]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Minor clarification.
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 0.9.8a and 0.9.9 [xx XXX xxxx]
6
7 *) Add initial support for TLS extensions, specifically for the server_name
8 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
9 have new members for a host name. The SSL data structure has an
10 additional member SSL_CTX *initial_ctx so that new sessions can be
11 stored in that context to allow for session resumption, even after the
12 SSL has been switched to a new SSL_CTX in reaction to a client's
13 server_name extension.
14
15 New functions (subject to change):
16
17 SSL_get_servername()
18 SSL_get_servername_type()
19 SSL_set_SSL_CTX()
20
21 New CTRL codes and macros (subject to change):
22
23 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
24 - SSL_CTX_set_tlsext_servername_callback()
25 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
26 - SSL_CTX_set_tlsext_servername_arg()
27 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_hostname()
28
29 openssl s_client has a new '-servername ...' option.
30
31 openssl s_server has new options '-servername_host ...', '-cert2 ...',
32 '-key2 ...', '-servername_fatal' (subject to change). This allows
33 testing the HostName extension for a specific single host name ('-cert'
34 and '-key' remain fallbacks for handshakes without HostName
35 negotiation). If the unrecogninzed_name alert has to be sent, this by
36 default is a warning; it becomes fatal with the '-servername_fatal'
37 option.
38
39 [Peter Sylvester, Remy Allais, Christophe Renou]
40
41 *) Whirlpool hash implementation is added.
42 [Andy Polyakov]
43
44 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
45 bn(64,32). Because of instruction set limitations it doesn't have
46 any negative impact on performance. This was done mostly in order
47 to make it possible to share assembler modules, such as bn_mul_mont
48 implementations, between 32- and 64-bit builds without hassle.
49 [Andy Polyakov]
50
51 *) Disable rogue ciphersuites:
52
53 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
54 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
55 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
56
57 The latter two were purportedly from
58 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
59 appear there.
60
61 Other ciphersuites from draft-ietf-tls-56-bit-ciphersuites-01.txt
62 remain enabled for now, but are just as unofficial, and the ID
63 has long expired; these will probably disappear soon.
64 [Bodo Moeller]
65
66 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
67 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
68 macro.
69 [Bodo Moeller]
70
71 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
72 dedicated Montgomery multiplication procedure, is introduced.
73 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
74 "64-bit" performance on certain 32-bit targets.
75 [Andy Polyakov]
76
77 *) New option SSL_OP_NO_COMP to disable use of compression selectively
78 in SSL structures. New SSL ctrl to set maximum send fragment size.
79 Save memory by seeting the I/O buffer sizes dynamically instead of
80 using the maximum available value.
81 [Steve Henson]
82
83 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
84 in addition to the text details.
85 [Bodo Moeller]
86
87 *) Very, very preliminary EXPERIMENTAL support for printing of general
88 ASN1 structures. This currently produces rather ugly output and doesn't
89 handle several customised structures at all.
90 [Steve Henson]
91
92 *) Integrated support for PVK file format and some related formats such
93 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
94 these in the 'rsa' and 'dsa' utilities.
95 [Steve Henson]
96
97 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
98 [Steve Henson]
99
100 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
101 place for the (very old) "NETSCAPE" format certificates which are now
102 handled using new ASN1 code equivalents.
103 [Steve Henson]
104
105 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
106 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
107 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
108 [Nils Larsch]
109
110 *) Modify CRL distribution points extension code to print out previously
111 unsupported fields. Enhance extension setting code to allow setting of
112 all fields.
113 [Steve Henson]
114
115 *) Add print and set support for Issuing Distribution Point CRL extension.
116 [Steve Henson]
117
118 Changes between 0.9.8a and 0.9.8b [XX xxx XXXX]
119
120 *) Link in manifests for VC++ if needed.
121 [Austin Ziegler <halostatue@gmail.com>]
122
123 *) Update support for ECC-based TLS ciphersuites according to
124 draft-ietf-tls-ecc-12.txt with proposed changes.
125 [Douglas Stebila]
126
127 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
128 opaque EVP_CIPHER_CTX handling.
129 [Steve Henson]
130
131 *) Fixes and enhancements to zlib compression code. We now only use
132 "zlib1.dll" and use the default __cdecl calling convention on Win32
133 to conform with the standards mentioned here:
134 http://www.zlib.net/DLL_FAQ.txt
135 Static zlib linking now works on Windows and the new --with-zlib-include
136 --with-zlib-lib options to Configure can be used to supply the location
137 of the headers and library. Gracefully handle case where zlib library
138 can't be loaded.
139 [Steve Henson]
140
141 *) Several fixes and enhancements to the OID generation code. The old code
142 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
143 handle numbers larger than ULONG_MAX, truncated printing and had a
144 non standard OBJ_obj2txt() behaviour.
145 [Steve Henson]
146
147 *) Add support for building of engines under engine/ as shared libraries
148 under VC++ build system.
149 [Steve Henson]
150
151 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
152 Hopefully, we will not see any false combination of paths any more.
153 [Richard Levitte]
154
155 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
156
157 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
158 (part of SSL_OP_ALL). This option used to disable the
159 countermeasure against man-in-the-middle protocol-version
160 rollback in the SSL 2.0 server implementation, which is a bad
161 idea. (CVE-2005-2969)
162
163 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
164 for Information Security, National Institute of Advanced Industrial
165 Science and Technology [AIST], Japan)]
166
167 *) Add two function to clear and return the verify parameter flags.
168 [Steve Henson]
169
170 *) Keep cipherlists sorted in the source instead of sorting them at
171 runtime, thus removing the need for a lock.
172 [Nils Larsch]
173
174 *) Avoid some small subgroup attacks in Diffie-Hellman.
175 [Nick Mathewson and Ben Laurie]
176
177 *) Add functions for well-known primes.
178 [Nick Mathewson]
179
180 *) Extended Windows CE support.
181 [Satoshi Nakamura and Andy Polyakov]
182
183 *) Initialize SSL_METHOD structures at compile time instead of during
184 runtime, thus removing the need for a lock.
185 [Steve Henson]
186
187 *) Make PKCS7_decrypt() work even if no certificate is supplied by
188 attempting to decrypt each encrypted key in turn. Add support to
189 smime utility.
190 [Steve Henson]
191
192 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
193
194 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
195 [Richard Levitte]
196
197 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
198 key into the same file any more.
199 [Richard Levitte]
200
201 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
202 [Andy Polyakov]
203
204 *) Add -utf8 command line and config file option to 'ca'.
205 [Stefan <stf@udoma.org]
206
207 *) Removed the macro des_crypt(), as it seems to conflict with some
208 libraries. Use DES_crypt().
209 [Richard Levitte]
210
211 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
212 involves renaming the source and generated shared-libs for
213 both. The engines will accept the corrected or legacy ids
214 ('ncipher' and '4758_cca' respectively) when binding. NB,
215 this only applies when building 'shared'.
216 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
217
218 *) Add attribute functions to EVP_PKEY structure. Modify
219 PKCS12_create() to recognize a CSP name attribute and
220 use it. Make -CSP option work again in pkcs12 utility.
221 [Steve Henson]
222
223 *) Add new functionality to the bn blinding code:
224 - automatic re-creation of the BN_BLINDING parameters after
225 a fixed number of uses (currently 32)
226 - add new function for parameter creation
227 - introduce flags to control the update behaviour of the
228 BN_BLINDING parameters
229 - hide BN_BLINDING structure
230 Add a second BN_BLINDING slot to the RSA structure to improve
231 performance when a single RSA object is shared among several
232 threads.
233 [Nils Larsch]
234
235 *) Add support for DTLS.
236 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
237
238 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
239 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
240 [Walter Goulet]
241
242 *) Remove buggy and incompletet DH cert support from
243 ssl/ssl_rsa.c and ssl/s3_both.c
244 [Nils Larsch]
245
246 *) Use SHA-1 instead of MD5 as the default digest algorithm for
247 the apps/openssl applications.
248 [Nils Larsch]
249
250 *) Compile clean with "-Wall -Wmissing-prototypes
251 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
252 DEBUG_SAFESTACK must also be set.
253 [Ben Laurie]
254
255 *) Change ./Configure so that certain algorithms can be disabled by default.
256 The new counterpiece to "no-xxx" is "enable-xxx".
257
258 The patented RC5 and MDC2 algorithms will now be disabled unless
259 "enable-rc5" and "enable-mdc2", respectively, are specified.
260
261 (IDEA remains enabled despite being patented. This is because IDEA
262 is frequently required for interoperability, and there is no license
263 fee for non-commercial use. As before, "no-idea" can be used to
264 avoid this algorithm.)
265
266 [Bodo Moeller]
267
268 *) Add processing of proxy certificates (see RFC 3820). This work was
269 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
270 EGEE (Enabling Grids for E-science in Europe).
271 [Richard Levitte]
272
273 *) RC4 performance overhaul on modern architectures/implementations, such
274 as Intel P4, IA-64 and AMD64.
275 [Andy Polyakov]
276
277 *) New utility extract-section.pl. This can be used specify an alternative
278 section number in a pod file instead of having to treat each file as
279 a separate case in Makefile. This can be done by adding two lines to the
280 pod file:
281
282 =for comment openssl_section:XXX
283
284 The blank line is mandatory.
285
286 [Steve Henson]
287
288 *) New arguments -certform, -keyform and -pass for s_client and s_server
289 to allow alternative format key and certificate files and passphrase
290 sources.
291 [Steve Henson]
292
293 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
294 update associated structures and add various utility functions.
295
296 Add new policy related verify parameters, include policy checking in
297 standard verify code. Enhance 'smime' application with extra parameters
298 to support policy checking and print out.
299 [Steve Henson]
300
301 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
302 Nehemiah processors. These extensions support AES encryption in hardware
303 as well as RNG (though RNG support is currently disabled).
304 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
305
306 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
307 [Geoff Thorpe]
308
309 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
310 [Andy Polyakov and a number of other people]
311
312 *) Improved PowerPC platform support. Most notably BIGNUM assembler
313 implementation contributed by IBM.
314 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
315
316 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
317 exponent rather than 'unsigned long'. There is a corresponding change to
318 the new 'rsa_keygen' element of the RSA_METHOD structure.
319 [Jelte Jansen, Geoff Thorpe]
320
321 *) Functionality for creating the initial serial number file is now
322 moved from CA.pl to the 'ca' utility with a new option -create_serial.
323
324 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
325 number file to 1, which is bound to cause problems. To avoid
326 the problems while respecting compatibility between different 0.9.7
327 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
328 CA.pl for serial number initialization. With the new release 0.9.8,
329 we can fix the problem directly in the 'ca' utility.)
330 [Steve Henson]
331
332 *) Reduced header interdepencies by declaring more opaque objects in
333 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
334 give fewer recursive includes, which could break lazy source code - so
335 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
336 developers should define this symbol when building and using openssl to
337 ensure they track the recommended behaviour, interfaces, [etc], but
338 backwards-compatible behaviour prevails when this isn't defined.
339 [Geoff Thorpe]
340
341 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
342 [Steve Henson]
343
344 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
345 This will generate a random key of the appropriate length based on the
346 cipher context. The EVP_CIPHER can provide its own random key generation
347 routine to support keys of a specific form. This is used in the des and
348 3des routines to generate a key of the correct parity. Update S/MIME
349 code to use new functions and hence generate correct parity DES keys.
350 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
351 valid (weak or incorrect parity).
352 [Steve Henson]
353
354 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
355 as looking them up. This is useful when the verified structure may contain
356 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
357 present unless the new PKCS7_NO_CRL flag is asserted.
358 [Steve Henson]
359
360 *) Extend ASN1 oid configuration module. It now additionally accepts the
361 syntax:
362
363 shortName = some long name, 1.2.3.4
364 [Steve Henson]
365
366 *) Reimplemented the BN_CTX implementation. There is now no more static
367 limitation on the number of variables it can handle nor the depth of the
368 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
369 information can now expand as required, and rather than having a single
370 static array of bignums, BN_CTX now uses a linked-list of such arrays
371 allowing it to expand on demand whilst maintaining the usefulness of
372 BN_CTX's "bundling".
373 [Geoff Thorpe]
374
375 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
376 to allow all RSA operations to function using a single BN_CTX.
377 [Geoff Thorpe]
378
379 *) Preliminary support for certificate policy evaluation and checking. This
380 is initially intended to pass the tests outlined in "Conformance Testing
381 of Relying Party Client Certificate Path Processing Logic" v1.07.
382 [Steve Henson]
383
384 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
385 remained unused and not that useful. A variety of other little bignum
386 tweaks and fixes have also been made continuing on from the audit (see
387 below).
388 [Geoff Thorpe]
389
390 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
391 associated ASN1, EVP and SSL functions and old ASN1 macros.
392 [Richard Levitte]
393
394 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
395 and this should never fail. So the return value from the use of
396 BN_set_word() (which can fail due to needless expansion) is now deprecated;
397 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
398 [Geoff Thorpe]
399
400 *) BN_CTX_get() should return zero-valued bignums, providing the same
401 initialised value as BN_new().
402 [Geoff Thorpe, suggested by Ulf Möller]
403
404 *) Support for inhibitAnyPolicy certificate extension.
405 [Steve Henson]
406
407 *) An audit of the BIGNUM code is underway, for which debugging code is
408 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
409 is considered valid when processing BIGNUMs, and causes execution to
410 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
411 further steps are taken to deliberately pollute unused data in BIGNUM
412 structures to try and expose faulty code further on. For now, openssl will
413 (in its default mode of operation) continue to tolerate the inconsistent
414 forms that it has tolerated in the past, but authors and packagers should
415 consider trying openssl and their own applications when compiled with
416 these debugging symbols defined. It will help highlight potential bugs in
417 their own code, and will improve the test coverage for OpenSSL itself. At
418 some point, these tighter rules will become openssl's default to improve
419 maintainability, though the assert()s and other overheads will remain only
420 in debugging configurations. See bn.h for more details.
421 [Geoff Thorpe, Nils Larsch, Ulf Möller]
422
423 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
424 that can only be obtained through BN_CTX_new() (which implicitly
425 initialises it). The presence of this function only made it possible
426 to overwrite an existing structure (and cause memory leaks).
427 [Geoff Thorpe]
428
429 *) Because of the callback-based approach for implementing LHASH as a
430 template type, lh_insert() adds opaque objects to hash-tables and
431 lh_doall() or lh_doall_arg() are typically used with a destructor callback
432 to clean up those corresponding objects before destroying the hash table
433 (and losing the object pointers). So some over-zealous constifications in
434 LHASH have been relaxed so that lh_insert() does not take (nor store) the
435 objects as "const" and the lh_doall[_arg] callback wrappers are not
436 prototyped to have "const" restrictions on the object pointers they are
437 given (and so aren't required to cast them away any more).
438 [Geoff Thorpe]
439
440 *) The tmdiff.h API was so ugly and minimal that our own timing utility
441 (speed) prefers to use its own implementation. The two implementations
442 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
443 its object type properly exposed (MS_TM) instead of casting to/from "char
444 *". This may still change yet if someone realises MS_TM and "ms_time_***"
445 aren't necessarily the greatest nomenclatures - but this is what was used
446 internally to the implementation so I've used that for now.
447 [Geoff Thorpe]
448
449 *) Ensure that deprecated functions do not get compiled when
450 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
451 the self-tests were still using deprecated key-generation functions so
452 these have been updated also.
453 [Geoff Thorpe]
454
455 *) Reorganise PKCS#7 code to separate the digest location functionality
456 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
457 New function PKCS7_set_digest() to set the digest type for PKCS#7
458 digestedData type. Add additional code to correctly generate the
459 digestedData type and add support for this type in PKCS7 initialization
460 functions.
461 [Steve Henson]
462
463 *) New function PKCS7_set0_type_other() this initializes a PKCS7
464 structure of type "other".
465 [Steve Henson]
466
467 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
468 sure the loop does correctly stop and breaking ("division by zero")
469 modulus operations are not performed. The (pre-generated) prime
470 table crypto/bn/bn_prime.h was already correct, but it could not be
471 re-generated on some platforms because of the "division by zero"
472 situation in the script.
473 [Ralf S. Engelschall]
474
475 *) Update support for ECC-based TLS ciphersuites according to
476 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
477 SHA-1 now is only used for "small" curves (where the
478 representation of a field element takes up to 24 bytes); for
479 larger curves, the field element resulting from ECDH is directly
480 used as premaster secret.
481 [Douglas Stebila (Sun Microsystems Laboratories)]
482
483 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
484 curve secp160r1 to the tests.
485 [Douglas Stebila (Sun Microsystems Laboratories)]
486
487 *) Add the possibility to load symbols globally with DSO.
488 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
489
490 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
491 control of the error stack.
492 [Richard Levitte]
493
494 *) Add support for STORE in ENGINE.
495 [Richard Levitte]
496
497 *) Add the STORE type. The intention is to provide a common interface
498 to certificate and key stores, be they simple file-based stores, or
499 HSM-type store, or LDAP stores, or...
500 NOTE: The code is currently UNTESTED and isn't really used anywhere.
501 [Richard Levitte]
502
503 *) Add a generic structure called OPENSSL_ITEM. This can be used to
504 pass a list of arguments to any function as well as provide a way
505 for a function to pass data back to the caller.
506 [Richard Levitte]
507
508 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
509 works like BUF_strdup() but can be used to duplicate a portion of
510 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
511 a memory area.
512 [Richard Levitte]
513
514 *) Add the function sk_find_ex() which works like sk_find(), but will
515 return an index to an element even if an exact match couldn't be
516 found. The index is guaranteed to point at the element where the
517 searched-for key would be inserted to preserve sorting order.
518 [Richard Levitte]
519
520 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
521 takes an extra flags argument for optional functionality. Currently,
522 the following flags are defined:
523
524 OBJ_BSEARCH_VALUE_ON_NOMATCH
525 This one gets OBJ_bsearch_ex() to return a pointer to the first
526 element where the comparing function returns a negative or zero
527 number.
528
529 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
530 This one gets OBJ_bsearch_ex() to return a pointer to the first
531 element where the comparing function returns zero. This is useful
532 if there are more than one element where the comparing function
533 returns zero.
534 [Richard Levitte]
535
536 *) Make it possible to create self-signed certificates with 'openssl ca'
537 in such a way that the self-signed certificate becomes part of the
538 CA database and uses the same mechanisms for serial number generation
539 as all other certificate signing. The new flag '-selfsign' enables
540 this functionality. Adapt CA.sh and CA.pl.in.
541 [Richard Levitte]
542
543 *) Add functionality to check the public key of a certificate request
544 against a given private. This is useful to check that a certificate
545 request can be signed by that key (self-signing).
546 [Richard Levitte]
547
548 *) Make it possible to have multiple active certificates with the same
549 subject in the CA index file. This is done only if the keyword
550 'unique_subject' is set to 'no' in the main CA section (default
551 if 'CA_default') of the configuration file. The value is saved
552 with the database itself in a separate index attribute file,
553 named like the index file with '.attr' appended to the name.
554 [Richard Levitte]
555
556 *) Generate muti valued AVAs using '+' notation in config files for
557 req and dirName.
558 [Steve Henson]
559
560 *) Support for nameConstraints certificate extension.
561 [Steve Henson]
562
563 *) Support for policyConstraints certificate extension.
564 [Steve Henson]
565
566 *) Support for policyMappings certificate extension.
567 [Steve Henson]
568
569 *) Make sure the default DSA_METHOD implementation only uses its
570 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
571 and change its own handlers to be NULL so as to remove unnecessary
572 indirection. This lets alternative implementations fallback to the
573 default implementation more easily.
574 [Geoff Thorpe]
575
576 *) Support for directoryName in GeneralName related extensions
577 in config files.
578 [Steve Henson]
579
580 *) Make it possible to link applications using Makefile.shared.
581 Make that possible even when linking against static libraries!
582 [Richard Levitte]
583
584 *) Support for single pass processing for S/MIME signing. This now
585 means that S/MIME signing can be done from a pipe, in addition
586 cleartext signing (multipart/signed type) is effectively streaming
587 and the signed data does not need to be all held in memory.
588
589 This is done with a new flag PKCS7_STREAM. When this flag is set
590 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
591 is done after the data is output (and digests calculated) in
592 SMIME_write_PKCS7().
593 [Steve Henson]
594
595 *) Add full support for -rpath/-R, both in shared libraries and
596 applications, at least on the platforms where it's known how
597 to do it.
598 [Richard Levitte]
599
600 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
601 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
602 will now compute a table of multiples of the generator that
603 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
604 faster (notably in the case of a single point multiplication,
605 scalar * generator).
606 [Nils Larsch, Bodo Moeller]
607
608 *) IPv6 support for certificate extensions. The various extensions
609 which use the IP:a.b.c.d can now take IPv6 addresses using the
610 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
611 correctly.
612 [Steve Henson]
613
614 *) Added an ENGINE that implements RSA by performing private key
615 exponentiations with the GMP library. The conversions to and from
616 GMP's mpz_t format aren't optimised nor are any montgomery forms
617 cached, and on x86 it appears OpenSSL's own performance has caught up.
618 However there are likely to be other architectures where GMP could
619 provide a boost. This ENGINE is not built in by default, but it can be
620 specified at Configure time and should be accompanied by the necessary
621 linker additions, eg;
622 ./config -DOPENSSL_USE_GMP -lgmp
623 [Geoff Thorpe]
624
625 *) "openssl engine" will not display ENGINE/DSO load failure errors when
626 testing availability of engines with "-t" - the old behaviour is
627 produced by increasing the feature's verbosity with "-tt".
628 [Geoff Thorpe]
629
630 *) ECDSA routines: under certain error conditions uninitialized BN objects
631 could be freed. Solution: make sure initialization is performed early
632 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
633 via PR#459)
634 [Lutz Jaenicke]
635
636 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
637 and DH_METHOD (eg. by ENGINE implementations) to override the normal
638 software implementations. For DSA and DH, parameter generation can
639 also be overriden by providing the appropriate method callbacks.
640 [Geoff Thorpe]
641
642 *) Change the "progress" mechanism used in key-generation and
643 primality testing to functions that take a new BN_GENCB pointer in
644 place of callback/argument pairs. The new API functions have "_ex"
645 postfixes and the older functions are reimplemented as wrappers for
646 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
647 declarations of the old functions to help (graceful) attempts to
648 migrate to the new functions. Also, the new key-generation API
649 functions operate on a caller-supplied key-structure and return
650 success/failure rather than returning a key or NULL - this is to
651 help make "keygen" another member function of RSA_METHOD etc.
652
653 Example for using the new callback interface:
654
655 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
656 void *my_arg = ...;
657 BN_GENCB my_cb;
658
659 BN_GENCB_set(&my_cb, my_callback, my_arg);
660
661 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
662 /* For the meaning of a, b in calls to my_callback(), see the
663 * documentation of the function that calls the callback.
664 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
665 * my_callback should return 1 if it wants BN_is_prime_ex()
666 * to continue, or 0 to stop.
667 */
668
669 [Geoff Thorpe]
670
671 *) Change the ZLIB compression method to be stateful, and make it
672 available to TLS with the number defined in
673 draft-ietf-tls-compression-04.txt.
674 [Richard Levitte]
675
676 *) Add the ASN.1 structures and functions for CertificatePair, which
677 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
678
679 CertificatePair ::= SEQUENCE {
680 forward [0] Certificate OPTIONAL,
681 reverse [1] Certificate OPTIONAL,
682 -- at least one of the pair shall be present -- }
683
684 Also implement the PEM functions to read and write certificate
685 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
686
687 This needed to be defined, mostly for the sake of the LDAP
688 attribute crossCertificatePair, but may prove useful elsewhere as
689 well.
690 [Richard Levitte]
691
692 *) Make it possible to inhibit symlinking of shared libraries in
693 Makefile.shared, for Cygwin's sake.
694 [Richard Levitte]
695
696 *) Extend the BIGNUM API by creating a function
697 void BN_set_negative(BIGNUM *a, int neg);
698 and a macro that behave like
699 int BN_is_negative(const BIGNUM *a);
700
701 to avoid the need to access 'a->neg' directly in applications.
702 [Nils Larsch]
703
704 *) Implement fast modular reduction for pseudo-Mersenne primes
705 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
706 EC_GROUP_new_curve_GFp() will now automatically use this
707 if applicable.
708 [Nils Larsch <nla@trustcenter.de>]
709
710 *) Add new lock type (CRYPTO_LOCK_BN).
711 [Bodo Moeller]
712
713 *) Change the ENGINE framework to automatically load engines
714 dynamically from specific directories unless they could be
715 found to already be built in or loaded. Move all the
716 current engines except for the cryptodev one to a new
717 directory engines/.
718 The engines in engines/ are built as shared libraries if
719 the "shared" options was given to ./Configure or ./config.
720 Otherwise, they are inserted in libcrypto.a.
721 /usr/local/ssl/engines is the default directory for dynamic
722 engines, but that can be overriden at configure time through
723 the usual use of --prefix and/or --openssldir, and at run
724 time with the environment variable OPENSSL_ENGINES.
725 [Geoff Thorpe and Richard Levitte]
726
727 *) Add Makefile.shared, a helper makefile to build shared
728 libraries. Addapt Makefile.org.
729 [Richard Levitte]
730
731 *) Add version info to Win32 DLLs.
732 [Peter 'Luna' Runestig" <peter@runestig.com>]
733
734 *) Add new 'medium level' PKCS#12 API. Certificates and keys
735 can be added using this API to created arbitrary PKCS#12
736 files while avoiding the low level API.
737
738 New options to PKCS12_create(), key or cert can be NULL and
739 will then be omitted from the output file. The encryption
740 algorithm NIDs can be set to -1 for no encryption, the mac
741 iteration count can be set to 0 to omit the mac.
742
743 Enhance pkcs12 utility by making the -nokeys and -nocerts
744 options work when creating a PKCS#12 file. New option -nomac
745 to omit the mac, NONE can be set for an encryption algorithm.
746 New code is modified to use the enhanced PKCS12_create()
747 instead of the low level API.
748 [Steve Henson]
749
750 *) Extend ASN1 encoder to support indefinite length constructed
751 encoding. This can output sequences tags and octet strings in
752 this form. Modify pk7_asn1.c to support indefinite length
753 encoding. This is experimental and needs additional code to
754 be useful, such as an ASN1 bio and some enhanced streaming
755 PKCS#7 code.
756
757 Extend template encode functionality so that tagging is passed
758 down to the template encoder.
759 [Steve Henson]
760
761 *) Let 'openssl req' fail if an argument to '-newkey' is not
762 recognized instead of using RSA as a default.
763 [Bodo Moeller]
764
765 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
766 As these are not official, they are not included in "ALL";
767 the "ECCdraft" ciphersuite group alias can be used to select them.
768 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
769
770 *) Add ECDH engine support.
771 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
772
773 *) Add ECDH in new directory crypto/ecdh/.
774 [Douglas Stebila (Sun Microsystems Laboratories)]
775
776 *) Let BN_rand_range() abort with an error after 100 iterations
777 without success (which indicates a broken PRNG).
778 [Bodo Moeller]
779
780 *) Change BN_mod_sqrt() so that it verifies that the input value
781 is really the square of the return value. (Previously,
782 BN_mod_sqrt would show GIGO behaviour.)
783 [Bodo Moeller]
784
785 *) Add named elliptic curves over binary fields from X9.62, SECG,
786 and WAP/WTLS; add OIDs that were still missing.
787
788 [Sheueling Chang Shantz and Douglas Stebila
789 (Sun Microsystems Laboratories)]
790
791 *) Extend the EC library for elliptic curves over binary fields
792 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
793 New EC_METHOD:
794
795 EC_GF2m_simple_method
796
797 New API functions:
798
799 EC_GROUP_new_curve_GF2m
800 EC_GROUP_set_curve_GF2m
801 EC_GROUP_get_curve_GF2m
802 EC_POINT_set_affine_coordinates_GF2m
803 EC_POINT_get_affine_coordinates_GF2m
804 EC_POINT_set_compressed_coordinates_GF2m
805
806 Point compression for binary fields is disabled by default for
807 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
808 enable it).
809
810 As binary polynomials are represented as BIGNUMs, various members
811 of the EC_GROUP and EC_POINT data structures can be shared
812 between the implementations for prime fields and binary fields;
813 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
814 are essentially identical to their ..._GFp counterparts.
815 (For simplicity, the '..._GFp' prefix has been dropped from
816 various internal method names.)
817
818 An internal 'field_div' method (similar to 'field_mul' and
819 'field_sqr') has been added; this is used only for binary fields.
820
821 [Sheueling Chang Shantz and Douglas Stebila
822 (Sun Microsystems Laboratories)]
823
824 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
825 through methods ('mul', 'precompute_mult').
826
827 The generic implementations (now internally called 'ec_wNAF_mul'
828 and 'ec_wNAF_precomputed_mult') remain the default if these
829 methods are undefined.
830
831 [Sheueling Chang Shantz and Douglas Stebila
832 (Sun Microsystems Laboratories)]
833
834 *) New function EC_GROUP_get_degree, which is defined through
835 EC_METHOD. For curves over prime fields, this returns the bit
836 length of the modulus.
837
838 [Sheueling Chang Shantz and Douglas Stebila
839 (Sun Microsystems Laboratories)]
840
841 *) New functions EC_GROUP_dup, EC_POINT_dup.
842 (These simply call ..._new and ..._copy).
843
844 [Sheueling Chang Shantz and Douglas Stebila
845 (Sun Microsystems Laboratories)]
846
847 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
848 Polynomials are represented as BIGNUMs (where the sign bit is not
849 used) in the following functions [macros]:
850
851 BN_GF2m_add
852 BN_GF2m_sub [= BN_GF2m_add]
853 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
854 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
855 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
856 BN_GF2m_mod_inv
857 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
858 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
859 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
860 BN_GF2m_cmp [= BN_ucmp]
861
862 (Note that only the 'mod' functions are actually for fields GF(2^m).
863 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
864
865 For some functions, an the irreducible polynomial defining a
866 field can be given as an 'unsigned int[]' with strictly
867 decreasing elements giving the indices of those bits that are set;
868 i.e., p[] represents the polynomial
869 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
870 where
871 p[0] > p[1] > ... > p[k] = 0.
872 This applies to the following functions:
873
874 BN_GF2m_mod_arr
875 BN_GF2m_mod_mul_arr
876 BN_GF2m_mod_sqr_arr
877 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
878 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
879 BN_GF2m_mod_exp_arr
880 BN_GF2m_mod_sqrt_arr
881 BN_GF2m_mod_solve_quad_arr
882 BN_GF2m_poly2arr
883 BN_GF2m_arr2poly
884
885 Conversion can be performed by the following functions:
886
887 BN_GF2m_poly2arr
888 BN_GF2m_arr2poly
889
890 bntest.c has additional tests for binary polynomial arithmetic.
891
892 Two implementations for BN_GF2m_mod_div() are available.
893 The default algorithm simply uses BN_GF2m_mod_inv() and
894 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
895 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
896 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
897
898 [Sheueling Chang Shantz and Douglas Stebila
899 (Sun Microsystems Laboratories)]
900
901 *) Add new error code 'ERR_R_DISABLED' that can be used when some
902 functionality is disabled at compile-time.
