello message has been
3773 received. As the flag value will be lost between multiple
3774 invocations of ssl3_accept when using non-blocking I/O, the
3775 function may not be aware that a handshake has actually taken
3776 place, thus preventing a new session from being added to the
3779 To avoid this problem, we now set s->new_session to 2 instead of
3780 using a local variable.
3781 [Lutz Jaenicke, Bodo Moeller]
3783 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3784 if the SSL_R_LENGTH_MISMATCH error is detected.
3785 [Geoff Thorpe, Bodo Moeller]
3787 *) New 'shared_ldflag' column in Configure platform table.
3790 *) Fix EVP_CIPHER_mode macro.
3791 ["Dan S. Camper" <dan@bti.net>]
3793 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3794 type, we must throw them away by setting rr->length to 0.
3795 [D P Chang <dpc@qualys.com>]
3797 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3799 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3800 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3801 worked incorrectly for those cases where range = 10..._2 and
3802 3*range is two bits longer than range.)
3805 *) Only add signing time to PKCS7 structures if it is not already
3809 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3810 OBJ_ld_ce should be OBJ_id_ce.
3811 Also some ip-pda OIDs in crypto/objects/objects.txt were
3812 incorrect (cf. RFC 3039).
3813 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3815 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3816 returns early because it has nothing to do.
3817 [Andy Schneider <andy.schneider@bjss.co.uk>]
3819 *) [In 0.9.6c-engine release:]
3820 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3821 [Andy Schneider <andy.schneider@bjss.co.uk>]
3823 *) [In 0.9.6c-engine release:]
3824 Add support for Cryptographic Appliance's keyserver technology.
3825 (Use engine 'keyclient')
3826 [Cryptographic Appliances and Geoff Thorpe]
3828 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3829 is called via tools/c89.sh because arguments have to be
3830 rearranged (all '-L' options must appear before the first object
3832 [Richard Shapiro <rshapiro@abinitio.com>]
3834 *) [In 0.9.6c-engine release:]
3835 Add support for Broadcom crypto accelerator cards, backported
3837 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3839 *) [In 0.9.6c-engine release:]
3840 Add support for SureWare crypto accelerator cards from
3841 Baltimore Technologies. (Use engine 'sureware')
3842 [Baltimore Technologies and Mark Cox]
3844 *) [In 0.9.6c-engine release:]
3845 Add support for crypto accelerator cards from Accelerated
3846 Encryption Processing, www.aep.ie. (Use engine 'aep')
3847 [AEP Inc. and Mark Cox]
3849 *) Add a configuration entry for gcc on UnixWare.
3850 [Gary Benson <gbenson@redhat.com>]
3852 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3853 messages are stored in a single piece (fixed-length part and
3854 variable-length part combined) and fix various bugs found on the way.
3857 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3858 instead. BIO_gethostbyname() does not know what timeouts are
3859 appropriate, so entries would stay in cache even when they have
3861 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3863 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3864 faced with a pathologically small ClientHello fragment that does
3865 not contain client_version: Instead of aborting with an error,
3866 simply choose the highest available protocol version (i.e.,
3867 TLS 1.0 unless it is disabled). In practice, ClientHello
3868 messages are never sent like this, but this change gives us
3869 strictly correct behaviour at least for TLS.
3872 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3873 never resets s->method to s->ctx->method when called from within
3874 one of the SSL handshake functions.
3875 [Bodo Moeller; problem pointed out by Niko Baric]
3877 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3878 (sent using the client's version number) if client_version is
3879 smaller than the protocol version in use. Also change
3880 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3881 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3882 the client will at least see that alert.
3885 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3889 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3890 client receives HelloRequest while in a handshake.
3891 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3893 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3894 should end in 'break', not 'goto end' which circuments various
3895 cleanups done in state SSL_ST_OK. But session related stuff
3896 must be disabled for SSL_ST_OK in the case that we just sent a
3899 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3900 before just sending a HelloRequest.
3901 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3903 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3904 reveal whether illegal block cipher padding was found or a MAC
3905 verification error occured. (Neither SSLerr() codes nor alerts
3906 are directly visible to potential attackers, but the information
3907 may leak via logfiles.)
3909 Similar changes are not required for the SSL 2.0 implementation
3910 because the number of padding bytes is sent in clear for SSL 2.0,
3911 and the extra bytes are just ignored. However ssl/s2_pkt.c
3912 failed to verify that the purported number of padding bytes is in
3916 *) Add OpenUNIX-8 support including shared libraries
3917 (Boyd Lynn Gerber <gerberb@zenez.com>).
3920 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3921 'wristwatch attack' using huge encoding parameters (cf.
3922 James H. Manger's CRYPTO 2001 paper). Note that the
3923 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3924 encoding parameters and hence was not vulnerable.
3927 *) BN_sqr() bug fix.
3928 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3930 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3931 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3932 followed by modular reduction.
3933 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3935 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3936 equivalent based on BN_pseudo_rand() instead of BN_rand().
3939 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3940 This function was broken, as the check for a new client hello message
3941 to handle SGC did not allow these large messages.
3942 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3945 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3948 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3949 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3952 *) Rework the configuration and shared library support for Tru64 Unix.
3953 The configuration part makes use of modern compiler features and
3954 still retains old compiler behavior for those that run older versions
3955 of the OS. The shared library support part includes a variant that
3956 uses the RPATH feature, and is available through the special
3957 configuration target "alpha-cc-rpath", which will never be selected
3959 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3961 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3962 with the same message size as in ssl3_get_certificate_request().
3963 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3964 messages might inadvertently be reject as too long.
3965 [Petr Lampa <lampa@fee.vutbr.cz>]
3967 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3970 *) Modified SSL library such that the verify_callback that has been set
3971 specificly for an SSL object with SSL_set_verify() is actually being
3972 used. Before the change, a verify_callback set with this function was
3973 ignored and the verify_callback() set in the SSL_CTX at the time of
3974 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3975 to allow the necessary settings.
3978 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3979 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3980 done automatically (in contradiction to the requirements of the C
3981 standard). This made problems when used from OpenSSH.
3984 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3985 dh->length and always used
3987 BN_rand_range(priv_key, dh->p).
3989 BN_rand_range() is not necessary for Diffie-Hellman, and this
3990 specific range makes Diffie-Hellman unnecessarily inefficient if
3991 dh->length (recommended exponent length) is much smaller than the
3992 length of dh->p. We could use BN_rand_range() if the order of
3993 the subgroup was stored in the DH structure, but we only have
3998 BN_rand(priv_key, l, ...)
4000 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4006 RSA_eay_public_encrypt
4007 RSA_eay_private_decrypt
4008 RSA_eay_private_encrypt (signing)
4009 RSA_eay_public_decrypt (signature verification)
4011 (default implementations for RSA_public_encrypt,
4012 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4013 always reject numbers >= n.
4016 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4017 to synchronize access to 'locking_thread'. This is necessary on
4018 systems where access to 'locking_thread' (an 'unsigned long'
4019 variable) is not atomic.
4022 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4023 *before* setting the 'crypto_lock_rand' flag. The previous code had
4024 a race condition if 0 is a valid thread ID.
4025 [Travis Vitek <vitek@roguewave.com>]
4027 *) Add support for shared libraries under Irix.
4028 [Albert Chin-A-Young <china@thewrittenword.com>]
4030 *) Add configuration option to build on Linux on both big-endian and
4032 [Ralf Baechle <ralf@uni-koblenz.de>]
4034 *) Add the possibility to create shared libraries on HP-UX.
4037 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4039 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4040 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4041 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4042 PRNG state recovery was possible based on the output of
4043 one PRNG request appropriately sized to gain knowledge on
4044 'md' followed by enough consecutive 1-byte PRNG requests
4045 to traverse all of 'state'.
4047 1. When updating 'md_local' (the current thread's copy of 'md')
4048 during PRNG output generation, hash all of the previous
4049 'md_local' value, not just the half used for PRNG output.
4051 2. Make the number of bytes from 'state' included into the hash
4052 independent from the number of PRNG bytes requested.
4054 The first measure alone would be sufficient to avoid
4055 Markku-Juhani's attack. (Actually it had never occurred
4056 to me that the half of 'md_local' used for chaining was the
4057 half from which PRNG output bytes were taken -- I had always
4058 assumed that the secret half would be used.) The second
4059 measure makes sure that additional data from 'state' is never
4060 mixed into 'md_local' in small portions; this heuristically
4061 further strengthens the PRNG.
4064 *) Fix crypto/bn/asm/mips3.s.
4067 *) When only the key is given to "enc", the IV is undefined. Print out
4068 an error message in this case.
4071 *) Handle special case when X509_NAME is empty in X509 printing routines.
4074 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4075 positive and less than q.
4078 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4079 used: it isn't thread safe and the add_lock_callback should handle
4081 [Paul Rose <Paul.Rose@bridge.com>]
4083 *) Verify that incoming data obeys the block size in
4084 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4088 [Ulf Möller, Bodo Möller]
4090 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4091 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4092 when fixing the server behaviour for backwards-compatible 'client
4093 hello' messages. (Note that the attack is impractical against
4094 SSL 3.0 and TLS 1.0 anyway because length and version checking
4095 means that the probability of guessing a valid ciphertext is
4096 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4099 Before 0.9.5, the countermeasure (hide the error by generating a
4100 random 'decryption result') did not work properly because
4101 ERR_clear_error() was missing, meaning that SSL_get_error() would
4102 detect the supposedly ignored error.
4104 Both problems are now fixed.
4107 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4108 (previously it was 1024).
4111 *) Fix for compatibility mode trust settings: ignore trust settings
4112 unless some valid trust or reject settings are present.
4115 *) Fix for blowfish EVP: its a variable length cipher.
4118 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4119 parameters in DSA public key structures and return an error in the
4120 DSA routines if parameters are absent.
4123 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4124 in the current directory if neither $RANDFILE nor $HOME was set.
4125 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4126 caused some confusion to Windows users who haven't defined $HOME.
4127 Thus RAND_file_name() is changed again: e_os.h can define a
4128 DEFAULT_HOME, which will be used if $HOME is not set.
4129 For Windows, we use "C:"; on other platforms, we still require
4130 environment variables.
4132 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4133 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4134 having multiple threads call RAND_poll() concurrently.
4137 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4138 combination of a flag and a thread ID variable.
4139 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4140 flag), *other* threads can enter ssleay_add_bytes without obeying
4141 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4142 that they do not hold after the first thread unsets add_do_not_lock).
4145 *) Change bctest again: '-x' expressions are not available in all
4149 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4151 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4152 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4154 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4155 the default extension for executables, if any. Also, make the perl
4156 scripts that use symlink() to test if it really exists and use "cp"
4157 if it doesn't. All this made OpenSSL compilable and installable in
4161 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4162 If SEQUENCE is length is indefinite just set c->slen to the total
4163 amount of data available.
4164 [Steve Henson, reported by shige@FreeBSD.org]
4165 [This change does not apply to 0.9.7.]
4167 *) Change bctest to avoid here-documents inside command substitution
4168 (workaround for FreeBSD /bin/sh bug).
4169 For compatibility with Ultrix, avoid shell functions (introduced
4170 in the bctest version that searches along $PATH).
4173 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4174 with des_encrypt() defined on some operating systems, like Solaris
4178 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4179 On the Importance of Eliminating Errors in Cryptographic
4180 Computations, J. Cryptology 14 (2001) 2, 101-119,
4181 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4184 *) MIPS assembler BIGNUM division bug fix.
4187 *) Disabled incorrect Alpha assembler code.
4190 *) Fix PKCS#7 decode routines so they correctly update the length
4191 after reading an EOC for the EXPLICIT tag.
4193 [This change does not apply to 0.9.7.]
4195 *) Fix bug in PKCS#12 key generation routines. This was triggered
4196 if a 3DES key was generated with a 0 initial byte. Include
4197 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4198 (but broken) behaviour.
4201 *) Enhance bctest to search for a working bc along $PATH and print
4203 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4205 *) Fix memory leaks in err.c: free err_data string if necessary;
4206 don't write to the wrong index in ERR_set_error_data.
4209 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4213 *) Replace rdtsc with _emit statements for VC++ version 5.
4214 [Jeremy Cooper <jeremy@baymoo.org>]
4216 *) Make it possible to reuse SSLv2 sessions.
4219 *) In copy_email() check for >= 0 as a return value for
4220 X509_NAME_get_index_by_NID() since 0 is a valid index.
4221 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4223 *) Avoid coredump with unsupported or invalid public keys by checking if
4224 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4225 PKCS7_verify() fails with non detached data.
4228 *) Don't use getenv in library functions when run as setuid/setgid.
4229 New function OPENSSL_issetugid().
4232 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4233 due to incorrect handling of multi-threading:
4235 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4237 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4239 3. Count how many times MemCheck_off() has been called so that
4240 nested use can be treated correctly. This also avoids
4241 inband-signalling in the previous code (which relied on the
4242 assumption that thread ID 0 is impossible).
4245 *) Add "-rand" option also to s_client and s_server.
4248 *) Fix CPU detection on Irix 6.x.
4249 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4250 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4252 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4255 [This change does not apply to 0.9.7.]
4257 *) Use the cached encoding of an X509_NAME structure rather than
4258 copying it. This is apparently the reason for the libsafe "errors"
4259 but the code is actually correct.
4262 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4263 Bleichenbacher's DSA attack.
4264 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4265 to be set and top=0 forces the highest bit to be set; top=-1 is new
4266 and leaves the highest bit random.
4267 [Ulf Moeller, Bodo Moeller]
4269 *) In the NCONF_...-based implementations for CONF_... queries
4270 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4271 a temporary CONF structure with the data component set to NULL
4272 (which gives segmentation faults in lh_retrieve).
4273 Instead, use NULL for the CONF pointer in CONF_get_string and
4274 CONF_get_number (which may use environment variables) and directly
4275 return NULL from CONF_get_section.
4278 *) Fix potential buffer overrun for EBCDIC.
4281 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4282 keyUsage if basicConstraints absent for a CA.
4285 *) Make SMIME_write_PKCS7() write mail header values with a format that
4286 is more generally accepted (no spaces before the semicolon), since
4287 some programs can't parse those values properly otherwise. Also make
4288 sure BIO's that break lines after each write do not create invalid
4292 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4293 macros previously used would not encode an empty SEQUENCE OF
4294 and break the signature.
4296 [This change does not apply to 0.9.7.]
4298 *) Zero the premaster secret after deriving the master secret in
4302 *) Add some EVP_add_digest_alias registrations (as found in
4303 OpenSSL_add_all_digests()) to SSL_library_init()
4304 aka OpenSSL_add_ssl_algorithms(). This provides improved
4305 compatibility with peers using X.509 certificates
4306 with unconventional AlgorithmIdentifier OIDs.
4309 *) Fix for Irix with NO_ASM.
4310 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4312 *) ./config script fixes.
4313 [Ulf Moeller, Richard Levitte]
4315 *) Fix 'openssl passwd -1'.
4318 *) Change PKCS12_key_gen_asc() so it can cope with non null
4319 terminated strings whose length is passed in the passlen
4320 parameter, for example from PEM callbacks. This was done
4321 by adding an extra length parameter to asc2uni().
4322 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4324 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4325 call failed, free the DSA structure.
4328 *) Fix to uni2asc() to cope with zero length Unicode strings.
4329 These are present in some PKCS#12 files.
4332 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4333 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4334 when writing a 32767 byte record.
4335 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4337 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4338 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4340 (RSA objects have a reference count access to which is protected
4341 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4342 so they are meant to be shared between threads.)
4343 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4344 "Reddie, Steven" <Steven.Reddie@ca.com>]
4346 *) Fix a deadlock in CRYPTO_mem_leaks().
4349 *) Use better test patterns in bntest.
4352 *) rand_win.c fix for Borland C.
4355 *) BN_rshift bugfix for n == 0.
4358 *) Add a 'bctest' script that checks for some known 'bc' bugs
4359 so that 'make test' does not abort just because 'bc' is broken.
4362 *) Store verify_result within SSL_SESSION also for client side to
4363 avoid potential security hole. (Re-used sessions on the client side
4364 always resulted in verify_result==X509_V_OK, not using the original
4365 result of the server certificate verification.)
4368 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4369 SSL3_RT_APPLICATION_DATA, return 0.
4370 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4374 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4375 releases, have been re-implemented by renaming the previous
4376 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4377 and ssl3_read_internal, respectively, and adding 'peek' parameters
4378 to them. The new ssl[23]_{read,peek} functions are calls to
4379 ssl[23]_read_internal with the 'peek' flag set appropriately.
4380 A 'peek' parameter has also been added to ssl3_read_bytes, which
4381 does the actual work for ssl3_read_internal.
4384 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4385 the method-specific "init()" handler. Also clean up ex_data after
4386 calling the method-specific "finish()" handler. Previously, this was
4387 happening the other way round.
4390 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4391 The previous value, 12, was not always sufficient for BN_mod_exp().
4394 *) Make sure that shared libraries get the internal name engine with
4395 the full version number and not just 0. This should mark the
4396 shared libraries as not backward compatible. Of course, this should
4397 be changed again when we can guarantee backward binary compatibility.
4400 *) Fix typo in get_cert_by_subject() in by_dir.c
4401 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4403 *) Rework the system to generate shared libraries:
4405 - Make note of the expected extension for the shared libraries and
4406 if there is a need for symbolic links from for example libcrypto.so.0
4407 to libcrypto.so.0.9.7. There is extended info in Configure for
4410 - Make as few rebuilds of the shared libraries as possible.
4412 - Still avoid linking the OpenSSL programs with the shared libraries.
4414 - When installing, install the shared libraries separately from the
4418 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4420 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4421 and not in SSL_clear because the latter is also used by the
4422 accept/connect functions; previously, the settings made by
4423 SSL_set_read_ahead would be lost during the handshake.
4424 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4426 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4427 Previously, it would create entries for disableed algorithms no
4431 *) Added several new manual pages for SSL_* function.
4434 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4436 *) In ssl23_get_client_hello, generate an error message when faced
4437 with an initial SSL 3.0/TLS record that is too small to contain the
4438 first two bytes of the ClientHello message, i.e. client_version.
4439 (Note that this is a pathologic case that probably has never happened
4440 in real life.) The previous approach was to use the version number
4441 from the record header as a substitute; but our protocol choice
4442 should not depend on that one because it is not authenticated
4443 by the Finished messages.
4446 *) More robust randomness gathering functions for Windows.
4447 [Jeffrey Altman <jaltman@columbia.edu>]
4449 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4450 not set then we don't setup the error code for issuer check errors
4451 to avoid possibly overwriting other errors which the callback does
4452 handle. If an application does set the flag then we assume it knows
4453 what it is doing and can handle the new informational codes
4457 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4458 a general "ANY" type, as such it should be able to decode anything
4459 including tagged types. However it didn't check the class so it would
4460 wrongly interpret tagged types in the same way as their universal
4461 counterpart and unknown types were just rejected. Changed so that the
4462 tagged and unknown types are handled in the same way as a SEQUENCE:
4463 that is the encoding is stored intact. There is also a new type
4464 "V_ASN1_OTHER" which is used when the class is not universal, in this
4465 case we have no idea what the actual type is so we just lump them all
4469 *) On VMS, stdout may very well lead to a file that is written to
4470 in a record-oriented fashion. That means that every write() will
4471 write a separate record, which will be read separately by the
4472 programs trying to read from it. This can be very confusing.
4474 The solution is to put a BIO filter in the way that will buffer
4475 text until a linefeed is reached, and then write everything a
4476 line at a time, so every record written will be an actual line,
4477 not chunks of lines and not (usually doesn't happen, but I've
4478 seen it once) several lines in one record. BIO_f_linebuffer() is
4481 Currently, it's a VMS-only method, because that's where it has
4482 been tested well enough.
4485 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4486 it can return incorrect results.
4487 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4488 but it was in 0.9.6-beta[12].)
4491 *) Disable the check for content being present when verifying detached
4492 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4493 include zero length content when signing messages.
4496 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4497 BIO_ctrl (for BIO pairs).
4500 *) Add DSO method for VMS.
4503 *) Bug fix: Montgomery multiplication could produce results with the
4507 *) Add RPM specification openssl.spec and modify it to build three
4508 packages. The default package contains applications, application
4509 documentation and run-time libraries. The devel package contains
4510 include files, static libraries and function documentation. The
4511 doc package contains the contents of the doc directory. The original
4512 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4515 *) Add a large number of documentation files for many SSL routines.
4516 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4518 *) Add a configuration entry for Sony News 4.
4519 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4521 *) Don't set the two most significant bits to one when generating a
4522 random number < q in the DSA library.
4525 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4526 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4527 the underlying transport is blocking) if a handshake took place.
4528 (The default behaviour is needed by applications such as s_client
4529 and s_server that use select() to determine when to use SSL_read;
4530 but for applications that know in advance when to expect data, it
4531 just makes things more complicated.)
4534 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4538 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4539 work better on such systems.
4540 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4542 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4543 Update PKCS12_parse() so it copies the friendlyName and the
4544 keyid to the certificates aux info.
4547 *) Fix bug in PKCS7_verify() which caused an infinite loop
4548 if there was more than one signature.
4549 [Sven Uszpelkat <su@celocom.de>]
4551 *) Major change in util/mkdef.pl to include extra information
4552 about each symbol, as well as presentig variables as well
4553 as functions. This change means that there's n more need
4554 to rebuild the .num files when some algorithms are excluded.
4557 *) Allow the verify time to be set by an application,
4558 rather than always using the current time.
4561 *) Phase 2 verify code reorganisation. The certificate
4562 verify code now looks up an issuer certificate by a
4563 number of criteria: subject name, authority key id
4564 and key usage. It also verifies self signed certificates
4565 by the same criteria. The main comparison function is
4566 X509_check_issued() which performs these checks.
4568 Lot of changes were necessary in order to support this
4569 without completely rewriting the lookup code.
4571 Authority and subject key identifier are now cached.
4573 The LHASH 'certs' is X509_STORE has now been replaced
4574 by a STACK_OF(X509_OBJECT). This is mainly because an
4575 LHASH can't store or retrieve multiple objects with
4576 the same hash value.
4578 As a result various functions (which were all internal
4579 use only) have changed to handle the new X509_STORE
4580 structure. This will break anything that messed round
4581 with X509_STORE internally.
4583 The functions X509_STORE_add_cert() now checks for an
4584 exact match, rather than just subject name.
4586 The X509_STORE API doesn't directly support the retrieval
4587 of multiple certificates matching a given criteria, however
4588 this can be worked round by performing a lookup first
4589 (which will fill the cache with candidate certificates)
4590 and then examining the cache for matches. This is probably
4591 the best we can do without throwing out X509_LOOKUP
4592 entirely (maybe later...).
4594 The X509_VERIFY_CTX structure has been enhanced considerably.
4596 All certificate lookup operations now go via a get_issuer()
4597 callback. Although this currently uses an X509_STORE it
4598 can be replaced by custom lookups. This is a simple way
4599 to bypass the X509_STORE hackery necessary to make this
4600 work and makes it possible to use more efficient techniques
4601 in future. A very simple version which uses a simple
4602 STACK for its trusted certificate store is also provided
4603 using X509_STORE_CTX_trusted_stack().
4605 The verify_cb() and verify() callbacks now have equivalents
4606 in the X509_STORE_CTX structure.
4608 X509_STORE_CTX also has a 'flags' field which can be used
4609 to customise the verify behaviour.
4612 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4613 excludes S/MIME capabilities.
4616 *) When a certificate request is read in keep a copy of the
4617 original encoding of the signed data and use it when outputing
4618 again. Signatures then use the original encoding rather than
4619 a decoded, encoded version which may cause problems if the
4620 request is improperly encoded.
4623 *) For consistency with other BIO_puts implementations, call
4624 buffer_write(b, ...) directly in buffer_puts instead of calling
4627 In BIO_puts, increment b->num_write as in BIO_write.
4628 [Peter.Sylvester@EdelWeb.fr]
4630 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4631 BN_zero, we may not return a BIGNUM with an array consisting of
4635 *) Avoid calling abort() from within the library when problems are
4636 detected, except if preprocessor symbols have been defined
4637 (such as REF_CHECK, BN_DEBUG etc.).
4640 *) New openssl application 'rsautl'. This utility can be
4641 used for low level RSA operations. DER public key
4642 BIO/fp routines also added.
4645 *) New Configure entry and patches for compiling on QNX 4.
4646 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4648 *) A demo state-machine implementation was sponsored by
4649 Nuron (http://www.nuron.com/) and is now available in
4650 demos/state_machine.
4653 *) New options added to the 'dgst' utility for signature
4654 generation and verification.
4657 *) Unrecognized PKCS#7 content types are now handled via a
4658 catch all ASN1_TYPE structure. This allows unsupported
4659 types to be stored as a "blob" and an application can
4660 encode and decode it manually.
4663 *) Fix various signed/unsigned issues to make a_strex.c
4665 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4667 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4668 length if passed a buffer. ASN1_INTEGER_to_BN failed
4669 if passed a NULL BN and its argument was negative.
4670 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4672 *) Modification to PKCS#7 encoding routines to output definite
4673 length encoding. Since currently the whole structures are in
4674 memory there's not real point in using indefinite length
4675 constructed encoding. However if OpenSSL is compiled with
4676 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4679 *) Added BIO_vprintf() and BIO_vsnprintf().
4682 *) Added more prefixes to parse for in the the strings written
4683 through a logging bio, to cover all the levels that are available
4684 through syslog. The prefixes are now:
4686 PANIC, EMERG, EMR => LOG_EMERG
4687 ALERT, ALR => LOG_ALERT
4688 CRIT, CRI => LOG_CRIT
4689 ERROR, ERR => LOG_ERR
4690 WARNING, WARN, WAR => LOG_WARNING
4691 NOTICE, NOTE, NOT => LOG_NOTICE
4692 INFO, INF => LOG_INFO
4693 DEBUG, DBG => LOG_DEBUG
4695 and as before, if none of those prefixes are present at the
4696 beginning of the string, LOG_ERR is chosen.
4698 On Win32, the LOG_* levels are mapped according to this:
4700 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4701 LOG_WARNING => EVENTLOG_WARNING_TYPE
4702 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4706 *) Made it possible to reconfigure with just the configuration
4707 argument "reconf" or "reconfigure". The command line arguments
4708 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4709 and are retrieved from there when reconfiguring.
4713 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4715 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4718 *) The obj_dat.pl script was messing up the sorting of object
4719 names. The reason was that it compared the quoted version
4720 of strings as a result "OCSP" > "OCSP Signing" because
4721 " > SPACE. Changed script to store unquoted versions of
4722 names and add quotes on output. It was also omitting some
4723 names from the lookup table if they were given a default
4724 value (that is if SN is missing it is given the same
4725 value as LN and vice versa), these are now added on the
4726 grounds that if an object has a name we should be able to
4727 look it up. Finally added warning output when duplicate
4728 short or long names are found.
4731 *) Changes needed for Tandem NSK.
4732 [Scott Uroff <scott@xypro.com>]
4734 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4735 RSA_padding_check_SSLv23(), special padding was never detected
4736 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4737 version rollback attacks was not effective.
4739 In s23_clnt.c, don't use special rollback-attack detection padding
4740 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4741 client; similarly, in s23_srvr.c, don't do the rollback check if
4742 SSL 2.0 is the only protocol enabled in the server.
4745 *) Make it possible to get hexdumps of unprintable data with 'openssl
4746 asn1parse'. By implication, the functions ASN1_parse_dump() and
4747 BIO_dump_indent() are added.
4750 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4751 these print out strings and name structures based on various
4752 flags including RFC2253 support and proper handling of
4753 multibyte characters. Added options to the 'x509' utility
4754 to allow the various flags to be set.
4757 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4758 Also change the functions X509_cmp_current_time() and
4759 X509_gmtime_adj() work with an ASN1_TIME structure,
4760 this will enable certificates using GeneralizedTime in validity
4761 dates to be checked.
4764 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4765 negative public key encodings) on by default,
4766 NO_NEG_PUBKEY_BUG can be set to disable it.
4769 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4770 content octets. An i2c_ASN1_OBJECT is unnecessary because
4771 the encoding can be trivially obtained from the structure.
4774 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4775 not read locks (CRYPTO_r_[un]lock).
4778 *) A first attempt at creating official support for shared
4779 libraries through configuration. I've kept it so the
4780 default is static libraries only, and the OpenSSL programs
4781 are always statically linked for now, but there are
4782 preparations for dynamic linking in place.
4783 This has been tested on Linux and Tru64.
4786 *) Randomness polling function for Win9x, as described in:
4787 Peter Gutmann, Software Generation of Practically Strong
4791 *) Fix so PRNG is seeded in req if using an already existing
4795 *) New options to smime application. -inform and -outform
4796 allow alternative formats for the S/MIME message including
4797 PEM and DER. The -content option allows the content to be
4798 specified separately. This should allow things like Netscape
4799 form signing output easier to verify.
4802 *) Fix the ASN1 encoding of tags using the 'long form'.