903 [Douglas Stebila <douglas.stebila@sun.com>]
904
905 *) Change default behaviour of 'openssl asn1parse' so that more
906 information is visible when viewing, e.g., a certificate:
907
908 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
909 mode the content of non-printable OCTET STRINGs is output in a
910 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
911 avoid the appearance of a printable string.
912 [Nils Larsch <nla@trustcenter.de>]
913
914 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
915 functions
916 EC_GROUP_set_asn1_flag()
917 EC_GROUP_get_asn1_flag()
918 EC_GROUP_set_point_conversion_form()
919 EC_GROUP_get_point_conversion_form()
920 These control ASN1 encoding details:
921 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
922 has been set to OPENSSL_EC_NAMED_CURVE.
923 - Points are encoded in uncompressed form by default; options for
924 asn1_for are as for point2oct, namely
925 POINT_CONVERSION_COMPRESSED
926 POINT_CONVERSION_UNCOMPRESSED
927 POINT_CONVERSION_HYBRID
928
929 Also add 'seed' and 'seed_len' members to EC_GROUP with access
930 functions
931 EC_GROUP_set_seed()
932 EC_GROUP_get0_seed()
933 EC_GROUP_get_seed_len()
934 This is used only for ASN1 purposes (so far).
935 [Nils Larsch <nla@trustcenter.de>]
936
937 *) Add 'field_type' member to EC_METHOD, which holds the NID
938 of the appropriate field type OID. The new function
939 EC_METHOD_get_field_type() returns this value.
940 [Nils Larsch <nla@trustcenter.de>]
941
942 *) Add functions
943 EC_POINT_point2bn()
944 EC_POINT_bn2point()
945 EC_POINT_point2hex()
946 EC_POINT_hex2point()
947 providing useful interfaces to EC_POINT_point2oct() and
948 EC_POINT_oct2point().
949 [Nils Larsch <nla@trustcenter.de>]
950
951 *) Change internals of the EC library so that the functions
952 EC_GROUP_set_generator()
953 EC_GROUP_get_generator()
954 EC_GROUP_get_order()
955 EC_GROUP_get_cofactor()
956 are implemented directly in crypto/ec/ec_lib.c and not dispatched
957 to methods, which would lead to unnecessary code duplication when
958 adding different types of curves.
959 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
960
961 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
962 arithmetic, and such that modified wNAFs are generated
963 (which avoid length expansion in many cases).
964 [Bodo Moeller]
965
966 *) Add a function EC_GROUP_check_discriminant() (defined via
967 EC_METHOD) that verifies that the curve discriminant is non-zero.
968
969 Add a function EC_GROUP_check() that makes some sanity tests
970 on a EC_GROUP, its generator and order. This includes
971 EC_GROUP_check_discriminant().
972 [Nils Larsch <nla@trustcenter.de>]
973
974 *) Add ECDSA in new directory crypto/ecdsa/.
975
976 Add applications 'openssl ecparam' and 'openssl ecdsa'
977 (these are based on 'openssl dsaparam' and 'openssl dsa').
978
979 ECDSA support is also included in various other files across the
980 library. Most notably,
981 - 'openssl req' now has a '-newkey ecdsa:file' option;
982 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
983 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
984 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
985 them suitable for ECDSA where domain parameters must be
986 extracted before the specific public key;
987 - ECDSA engine support has been added.
988 [Nils Larsch <nla@trustcenter.de>]
989
990 *) Include some named elliptic curves, and add OIDs from X9.62,
991 SECG, and WAP/WTLS. Each curve can be obtained from the new
992 function
993 EC_GROUP_new_by_curve_name(),
994 and the list of available named curves can be obtained with
995 EC_get_builtin_curves().
996 Also add a 'curve_name' member to EC_GROUP objects, which can be
997 accessed via
998 EC_GROUP_set_curve_name()
999 EC_GROUP_get_curve_name()
1000 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1001
1002 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1003 was actually never needed) and in BN_mul(). The removal in BN_mul()
1004 required a small change in bn_mul_part_recursive() and the addition
1005 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1006 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1007 bn_sub_words() and bn_add_words() except they take arrays with
1008 differing sizes.
1009 [Richard Levitte]
1010
1011 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1012
1013 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1014 The value now differs depending on if you build for FIPS or not.
1015 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1016 safely run with a non-FIPSed libcrypto, as it may crash because of
1017 the difference induced by this change.
1018 [Andy Polyakov]
1019
1020 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1021
1022 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1023 (part of SSL_OP_ALL). This option used to disable the
1024 countermeasure against man-in-the-middle protocol-version
1025 rollback in the SSL 2.0 server implementation, which is a bad
1026 idea. (CVE-2005-2969)
1027
1028 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1029 for Information Security, National Institute of Advanced Industrial
1030 Science and Technology [AIST], Japan)]
1031
1032 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1033 mainly for FIPS compliance and not fully integrated at this stage.
1034 [Steve Henson]
1035
1036 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1037 the exponentiation using a fixed-length exponent. (Otherwise,
1038 the information leaked through timing could expose the secret key
1039 after many signatures; cf. Bleichenbacher's attack on DSA with
1040 biased k.)
1041 [Bodo Moeller]
1042
1043 *) Make a new fixed-window mod_exp implementation the default for
1044 RSA, DSA, and DH private-key operations so that the sequence of
1045 squares and multiplies and the memory access pattern are
1046 independent of the particular secret key. This will mitigate
1047 cache-timing and potential related attacks.
1048
1049 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1050 and this is automatically used by BN_mod_exp_mont() if the new flag
1051 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1052 will use this BN flag for private exponents unless the flag
1053 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1054 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1055
1056 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1057
1058 *) Change the client implementation for SSLv23_method() and
1059 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1060 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1061 (Previously, the SSL 2.0 backwards compatible Client Hello
1062 message format would be used even with SSL_OP_NO_SSLv2.)
1063 [Bodo Moeller]
1064
1065 *) Add support for smime-type MIME parameter in S/MIME messages which some
1066 clients need.
1067 [Steve Henson]
1068
1069 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1070 a threadsafe manner. Modify rsa code to use new function and add calls
1071 to dsa and dh code (which had race conditions before).
1072 [Steve Henson]
1073
1074 *) Include the fixed error library code in the C error file definitions
1075 instead of fixing them up at runtime. This keeps the error code
1076 structures constant.
1077 [Steve Henson]
1078
1079 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1080
1081 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1082 OpenSSL 0.9.8.]
1083
1084 *) Fixes for newer kerberos headers. NB: the casts are needed because
1085 the 'length' field is signed on one version and unsigned on another
1086 with no (?) obvious way to tell the difference, without these VC++
1087 complains. Also the "definition" of FAR (blank) is no longer included
1088 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1089 some needed definitions.
1090 [Steve Henson]
1091
1092 *) Undo Cygwin change.
1093 [Ulf Möller]
1094
1095 *) Added support for proxy certificates according to RFC 3820.
1096 Because they may be a security thread to unaware applications,
1097 they must be explicitely allowed in run-time. See
1098 docs/HOWTO/proxy_certificates.txt for further information.
1099 [Richard Levitte]
1100
1101 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1102
1103 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1104 server and client random values. Previously
1105 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1106 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1107
1108 This change has negligible security impact because:
1109
1110 1. Server and client random values still have 24 bytes of pseudo random
1111 data.
1112
1113 2. Server and client random values are sent in the clear in the initial
1114 handshake.
1115
1116 3. The master secret is derived using the premaster secret (48 bytes in
1117 size for static RSA ciphersuites) as well as client server and random
1118 values.
1119
1120 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1121 to our attention.
1122
1123 [Stephen Henson, reported by UK NISCC]
1124
1125 *) Use Windows randomness collection on Cygwin.
1126 [Ulf Möller]
1127
1128 *) Fix hang in EGD/PRNGD query when communication socket is closed
1129 prematurely by EGD/PRNGD.
1130 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1131
1132 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1133 [Steve Henson]
1134
1135 *) Back-port of selected performance improvements from development
1136 branch, as well as improved support for PowerPC platforms.
1137 [Andy Polyakov]
1138
1139 *) Add lots of checks for memory allocation failure, error codes to indicate
1140 failure and freeing up memory if a failure occurs.
1141 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1142
1143 *) Add new -passin argument to dgst.
1144 [Steve Henson]
1145
1146 *) Perform some character comparisons of different types in X509_NAME_cmp:
1147 this is needed for some certificates that reencode DNs into UTF8Strings
1148 (in violation of RFC3280) and can't or wont issue name rollover
1149 certificates.
1150 [Steve Henson]
1151
1152 *) Make an explicit check during certificate validation to see that
1153 the CA setting in each certificate on the chain is correct. As a
1154 side effect always do the following basic checks on extensions,
1155 not just when there's an associated purpose to the check:
1156
1157 - if there is an unhandled critical extension (unless the user
1158 has chosen to ignore this fault)
1159 - if the path length has been exceeded (if one is set at all)
1160 - that certain extensions fit the associated purpose (if one has
1161 been given)
1162 [Richard Levitte]
1163
1164 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1165
1166 *) Avoid a race condition when CRLs are checked in a multi threaded
1167 environment. This would happen due to the reordering of the revoked
1168 entries during signature checking and serial number lookup. Now the
1169 encoding is cached and the serial number sort performed under a lock.
1170 Add new STACK function sk_is_sorted().
1171 [Steve Henson]
1172
1173 *) Add Delta CRL to the extension code.
1174 [Steve Henson]
1175
1176 *) Various fixes to s3_pkt.c so alerts are sent properly.
1177 [David Holmes <d.holmes@f5.com>]
1178
1179 *) Reduce the chances of duplicate issuer name and serial numbers (in
1180 violation of RFC3280) using the OpenSSL certificate creation utilities.
1181 This is done by creating a random 64 bit value for the initial serial
1182 number when a serial number file is created or when a self signed
1183 certificate is created using 'openssl req -x509'. The initial serial
1184 number file is created using 'openssl x509 -next_serial' in CA.pl
1185 rather than being initialized to 1.
1186 [Steve Henson]
1187
1188 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1189
1190 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1191 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1192 [Joe Orton, Steve Henson]
1193
1194 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1195 (CVE-2004-0112)
1196 [Joe Orton, Steve Henson]
1197
1198 *) Make it possible to have multiple active certificates with the same
1199 subject in the CA index file. This is done only if the keyword
1200 'unique_subject' is set to 'no' in the main CA section (default
1201 if 'CA_default') of the configuration file. The value is saved
1202 with the database itself in a separate index attribute file,
1203 named like the index file with '.attr' appended to the name.
1204 [Richard Levitte]
1205
1206 *) X509 verify fixes. Disable broken certificate workarounds when
1207 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1208 keyUsage extension present. Don't accept CRLs with unhandled critical
1209 extensions: since verify currently doesn't process CRL extensions this
1210 rejects a CRL with *any* critical extensions. Add new verify error codes
1211 for these cases.
1212 [Steve Henson]
1213
1214 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1215 A clarification of RFC2560 will require the use of OCTET STRINGs and
1216 some implementations cannot handle the current raw format. Since OpenSSL
1217 copies and compares OCSP nonces as opaque blobs without any attempt at
1218 parsing them this should not create any compatibility issues.
1219 [Steve Henson]
1220
1221 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1222 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1223 this HMAC (and other) operations are several times slower than OpenSSL
1224 < 0.9.7.
1225 [Steve Henson]
1226
1227 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1228 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1229
1230 *) Use the correct content when signing type "other".
1231 [Steve Henson]
1232
1233 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1234
1235 *) Fix various bugs revealed by running the NISCC test suite:
1236
1237 Stop out of bounds reads in the ASN1 code when presented with
1238 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1239
1240 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1241
1242 If verify callback ignores invalid public key errors don't try to check
1243 certificate signature with the NULL public key.
1244
1245 [Steve Henson]
1246
1247 *) New -ignore_err option in ocsp application to stop the server
1248 exiting on the first error in a request.
1249 [Steve Henson]
1250
1251 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1252 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1253 specifications.
1254 [Steve Henson]
1255
1256 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1257 extra data after the compression methods not only for TLS 1.0
1258 but also for SSL 3.0 (as required by the specification).
1259 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1260
1261 *) Change X509_certificate_type() to mark the key as exported/exportable
1262 when it's 512 *bits* long, not 512 bytes.
1263 [Richard Levitte]
1264
1265 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1266 blocks during encryption.
1267 [Richard Levitte]
1268
1269 *) Various fixes to base64 BIO and non blocking I/O. On write
1270 flushes were not handled properly if the BIO retried. On read
1271 data was not being buffered properly and had various logic bugs.
1272 This also affects blocking I/O when the data being decoded is a
1273 certain size.
1274 [Steve Henson]
1275
1276 *) Various S/MIME bugfixes and compatibility changes:
1277 output correct application/pkcs7 MIME type if
1278 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1279 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1280 of files as .eml work). Correctly handle very long lines in MIME
1281 parser.
1282 [Steve Henson]
1283
1284 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1285
1286 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1287 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1288 a protocol version number mismatch like a decryption error
1289 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1290 [Bodo Moeller]
1291
1292 *) Turn on RSA blinding by default in the default implementation
1293 to avoid a timing attack. Applications that don't want it can call
1294 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1295 They would be ill-advised to do so in most cases.
1296 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1297
1298 *) Change RSA blinding code so that it works when the PRNG is not
1299 seeded (in this case, the secret RSA exponent is abused as
1300 an unpredictable seed -- if it is not unpredictable, there
1301 is no point in blinding anyway). Make RSA blinding thread-safe
1302 by remembering the creator's thread ID in rsa->blinding and
1303 having all other threads use local one-time blinding factors
1304 (this requires more computation than sharing rsa->blinding, but
1305 avoids excessive locking; and if an RSA object is not shared
1306 between threads, blinding will still be very fast).
1307 [Bodo Moeller]
1308
1309 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1310 ENGINE as defaults for all supported algorithms irrespective of
1311 the 'flags' parameter. 'flags' is now honoured, so applications
1312 should make sure they are passing it correctly.
1313 [Geoff Thorpe]
1314
1315 *) Target "mingw" now allows native Windows code to be generated in
1316 the Cygwin environment as well as with the MinGW compiler.
1317 [Ulf Moeller]
1318
1319 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1320
1321 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1322 via timing by performing a MAC computation even if incorrrect
1323 block cipher padding has been found. This is a countermeasure
1324 against active attacks where the attacker has to distinguish
1325 between bad padding and a MAC verification error. (CVE-2003-0078)
1326
1327 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1328 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1329 Martin Vuagnoux (EPFL, Ilion)]
1330
1331 *) Make the no-err option work as intended. The intention with no-err
1332 is not to have the whole error stack handling routines removed from
1333 libcrypto, it's only intended to remove all the function name and
1334 reason texts, thereby removing some of the footprint that may not
1335 be interesting if those errors aren't displayed anyway.
1336
1337 NOTE: it's still possible for any application or module to have it's
1338 own set of error texts inserted. The routines are there, just not
1339 used by default when no-err is given.
1340 [Richard Levitte]
1341
1342 *) Add support for FreeBSD on IA64.
1343 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1344
1345 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1346 Kerberos function mit_des_cbc_cksum(). Before this change,
1347 the value returned by DES_cbc_cksum() was like the one from
1348 mit_des_cbc_cksum(), except the bytes were swapped.
1349 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1350
1351 *) Allow an application to disable the automatic SSL chain building.
1352 Before this a rather primitive chain build was always performed in
1353 ssl3_output_cert_chain(): an application had no way to send the
1354 correct chain if the automatic operation produced an incorrect result.
1355
1356 Now the chain builder is disabled if either:
1357
1358 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1359
1360 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1361
1362 The reasoning behind this is that an application would not want the
1363 auto chain building to take place if extra chain certificates are
1364 present and it might also want a means of sending no additional
1365 certificates (for example the chain has two certificates and the
1366 root is omitted).
1367 [Steve Henson]
1368
1369 *) Add the possibility to build without the ENGINE framework.
1370 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1371
1372 *) Under Win32 gmtime() can return NULL: check return value in
1373 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1374 [Steve Henson]
1375
1376 *) DSA routines: under certain error conditions uninitialized BN objects
1377 could be freed. Solution: make sure initialization is performed early
1378 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1379 Nils Larsch <nla@trustcenter.de> via PR#459)
1380 [Lutz Jaenicke]
1381
1382 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1383 checked on reconnect on the client side, therefore session resumption
1384 could still fail with a "ssl session id is different" error. This
1385 behaviour is masked when SSL_OP_ALL is used due to
1386 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1387 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1388 followup to PR #377.
1389 [Lutz Jaenicke]
1390
1391 *) IA-32 assembler support enhancements: unified ELF targets, support
1392 for SCO/Caldera platforms, fix for Cygwin shared build.
1393 [Andy Polyakov]
1394
1395 *) Add support for FreeBSD on sparc64. As a consequence, support for
1396 FreeBSD on non-x86 processors is separate from x86 processors on
1397 the config script, much like the NetBSD support.
1398 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1399
1400 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1401
1402 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1403 OpenSSL 0.9.7.]
1404
1405 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1406 code (06) was taken as the first octet of the session ID and the last
1407 octet was ignored consequently. As a result SSLv2 client side session
1408 caching could not have worked due to the session ID mismatch between
1409 client and server.
1410 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1411 PR #377.
1412 [Lutz Jaenicke]
1413
1414 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1415 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1416 removed entirely.
1417 [Richard Levitte]
1418
1419 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1420 seems that in spite of existing for more than a year, many application
1421 author have done nothing to provide the necessary callbacks, which
1422 means that this particular engine will not work properly anywhere.
1423 This is a very unfortunate situation which forces us, in the name
1424 of usability, to give the hw_ncipher.c a static lock, which is part
1425 of libcrypto.
1426 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1427 appear in 0.9.8 or later. We EXPECT application authors to have
1428 dealt properly with this when 0.9.8 is released (unless we actually
1429 make such changes in the libcrypto locking code that changes will
1430 have to be made anyway).
1431 [Richard Levitte]
1432
1433 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1434 octets have been read, EOF or an error occurs. Without this change
1435 some truncated ASN1 structures will not produce an error.
1436 [Steve Henson]
1437
1438 *) Disable Heimdal support, since it hasn't been fully implemented.
1439 Still give the possibility to force the use of Heimdal, but with
1440 warnings and a request that patches get sent to openssl-dev.
1441 [Richard Levitte]
1442
1443 *) Add the VC-CE target, introduce the WINCE sysname, and add
1444 INSTALL.WCE and appropriate conditionals to make it build.
1445 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1446
1447 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1448 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1449 edit numbers of the version.
1450 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1451
1452 *) Introduce safe string copy and catenation functions
1453 (BUF_strlcpy() and BUF_strlcat()).
1454 [Ben Laurie (CHATS) and Richard Levitte]
1455
1456 *) Avoid using fixed-size buffers for one-line DNs.
1457 [Ben Laurie (CHATS)]
1458
1459 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1460 resizing buffers containing secrets, and use where appropriate.
1461 [Ben Laurie (CHATS)]
1462
1463 *) Avoid using fixed size buffers for configuration file location.
1464 [Ben Laurie (CHATS)]
1465
1466 *) Avoid filename truncation for various CA files.
1467 [Ben Laurie (CHATS)]
1468
1469 *) Use sizeof in preference to magic numbers.
1470 [Ben Laurie (CHATS)]
1471
1472 *) Avoid filename truncation in cert requests.
1473 [Ben Laurie (CHATS)]
1474
1475 *) Add assertions to check for (supposedly impossible) buffer
1476 overflows.
1477 [Ben Laurie (CHATS)]
1478
1479 *) Don't cache truncated DNS entries in the local cache (this could
1480 potentially lead to a spoofing attack).
1481 [Ben Laurie (CHATS)]
1482
1483 *) Fix various buffers to be large enough for hex/decimal
1484 representations in a platform independent manner.
1485 [Ben Laurie (CHATS)]
1486
1487 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1488 resizing buffers containing secrets, and use where appropriate.
1489 [Ben Laurie (CHATS)]
1490
1491 *) Add BIO_indent() to avoid much slightly worrying code to do
1492 indents.
1493 [Ben Laurie (CHATS)]
1494
1495 *) Convert sprintf()/BIO_puts() to BIO_printf().
1496 [Ben Laurie (CHATS)]
1497
1498 *) buffer_gets() could terminate with the buffer only half
1499 full. Fixed.
1500 [Ben Laurie (CHATS)]
1501
1502 *) Add assertions to prevent user-supplied crypto functions from
1503 overflowing internal buffers by having large block sizes, etc.
1504 [Ben Laurie (CHATS)]
1505
1506 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1507 unconditionally).
1508 [Ben Laurie (CHATS)]
1509
1510 *) Eliminate unused copy of key in RC4.
1511 [Ben Laurie (CHATS)]
1512
1513 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1514 [Ben Laurie (CHATS)]
1515
1516 *) Fix off-by-one error in EGD path.
1517 [Ben Laurie (CHATS)]
1518
1519 *) If RANDFILE path is too long, ignore instead of truncating.
1520 [Ben Laurie (CHATS)]
1521
1522 *) Eliminate unused and incorrectly sized X.509 structure
1523 CBCParameter.
1524 [Ben Laurie (CHATS)]
1525
1526 *) Eliminate unused and dangerous function knumber().
1527 [Ben Laurie (CHATS)]
1528
1529 *) Eliminate unused and dangerous structure, KSSL_ERR.
1530 [Ben Laurie (CHATS)]
1531
1532 *) Protect against overlong session ID context length in an encoded
1533 session object. Since these are local, this does not appear to be
1534 exploitable.
1535 [Ben Laurie (CHATS)]
1536
1537 *) Change from security patch (see 0.9.6e below) that did not affect
1538 the 0.9.6 release series:
1539
1540 Remote buffer overflow in SSL3 protocol - an attacker could
1541 supply an oversized master key in Kerberos-enabled versions.
1542 (CVE-2002-0657)
1543 [Ben Laurie (CHATS)]
1544
1545 *) Change the SSL kerb5 codes to match RFC 2712.
1546 [Richard Levitte]
1547
1548 *) Make -nameopt work fully for req and add -reqopt switch.
1549 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1550
1551 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1552 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1553
1554 *) Make sure tests can be performed even if the corresponding algorithms
1555 have been removed entirely. This was also the last step to make
1556 OpenSSL compilable with DJGPP under all reasonable conditions.
1557 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1558
1559 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1560 to allow version independent disabling of normally unselected ciphers,
1561 which may be activated as a side-effect of selecting a single cipher.
1562
1563 (E.g., cipher list string "RSA" enables ciphersuites that are left
1564 out of "ALL" because they do not provide symmetric encryption.
1565 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1566 [Lutz Jaenicke, Bodo Moeller]
1567
1568 *) Add appropriate support for separate platform-dependent build
1569 directories. The recommended way to make a platform-dependent
1570 build directory is the following (tested on Linux), maybe with
1571 some local tweaks:
1572
1573 # Place yourself outside of the OpenSSL source tree. In
1574 # this example, the environment variable OPENSSL_SOURCE
1575 # is assumed to contain the absolute OpenSSL source directory.
1576 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1577 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1578 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1579 mkdir -p `dirname $F`
1580 ln -s $OPENSSL_SOURCE/$F $F
1581 done
1582
1583 To be absolutely sure not to disturb the source tree, a "make clean"
1584 is a good thing. If it isn't successfull, don't worry about it,
1585 it probably means the source directory is very clean.
1586 [Richard Levitte]
1587
1588 *) Make sure any ENGINE control commands make local copies of string
1589 pointers passed to them whenever necessary. Otherwise it is possible
1590 the caller may have overwritten (or deallocated) the original string
1591 data when a later ENGINE operation tries to use the stored values.
1592 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1593
1594 *) Improve diagnostics in file reading and command-line digests.
1595 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1596
1597 *) Add AES modes CFB and OFB to the object database. Correct an
1598 error in AES-CFB decryption.
1599 [Richard Levitte]
1600
1601 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1602 allows existing EVP_CIPHER_CTX structures to be reused after
1603 calling EVP_*Final(). This behaviour is used by encryption
1604 BIOs and some applications. This has the side effect that
1605 applications must explicitly clean up cipher contexts with
1606 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1607 [Steve Henson]
1608
1609 *) Check the values of dna and dnb in bn_mul_recursive before calling
1610 bn_mul_comba (a non zero value means the a or b arrays do not contain
1611 n2 elements) and fallback to bn_mul_normal if either is not zero.
1612 [Steve Henson]
1613
1614 *) Fix escaping of non-ASCII characters when using the -subj option
1615 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1616 [Lutz Jaenicke]
1617
1618 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1619 form for "surname", serialNumber has no short form.
1620 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1621 therefore remove "mail" short name for "internet 7".
1622 The OID for unique identifiers in X509 certificates is
1623 x500UniqueIdentifier, not uniqueIdentifier.
1624 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1625 [Lutz Jaenicke]
1626
1627 *) Add an "init" command to the ENGINE config module and auto initialize
1628 ENGINEs. Without any "init" command the ENGINE will be initialized
1629 after all ctrl commands have been executed on it. If init=1 the
1630 ENGINE is initailized at that point (ctrls before that point are run
1631 on the uninitialized ENGINE and after on the initialized one). If
1632 init=0 then the ENGINE will not be iniatialized at all.
1633 [Steve Henson]
1634
1635 *) Fix the 'app_verify_callback' interface so that the user-defined
1636 argument is actually passed to the callback: In the
1637 SSL_CTX_set_cert_verify_callback() prototype, the callback
1638 declaration has been changed from
1639 int (*cb)()
1640 into
1641 int (*cb)(X509_STORE_CTX *,void *);
1642 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1643 i=s->ctx->app_verify_callback(&ctx)
1644 has been changed into
1645 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1646
1647 To update applications using SSL_CTX_set_cert_verify_callback(),
1648 a dummy argument can be added to their callback functions.
1649 [D. K. Smetters <smetters@parc.xerox.com>]
1650
1651 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1652 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1653
1654 *) Add and OPENSSL_LOAD_CONF define which will cause
1655 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1656 This allows older applications to transparently support certain
1657 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1658 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1659 load the config file and OPENSSL_add_all_algorithms_conf() which will
1660 always load it have also been added.
1661 [Steve Henson]
1662
1663 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1664 Adjust NIDs and EVP layer.
1665 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1666
1667 *) Config modules support in openssl utility.
1668
1669 Most commands now load modules from the config file,
1670 though in a few (such as version) this isn't done
1671 because it couldn't be used for anything.
1672
1673 In the case of ca and req the config file used is
1674 the same as the utility itself: that is the -config
1675 command line option can be used to specify an
1676 alternative file.
1677 [Steve Henson]
1678
1679 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1680 use "openssl_conf" if filename is NULL use default openssl config file.
1681 [Steve Henson]
1682
1683 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1684 config section name. Add a new flag to tolerate a missing config file
1685 and move code to CONF_modules_load_file().
1686 [Steve Henson]
1687
1688 *) Support for crypto accelerator cards from Accelerated Encryption
1689 Processing, www.aep.ie. (Use engine 'aep')
1690 The support was copied from 0.9.6c [engine] and adapted/corrected
1691 to work with the new engine framework.
1692 [AEP Inc. and Richard Levitte]
1693
1694 *) Support for SureWare crypto accelerator cards from Baltimore
1695 Technologies. (Use engine 'sureware')
1696 The support was copied from 0.9.6c [engine] and adapted
1697 to work with the new engine framework.
1698 [Richard Levitte]
1699
1700 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1701 make the newer ENGINE framework commands for the CHIL engine work.
1702 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1703
1704 *) Make it possible to produce shared libraries on ReliantUNIX.
1705 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1706
1707 *) Add the configuration target debug-linux-ppro.
1708 Make 'openssl rsa' use the general key loading routines
1709 implemented in apps.c, and make those routines able to
1710 handle the key format FORMAT_NETSCAPE and the variant
1711 FORMAT_IISSGC.
1712 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1713
1714 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1715 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1716
1717 *) Add -keyform to rsautl, and document -engine.
1718 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1719
1720 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1721 BIO_R_NO_SUCH_FILE error code rather than the generic
1722 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1723 [Ben Laurie]
1724
1725 *) Add new functions
1726 ERR_peek_last_error
1727 ERR_peek_last_error_line
1728 ERR_peek_last_error_line_data.
1729 These are similar to
1730 ERR_peek_error
1731 ERR_peek_error_line
1732 ERR_peek_error_line_data,
1733 but report on the latest error recorded rather than the first one
1734 still in the error queue.
1735 [Ben Laurie, Bodo Moeller]
1736
1737 *) default_algorithms option in ENGINE config module. This allows things
1738 like:
1739 default_algorithms = ALL
1740 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1741 [Steve Henson]
1742
1743 *) Prelminary ENGINE config module.
1744 [Steve Henson]
1745
1746 *) New experimental application configuration code.
1747 [Steve Henson]
1748
1749 *) Change the AES code to follow the same name structure as all other
1750 symmetric ciphers, and behave the same way. Move everything to
1751 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1752 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1753
1754 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1755 [Ben Laurie and Theo de Raadt]
1756
1757 *) Add option to output public keys in req command.
1758 [Massimiliano Pala madwolf@openca.org]
1759
1760 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1761 (up to about 10% better than before for P-192 and P-224).
1762 [Bodo Moeller]
1763
1764 *) New functions/macros
1765
1766 SSL_CTX_set_msg_callback(ctx, cb)
1767 SSL_CTX_set_msg_callback_arg(ctx, arg)
1768 SSL_set_msg_callback(ssl, cb)
1769 SSL_set_msg_callback_arg(ssl, arg)
1770
1771 to request calling a callback function
1772
1773 void cb(int write_p, int version, int content_type,
1774 const void *buf, size_t len, SSL *ssl, void *arg)
1775
1776 whenever a protocol message has been completely received
1777 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1778 protocol version according to which the SSL library interprets
1779 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1780 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1781 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1782 specification (change_cipher_spec(20), alert(21), handshake(22)).
1783 'buf' and 'len' point to the actual message, 'ssl' to the
1784 SSL object, and 'arg' is the application-defined value set by
1785 SSL[_CTX]_set_msg_callback_arg().
1786
1787 'openssl s_client' and 'openssl s_server' have new '-msg' options
1788 to enable a callback that displays all protocol messages.
1789 [Bodo Moeller]
1790
1791 *) Change the shared library support so shared libraries are built as
1792 soon as the corresponding static library is finished, and thereby get
1793 openssl and the test programs linked against the shared library.
1794 This still only happens when the keyword "shard" has been given to
1795 the configuration scripts.
1796
1797 NOTE: shared library support is still an experimental thing, and
1798 backward binary compatibility is still not guaranteed.
1799 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1800
1801 *) Add support for Subject Information Access extension.
1802 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1803
1804 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1805 additional bytes when new memory had to be allocated, not just
1806 when reusing an existing buffer.
1807 [Bodo Moeller]
1808
1809 *) New command line and configuration option 'utf8' for the req command.
1810 This allows field values to be specified as UTF8 strings.
1811 [Steve Henson]
1812
1813 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1814 runs for the former and machine-readable output for the latter.
1815 [Ben Laurie]
1816
1817 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1818 of the e-mail address in the DN (i.e., it will go into a certificate
1819 extension only). The new configuration file option 'email_in_dn = no'
1820 has the same effect.