4805 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4806 STRING types. These convert content octets to and from the
4807 underlying type. The actual tag and length octets are
4808 already assumed to have been read in and checked. These
4809 are needed because all other string types have virtually
4810 identical handling apart from the tag. By having versions
4811 of the ASN1 functions that just operate on content octets
4812 IMPLICIT tagging can be handled properly. It also allows
4813 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4814 and ASN1_INTEGER are identical apart from the tag.
4817 *) Change the handling of OID objects as follows:
4819 - New object identifiers are inserted in objects.txt, following
4820 the syntax given in objects.README.
4821 - objects.pl is used to process obj_mac.num and create a new
4823 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4826 This is currently kind of a hack, and the perl code in objects.pl
4827 isn't very elegant, but it works as I intended. The simplest way
4828 to check that it worked correctly is to look in obj_dat.h and
4829 check the array nid_objs and make sure the objects haven't moved
4830 around (this is important!). Additions are OK, as well as
4831 consistent name changes.
4834 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4837 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4838 The given file adds to whatever has already been seeded into the
4839 random pool through the RANDFILE configuration file option or
4840 environment variable, or the default random state file.
4843 *) mkstack.pl now sorts each macro group into lexical order.
4844 Previously the output order depended on the order the files
4845 appeared in the directory, resulting in needless rewriting
4849 *) Patches to make OpenSSL compile under Win32 again. Mostly
4850 work arounds for the VC++ problem that it treats func() as
4851 func(void). Also stripped out the parts of mkdef.pl that
4852 added extra typesafe functions: these no longer exist.
4855 *) Reorganisation of the stack code. The macros are now all
4856 collected in safestack.h . Each macro is defined in terms of
4857 a "stack macro" of the form SKM_<name>(type, a, b). The
4858 DEBUG_SAFESTACK is now handled in terms of function casts,
4859 this has the advantage of retaining type safety without the
4860 use of additional functions. If DEBUG_SAFESTACK is not defined
4861 then the non typesafe macros are used instead. Also modified the
4862 mkstack.pl script to handle the new form. Needs testing to see
4863 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4864 the default if no major problems. Similar behaviour for ASN1_SET_OF
4865 and PKCS12_STACK_OF.
4868 *) When some versions of IIS use the 'NET' form of private key the
4869 key derivation algorithm is different. Normally MD5(password) is
4870 used as a 128 bit RC4 key. In the modified case
4871 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4872 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4873 as the old Netscape_RSA functions except they have an additional
4874 'sgckey' parameter which uses the modified algorithm. Also added
4875 an -sgckey command line option to the rsa utility. Thanks to
4876 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4877 algorithm to openssl-dev.
4880 *) The evp_local.h macros were using 'c.##kname' which resulted in
4881 invalid expansion on some systems (SCO 5.0.5 for example).
4882 Corrected to 'c.kname'.
4883 [Phillip Porch <root@theporch.com>]
4885 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4886 a STACK of email addresses from a certificate or request, these look
4887 in the subject name and the subject alternative name extensions and
4888 omit any duplicate addresses.
4891 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4892 This makes DSA verification about 2 % faster.
4895 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4896 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4897 plus overhead for 1024 bit moduli).
4898 This makes exponentiations about 0.5 % faster for 1024 bit
4899 exponents (as measured by "openssl speed rsa2048").
4902 *) Rename memory handling macros to avoid conflicts with other
4904 Malloc => OPENSSL_malloc
4905 Malloc_locked => OPENSSL_malloc_locked
4906 Realloc => OPENSSL_realloc
4907 Free => OPENSSL_free
4910 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4911 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4914 *) CygWin32 support.
4915 [John Jarvie <jjarvie@newsguy.com>]
4917 *) The type-safe stack code has been rejigged. It is now only compiled
4918 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4919 by default all type-specific stack functions are "#define"d back to
4920 standard stack functions. This results in more streamlined output
4921 but retains the type-safety checking possibilities of the original
4925 *) The STACK code has been cleaned up, and certain type declarations
4926 that didn't make a lot of sense have been brought in line. This has
4927 also involved a cleanup of sorts in safestack.h to more correctly
4928 map type-safe stack functions onto their plain stack counterparts.
4929 This work has also resulted in a variety of "const"ifications of
4930 lots of the code, especially "_cmp" operations which should normally
4931 be prototyped with "const" parameters anyway.
4934 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4935 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4936 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4937 where all of 'md' is used each time the PRNG is used, but 'state'
4938 is used only indexed by a cyclic counter. As entropy may not be
4939 well distributed from the beginning, 'md' is important as a
4940 chaining variable. However, the output function chains only half
4941 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4942 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4943 in all of 'state' being rewritten, with the new values depending
4944 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4947 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4948 the handshake is continued after ssl_verify_cert_chain();
4949 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4950 can lead to 'unexplainable' connection aborts later.
4951 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4953 *) Major EVP API cipher revision.
4954 Add hooks for extra EVP features. This allows various cipher
4955 parameters to be set in the EVP interface. Support added for variable
4956 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4957 setting of RC2 and RC5 parameters.
4959 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4962 Remove lots of duplicated code from the EVP library. For example *every*
4963 cipher init() function handles the 'iv' in the same way according to the
4964 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4965 for CFB and OFB modes they zero ctx->num.
4967 New functionality allows removal of S/MIME code RC2 hack.
4969 Most of the routines have the same form and so can be declared in terms
4972 By shifting this to the top level EVP_CipherInit() it can be removed from
4973 all individual ciphers. If the cipher wants to handle IVs or keys
4974 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4977 Change lots of functions like EVP_EncryptUpdate() to now return a
4978 value: although software versions of the algorithms cannot fail
4979 any installed hardware versions can.
4982 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4983 this option is set, tolerate broken clients that send the negotiated
4984 protocol version number instead of the requested protocol version
4988 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4989 i.e. non-zero for export ciphersuites, zero otherwise.
4990 Previous versions had this flag inverted, inconsistent with
4991 rsa_tmp_cb (..._TMP_RSA_CB).
4992 [Bodo Moeller; problem reported by Amit Chopra]
4994 *) Add missing DSA library text string. Work around for some IIS
4995 key files with invalid SEQUENCE encoding.
4998 *) Add a document (doc/standards.txt) that list all kinds of standards
4999 and so on that are implemented in OpenSSL.
5002 *) Enhance c_rehash script. Old version would mishandle certificates
5003 with the same subject name hash and wouldn't handle CRLs at all.
5004 Added -fingerprint option to crl utility, to support new c_rehash
5008 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5011 *) Fix for SSL server purpose checking. Server checking was
5012 rejecting certificates which had extended key usage present
5013 but no ssl client purpose.
5014 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5016 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5017 is a little unclear about how a blank password is handled.
5018 Since the password in encoded as a BMPString with terminating
5019 double NULL a zero length password would end up as just the
5020 double NULL. However no password at all is different and is
5021 handled differently in the PKCS#12 key generation code. NS
5022 treats a blank password as zero length. MSIE treats it as no
5023 password on export: but it will try both on import. We now do
5024 the same: PKCS12_parse() tries zero length and no password if
5025 the password is set to "" or NULL (NULL is now a valid password:
5026 it wasn't before) as does the pkcs12 application.
5029 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5030 perror when PEM_read_bio_X509_REQ fails, the error message must
5031 be obtained from the error queue.
5034 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5035 it in ERR_remove_state if appropriate, and change ERR_get_state
5036 accordingly to avoid race conditions (this is necessary because
5037 thread_hash is no longer constant once set).
5040 *) Bugfix for linux-elf makefile.one.
5043 *) RSA_get_default_method() will now cause a default
5044 RSA_METHOD to be chosen if one doesn't exist already.
5045 Previously this was only set during a call to RSA_new()
5046 or RSA_new_method(NULL) meaning it was possible for
5047 RSA_get_default_method() to return NULL.
5050 *) Added native name translation to the existing DSO code
5051 that will convert (if the flag to do so is set) filenames
5052 that are sufficiently small and have no path information
5053 into a canonical native form. Eg. "blah" converted to
5054 "libblah.so" or "blah.dll" etc.
5057 *) New function ERR_error_string_n(e, buf, len) which is like
5058 ERR_error_string(e, buf), but writes at most 'len' bytes
5059 including the 0 terminator. For ERR_error_string_n, 'buf'
5061 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5063 *) CONF library reworked to become more general. A new CONF
5064 configuration file reader "class" is implemented as well as a
5065 new functions (NCONF_*, for "New CONF") to handle it. The now
5066 old CONF_* functions are still there, but are reimplemented to
5067 work in terms of the new functions. Also, a set of functions
5068 to handle the internal storage of the configuration data is
5069 provided to make it easier to write new configuration file
5070 reader "classes" (I can definitely see something reading a
5071 configuration file in XML format, for example), called _CONF_*,
5072 or "the configuration storage API"...
5074 The new configuration file reading functions are:
5076 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5077 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5079 NCONF_default, NCONF_WIN32
5081 NCONF_dump_fp, NCONF_dump_bio
5083 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5084 NCONF_new creates a new CONF object. This works in the same way
5085 as other interfaces in OpenSSL, like the BIO interface.
5086 NCONF_dump_* dump the internal storage of the configuration file,
5087 which is useful for debugging. All other functions take the same
5088 arguments as the old CONF_* functions wth the exception of the
5089 first that must be a `CONF *' instead of a `LHASH *'.
5091 To make it easer to use the new classes with the old CONF_* functions,
5092 the function CONF_set_default_method is provided.
5095 *) Add '-tls1' option to 'openssl ciphers', which was already
5096 mentioned in the documentation but had not been implemented.
5097 (This option is not yet really useful because even the additional
5098 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5101 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5102 OpenSSL-based applications) load shared libraries and bind to
5103 them in a portable way.
5104 [Geoff Thorpe, with contributions from Richard Levitte]
5106 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5108 *) Make sure _lrotl and _lrotr are only used with MSVC.
5110 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5111 (the default implementation of RAND_status).
5113 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5114 to '-clrext' (= clear extensions), as intended and documented.
5115 [Bodo Moeller; inconsistency pointed out by Michael Attili
5118 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5119 was larger than the MD block size.
5120 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5122 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5123 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5124 using the passed key: if the passed key was a private key the result
5125 of X509_print(), for example, would be to print out all the private key
5129 *) des_quad_cksum() byte order bug fix.
5130 [Ulf Möller, using the problem description in krb4-0.9.7, where
5131 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5133 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5135 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5137 *) For easily testing in shell scripts whether some command
5138 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5139 returns with exit code 0 iff no command of the given name is available.
5140 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5141 the output goes to stdout and nothing is printed to stderr.
5142 Additional arguments are always ignored.
5144 Since for each cipher there is a command of the same name,
5145 the 'no-cipher' compilation switches can be tested this way.
5147 ('openssl no-XXX' is not able to detect pseudo-commands such
5148 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5151 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5154 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5155 is set; it will be thrown away anyway because each handshake creates
5157 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5158 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5159 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5160 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5163 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5164 'Q' and 'R' lose their special meanings (quit/renegotiate).
5165 This is part of what -quiet does; unlike -quiet, -ign_eof
5166 does not suppress any output.
5169 *) Add compatibility options to the purpose and trust code. The
5170 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5171 accepts a certificate or CA, this was the previous behaviour,
5172 with all the associated security issues.
5174 X509_TRUST_COMPAT is the old trust behaviour: only and
5175 automatically trust self signed roots in certificate store. A
5176 new trust setting X509_TRUST_DEFAULT is used to specify that
5177 a purpose has no associated trust setting and it should instead
5178 use the value in the default purpose.
5181 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5182 and fix a memory leak.
5185 *) In util/mkerr.pl (which implements 'make errors'), preserve
5186 reason strings from the previous version of the .c file, as
5187 the default to have only downcase letters (and digits) in
5188 automatically generated reasons codes is not always appropriate.
5191 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5192 using strerror. Previously, ERR_reason_error_string() returned
5193 library names as reason strings for SYSerr; but SYSerr is a special
5194 case where small numbers are errno values, not library numbers.
5197 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5198 converts DSA parameters into DH parameters. (When creating parameters,
5199 DSA_generate_parameters is used.)
5202 *) Include 'length' (recommended exponent length) in C code generated
5203 by 'openssl dhparam -C'.
5206 *) The second argument to set_label in perlasm was already being used
5207 so couldn't be used as a "file scope" flag. Moved to third argument
5211 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5212 instead of RAND_bytes for encryption IVs and salts.
5215 *) Include RAND_status() into RAND_METHOD instead of implementing
5216 it only for md_rand.c Otherwise replacing the PRNG by calling
5217 RAND_set_rand_method would be impossible.
5220 *) Don't let DSA_generate_key() enter an infinite loop if the random
5221 number generation fails.
5224 *) New 'rand' application for creating pseudo-random output.
5227 *) Added configuration support for Linux/IA64
5228 [Rolf Haberrecker <rolf@suse.de>]
5230 *) Assembler module support for Mingw32.
5233 *) Shared library support for HPUX (in shlib/).
5234 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5236 *) Shared library support for Solaris gcc.
5237 [Lutz Behnke <behnke@trustcenter.de>]
5239 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5241 *) PKCS7_encrypt() was adding text MIME headers twice because they
5242 were added manually and by SMIME_crlf_copy().
5245 *) In bntest.c don't call BN_rand with zero bits argument.
5246 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5248 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5249 case was implemented. This caused BN_div_recp() to fail occasionally.
5252 *) Add an optional second argument to the set_label() in the perl
5253 assembly language builder. If this argument exists and is set
5254 to 1 it signals that the assembler should use a symbol whose
5255 scope is the entire file, not just the current function. This
5256 is needed with MASM which uses the format label:: for this scope.
5257 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5259 *) Change the ASN1 types so they are typedefs by default. Before
5260 almost all types were #define'd to ASN1_STRING which was causing
5261 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5265 *) Change names of new functions to the new get1/get0 naming
5266 convention: After 'get1', the caller owns a reference count
5267 and has to call ..._free; 'get0' returns a pointer to some
5268 data structure without incrementing reference counters.
5269 (Some of the existing 'get' functions increment a reference
5270 counter, some don't.)
5271 Similarly, 'set1' and 'add1' functions increase reference
5272 counters or duplicate objects.
5275 *) Allow for the possibility of temp RSA key generation failure:
5276 the code used to assume it always worked and crashed on failure.
5279 *) Fix potential buffer overrun problem in BIO_printf().
5280 [Ulf Möller, using public domain code by Patrick Powell; problem
5281 pointed out by David Sacerdote <das33@cornell.edu>]
5283 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5284 RAND_egd() and RAND_status(). In the command line application,
5285 the EGD socket can be specified like a seed file using RANDFILE
5289 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5290 Some CAs (e.g. Verisign) distribute certificates in this form.
5293 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5294 list to exclude them. This means that no special compilation option
5295 is needed to use anonymous DH: it just needs to be included in the
5299 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5300 EVP_MD_type. The old functionality is available in a new macro called
5301 EVP_MD_md(). Change code that uses it and update docs.
5304 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5305 where the 'void *' argument is replaced by a function pointer argument.
5306 Previously 'void *' was abused to point to functions, which works on
5307 many platforms, but is not correct. As these functions are usually
5308 called by macros defined in OpenSSL header files, most source code
5309 should work without changes.
5312 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5313 sections with information on -D... compiler switches used for
5314 compiling the library so that applications can see them. To enable
5315 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5316 must be defined. E.g.,
5317 #define OPENSSL_ALGORITHM_DEFINES
5318 #include <openssl/opensslconf.h>
5319 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5320 [Richard Levitte, Ulf and Bodo Möller]
5322 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5326 *) Change the 'other' type in certificate aux info to a STACK_OF
5327 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5328 the required ASN1 format: arbitrary types determined by an OID.
5331 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5332 argument to 'req'. This is not because the function is newer or
5333 better than others it just uses the work 'NEW' in the certificate
5334 request header lines. Some software needs this.
5337 *) Reorganise password command line arguments: now passwords can be
5338 obtained from various sources. Delete the PEM_cb function and make
5339 it the default behaviour: i.e. if the callback is NULL and the
5340 usrdata argument is not NULL interpret it as a null terminated pass
5341 phrase. If usrdata and the callback are NULL then the pass phrase
5342 is prompted for as usual.
5345 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5346 the support is automatically enabled. The resulting binaries will
5347 autodetect the card and use it if present.
5348 [Ben Laurie and Compaq Inc.]
5350 *) Work around for Netscape hang bug. This sends certificate request
5351 and server done in one record. Since this is perfectly legal in the
5352 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5353 the bugs/SSLv3 entry for more info.
5356 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5359 *) Add -rand argument to smime and pkcs12 applications and read/write
5363 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5366 *) Add command line password options to the remaining applications.
5369 *) Bug fix for BN_div_recp() for numerators with an even number of
5373 *) More tests in bntest.c, and changed test_bn output.
5376 *) ./config recognizes MacOS X now.
5379 *) Bug fix for BN_div() when the first words of num and divsor are
5380 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5383 *) Add support for various broken PKCS#8 formats, and command line
5384 options to produce them.
5387 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5388 get temporary BIGNUMs from a BN_CTX.
5391 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5395 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5396 include a #define from the old name to the new. The original intent
5397 was that statically linked binaries could for example just call
5398 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5399 link with digests. This never worked becayse SSLeay_add_all_digests()
5400 and SSLeay_add_all_ciphers() were in the same source file so calling
5401 one would link with the other. They are now in separate source files.
5404 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5407 *) Use a less unusual form of the Miller-Rabin primality test (it used
5408 a binary algorithm for exponentiation integrated into the Miller-Rabin
5409 loop, our standard modexp algorithms are faster).
5412 *) Support for the EBCDIC character set completed.
5413 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5415 *) Source code cleanups: use const where appropriate, eliminate casts,
5416 use void * instead of char * in lhash.
5419 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5420 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5421 this the server could overwrite ephemeral keys that the client
5425 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5426 using 50 iterations of the Rabin-Miller test.
5428 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5429 iterations of the Rabin-Miller test as required by the appendix
5430 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5431 As BN_is_prime_fasttest includes trial division, DSA parameter
5432 generation becomes much faster.
5434 This implies a change for the callback functions in DSA_is_prime
5435 and DSA_generate_parameters: The callback function is called once
5436 for each positive witness in the Rabin-Miller test, not just
5437 occasionally in the inner loop; and the parameters to the
5438 callback function now provide an iteration count for the outer
5439 loop rather than for the current invocation of the inner loop.
5440 DSA_generate_parameters additionally can call the callback
5441 function with an 'iteration count' of -1, meaning that a
5442 candidate has passed the trial division test (when q is generated
5443 from an application-provided seed, trial division is skipped).
5446 *) New function BN_is_prime_fasttest that optionally does trial
5447 division before starting the Rabin-Miller test and has
5448 an additional BN_CTX * argument (whereas BN_is_prime always
5449 has to allocate at least one BN_CTX).
5450 'callback(1, -1, cb_arg)' is called when a number has passed the
5451 trial division stage.
5454 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5458 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5461 *) New function BN_pseudo_rand().
5464 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5465 bignum version of BN_from_montgomery() with the working code from
5466 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5470 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5471 made it impossible to use the same SSL_SESSION data structure in
5472 SSL2 clients in multiple threads.
5475 *) The return value of RAND_load_file() no longer counts bytes obtained
5476 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5477 to seed the PRNG (previously an explicit byte count was required).
5478 [Ulf Möller, Bodo Möller]
5480 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5481 used (char *) instead of (void *) and had casts all over the place.
5484 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5487 *) Retain source code compatibility for BN_prime_checks macro:
5488 BN_is_prime(..., BN_prime_checks, ...) now uses
5489 BN_prime_checks_for_size to determine the appropriate number of
5490 Rabin-Miller iterations.
5493 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5494 DH_CHECK_P_NOT_SAFE_PRIME.
5495 (Check if this is true? OpenPGP calls them "strong".)
5498 *) Merge the functionality of "dh" and "gendh" programs into a new program
5499 "dhparam". The old programs are retained for now but will handle DH keys
5500 (instead of parameters) in future.
5503 *) Make the ciphers, s_server and s_client programs check the return values
5504 when a new cipher list is set.
5507 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5508 ciphers. Before when the 56bit ciphers were enabled the sorting was
5511 The syntax for the cipher sorting has been extended to support sorting by
5512 cipher-strength (using the strength_bits hard coded in the tables).
5513 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5515 Fix a bug in the cipher-command parser: when supplying a cipher command
5516 string with an "undefined" symbol (neither command nor alphanumeric
5517 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5518 an error is flagged.
5520 Due to the strength-sorting extension, the code of the
5521 ssl_create_cipher_list() function was completely rearranged. I hope that
5522 the readability was also increased :-)
5523 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5525 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5526 for the first serial number and places 2 in the serial number file. This
5527 avoids problems when the root CA is created with serial number zero and
5528 the first user certificate has the same issuer name and serial number
5532 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5533 the new code. Add documentation for this stuff.
5536 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5537 X509_*() to X509at_*() on the grounds that they don't handle X509
5538 structures and behave in an analagous way to the X509v3 functions:
5539 they shouldn't be called directly but wrapper functions should be used
5542 So we also now have some wrapper functions that call the X509at functions
5543 when passed certificate requests. (TO DO: similar things can be done with
5544 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5545 things. Some of these need some d2i or i2d and print functionality
5546 because they handle more complex structures.)
5549 *) Add missing #ifndefs that caused missing symbols when building libssl
5550 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5551 NO_RSA in ssl/s2*.c.
5552 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5554 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5555 has a return value which indicates the quality of the random data
5556 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5557 error queue. New function RAND_pseudo_bytes() generates output that is
5558 guaranteed to be unique but not unpredictable. RAND_add is like
5559 RAND_seed, but takes an extra argument for an entropy estimate
5560 (RAND_seed always assumes full entropy).
5563 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5564 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5565 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5566 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5567 false-positive rate of at most 2^-80 for random input.
5570 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5573 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5574 in the 0.9.5 release), this returns the chain
5575 from an X509_CTX structure with a dup of the stack and all
5576 the X509 reference counts upped: so the stack will exist
5577 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5580 Also make SSL_SESSION_print() print out the verify return
5584 *) Add manpage for the pkcs12 command. Also change the default
5585 behaviour so MAC iteration counts are used unless the new
5586 -nomaciter option is used. This improves file security and
5587 only older versions of MSIE (4.0 for example) need it.
5590 *) Honor the no-xxx Configure options when creating .DEF files.
5593 *) Add PKCS#10 attributes to field table: challengePassword,
5594 unstructuredName and unstructuredAddress. These are taken from
5595 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5596 international characters are used.
5598 More changes to X509_ATTRIBUTE code: allow the setting of types
5599 based on strings. Remove the 'loc' parameter when adding
5600 attributes because these will be a SET OF encoding which is sorted
5604 *) Initial changes to the 'req' utility to allow request generation
5605 automation. This will allow an application to just generate a template
5606 file containing all the field values and have req construct the
5609 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5610 used all over the place including certificate requests and PKCS#7
5611 structures. They are currently handled manually where necessary with
5612 some primitive wrappers for PKCS#7. The new functions behave in a
5613 manner analogous to the X509 extension functions: they allow
5614 attributes to be looked up by NID and added.
5616 Later something similar to the X509V3 code would be desirable to
5617 automatically handle the encoding, decoding and printing of the
5618 more complex types. The string types like challengePassword can
5619 be handled by the string table functions.
5621 Also modified the multi byte string table handling. Now there is
5622 a 'global mask' which masks out certain types. The table itself
5623 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5624 is useful when for example there is only one permissible type
5625 (as in countryName) and using the mask might result in no valid
5629 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5630 SSL_get_peer_finished to allow applications to obtain the latest
5631 Finished messages sent to the peer or expected from the peer,
5632 respectively. (SSL_get_peer_finished is usually the Finished message
5633 actually received from the peer, otherwise the protocol will be aborted.)
5635 As the Finished message are message digests of the complete handshake
5636 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5637 be used for external authentication procedures when the authentication
5638 provided by SSL/TLS is not desired or is not enough.
5641 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5642 the host supports BWX extension and if Compaq C is present on the
5643 $PATH. Just exploiting of the BWX extension results in 20-30%
5644 performance kick for some algorithms, e.g. DES and RC4 to mention
5645 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5649 *) Add support for MS "fast SGC". This is arguably a violation of the
5650 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5651 weak crypto and after checking the certificate is SGC a second one
5652 with strong crypto. MS SGC stops the first handshake after receiving
5653 the server certificate message and sends a second client hello. Since
5654 a server will typically do all the time consuming operations before
5655 expecting any further messages from the client (server key exchange
5656 is the most expensive) there is little difference between the two.
5658 To get OpenSSL to support MS SGC we have to permit a second client
5659 hello message after we have sent server done. In addition we have to
5660 reset the MAC if we do get this second client hello.
5663 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5664 if a DER encoded private key is RSA or DSA traditional format. Changed
5665 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5666 format DER encoded private key. Newer code should use PKCS#8 format which
5667 has the key type encoded in the ASN1 structure. Added DER private key
5668 support to pkcs8 application.
5671 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5672 ciphersuites has been selected (as required by the SSL 3/TLS 1
5673 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5674 is set, we interpret this as a request to violate the specification
5675 (the worst that can happen is a handshake failure, and 'correct'
5676 behaviour would result in a handshake failure anyway).
5679 *) In SSL_CTX_add_session, take into account that there might be multiple
5680 SSL_SESSION structures with the same session ID (e.g. when two threads
5681 concurrently obtain them from an external cache).
5682 The internal cache can handle only one SSL_SESSION with a given ID,
5683 so if there's a conflict, we now throw out the old one to achieve
5687 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5688 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5689 some routines that use cipher OIDs: some ciphers do not have OIDs
5690 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5694 *) Simplify the trust setting structure and code. Now we just have
5695 two sequences of OIDs for trusted and rejected settings. These will
5696 typically have values the same as the extended key usage extension
5697 and any application specific purposes.
5699 The trust checking code now has a default behaviour: it will just
5700 check for an object with the same NID as the passed id. Functions can
5701 be provided to override either the default behaviour or the behaviour
5702 for a given id. SSL client, server and email already have functions
5703 in place for compatibility: they check the NID and also return "trusted"
5704 if the certificate is self signed.
5707 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5708 traditional format into an EVP_PKEY structure.
5711 *) Add a password callback function PEM_cb() which either prompts for
5712 a password if usr_data is NULL or otherwise assumes it is a null
5713 terminated password. Allow passwords to be passed on command line
5714 environment or config files in a few more utilities.
5717 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5718 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5719 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5720 Update documentation.
5723 *) Support for ASN1 "NULL" type. This could be handled before by using
5724 ASN1_TYPE but there wasn't any function that would try to read a NULL
5725 and produce an error if it couldn't. For compatibility we also have
5726 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5727 don't allocate anything because they don't need to.
5730 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5732 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5734 *) Rebuild of the memory allocation routines used by OpenSSL code and
5735 possibly others as well. The purpose is to make an interface that
5736 provide hooks so anyone can build a separate set of allocation and
5737 deallocation routines to be used by OpenSSL, for example memory
5738 pool implementations, or something else, which was previously hard
5739 since Malloc(), Realloc() and Free() were defined as macros having
5740 the values malloc, realloc and free, respectively (except for Win32
5741 compilations). The same is provided for memory debugging code.
5742 OpenSSL already comes with functionality to find memory leaks, but
5743 this gives people a chance to debug other memory problems.
5745 With these changes, a new set of functions and macros have appeared:
5747 CRYPTO_set_mem_debug_functions() [F]
5748 CRYPTO_get_mem_debug_functions() [F]
5749 CRYPTO_dbg_set_options() [F]
5750 CRYPTO_dbg_get_options() [F]
5751 CRYPTO_malloc_debug_init() [M]
5753 The memory debug functions are NULL by default, unless the library
5754 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5755 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5756 gives the standard debugging functions that come with OpenSSL) or
5757 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5758 provided by the library user) must be used. When the standard
5759 debugging functions are used, CRYPTO_dbg_set_options can be used to
5760 request additional information:
5761 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5762 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5764 Also, things like CRYPTO_set_mem_functions will always give the
5765 expected result (the new set of functions is used for allocation
5766 and deallocation) at all times, regardless of platform and compiler
5769 To finish it up, some functions that were never use in any other
5770 way than through macros have a new API and new semantic:
5773 CRYPTO_dbg_realloc()
5776 All macros of value have retained their old syntax.
5777 [Richard Levitte and Bodo Moeller]
5779 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5780 ordering of SMIMECapabilities wasn't in "strength order" and there
5781 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5785 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5786 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5787 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5789 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5790 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5791 functionality to handle multipart/signed properly) and a utility
5792 called 'smime' to call all this stuff. This is based on code I
5793 originally wrote for Celo who have kindly allowed it to be
5794 included in OpenSSL.
5797 *) Add variants des_set_key_checked and des_set_key_unchecked of
5798 des_set_key (aka des_key_sched). Global variable des_check_key
5799 decides which of these is called by des_set_key; this way
5800 des_check_key behaves as it always did, but applications and
5801 the library itself, which was buggy for des_check_key == 1,
5802 have a cleaner way to pick the version they need.