1821 [Massimiliano Pala madwolf@openca.org]
1822
1823 *) Change all functions with names starting with des_ to be starting
1824 with DES_ instead. Add wrappers that are compatible with libdes,
1825 but are named _ossl_old_des_*. Finally, add macros that map the
1826 des_* symbols to the corresponding _ossl_old_des_* if libdes
1827 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1828 desired, the des_* symbols will be mapped to DES_*, with one
1829 exception.
1830
1831 Since we provide two compatibility mappings, the user needs to
1832 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1833 compatibility is desired. The default (i.e., when that macro
1834 isn't defined) is OpenSSL 0.9.6c compatibility.
1835
1836 There are also macros that enable and disable the support of old
1837 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1838 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1839 are defined, the default will apply: to support the old des routines.
1840
1841 In either case, one must include openssl/des.h to get the correct
1842 definitions. Do not try to just include openssl/des_old.h, that
1843 won't work.
1844
1845 NOTE: This is a major break of an old API into a new one. Software
1846 authors are encouraged to switch to the DES_ style functions. Some
1847 time in the future, des_old.h and the libdes compatibility functions
1848 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1849 default), and then completely removed.
1850 [Richard Levitte]
1851
1852 *) Test for certificates which contain unsupported critical extensions.
1853 If such a certificate is found during a verify operation it is
1854 rejected by default: this behaviour can be overridden by either
1855 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1856 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1857 X509_supported_extension() has also been added which returns 1 if a
1858 particular extension is supported.
1859 [Steve Henson]
1860
1861 *) Modify the behaviour of EVP cipher functions in similar way to digests
1862 to retain compatibility with existing code.
1863 [Steve Henson]
1864
1865 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1866 compatibility with existing code. In particular the 'ctx' parameter does
1867 not have to be to be initialized before the call to EVP_DigestInit() and
1868 it is tidied up after a call to EVP_DigestFinal(). New function
1869 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1870 EVP_MD_CTX_copy() changed to not require the destination to be
1871 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1872 requires the destination to be valid.
1873
1874 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1875 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1876 [Steve Henson]
1877
1878 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1879 so that complete 'Handshake' protocol structures are kept in memory
1880 instead of overwriting 'msg_type' and 'length' with 'body' data.
1881 [Bodo Moeller]
1882
1883 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1884 [Massimo Santin via Richard Levitte]
1885
1886 *) Major restructuring to the underlying ENGINE code. This includes
1887 reduction of linker bloat, separation of pure "ENGINE" manipulation
1888 (initialisation, etc) from functionality dealing with implementations
1889 of specific crypto iterfaces. This change also introduces integrated
1890 support for symmetric ciphers and digest implementations - so ENGINEs
1891 can now accelerate these by providing EVP_CIPHER and EVP_MD
1892 implementations of their own. This is detailed in crypto/engine/README
1893 as it couldn't be adequately described here. However, there are a few
1894 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1895 were changed in the original introduction of ENGINE code have now
1896 reverted back - the hooking from this code to ENGINE is now a good
1897 deal more passive and at run-time, operations deal directly with
1898 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1899 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1900 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1901 they were not being used by the framework as there is no concept of a
1902 BIGNUM_METHOD and they could not be generalised to the new
1903 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1904 ENGINE_cpy() has been removed as it cannot be consistently defined in
1905 the new code.
1906 [Geoff Thorpe]
1907
1908 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1909 [Steve Henson]
1910
1911 *) Change mkdef.pl to sort symbols that get the same entry number,
1912 and make sure the automatically generated functions ERR_load_*
1913 become part of libeay.num as well.
1914 [Richard Levitte]
1915
1916 *) New function SSL_renegotiate_pending(). This returns true once
1917 renegotiation has been requested (either SSL_renegotiate() call
1918 or HelloRequest/ClientHello receveived from the peer) and becomes
1919 false once a handshake has been completed.
1920 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1921 sends a HelloRequest, but does not ensure that a handshake takes
1922 place. SSL_renegotiate_pending() is useful for checking if the
1923 client has followed the request.)
1924 [Bodo Moeller]
1925
1926 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1927 By default, clients may request session resumption even during
1928 renegotiation (if session ID contexts permit); with this option,
1929 session resumption is possible only in the first handshake.
1930
1931 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1932 more bits available for options that should not be part of
1933 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1934 [Bodo Moeller]
1935
1936 *) Add some demos for certificate and certificate request creation.
1937 [Steve Henson]
1938
1939 *) Make maximum certificate chain size accepted from the peer application
1940 settable (SSL*_get/set_max_cert_list()), as proposed by
1941 "Douglas E. Engert" <deengert@anl.gov>.
1942 [Lutz Jaenicke]
1943
1944 *) Add support for shared libraries for Unixware-7
1945 (Boyd Lynn Gerber <gerberb@zenez.com>).
1946 [Lutz Jaenicke]
1947
1948 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1949 be done prior to destruction. Use this to unload error strings from
1950 ENGINEs that load their own error strings. NB: This adds two new API
1951 functions to "get" and "set" this destroy handler in an ENGINE.
1952 [Geoff Thorpe]
1953
1954 *) Alter all existing ENGINE implementations (except "openssl" and
1955 "openbsd") to dynamically instantiate their own error strings. This
1956 makes them more flexible to be built both as statically-linked ENGINEs
1957 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
1958 Also, add stub code to each that makes building them as self-contained
1959 shared-libraries easier (see README.ENGINE).
1960 [Geoff Thorpe]
1961
1962 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
1963 implementations into applications that are completely implemented in
1964 self-contained shared-libraries. The "dynamic" ENGINE exposes control
1965 commands that can be used to configure what shared-library to load and
1966 to control aspects of the way it is handled. Also, made an update to
1967 the README.ENGINE file that brings its information up-to-date and
1968 provides some information and instructions on the "dynamic" ENGINE
1969 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
1970 [Geoff Thorpe]
1971
1972 *) Make it possible to unload ranges of ERR strings with a new
1973 "ERR_unload_strings" function.
1974 [Geoff Thorpe]
1975
1976 *) Add a copy() function to EVP_MD.
1977 [Ben Laurie]
1978
1979 *) Make EVP_MD routines take a context pointer instead of just the
1980 md_data void pointer.
1981 [Ben Laurie]
1982
1983 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
1984 that the digest can only process a single chunk of data
1985 (typically because it is provided by a piece of
1986 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
1987 is only going to provide a single chunk of data, and hence the
1988 framework needn't accumulate the data for oneshot drivers.
1989 [Ben Laurie]
1990
1991 *) As with "ERR", make it possible to replace the underlying "ex_data"
1992 functions. This change also alters the storage and management of global
1993 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
1994 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
1995 index counters. The API functions that use this state have been changed
1996 to take a "class_index" rather than pointers to the class's local STACK
1997 and counter, and there is now an API function to dynamically create new
1998 classes. This centralisation allows us to (a) plug a lot of the
1999 thread-safety problems that existed, and (b) makes it possible to clean
2000 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2001 such data would previously have always leaked in application code and
2002 workarounds were in place to make the memory debugging turn a blind eye
2003 to it. Application code that doesn't use this new function will still
2004 leak as before, but their memory debugging output will announce it now
2005 rather than letting it slide.
2006
2007 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2008 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2009 has a return value to indicate success or failure.
2010 [Geoff Thorpe]
2011
2012 *) Make it possible to replace the underlying "ERR" functions such that the
2013 global state (2 LHASH tables and 2 locks) is only used by the "default"
2014 implementation. This change also adds two functions to "get" and "set"
2015 the implementation prior to it being automatically set the first time
2016 any other ERR function takes place. Ie. an application can call "get",
2017 pass the return value to a module it has just loaded, and that module
2018 can call its own "set" function using that value. This means the
2019 module's "ERR" operations will use (and modify) the error state in the
2020 application and not in its own statically linked copy of OpenSSL code.
2021 [Geoff Thorpe]
2022
2023 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2024 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2025 the operation, and provides a more encapsulated way for external code
2026 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2027 to use these functions rather than manually incrementing the counts.
2028
2029 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2030 [Geoff Thorpe]
2031
2032 *) Add EVP test program.
2033 [Ben Laurie]
2034
2035 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2036 [Ben Laurie]
2037
2038 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2039 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2040 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2041 These allow a CRL to be built without having to access X509_CRL fields
2042 directly. Modify 'ca' application to use new functions.
2043 [Steve Henson]
2044
2045 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2046 bug workarounds. Rollback attack detection is a security feature.
2047 The problem will only arise on OpenSSL servers when TLSv1 is not
2048 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2049 Software authors not wanting to support TLSv1 will have special reasons
2050 for their choice and can explicitly enable this option.
2051 [Bodo Moeller, Lutz Jaenicke]
2052
2053 *) Rationalise EVP so it can be extended: don't include a union of
2054 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2055 (similar to those existing for EVP_CIPHER_CTX).
2056 Usage example:
2057
2058 EVP_MD_CTX md;
2059
2060 EVP_MD_CTX_init(&md); /* new function call */
2061 EVP_DigestInit(&md, EVP_sha1());
2062 EVP_DigestUpdate(&md, in, len);
2063 EVP_DigestFinal(&md, out, NULL);
2064 EVP_MD_CTX_cleanup(&md); /* new function call */
2065
2066 [Ben Laurie]
2067
2068 *) Make DES key schedule conform to the usual scheme, as well as
2069 correcting its structure. This means that calls to DES functions
2070 now have to pass a pointer to a des_key_schedule instead of a
2071 plain des_key_schedule (which was actually always a pointer
2072 anyway): E.g.,
2073
2074 des_key_schedule ks;
2075
2076 des_set_key_checked(..., &ks);
2077 des_ncbc_encrypt(..., &ks, ...);
2078
2079 (Note that a later change renames 'des_...' into 'DES_...'.)
2080 [Ben Laurie]
2081
2082 *) Initial reduction of linker bloat: the use of some functions, such as
2083 PEM causes large amounts of unused functions to be linked in due to
2084 poor organisation. For example pem_all.c contains every PEM function
2085 which has a knock on effect of linking in large amounts of (unused)
2086 ASN1 code. Grouping together similar functions and splitting unrelated
2087 functions prevents this.
2088 [Steve Henson]
2089
2090 *) Cleanup of EVP macros.
2091 [Ben Laurie]
2092
2093 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2094 correct _ecb suffix.
2095 [Ben Laurie]
2096
2097 *) Add initial OCSP responder support to ocsp application. The
2098 revocation information is handled using the text based index
2099 use by the ca application. The responder can either handle
2100 requests generated internally, supplied in files (for example
2101 via a CGI script) or using an internal minimal server.
2102 [Steve Henson]
2103
2104 *) Add configuration choices to get zlib compression for TLS.
2105 [Richard Levitte]
2106
2107 *) Changes to Kerberos SSL for RFC 2712 compliance:
2108 1. Implemented real KerberosWrapper, instead of just using
2109 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2110 2. Implemented optional authenticator field of KerberosWrapper.
2111
2112 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2113 and authenticator structs; see crypto/krb5/.
2114
2115 Generalized Kerberos calls to support multiple Kerberos libraries.
2116 [Vern Staats <staatsvr@asc.hpc.mil>,
2117 Jeffrey Altman <jaltman@columbia.edu>
2118 via Richard Levitte]
2119
2120 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2121 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2122 values for each of the key sizes rather than having just
2123 parameters (and 'speed' generating keys each time).
2124 [Geoff Thorpe]
2125
2126 *) Speed up EVP routines.
2127 Before:
2128 encrypt
2129 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2130 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2131 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2132 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2133 decrypt
2134 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2135 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2136 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2137 After:
2138 encrypt
2139 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2140 decrypt
2141 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2142 [Ben Laurie]
2143
2144 *) Added the OS2-EMX target.
2145 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2146
2147 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2148 to support NCONF routines in extension code. New function CONF_set_nconf()
2149 to allow functions which take an NCONF to also handle the old LHASH
2150 structure: this means that the old CONF compatible routines can be
2151 retained (in particular wrt extensions) without having to duplicate the
2152 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2153 [Steve Henson]
2154
2155 *) Enhance the general user interface with mechanisms for inner control
2156 and with possibilities to have yes/no kind of prompts.
2157 [Richard Levitte]
2158
2159 *) Change all calls to low level digest routines in the library and
2160 applications to use EVP. Add missing calls to HMAC_cleanup() and
2161 don't assume HMAC_CTX can be copied using memcpy().
2162 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2163
2164 *) Add the possibility to control engines through control names but with
2165 arbitrary arguments instead of just a string.
2166 Change the key loaders to take a UI_METHOD instead of a callback
2167 function pointer. NOTE: this breaks binary compatibility with earlier
2168 versions of OpenSSL [engine].
2169 Adapt the nCipher code for these new conditions and add a card insertion
2170 callback.
2171 [Richard Levitte]
2172
2173 *) Enhance the general user interface with mechanisms to better support
2174 dialog box interfaces, application-defined prompts, the possibility
2175 to use defaults (for example default passwords from somewhere else)
2176 and interrupts/cancellations.
2177 [Richard Levitte]
2178
2179 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2180 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2181 [Steve Henson]
2182
2183 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2184 tidy up some unnecessarily weird code in 'sk_new()').
2185 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2186
2187 *) Change the key loading routines for ENGINEs to use the same kind
2188 callback (pem_password_cb) as all other routines that need this
2189 kind of callback.
2190 [Richard Levitte]
2191
2192 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2193 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2194 than this minimum value is recommended.
2195 [Lutz Jaenicke]
2196
2197 *) New random seeder for OpenVMS, using the system process statistics
2198 that are easily reachable.
2199 [Richard Levitte]
2200
2201 *) Windows apparently can't transparently handle global
2202 variables defined in DLLs. Initialisations such as:
2203
2204 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2205
2206 wont compile. This is used by the any applications that need to
2207 declare their own ASN1 modules. This was fixed by adding the option
2208 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2209 needed for static libraries under Win32.
2210 [Steve Henson]
2211
2212 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2213 setting of purpose and trust fields. New X509_STORE trust and
2214 purpose functions and tidy up setting in other SSL functions.
2215 [Steve Henson]
2216
2217 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2218 structure. These are inherited by X509_STORE_CTX when it is
2219 initialised. This allows various defaults to be set in the
2220 X509_STORE structure (such as flags for CRL checking and custom
2221 purpose or trust settings) for functions which only use X509_STORE_CTX
2222 internally such as S/MIME.
2223
2224 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2225 trust settings if they are not set in X509_STORE. This allows X509_STORE
2226 purposes and trust (in S/MIME for example) to override any set by default.
2227
2228 Add command line options for CRL checking to smime, s_client and s_server
2229 applications.
2230 [Steve Henson]
2231
2232 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2233 are set then the CRL is looked up in the X509_STORE structure and
2234 its validity and signature checked, then if the certificate is found
2235 in the CRL the verify fails with a revoked error.
2236
2237 Various new CRL related callbacks added to X509_STORE_CTX structure.
2238
2239 Command line options added to 'verify' application to support this.
2240
2241 This needs some additional work, such as being able to handle multiple
2242 CRLs with different times, extension based lookup (rather than just
2243 by subject name) and ultimately more complete V2 CRL extension
2244 handling.
2245 [Steve Henson]
2246
2247 *) Add a general user interface API (crypto/ui/). This is designed
2248 to replace things like des_read_password and friends (backward
2249 compatibility functions using this new API are provided).
2250 The purpose is to remove prompting functions from the DES code
2251 section as well as provide for prompting through dialog boxes in
2252 a window system and the like.
2253 [Richard Levitte]
2254
2255 *) Add "ex_data" support to ENGINE so implementations can add state at a
2256 per-structure level rather than having to store it globally.
2257 [Geoff]
2258
2259 *) Make it possible for ENGINE structures to be copied when retrieved by
2260 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2261 This causes the "original" ENGINE structure to act like a template,
2262 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2263 operational state can be localised to each ENGINE structure, despite the
2264 fact they all share the same "methods". New ENGINE structures returned in
2265 this case have no functional references and the return value is the single
2266 structural reference. This matches the single structural reference returned
2267 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2268 ENGINE structure.
2269 [Geoff]
2270
2271 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2272 needs to match any other type at all we need to manually clear the
2273 tag cache.
2274 [Steve Henson]
2275
2276 *) Changes to the "openssl engine" utility to include;
2277 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2278 about an ENGINE's available control commands.
2279 - executing control commands from command line arguments using the
2280 '-pre' and '-post' switches. '-post' is only used if '-t' is
2281 specified and the ENGINE is successfully initialised. The syntax for
2282 the individual commands are colon-separated, for example;
2283 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2284 [Geoff]
2285
2286 *) New dynamic control command support for ENGINEs. ENGINEs can now
2287 declare their own commands (numbers), names (strings), descriptions,
2288 and input types for run-time discovery by calling applications. A
2289 subset of these commands are implicitly classed as "executable"
2290 depending on their input type, and only these can be invoked through
2291 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2292 can be based on user input, config files, etc). The distinction is
2293 that "executable" commands cannot return anything other than a boolean
2294 result and can only support numeric or string input, whereas some
2295 discoverable commands may only be for direct use through
2296 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2297 pointers, or other custom uses. The "executable" commands are to
2298 support parameterisations of ENGINE behaviour that can be
2299 unambiguously defined by ENGINEs and used consistently across any
2300 OpenSSL-based application. Commands have been added to all the
2301 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2302 control over shared-library paths without source code alterations.
2303 [Geoff]
2304
2305 *) Changed all ENGINE implementations to dynamically allocate their
2306 ENGINEs rather than declaring them statically. Apart from this being
2307 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2308 this also allows the implementations to compile without using the
2309 internal engine_int.h header.
2310 [Geoff]
2311
2312 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2313 'const' value. Any code that should be able to modify a RAND_METHOD
2314 should already have non-const pointers to it (ie. they should only
2315 modify their own ones).
2316 [Geoff]
2317
2318 *) Made a variety of little tweaks to the ENGINE code.
2319 - "atalla" and "ubsec" string definitions were moved from header files
2320 to C code. "nuron" string definitions were placed in variables
2321 rather than hard-coded - allowing parameterisation of these values
2322 later on via ctrl() commands.
2323 - Removed unused "#if 0"'d code.
2324 - Fixed engine list iteration code so it uses ENGINE_free() to release
2325 structural references.
2326 - Constified the RAND_METHOD element of ENGINE structures.
2327 - Constified various get/set functions as appropriate and added
2328 missing functions (including a catch-all ENGINE_cpy that duplicates
2329 all ENGINE values onto a new ENGINE except reference counts/state).
2330 - Removed NULL parameter checks in get/set functions. Setting a method
2331 or function to NULL is a way of cancelling out a previously set
2332 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2333 and doesn't justify the extra error symbols and code.
2334 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2335 flags from engine_int.h to engine.h.
2336 - Changed prototypes for ENGINE handler functions (init(), finish(),
2337 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2338 [Geoff]
2339
2340 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2341 to the algorithm using long division. The binary algorithm can be
2342 used only if the modulus is odd. On 32-bit systems, it is faster
2343 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2344 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2345 up to 450 bits. In 64-bit environments, the binary algorithm
2346 appears to be advantageous for much longer moduli; here we use it
2347 for moduli up to 2048 bits.
2348 [Bodo Moeller]
2349
2350 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2351 could not support the combine flag in choice fields.
2352 [Steve Henson]
2353
2354 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2355 extensions from a certificate request to the certificate.
2356 [Steve Henson]
2357
2358 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2359 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2360 file: this allows the display of the certificate about to be
2361 signed to be customised, to allow certain fields to be included
2362 or excluded and extension details. The old system didn't display
2363 multicharacter strings properly, omitted fields not in the policy
2364 and couldn't display additional details such as extensions.
2365 [Steve Henson]
2366
2367 *) Function EC_POINTs_mul for multiple scalar multiplication
2368 of an arbitrary number of elliptic curve points
2369 \sum scalars[i]*points[i],
2370 optionally including the generator defined for the EC_GROUP:
2371 scalar*generator + \sum scalars[i]*points[i].
2372
2373 EC_POINT_mul is a simple wrapper function for the typical case
2374 that the point list has just one item (besides the optional
2375 generator).
2376 [Bodo Moeller]
2377
2378 *) First EC_METHODs for curves over GF(p):
2379
2380 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2381 operations and provides various method functions that can also
2382 operate with faster implementations of modular arithmetic.
2383
2384 EC_GFp_mont_method() reuses most functions that are part of
2385 EC_GFp_simple_method, but uses Montgomery arithmetic.
2386
2387 [Bodo Moeller; point addition and point doubling
2388 implementation directly derived from source code provided by
2389 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2390
2391 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2392 crypto/ec/ec_lib.c):
2393
2394 Curves are EC_GROUP objects (with an optional group generator)
2395 based on EC_METHODs that are built into the library.
2396
2397 Points are EC_POINT objects based on EC_GROUP objects.
2398
2399 Most of the framework would be able to handle curves over arbitrary
2400 finite fields, but as there are no obvious types for fields other
2401 than GF(p), some functions are limited to that for now.
2402 [Bodo Moeller]
2403
2404 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2405 that the file contains a complete HTTP response.
2406 [Richard Levitte]
2407
2408 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2409 change the def and num file printf format specifier from "%-40sXXX"
2410 to "%-39s XXX". The latter will always guarantee a space after the
2411 field while the former will cause them to run together if the field
2412 is 40 of more characters long.
2413 [Steve Henson]
2414
2415 *) Constify the cipher and digest 'method' functions and structures
2416 and modify related functions to take constant EVP_MD and EVP_CIPHER
2417 pointers.
2418 [Steve Henson]
2419
2420 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2421 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2422 [Bodo Moeller]
2423
2424 *) Modify EVP_Digest*() routines so they now return values. Although the
2425 internal software routines can never fail additional hardware versions
2426 might.
2427 [Steve Henson]
2428
2429 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2430
2431 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2432 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2433
2434 ASN1 error codes
2435 ERR_R_NESTED_ASN1_ERROR
2436 ...
2437 ERR_R_MISSING_ASN1_EOS
2438 were 4 .. 9, conflicting with
2439 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2440 ...
2441 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2442 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2443
2444 Add new error code 'ERR_R_INTERNAL_ERROR'.
2445 [Bodo Moeller]
2446
2447 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2448 suffices.
2449 [Bodo Moeller]
2450
2451 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2452 sets the subject name for a new request or supersedes the
2453 subject name in a given request. Formats that can be parsed are
2454 'CN=Some Name, OU=myOU, C=IT'
2455 and
2456 'CN=Some Name/OU=myOU/C=IT'.
2457
2458 Add options '-batch' and '-verbose' to 'openssl req'.
2459 [Massimiliano Pala <madwolf@hackmasters.net>]
2460
2461 *) Introduce the possibility to access global variables through
2462 functions on platform were that's the best way to handle exporting
2463 global variables in shared libraries. To enable this functionality,
2464 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2465 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2466 is normally done by Configure or something similar).
2467
2468 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2469 in the source file (foo.c) like this:
2470
2471 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2472 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2473
2474 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2475 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2476
2477 OPENSSL_DECLARE_GLOBAL(int,foo);
2478 #define foo OPENSSL_GLOBAL_REF(foo)
2479 OPENSSL_DECLARE_GLOBAL(double,bar);
2480 #define bar OPENSSL_GLOBAL_REF(bar)
2481
2482 The #defines are very important, and therefore so is including the
2483 header file everywhere where the defined globals are used.
2484
2485 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2486 of ASN.1 items, but that structure is a bit different.
2487
2488 The largest change is in util/mkdef.pl which has been enhanced with
2489 better and easier to understand logic to choose which symbols should
2490 go into the Windows .def files as well as a number of fixes and code
2491 cleanup (among others, algorithm keywords are now sorted
2492 lexicographically to avoid constant rewrites).
2493 [Richard Levitte]
2494
2495 *) In BN_div() keep a copy of the sign of 'num' before writing the
2496 result to 'rm' because if rm==num the value will be overwritten
2497 and produce the wrong result if 'num' is negative: this caused
2498 problems with BN_mod() and BN_nnmod().
2499 [Steve Henson]
2500
2501 *) Function OCSP_request_verify(). This checks the signature on an
2502 OCSP request and verifies the signer certificate. The signer
2503 certificate is just checked for a generic purpose and OCSP request
2504 trust settings.
2505 [Steve Henson]
2506
2507 *) Add OCSP_check_validity() function to check the validity of OCSP
2508 responses. OCSP responses are prepared in real time and may only
2509 be a few seconds old. Simply checking that the current time lies
2510 between thisUpdate and nextUpdate max reject otherwise valid responses
2511 caused by either OCSP responder or client clock inaccuracy. Instead
2512 we allow thisUpdate and nextUpdate to fall within a certain period of
2513 the current time. The age of the response can also optionally be
2514 checked. Two new options -validity_period and -status_age added to
2515 ocsp utility.
2516 [Steve Henson]
2517
2518 *) If signature or public key algorithm is unrecognized print out its
2519 OID rather that just UNKNOWN.
2520 [Steve Henson]
2521
2522 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2523 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2524 ID to be generated from the issuer certificate alone which can then be
2525 passed to OCSP_id_issuer_cmp().
2526 [Steve Henson]
2527
2528 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2529 ASN1 modules to export functions returning ASN1_ITEM pointers
2530 instead of the ASN1_ITEM structures themselves. This adds several
2531 new macros which allow the underlying ASN1 function/structure to
2532 be accessed transparently. As a result code should not use ASN1_ITEM
2533 references directly (such as &X509_it) but instead use the relevant
2534 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2535 use of the new ASN1 code on platforms where exporting structures
2536 is problematical (for example in shared libraries) but exporting
2537 functions returning pointers to structures is not.
2538 [Steve Henson]
2539
2540 *) Add support for overriding the generation of SSL/TLS session IDs.
2541 These callbacks can be registered either in an SSL_CTX or per SSL.
2542 The purpose of this is to allow applications to control, if they wish,
2543 the arbitrary values chosen for use as session IDs, particularly as it
2544 can be useful for session caching in multiple-server environments. A
2545 command-line switch for testing this (and any client code that wishes
2546 to use such a feature) has been added to "s_server".
2547 [Geoff Thorpe, Lutz Jaenicke]
2548
2549 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2550 of the form '#if defined(...) || defined(...) || ...' and
2551 '#if !defined(...) && !defined(...) && ...'. This also avoids
2552 the growing number of special cases it was previously handling.
2553 [Richard Levitte]
2554
2555 *) Make all configuration macros available for application by making
2556 sure they are available in opensslconf.h, by giving them names starting
2557 with "OPENSSL_" to avoid conflicts with other packages and by making
2558 sure e_os2.h will cover all platform-specific cases together with
2559 opensslconf.h.
2560 Additionally, it is now possible to define configuration/platform-
2561 specific names (called "system identities"). In the C code, these
2562 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2563 macro with the name beginning with "OPENSSL_SYS_", which is determined
2564 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2565 what is available.
2566 [Richard Levitte]
2567
2568 *) New option -set_serial to 'req' and 'x509' this allows the serial
2569 number to use to be specified on the command line. Previously self
2570 signed certificates were hard coded with serial number 0 and the
2571 CA options of 'x509' had to use a serial number in a file which was
2572 auto incremented.
2573 [Steve Henson]
2574
2575 *) New options to 'ca' utility to support V2 CRL entry extensions.
2576 Currently CRL reason, invalidity date and hold instruction are
2577 supported. Add new CRL extensions to V3 code and some new objects.
2578 [Steve Henson]
2579
2580 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2581 disable standard block padding (aka PKCS#5 padding) in the EVP
2582 API, which was previously mandatory. This means that the data is
2583 not padded in any way and so the total length much be a multiple
2584 of the block size, otherwise an error occurs.
2585 [Steve Henson]
2586
2587 *) Initial (incomplete) OCSP SSL support.
2588 [Steve Henson]
2589
2590 *) New function OCSP_parse_url(). This splits up a URL into its host,
2591 port and path components: primarily to parse OCSP URLs. New -url
2592 option to ocsp utility.
2593 [Steve Henson]
2594
2595 *) New nonce behavior. The return value of OCSP_check_nonce() now
2596 reflects the various checks performed. Applications can decide
2597 whether to tolerate certain situations such as an absent nonce
2598 in a response when one was present in a request: the ocsp application
2599 just prints out a warning. New function OCSP_add1_basic_nonce()
2600 this is to allow responders to include a nonce in a response even if
2601 the request is nonce-less.
2602 [Steve Henson]
2603
2604 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2605 skipped when using openssl x509 multiple times on a single input file,
2606 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2607 [Bodo Moeller]
2608
2609 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2610 set string type: to handle setting ASN1_TIME structures. Fix ca
2611 utility to correctly initialize revocation date of CRLs.
2612 [Steve Henson]
2613
2614 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2615 the clients preferred ciphersuites and rather use its own preferences.
2616 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2617 Internet Explorer by ensuring unchanged hash method during stepup.
2618 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2619 [Lutz Jaenicke]
2620
2621 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2622 to aes and add a new 'exist' option to print out symbols that don't
2623 appear to exist.
2624 [Steve Henson]
2625
2626 *) Additional options to ocsp utility to allow flags to be set and
2627 additional certificates supplied.
2628 [Steve Henson]
2629
2630 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2631 OCSP client a number of certificate to only verify the response
2632 signature against.
2633 [Richard Levitte]
2634
2635 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2636 handle the new API. Currently only ECB, CBC modes supported. Add new
2637 AES OIDs.
2638
2639 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2640 Encryption Standard (AES) Ciphersuites for Transport Layer
2641 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2642 not enabled by default and were not part of the "ALL" ciphersuite
2643 alias because they were not yet official; they could be
2644 explicitly requested by specifying the "AESdraft" ciphersuite
2645 group alias. In the final release of OpenSSL 0.9.7, the group
2646 alias is called "AES" and is part of "ALL".)
2647 [Ben Laurie, Steve Henson, Bodo Moeller]
2648
2649 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2650 request to response.
2651 [Steve Henson]
2652
2653 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2654 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2655 extract information from a certificate request. OCSP_response_create()
2656 creates a response and optionally adds a basic response structure.
2657 OCSP_basic_add1_status() adds a complete single response to a basic
2658 response and returns the OCSP_SINGLERESP structure just added (to allow
2659 extensions to be included for example). OCSP_basic_add1_cert() adds a
2660 certificate to a basic response and OCSP_basic_sign() signs a basic
2661 response with various flags. New helper functions ASN1_TIME_check()
2662 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2663 (converts ASN1_TIME to GeneralizedTime).
2664 [Steve Henson]
2665
2666 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2667 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2668 structure from a certificate. X509_pubkey_digest() digests the public_key
2669 contents: this is used in various key identifiers.
2670 [Steve Henson]
2671
2672 *) Make sk_sort() tolerate a NULL argument.
2673 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2674
2675 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2676 passed by the function are trusted implicitly. If any of them signed the
2677 response then it is assumed to be valid and is not verified.
2678 [Steve Henson]
2679
2680 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2681 to data. This was previously part of the PKCS7 ASN1 code. This
2682 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2683 [Steve Henson, reported by Kenneth R. Robinette
2684 <support@securenetterm.com>]
2685
2686 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2687 routines: without these tracing memory leaks is very painful.
2688 Fix leaks in PKCS12 and PKCS7 routines.
2689 [Steve Henson]
2690
2691 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2692 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2693 effectively meant GeneralizedTime would never be used. Now it
2694 is initialised to -1 but X509_time_adj() now has to check the value
2695 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2696 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2697 [Steve Henson, reported by Kenneth R. Robinette
2698 <support@securenetterm.com>]
2699
2700 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2701 result in a zero length in the ASN1_INTEGER structure which was
2702 not consistent with the structure when d2i_ASN1_INTEGER() was used
2703 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2704 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2705 where it did not print out a minus for negative ASN1_INTEGER.
2706 [Steve Henson]
2707
2708 *) Add summary printout to ocsp utility. The various functions which
2709 convert status values to strings have been renamed to:
2710 OCSP_response_status_str(), OCSP_cert_status_str() and
2711 OCSP_crl_reason_str() and are no longer static. New options
2712 to verify nonce values and to disable verification. OCSP response
2713 printout format cleaned up.
2714 [Steve Henson]
2715
2716 *) Add additional OCSP certificate checks. These are those specified
2717 in RFC2560. This consists of two separate checks: the CA of the
2718 certificate being checked must either be the OCSP signer certificate
2719 or the issuer of the OCSP signer certificate. In the latter case the
2720 OCSP signer certificate must contain the OCSP signing extended key
2721 usage. This check is performed by attempting to match the OCSP
2722 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2723 in the OCSP_CERTID structures of the response.
2724 [Steve Henson]
2725
2726 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2727 and related routines. This uses the standard OpenSSL certificate
2728 verify routines to perform initial checks (just CA validity) and
2729 to obtain the certificate chain. Then additional checks will be
2730 performed on the chain. Currently the root CA is checked to see
2731 if it is explicitly trusted for OCSP signing. This is used to set
2732 a root CA as a global signing root: that is any certificate that
2733 chains to that CA is an acceptable OCSP signing certificate.
2734 [Steve Henson]
2735
2736 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2737 extensions from a separate configuration file.
2738 As when reading extensions from the main configuration file,
2739 the '-extensions ...' option may be used for specifying the
2740 section to use.
2741 [Massimiliano Pala <madwolf@comune.modena.it>]
2742
2743 *) New OCSP utility. Allows OCSP requests to be generated or
2744 read. The request can be sent to a responder and the output
2745 parsed, outputed or printed in text form. Not complete yet:
2746 still needs to check the OCSP response validity.
2747 [Steve Henson]
2748
2749 *) New subcommands for 'openssl ca':
2750 'openssl ca -status <serial>' prints the status of the cert with
2751 the given serial number (according to the index file).
2752 'openssl ca -updatedb' updates the expiry status of certificates
2753 in the index file.
2754 [Massimiliano Pala <madwolf@comune.modena.it>]
2755
2756 *) New '-newreq-nodes' command option to CA.pl. This is like
2757 '-newreq', but calls 'openssl req' with the '-nodes' option
2758 so that the resulting key is not encrypted.
2759 [Damien Miller <djm@mindrot.org>]
2760
2761 *) New configuration for the GNU Hurd.
2762 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2763
2764 *) Initial code to implement OCSP basic response verify. This
2765 is currently incomplete. Currently just finds the signer's
2766 certificate and verifies the signature on the response.
2767 [Steve Henson]
2768
2769 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2770 value of OPENSSLDIR. This is available via the new '-d' option
2771 to 'openssl version', and is also included in 'openssl version -a'.
2772 [Bodo Moeller]
2773
2774 *) Allowing defining memory allocation callbacks that will be given
2775 file name and line number information in additional arguments
2776 (a const char* and an int). The basic functionality remains, as
2777 well as the original possibility to just replace malloc(),
2778 realloc() and free() by functions that do not know about these
2779 additional arguments. To register and find out the current
2780 settings for extended allocation functions, the following
2781 functions are provided:
2782
2783 CRYPTO_set_mem_ex_functions
2784 CRYPTO_set_locked_mem_ex_functions
2785 CRYPTO_get_mem_ex_functions
2786 CRYPTO_get_locked_mem_ex_functions
2787
2788 These work the same way as CRYPTO_set_mem_functions and friends.
2789 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2790 extended allocation function is enabled.
2791 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2792 a conventional allocation function is enabled.
2793 [Richard Levitte, Bodo Moeller]
2794
2795 *) Finish off removing the remaining LHASH function pointer casts.
2796 There should no longer be any prototype-casting required when using
2797 the LHASH abstraction, and any casts that remain are "bugs". See
2798 the callback types and macros at the head of lhash.h for details
2799 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2800 [Geoff Thorpe]
2801
2802 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2803 If /dev/[u]random devices are not available or do not return enough
2804 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2805 be queried.
2806 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2807 /etc/entropy will be queried once each in this sequence, quering stops
2808 when enough entropy was collected without querying more sockets.
2809 [Lutz Jaenicke]
2810
2811 *) Change the Unix RAND_poll() variant to be able to poll several
2812 random devices, as specified by DEVRANDOM, until a sufficient amount
2813 of data has been collected. We spend at most 10 ms on each file
2814 (select timeout) and read in non-blocking mode. DEVRANDOM now
2815 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2816 (previously it was just the string "/dev/urandom"), so on typical
2817 platforms the 10 ms delay will never occur.
2818 Also separate out the Unix variant to its own file, rand_unix.c.
2819 For VMS, there's a currently-empty rand_vms.c.
2820 [Richard Levitte]
2821
2822 *) Move OCSP client related routines to ocsp_cl.c. These
2823 provide utility functions which an application needing
2824 to issue a request to an OCSP responder and analyse the
2825 response will typically need: as opposed to those which an
2826 OCSP responder itself would need which will be added later.
2827
2828 OCSP_request_sign() signs an OCSP request with an API similar
2829 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2830 response. OCSP_response_get1_basic() extracts basic response
2831 from response. OCSP_resp_find_status(): finds and extracts status
2832 information from an OCSP_CERTID structure (which will be created
2833 when the request structure is built). These are built from lower
2834 level functions which work on OCSP_SINGLERESP structures but
2835 wont normally be used unless the application wishes to examine
2836 extensions in the OCSP response for example.
2837
2838 Replace nonce routines with a pair of functions.
2839 OCSP_request_add1_nonce() adds a nonce value and optionally
2840 generates a random value. OCSP_check_nonce() checks the
2841 validity of the nonce in an OCSP response.
2842 [Steve Henson]
2843
2844 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2845 This doesn't copy the supplied OCSP_CERTID and avoids the
2846 need to free up the newly created id. Change return type
2847 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2848 This can then be used to add extensions to the request.
2849 Deleted OCSP_request_new(), since most of its functionality
2850 is now in OCSP_REQUEST_new() (and the case insensitive name
2851 clash) apart from the ability to set the request name which
2852 will be added elsewhere.
2853 [Steve Henson]
2854
2855 *) Update OCSP API. Remove obsolete extensions argument from
2856 various functions. Extensions are now handled using the new
2857 OCSP extension code. New simple OCSP HTTP function which
2858 can be used to send requests and parse the response.
2859 [Steve Henson]
2860
2861 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2862 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2863 uses the special reorder version of SET OF to sort the attributes
2864 and reorder them to match the encoded order. This resolves a long
2865 standing problem: a verify on a PKCS7 structure just after signing
2866 it used to fail because the attribute order did not match the
2867 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2868 it uses the received order. This is necessary to tolerate some broken
2869 software that does not order SET OF. This is handled by encoding
2870 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2871 to produce the required SET OF.
2872 [Steve Henson]
2873
2874 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2875 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2876 files to get correct declarations of the ASN.1 item variables.
2877 [Richard Levitte]
2878
2879 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2880 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2881 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2882 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2883 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2884 ASN1_ITEM and no wrapper functions.
2885 [Steve Henson]
2886
2887 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2888 replace the old function pointer based I/O routines. Change most of
2889 the *_d2i_bio() and *_d2i_fp() functions to use these.
2890 [Steve Henson]
2891
2892 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2893 lines, recognice more "algorithms" that can be deselected, and make
2894 it complain about algorithm deselection that isn't recognised.
2895 [Richard Levitte]
2896
2897 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2898 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2899 to use new functions. Add NO_ASN1_OLD which can be set to remove
2900 some old style ASN1 functions: this can be used to determine if old
2901 code will still work when these eventually go away.
2902 [Steve Henson]
2903
2904 *) New extension functions for OCSP structures, these follow the
2905 same conventions as certificates and CRLs.
2906 [Steve Henson]
2907
2908 *) New function X509V3_add1_i2d(). This automatically encodes and
2909 adds an extension. Its behaviour can be customised with various
2910 flags to append, replace or delete. Various wrappers added for
2911 certifcates and CRLs.
2912 [Steve Henson]
2913
2914 *) Fix to avoid calling the underlying ASN1 print routine when
2915 an extension cannot be parsed. Correct a typo in the
2916 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2917 [Steve Henson]
2918
2919 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2920 entries for variables.
2921 [Steve Henson]
2922
2923 *) Add functionality to apps/openssl.c for detecting locking
2924 problems: As the program is single-threaded, all we have
2925 to do is register a locking callback using an array for
2926 storing which locks are currently held by the program.
2927 [Bodo Moeller]
2928
2929 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2930 SSL_get_ex_data_X509_STORE_idx(), which is used in
2931 ssl_verify_cert_chain() and thus can be called at any time
2932 during TLS/SSL handshakes so that thread-safety is essential.
2933 Unfortunately, the ex_data design is not at all suited
2934 for multi-threaded use, so it probably should be abolished.
2935 [Bodo Moeller]
2936
2937 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2938 [Broadcom, tweaked and integrated by Geoff Thorpe]
2939
2940 *) Move common extension printing code to new function
2941 X509V3_print_extensions(). Reorganise OCSP print routines and
2942 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2943 [Steve Henson]
2944
2945 *) New function X509_signature_print() to remove duplication in some
2946 print routines.
2947 [Steve Henson]
2948
2949 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2950 set (this was treated exactly the same as SET OF previously). This
2951 is used to reorder the STACK representing the structure to match the
2952 encoding. This will be used to get round a problem where a PKCS7
2953 structure which was signed could not be verified because the STACK
2954 order did not reflect the encoded order.
2955 [Steve Henson]
2956
2957 *) Reimplement the OCSP ASN1 module using the new code.
2958 [Steve Henson]
2959
2960 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
2961 for its ASN1 operations. The old style function pointers still exist
2962 for now but they will eventually go away.
2963 [Steve Henson]
2964
2965 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
2966 completely replaces the old ASN1 functionality with a table driven
2967 encoder and decoder which interprets an ASN1_ITEM structure describing
2968 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
2969 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
2970 has also been converted to the new form.
2971 [Steve Henson]
2972
2973 *) Change BN_mod_exp_recp so that negative moduli are tolerated
2974 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
2975 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
2976 for negative moduli.
2977 [Bodo Moeller]
2978
2979 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
2980 of not touching the result's sign bit.
2981 [Bodo Moeller]
2982
2983 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2984 set.
2985 [Bodo Moeller]
2986
2987 *) Changed the LHASH code to use prototypes for callbacks, and created
2988 macros to declare and implement thin (optionally static) functions
2989 that provide type-safety and avoid function pointer casting for the
2990 type-specific callbacks.
2991 [Geoff Thorpe]
2992
2993 *) Added Kerberos Cipher Suites to be used with TLS, as written in
2994 RFC 2712.
2995 [Veers Staats <staatsvr@asc.hpc.mil>,
2996 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2997
2998 *) Reformat the FAQ so the different questions and answers can be divided
2999 in sections depending on the subject.
3000 [Richard Levitte]
3001
3002 *) Have the zlib compression code load ZLIB.DLL dynamically under
3003 Windows.
3004 [Richard Levitte]
3005
3006 *) New function BN_mod_sqrt for computing square roots modulo a prime
3007 (using the probabilistic Tonelli-Shanks algorithm unless
3008 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3009 be handled deterministically).
3010 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3011
3012 *) Make BN_mod_inverse faster by explicitly handling small quotients
3013 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3014 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3015 [Bodo Moeller]
3016
3017 *) New function BN_kronecker.
3018 [Bodo Moeller]
3019
3020 *) Fix BN_gcd so that it works on negative inputs; the result is
3021 positive unless both parameters are zero.
3022 Previously something reasonably close to an infinite loop was
3023 possible because numbers could be growing instead of shrinking
3024 in the implementation of Euclid's algorithm.
3025 [Bodo Moeller]
3026
3027 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3028 sign of the number in question.
3029
3030 Fix BN_is_word(a,w) to work correctly for w == 0.
3031
3032 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3033 because its test if the absolute value of 'a' equals 'w'.
3034 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3035 it exists mostly for use in the implementations of BN_is_zero(),
3036 BN_is_one(), and BN_is_word().
3037 [Bodo Moeller]
3038
3039 *) New function BN_swap.
3040 [Bodo Moeller]
3041
3042 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3043 the exponentiation functions are more likely to produce reasonable
3044 results on negative inputs.
3045 [Bodo Moeller]
3046
3047 *) Change BN_mod_mul so that the result is always non-negative.
3048 Previously, it could be negative if one of the factors was negative;
3049 I don't think anyone really wanted that behaviour.
3050 [Bodo Moeller]
3051
3052 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3053 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3054 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3055 and add new functions:
3056
3057 BN_nnmod
3058 BN_mod_sqr
3059 BN_mod_add
3060 BN_mod_add_quick
3061 BN_mod_sub
3062 BN_mod_sub_quick
3063 BN_mod_lshift1
3064 BN_mod_lshift1_quick
3065 BN_mod_lshift
3066 BN_mod_lshift_quick
3067
3068 These functions always generate non-negative results.
3069
3070 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3071 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3072
3073 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3074 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3075 be reduced modulo m.
3076 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3077
3078 #if 0
3079 The following entry accidentily appeared in the CHANGES file
3080 distributed with OpenSSL 0.9.7. The modifications described in
3081 it do *not* apply to OpenSSL 0.9.7.
3082
3083 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3084 was actually never needed) and in BN_mul(). The removal in BN_mul()
3085 required a small change in bn_mul_part_recursive() and the addition
3086 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3087 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3088 bn_sub_words() and bn_add_words() except they take arrays with
3089 differing sizes.
3090 [Richard Levitte]
3091 #endif
3092
3093 *) In 'openssl passwd', verify passwords read from the terminal
3094 unless the '-salt' option is used (which usually means that
3095 verification would just waste user's time since the resulting
3096 hash is going to be compared with some given password hash)
3097 or the new '-noverify' option is used.
3098
3099 This is an incompatible change, but it does not affect
3100 non-interactive use of 'openssl passwd' (passwords on the command
3101 line, '-stdin' option, '-in ...' option) and thus should not
3102 cause any problems.
3103 [Bodo Moeller]
3104
3105 *) Remove all references to RSAref, since there's no more need for it.
3106 [Richard Levitte]
3107
3108 *) Make DSO load along a path given through an environment variable
3109 (SHLIB_PATH) with shl_load().
3110 [Richard Levitte]
3111
3112 *) Constify the ENGINE code as a result of BIGNUM constification.
3113 Also constify the RSA code and most things related to it. In a
3114 few places, most notable in the depth of the ASN.1 code, ugly
3115 casts back to non-const were required (to be solved at a later
3116 time)
3117 [Richard Levitte]
3118
3119 *) Make it so the openssl application has all engines loaded by default.
3120 [Richard Levitte]
3121
3122 *) Constify the BIGNUM routines a little more.
3123 [Richard Levitte]
3124
3125 *) Add the following functions:
3126
3127 ENGINE_load_cswift()
3128 ENGINE_load_chil()
3129 ENGINE_load_atalla()
3130 ENGINE_load_nuron()
3131 ENGINE_load_builtin_engines()
3132
3133 That way, an application can itself choose if external engines that
3134 are built-in in OpenSSL shall ever be used or not. The benefit is
3135 that applications won't have to be linked with libdl or other dso
3136 libraries unless it's really needed.
3137
3138 Changed 'openssl engine' to load all engines on demand.
3139 Changed the engine header files to avoid the duplication of some
3140 declarations (they differed!).
3141 [Richard Levitte]
3142
3143 *) 'openssl engine' can now list capabilities.
3144 [Richard Levitte]
3145
3146 *) Better error reporting in 'openssl engine'.
3147 [Richard Levitte]
3148
3149 *) Never call load_dh_param(NULL) in s_server.
3150 [Bodo Moeller]
3151
3152 *) Add engine application. It can currently list engines by name and
3153 identity, and test if they are actually available.
3154 [Richard Levitte]
3155
3156 *) Improve RPM specification file by forcing symbolic linking and making
3157 sure the installed documentation is also owned by root.root.
3158 [Damien Miller <djm@mindrot.org>]
3159
3160 *) Give the OpenSSL applications more possibilities to make use of
3161 keys (public as well as private) handled by engines.
3162 [Richard Levitte]
3163
3164 *) Add OCSP code that comes from CertCo.
3165 [Richard Levitte]
3166
3167 *) Add VMS support for the Rijndael code.
3168 [Richard Levitte]
3169
3170 *) Added untested support for Nuron crypto accelerator.
3171 [Ben Laurie]
3172
3173 *) Add support for external cryptographic devices. This code was
3174 previously distributed separately as the "engine" branch.
3175 [Geoff Thorpe, Richard Levitte]
3176
3177 *) Rework the filename-translation in the DSO code. It is now possible to
3178 have far greater control over how a "name" is turned into a filename
3179 depending on the operating environment and any oddities about the
3180 different shared library filenames on each system.
3181 [Geoff Thorpe]
3182
3183 *) Support threads on FreeBSD-elf in Configure.
3184 [Richard Levitte]
3185
3186 *) Fix for SHA1 assembly problem with MASM: it produces
3187 warnings about corrupt line number information when assembling
3188 with debugging information. This is caused by the overlapping
3189 of two sections.
3190 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3191
3192 *) NCONF changes.
3193 NCONF_get_number() has no error checking at all. As a replacement,
3194 NCONF_get_number_e() is defined (_e for "error checking") and is
3195 promoted strongly. The old NCONF_get_number is kept around for
3196 binary backward compatibility.
3197 Make it possible for methods to load from something other than a BIO,
3198 by providing a function pointer that is given a name instead of a BIO.
3199 For example, this could be used to load configuration data from an
3200 LDAP server.
3201 [Richard Levitte]
3202
3203 *) Fix for non blocking accept BIOs. Added new I/O special reason
3204 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3205 with non blocking I/O was not possible because no retry code was
3206 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3207 this case.
3208 [Steve Henson]
3209
3210 *) Added the beginnings of Rijndael support.
3211 [Ben Laurie]
3212
3213 *) Fix for bug in DirectoryString mask setting. Add support for
3214 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3215 to allow certificate printing to more controllable, additional
3216 'certopt' option to 'x509' to allow new printing options to be
3217 set.
3218 [Steve Henson]
3219
3220 *) Clean old EAY MD5 hack from e_os.h.
3221 [Richard Levitte]
3222
3223 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3224
3225 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3226 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3227 [Joe Orton, Steve Henson]
3228
3229 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3230
3231 *) Fix additional bug revealed by the NISCC test suite:
3232
3233 Stop bug triggering large recursion when presented with
3234 certain ASN.1 tags (CVE-2003-0851)
3235 [Steve Henson]
3236
3237 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3238
3239 *) Fix various bugs revealed by running the NISCC test suite:
3240
3241 Stop out of bounds reads in the ASN1 code when presented with
3242 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3243
3244 If verify callback ignores invalid public key errors don't try to check
3245 certificate signature with the NULL public key.
3246
3247 [Steve Henson]
3248
3249 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3250 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3251 specifications.
3252 [Steve Henson]
3253
3254 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3255 extra data after the compression methods not only for TLS 1.0
3256 but also for SSL 3.0 (as required by the specification).
3257 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3258
3259 *) Change X509_certificate_type() to mark the key as exported/exportable
3260 when it's 512 *bits* long, not 512 bytes.
3261 [Richard Levitte]
3262
3263 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3264
3265 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3266 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3267 a protocol version number mismatch like a decryption error
3268 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3269 [Bodo Moeller]
3270
3271 *) Turn on RSA blinding by default in the default implementation
3272 to avoid a timing attack. Applications that don't want it can call
3273 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3274 They would be ill-advised to do so in most cases.
3275 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3276
3277 *) Change RSA blinding code so that it works when the PRNG is not
3278 seeded (in this case, the secret RSA exponent is abused as
3279 an unpredictable seed -- if it is not unpredictable, there
3280 is no point in blinding anyway). Make RSA blinding thread-safe
3281 by remembering the creator's thread ID in rsa->blinding and
3282 having all other threads use local one-time blinding factors
3283 (this requires more computation than sharing rsa->blinding, but
3284 avoids excessive locking; and if an RSA object is not shared
3285 between threads, blinding will still be very fast).
3286 [Bodo Moeller]
3287
3288 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3289
3290 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3291 via timing by performing a MAC computation even if incorrrect
3292 block cipher padding has been found. This is a countermeasure
3293 against active attacks where the attacker has to distinguish
3294 between bad padding and a MAC verification error. (CVE-2003-0078)
3295
3296 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3297 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3298 Martin Vuagnoux (EPFL, Ilion)]
3299
3300 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3301
3302 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3303 memory from it's contents. This is done with a counter that will
3304 place alternating values in each byte. This can be used to solve
3305 two issues: 1) the removal of calls to memset() by highly optimizing
3306 compilers, and 2) cleansing with other values than 0, since those can
3307 be read through on certain media, for example a swap space on disk.
3308 [Geoff Thorpe]
3309
3310 *) Bugfix: client side session caching did not work with external caching,
3311 because the session->cipher setting was not restored when reloading
3312 from the external cache. This problem was masked, when
3313 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3314 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3315 [Lutz Jaenicke]
3316
3317 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3318 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3319 [Zeev Lieber <zeev-l@yahoo.com>]
3320
3321 *) Undo an undocumented change introduced in 0.9.6e which caused
3322 repeated calls to OpenSSL_add_all_ciphers() and
3323 OpenSSL_add_all_digests() to be ignored, even after calling
3324 EVP_cleanup().
3325 [Richard Levitte]
3326
3327 *) Change the default configuration reader to deal with last line not
3328 being properly terminated.
3329 [Richard Levitte]
3330
3331 *) Change X509_NAME_cmp() so it applies the special rules on handling
3332 DN values that are of type PrintableString, as well as RDNs of type
3333 emailAddress where the value has the type ia5String.
3334 [stefank@valicert.com via Richard Levitte]
3335
3336 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3337 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3338 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3339 the bitwise-OR of the two for use by the majority of applications
3340 wanting this behaviour, and update the docs. The documented
3341 behaviour and actual behaviour were inconsistent and had been
3342 changing anyway, so this is more a bug-fix than a behavioural
3343 change.
3344 [Geoff Thorpe, diagnosed by Nadav Har'El]
3345
3346 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3347 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3348 [Bodo Moeller]
3349
3350 *) Fix initialization code race conditions in
3351 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3352 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3353 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3354 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3355 ssl2_get_cipher_by_char(),
3356 ssl3_get_cipher_by_char().
3357 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3358
3359 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3360 the cached sessions are flushed, as the remove_cb() might use ex_data
3361 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3362 (see [openssl.org #212]).
3363 [Geoff Thorpe, Lutz Jaenicke]
3364
3365 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3366 length, instead of the encoding length to d2i_ASN1_OBJECT.
3367 [Steve Henson]
3368
3369 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3370
3371 *) [In 0.9.6g-engine release:]
3372 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3373 [Lynn Gazis <lgazis@rainbow.com>]
3374
3375 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3376
3377 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3378 and get fix the header length calculation.
3379 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3380 Alon Kantor <alonk@checkpoint.com> (and others),
3381 Steve Henson]
3382
3383 *) Use proper error handling instead of 'assertions' in buffer
3384 overflow checks added in 0.9.6e. This prevents DoS (the
3385 assertions could call abort()).
3386 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3387
3388 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3389
3390 *) Add various sanity checks to asn1_get_length() to reject
3391 the ASN1 length bytes if they exceed sizeof(long), will appear
3392 negative or the content length exceeds the length of the
3393 supplied buffer.
3394 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3395
3396 *) Fix cipher selection routines: ciphers without encryption had no flags
3397 for the cipher strength set and where therefore not handled correctly
3398 by the selection routines (PR #130).
3399 [Lutz Jaenicke]
3400
3401 *) Fix EVP_dsa_sha macro.
3402 [Nils Larsch]
3403
3404 *) New option
3405 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3406 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3407 that was added in OpenSSL 0.9.6d.
3408
3409 As the countermeasure turned out to be incompatible with some
3410 broken SSL implementations, the new option is part of SSL_OP_ALL.
3411 SSL_OP_ALL is usually employed when compatibility with weird SSL
3412 implementations is desired (e.g. '-bugs' option to 's_client' and
3413 's_server'), so the new option is automatically set in many
3414 applications.
3415 [Bodo Moeller]
3416
3417 *) Changes in security patch:
3418
3419 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3420 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3421 Air Force Materiel Command, USAF, under agreement number
3422 F30602-01-2-0537.
3423
3424 *) Add various sanity checks to asn1_get_length() to reject
3425 the ASN1 length bytes if they exceed sizeof(long), will appear
3426 negative or the content length exceeds the length of the
3427 supplied buffer. (CVE-2002-0659)
3428 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3429
3430 *) Assertions for various potential buffer overflows, not known to
3431 happen in practice.
3432 [Ben Laurie (CHATS)]
3433
3434 *) Various temporary buffers to hold ASCII versions of integers were
3435 too small for 64 bit platforms. (CVE-2002-0655)
3436 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3437
3438 *) Remote buffer overflow in SSL3 protocol - an attacker could
3439 supply an oversized session ID to a client. (CVE-2002-0656)
3440 [Ben Laurie (CHATS)]
3441
3442 *) Remote buffer overflow in SSL2 protocol - an attacker could
3443 supply an oversized client master key. (CVE-2002-0656)
3444 [Ben Laurie (CHATS)]
3445
3446 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3447
3448 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3449 encoded as NULL) with id-dsa-with-sha1.
3450 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3451
3452 *) Check various X509_...() return values in apps/req.c.
3453 [Nils Larsch <nla@trustcenter.de>]
3454
3455 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3456 an end-of-file condition would erronously be flagged, when the CRLF
3457 was just at the end of a processed block. The bug was discovered when
3458 processing data through a buffering memory BIO handing the data to a
3459 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3460 <ptsekov@syntrex.com> and Nedelcho Stanev.
3461 [Lutz Jaenicke]
3462
3463 *) Implement a countermeasure against a vulnerability recently found
3464 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3465 before application data chunks to avoid the use of known IVs
3466 with data potentially chosen by the attacker.
3467 [Bodo Moeller]
3468
3469 *) Fix length checks in ssl3_get_client_hello().
3470 [Bodo Moeller]
3471
3472 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3473 to prevent ssl3_read_internal() from incorrectly assuming that
3474 ssl3_read_bytes() found application data while handshake
3475 processing was enabled when in fact s->s3->in_read_app_data was
3476 merely automatically cleared during the initial handshake.
3477 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3478
3479 *) Fix object definitions for Private and Enterprise: they were not
3480 recognized in their shortname (=lowercase) representation. Extend
3481 obj_dat.pl to issue an error when using undefined keywords instead
3482 of silently ignoring the problem (Svenning Sorensen
3483 <sss@sss.dnsalias.net>).
3484 [Lutz Jaenicke]
3485
3486 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3487 generators, i.e. generators other than 2 and 5. (Previously, the
3488 code did not properly initialise the 'add' and 'rem' values to
3489 BN_generate_prime().)
3490
3491 In the new general case, we do not insist that 'generator' is
3492 actually a primitive root: This requirement is rather pointless;
3493 a generator of the order-q subgroup is just as good, if not
3494 better.
3495 [Bodo Moeller]
3496
3497 *) Map new X509 verification errors to alerts. Discovered and submitted by
3498 Tom Wu <tom@arcot.com>.
3499 [Lutz Jaenicke]
3500
3501 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3502 returning non-zero before the data has been completely received
3503 when using non-blocking I/O.
3504 [Bodo Moeller; problem pointed out by John Hughes]
3505
3506 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3507 [Ben Laurie, Lutz Jaenicke]
3508
3509 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3510 Yoram Zahavi <YoramZ@gilian.com>).
3511 [Lutz Jaenicke]
3512
3513 *) Add information about CygWin 1.3 and on, and preserve proper
3514 configuration for the versions before that.
3515 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3516
3517 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3518 check whether we deal with a copy of a session and do not delete from
3519 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3520 <izhar@checkpoint.com>.
3521 [Lutz Jaenicke]
3522
3523 *) Do not store session data into the internal session cache, if it
3524 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3525 flag is set). Proposed by Aslam <aslam@funk.com>.
3526 [Lutz Jaenicke]
3527
3528 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3529 value is 0.
3530 [Richard Levitte]
3531
3532 *) [In 0.9.6d-engine release:]
3533 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3534 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3535
3536 *) Add the configuration target linux-s390x.
3537 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3538
3539 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3540 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3541 variable as an indication that a ClientHello message has been
3542 received. As the flag value will be lost between multiple
3543 invocations of ssl3_accept when using non-blocking I/O, the
3544 function may not be aware that a handshake has actually taken
3545 place, thus preventing a new session from being added to the
3546 session cache.
3547
3548 To avoid this problem, we now set s->new_session to 2 instead of
3549 using a local variable.
3550 [Lutz Jaenicke, Bodo Moeller]
3551
3552 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3553 if the SSL_R_LENGTH_MISMATCH error is detected.
3554 [Geoff Thorpe, Bodo Moeller]
3555
3556 *) New 'shared_ldflag' column in Configure platform table.
3557 [Richard Levitte]
3558
3559 *) Fix EVP_CIPHER_mode macro.
3560 ["Dan S. Camper" <dan@bti.net>]
3561
3562 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3563 type, we must throw them away by setting rr->length to 0.
3564 [D P Chang <dpc@qualys.com>]
3565
3566 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3567
3568 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3569 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3570 worked incorrectly for those cases where range = 10..._2 and
3571 3*range is two bits longer than range.)
3572 [Bodo Moeller]
3573
3574 *) Only add signing time to PKCS7 structures if it is not already
3575 present.
3576 [Steve Henson]
3577
3578 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3579 OBJ_ld_ce should be OBJ_id_ce.
3580 Also some ip-pda OIDs in crypto/objects/objects.txt were
3581 incorrect (cf. RFC 3039).
3582 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3583
3584 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3585 returns early because it has nothing to do.
3586 [Andy Schneider <andy.schneider@bjss.co.uk>]
3587
3588 *) [In 0.9.6c-engine release:]
3589 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3590 [Andy Schneider <andy.schneider@bjss.co.uk>]
3591
3592 *) [In 0.9.6c-engine release:]
3593 Add support for Cryptographic Appliance's keyserver technology.
3594 (Use engine 'keyclient')
3595 [Cryptographic Appliances and Geoff Thorpe]
3596
3597 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3598 is called via tools/c89.sh because arguments have to be
3599 rearranged (all '-L' options must appear before the first object
3600 modules).