5805 *) New function PKCS12_newpass() which changes the password of a
5809 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5810 dynamic mix. In both cases the ids can be used as an index into the
5811 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5812 functions so they accept a list of the field values and the
5813 application doesn't need to directly manipulate the X509_TRUST
5817 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5821 *) Modify the way the V3 extension code looks up extensions. This now
5822 works in a similar way to the object code: we have some "standard"
5823 extensions in a static table which is searched with OBJ_bsearch()
5824 and the application can add dynamic ones if needed. The file
5825 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5826 updated whenever a new extension is added to the core code and kept
5827 in ext_nid order. There is a simple program 'tabtest.c' which checks
5828 this. New extensions are not added too often so this file can readily
5829 be maintained manually.
5831 There are two big advantages in doing things this way. The extensions
5832 can be looked up immediately and no longer need to be "added" using
5833 X509V3_add_standard_extensions(): this function now does nothing.
5834 [Side note: I get *lots* of email saying the extension code doesn't
5835 work because people forget to call this function]
5836 Also no dynamic allocation is done unless new extensions are added:
5837 so if we don't add custom extensions there is no need to call
5838 X509V3_EXT_cleanup().
5841 *) Modify enc utility's salting as follows: make salting the default. Add a
5842 magic header, so unsalted files fail gracefully instead of just decrypting
5843 to garbage. This is because not salting is a big security hole, so people
5844 should be discouraged from doing it.
5847 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5848 digest to be passed on the command line but it only used this
5849 parameter when signing a certificate. Modified so all relevant
5850 operations are affected by the digest parameter including the
5851 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5852 DSA key was used because it didn't fix the digest.
5855 *) Initial certificate chain verify code. Currently tests the untrusted
5856 certificates for consistency with the verify purpose (which is set
5857 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5859 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5860 this is because it will reject chains with invalid extensions whereas
5861 every previous version of OpenSSL and SSLeay made no checks at all.
5863 Trust code: checks the root CA for the relevant trust settings. Trust
5864 settings have an initial value consistent with the verify purpose: e.g.
5865 if the verify purpose is for SSL client use it expects the CA to be
5866 trusted for SSL client use. However the default value can be changed to
5867 permit custom trust settings: one example of this would be to only trust
5868 certificates from a specific "secure" set of CAs.
5870 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5871 which should be used for version portability: especially since the
5872 verify structure is likely to change more often now.
5874 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5875 to set them. If not set then assume SSL clients will verify SSL servers
5878 Two new options to the verify program: -untrusted allows a set of
5879 untrusted certificates to be passed in and -purpose which sets the
5880 intended purpose of the certificate. If a purpose is set then the
5881 new chain verify code is used to check extension consistency.
5884 *) Support for the authority information access extension.
5887 *) Modify RSA and DSA PEM read routines to transparently handle
5888 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5889 public keys in a format compatible with certificate
5890 SubjectPublicKeyInfo structures. Unfortunately there were already
5891 functions called *_PublicKey_* which used various odd formats so
5892 these are retained for compatibility: however the DSA variants were
5893 never in a public release so they have been deleted. Changed dsa/rsa
5894 utilities to handle the new format: note no releases ever handled public
5895 keys so we should be OK.
5897 The primary motivation for this change is to avoid the same fiasco
5898 that dogs private keys: there are several incompatible private key
5899 formats some of which are standard and some OpenSSL specific and
5900 require various evil hacks to allow partial transparent handling and
5901 even then it doesn't work with DER formats. Given the option anything
5902 other than PKCS#8 should be dumped: but the other formats have to
5903 stay in the name of compatibility.
5905 With public keys and the benefit of hindsight one standard format
5906 is used which works with EVP_PKEY, RSA or DSA structures: though
5907 it clearly returns an error if you try to read the wrong kind of key.
5909 Added a -pubkey option to the 'x509' utility to output the public key.
5910 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5911 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5912 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5913 that do the same as the EVP_PKEY_assign_*() except they up the
5914 reference count of the added key (they don't "swallow" the
5918 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5919 CRLs would fail if the file contained no certificates or no CRLs:
5920 added a new function to read in both types and return the number
5921 read: this means that if none are read it will be an error. The
5922 DER versions of the certificate and CRL reader would always fail
5923 because it isn't possible to mix certificates and CRLs in DER format
5924 without choking one or the other routine. Changed this to just read
5925 a certificate: this is the best we can do. Also modified the code
5926 in apps/verify.c to take notice of return codes: it was previously
5927 attempting to read in certificates from NULL pointers and ignoring
5928 any errors: this is one reason why the cert and CRL reader seemed
5929 to work. It doesn't check return codes from the default certificate
5930 routines: these may well fail if the certificates aren't installed.
5933 *) Code to support otherName option in GeneralName.
5936 *) First update to verify code. Change the verify utility
5937 so it warns if it is passed a self signed certificate:
5938 for consistency with the normal behaviour. X509_verify
5939 has been modified to it will now verify a self signed
5940 certificate if *exactly* the same certificate appears
5941 in the store: it was previously impossible to trust a
5942 single self signed certificate. This means that:
5943 openssl verify ss.pem
5944 now gives a warning about a self signed certificate but
5945 openssl verify -CAfile ss.pem ss.pem
5949 *) For servers, store verify_result in SSL_SESSION data structure
5950 (and add it to external session representation).
5951 This is needed when client certificate verifications fails,
5952 but an application-provided verification callback (set by
5953 SSL_CTX_set_cert_verify_callback) allows accepting the session
5954 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5955 but returns 1): When the session is reused, we have to set
5956 ssl->verify_result to the appropriate error code to avoid
5958 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5960 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5961 case in PKCS7_dataInit() where the signed PKCS7 structure
5962 didn't contain any existing data because it was being created.
5963 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5965 *) Add a salt to the key derivation routines in enc.c. This
5966 forms the first 8 bytes of the encrypted file. Also add a
5967 -S option to allow a salt to be input on the command line.
5970 *) New function X509_cmp(). Oddly enough there wasn't a function
5971 to compare two certificates. We do this by working out the SHA1
5972 hash and comparing that. X509_cmp() will be needed by the trust
5976 *) SSL_get1_session() is like SSL_get_session(), but increments
5977 the reference count in the SSL_SESSION returned.
5978 [Geoff Thorpe <geoff@eu.c2.net>]
5980 *) Fix for 'req': it was adding a null to request attributes.
5981 Also change the X509_LOOKUP and X509_INFO code to handle
5982 certificate auxiliary information.
5985 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5989 *) Add the possibility to add extra information to the memory leak
5990 detecting output, to form tracebacks, showing from where each
5991 allocation was originated: CRYPTO_push_info("constant string") adds
5992 the string plus current file name and line number to a per-thread
5993 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5994 is like calling CYRPTO_pop_info() until the stack is empty.
5995 Also updated memory leak detection code to be multi-thread-safe.
5998 *) Add options -text and -noout to pkcs7 utility and delete the
5999 encryption options which never did anything. Update docs.
6002 *) Add options to some of the utilities to allow the pass phrase
6003 to be included on either the command line (not recommended on
6004 OSes like Unix) or read from the environment. Update the
6005 manpages and fix a few bugs.
6008 *) Add a few manpages for some of the openssl commands.
6011 *) Fix the -revoke option in ca. It was freeing up memory twice,
6012 leaking and not finding already revoked certificates.
6015 *) Extensive changes to support certificate auxiliary information.
6016 This involves the use of X509_CERT_AUX structure and X509_AUX
6017 functions. An X509_AUX function such as PEM_read_X509_AUX()
6018 can still read in a certificate file in the usual way but it
6019 will also read in any additional "auxiliary information". By
6020 doing things this way a fair degree of compatibility can be
6021 retained: existing certificates can have this information added
6022 using the new 'x509' options.
6024 Current auxiliary information includes an "alias" and some trust
6025 settings. The trust settings will ultimately be used in enhanced
6026 certificate chain verification routines: currently a certificate
6027 can only be trusted if it is self signed and then it is trusted
6031 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6032 The problem was that one of the replacement routines had not been working
6033 since SSLeay releases. For now the offending routine has been replaced
6034 with non-optimised assembler. Even so, this now gives around 95%
6035 performance improvement for 1024 bit RSA signs.
6038 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6039 handling. Most clients have the effective key size in bits equal to
6040 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6041 A few however don't do this and instead use the size of the decrypted key
6042 to determine the RC2 key length and the AlgorithmIdentifier to determine
6043 the effective key length. In this case the effective key length can still
6044 be 40 bits but the key length can be 168 bits for example. This is fixed
6045 by manually forcing an RC2 key into the EVP_PKEY structure because the
6046 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6047 the key length and effective key length are equal.
6050 *) Add a bunch of functions that should simplify the creation of
6051 X509_NAME structures. Now you should be able to do:
6052 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6053 and have it automatically work out the correct field type and fill in
6054 the structures. The more adventurous can try:
6055 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6056 and it will (hopefully) work out the correct multibyte encoding.
6059 *) Change the 'req' utility to use the new field handling and multibyte
6060 copy routines. Before the DN field creation was handled in an ad hoc
6061 way in req, ca, and x509 which was rather broken and didn't support
6062 BMPStrings or UTF8Strings. Since some software doesn't implement
6063 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6064 using the dirstring_type option. See the new comment in the default
6065 openssl.cnf for more info.
6068 *) Make crypto/rand/md_rand.c more robust:
6069 - Assure unique random numbers after fork().
6070 - Make sure that concurrent threads access the global counter and
6071 md serializably so that we never lose entropy in them
6072 or use exactly the same state in multiple threads.
6073 Access to the large state is not always serializable because
6074 the additional locking could be a performance killer, and
6075 md should be large enough anyway.
6078 *) New file apps/app_rand.c with commonly needed functionality
6079 for handling the random seed file.
6081 Use the random seed file in some applications that previously did not:
6083 dsaparam -genkey (which also ignored its '-rand' option),
6086 x509 (when signing).
6087 Except on systems with /dev/urandom, it is crucial to have a random
6088 seed file at least for key creation, DSA signing, and for DH exchanges;
6089 for RSA signatures we could do without one.
6091 gendh and gendsa (unlike genrsa) used to read only the first byte
6092 of each file listed in the '-rand' option. The function as previously
6093 found in genrsa is now in app_rand.c and is used by all programs
6094 that support '-rand'.
6097 *) In RAND_write_file, use mode 0600 for creating files;
6098 don't just chmod when it may be too late.
6101 *) Report an error from X509_STORE_load_locations
6102 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6105 *) New function ASN1_mbstring_copy() this copies a string in either
6106 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6107 into an ASN1_STRING type. A mask of permissible types is passed
6108 and it chooses the "minimal" type to use or an error if not type
6112 *) Add function equivalents to the various macros in asn1.h. The old
6113 macros are retained with an M_ prefix. Code inside the library can
6114 use the M_ macros. External code (including the openssl utility)
6115 should *NOT* in order to be "shared library friendly".
6118 *) Add various functions that can check a certificate's extensions
6119 to see if it usable for various purposes such as SSL client,
6120 server or S/MIME and CAs of these types. This is currently
6121 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6122 verification. Also added a -purpose flag to x509 utility to
6123 print out all the purposes.
6126 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6130 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6131 for, obtain and decode and extension and obtain its critical flag.
6132 This allows all the necessary extension code to be handled in a
6133 single function call.
6136 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6137 platforms. See crypto/rc4/rc4_enc.c for further details.
6140 *) New -noout option to asn1parse. This causes no output to be produced
6141 its main use is when combined with -strparse and -out to extract data
6142 from a file (which may not be in ASN.1 format).
6145 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6146 when producing the local key id.
6147 [Richard Levitte <levitte@stacken.kth.se>]
6149 *) New option -dhparam in s_server. This allows a DH parameter file to be
6150 stated explicitly. If it is not stated then it tries the first server
6151 certificate file. The previous behaviour hard coded the filename
6155 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6156 a public key to be input or output. For example:
6157 openssl rsa -in key.pem -pubout -out pubkey.pem
6158 Also added necessary DSA public key functions to handle this.
6161 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6162 in the message. This was handled by allowing
6163 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6164 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6166 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6167 to the end of the strings whereas this didn't. This would cause problems
6168 if strings read with d2i_ASN1_bytes() were later modified.
6169 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6171 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6172 data and it contains EOF it will end up returning an error. This is
6173 caused by input 46 bytes long. The cause is due to the way base64
6174 BIOs find the start of base64 encoded data. They do this by trying a
6175 trial decode on each line until they find one that works. When they
6176 do a flag is set and it starts again knowing it can pass all the
6177 data directly through the decoder. Unfortunately it doesn't reset
6178 the context it uses. This means that if EOF is reached an attempt
6179 is made to pass two EOFs through the context and this causes the
6180 resulting error. This can also cause other problems as well. As is
6181 usual with these problems it takes *ages* to find and the fix is
6182 trivial: move one line.
6183 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6185 *) Ugly workaround to get s_client and s_server working under Windows. The
6186 old code wouldn't work because it needed to select() on sockets and the
6187 tty (for keypresses and to see if data could be written). Win32 only
6188 supports select() on sockets so we select() with a 1s timeout on the
6189 sockets and then see if any characters are waiting to be read, if none
6190 are present then we retry, we also assume we can always write data to
6191 the tty. This isn't nice because the code then blocks until we've
6192 received a complete line of data and it is effectively polling the
6193 keyboard at 1s intervals: however it's quite a bit better than not
6194 working at all :-) A dedicated Windows application might handle this
6195 with an event loop for example.
6198 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6199 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6200 will be called when RSA_sign() and RSA_verify() are used. This is useful
6201 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6202 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6203 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6204 This necessitated the support of an extra signature type NID_md5_sha1
6205 for SSL signatures and modifications to the SSL library to use it instead
6206 of calling RSA_public_decrypt() and RSA_private_encrypt().
6209 *) Add new -verify -CAfile and -CApath options to the crl program, these
6210 will lookup a CRL issuers certificate and verify the signature in a
6211 similar way to the verify program. Tidy up the crl program so it
6212 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6213 less strict. It will now permit CRL extensions even if it is not
6214 a V2 CRL: this will allow it to tolerate some broken CRLs.
6217 *) Initialize all non-automatic variables each time one of the openssl
6218 sub-programs is started (this is necessary as they may be started
6219 multiple times from the "OpenSSL>" prompt).
6220 [Lennart Bang, Bodo Moeller]
6222 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6223 removing all other RSA functionality (this is what NO_RSA does). This
6224 is so (for example) those in the US can disable those operations covered
6225 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6229 *) Non-copying interface to BIO pairs.