3601 [Richard Shapiro <rshapiro@abinitio.com>]
3602
3603 *) [In 0.9.6c-engine release:]
3604 Add support for Broadcom crypto accelerator cards, backported
3605 from 0.9.7.
3606 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3607
3608 *) [In 0.9.6c-engine release:]
3609 Add support for SureWare crypto accelerator cards from
3610 Baltimore Technologies. (Use engine 'sureware')
3611 [Baltimore Technologies and Mark Cox]
3612
3613 *) [In 0.9.6c-engine release:]
3614 Add support for crypto accelerator cards from Accelerated
3615 Encryption Processing, www.aep.ie. (Use engine 'aep')
3616 [AEP Inc. and Mark Cox]
3617
3618 *) Add a configuration entry for gcc on UnixWare.
3619 [Gary Benson <gbenson@redhat.com>]
3620
3621 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3622 messages are stored in a single piece (fixed-length part and
3623 variable-length part combined) and fix various bugs found on the way.
3624 [Bodo Moeller]
3625
3626 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3627 instead. BIO_gethostbyname() does not know what timeouts are
3628 appropriate, so entries would stay in cache even when they have
3629 become invalid.
3630 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3631
3632 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3633 faced with a pathologically small ClientHello fragment that does
3634 not contain client_version: Instead of aborting with an error,
3635 simply choose the highest available protocol version (i.e.,
3636 TLS 1.0 unless it is disabled). In practice, ClientHello
3637 messages are never sent like this, but this change gives us
3638 strictly correct behaviour at least for TLS.
3639 [Bodo Moeller]
3640
3641 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3642 never resets s->method to s->ctx->method when called from within
3643 one of the SSL handshake functions.
3644 [Bodo Moeller; problem pointed out by Niko Baric]
3645
3646 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3647 (sent using the client's version number) if client_version is
3648 smaller than the protocol version in use. Also change
3649 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3650 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3651 the client will at least see that alert.
3652 [Bodo Moeller]
3653
3654 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3655 correctly.
3656 [Bodo Moeller]
3657
3658 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3659 client receives HelloRequest while in a handshake.
3660 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3661
3662 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3663 should end in 'break', not 'goto end' which circuments various
3664 cleanups done in state SSL_ST_OK. But session related stuff
3665 must be disabled for SSL_ST_OK in the case that we just sent a
3666 HelloRequest.
3667
3668 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3669 before just sending a HelloRequest.
3670 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3671
3672 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3673 reveal whether illegal block cipher padding was found or a MAC
3674 verification error occured. (Neither SSLerr() codes nor alerts
3675 are directly visible to potential attackers, but the information
3676 may leak via logfiles.)
3677
3678 Similar changes are not required for the SSL 2.0 implementation
3679 because the number of padding bytes is sent in clear for SSL 2.0,
3680 and the extra bytes are just ignored. However ssl/s2_pkt.c
3681 failed to verify that the purported number of padding bytes is in
3682 the legal range.
3683 [Bodo Moeller]
3684
3685 *) Add OpenUNIX-8 support including shared libraries
3686 (Boyd Lynn Gerber <gerberb@zenez.com>).
3687 [Lutz Jaenicke]
3688
3689 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3690 'wristwatch attack' using huge encoding parameters (cf.
3691 James H. Manger's CRYPTO 2001 paper). Note that the
3692 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3693 encoding parameters and hence was not vulnerable.
3694 [Bodo Moeller]
3695
3696 *) BN_sqr() bug fix.
3697 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3698
3699 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3700 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3701 followed by modular reduction.
3702 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3703
3704 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3705 equivalent based on BN_pseudo_rand() instead of BN_rand().
3706 [Bodo Moeller]
3707
3708 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3709 This function was broken, as the check for a new client hello message
3710 to handle SGC did not allow these large messages.
3711 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3712 [Lutz Jaenicke]
3713
3714 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3715 [Lutz Jaenicke]
3716
3717 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3718 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3719 [Lutz Jaenicke]
3720
3721 *) Rework the configuration and shared library support for Tru64 Unix.
3722 The configuration part makes use of modern compiler features and
3723 still retains old compiler behavior for those that run older versions
3724 of the OS. The shared library support part includes a variant that
3725 uses the RPATH feature, and is available through the special
3726 configuration target "alpha-cc-rpath", which will never be selected
3727 automatically.
3728 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3729
3730 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3731 with the same message size as in ssl3_get_certificate_request().
3732 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3733 messages might inadvertently be reject as too long.
3734 [Petr Lampa <lampa@fee.vutbr.cz>]
3735
3736 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3737 [Andy Polyakov]
3738
3739 *) Modified SSL library such that the verify_callback that has been set
3740 specificly for an SSL object with SSL_set_verify() is actually being
3741 used. Before the change, a verify_callback set with this function was
3742 ignored and the verify_callback() set in the SSL_CTX at the time of
3743 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3744 to allow the necessary settings.
3745 [Lutz Jaenicke]
3746
3747 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3748 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3749 done automatically (in contradiction to the requirements of the C
3750 standard). This made problems when used from OpenSSH.
3751 [Lutz Jaenicke]
3752
3753 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3754 dh->length and always used
3755
3756 BN_rand_range(priv_key, dh->p).
3757
3758 BN_rand_range() is not necessary for Diffie-Hellman, and this
3759 specific range makes Diffie-Hellman unnecessarily inefficient if
3760 dh->length (recommended exponent length) is much smaller than the
3761 length of dh->p. We could use BN_rand_range() if the order of
3762 the subgroup was stored in the DH structure, but we only have
3763 dh->length.
3764
3765 So switch back to
3766
3767 BN_rand(priv_key, l, ...)
3768
3769 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3770 otherwise.
3771 [Bodo Moeller]
3772
3773 *) In
3774
3775 RSA_eay_public_encrypt
3776 RSA_eay_private_decrypt
3777 RSA_eay_private_encrypt (signing)
3778 RSA_eay_public_decrypt (signature verification)
3779
3780 (default implementations for RSA_public_encrypt,
3781 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3782 always reject numbers >= n.
3783 [Bodo Moeller]
3784
3785 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3786 to synchronize access to 'locking_thread'. This is necessary on
3787 systems where access to 'locking_thread' (an 'unsigned long'
3788 variable) is not atomic.
3789 [Bodo Moeller]
3790
3791 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3792 *before* setting the 'crypto_lock_rand' flag. The previous code had
3793 a race condition if 0 is a valid thread ID.
3794 [Travis Vitek <vitek@roguewave.com>]
3795
3796 *) Add support for shared libraries under Irix.
3797 [Albert Chin-A-Young <china@thewrittenword.com>]
3798
3799 *) Add configuration option to build on Linux on both big-endian and
3800 little-endian MIPS.
3801 [Ralf Baechle <ralf@uni-koblenz.de>]
3802
3803 *) Add the possibility to create shared libraries on HP-UX.
3804 [Richard Levitte]
3805
3806 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3807
3808 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3809 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3810 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3811 PRNG state recovery was possible based on the output of
3812 one PRNG request appropriately sized to gain knowledge on
3813 'md' followed by enough consecutive 1-byte PRNG requests
3814 to traverse all of 'state'.
3815
3816 1. When updating 'md_local' (the current thread's copy of 'md')
3817 during PRNG output generation, hash all of the previous
3818 'md_local' value, not just the half used for PRNG output.
3819
3820 2. Make the number of bytes from 'state' included into the hash
3821 independent from the number of PRNG bytes requested.
3822
3823 The first measure alone would be sufficient to avoid
3824 Markku-Juhani's attack. (Actually it had never occurred
3825 to me that the half of 'md_local' used for chaining was the
3826 half from which PRNG output bytes were taken -- I had always
3827 assumed that the secret half would be used.) The second
3828 measure makes sure that additional data from 'state' is never
3829 mixed into 'md_local' in small portions; this heuristically
3830 further strengthens the PRNG.
3831 [Bodo Moeller]
3832
3833 *) Fix crypto/bn/asm/mips3.s.
3834 [Andy Polyakov]
3835
3836 *) When only the key is given to "enc", the IV is undefined. Print out
3837 an error message in this case.
3838 [Lutz Jaenicke]
3839
3840 *) Handle special case when X509_NAME is empty in X509 printing routines.
3841 [Steve Henson]
3842
3843 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3844 positive and less than q.
3845 [Bodo Moeller]
3846
3847 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3848 used: it isn't thread safe and the add_lock_callback should handle
3849 that itself.
3850 [Paul Rose <Paul.Rose@bridge.com>]
3851
3852 *) Verify that incoming data obeys the block size in
3853 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3854 [Bodo Moeller]
3855
3856 *) Fix OAEP check.
3857 [Ulf Möller, Bodo Möller]
3858
3859 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3860 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3861 when fixing the server behaviour for backwards-compatible 'client
3862 hello' messages. (Note that the attack is impractical against
3863 SSL 3.0 and TLS 1.0 anyway because length and version checking
3864 means that the probability of guessing a valid ciphertext is
3865 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3866 paper.)
3867
3868 Before 0.9.5, the countermeasure (hide the error by generating a
3869 random 'decryption result') did not work properly because
3870 ERR_clear_error() was missing, meaning that SSL_get_error() would
3871 detect the supposedly ignored error.
3872
3873 Both problems are now fixed.
3874 [Bodo Moeller]
3875
3876 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3877 (previously it was 1024).
3878 [Bodo Moeller]
3879
3880 *) Fix for compatibility mode trust settings: ignore trust settings
3881 unless some valid trust or reject settings are present.
3882 [Steve Henson]
3883
3884 *) Fix for blowfish EVP: its a variable length cipher.
3885 [Steve Henson]
3886
3887 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3888 parameters in DSA public key structures and return an error in the
3889 DSA routines if parameters are absent.
3890 [Steve Henson]
3891
3892 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3893 in the current directory if neither $RANDFILE nor $HOME was set.
3894 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3895 caused some confusion to Windows users who haven't defined $HOME.
3896 Thus RAND_file_name() is changed again: e_os.h can define a
3897 DEFAULT_HOME, which will be used if $HOME is not set.
3898 For Windows, we use "C:"; on other platforms, we still require
3899 environment variables.
3900
3901 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3902 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3903 having multiple threads call RAND_poll() concurrently.
3904 [Bodo Moeller]
3905
3906 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3907 combination of a flag and a thread ID variable.
3908 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3909 flag), *other* threads can enter ssleay_add_bytes without obeying
3910 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3911 that they do not hold after the first thread unsets add_do_not_lock).
3912 [Bodo Moeller]
3913
3914 *) Change bctest again: '-x' expressions are not available in all
3915 versions of 'test'.
3916 [Bodo Moeller]
3917
3918 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3919
3920 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3921 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3922
3923 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3924 the default extension for executables, if any. Also, make the perl
3925 scripts that use symlink() to test if it really exists and use "cp"
3926 if it doesn't. All this made OpenSSL compilable and installable in
3927 CygWin.
3928 [Richard Levitte]
3929
3930 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3931 If SEQUENCE is length is indefinite just set c->slen to the total
3932 amount of data available.
3933 [Steve Henson, reported by shige@FreeBSD.org]
3934 [This change does not apply to 0.9.7.]
3935
3936 *) Change bctest to avoid here-documents inside command substitution
3937 (workaround for FreeBSD /bin/sh bug).
3938 For compatibility with Ultrix, avoid shell functions (introduced
3939 in the bctest version that searches along $PATH).
3940 [Bodo Moeller]
3941
3942 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3943 with des_encrypt() defined on some operating systems, like Solaris
3944 and UnixWare.
3945 [Richard Levitte]
3946
3947 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3948 On the Importance of Eliminating Errors in Cryptographic
3949 Computations, J. Cryptology 14 (2001) 2, 101-119,
3950 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
3951 [Ulf Moeller]
3952
3953 *) MIPS assembler BIGNUM division bug fix.
3954 [Andy Polyakov]
3955
3956 *) Disabled incorrect Alpha assembler code.
3957 [Richard Levitte]
3958
3959 *) Fix PKCS#7 decode routines so they correctly update the length
3960 after reading an EOC for the EXPLICIT tag.
3961 [Steve Henson]
3962 [This change does not apply to 0.9.7.]
3963
3964 *) Fix bug in PKCS#12 key generation routines. This was triggered
3965 if a 3DES key was generated with a 0 initial byte. Include
3966 PKCS12_BROKEN_KEYGEN compilation option to retain the old
3967 (but broken) behaviour.
3968 [Steve Henson]
3969
3970 *) Enhance bctest to search for a working bc along $PATH and print
3971 it when found.
3972 [Tim Rice <tim@multitalents.net> via Richard Levitte]
3973
3974 *) Fix memory leaks in err.c: free err_data string if necessary;
3975 don't write to the wrong index in ERR_set_error_data.
3976 [Bodo Moeller]
3977
3978 *) Implement ssl23_peek (analogous to ssl23_read), which previously
3979 did not exist.
3980 [Bodo Moeller]
3981
3982 *) Replace rdtsc with _emit statements for VC++ version 5.
3983 [Jeremy Cooper <jeremy@baymoo.org>]
3984
3985 *) Make it possible to reuse SSLv2 sessions.
3986 [Richard Levitte]
3987
3988 *) In copy_email() check for >= 0 as a return value for
3989 X509_NAME_get_index_by_NID() since 0 is a valid index.
3990 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
3991
3992 *) Avoid coredump with unsupported or invalid public keys by checking if
3993 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
3994 PKCS7_verify() fails with non detached data.
3995 [Steve Henson]
3996
3997 *) Don't use getenv in library functions when run as setuid/setgid.
3998 New function OPENSSL_issetugid().
3999 [Ulf Moeller]
4000
4001 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4002 due to incorrect handling of multi-threading:
4003
4004 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4005
4006 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4007
4008 3. Count how many times MemCheck_off() has been called so that
4009 nested use can be treated correctly. This also avoids
4010 inband-signalling in the previous code (which relied on the
4011 assumption that thread ID 0 is impossible).
4012 [Bodo Moeller]
4013
4014 *) Add "-rand" option also to s_client and s_server.
4015 [Lutz Jaenicke]
4016
4017 *) Fix CPU detection on Irix 6.x.
4018 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4019 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4020
4021 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4022 was empty.
4023 [Steve Henson]
4024 [This change does not apply to 0.9.7.]
4025
4026 *) Use the cached encoding of an X509_NAME structure rather than
4027 copying it. This is apparently the reason for the libsafe "errors"
4028 but the code is actually correct.
4029 [Steve Henson]
4030
4031 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4032 Bleichenbacher's DSA attack.
4033 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4034 to be set and top=0 forces the highest bit to be set; top=-1 is new
4035 and leaves the highest bit random.
4036 [Ulf Moeller, Bodo Moeller]
4037
4038 *) In the NCONF_...-based implementations for CONF_... queries
4039 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4040 a temporary CONF structure with the data component set to NULL
4041 (which gives segmentation faults in lh_retrieve).
4042 Instead, use NULL for the CONF pointer in CONF_get_string and
4043 CONF_get_number (which may use environment variables) and directly
4044 return NULL from CONF_get_section.
4045 [Bodo Moeller]
4046
4047 *) Fix potential buffer overrun for EBCDIC.
4048 [Ulf Moeller]
4049
4050 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4051 keyUsage if basicConstraints absent for a CA.
4052 [Steve Henson]
4053
4054 *) Make SMIME_write_PKCS7() write mail header values with a format that
4055 is more generally accepted (no spaces before the semicolon), since
4056 some programs can't parse those values properly otherwise. Also make
4057 sure BIO's that break lines after each write do not create invalid
4058 headers.
4059 [Richard Levitte]
4060
4061 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4062 macros previously used would not encode an empty SEQUENCE OF
4063 and break the signature.
4064 [Steve Henson]
4065 [This change does not apply to 0.9.7.]
4066
4067 *) Zero the premaster secret after deriving the master secret in
4068 DH ciphersuites.
4069 [Steve Henson]
4070
4071 *) Add some EVP_add_digest_alias registrations (as found in
4072 OpenSSL_add_all_digests()) to SSL_library_init()
4073 aka OpenSSL_add_ssl_algorithms(). This provides improved
4074 compatibility with peers using X.509 certificates
4075 with unconventional AlgorithmIdentifier OIDs.
4076 [Bodo Moeller]
4077
4078 *) Fix for Irix with NO_ASM.
4079 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4080
4081 *) ./config script fixes.
4082 [Ulf Moeller, Richard Levitte]
4083
4084 *) Fix 'openssl passwd -1'.
4085 [Bodo Moeller]
4086
4087 *) Change PKCS12_key_gen_asc() so it can cope with non null
4088 terminated strings whose length is passed in the passlen
4089 parameter, for example from PEM callbacks. This was done
4090 by adding an extra length parameter to asc2uni().
4091 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4092
4093 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4094 call failed, free the DSA structure.
4095 [Bodo Moeller]
4096
4097 *) Fix to uni2asc() to cope with zero length Unicode strings.
4098 These are present in some PKCS#12 files.
4099 [Steve Henson]
4100
4101 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4102 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4103 when writing a 32767 byte record.
4104 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4105
4106 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4107 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4108
4109 (RSA objects have a reference count access to which is protected
4110 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4111 so they are meant to be shared between threads.)
4112 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4113 "Reddie, Steven" <Steven.Reddie@ca.com>]
4114
4115 *) Fix a deadlock in CRYPTO_mem_leaks().
4116 [Bodo Moeller]
4117
4118 *) Use better test patterns in bntest.
4119 [Ulf Möller]
4120
4121 *) rand_win.c fix for Borland C.
4122 [Ulf Möller]
4123
4124 *) BN_rshift bugfix for n == 0.
4125 [Bodo Moeller]
4126
4127 *) Add a 'bctest' script that checks for some known 'bc' bugs
4128 so that 'make test' does not abort just because 'bc' is broken.
4129 [Bodo Moeller]
4130
4131 *) Store verify_result within SSL_SESSION also for client side to
4132 avoid potential security hole. (Re-used sessions on the client side
4133 always resulted in verify_result==X509_V_OK, not using the original
4134 result of the server certificate verification.)
4135 [Lutz Jaenicke]
4136
4137 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4138 SSL3_RT_APPLICATION_DATA, return 0.
4139 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4140 [Bodo Moeller]
4141
4142 *) Fix SSL_peek:
4143 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4144 releases, have been re-implemented by renaming the previous
4145 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4146 and ssl3_read_internal, respectively, and adding 'peek' parameters
4147 to them. The new ssl[23]_{read,peek} functions are calls to
4148 ssl[23]_read_internal with the 'peek' flag set appropriately.
4149 A 'peek' parameter has also been added to ssl3_read_bytes, which
4150 does the actual work for ssl3_read_internal.
4151 [Bodo Moeller]
4152
4153 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4154 the method-specific "init()" handler. Also clean up ex_data after
4155 calling the method-specific "finish()" handler. Previously, this was
4156 happening the other way round.
4157 [Geoff Thorpe]
4158
4159 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4160 The previous value, 12, was not always sufficient for BN_mod_exp().
4161 [Bodo Moeller]
4162
4163 *) Make sure that shared libraries get the internal name engine with
4164 the full version number and not just 0. This should mark the
4165 shared libraries as not backward compatible. Of course, this should
4166 be changed again when we can guarantee backward binary compatibility.
4167 [Richard Levitte]
4168
4169 *) Fix typo in get_cert_by_subject() in by_dir.c
4170 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4171
4172 *) Rework the system to generate shared libraries:
4173
4174 - Make note of the expected extension for the shared libraries and
4175 if there is a need for symbolic links from for example libcrypto.so.0
4176 to libcrypto.so.0.9.7. There is extended info in Configure for
4177 that.
4178
4179 - Make as few rebuilds of the shared libraries as possible.
4180
4181 - Still avoid linking the OpenSSL programs with the shared libraries.
4182
4183 - When installing, install the shared libraries separately from the
4184 static ones.
4185 [Richard Levitte]
4186
4187 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4188
4189 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4190 and not in SSL_clear because the latter is also used by the
4191 accept/connect functions; previously, the settings made by
4192 SSL_set_read_ahead would be lost during the handshake.
4193 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4194
4195 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4196 Previously, it would create entries for disableed algorithms no
4197 matter what.
4198 [Richard Levitte]
4199
4200 *) Added several new manual pages for SSL_* function.
4201 [Lutz Jaenicke]
4202
4203 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4204
4205 *) In ssl23_get_client_hello, generate an error message when faced
4206 with an initial SSL 3.0/TLS record that is too small to contain the
4207 first two bytes of the ClientHello message, i.e. client_version.
4208 (Note that this is a pathologic case that probably has never happened
4209 in real life.) The previous approach was to use the version number
4210 from the record header as a substitute; but our protocol choice
4211 should not depend on that one because it is not authenticated
4212 by the Finished messages.
4213 [Bodo Moeller]
4214
4215 *) More robust randomness gathering functions for Windows.
4216 [Jeffrey Altman <jaltman@columbia.edu>]
4217
4218 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4219 not set then we don't setup the error code for issuer check errors
4220 to avoid possibly overwriting other errors which the callback does
4221 handle. If an application does set the flag then we assume it knows
4222 what it is doing and can handle the new informational codes
4223 appropriately.
4224 [Steve Henson]
4225
4226 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4227 a general "ANY" type, as such it should be able to decode anything
4228 including tagged types. However it didn't check the class so it would
4229 wrongly interpret tagged types in the same way as their universal
4230 counterpart and unknown types were just rejected. Changed so that the
4231 tagged and unknown types are handled in the same way as a SEQUENCE:
4232 that is the encoding is stored intact. There is also a new type
4233 "V_ASN1_OTHER" which is used when the class is not universal, in this
4234 case we have no idea what the actual type is so we just lump them all
4235 together.
4236 [Steve Henson]
4237
4238 *) On VMS, stdout may very well lead to a file that is written to
4239 in a record-oriented fashion. That means that every write() will
4240 write a separate record, which will be read separately by the
4241 programs trying to read from it. This can be very confusing.
4242
4243 The solution is to put a BIO filter in the way that will buffer
4244 text until a linefeed is reached, and then write everything a
4245 line at a time, so every record written will be an actual line,
4246 not chunks of lines and not (usually doesn't happen, but I've
4247 seen it once) several lines in one record. BIO_f_linebuffer() is
4248 the answer.
4249
4250 Currently, it's a VMS-only method, because that's where it has
4251 been tested well enough.
4252 [Richard Levitte]
4253
4254 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4255 it can return incorrect results.
4256 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4257 but it was in 0.9.6-beta[12].)
4258 [Bodo Moeller]
4259
4260 *) Disable the check for content being present when verifying detached
4261 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4262 include zero length content when signing messages.
4263 [Steve Henson]
4264
4265 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4266 BIO_ctrl (for BIO pairs).
4267 [Bodo Möller]
4268
4269 *) Add DSO method for VMS.
4270 [Richard Levitte]
4271
4272 *) Bug fix: Montgomery multiplication could produce results with the
4273 wrong sign.
4274 [Ulf Möller]
4275
4276 *) Add RPM specification openssl.spec and modify it to build three
4277 packages. The default package contains applications, application
4278 documentation and run-time libraries. The devel package contains
4279 include files, static libraries and function documentation. The
4280 doc package contains the contents of the doc directory. The original
4281 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4282 [Richard Levitte]
4283
4284 *) Add a large number of documentation files for many SSL routines.
4285 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4286
4287 *) Add a configuration entry for Sony News 4.
4288 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4289
4290 *) Don't set the two most significant bits to one when generating a
4291 random number < q in the DSA library.
4292 [Ulf Möller]
4293
4294 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4295 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4296 the underlying transport is blocking) if a handshake took place.
4297 (The default behaviour is needed by applications such as s_client
4298 and s_server that use select() to determine when to use SSL_read;
4299 but for applications that know in advance when to expect data, it
4300 just makes things more complicated.)
4301 [Bodo Moeller]
4302
4303 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4304 from EGD.
4305 [Ben Laurie]
4306
4307 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4308 work better on such systems.
4309 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4310
4311 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4312 Update PKCS12_parse() so it copies the friendlyName and the
4313 keyid to the certificates aux info.
4314 [Steve Henson]
4315
4316 *) Fix bug in PKCS7_verify() which caused an infinite loop
4317 if there was more than one signature.
4318 [Sven Uszpelkat <su@celocom.de>]
4319
4320 *) Major change in util/mkdef.pl to include extra information
4321 about each symbol, as well as presentig variables as well
4322 as functions. This change means that there's n more need
4323 to rebuild the .num files when some algorithms are excluded.
4324 [Richard Levitte]
4325
4326 *) Allow the verify time to be set by an application,
4327 rather than always using the current time.
4328 [Steve Henson]
4329
4330 *) Phase 2 verify code reorganisation. The certificate
4331 verify code now looks up an issuer certificate by a
4332 number of criteria: subject name, authority key id
4333 and key usage. It also verifies self signed certificates
4334 by the same criteria. The main comparison function is
4335 X509_check_issued() which performs these checks.
4336
4337 Lot of changes were necessary in order to support this
4338 without completely rewriting the lookup code.
4339
4340 Authority and subject key identifier are now cached.
4341
4342 The LHASH 'certs' is X509_STORE has now been replaced
4343 by a STACK_OF(X509_OBJECT). This is mainly because an
4344 LHASH can't store or retrieve multiple objects with
4345 the same hash value.
4346
4347 As a result various functions (which were all internal
4348 use only) have changed to handle the new X509_STORE
4349 structure. This will break anything that messed round
4350 with X509_STORE internally.
4351
4352 The functions X509_STORE_add_cert() now checks for an
4353 exact match, rather than just subject name.
4354
4355 The X509_STORE API doesn't directly support the retrieval
4356 of multiple certificates matching a given criteria, however
4357 this can be worked round by performing a lookup first
4358 (which will fill the cache with candidate certificates)
4359 and then examining the cache for matches. This is probably
4360 the best we can do without throwing out X509_LOOKUP
4361 entirely (maybe later...).
4362
4363 The X509_VERIFY_CTX structure has been enhanced considerably.
4364
4365 All certificate lookup operations now go via a get_issuer()
4366 callback. Although this currently uses an X509_STORE it
4367 can be replaced by custom lookups. This is a simple way
4368 to bypass the X509_STORE hackery necessary to make this
4369 work and makes it possible to use more efficient techniques
4370 in future. A very simple version which uses a simple
4371 STACK for its trusted certificate store is also provided
4372 using X509_STORE_CTX_trusted_stack().
4373
4374 The verify_cb() and verify() callbacks now have equivalents
4375 in the X509_STORE_CTX structure.
4376
4377 X509_STORE_CTX also has a 'flags' field which can be used
4378 to customise the verify behaviour.
4379 [Steve Henson]
4380
4381 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4382 excludes S/MIME capabilities.
4383 [Steve Henson]
4384
4385 *) When a certificate request is read in keep a copy of the
4386 original encoding of the signed data and use it when outputing
4387 again. Signatures then use the original encoding rather than
4388 a decoded, encoded version which may cause problems if the
4389 request is improperly encoded.
4390 [Steve Henson]
4391
4392 *) For consistency with other BIO_puts implementations, call
4393 buffer_write(b, ...) directly in buffer_puts instead of calling
4394 BIO_write(b, ...).
4395
4396 In BIO_puts, increment b->num_write as in BIO_write.
4397 [Peter.Sylvester@EdelWeb.fr]
4398
4399 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4400 BN_zero, we may not return a BIGNUM with an array consisting of
4401 words set to zero.)
4402 [Bodo Moeller]
4403
4404 *) Avoid calling abort() from within the library when problems are
4405 detected, except if preprocessor symbols have been defined
4406 (such as REF_CHECK, BN_DEBUG etc.).
4407 [Bodo Moeller]
4408
4409 *) New openssl application 'rsautl'. This utility can be
4410 used for low level RSA operations. DER public key
4411 BIO/fp routines also added.
4412 [Steve Henson]
4413
4414 *) New Configure entry and patches for compiling on QNX 4.
4415 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4416
4417 *) A demo state-machine implementation was sponsored by
4418 Nuron (http://www.nuron.com/) and is now available in
4419 demos/state_machine.
4420 [Ben Laurie]
4421
4422 *) New options added to the 'dgst' utility for signature
4423 generation and verification.
4424 [Steve Henson]
4425
4426 *) Unrecognized PKCS#7 content types are now handled via a
4427 catch all ASN1_TYPE structure. This allows unsupported
4428 types to be stored as a "blob" and an application can
4429 encode and decode it manually.
4430 [Steve Henson]
4431
4432 *) Fix various signed/unsigned issues to make a_strex.c
4433 compile under VC++.
4434 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4435
4436 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4437 length if passed a buffer. ASN1_INTEGER_to_BN failed
4438 if passed a NULL BN and its argument was negative.
4439 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4440
4441 *) Modification to PKCS#7 encoding routines to output definite
4442 length encoding. Since currently the whole structures are in
4443 memory there's not real point in using indefinite length
4444 constructed encoding. However if OpenSSL is compiled with
4445 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4446 [Steve Henson]
4447
4448 *) Added BIO_vprintf() and BIO_vsnprintf().
4449 [Richard Levitte]
4450
4451 *) Added more prefixes to parse for in the the strings written
4452 through a logging bio, to cover all the levels that are available
4453 through syslog. The prefixes are now:
4454
4455 PANIC, EMERG, EMR => LOG_EMERG
4456 ALERT, ALR => LOG_ALERT
4457 CRIT, CRI => LOG_CRIT
4458 ERROR, ERR => LOG_ERR
4459 WARNING, WARN, WAR => LOG_WARNING
4460 NOTICE, NOTE, NOT => LOG_NOTICE
4461 INFO, INF => LOG_INFO
4462 DEBUG, DBG => LOG_DEBUG
4463
4464 and as before, if none of those prefixes are present at the
4465 beginning of the string, LOG_ERR is chosen.
4466
4467 On Win32, the LOG_* levels are mapped according to this:
4468
4469 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4470 LOG_WARNING => EVENTLOG_WARNING_TYPE
4471 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4472
4473 [Richard Levitte]
4474
4475 *) Made it possible to reconfigure with just the configuration
4476 argument "reconf" or "reconfigure". The command line arguments
4477 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4478 and are retrieved from there when reconfiguring.
4479 [Richard Levitte]
4480
4481 *) MD4 implemented.
4482 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4483
4484 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4485 [Richard Levitte]
4486
4487 *) The obj_dat.pl script was messing up the sorting of object
4488 names. The reason was that it compared the quoted version
4489 of strings as a result "OCSP" > "OCSP Signing" because
4490 " > SPACE. Changed script to store unquoted versions of
4491 names and add quotes on output. It was also omitting some
4492 names from the lookup table if they were given a default
4493 value (that is if SN is missing it is given the same
4494 value as LN and vice versa), these are now added on the
4495 grounds that if an object has a name we should be able to
4496 look it up. Finally added warning output when duplicate
4497 short or long names are found.
4498 [Steve Henson]
4499
4500 *) Changes needed for Tandem NSK.
4501 [Scott Uroff <scott@xypro.com>]
4502
4503 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4504 RSA_padding_check_SSLv23(), special padding was never detected
4505 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4506 version rollback attacks was not effective.
4507
4508 In s23_clnt.c, don't use special rollback-attack detection padding
4509 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4510 client; similarly, in s23_srvr.c, don't do the rollback check if
4511 SSL 2.0 is the only protocol enabled in the server.
4512 [Bodo Moeller]
4513
4514 *) Make it possible to get hexdumps of unprintable data with 'openssl
4515 asn1parse'. By implication, the functions ASN1_parse_dump() and
4516 BIO_dump_indent() are added.
4517 [Richard Levitte]
4518
4519 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4520 these print out strings and name structures based on various
4521 flags including RFC2253 support and proper handling of
4522 multibyte characters. Added options to the 'x509' utility
4523 to allow the various flags to be set.
4524 [Steve Henson]
4525
4526 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4527 Also change the functions X509_cmp_current_time() and
4528 X509_gmtime_adj() work with an ASN1_TIME structure,
4529 this will enable certificates using GeneralizedTime in validity
4530 dates to be checked.
4531 [Steve Henson]
4532
4533 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4534 negative public key encodings) on by default,
4535 NO_NEG_PUBKEY_BUG can be set to disable it.
4536 [Steve Henson]
4537
4538 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4539 content octets. An i2c_ASN1_OBJECT is unnecessary because
4540 the encoding can be trivially obtained from the structure.
4541 [Steve Henson]
4542
4543 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4544 not read locks (CRYPTO_r_[un]lock).
4545 [Bodo Moeller]
4546
4547 *) A first attempt at creating official support for shared
4548 libraries through configuration. I've kept it so the
4549 default is static libraries only, and the OpenSSL programs
4550 are always statically linked for now, but there are
4551 preparations for dynamic linking in place.
4552 This has been tested on Linux and Tru64.
4553 [Richard Levitte]
4554
4555 *) Randomness polling function for Win9x, as described in:
4556 Peter Gutmann, Software Generation of Practically Strong
4557 Random Numbers.
4558 [Ulf Möller]
4559
4560 *) Fix so PRNG is seeded in req if using an already existing
4561 DSA key.
4562 [Steve Henson]
4563
4564 *) New options to smime application. -inform and -outform
4565 allow alternative formats for the S/MIME message including
4566 PEM and DER. The -content option allows the content to be
4567 specified separately. This should allow things like Netscape
4568 form signing output easier to verify.
4569 [Steve Henson]
4570
4571 *) Fix the ASN1 encoding of tags using the 'long form'.
4572 [Steve Henson]
4573
4574 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4575 STRING types. These convert content octets to and from the
4576 underlying type. The actual tag and length octets are
4577 already assumed to have been read in and checked. These
4578 are needed because all other string types have virtually
4579 identical handling apart from the tag. By having versions
4580 of the ASN1 functions that just operate on content octets
4581 IMPLICIT tagging can be handled properly. It also allows
4582 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4583 and ASN1_INTEGER are identical apart from the tag.
4584 [Steve Henson]
4585
4586 *) Change the handling of OID objects as follows:
4587
4588 - New object identifiers are inserted in objects.txt, following
4589 the syntax given in objects.README.
4590 - objects.pl is used to process obj_mac.num and create a new
4591 obj_mac.h.
4592 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4593 obj_mac.h.
4594
4595 This is currently kind of a hack, and the perl code in objects.pl
4596 isn't very elegant, but it works as I intended. The simplest way
4597 to check that it worked correctly is to look in obj_dat.h and
4598 check the array nid_objs and make sure the objects haven't moved
4599 around (this is important!). Additions are OK, as well as
4600 consistent name changes.
4601 [Richard Levitte]
4602
4603 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4604 [Bodo Moeller]
4605
4606 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4607 The given file adds to whatever has already been seeded into the
4608 random pool through the RANDFILE configuration file option or
4609 environment variable, or the default random state file.
4610 [Richard Levitte]
4611
4612 *) mkstack.pl now sorts each macro group into lexical order.
4613 Previously the output order depended on the order the files
4614 appeared in the directory, resulting in needless rewriting
4615 of safestack.h .
4616 [Steve Henson]
4617
4618 *) Patches to make OpenSSL compile under Win32 again. Mostly
4619 work arounds for the VC++ problem that it treats func() as
4620 func(void). Also stripped out the parts of mkdef.pl that
4621 added extra typesafe functions: these no longer exist.
4622 [Steve Henson]
4623
4624 *) Reorganisation of the stack code. The macros are now all
4625 collected in safestack.h . Each macro is defined in terms of
4626 a "stack macro" of the form SKM_<name>(type, a, b). The
4627 DEBUG_SAFESTACK is now handled in terms of function casts,
4628 this has the advantage of retaining type safety without the
4629 use of additional functions. If DEBUG_SAFESTACK is not defined
4630 then the non typesafe macros are used instead. Also modified the
4631 mkstack.pl script to handle the new form. Needs testing to see
4632 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4633 the default if no major problems. Similar behaviour for ASN1_SET_OF
4634 and PKCS12_STACK_OF.
4635 [Steve Henson]
4636
4637 *) When some versions of IIS use the 'NET' form of private key the
4638 key derivation algorithm is different. Normally MD5(password) is
4639 used as a 128 bit RC4 key. In the modified case
4640 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4641 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4642 as the old Netscape_RSA functions except they have an additional
4643 'sgckey' parameter which uses the modified algorithm. Also added
4644 an -sgckey command line option to the rsa utility. Thanks to
4645 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4646 algorithm to openssl-dev.
4647 [Steve Henson]
4648
4649 *) The evp_local.h macros were using 'c.##kname' which resulted in
4650 invalid expansion on some systems (SCO 5.0.5 for example).
4651 Corrected to 'c.kname'.
4652 [Phillip Porch <root@theporch.com>]
4653
4654 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4655 a STACK of email addresses from a certificate or request, these look
4656 in the subject name and the subject alternative name extensions and
4657 omit any duplicate addresses.
4658 [Steve Henson]
4659
4660 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4661 This makes DSA verification about 2 % faster.
4662 [Bodo Moeller]
4663
4664 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4665 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4666 plus overhead for 1024 bit moduli).
4667 This makes exponentiations about 0.5 % faster for 1024 bit
4668 exponents (as measured by "openssl speed rsa2048").
4669 [Bodo Moeller]
4670
4671 *) Rename memory handling macros to avoid conflicts with other
4672 software:
4673 Malloc => OPENSSL_malloc
4674 Malloc_locked => OPENSSL_malloc_locked
4675 Realloc => OPENSSL_realloc
4676 Free => OPENSSL_free
4677 [Richard Levitte]
4678
4679 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4680 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4681 [Bodo Moeller]
4682
4683 *) CygWin32 support.
4684 [John Jarvie <jjarvie@newsguy.com>]
4685
4686 *) The type-safe stack code has been rejigged. It is now only compiled
4687 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4688 by default all type-specific stack functions are "#define"d back to
4689 standard stack functions. This results in more streamlined output
4690 but retains the type-safety checking possibilities of the original
4691 approach.
4692 [Geoff Thorpe]
4693
4694 *) The STACK code has been cleaned up, and certain type declarations
4695 that didn't make a lot of sense have been brought in line. This has
4696 also involved a cleanup of sorts in safestack.h to more correctly
4697 map type-safe stack functions onto their plain stack counterparts.
4698 This work has also resulted in a variety of "const"ifications of
4699 lots of the code, especially "_cmp" operations which should normally
4700 be prototyped with "const" parameters anyway.
4701 [Geoff Thorpe]
4702
4703 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4704 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4705 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4706 where all of 'md' is used each time the PRNG is used, but 'state'
4707 is used only indexed by a cyclic counter. As entropy may not be
4708 well distributed from the beginning, 'md' is important as a
4709 chaining variable. However, the output function chains only half
4710 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4711 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4712 in all of 'state' being rewritten, with the new values depending
4713 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4714 [Bodo Moeller]
4715
4716 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4717 the handshake is continued after ssl_verify_cert_chain();
4718 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4719 can lead to 'unexplainable' connection aborts later.
4720 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4721
4722 *) Major EVP API cipher revision.
4723 Add hooks for extra EVP features. This allows various cipher
4724 parameters to be set in the EVP interface. Support added for variable
4725 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4726 setting of RC2 and RC5 parameters.
4727
4728 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4729 ciphers.
4730
4731 Remove lots of duplicated code from the EVP library. For example *every*
4732 cipher init() function handles the 'iv' in the same way according to the
4733 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4734 for CFB and OFB modes they zero ctx->num.
4735
4736 New functionality allows removal of S/MIME code RC2 hack.
4737
4738 Most of the routines have the same form and so can be declared in terms
4739 of macros.
4740
4741 By shifting this to the top level EVP_CipherInit() it can be removed from
4742 all individual ciphers. If the cipher wants to handle IVs or keys
4743 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4744 flags.
4745
4746 Change lots of functions like EVP_EncryptUpdate() to now return a
4747 value: although software versions of the algorithms cannot fail
4748 any installed hardware versions can.
4749 [Steve Henson]
4750
4751 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4752 this option is set, tolerate broken clients that send the negotiated
4753 protocol version number instead of the requested protocol version
4754 number.
4755 [Bodo Moeller]
4756
4757 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4758 i.e. non-zero for export ciphersuites, zero otherwise.
4759 Previous versions had this flag inverted, inconsistent with
4760 rsa_tmp_cb (..._TMP_RSA_CB).
4761 [Bodo Moeller; problem reported by Amit Chopra]
4762
4763 *) Add missing DSA library text string. Work around for some IIS
4764 key files with invalid SEQUENCE encoding.
4765 [Steve Henson]
4766
4767 *) Add a document (doc/standards.txt) that list all kinds of standards
4768 and so on that are implemented in OpenSSL.
4769 [Richard Levitte]
4770
4771 *) Enhance c_rehash script. Old version would mishandle certificates
4772 with the same subject name hash and wouldn't handle CRLs at all.
4773 Added -fingerprint option to crl utility, to support new c_rehash
4774 features.
4775 [Steve Henson]
4776
4777 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4778 [Ulf Möller]
4779
4780 *) Fix for SSL server purpose checking. Server checking was
4781 rejecting certificates which had extended key usage present
4782 but no ssl client purpose.
4783 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4784
4785 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4786 is a little unclear about how a blank password is handled.
4787 Since the password in encoded as a BMPString with terminating
4788 double NULL a zero length password would end up as just the
4789 double NULL. However no password at all is different and is
4790 handled differently in the PKCS#12 key generation code. NS
4791 treats a blank password as zero length. MSIE treats it as no
4792 password on export: but it will try both on import. We now do
4793 the same: PKCS12_parse() tries zero length and no password if
4794 the password is set to "" or NULL (NULL is now a valid password:
4795 it wasn't before) as does the pkcs12 application.
4796 [Steve Henson]
4797
4798 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4799 perror when PEM_read_bio_X509_REQ fails, the error message must
4800 be obtained from the error queue.
4801 [Bodo Moeller]
4802
4803 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4804 it in ERR_remove_state if appropriate, and change ERR_get_state
4805 accordingly to avoid race conditions (this is necessary because
4806 thread_hash is no longer constant once set).
4807 [Bodo Moeller]
4808
4809 *) Bugfix for linux-elf makefile.one.
4810 [Ulf Möller]
4811
4812 *) RSA_get_default_method() will now cause a default
4813 RSA_METHOD to be chosen if one doesn't exist already.
4814 Previously this was only set during a call to RSA_new()
4815 or RSA_new_method(NULL) meaning it was possible for
4816 RSA_get_default_method() to return NULL.
4817 [Geoff Thorpe]
4818
4819 *) Added native name translation to the existing DSO code
4820 that will convert (if the flag to do so is set) filenames
4821 that are sufficiently small and have no path information
4822 into a canonical native form. Eg. "blah" converted to
4823 "libblah.so" or "blah.dll" etc.
4824 [Geoff Thorpe]
4825
4826 *) New function ERR_error_string_n(e, buf, len) which is like
4827 ERR_error_string(e, buf), but writes at most 'len' bytes
4828 including the 0 terminator. For ERR_error_string_n, 'buf'
4829 may not be NULL.
4830 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4831
4832 *) CONF library reworked to become more general. A new CONF
4833 configuration file reader "class" is implemented as well as a
4834 new functions (NCONF_*, for "New CONF") to handle it. The now
4835 old CONF_* functions are still there, but are reimplemented to
4836 work in terms of the new functions. Also, a set of functions
4837 to handle the internal storage of the configuration data is
4838 provided to make it easier to write new configuration file
4839 reader "classes" (I can definitely see something reading a
4840 configuration file in XML format, for example), called _CONF_*,
4841 or "the configuration storage API"...
4842
4843 The new configuration file reading functions are:
4844
4845 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4846 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4847
4848 NCONF_default, NCONF_WIN32
4849
4850 NCONF_dump_fp, NCONF_dump_bio
4851
4852 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4853 NCONF_new creates a new CONF object. This works in the same way
4854 as other interfaces in OpenSSL, like the BIO interface.
4855 NCONF_dump_* dump the internal storage of the configuration file,
4856 which is useful for debugging. All other functions take the same
4857 arguments as the old CONF_* functions wth the exception of the
4858 first that must be a `CONF *' instead of a `LHASH *'.
4859
4860 To make it easer to use the new classes with the old CONF_* functions,
4861 the function CONF_set_default_method is provided.
4862 [Richard Levitte]
4863
4864 *) Add '-tls1' option to 'openssl ciphers', which was already
4865 mentioned in the documentation but had not been implemented.
4866 (This option is not yet really useful because even the additional
4867 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4868 [Bodo Moeller]
4869
4870 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4871 OpenSSL-based applications) load shared libraries and bind to
4872 them in a portable way.
4873 [Geoff Thorpe, with contributions from Richard Levitte]
4874
4875 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4876
4877 *) Make sure _lrotl and _lrotr are only used with MSVC.
4878
4879 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4880 (the default implementation of RAND_status).
4881
4882 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4883 to '-clrext' (= clear extensions), as intended and documented.
4884 [Bodo Moeller; inconsistency pointed out by Michael Attili
4885 <attili@amaxo.com>]
4886
4887 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4888 was larger than the MD block size.
4889 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4890
4891 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4892 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4893 using the passed key: if the passed key was a private key the result
4894 of X509_print(), for example, would be to print out all the private key
4895 components.
4896 [Steve Henson]
4897
4898 *) des_quad_cksum() byte order bug fix.
4899 [Ulf Möller, using the problem description in krb4-0.9.7, where
4900 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4901
4902 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4903 discouraged.
4904 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4905
4906 *) For easily testing in shell scripts whether some command
4907 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4908 returns with exit code 0 iff no command of the given name is available.
4909 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4910 the output goes to stdout and nothing is printed to stderr.
4911 Additional arguments are always ignored.
4912
4913 Since for each cipher there is a command of the same name,
4914 the 'no-cipher' compilation switches can be tested this way.
4915
4916 ('openssl no-XXX' is not able to detect pseudo-commands such
4917 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4918 [Bodo Moeller]
4919
4920 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4921 [Bodo Moeller]
4922
4923 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4924 is set; it will be thrown away anyway because each handshake creates
4925 its own key.
4926 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4927 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4928 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4929 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4930 [Bodo Moeller]
4931
4932 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4933 'Q' and 'R' lose their special meanings (quit/renegotiate).
4934 This is part of what -quiet does; unlike -quiet, -ign_eof
4935 does not suppress any output.
4936 [Richard Levitte]
4937
4938 *) Add compatibility options to the purpose and trust code. The
4939 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4940 accepts a certificate or CA, this was the previous behaviour,
4941 with all the associated security issues.
4942
4943 X509_TRUST_COMPAT is the old trust behaviour: only and
4944 automatically trust self signed roots in certificate store. A
4945 new trust setting X509_TRUST_DEFAULT is used to specify that
4946 a purpose has no associated trust setting and it should instead
4947 use the value in the default purpose.
4948 [Steve Henson]
4949
4950 *) Fix the PKCS#8 DSA private key code so it decodes keys again
4951 and fix a memory leak.
4952 [Steve Henson]
4953
4954 *) In util/mkerr.pl (which implements 'make errors'), preserve
4955 reason strings from the previous version of the .c file, as
4956 the default to have only downcase letters (and digits) in
4957 automatically generated reasons codes is not always appropriate.
4958 [Bodo Moeller]
4959
4960 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
4961 using strerror. Previously, ERR_reason_error_string() returned
4962 library names as reason strings for SYSerr; but SYSerr is a special
4963 case where small numbers are errno values, not library numbers.
4964 [Bodo Moeller]
4965
4966 *) Add '-dsaparam' option to 'openssl dhparam' application. This
4967 converts DSA parameters into DH parameters. (When creating parameters,
4968 DSA_generate_parameters is used.)
4969 [Bodo Moeller]
4970
4971 *) Include 'length' (recommended exponent length) in C code generated
4972 by 'openssl dhparam -C'.
4973 [Bodo Moeller]
4974
4975 *) The second argument to set_label in perlasm was already being used
4976 so couldn't be used as a "file scope" flag. Moved to third argument
4977 which was free.
4978 [Steve Henson]
4979
4980 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
4981 instead of RAND_bytes for encryption IVs and salts.
4982 [Bodo Moeller]
4983
4984 *) Include RAND_status() into RAND_METHOD instead of implementing
4985 it only for md_rand.c Otherwise replacing the PRNG by calling
4986 RAND_set_rand_method would be impossible.
4987 [Bodo Moeller]
4988
4989 *) Don't let DSA_generate_key() enter an infinite loop if the random
4990 number generation fails.
4991 [Bodo Moeller]
4992
4993 *) New 'rand' application for creating pseudo-random output.
4994 [Bodo Moeller]
4995
4996 *) Added configuration support for Linux/IA64
4997 [Rolf Haberrecker <rolf@suse.de>]
4998
4999 *) Assembler module support for Mingw32.
5000 [Ulf Möller]
5001
5002 *) Shared library support for HPUX (in shlib/).
5003 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5004
5005 *) Shared library support for Solaris gcc.
5006 [Lutz Behnke <behnke@trustcenter.de>]
5007
5008 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5009
5010 *) PKCS7_encrypt() was adding text MIME headers twice because they
5011 were added manually and by SMIME_crlf_copy().
5012 [Steve Henson]
5013
5014 *) In bntest.c don't call BN_rand with zero bits argument.
5015 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5016
5017 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5018 case was implemented. This caused BN_div_recp() to fail occasionally.
5019 [Ulf Möller]
5020
5021 *) Add an optional second argument to the set_label() in the perl
5022 assembly language builder. If this argument exists and is set
5023 to 1 it signals that the assembler should use a symbol whose
5024 scope is the entire file, not just the current function. This
5025 is needed with MASM which uses the format label:: for this scope.
5026 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5027
5028 *) Change the ASN1 types so they are typedefs by default. Before
5029 almost all types were #define'd to ASN1_STRING which was causing
5030 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5031 for example.
5032 [Steve Henson]
5033
5034 *) Change names of new functions to the new get1/get0 naming
5035 convention: After 'get1', the caller owns a reference count
5036 and has to call ..._free; 'get0' returns a pointer to some
5037 data structure without incrementing reference counters.
5038 (Some of the existing 'get' functions increment a reference
5039 counter, some don't.)
5040 Similarly, 'set1' and 'add1' functions increase reference
5041 counters or duplicate objects.
5042 [Steve Henson]
5043
5044 *) Allow for the possibility of temp RSA key generation failure:
5045 the code used to assume it always worked and crashed on failure.
5046 [Steve Henson]
5047
5048 *) Fix potential buffer overrun problem in BIO_printf().
5049 [Ulf Möller, using public domain code by Patrick Powell; problem
5050 pointed out by David Sacerdote <das33@cornell.edu>]
5051
5052 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5053 RAND_egd() and RAND_status(). In the command line application,
5054 the EGD socket can be specified like a seed file using RANDFILE
5055 or -rand.
5056 [Ulf Möller]
5057
5058 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5059 Some CAs (e.g. Verisign) distribute certificates in this form.
5060 [Steve Henson]
5061
5062 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5063 list to exclude them. This means that no special compilation option
5064 is needed to use anonymous DH: it just needs to be included in the
5065 cipher list.
5066 [Steve Henson]
5067
5068 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5069 EVP_MD_type. The old functionality is available in a new macro called
5070 EVP_MD_md(). Change code that uses it and update docs.
5071 [Steve Henson]
5072
5073 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5074 where the 'void *' argument is replaced by a function pointer argument.
5075 Previously 'void *' was abused to point to functions, which works on
5076 many platforms, but is not correct. As these functions are usually
5077 called by macros defined in OpenSSL header files, most source code
5078 should work without changes.
5079 [Richard Levitte]
5080
5081 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5082 sections with information on -D... compiler switches used for
5083 compiling the library so that applications can see them. To enable
5084 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5085 must be defined. E.g.,
5086 #define OPENSSL_ALGORITHM_DEFINES
5087 #include <openssl/opensslconf.h>
5088 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5089 [Richard Levitte, Ulf and Bodo Möller]
5090
5091 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5092 record layer.
5093 [Bodo Moeller]
5094
5095 *) Change the 'other' type in certificate aux info to a STACK_OF
5096 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5097 the required ASN1 format: arbitrary types determined by an OID.
5098 [Steve Henson]
5099
5100 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5101 argument to 'req'. This is not because the function is newer or
5102 better than others it just uses the work 'NEW' in the certificate
5103 request header lines. Some software needs this.
5104 [Steve Henson]
5105
5106 *) Reorganise password command line arguments: now passwords can be
5107 obtained from various sources. Delete the PEM_cb function and make
5108 it the default behaviour: i.e. if the callback is NULL and the
5109 usrdata argument is not NULL interpret it as a null terminated pass
5110 phrase. If usrdata and the callback are NULL then the pass phrase
5111 is prompted for as usual.
5112 [Steve Henson]
5113
5114 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5115 the support is automatically enabled. The resulting binaries will
5116 autodetect the card and use it if present.
5117 [Ben Laurie and Compaq Inc.]
5118
5119 *) Work around for Netscape hang bug. This sends certificate request
5120 and server done in one record. Since this is perfectly legal in the
5121 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5122 the bugs/SSLv3 entry for more info.
5123 [Steve Henson]
5124
5125 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5126 [Andy Polyakov]
5127
5128 *) Add -rand argument to smime and pkcs12 applications and read/write
5129 of seed file.
5130 [Steve Henson]
5131
5132 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5133 [Bodo Moeller]
5134
5135 *) Add command line password options to the remaining applications.
5136 [Steve Henson]
5137
5138 *) Bug fix for BN_div_recp() for numerators with an even number of
5139 bits.
5140 [Ulf Möller]
5141
5142 *) More tests in bntest.c, and changed test_bn output.
5143 [Ulf Möller]
5144
5145 *) ./config recognizes MacOS X now.
5146 [Andy Polyakov]
5147
5148 *) Bug fix for BN_div() when the first words of num and divsor are
5149 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5150 [Ulf Möller]
5151
5152 *) Add support for various broken PKCS#8 formats, and command line
5153 options to produce them.
5154 [Steve Henson]
5155
5156 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5157 get temporary BIGNUMs from a BN_CTX.
5158 [Ulf Möller]
5159
5160 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5161 for p == 0.
5162 [Ulf Möller]
5163
5164 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5165 include a #define from the old name to the new. The original intent
5166 was that statically linked binaries could for example just call
5167 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5168 link with digests. This never worked becayse SSLeay_add_all_digests()
5169 and SSLeay_add_all_ciphers() were in the same source file so calling
5170 one would link with the other. They are now in separate source files.
5171 [Steve Henson]
5172
5173 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5174 [Steve Henson]
5175
5176 *) Use a less unusual form of the Miller-Rabin primality test (it used
5177 a binary algorithm for exponentiation integrated into the Miller-Rabin
5178 loop, our standard modexp algorithms are faster).
5179 [Bodo Moeller]
5180
5181 *) Support for the EBCDIC character set completed.
5182 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5183
5184 *) Source code cleanups: use const where appropriate, eliminate casts,
5185 use void * instead of char * in lhash.
5186 [Ulf Möller]
5187
5188 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5189 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5190 this the server could overwrite ephemeral keys that the client
5191 has already seen).
5192 [Bodo Moeller]
5193
5194 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5195 using 50 iterations of the Rabin-Miller test.
5196
5197 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5198 iterations of the Rabin-Miller test as required by the appendix
5199 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5200 As BN_is_prime_fasttest includes trial division, DSA parameter
5201 generation becomes much faster.
5202
5203 This implies a change for the callback functions in DSA_is_prime
5204 and DSA_generate_parameters: The callback function is called once
5205 for each positive witness in the Rabin-Miller test, not just
5206 occasionally in the inner loop; and the parameters to the
5207 callback function now provide an iteration count for the outer
5208 loop rather than for the current invocation of the inner loop.
5209 DSA_generate_parameters additionally can call the callback
5210 function with an 'iteration count' of -1, meaning that a
5211 candidate has passed the trial division test (when q is generated
5212 from an application-provided seed, trial division is skipped).
5213 [Bodo Moeller]
5214
5215 *) New function BN_is_prime_fasttest that optionally does trial
5216 division before starting the Rabin-Miller test and has
5217 an additional BN_CTX * argument (whereas BN_is_prime always
5218 has to allocate at least one BN_CTX).
5219 'callback(1, -1, cb_arg)' is called when a number has passed the
5220 trial division stage.
5221 [Bodo Moeller]
5222
5223 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5224 as ASN1_TIME.
5225 [Steve Henson]
5226
5227 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5228 [Steve Henson]
5229
5230 *) New function BN_pseudo_rand().
5231 [Ulf Möller]
5232
5233 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5234 bignum version of BN_from_montgomery() with the working code from
5235 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5236 the comments.
5237 [Ulf Möller]
5238
5239 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5240 made it impossible to use the same SSL_SESSION data structure in
5241 SSL2 clients in multiple threads.
5242 [Bodo Moeller]
5243
5244 *) The return value of RAND_load_file() no longer counts bytes obtained
5245 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5246 to seed the PRNG (previously an explicit byte count was required).
5247 [Ulf Möller, Bodo Möller]
5248
5249 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5250 used (char *) instead of (void *) and had casts all over the place.
5251 [Steve Henson]
5252
5253 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5254 [Ulf Möller]
5255
5256 *) Retain source code compatibility for BN_prime_checks macro:
5257 BN_is_prime(..., BN_prime_checks, ...) now uses
5258 BN_prime_checks_for_size to determine the appropriate number of
5259 Rabin-Miller iterations.
5260 [Ulf Möller]
5261
5262 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5263 DH_CHECK_P_NOT_SAFE_PRIME.
5264 (Check if this is true? OpenPGP calls them "strong".)
5265 [Ulf Möller]
5266
5267 *) Merge the functionality of "dh" and "gendh" programs into a new program
5268 "dhparam". The old programs are retained for now but will handle DH keys
5269 (instead of parameters) in future.
5270 [Steve Henson]
5271
5272 *) Make the ciphers, s_server and s_client programs check the return values
5273 when a new cipher list is set.
5274 [Steve Henson]
5275
5276 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5277 ciphers. Before when the 56bit ciphers were enabled the sorting was
5278 wrong.
5279
5280 The syntax for the cipher sorting has been extended to support sorting by
5281 cipher-strength (using the strength_bits hard coded in the tables).
5282 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5283
5284 Fix a bug in the cipher-command parser: when supplying a cipher command
5285 string with an "undefined" symbol (neither command nor alphanumeric
5286 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5287 an error is flagged.
5288
5289 Due to the strength-sorting extension, the code of the
5290 ssl_create_cipher_list() function was completely rearranged. I hope that
5291 the readability was also increased :-)
5292 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5293
5294 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5295 for the first serial number and places 2 in the serial number file. This
5296 avoids problems when the root CA is created with serial number zero and
5297 the first user certificate has the same issuer name and serial number
5298 as the root CA.
5299 [Steve Henson]
5300
5301 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5302 the new code. Add documentation for this stuff.
5303 [Steve Henson]
5304
5305 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5306 X509_*() to X509at_*() on the grounds that they don't handle X509
5307 structures and behave in an analagous way to the X509v3 functions:
5308 they shouldn't be called directly but wrapper functions should be used
5309 instead.
5310
5311 So we also now have some wrapper functions that call the X509at functions
5312 when passed certificate requests. (TO DO: similar things can be done with
5313 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5314 things. Some of these need some d2i or i2d and print functionality
5315 because they handle more complex structures.)
5316 [Steve Henson]
5317
5318 *) Add missing #ifndefs that caused missing symbols when building libssl
5319 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5320 NO_RSA in ssl/s2*.c.
5321 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5322
5323 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5324 has a return value which indicates the quality of the random data
5325 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5326 error queue. New function RAND_pseudo_bytes() generates output that is
5327 guaranteed to be unique but not unpredictable. RAND_add is like
5328 RAND_seed, but takes an extra argument for an entropy estimate
5329 (RAND_seed always assumes full entropy).
5330 [Ulf Möller]
5331
5332 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5333 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5334 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5335 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5336 false-positive rate of at most 2^-80 for random input.
5337 [Bodo Moeller]
5338
5339 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5340 [Bodo Moeller]
5341
5342 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5343 in the 0.9.5 release), this returns the chain
5344 from an X509_CTX structure with a dup of the stack and all
5345 the X509 reference counts upped: so the stack will exist
5346 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5347 to use this.
5348
5349 Also make SSL_SESSION_print() print out the verify return
5350 code.
5351 [Steve Henson]
5352
5353 *) Add manpage for the pkcs12 command. Also change the default
5354 behaviour so MAC iteration counts are used unless the new
5355 -nomaciter option is used. This improves file security and
5356 only older versions of MSIE (4.0 for example) need it.
5357 [Steve Henson]
5358
5359 *) Honor the no-xxx Configure options when creating .DEF files.
5360 [Ulf Möller]
5361
5362 *) Add PKCS#10 attributes to field table: challengePassword,
5363 unstructuredName and unstructuredAddress. These are taken from
5364 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5365 international characters are used.
5366
5367 More changes to X509_ATTRIBUTE code: allow the setting of types
5368 based on strings. Remove the 'loc' parameter when adding
5369 attributes because these will be a SET OF encoding which is sorted
5370 in ASN1 order.
5371 [Steve Henson]
5372
5373 *) Initial changes to the 'req' utility to allow request generation
5374 automation. This will allow an application to just generate a template
5375 file containing all the field values and have req construct the
5376 request.
5377
5378 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5379 used all over the place including certificate requests and PKCS#7
5380 structures. They are currently handled manually where necessary with
5381 some primitive wrappers for PKCS#7. The new functions behave in a
5382 manner analogous to the X509 extension functions: they allow
5383 attributes to be looked up by NID and added.
5384
5385 Later something similar to the X509V3 code would be desirable to
5386 automatically handle the encoding, decoding and printing of the
5387 more complex types. The string types like challengePassword can
5388 be handled by the string table functions.
5389
5390 Also modified the multi byte string table handling. Now there is
5391 a 'global mask' which masks out certain types. The table itself
5392 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5393 is useful when for example there is only one permissible type
5394 (as in countryName) and using the mask might result in no valid
5395 types at all.