6230 (still largely untested)
6233 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6234 ASCII string. This was handled independently in various places before.
6237 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6238 UTF8 strings a character at a time.
6241 *) Use client_version from client hello to select the protocol
6242 (s23_srvr.c) and for RSA client key exchange verification
6243 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6246 *) Add various utility functions to handle SPKACs, these were previously
6247 handled by poking round in the structure internals. Added new function
6248 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6249 print, verify and generate SPKACs. Based on an original idea from
6250 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6253 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6256 *) Allow the config file extension section to be overwritten on the
6257 command line. Based on an original idea from Massimiliano Pala
6258 <madwolf@comune.modena.it>. The new option is called -extensions
6259 and can be applied to ca, req and x509. Also -reqexts to override
6260 the request extensions in req and -crlexts to override the crl extensions
6264 *) Add new feature to the SPKAC handling in ca. Now you can include
6265 the same field multiple times by preceding it by "XXXX." for example:
6268 this is the same syntax as used in the req config file.
6271 *) Allow certificate extensions to be added to certificate requests. These
6272 are specified in a 'req_extensions' option of the req section of the
6273 config file. They can be printed out with the -text option to req but
6274 are otherwise ignored at present.
6277 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6278 data read consists of only the final block it would not decrypted because
6279 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6280 A misplaced 'break' also meant the decrypted final block might not be
6281 copied until the next read.
6284 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6285 a few extra parameters to the DH structure: these will be useful if
6286 for example we want the value of 'q' or implement X9.42 DH.
6289 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6290 provides hooks that allow the default DSA functions or functions on a
6291 "per key" basis to be replaced. This allows hardware acceleration and
6292 hardware key storage to be handled without major modification to the
6293 library. Also added low level modexp hooks and CRYPTO_EX structure and
6294 associated functions.
6297 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6298 as "read only": it can't be written to and the buffer it points to will
6299 not be freed. Reading from a read only BIO is much more efficient than
6300 a normal memory BIO. This was added because there are several times when
6301 an area of memory needs to be read from a BIO. The previous method was
6302 to create a memory BIO and write the data to it, this results in two
6303 copies of the data and an O(n^2) reading algorithm. There is a new
6304 function BIO_new_mem_buf() which creates a read only memory BIO from
6305 an area of memory. Also modified the PKCS#7 routines to use read only
6309 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6310 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6311 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6312 but a retry condition occured while trying to read the rest.
6315 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6316 NID_pkcs7_encrypted by default: this was wrong since this should almost
6317 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6318 the encrypted data type: this is a more sensible place to put it and it
6319 allows the PKCS#12 code to be tidied up that duplicated this
6323 *) Changed obj_dat.pl script so it takes its input and output files on
6324 the command line. This should avoid shell escape redirection problems
6328 *) Initial support for certificate extension requests, these are included
6329 in things like Xenroll certificate requests. Included functions to allow
6330 extensions to be obtained and added.
6333 *) -crlf option to s_client and s_server for sending newlines as
6334 CRLF (as required by many protocols).
6337 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6339 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6340 [Ralf S. Engelschall]
6342 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6343 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6345 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6349 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6350 DH parameters/keys (q is lost during that conversion, but the resulting
6351 DH parameters contain its length).
6353 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6354 much faster than DH_generate_parameters (which creates parameters
6355 where p = 2*q + 1), and also the smaller q makes DH computations
6356 much more efficient (160-bit exponentiation instead of 1024-bit
6357 exponentiation); so this provides a convenient way to support DHE
6358 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6359 utter importance to use
6360 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6362 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6363 when such DH parameters are used, because otherwise small subgroup
6364 attacks may become possible!
6367 *) Avoid memory leak in i2d_DHparams.
6370 *) Allow the -k option to be used more than once in the enc program:
6371 this allows the same encrypted message to be read by multiple recipients.
6374 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6375 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6376 it will always use the numerical form of the OID, even if it has a short
6380 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6381 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6382 otherwise bn_mod_exp was called. In the case of hardware keys for example
6383 no private key components need be present and it might store extra data
6384 in the RSA structure, which cannot be accessed from bn_mod_exp.
6385 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6386 private key operations.
6389 *) Added support for SPARC Linux.
6392 *) pem_password_cb function type incompatibly changed from
6393 typedef int pem_password_cb(char *buf, int size, int rwflag);
6395 ....(char *buf, int size, int rwflag, void *userdata);
6396 so that applications can pass data to their callbacks:
6397 The PEM[_ASN1]_{read,write}... functions and macros now take an
6398 additional void * argument, which is just handed through whenever
6399 the password callback is called.
6400 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6402 New function SSL_CTX_set_default_passwd_cb_userdata.
6404 Compatibility note: As many C implementations push function arguments
6405 onto the stack in reverse order, the new library version is likely to
6406 interoperate with programs that have been compiled with the old
6407 pem_password_cb definition (PEM_whatever takes some data that
6408 happens to be on the stack as its last argument, and the callback
6409 just ignores this garbage); but there is no guarantee whatsoever that
6412 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6413 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6414 problems not only on Windows, but also on some Unix platforms.
6415 To avoid problematic command lines, these definitions are now in an
6416 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6417 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6420 *) MIPS III/IV assembler module is reimplemented.
6423 *) More DES library cleanups: remove references to srand/rand and
6424 delete an unused file.
6427 *) Add support for the the free Netwide assembler (NASM) under Win32,
6428 since not many people have MASM (ml) and it can be hard to obtain.
6429 This is currently experimental but it seems to work OK and pass all
6430 the tests. Check out INSTALL.W32 for info.
6433 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6434 without temporary keys kept an extra copy of the server key,
6435 and connections with temporary keys did not free everything in case
6439 *) New function RSA_check_key and new openssl rsa option -check
6440 for verifying the consistency of RSA keys.
6441 [Ulf Moeller, Bodo Moeller]
6443 *) Various changes to make Win32 compile work:
6444 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6445 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6446 comparison" warnings.
6447 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6450 *) Add a debugging option to PKCS#5 v2 key generation function: when
6451 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6452 derived keys are printed to stderr.
6455 *) Copy the flags in ASN1_STRING_dup().
6456 [Roman E. Pavlov <pre@mo.msk.ru>]
6458 *) The x509 application mishandled signing requests containing DSA
6459 keys when the signing key was also DSA and the parameters didn't match.
6461 It was supposed to omit the parameters when they matched the signing key:
6462 the verifying software was then supposed to automatically use the CA's
6463 parameters if they were absent from the end user certificate.
6465 Omitting parameters is no longer recommended. The test was also
6466 the wrong way round! This was probably due to unusual behaviour in
6467 EVP_cmp_parameters() which returns 1 if the parameters match.
6468 This meant that parameters were omitted when they *didn't* match and
6469 the certificate was useless. Certificates signed with 'ca' didn't have
6471 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6473 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6474 The interface is as follows:
6475 Applications can use
6476 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6477 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6478 "off" is now the default.
6479 The library internally uses
6480 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6481 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6482 to disable memory-checking temporarily.
6484 Some inconsistent states that previously were possible (and were
6485 even the default) are now avoided.
6487 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6488 with each memory chunk allocated; this is occasionally more helpful
6489 than just having a counter.
6491 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6493 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6497 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6498 which largely parallels "options", but is for changing API behaviour,
6499 whereas "options" are about protocol behaviour.
6500 Initial "mode" flags are:
6502 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6503 a single record has been written.
6504 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6505 retries use the same buffer location.
6506 (But all of the contents must be
6510 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6513 *) Fix problems with no-hmac etc.
6514 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6516 *) New functions RSA_get_default_method(), RSA_set_method() and
6517 RSA_get_method(). These allows replacement of RSA_METHODs without having
6518 to mess around with the internals of an RSA structure.
6521 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6522 Also really enable memory leak checks in openssl.c and in some
6524 [Chad C. Mulligan, Bodo Moeller]
6526 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6527 up the length of negative integers. This has now been simplified to just
6528 store the length when it is first determined and use it later, rather
6529 than trying to keep track of where data is copied and updating it to
6531 [Steve Henson, reported by Brien Wheeler
6532 <bwheeler@authentica-security.com>]
6534 *) Add a new function PKCS7_signatureVerify. This allows the verification
6535 of a PKCS#7 signature but with the signing certificate passed to the
6536 function itself. This contrasts with PKCS7_dataVerify which assumes the
6537 certificate is present in the PKCS#7 structure. This isn't always the
6538 case: certificates can be omitted from a PKCS#7 structure and be
6539 distributed by "out of band" means (such as a certificate database).
6542 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6543 function prototypes in pem.h, also change util/mkdef.pl to add the
6544 necessary function names.
6547 *) mk1mf.pl (used by Windows builds) did not properly read the
6548 options set by Configure in the top level Makefile, and Configure
6549 was not even able to write more than one option correctly.
6550 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6553 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6554 file to be loaded from a BIO or FILE pointer. The BIO version will
6555 for example allow memory BIOs to contain config info.
6558 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6559 Whoever hopes to achieve shared-library compatibility across versions
6560 must use this, not the compile-time macro.
6561 (Exercise 0.9.4: Which is the minimum library version required by
6563 Note: All this applies only to multi-threaded programs, others don't
6567 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6568 through a BIO pair triggered the default case, i.e.
6569 SSLerr(...,SSL_R_UNKNOWN_STATE).
6572 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6573 can use the SSL library even if none of the specific BIOs is
6577 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6578 for the encoded length.
6579 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6581 *) Add initial documentation of the X509V3 functions.
6584 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6585 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6586 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6587 secure PKCS#8 private key format with a high iteration count.
6590 *) Fix determination of Perl interpreter: A perl or perl5
6591 _directory_ in $PATH was also accepted as the interpreter.
6592 [Ralf S. Engelschall]
6594 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6595 wrong with it but it was very old and did things like calling
6596 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6600 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6601 to use the new extension code.
6604 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6605 with macros. This should make it easier to change their form, add extra
6606 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6610 *) Add to configuration table a new entry that can specify an alternative
6611 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6612 according to Mark Crispin <MRC@Panda.COM>.
6616 *) DES CBC did not update the IV. Weird.
6619 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6620 Changing the behaviour of the former might break existing programs --
6621 where IV updating is needed, des_ncbc_encrypt can be used.
6624 *) When bntest is run from "make test" it drives bc to check its
6625 calculations, as well as internally checking them. If an internal check
6626 fails, it needs to cause bc to give a non-zero result or make test carries
6627 on without noticing the failure. Fixed.
6630 *) DES library cleanups.
6633 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6634 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6635 ciphers. NOTE: although the key derivation function has been verified
6636 against some published test vectors it has not been extensively tested
6637 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6641 *) Instead of "mkdir -p", which is not fully portable, use new
6642 Perl script "util/mkdir-p.pl".
6645 *) Rewrite the way password based encryption (PBE) is handled. It used to
6646 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6647 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6648 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6649 the 'parameter' field of the AlgorithmIdentifier is passed to the
6650 underlying key generation function so it must do its own ASN1 parsing.
6651 This has also changed the EVP_PBE_CipherInit() function which now has a
6652 'parameter' argument instead of literal salt and iteration count values
6653 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6656 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6657 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6658 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6659 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6660 value was just used as a "magic string" and not used directly its
6661 value doesn't matter.
6664 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6668 *) "linux-sparc64" configuration (ultrapenguin).
6669 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6670 "linux-sparc" configuration.
6671 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6673 *) config now generates no-xxx options for missing ciphers.
6676 *) Support the EBCDIC character set (work in progress).
6677 File ebcdic.c not yet included because it has a different license.
6678 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6680 *) Support BS2000/OSD-POSIX.
6681 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6683 *) Make callbacks for key generation use void * instead of char *.
6686 *) Make S/MIME samples compile (not yet tested).
6689 *) Additional typesafe stacks.
6692 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6696 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6698 *) New configuration variant "sco5-gcc".
6700 *) Updated some demos.
6701 [Sean O Riordain, Wade Scholine]
6703 *) Add missing BIO_free at exit of pkcs12 application.
6706 *) Fix memory leak in conf.c.
6709 *) Updates for Win32 to assembler version of MD5.
6712 *) Set #! path to perl in apps/der_chop to where we found it
6713 instead of using a fixed path.
6716 *) SHA library changes for irix64-mips4-cc.
6719 *) Improvements for VMS support.
6723 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6725 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6726 This also avoids the problems with SC4.2 and unpatched SC5.
6727 [Andy Polyakov <appro@fy.chalmers.se>]
6729 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6730 These are required because of the typesafe stack would otherwise break
6731 existing code. If old code used a structure member which used to be STACK
6732 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6733 sk_num or sk_value it would produce an error because the num, data members
6734 are not present in STACK_OF. Now it just produces a warning. sk_set
6735 replaces the old method of assigning a value to sk_value
6736 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6737 that does this will no longer work (and should use sk_set instead) but
6738 this could be regarded as a "questionable" behaviour anyway.
6741 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6742 correctly handle encrypted S/MIME data.
6745 *) Change type of various DES function arguments from des_cblock
6746 (which means, in function argument declarations, pointer to char)
6747 to des_cblock * (meaning pointer to array with 8 char elements),
6748 which allows the compiler to do more typechecking; it was like
6749 that back in SSLeay, but with lots of ugly casts.
6751 Introduce new type const_des_cblock.
6754 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6755 problems: find RecipientInfo structure that matches recipient certificate
6756 and initialise the ASN1 structures properly based on passed cipher.
6759 *) Belatedly make the BN tests actually check the results.
6762 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6763 to and from BNs: it was completely broken. New compilation option
6764 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6765 key elements as negative integers.
6768 *) Reorganize and speed up MD5.
6769 [Andy Polyakov <appro@fy.chalmers.se>]
6772 [Richard Levitte <richard@levitte.org>]
6774 *) New option -out to asn1parse to allow the parsed structure to be
6775 output to a file. This is most useful when combined with the -strparse
6776 option to examine the output of things like OCTET STRINGS.
6779 *) Make SSL library a little more fool-proof by not requiring any longer
6780 that SSL_set_{accept,connect}_state be called before
6781 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6782 in many applications because usually everything *appeared* to work as
6783 intended anyway -- now it really works as intended).
6786 *) Move openssl.cnf out of lib/.
6789 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6790 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6791 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6792 [Ralf S. Engelschall]
6794 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6795 handle PKCS#7 enveloped data properly.
6796 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6798 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6799 copying pointers. The cert_st handling is changed by this in
6800 various ways (and thus what used to be known as ctx->default_cert
6801 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6802 any longer when s->cert does not give us what we need).
6803 ssl_cert_instantiate becomes obsolete by this change.