5396 [Steve Henson]
5397
5398 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5399 SSL_get_peer_finished to allow applications to obtain the latest
5400 Finished messages sent to the peer or expected from the peer,
5401 respectively. (SSL_get_peer_finished is usually the Finished message
5402 actually received from the peer, otherwise the protocol will be aborted.)
5403
5404 As the Finished message are message digests of the complete handshake
5405 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5406 be used for external authentication procedures when the authentication
5407 provided by SSL/TLS is not desired or is not enough.
5408 [Bodo Moeller]
5409
5410 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5411 the host supports BWX extension and if Compaq C is present on the
5412 $PATH. Just exploiting of the BWX extension results in 20-30%
5413 performance kick for some algorithms, e.g. DES and RC4 to mention
5414 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5415 SHA1.
5416 [Andy Polyakov]
5417
5418 *) Add support for MS "fast SGC". This is arguably a violation of the
5419 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5420 weak crypto and after checking the certificate is SGC a second one
5421 with strong crypto. MS SGC stops the first handshake after receiving
5422 the server certificate message and sends a second client hello. Since
5423 a server will typically do all the time consuming operations before
5424 expecting any further messages from the client (server key exchange
5425 is the most expensive) there is little difference between the two.
5426
5427 To get OpenSSL to support MS SGC we have to permit a second client
5428 hello message after we have sent server done. In addition we have to
5429 reset the MAC if we do get this second client hello.
5430 [Steve Henson]
5431
5432 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5433 if a DER encoded private key is RSA or DSA traditional format. Changed
5434 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5435 format DER encoded private key. Newer code should use PKCS#8 format which
5436 has the key type encoded in the ASN1 structure. Added DER private key
5437 support to pkcs8 application.
5438 [Steve Henson]
5439
5440 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5441 ciphersuites has been selected (as required by the SSL 3/TLS 1
5442 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5443 is set, we interpret this as a request to violate the specification
5444 (the worst that can happen is a handshake failure, and 'correct'
5445 behaviour would result in a handshake failure anyway).
5446 [Bodo Moeller]
5447
5448 *) In SSL_CTX_add_session, take into account that there might be multiple
5449 SSL_SESSION structures with the same session ID (e.g. when two threads
5450 concurrently obtain them from an external cache).
5451 The internal cache can handle only one SSL_SESSION with a given ID,
5452 so if there's a conflict, we now throw out the old one to achieve
5453 consistency.
5454 [Bodo Moeller]
5455
5456 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5457 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5458 some routines that use cipher OIDs: some ciphers do not have OIDs
5459 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5460 example.
5461 [Steve Henson]
5462
5463 *) Simplify the trust setting structure and code. Now we just have
5464 two sequences of OIDs for trusted and rejected settings. These will
5465 typically have values the same as the extended key usage extension
5466 and any application specific purposes.
5467
5468 The trust checking code now has a default behaviour: it will just
5469 check for an object with the same NID as the passed id. Functions can
5470 be provided to override either the default behaviour or the behaviour
5471 for a given id. SSL client, server and email already have functions
5472 in place for compatibility: they check the NID and also return "trusted"
5473 if the certificate is self signed.
5474 [Steve Henson]
5475
5476 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5477 traditional format into an EVP_PKEY structure.
5478 [Steve Henson]
5479
5480 *) Add a password callback function PEM_cb() which either prompts for
5481 a password if usr_data is NULL or otherwise assumes it is a null
5482 terminated password. Allow passwords to be passed on command line
5483 environment or config files in a few more utilities.
5484 [Steve Henson]
5485
5486 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5487 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5488 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5489 Update documentation.
5490 [Steve Henson]
5491
5492 *) Support for ASN1 "NULL" type. This could be handled before by using
5493 ASN1_TYPE but there wasn't any function that would try to read a NULL
5494 and produce an error if it couldn't. For compatibility we also have
5495 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5496 don't allocate anything because they don't need to.
5497 [Steve Henson]
5498
5499 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5500 for details.
5501 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5502
5503 *) Rebuild of the memory allocation routines used by OpenSSL code and
5504 possibly others as well. The purpose is to make an interface that
5505 provide hooks so anyone can build a separate set of allocation and
5506 deallocation routines to be used by OpenSSL, for example memory
5507 pool implementations, or something else, which was previously hard
5508 since Malloc(), Realloc() and Free() were defined as macros having
5509 the values malloc, realloc and free, respectively (except for Win32
5510 compilations). The same is provided for memory debugging code.
5511 OpenSSL already comes with functionality to find memory leaks, but
5512 this gives people a chance to debug other memory problems.
5513
5514 With these changes, a new set of functions and macros have appeared:
5515
5516 CRYPTO_set_mem_debug_functions() [F]
5517 CRYPTO_get_mem_debug_functions() [F]
5518 CRYPTO_dbg_set_options() [F]
5519 CRYPTO_dbg_get_options() [F]
5520 CRYPTO_malloc_debug_init() [M]
5521
5522 The memory debug functions are NULL by default, unless the library
5523 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5524 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5525 gives the standard debugging functions that come with OpenSSL) or
5526 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5527 provided by the library user) must be used. When the standard
5528 debugging functions are used, CRYPTO_dbg_set_options can be used to
5529 request additional information:
5530 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5531 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5532
5533 Also, things like CRYPTO_set_mem_functions will always give the
5534 expected result (the new set of functions is used for allocation
5535 and deallocation) at all times, regardless of platform and compiler
5536 options.
5537
5538 To finish it up, some functions that were never use in any other
5539 way than through macros have a new API and new semantic:
5540
5541 CRYPTO_dbg_malloc()
5542 CRYPTO_dbg_realloc()
5543 CRYPTO_dbg_free()
5544
5545 All macros of value have retained their old syntax.
5546 [Richard Levitte and Bodo Moeller]
5547
5548 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5549 ordering of SMIMECapabilities wasn't in "strength order" and there
5550 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5551 algorithm.
5552 [Steve Henson]
5553
5554 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5555 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5556 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5557
5558 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5559 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5560 functionality to handle multipart/signed properly) and a utility
5561 called 'smime' to call all this stuff. This is based on code I
5562 originally wrote for Celo who have kindly allowed it to be
5563 included in OpenSSL.
5564 [Steve Henson]
5565
5566 *) Add variants des_set_key_checked and des_set_key_unchecked of
5567 des_set_key (aka des_key_sched). Global variable des_check_key
5568 decides which of these is called by des_set_key; this way
5569 des_check_key behaves as it always did, but applications and
5570 the library itself, which was buggy for des_check_key == 1,
5571 have a cleaner way to pick the version they need.
5572 [Bodo Moeller]
5573
5574 *) New function PKCS12_newpass() which changes the password of a
5575 PKCS12 structure.
5576 [Steve Henson]
5577
5578 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5579 dynamic mix. In both cases the ids can be used as an index into the
5580 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5581 functions so they accept a list of the field values and the
5582 application doesn't need to directly manipulate the X509_TRUST
5583 structure.
5584 [Steve Henson]
5585
5586 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5587 need initialising.
5588 [Steve Henson]
5589
5590 *) Modify the way the V3 extension code looks up extensions. This now
5591 works in a similar way to the object code: we have some "standard"
5592 extensions in a static table which is searched with OBJ_bsearch()
5593 and the application can add dynamic ones if needed. The file
5594 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5595 updated whenever a new extension is added to the core code and kept
5596 in ext_nid order. There is a simple program 'tabtest.c' which checks
5597 this. New extensions are not added too often so this file can readily
5598 be maintained manually.
5599
5600 There are two big advantages in doing things this way. The extensions
5601 can be looked up immediately and no longer need to be "added" using
5602 X509V3_add_standard_extensions(): this function now does nothing.
5603 [Side note: I get *lots* of email saying the extension code doesn't
5604 work because people forget to call this function]
5605 Also no dynamic allocation is done unless new extensions are added:
5606 so if we don't add custom extensions there is no need to call
5607 X509V3_EXT_cleanup().
5608 [Steve Henson]
5609
5610 *) Modify enc utility's salting as follows: make salting the default. Add a
5611 magic header, so unsalted files fail gracefully instead of just decrypting
5612 to garbage. This is because not salting is a big security hole, so people
5613 should be discouraged from doing it.
5614 [Ben Laurie]
5615
5616 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5617 digest to be passed on the command line but it only used this
5618 parameter when signing a certificate. Modified so all relevant
5619 operations are affected by the digest parameter including the
5620 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5621 DSA key was used because it didn't fix the digest.
5622 [Steve Henson]
5623
5624 *) Initial certificate chain verify code. Currently tests the untrusted
5625 certificates for consistency with the verify purpose (which is set
5626 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5627
5628 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5629 this is because it will reject chains with invalid extensions whereas
5630 every previous version of OpenSSL and SSLeay made no checks at all.
5631
5632 Trust code: checks the root CA for the relevant trust settings. Trust
5633 settings have an initial value consistent with the verify purpose: e.g.
5634 if the verify purpose is for SSL client use it expects the CA to be
5635 trusted for SSL client use. However the default value can be changed to
5636 permit custom trust settings: one example of this would be to only trust
5637 certificates from a specific "secure" set of CAs.
5638
5639 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5640 which should be used for version portability: especially since the
5641 verify structure is likely to change more often now.
5642
5643 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5644 to set them. If not set then assume SSL clients will verify SSL servers
5645 and vice versa.
5646
5647 Two new options to the verify program: -untrusted allows a set of
5648 untrusted certificates to be passed in and -purpose which sets the
5649 intended purpose of the certificate. If a purpose is set then the
5650 new chain verify code is used to check extension consistency.
5651 [Steve Henson]
5652
5653 *) Support for the authority information access extension.
5654 [Steve Henson]
5655
5656 *) Modify RSA and DSA PEM read routines to transparently handle
5657 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5658 public keys in a format compatible with certificate
5659 SubjectPublicKeyInfo structures. Unfortunately there were already
5660 functions called *_PublicKey_* which used various odd formats so
5661 these are retained for compatibility: however the DSA variants were
5662 never in a public release so they have been deleted. Changed dsa/rsa
5663 utilities to handle the new format: note no releases ever handled public
5664 keys so we should be OK.
5665
5666 The primary motivation for this change is to avoid the same fiasco
5667 that dogs private keys: there are several incompatible private key
5668 formats some of which are standard and some OpenSSL specific and
5669 require various evil hacks to allow partial transparent handling and
5670 even then it doesn't work with DER formats. Given the option anything
5671 other than PKCS#8 should be dumped: but the other formats have to
5672 stay in the name of compatibility.
5673
5674 With public keys and the benefit of hindsight one standard format
5675 is used which works with EVP_PKEY, RSA or DSA structures: though
5676 it clearly returns an error if you try to read the wrong kind of key.
5677
5678 Added a -pubkey option to the 'x509' utility to output the public key.
5679 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5680 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5681 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5682 that do the same as the EVP_PKEY_assign_*() except they up the
5683 reference count of the added key (they don't "swallow" the
5684 supplied key).
5685 [Steve Henson]
5686
5687 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5688 CRLs would fail if the file contained no certificates or no CRLs:
5689 added a new function to read in both types and return the number
5690 read: this means that if none are read it will be an error. The
5691 DER versions of the certificate and CRL reader would always fail
5692 because it isn't possible to mix certificates and CRLs in DER format
5693 without choking one or the other routine. Changed this to just read
5694 a certificate: this is the best we can do. Also modified the code
5695 in apps/verify.c to take notice of return codes: it was previously
5696 attempting to read in certificates from NULL pointers and ignoring
5697 any errors: this is one reason why the cert and CRL reader seemed
5698 to work. It doesn't check return codes from the default certificate
5699 routines: these may well fail if the certificates aren't installed.
5700 [Steve Henson]
5701
5702 *) Code to support otherName option in GeneralName.
5703 [Steve Henson]
5704
5705 *) First update to verify code. Change the verify utility
5706 so it warns if it is passed a self signed certificate:
5707 for consistency with the normal behaviour. X509_verify
5708 has been modified to it will now verify a self signed
5709 certificate if *exactly* the same certificate appears
5710 in the store: it was previously impossible to trust a
5711 single self signed certificate. This means that:
5712 openssl verify ss.pem
5713 now gives a warning about a self signed certificate but
5714 openssl verify -CAfile ss.pem ss.pem
5715 is OK.
5716 [Steve Henson]
5717
5718 *) For servers, store verify_result in SSL_SESSION data structure
5719 (and add it to external session representation).
5720 This is needed when client certificate verifications fails,
5721 but an application-provided verification callback (set by
5722 SSL_CTX_set_cert_verify_callback) allows accepting the session
5723 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5724 but returns 1): When the session is reused, we have to set
5725 ssl->verify_result to the appropriate error code to avoid
5726 security holes.
5727 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5728
5729 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5730 case in PKCS7_dataInit() where the signed PKCS7 structure
5731 didn't contain any existing data because it was being created.
5732 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5733
5734 *) Add a salt to the key derivation routines in enc.c. This
5735 forms the first 8 bytes of the encrypted file. Also add a
5736 -S option to allow a salt to be input on the command line.
5737 [Steve Henson]
5738
5739 *) New function X509_cmp(). Oddly enough there wasn't a function
5740 to compare two certificates. We do this by working out the SHA1
5741 hash and comparing that. X509_cmp() will be needed by the trust
5742 code.
5743 [Steve Henson]
5744
5745 *) SSL_get1_session() is like SSL_get_session(), but increments
5746 the reference count in the SSL_SESSION returned.
5747 [Geoff Thorpe <geoff@eu.c2.net>]
5748
5749 *) Fix for 'req': it was adding a null to request attributes.
5750 Also change the X509_LOOKUP and X509_INFO code to handle
5751 certificate auxiliary information.
5752 [Steve Henson]
5753
5754 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5755 the 'enc' command.
5756 [Steve Henson]
5757
5758 *) Add the possibility to add extra information to the memory leak
5759 detecting output, to form tracebacks, showing from where each
5760 allocation was originated: CRYPTO_push_info("constant string") adds
5761 the string plus current file name and line number to a per-thread
5762 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5763 is like calling CYRPTO_pop_info() until the stack is empty.
5764 Also updated memory leak detection code to be multi-thread-safe.
5765 [Richard Levitte]
5766
5767 *) Add options -text and -noout to pkcs7 utility and delete the
5768 encryption options which never did anything. Update docs.
5769 [Steve Henson]
5770
5771 *) Add options to some of the utilities to allow the pass phrase
5772 to be included on either the command line (not recommended on
5773 OSes like Unix) or read from the environment. Update the
5774 manpages and fix a few bugs.
5775 [Steve Henson]
5776
5777 *) Add a few manpages for some of the openssl commands.
5778 [Steve Henson]
5779
5780 *) Fix the -revoke option in ca. It was freeing up memory twice,
5781 leaking and not finding already revoked certificates.
5782 [Steve Henson]
5783
5784 *) Extensive changes to support certificate auxiliary information.
5785 This involves the use of X509_CERT_AUX structure and X509_AUX
5786 functions. An X509_AUX function such as PEM_read_X509_AUX()
5787 can still read in a certificate file in the usual way but it
5788 will also read in any additional "auxiliary information". By
5789 doing things this way a fair degree of compatibility can be
5790 retained: existing certificates can have this information added
5791 using the new 'x509' options.
5792
5793 Current auxiliary information includes an "alias" and some trust
5794 settings. The trust settings will ultimately be used in enhanced
5795 certificate chain verification routines: currently a certificate
5796 can only be trusted if it is self signed and then it is trusted
5797 for all purposes.
5798 [Steve Henson]
5799
5800 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5801 The problem was that one of the replacement routines had not been working
5802 since SSLeay releases. For now the offending routine has been replaced
5803 with non-optimised assembler. Even so, this now gives around 95%
5804 performance improvement for 1024 bit RSA signs.
5805 [Mark Cox]
5806
5807 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5808 handling. Most clients have the effective key size in bits equal to
5809 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5810 A few however don't do this and instead use the size of the decrypted key
5811 to determine the RC2 key length and the AlgorithmIdentifier to determine
5812 the effective key length. In this case the effective key length can still
5813 be 40 bits but the key length can be 168 bits for example. This is fixed
5814 by manually forcing an RC2 key into the EVP_PKEY structure because the
5815 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5816 the key length and effective key length are equal.
5817 [Steve Henson]
5818
5819 *) Add a bunch of functions that should simplify the creation of
5820 X509_NAME structures. Now you should be able to do:
5821 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5822 and have it automatically work out the correct field type and fill in
5823 the structures. The more adventurous can try:
5824 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5825 and it will (hopefully) work out the correct multibyte encoding.
5826 [Steve Henson]
5827
5828 *) Change the 'req' utility to use the new field handling and multibyte
5829 copy routines. Before the DN field creation was handled in an ad hoc
5830 way in req, ca, and x509 which was rather broken and didn't support
5831 BMPStrings or UTF8Strings. Since some software doesn't implement
5832 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5833 using the dirstring_type option. See the new comment in the default
5834 openssl.cnf for more info.
5835 [Steve Henson]
5836
5837 *) Make crypto/rand/md_rand.c more robust:
5838 - Assure unique random numbers after fork().
5839 - Make sure that concurrent threads access the global counter and
5840 md serializably so that we never lose entropy in them
5841 or use exactly the same state in multiple threads.
5842 Access to the large state is not always serializable because
5843 the additional locking could be a performance killer, and
5844 md should be large enough anyway.
5845 [Bodo Moeller]
5846
5847 *) New file apps/app_rand.c with commonly needed functionality
5848 for handling the random seed file.
5849
5850 Use the random seed file in some applications that previously did not:
5851 ca,
5852 dsaparam -genkey (which also ignored its '-rand' option),
5853 s_client,
5854 s_server,
5855 x509 (when signing).
5856 Except on systems with /dev/urandom, it is crucial to have a random
5857 seed file at least for key creation, DSA signing, and for DH exchanges;
5858 for RSA signatures we could do without one.
5859
5860 gendh and gendsa (unlike genrsa) used to read only the first byte
5861 of each file listed in the '-rand' option. The function as previously
5862 found in genrsa is now in app_rand.c and is used by all programs
5863 that support '-rand'.
5864 [Bodo Moeller]
5865
5866 *) In RAND_write_file, use mode 0600 for creating files;
5867 don't just chmod when it may be too late.
5868 [Bodo Moeller]
5869
5870 *) Report an error from X509_STORE_load_locations
5871 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5872 [Bill Perry]
5873
5874 *) New function ASN1_mbstring_copy() this copies a string in either
5875 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5876 into an ASN1_STRING type. A mask of permissible types is passed
5877 and it chooses the "minimal" type to use or an error if not type
5878 is suitable.
5879 [Steve Henson]
5880
5881 *) Add function equivalents to the various macros in asn1.h. The old
5882 macros are retained with an M_ prefix. Code inside the library can
5883 use the M_ macros. External code (including the openssl utility)
5884 should *NOT* in order to be "shared library friendly".
5885 [Steve Henson]
5886
5887 *) Add various functions that can check a certificate's extensions
5888 to see if it usable for various purposes such as SSL client,
5889 server or S/MIME and CAs of these types. This is currently
5890 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5891 verification. Also added a -purpose flag to x509 utility to
5892 print out all the purposes.
5893 [Steve Henson]
5894
5895 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5896 functions.
5897 [Steve Henson]
5898
5899 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5900 for, obtain and decode and extension and obtain its critical flag.
5901 This allows all the necessary extension code to be handled in a
5902 single function call.
5903 [Steve Henson]
5904
5905 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5906 platforms. See crypto/rc4/rc4_enc.c for further details.
5907 [Andy Polyakov]
5908
5909 *) New -noout option to asn1parse. This causes no output to be produced
5910 its main use is when combined with -strparse and -out to extract data
5911 from a file (which may not be in ASN.1 format).
5912 [Steve Henson]
5913
5914 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5915 when producing the local key id.
5916 [Richard Levitte <levitte@stacken.kth.se>]
5917
5918 *) New option -dhparam in s_server. This allows a DH parameter file to be
5919 stated explicitly. If it is not stated then it tries the first server
5920 certificate file. The previous behaviour hard coded the filename
5921 "server.pem".
5922 [Steve Henson]
5923
5924 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5925 a public key to be input or output. For example:
5926 openssl rsa -in key.pem -pubout -out pubkey.pem
5927 Also added necessary DSA public key functions to handle this.
5928 [Steve Henson]
5929
5930 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5931 in the message. This was handled by allowing
5932 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5933 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5934
5935 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5936 to the end of the strings whereas this didn't. This would cause problems
5937 if strings read with d2i_ASN1_bytes() were later modified.
5938 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5939
5940 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5941 data and it contains EOF it will end up returning an error. This is
5942 caused by input 46 bytes long. The cause is due to the way base64
5943 BIOs find the start of base64 encoded data. They do this by trying a
5944 trial decode on each line until they find one that works. When they
5945 do a flag is set and it starts again knowing it can pass all the
5946 data directly through the decoder. Unfortunately it doesn't reset
5947 the context it uses. This means that if EOF is reached an attempt
5948 is made to pass two EOFs through the context and this causes the
5949 resulting error. This can also cause other problems as well. As is
5950 usual with these problems it takes *ages* to find and the fix is
5951 trivial: move one line.
5952 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
5953
5954 *) Ugly workaround to get s_client and s_server working under Windows. The
5955 old code wouldn't work because it needed to select() on sockets and the
5956 tty (for keypresses and to see if data could be written). Win32 only
5957 supports select() on sockets so we select() with a 1s timeout on the
5958 sockets and then see if any characters are waiting to be read, if none
5959 are present then we retry, we also assume we can always write data to
5960 the tty. This isn't nice because the code then blocks until we've
5961 received a complete line of data and it is effectively polling the
5962 keyboard at 1s intervals: however it's quite a bit better than not
5963 working at all :-) A dedicated Windows application might handle this
5964 with an event loop for example.
5965 [Steve Henson]
5966
5967 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
5968 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
5969 will be called when RSA_sign() and RSA_verify() are used. This is useful
5970 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
5971 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
5972 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
5973 This necessitated the support of an extra signature type NID_md5_sha1
5974 for SSL signatures and modifications to the SSL library to use it instead
5975 of calling RSA_public_decrypt() and RSA_private_encrypt().
5976 [Steve Henson]
5977
5978 *) Add new -verify -CAfile and -CApath options to the crl program, these
5979 will lookup a CRL issuers certificate and verify the signature in a
5980 similar way to the verify program. Tidy up the crl program so it
5981 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
5982 less strict. It will now permit CRL extensions even if it is not
5983 a V2 CRL: this will allow it to tolerate some broken CRLs.
5984 [Steve Henson]
5985
5986 *) Initialize all non-automatic variables each time one of the openssl
5987 sub-programs is started (this is necessary as they may be started
5988 multiple times from the "OpenSSL>" prompt).
5989 [Lennart Bang, Bodo Moeller]
5990
5991 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
5992 removing all other RSA functionality (this is what NO_RSA does). This
5993 is so (for example) those in the US can disable those operations covered
5994 by the RSA patent while allowing storage and parsing of RSA keys and RSA
5995 key generation.
5996 [Steve Henson]
5997
5998 *) Non-copying interface to BIO pairs.
5999 (still largely untested)
6000 [Bodo Moeller]
6001
6002 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6003 ASCII string. This was handled independently in various places before.
6004 [Steve Henson]
6005
6006 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6007 UTF8 strings a character at a time.
6008 [Steve Henson]
6009
6010 *) Use client_version from client hello to select the protocol
6011 (s23_srvr.c) and for RSA client key exchange verification
6012 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6013 [Bodo Moeller]
6014
6015 *) Add various utility functions to handle SPKACs, these were previously
6016 handled by poking round in the structure internals. Added new function
6017 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6018 print, verify and generate SPKACs. Based on an original idea from
6019 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6020 [Steve Henson]
6021
6022 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6023 [Andy Polyakov]
6024
6025 *) Allow the config file extension section to be overwritten on the
6026 command line. Based on an original idea from Massimiliano Pala
6027 <madwolf@comune.modena.it>. The new option is called -extensions
6028 and can be applied to ca, req and x509. Also -reqexts to override
6029 the request extensions in req and -crlexts to override the crl extensions
6030 in ca.
6031 [Steve Henson]
6032
6033 *) Add new feature to the SPKAC handling in ca. Now you can include
6034 the same field multiple times by preceding it by "XXXX." for example:
6035 1.OU="Unit name 1"
6036 2.OU="Unit name 2"
6037 this is the same syntax as used in the req config file.
6038 [Steve Henson]
6039
6040 *) Allow certificate extensions to be added to certificate requests. These
6041 are specified in a 'req_extensions' option of the req section of the
6042 config file. They can be printed out with the -text option to req but
6043 are otherwise ignored at present.
6044 [Steve Henson]
6045
6046 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6047 data read consists of only the final block it would not decrypted because
6048 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6049 A misplaced 'break' also meant the decrypted final block might not be
6050 copied until the next read.
6051 [Steve Henson]
6052
6053 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6054 a few extra parameters to the DH structure: these will be useful if
6055 for example we want the value of 'q' or implement X9.42 DH.
6056 [Steve Henson]
6057
6058 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6059 provides hooks that allow the default DSA functions or functions on a
6060 "per key" basis to be replaced. This allows hardware acceleration and
6061 hardware key storage to be handled without major modification to the
6062 library. Also added low level modexp hooks and CRYPTO_EX structure and
6063 associated functions.
6064 [Steve Henson]
6065
6066 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6067 as "read only": it can't be written to and the buffer it points to will
6068 not be freed. Reading from a read only BIO is much more efficient than
6069 a normal memory BIO. This was added because there are several times when
6070 an area of memory needs to be read from a BIO. The previous method was
6071 to create a memory BIO and write the data to it, this results in two
6072 copies of the data and an O(n^2) reading algorithm. There is a new
6073 function BIO_new_mem_buf() which creates a read only memory BIO from
6074 an area of memory. Also modified the PKCS#7 routines to use read only
6075 memory BIOs.
6076 [Steve Henson]
6077
6078 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6079 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6080 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6081 but a retry condition occured while trying to read the rest.
6082 [Bodo Moeller]
6083
6084 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6085 NID_pkcs7_encrypted by default: this was wrong since this should almost
6086 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6087 the encrypted data type: this is a more sensible place to put it and it
6088 allows the PKCS#12 code to be tidied up that duplicated this
6089 functionality.
6090 [Steve Henson]
6091
6092 *) Changed obj_dat.pl script so it takes its input and output files on
6093 the command line. This should avoid shell escape redirection problems
6094 under Win32.
6095 [Steve Henson]
6096
6097 *) Initial support for certificate extension requests, these are included
6098 in things like Xenroll certificate requests. Included functions to allow
6099 extensions to be obtained and added.
6100 [Steve Henson]
6101
6102 *) -crlf option to s_client and s_server for sending newlines as
6103 CRLF (as required by many protocols).
6104 [Bodo Moeller]
6105
6106 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6107
6108 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6109 [Ralf S. Engelschall]
6110
6111 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6112 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6113
6114 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6115 program.
6116 [Steve Henson]
6117
6118 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6119 DH parameters/keys (q is lost during that conversion, but the resulting
6120 DH parameters contain its length).
6121
6122 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6123 much faster than DH_generate_parameters (which creates parameters
6124 where p = 2*q + 1), and also the smaller q makes DH computations
6125 much more efficient (160-bit exponentiation instead of 1024-bit
6126 exponentiation); so this provides a convenient way to support DHE
6127 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6128 utter importance to use
6129 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6130 or
6131 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6132 when such DH parameters are used, because otherwise small subgroup
6133 attacks may become possible!
6134 [Bodo Moeller]
6135
6136 *) Avoid memory leak in i2d_DHparams.
6137 [Bodo Moeller]
6138
6139 *) Allow the -k option to be used more than once in the enc program:
6140 this allows the same encrypted message to be read by multiple recipients.
6141 [Steve Henson]
6142
6143 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6144 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6145 it will always use the numerical form of the OID, even if it has a short
6146 or long name.
6147 [Steve Henson]
6148
6149 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6150 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6151 otherwise bn_mod_exp was called. In the case of hardware keys for example
6152 no private key components need be present and it might store extra data
6153 in the RSA structure, which cannot be accessed from bn_mod_exp.
6154 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6155 private key operations.
6156 [Steve Henson]
6157
6158 *) Added support for SPARC Linux.
6159 [Andy Polyakov]
6160
6161 *) pem_password_cb function type incompatibly changed from
6162 typedef int pem_password_cb(char *buf, int size, int rwflag);
6163 to
6164 ....(char *buf, int size, int rwflag, void *userdata);
6165 so that applications can pass data to their callbacks:
6166 The PEM[_ASN1]_{read,write}... functions and macros now take an
6167 additional void * argument, which is just handed through whenever
6168 the password callback is called.
6169 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6170
6171 New function SSL_CTX_set_default_passwd_cb_userdata.
6172
6173 Compatibility note: As many C implementations push function arguments
6174 onto the stack in reverse order, the new library version is likely to
6175 interoperate with programs that have been compiled with the old
6176 pem_password_cb definition (PEM_whatever takes some data that
6177 happens to be on the stack as its last argument, and the callback
6178 just ignores this garbage); but there is no guarantee whatsoever that
6179 this will work.
6180
6181 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6182 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6183 problems not only on Windows, but also on some Unix platforms.
6184 To avoid problematic command lines, these definitions are now in an
6185 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6186 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6187 [Bodo Moeller]
6188
6189 *) MIPS III/IV assembler module is reimplemented.
6190 [Andy Polyakov]
6191
6192 *) More DES library cleanups: remove references to srand/rand and
6193 delete an unused file.
6194 [Ulf Möller]
6195
6196 *) Add support for the the free Netwide assembler (NASM) under Win32,
6197 since not many people have MASM (ml) and it can be hard to obtain.
6198 This is currently experimental but it seems to work OK and pass all
6199 the tests. Check out INSTALL.W32 for info.
6200 [Steve Henson]
6201
6202 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6203 without temporary keys kept an extra copy of the server key,
6204 and connections with temporary keys did not free everything in case
6205 of an error.
6206 [Bodo Moeller]
6207
6208 *) New function RSA_check_key and new openssl rsa option -check
6209 for verifying the consistency of RSA keys.
6210 [Ulf Moeller, Bodo Moeller]
6211
6212 *) Various changes to make Win32 compile work:
6213 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6214 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6215 comparison" warnings.
6216 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6217 [Steve Henson]
6218
6219 *) Add a debugging option to PKCS#5 v2 key generation function: when
6220 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6221 derived keys are printed to stderr.
6222 [Steve Henson]
6223
6224 *) Copy the flags in ASN1_STRING_dup().
6225 [Roman E. Pavlov <pre@mo.msk.ru>]
6226
6227 *) The x509 application mishandled signing requests containing DSA
6228 keys when the signing key was also DSA and the parameters didn't match.
6229
6230 It was supposed to omit the parameters when they matched the signing key:
6231 the verifying software was then supposed to automatically use the CA's
6232 parameters if they were absent from the end user certificate.
6233
6234 Omitting parameters is no longer recommended. The test was also
6235 the wrong way round! This was probably due to unusual behaviour in
6236 EVP_cmp_parameters() which returns 1 if the parameters match.
6237 This meant that parameters were omitted when they *didn't* match and
6238 the certificate was useless. Certificates signed with 'ca' didn't have
6239 this bug.