6804 As soon as we've got the new code right (possibly it already is?),
6805 we have solved a couple of bugs of the earlier code where s->cert
6806 was used as if it could not have been shared with other SSL structures.
6808 Note that using the SSL API in certain dirty ways now will result
6809 in different behaviour than observed with earlier library versions:
6810 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6811 does not influence s as it used to.
6813 In order to clean up things more thoroughly, inside SSL_SESSION
6814 we don't use CERT any longer, but a new structure SESS_CERT
6815 that holds per-session data (if available); currently, this is
6816 the peer's certificate chain and, for clients, the server's certificate
6817 and temporary key. CERT holds only those values that can have
6818 meaningful defaults in an SSL_CTX.
6821 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6822 from the internal representation. Various PKCS#7 fixes: remove some
6823 evil casts and set the enc_dig_alg field properly based on the signing
6827 *) Allow PKCS#12 password to be set from the command line or the
6828 environment. Let 'ca' get its config file name from the environment
6829 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6833 *) Allow certificate policies extension to use an IA5STRING for the
6834 organization field. This is contrary to the PKIX definition but
6835 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6839 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6840 without disallowing inline assembler and the like for non-pedantic builds.
6843 *) Support Borland C++ builder.
6844 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6849 *) SHA-1 cleanups and performance enhancements.
6850 [Andy Polyakov <appro@fy.chalmers.se>]
6852 *) Sparc v8plus assembler for the bignum library.
6853 [Andy Polyakov <appro@fy.chalmers.se>]
6855 *) Accept any -xxx and +xxx compiler options in Configure.
6858 *) Update HPUX configuration.
6861 *) Add missing sk_<type>_unshift() function to safestack.h
6862 [Ralf S. Engelschall]
6864 *) New function SSL_CTX_use_certificate_chain_file that sets the
6865 "extra_cert"s in addition to the certificate. (This makes sense
6866 only for "PEM" format files, as chains as a whole are not
6870 *) Support verify_depth from the SSL API.
6871 x509_vfy.c had what can be considered an off-by-one-error:
6872 Its depth (which was not part of the external interface)
6873 was actually counting the number of certificates in a chain;
6874 now it really counts the depth.
6877 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6878 instead of X509err, which often resulted in confusing error
6879 messages since the error codes are not globally unique
6880 (e.g. an alleged error in ssl3_accept when a certificate
6881 didn't match the private key).
6883 *) New function SSL_CTX_set_session_id_context that allows to set a default
6884 value (so that you don't need SSL_set_session_id_context for each
6885 connection using the SSL_CTX).
6888 *) OAEP decoding bug fix.
6891 *) Support INSTALL_PREFIX for package builders, as proposed by
6895 *) New Configure options "threads" and "no-threads". For systems
6896 where the proper compiler options are known (currently Solaris
6897 and Linux), "threads" is the default.
6900 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6903 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6904 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6905 such as /usr/local/bin.
6908 *) "make linux-shared" to build shared libraries.
6909 [Niels Poppe <niels@netbox.org>]
6911 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6914 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6915 extension adding in x509 utility.
6918 *) Remove NOPROTO sections and error code comments.
6921 *) Partial rewrite of the DEF file generator to now parse the ANSI
6925 *) New Configure options --prefix=DIR and --openssldir=DIR.
6928 *) Complete rewrite of the error code script(s). It is all now handled
6929 by one script at the top level which handles error code gathering,
6930 header rewriting and C source file generation. It should be much better
6931 than the old method: it now uses a modified version of Ulf's parser to
6932 read the ANSI prototypes in all header files (thus the old K&R definitions
6933 aren't needed for error creation any more) and do a better job of
6934 translating function codes into names. The old 'ASN1 error code imbedded
6935 in a comment' is no longer necessary and it doesn't use .err files which
6936 have now been deleted. Also the error code call doesn't have to appear all
6937 on one line (which resulted in some large lines...).
6940 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6943 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6944 0 (which usually indicates a closed connection), but continue reading.
6947 *) Fix some race conditions.
6950 *) Add support for CRL distribution points extension. Add Certificate
6951 Policies and CRL distribution points documentation.
6954 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6957 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6958 8 of keying material. Merlin has also confirmed interop with this fix
6959 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6960 [Merlin Hughes <merlin@baltimore.ie>]
6962 *) Fix lots of warnings.
6963 [Richard Levitte <levitte@stacken.kth.se>]
6965 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6966 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6967 [Richard Levitte <levitte@stacken.kth.se>]
6969 *) Fix problems with sizeof(long) == 8.
6970 [Andy Polyakov <appro@fy.chalmers.se>]
6972 *) Change functions to ANSI C.
6975 *) Fix typos in error codes.
6976 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6978 *) Remove defunct assembler files from Configure.
6981 *) SPARC v8 assembler BIGNUM implementation.
6982 [Andy Polyakov <appro@fy.chalmers.se>]
6984 *) Support for Certificate Policies extension: both print and set.
6985 Various additions to support the r2i method this uses.
6988 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6989 return a const string when you are expecting an allocated buffer.
6992 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6993 types DirectoryString and DisplayText.
6996 *) Add code to allow r2i extensions to access the configuration database,
6997 add an LHASH database driver and add several ctx helper functions.
7000 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7001 fail when they extended the size of a BIGNUM.
7004 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7005 support typesafe stack.
7008 *) Fix typo in SSL_[gs]et_options().
7009 [Nils Frostberg <nils@medcom.se>]
7011 *) Delete various functions and files that belonged to the (now obsolete)
7012 old X509V3 handling code.
7015 *) New Configure option "rsaref".
7018 *) Don't auto-generate pem.h.
7021 *) Introduce type-safe ASN.1 SETs.
7024 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7025 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7027 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7028 that links with OpenSSL (well at least cause lots of warnings), but fear
7029 not: the conversion is trivial, and it eliminates loads of evil casts. A
7030 few STACKed things have been converted already. Feel free to convert more.
7031 In the fullness of time, I'll do away with the STACK type altogether.
7034 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7035 specified in <certfile> by updating the entry in the index.txt file.
7036 This way one no longer has to edit the index.txt file manually for
7037 revoking a certificate. The -revoke option does the gory details now.
7038 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7040 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7041 `-text' option at all and this way the `-noout -text' combination was
7042 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7043 [Ralf S. Engelschall]
7045 *) Make sure a corresponding plain text error message exists for the
7046 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7047 verify callback function determined that a certificate was revoked.
7048 [Ralf S. Engelschall]
7050 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7051 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7052 all available cipers including rc5, which was forgotten until now.
7053 In order to let the testing shell script know which algorithms
7054 are available, a new (up to now undocumented) command
7055 "openssl list-cipher-commands" is used.
7058 *) Bugfix: s_client occasionally would sleep in select() when
7059 it should have checked SSL_pending() first.
7062 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7063 the raw DSA values prior to ASN.1 encoding.
7066 *) Tweaks to Configure
7067 [Niels Poppe <niels@netbox.org>]
7069 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7073 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7076 *) New config option to avoid instructions that are illegal on the 80386.
7077 The default code is faster, but requires at least a 486.
7080 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7081 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7082 same as SSL2_VERSION anyway.
7085 *) New "-showcerts" option for s_client.
7088 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7089 application. Various cleanups and fixes.
7092 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7093 modify error routines to work internally. Add error codes and PBE init
7094 to library startup routines.
7097 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7098 packing functions to asn1 and evp. Changed function names and error
7099 codes along the way.
7102 *) PKCS12 integration: and so it begins... First of several patches to
7103 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7104 objects to objects.h
7107 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7108 and display support for Thawte strong extranet extension.
7111 *) Add LinuxPPC support.
7112 [Jeff Dubrule <igor@pobox.org>]
7114 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7115 bn_div_words in alpha.s.
7116 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7118 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7119 OAEP isn't supported when OpenSSL is built with RSAref.
7120 [Ulf Moeller <ulf@fitug.de>]
7122 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7123 so they no longer are missing under -DNOPROTO.
7124 [Soren S. Jorvang <soren@t.dk>]
7127 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7129 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7130 doesn't work when the session is reused. Coming soon!
7133 *) Fix a security hole, that allows sessions to be reused in the wrong
7134 context thus bypassing client cert protection! All software that uses
7135 client certs and session caches in multiple contexts NEEDS PATCHING to
7136 allow session reuse! A fuller solution is in the works.
7137 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7139 *) Some more source tree cleanups (removed obsolete files
7140 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7141 permission on "config" script to be executable) and a fix for the INSTALL
7143 [Ulf Moeller <ulf@fitug.de>]
7145 *) Remove some legacy and erroneous uses of malloc, free instead of
7147 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7149 *) Make rsa_oaep_test return non-zero on error.
7150 [Ulf Moeller <ulf@fitug.de>]
7152 *) Add support for native Solaris shared libraries. Configure
7153 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7154 if someone would make that last step automatic.
7155 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7157 *) ctx_size was not built with the right compiler during "make links". Fixed.
7160 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7161 except NULL ciphers". This means the default cipher list will no longer
7162 enable NULL ciphers. They need to be specifically enabled e.g. with
7163 the string "DEFAULT:eNULL".
7166 *) Fix to RSA private encryption routines: if p < q then it would
7167 occasionally produce an invalid result. This will only happen with
7168 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7171 *) Be less restrictive and allow also `perl util/perlpath.pl
7172 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7173 because this way one can also use an interpreter named `perl5' (which is
7174 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7175 installed as `perl').
7176 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7178 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7179 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7181 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7182 advapi32.lib to Win32 build and change the pem test comparision
7183 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7184 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7185 and crypto/des/ede_cbcm_enc.c.
7188 *) DES quad checksum was broken on big-endian architectures. Fixed.
7191 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7192 Win32 test batch file so it (might) work again. The Win32 test batch file
7193 is horrible: I feel ill....
7196 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7197 in e_os.h. Audit of header files to check ANSI and non ANSI
7198 sections: 10 functions were absent from non ANSI section and not exported
7199 from Windows DLLs. Fixed up libeay.num for new functions.
7202 *) Make `openssl version' output lines consistent.
7203 [Ralf S. Engelschall]
7205 *) Fix Win32 symbol export lists for BIO functions: Added
7206 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7207 to ms/libeay{16,32}.def.
7208 [Ralf S. Engelschall]
7210 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7211 fine under Unix and passes some trivial tests I've now added. But the
7212 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7213 added to make sure no one expects that this stuff really works in the
7214 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7215 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7217 [Ralf S. Engelschall]
7219 *) Fix the generation of two part addresses in perl.
7220 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7222 *) Add config entry for Linux on MIPS.
7223 [John Tobey <jtobey@channel1.com>]
7225 *) Make links whenever Configure is run, unless we are on Windoze.
7228 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7229 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7233 *) Add a useful kludge to allow package maintainers to specify compiler and
7234 other platforms details on the command line without having to patch the
7235 Configure script everytime: One now can use ``perl Configure
7236 <id>:<details>'', i.e. platform ids are allowed to have details appended
7237 to them (seperated by colons). This is treated as there would be a static
7238 pre-configured entry in Configure's %table under key <id> with value
7239 <details> and ``perl Configure <id>'' is called. So, when you want to
7240 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7241 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7242 now, which overrides the FreeBSD-elf entry on-the-fly.
7243 [Ralf S. Engelschall]
7245 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7248 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7249 on the `perl Configure ...' command line. This way one can compile
7250 OpenSSL libraries with Position Independent Code (PIC) which is needed
7251 for linking it into DSOs.
7252 [Ralf S. Engelschall]
7254 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7258 *) Cleaned up the LICENSE document: The official contact for any license
7259 questions now is the OpenSSL core team under openssl-core@openssl.org.
7260 And add a paragraph about the dual-license situation to make sure people
7261 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7262 to the OpenSSL toolkit.
7263 [Ralf S. Engelschall]
7265 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7266 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7267 Additonally cleaned up the `make links' target: Remove unnecessary
7268 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7269 to speed processing and no longer clutter the display with confusing
7270 stuff. Instead only the actually done links are displayed.
7271 [Ralf S. Engelschall]
7273 *) Permit null encryption ciphersuites, used for authentication only. It used
7274 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7275 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7279 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7280 signed attributes when verifying signatures (this would break them),
7281 the detached data encoding was wrong and public keys obtained using
7282 X509_get_pubkey() weren't freed.
7285 *) Add text documentation for the BUFFER functions. Also added a work around
7286 to a Win95 console bug. This was triggered by the password read stuff: the
7287 last character typed gets carried over to the next fread(). If you were
7288 generating a new cert request using 'req' for example then the last
7289 character of the passphrase would be CR which would then enter the first
7293 *) Added the new `Includes OpenSSL Cryptography Software' button as
7294 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7295 button and can be used by applications based on OpenSSL to show the
7296 relationship to the OpenSSL project.
7297 [Ralf S. Engelschall]
7299 *) Remove confusing variables in function signatures in files
7300 ssl/ssl_lib.c and ssl/ssl.h.
7301 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7303 *) Don't install bss_file.c under PREFIX/include/
7304 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7306 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7307 functions that return function pointers and has support for NT specific
7308 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7309 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7310 unsigned to signed types: this was killing the Win32 compile.
7313 *) Add new certificate file to stack functions,
7314 SSL_add_dir_cert_subjects_to_stack() and
7315 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7316 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7317 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7318 This means that Apache-SSL and similar packages don't have to mess around
7319 to add as many CAs as they want to the preferred list.
7322 *) Experiment with doxygen documentation. Currently only partially applied to
7324 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7325 openssl.doxy as the configuration file.
7328 *) Get rid of remaining C++-style comments which strict C compilers hate.
7329 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7331 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7332 compiled in by default: it has problems with large keys.
7335 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7336 DH private keys and/or callback functions which directly correspond to
7337 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7338 is needed for applications which have to configure certificates on a
7339 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7341 For the RSA certificate situation is makes no difference, but
7342 for the DSA certificate situation this fixes the "no shared cipher"
7343 problem where the OpenSSL cipher selection procedure failed because the
7344 temporary keys were not overtaken from the context and the API provided
7345 no way to reconfigure them.
7346 The new functions now let applications reconfigure the stuff and they
7347 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7348 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7349 non-public-API function ssl_cert_instantiate() is used as a helper
7350 function and also to reduce code redundancy inside ssl_rsa.c.
7351 [Ralf S. Engelschall]
7353 *) Move s_server -dcert and -dkey options out of the undocumented feature
7354 area because they are useful for the DSA situation and should be
7355 recognized by the users.
7356 [Ralf S. Engelschall]
7358 *) Fix the cipher decision scheme for export ciphers: the export bits are
7359 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7360 SSL_EXP_MASK. So, the original variable has to be used instead of the
7361 already masked variable.
7362 [Richard Levitte <levitte@stacken.kth.se>]
7364 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7365 [Richard Levitte <levitte@stacken.kth.se>]
7367 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7368 from `int' to `unsigned int' because it's a length and initialized by
7369 EVP_DigestFinal() which expects an `unsigned int *'.
7370 [Richard Levitte <levitte@stacken.kth.se>]
7372 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7373 script. Instead use the usual Shell->Perl transition trick.
7374 [Ralf S. Engelschall]
7376 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7377 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7378 -noout -modulus' as it's already the case for `openssl rsa -noout
7379 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7380 currently the public key is printed (a decision which was already done by
7381 `openssl dsa -modulus' in the past) which serves a similar purpose.
7382 Additionally the NO_RSA no longer completely removes the whole -modulus
7383 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7385 [Ralf S. Engelschall]
7387 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7388 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7389 [Arne Ansper <arne@ats.cyber.ee>]
7391 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7392 to be added. Now both 'req' and 'ca' can use new objects defined in the
7396 *) Add cool BIO that does syslog (or event log on NT).
7397 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7399 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7400 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7401 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7402 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7405 *) Add preliminary config info for new extension code.
7408 *) Make RSA_NO_PADDING really use no padding.
7409 [Ulf Moeller <ulf@fitug.de>]
7411 *) Generate errors when private/public key check is done.
7414 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7415 for some CRL extensions and new objects added.
7418 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7419 key usage extension and fuller support for authority key id.
7422 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7423 padding method for RSA, which is recommended for new applications in PKCS
7424 #1 v2.0 (RFC 2437, October 1998).
7425 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7426 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7427 against Bleichbacher's attack on RSA.
7428 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7431 *) Updates to the new SSL compression code
7432 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7434 *) Fix so that the version number in the master secret, when passed
7435 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7436 (because the server will not accept higher), that the version number
7437 is 0x03,0x01, not 0x03,0x00
7438 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7440 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7441 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7442 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7445 *) Support for RAW extensions where an arbitrary extension can be
7446 created by including its DER encoding. See apps/openssl.cnf for
7450 *) Make sure latest Perl versions don't interpret some generated C array
7451 code as Perl array code in the crypto/err/err_genc.pl script.
7452 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7454 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7455 not many people have the assembler. Various Win32 compilation fixes and
7456 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7460 *) Modify configure script 'Configure' to automatically create crypto/date.h
7461 file under Win32 and also build pem.h from pem.org. New script
7462 util/mkfiles.pl to create the MINFO file on environments that can't do a
7463 'make files': perl util/mkfiles.pl >MINFO should work.
7466 *) Major rework of DES function declarations, in the pursuit of correctness
7467 and purity. As a result, many evil casts evaporated, and some weirdness,
7468 too. You may find this causes warnings in your code. Zapping your evil
7469 casts will probably fix them. Mostly.
7472 *) Fix for a typo in asn1.h. Bug fix to object creation script
7473 obj_dat.pl. It considered a zero in an object definition to mean
7474 "end of object": none of the objects in objects.h have any zeros
7475 so it wasn't spotted.
7476 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7478 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7479 Masking (CBCM). In the absence of test vectors, the best I have been able
7480 to do is check that the decrypt undoes the encrypt, so far. Send me test
7481 vectors if you have them.
7484 *) Correct calculation of key length for export ciphers (too much space was
7485 allocated for null ciphers). This has not been tested!
7488 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7489 message is now correct (it understands "crypto" and "ssl" on its
7490 command line). There is also now an "update" option. This will update
7491 the util/ssleay.num and util/libeay.num files with any new functions.
7493 perl util/mkdef.pl crypto ssl update
7494 it will update them.
7497 *) Overhauled the Perl interface (perl/*):
7498 - ported BN stuff to OpenSSL's different BN library
7499 - made the perl/ source tree CVS-aware
7500 - renamed the package from SSLeay to OpenSSL (the files still contain
7501 their history because I've copied them in the repository)
7502 - removed obsolete files (the test scripts will be replaced
7503 by better Test::Harness variants in the future)
7504 [Ralf S. Engelschall]
7506 *) First cut for a very conservative source tree cleanup:
7507 1. merge various obsolete readme texts into doc/ssleay.txt
7508 where we collect the old documents and readme texts.
7509 2. remove the first part of files where I'm already sure that we no
7510 longer need them because of three reasons: either they are just temporary
7511 files which were left by Eric or they are preserved original files where
7512 I've verified that the diff is also available in the CVS via "cvs diff
7513 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7514 the crypto/md/ stuff).
7515 [Ralf S. Engelschall]
7517 *) More extension code. Incomplete support for subject and issuer alt
7518 name, issuer and authority key id. Change the i2v function parameters
7519 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7520 what that's for :-) Fix to ASN1 macro which messed up
7521 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7524 *) Preliminary support for ENUMERATED type. This is largely copied from the
7528 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7529 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7531 *) Make sure `make rehash' target really finds the `openssl' program.
7532 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7534 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7535 like to hear about it if this slows down other processors.
7538 *) Add CygWin32 platform information to Configure script.
7539 [Alan Batie <batie@aahz.jf.intel.com>]
7541 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7542 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7544 *) New program nseq to manipulate netscape certificate sequences
7547 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7551 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7552 but the BN code had some problems that would cause failures when
7553 doing certificate verification and some other functions.
7554 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7556 *) Add ASN1 and PEM code to support netscape certificate sequences.
7559 *) Add ASN1 and PEM code to support netscape certificate sequences.
7562 *) Add several PKIX and private extended key usage OIDs.
7565 *) Modify the 'ca' program to handle the new extension code. Modify
7566 openssl.cnf for new extension format, add comments.
7569 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7570 and add a sample to openssl.cnf so req -x509 now adds appropriate
7574 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7575 error code, add initial support to X509_print() and x509 application.
7578 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7579 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7580 stuff is currently isolated and isn't even compiled yet.
7583 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7584 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7585 Removed the versions check from X509 routines when loading extensions:
7586 this allows certain broken certificates that don't set the version
7587 properly to be processed.
7590 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7591 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7592 can still be regenerated with "make depend".
7595 *) Spelling mistake in C version of CAST-128.
7596 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7598 *) Changes to the error generation code. The perl script err-code.pl
7599 now reads in the old error codes and retains the old numbers, only
7600 adding new ones if necessary. It also only changes the .err files if new
7601 codes are added. The makefiles have been modified to only insert errors
7602 when needed (to avoid needlessly modifying header files). This is done
7603 by only inserting errors if the .err file is newer than the auto generated
7604 C file. To rebuild all the error codes from scratch (the old behaviour)
7605 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7606 or delete all the .err files.
7609 *) CAST-128 was incorrectly implemented for short keys. The C version has
7610 been fixed, but is untested. The assembler versions are also fixed, but
7611 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7612 to regenerate it if needed.
7613 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7614 Hagino <itojun@kame.net>]
7616 *) File was opened incorrectly in randfile.c.
7617 [Ulf Möller <ulf@fitug.de>]
7619 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7620 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7621 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7622 al: it's just almost always a UTCTime. Note this patch adds new error
7623 codes so do a "make errors" if there are problems.
7626 *) Correct Linux 1 recognition in config.
7627 [Ulf Möller <ulf@fitug.de>]
7629 *) Remove pointless MD5 hash when using DSA keys in ca.
7630 [Anonymous <nobody@replay.com>]
7632 *) Generate an error if given an empty string as a cert directory. Also
7633 generate an error if handed NULL (previously returned 0 to indicate an
7634 error, but didn't set one).
7635 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7637 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7640 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7641 parameters. This was causing a warning which killed off the Win32 compile.
7644 *) Remove C++ style comments from crypto/bn/bn_local.h.
7645 [Neil Costigan <neil.costigan@celocom.com>]
7647 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7648 based on a text string, looking up short and long names and finally
7649 "dot" format. The "dot" format stuff didn't work. Added new function
7650 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7651 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7652 OID is not part of the table.
7655 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7656 X509_LOOKUP_by_alias().
7659 *) Sort openssl functions by name.
7662 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7663 encryption from sample DSA keys (in case anyone is interested the password
7667 *) Make _all_ *_free functions accept a NULL pointer.
7668 [Frans Heymans <fheymans@isaserver.be>]
7670 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7672 [Anonymous <nobody@replay.com>]
7674 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7675 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7677 *) Don't blow it for numeric -newkey arguments to apps/req.
7678 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7680 *) Temp key "for export" tests were wrong in s3_srvr.c.
7681 [Anonymous <nobody@replay.com>]
7683 *) Add prototype for temp key callback functions
7684 SSL_CTX_set_tmp_{rsa,dh}_callback().
7687 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7688 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7691 *) X509_name_add_entry() freed the wrong thing after an error.
7692 [Arne Ansper <arne@ats.cyber.ee>]
7694 *) rsa_eay.c would attempt to free a NULL context.
7695 [Arne Ansper <arne@ats.cyber.ee>]
7697 *) BIO_s_socket() had a broken should_retry() on Windoze.
7698 [Arne Ansper <arne@ats.cyber.ee>]
7700 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7701 [Arne Ansper <arne@ats.cyber.ee>]
7703 *) Make sure the already existing X509_STORE->depth variable is initialized
7704 in X509_STORE_new(), but document the fact that this variable is still
7705 unused in the certificate verification process.
7706 [Ralf S. Engelschall]
7708 *) Fix the various library and apps files to free up pkeys obtained from
7709 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7712 *) Fix reference counting in X509_PUBKEY_get(). This makes
7713 demos/maurice/example2.c work, amongst others, probably.
7714 [Steve Henson and Ben Laurie]
7716 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7717 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7718 are no longer created. This way we have a single and consistent command
7719 line interface `openssl <command>', similar to `cvs <command>'.
7720 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7722 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7723 BIT STRING wrapper always have zero unused bits.
7726 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7729 *) Make the top-level INSTALL documentation easier to understand.
7732 *) Makefiles updated to exit if an error occurs in a sub-directory
7733 make (including if user presses ^C) [Paul Sutton]
7735 *) Make Montgomery context stuff explicit in RSA data structure.
7738 *) Fix build order of pem and err to allow for generated pem.h.
7741 *) Fix renumbering bug in X509_NAME_delete_entry().
7744 *) Enhanced the err-ins.pl script so it makes the error library number
7745 global and can add a library name. This is needed for external ASN1 and
7746 other error libraries.
7749 *) Fixed sk_insert which never worked properly.
7752 *) Fix ASN1 macros so they can handle indefinite length construted
7753 EXPLICIT tags. Some non standard certificates use these: they can now
7757 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7758 into a single doc/ssleay.txt bundle. This way the information is still
7759 preserved but no longer messes up this directory. Now it's new room for
7760 the new set of documenation files.
7761 [Ralf S. Engelschall]
7763 *) SETs were incorrectly DER encoded. This was a major pain, because they
7764 shared code with SEQUENCEs, which aren't coded the same. This means that
7765 almost everything to do with SETs or SEQUENCEs has either changed name or
7766 number of arguments.
7767 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7769 *) Fix test data to work with the above.
7772 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7773 was already fixed by Eric for 0.9.1 it seems.
7774 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7776 *) Autodetect FreeBSD3.
7779 *) Fix various bugs in Configure. This affects the following platforms:
7783 unixware-2.0-pentium
7787 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7788 before they are needed.
7791 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7795 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7797 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7798 changed SSLeay to OpenSSL in version strings.
7799 [Ralf S. Engelschall]
7801 *) Some fixups to the top-level documents.
7804 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7805 because the symlink to include/ was missing.
7806 [Ralf S. Engelschall]
7808 *) Incorporated the popular no-RSA/DSA-only patches
7809 which allow to compile a RSA-free SSLeay.
7810 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7812 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7813 when "ssleay" is still not found.
7814 [Ralf S. Engelschall]
7816 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7817 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7819 *) Updated the README file.
7820 [Ralf S. Engelschall]
7822 *) Added various .cvsignore files in the CVS repository subdirs
7823 to make a "cvs update" really silent.
7824 [Ralf S. Engelschall]
7826 *) Recompiled the error-definition header files and added
7827 missing symbols to the Win32 linker tables.
7828 [Ralf S. Engelschall]
7830 *) Cleaned up the top-level documents;
7831 o new files: CHANGES and LICENSE
7832 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7833 o merged COPYRIGHT into LICENSE
7834 o removed obsolete TODO file
7835 o renamed MICROSOFT to INSTALL.W32
7836 [Ralf S. Engelschall]
7838 *) Removed dummy files from the 0.9.1b source tree:
7839 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7840 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7841 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7842 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7843 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7844 [Ralf S. Engelschall]
7846 *) Added various platform portability fixes.
7849 *) The Genesis of the OpenSSL rpject:
7850 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7851 Young and Tim J. Hudson created while they were working for C2Net until
7853 [The OpenSSL Project]
7856 Changes between 0.9.0b and 0.9.1b [not released]
7858 *) Updated a few CA certificates under certs/
7861 *) Changed some BIGNUM api stuff.
7864 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7865 DGUX x86, Linux Alpha, etc.
7868 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7869 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7873 *) Add -strparse option to asn1pars program which parses nested
7875 [Dr Stephen Henson <shenson@bigfoot.com>]
7877 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7880 *) DSA fix for "ca" program.
7883 *) Added "-genkey" option to "dsaparam" program.
7886 *) Added RIPE MD160 (rmd160) message digest.
7889 *) Added -a (all) option to "ssleay version" command.
7892 *) Added PLATFORM define which is the id given to Configure.
7895 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7898 *) Extended the ASN.1 parser routines.
7901 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7904 *) Added a BN_CTX to the BN library.
7907 *) Fixed the weak key values in DES library
7910 *) Changed API in EVP library for cipher aliases.
7913 *) Added support for RC2/64bit cipher.
7916 *) Converted the lhash library to the crypto/mem.c functions.
7919 *) Added more recognized ASN.1 object ids.
7922 *) Added more RSA padding checks for SSL/TLS.
7925 *) Added BIO proxy/filter functionality.
7928 *) Added extra_certs to SSL_CTX which can be used
7929 send extra CA certificates to the client in the CA cert chain sending
7930 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7933 *) Now Fortezza is denied in the authentication phase because
7934 this is key exchange mechanism is not supported by SSLeay at all.
7937 *) Additional PKCS1 checks.
7940 *) Support the string "TLSv1" for all TLS v1 ciphers.
7943 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7944 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7947 *) Fixed a few memory leaks.
7950 *) Fixed various code and comment typos.
7953 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7954 bytes sent in the client random.
7955 [Edward Bishop <ebishop@spyglass.com>]