6240 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6241
6242 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6243 The interface is as follows:
6244 Applications can use
6245 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6246 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6247 "off" is now the default.
6248 The library internally uses
6249 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6250 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6251 to disable memory-checking temporarily.
6252
6253 Some inconsistent states that previously were possible (and were
6254 even the default) are now avoided.
6255
6256 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6257 with each memory chunk allocated; this is occasionally more helpful
6258 than just having a counter.
6259
6260 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6261
6262 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6263 extensions.
6264 [Bodo Moeller]
6265
6266 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6267 which largely parallels "options", but is for changing API behaviour,
6268 whereas "options" are about protocol behaviour.
6269 Initial "mode" flags are:
6270
6271 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6272 a single record has been written.
6273 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6274 retries use the same buffer location.
6275 (But all of the contents must be
6276 copied!)
6277 [Bodo Moeller]
6278
6279 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6280 worked.
6281
6282 *) Fix problems with no-hmac etc.
6283 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6284
6285 *) New functions RSA_get_default_method(), RSA_set_method() and
6286 RSA_get_method(). These allows replacement of RSA_METHODs without having
6287 to mess around with the internals of an RSA structure.
6288 [Steve Henson]
6289
6290 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6291 Also really enable memory leak checks in openssl.c and in some
6292 test programs.
6293 [Chad C. Mulligan, Bodo Moeller]
6294
6295 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6296 up the length of negative integers. This has now been simplified to just
6297 store the length when it is first determined and use it later, rather
6298 than trying to keep track of where data is copied and updating it to
6299 point to the end.
6300 [Steve Henson, reported by Brien Wheeler
6301 <bwheeler@authentica-security.com>]
6302
6303 *) Add a new function PKCS7_signatureVerify. This allows the verification
6304 of a PKCS#7 signature but with the signing certificate passed to the
6305 function itself. This contrasts with PKCS7_dataVerify which assumes the
6306 certificate is present in the PKCS#7 structure. This isn't always the
6307 case: certificates can be omitted from a PKCS#7 structure and be
6308 distributed by "out of band" means (such as a certificate database).
6309 [Steve Henson]
6310
6311 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6312 function prototypes in pem.h, also change util/mkdef.pl to add the
6313 necessary function names.
6314 [Steve Henson]
6315
6316 *) mk1mf.pl (used by Windows builds) did not properly read the
6317 options set by Configure in the top level Makefile, and Configure
6318 was not even able to write more than one option correctly.
6319 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6320 [Bodo Moeller]
6321
6322 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6323 file to be loaded from a BIO or FILE pointer. The BIO version will
6324 for example allow memory BIOs to contain config info.
6325 [Steve Henson]
6326
6327 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6328 Whoever hopes to achieve shared-library compatibility across versions
6329 must use this, not the compile-time macro.
6330 (Exercise 0.9.4: Which is the minimum library version required by
6331 such programs?)
6332 Note: All this applies only to multi-threaded programs, others don't
6333 need locks.
6334 [Bodo Moeller]
6335
6336 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6337 through a BIO pair triggered the default case, i.e.
6338 SSLerr(...,SSL_R_UNKNOWN_STATE).
6339 [Bodo Moeller]
6340
6341 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6342 can use the SSL library even if none of the specific BIOs is
6343 appropriate.
6344 [Bodo Moeller]
6345
6346 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6347 for the encoded length.
6348 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6349
6350 *) Add initial documentation of the X509V3 functions.
6351 [Steve Henson]
6352
6353 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6354 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6355 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6356 secure PKCS#8 private key format with a high iteration count.
6357 [Steve Henson]
6358
6359 *) Fix determination of Perl interpreter: A perl or perl5
6360 _directory_ in $PATH was also accepted as the interpreter.
6361 [Ralf S. Engelschall]
6362
6363 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6364 wrong with it but it was very old and did things like calling
6365 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6366 unusual formatting.
6367 [Steve Henson]
6368
6369 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6370 to use the new extension code.
6371 [Steve Henson]
6372
6373 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6374 with macros. This should make it easier to change their form, add extra
6375 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6376 constant.
6377 [Steve Henson]
6378
6379 *) Add to configuration table a new entry that can specify an alternative
6380 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6381 according to Mark Crispin <MRC@Panda.COM>.
6382 [Bodo Moeller]
6383
6384 #if 0
6385 *) DES CBC did not update the IV. Weird.
6386 [Ben Laurie]
6387 #else
6388 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6389 Changing the behaviour of the former might break existing programs --
6390 where IV updating is needed, des_ncbc_encrypt can be used.
6391 #endif
6392
6393 *) When bntest is run from "make test" it drives bc to check its
6394 calculations, as well as internally checking them. If an internal check
6395 fails, it needs to cause bc to give a non-zero result or make test carries
6396 on without noticing the failure. Fixed.
6397 [Ben Laurie]
6398
6399 *) DES library cleanups.
6400 [Ulf Möller]
6401
6402 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6403 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6404 ciphers. NOTE: although the key derivation function has been verified
6405 against some published test vectors it has not been extensively tested
6406 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6407 of v2.0.
6408 [Steve Henson]
6409
6410 *) Instead of "mkdir -p", which is not fully portable, use new
6411 Perl script "util/mkdir-p.pl".
6412 [Bodo Moeller]
6413
6414 *) Rewrite the way password based encryption (PBE) is handled. It used to
6415 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6416 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6417 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6418 the 'parameter' field of the AlgorithmIdentifier is passed to the
6419 underlying key generation function so it must do its own ASN1 parsing.
6420 This has also changed the EVP_PBE_CipherInit() function which now has a
6421 'parameter' argument instead of literal salt and iteration count values
6422 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6423 [Steve Henson]
6424
6425 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6426 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6427 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6428 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6429 value was just used as a "magic string" and not used directly its
6430 value doesn't matter.
6431 [Steve Henson]
6432
6433 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6434 support mutable.
6435 [Ben Laurie]
6436
6437 *) "linux-sparc64" configuration (ultrapenguin).
6438 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6439 "linux-sparc" configuration.
6440 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6441
6442 *) config now generates no-xxx options for missing ciphers.
6443 [Ulf Möller]
6444
6445 *) Support the EBCDIC character set (work in progress).
6446 File ebcdic.c not yet included because it has a different license.
6447 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6448
6449 *) Support BS2000/OSD-POSIX.
6450 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6451
6452 *) Make callbacks for key generation use void * instead of char *.
6453 [Ben Laurie]
6454
6455 *) Make S/MIME samples compile (not yet tested).
6456 [Ben Laurie]
6457
6458 *) Additional typesafe stacks.
6459 [Ben Laurie]
6460
6461 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6462 [Bodo Moeller]
6463
6464
6465 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6466
6467 *) New configuration variant "sco5-gcc".
6468
6469 *) Updated some demos.
6470 [Sean O Riordain, Wade Scholine]
6471
6472 *) Add missing BIO_free at exit of pkcs12 application.
6473 [Wu Zhigang]
6474
6475 *) Fix memory leak in conf.c.
6476 [Steve Henson]
6477
6478 *) Updates for Win32 to assembler version of MD5.
6479 [Steve Henson]
6480
6481 *) Set #! path to perl in apps/der_chop to where we found it
6482 instead of using a fixed path.
6483 [Bodo Moeller]
6484
6485 *) SHA library changes for irix64-mips4-cc.
6486 [Andy Polyakov]
6487
6488 *) Improvements for VMS support.
6489 [Richard Levitte]
6490
6491
6492 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6493
6494 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6495 This also avoids the problems with SC4.2 and unpatched SC5.
6496 [Andy Polyakov <appro@fy.chalmers.se>]
6497
6498 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6499 These are required because of the typesafe stack would otherwise break
6500 existing code. If old code used a structure member which used to be STACK
6501 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6502 sk_num or sk_value it would produce an error because the num, data members
6503 are not present in STACK_OF. Now it just produces a warning. sk_set
6504 replaces the old method of assigning a value to sk_value
6505 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6506 that does this will no longer work (and should use sk_set instead) but
6507 this could be regarded as a "questionable" behaviour anyway.
6508 [Steve Henson]
6509
6510 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6511 correctly handle encrypted S/MIME data.
6512 [Steve Henson]
6513
6514 *) Change type of various DES function arguments from des_cblock
6515 (which means, in function argument declarations, pointer to char)
6516 to des_cblock * (meaning pointer to array with 8 char elements),
6517 which allows the compiler to do more typechecking; it was like
6518 that back in SSLeay, but with lots of ugly casts.
6519
6520 Introduce new type const_des_cblock.
6521 [Bodo Moeller]
6522
6523 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6524 problems: find RecipientInfo structure that matches recipient certificate
6525 and initialise the ASN1 structures properly based on passed cipher.
6526 [Steve Henson]
6527
6528 *) Belatedly make the BN tests actually check the results.
6529 [Ben Laurie]
6530
6531 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6532 to and from BNs: it was completely broken. New compilation option
6533 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6534 key elements as negative integers.
6535 [Steve Henson]
6536
6537 *) Reorganize and speed up MD5.
6538 [Andy Polyakov <appro@fy.chalmers.se>]
6539
6540 *) VMS support.
6541 [Richard Levitte <richard@levitte.org>]
6542
6543 *) New option -out to asn1parse to allow the parsed structure to be
6544 output to a file. This is most useful when combined with the -strparse
6545 option to examine the output of things like OCTET STRINGS.
6546 [Steve Henson]
6547
6548 *) Make SSL library a little more fool-proof by not requiring any longer
6549 that SSL_set_{accept,connect}_state be called before
6550 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6551 in many applications because usually everything *appeared* to work as
6552 intended anyway -- now it really works as intended).
6553 [Bodo Moeller]
6554
6555 *) Move openssl.cnf out of lib/.
6556 [Ulf Möller]
6557
6558 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6559 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6560 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6561 [Ralf S. Engelschall]
6562
6563 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6564 handle PKCS#7 enveloped data properly.
6565 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6566
6567 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6568 copying pointers. The cert_st handling is changed by this in
6569 various ways (and thus what used to be known as ctx->default_cert
6570 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6571 any longer when s->cert does not give us what we need).
6572 ssl_cert_instantiate becomes obsolete by this change.
6573 As soon as we've got the new code right (possibly it already is?),
6574 we have solved a couple of bugs of the earlier code where s->cert
6575 was used as if it could not have been shared with other SSL structures.
6576
6577 Note that using the SSL API in certain dirty ways now will result
6578 in different behaviour than observed with earlier library versions:
6579 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6580 does not influence s as it used to.
6581
6582 In order to clean up things more thoroughly, inside SSL_SESSION
6583 we don't use CERT any longer, but a new structure SESS_CERT
6584 that holds per-session data (if available); currently, this is
6585 the peer's certificate chain and, for clients, the server's certificate
6586 and temporary key. CERT holds only those values that can have
6587 meaningful defaults in an SSL_CTX.
6588 [Bodo Moeller]
6589
6590 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6591 from the internal representation. Various PKCS#7 fixes: remove some
6592 evil casts and set the enc_dig_alg field properly based on the signing
6593 key type.
6594 [Steve Henson]
6595
6596 *) Allow PKCS#12 password to be set from the command line or the
6597 environment. Let 'ca' get its config file name from the environment
6598 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6599 and 'x509').
6600 [Steve Henson]
6601
6602 *) Allow certificate policies extension to use an IA5STRING for the
6603 organization field. This is contrary to the PKIX definition but
6604 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6605 extension option.
6606 [Steve Henson]
6607
6608 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6609 without disallowing inline assembler and the like for non-pedantic builds.
6610 [Ben Laurie]
6611
6612 *) Support Borland C++ builder.
6613 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6614
6615 *) Support Mingw32.
6616 [Ulf Möller]
6617
6618 *) SHA-1 cleanups and performance enhancements.
6619 [Andy Polyakov <appro@fy.chalmers.se>]
6620
6621 *) Sparc v8plus assembler for the bignum library.
6622 [Andy Polyakov <appro@fy.chalmers.se>]
6623
6624 *) Accept any -xxx and +xxx compiler options in Configure.
6625 [Ulf Möller]
6626
6627 *) Update HPUX configuration.
6628 [Anonymous]
6629
6630 *) Add missing sk_<type>_unshift() function to safestack.h
6631 [Ralf S. Engelschall]
6632
6633 *) New function SSL_CTX_use_certificate_chain_file that sets the
6634 "extra_cert"s in addition to the certificate. (This makes sense
6635 only for "PEM" format files, as chains as a whole are not
6636 DER-encoded.)
6637 [Bodo Moeller]
6638
6639 *) Support verify_depth from the SSL API.
6640 x509_vfy.c had what can be considered an off-by-one-error:
6641 Its depth (which was not part of the external interface)
6642 was actually counting the number of certificates in a chain;
6643 now it really counts the depth.
6644 [Bodo Moeller]
6645
6646 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6647 instead of X509err, which often resulted in confusing error
6648 messages since the error codes are not globally unique
6649 (e.g. an alleged error in ssl3_accept when a certificate
6650 didn't match the private key).
6651
6652 *) New function SSL_CTX_set_session_id_context that allows to set a default
6653 value (so that you don't need SSL_set_session_id_context for each
6654 connection using the SSL_CTX).
6655 [Bodo Moeller]
6656
6657 *) OAEP decoding bug fix.
6658 [Ulf Möller]
6659
6660 *) Support INSTALL_PREFIX for package builders, as proposed by
6661 David Harris.
6662 [Bodo Moeller]
6663
6664 *) New Configure options "threads" and "no-threads". For systems
6665 where the proper compiler options are known (currently Solaris
6666 and Linux), "threads" is the default.
6667 [Bodo Moeller]
6668
6669 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6670 [Bodo Moeller]
6671
6672 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6673 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6674 such as /usr/local/bin.
6675 [Bodo Moeller]
6676
6677 *) "make linux-shared" to build shared libraries.
6678 [Niels Poppe <niels@netbox.org>]
6679
6680 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6681 [Ulf Möller]
6682
6683 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6684 extension adding in x509 utility.
6685 [Steve Henson]
6686
6687 *) Remove NOPROTO sections and error code comments.
6688 [Ulf Möller]
6689
6690 *) Partial rewrite of the DEF file generator to now parse the ANSI
6691 prototypes.
6692 [Steve Henson]
6693
6694 *) New Configure options --prefix=DIR and --openssldir=DIR.
6695 [Ulf Möller]
6696
6697 *) Complete rewrite of the error code script(s). It is all now handled
6698 by one script at the top level which handles error code gathering,
6699 header rewriting and C source file generation. It should be much better
6700 than the old method: it now uses a modified version of Ulf's parser to
6701 read the ANSI prototypes in all header files (thus the old K&R definitions
6702 aren't needed for error creation any more) and do a better job of
6703 translating function codes into names. The old 'ASN1 error code imbedded
6704 in a comment' is no longer necessary and it doesn't use .err files which
6705 have now been deleted. Also the error code call doesn't have to appear all
6706 on one line (which resulted in some large lines...).
6707 [Steve Henson]
6708
6709 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6710 [Bodo Moeller]
6711
6712 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6713 0 (which usually indicates a closed connection), but continue reading.
6714 [Bodo Moeller]
6715
6716 *) Fix some race conditions.
6717 [Bodo Moeller]
6718
6719 *) Add support for CRL distribution points extension. Add Certificate
6720 Policies and CRL distribution points documentation.
6721 [Steve Henson]
6722
6723 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6724 [Ulf Möller]
6725
6726 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6727 8 of keying material. Merlin has also confirmed interop with this fix
6728 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6729 [Merlin Hughes <merlin@baltimore.ie>]
6730
6731 *) Fix lots of warnings.
6732 [Richard Levitte <levitte@stacken.kth.se>]
6733
6734 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6735 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6736 [Richard Levitte <levitte@stacken.kth.se>]
6737
6738 *) Fix problems with sizeof(long) == 8.
6739 [Andy Polyakov <appro@fy.chalmers.se>]
6740
6741 *) Change functions to ANSI C.
6742 [Ulf Möller]
6743
6744 *) Fix typos in error codes.
6745 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6746
6747 *) Remove defunct assembler files from Configure.
6748 [Ulf Möller]
6749
6750 *) SPARC v8 assembler BIGNUM implementation.
6751 [Andy Polyakov <appro@fy.chalmers.se>]
6752
6753 *) Support for Certificate Policies extension: both print and set.
6754 Various additions to support the r2i method this uses.
6755 [Steve Henson]
6756
6757 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6758 return a const string when you are expecting an allocated buffer.
6759 [Ben Laurie]
6760
6761 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6762 types DirectoryString and DisplayText.
6763 [Steve Henson]
6764
6765 *) Add code to allow r2i extensions to access the configuration database,
6766 add an LHASH database driver and add several ctx helper functions.
6767 [Steve Henson]
6768
6769 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6770 fail when they extended the size of a BIGNUM.
6771 [Steve Henson]
6772
6773 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6774 support typesafe stack.
6775 [Steve Henson]
6776
6777 *) Fix typo in SSL_[gs]et_options().
6778 [Nils Frostberg <nils@medcom.se>]
6779
6780 *) Delete various functions and files that belonged to the (now obsolete)
6781 old X509V3 handling code.
6782 [Steve Henson]
6783
6784 *) New Configure option "rsaref".
6785 [Ulf Möller]
6786
6787 *) Don't auto-generate pem.h.
6788 [Bodo Moeller]
6789
6790 *) Introduce type-safe ASN.1 SETs.
6791 [Ben Laurie]
6792
6793 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6794 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6795
6796 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6797 that links with OpenSSL (well at least cause lots of warnings), but fear
6798 not: the conversion is trivial, and it eliminates loads of evil casts. A
6799 few STACKed things have been converted already. Feel free to convert more.
6800 In the fullness of time, I'll do away with the STACK type altogether.
6801 [Ben Laurie]
6802
6803 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6804 specified in <certfile> by updating the entry in the index.txt file.
6805 This way one no longer has to edit the index.txt file manually for
6806 revoking a certificate. The -revoke option does the gory details now.
6807 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6808
6809 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6810 `-text' option at all and this way the `-noout -text' combination was
6811 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6812 [Ralf S. Engelschall]
6813
6814 *) Make sure a corresponding plain text error message exists for the
6815 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6816 verify callback function determined that a certificate was revoked.
6817 [Ralf S. Engelschall]
6818
6819 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6820 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6821 all available cipers including rc5, which was forgotten until now.
6822 In order to let the testing shell script know which algorithms
6823 are available, a new (up to now undocumented) command
6824 "openssl list-cipher-commands" is used.
6825 [Bodo Moeller]
6826
6827 *) Bugfix: s_client occasionally would sleep in select() when
6828 it should have checked SSL_pending() first.
6829 [Bodo Moeller]
6830
6831 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6832 the raw DSA values prior to ASN.1 encoding.
6833 [Ulf Möller]
6834
6835 *) Tweaks to Configure
6836 [Niels Poppe <niels@netbox.org>]
6837
6838 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6839 yet...
6840 [Steve Henson]
6841
6842 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6843 [Ulf Möller]
6844
6845 *) New config option to avoid instructions that are illegal on the 80386.
6846 The default code is faster, but requires at least a 486.
6847 [Ulf Möller]
6848
6849 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6850 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6851 same as SSL2_VERSION anyway.
6852 [Bodo Moeller]
6853
6854 *) New "-showcerts" option for s_client.
6855 [Bodo Moeller]
6856
6857 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6858 application. Various cleanups and fixes.
6859 [Steve Henson]
6860
6861 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6862 modify error routines to work internally. Add error codes and PBE init
6863 to library startup routines.
6864 [Steve Henson]
6865
6866 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6867 packing functions to asn1 and evp. Changed function names and error
6868 codes along the way.
6869 [Steve Henson]
6870
6871 *) PKCS12 integration: and so it begins... First of several patches to
6872 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6873 objects to objects.h
6874 [Steve Henson]
6875
6876 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6877 and display support for Thawte strong extranet extension.
6878 [Steve Henson]
6879
6880 *) Add LinuxPPC support.
6881 [Jeff Dubrule <igor@pobox.org>]
6882
6883 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6884 bn_div_words in alpha.s.
6885 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6886
6887 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6888 OAEP isn't supported when OpenSSL is built with RSAref.
6889 [Ulf Moeller <ulf@fitug.de>]
6890
6891 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6892 so they no longer are missing under -DNOPROTO.
6893 [Soren S. Jorvang <soren@t.dk>]
6894
6895
6896 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6897
6898 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6899 doesn't work when the session is reused. Coming soon!
6900 [Ben Laurie]
6901
6902 *) Fix a security hole, that allows sessions to be reused in the wrong
6903 context thus bypassing client cert protection! All software that uses
6904 client certs and session caches in multiple contexts NEEDS PATCHING to
6905 allow session reuse! A fuller solution is in the works.
6906 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6907
6908 *) Some more source tree cleanups (removed obsolete files
6909 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6910 permission on "config" script to be executable) and a fix for the INSTALL
6911 document.
6912 [Ulf Moeller <ulf@fitug.de>]
6913
6914 *) Remove some legacy and erroneous uses of malloc, free instead of
6915 Malloc, Free.
6916 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6917
6918 *) Make rsa_oaep_test return non-zero on error.
6919 [Ulf Moeller <ulf@fitug.de>]
6920
6921 *) Add support for native Solaris shared libraries. Configure
6922 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6923 if someone would make that last step automatic.
6924 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6925
6926 *) ctx_size was not built with the right compiler during "make links". Fixed.
6927 [Ben Laurie]
6928
6929 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6930 except NULL ciphers". This means the default cipher list will no longer
6931 enable NULL ciphers. They need to be specifically enabled e.g. with
6932 the string "DEFAULT:eNULL".
6933 [Steve Henson]
6934
6935 *) Fix to RSA private encryption routines: if p < q then it would
6936 occasionally produce an invalid result. This will only happen with
6937 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6938 [Steve Henson]
6939
6940 *) Be less restrictive and allow also `perl util/perlpath.pl
6941 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6942 because this way one can also use an interpreter named `perl5' (which is
6943 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6944 installed as `perl').
6945 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6946
6947 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6948 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6949
6950 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
6951 advapi32.lib to Win32 build and change the pem test comparision
6952 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
6953 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
6954 and crypto/des/ede_cbcm_enc.c.
6955 [Steve Henson]
6956
6957 *) DES quad checksum was broken on big-endian architectures. Fixed.
6958 [Ben Laurie]
6959
6960 *) Comment out two functions in bio.h that aren't implemented. Fix up the
6961 Win32 test batch file so it (might) work again. The Win32 test batch file
6962 is horrible: I feel ill....
6963 [Steve Henson]
6964
6965 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
6966 in e_os.h. Audit of header files to check ANSI and non ANSI
6967 sections: 10 functions were absent from non ANSI section and not exported
6968 from Windows DLLs. Fixed up libeay.num for new functions.
6969 [Steve Henson]
6970
6971 *) Make `openssl version' output lines consistent.
6972 [Ralf S. Engelschall]
6973
6974 *) Fix Win32 symbol export lists for BIO functions: Added
6975 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
6976 to ms/libeay{16,32}.def.
6977 [Ralf S. Engelschall]
6978
6979 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
6980 fine under Unix and passes some trivial tests I've now added. But the
6981 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
6982 added to make sure no one expects that this stuff really works in the
6983 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
6984 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
6985 openssl_bio.xs.
6986 [Ralf S. Engelschall]
6987
6988 *) Fix the generation of two part addresses in perl.
6989 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
6990
6991 *) Add config entry for Linux on MIPS.
6992 [John Tobey <jtobey@channel1.com>]
6993
6994 *) Make links whenever Configure is run, unless we are on Windoze.
6995 [Ben Laurie]
6996
6997 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
6998 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
6999 in CRLs.
7000 [Steve Henson]
7001
7002 *) Add a useful kludge to allow package maintainers to specify compiler and
7003 other platforms details on the command line without having to patch the
7004 Configure script everytime: One now can use ``perl Configure
7005 <id>:<details>'', i.e. platform ids are allowed to have details appended
7006 to them (seperated by colons). This is treated as there would be a static
7007 pre-configured entry in Configure's %table under key <id> with value
7008 <details> and ``perl Configure <id>'' is called. So, when you want to
7009 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7010 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7011 now, which overrides the FreeBSD-elf entry on-the-fly.
7012 [Ralf S. Engelschall]
7013
7014 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7015 [Ben Laurie]
7016
7017 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7018 on the `perl Configure ...' command line. This way one can compile
7019 OpenSSL libraries with Position Independent Code (PIC) which is needed
7020 for linking it into DSOs.
7021 [Ralf S. Engelschall]
7022
7023 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7024 Fixed.
7025 [Ben Laurie]
7026
7027 *) Cleaned up the LICENSE document: The official contact for any license
7028 questions now is the OpenSSL core team under openssl-core@openssl.org.
7029 And add a paragraph about the dual-license situation to make sure people
7030 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7031 to the OpenSSL toolkit.
7032 [Ralf S. Engelschall]
7033
7034 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7035 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7036 Additonally cleaned up the `make links' target: Remove unnecessary
7037 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7038 to speed processing and no longer clutter the display with confusing
7039 stuff. Instead only the actually done links are displayed.
7040 [Ralf S. Engelschall]
7041
7042 *) Permit null encryption ciphersuites, used for authentication only. It used
7043 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7044 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7045 encryption.
7046 [Ben Laurie]
7047
7048 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7049 signed attributes when verifying signatures (this would break them),
7050 the detached data encoding was wrong and public keys obtained using
7051 X509_get_pubkey() weren't freed.
7052 [Steve Henson]
7053
7054 *) Add text documentation for the BUFFER functions. Also added a work around
7055 to a Win95 console bug. This was triggered by the password read stuff: the
7056 last character typed gets carried over to the next fread(). If you were
7057 generating a new cert request using 'req' for example then the last
7058 character of the passphrase would be CR which would then enter the first
7059 field as blank.
7060 [Steve Henson]
7061
7062 *) Added the new `Includes OpenSSL Cryptography Software' button as
7063 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7064 button and can be used by applications based on OpenSSL to show the
7065 relationship to the OpenSSL project.
7066 [Ralf S. Engelschall]
7067
7068 *) Remove confusing variables in function signatures in files
7069 ssl/ssl_lib.c and ssl/ssl.h.
7070 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7071
7072 *) Don't install bss_file.c under PREFIX/include/
7073 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7074
7075 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7076 functions that return function pointers and has support for NT specific
7077 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7078 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7079 unsigned to signed types: this was killing the Win32 compile.
7080 [Steve Henson]
7081
7082 *) Add new certificate file to stack functions,
7083 SSL_add_dir_cert_subjects_to_stack() and
7084 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7085 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7086 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7087 This means that Apache-SSL and similar packages don't have to mess around
7088 to add as many CAs as they want to the preferred list.
7089 [Ben Laurie]
7090
7091 *) Experiment with doxygen documentation. Currently only partially applied to
7092 ssl/ssl_lib.c.
7093 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7094 openssl.doxy as the configuration file.
7095 [Ben Laurie]
7096
7097 *) Get rid of remaining C++-style comments which strict C compilers hate.
7098 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7099
7100 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7101 compiled in by default: it has problems with large keys.
7102 [Steve Henson]
7103
7104 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7105 DH private keys and/or callback functions which directly correspond to
7106 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7107 is needed for applications which have to configure certificates on a
7108 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7109 (e.g. s_server).
7110 For the RSA certificate situation is makes no difference, but
7111 for the DSA certificate situation this fixes the "no shared cipher"
7112 problem where the OpenSSL cipher selection procedure failed because the
7113 temporary keys were not overtaken from the context and the API provided
7114 no way to reconfigure them.
7115 The new functions now let applications reconfigure the stuff and they
7116 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7117 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7118 non-public-API function ssl_cert_instantiate() is used as a helper
7119 function and also to reduce code redundancy inside ssl_rsa.c.
7120 [Ralf S. Engelschall]
7121
7122 *) Move s_server -dcert and -dkey options out of the undocumented feature
7123 area because they are useful for the DSA situation and should be
7124 recognized by the users.
7125 [Ralf S. Engelschall]
7126
7127 *) Fix the cipher decision scheme for export ciphers: the export bits are
7128 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7129 SSL_EXP_MASK. So, the original variable has to be used instead of the
7130 already masked variable.
7131 [Richard Levitte <levitte@stacken.kth.se>]
7132
7133 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7134 [Richard Levitte <levitte@stacken.kth.se>]
7135
7136 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7137 from `int' to `unsigned int' because it's a length and initialized by
7138 EVP_DigestFinal() which expects an `unsigned int *'.
7139 [Richard Levitte <levitte@stacken.kth.se>]
7140
7141 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7142 script. Instead use the usual Shell->Perl transition trick.
7143 [Ralf S. Engelschall]
7144
7145 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7146 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7147 -noout -modulus' as it's already the case for `openssl rsa -noout
7148 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7149 currently the public key is printed (a decision which was already done by
7150 `openssl dsa -modulus' in the past) which serves a similar purpose.
7151 Additionally the NO_RSA no longer completely removes the whole -modulus
7152 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7153 now, too.
7154 [Ralf S. Engelschall]
7155
7156 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7157 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7158 [Arne Ansper <arne@ats.cyber.ee>]
7159
7160 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7161 to be added. Now both 'req' and 'ca' can use new objects defined in the
7162 config file.
7163 [Steve Henson]
7164
7165 *) Add cool BIO that does syslog (or event log on NT).
7166 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7167
7168 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7169 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7170 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7171 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7172 [Ben Laurie]
7173
7174 *) Add preliminary config info for new extension code.
7175 [Steve Henson]
7176
7177 *) Make RSA_NO_PADDING really use no padding.
7178 [Ulf Moeller <ulf@fitug.de>]
7179
7180 *) Generate errors when private/public key check is done.
7181 [Ben Laurie]
7182
7183 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7184 for some CRL extensions and new objects added.
7185 [Steve Henson]
7186
7187 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7188 key usage extension and fuller support for authority key id.
7189 [Steve Henson]
7190
7191 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7192 padding method for RSA, which is recommended for new applications in PKCS
7193 #1 v2.0 (RFC 2437, October 1998).
7194 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7195 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7196 against Bleichbacher's attack on RSA.
7197 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7198 Ben Laurie]
7199
7200 *) Updates to the new SSL compression code
7201 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7202
7203 *) Fix so that the version number in the master secret, when passed
7204 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7205 (because the server will not accept higher), that the version number
7206 is 0x03,0x01, not 0x03,0x00
7207 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7208
7209 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7210 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7211 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7212 [Steve Henson]
7213
7214 *) Support for RAW extensions where an arbitrary extension can be
7215 created by including its DER encoding. See apps/openssl.cnf for
7216 an example.
7217 [Steve Henson]
7218
7219 *) Make sure latest Perl versions don't interpret some generated C array
7220 code as Perl array code in the crypto/err/err_genc.pl script.
7221 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7222
7223 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7224 not many people have the assembler. Various Win32 compilation fixes and
7225 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7226 build instructions.
7227 [Steve Henson]
7228
7229 *) Modify configure script 'Configure' to automatically create crypto/date.h
7230 file under Win32 and also build pem.h from pem.org. New script
7231 util/mkfiles.pl to create the MINFO file on environments that can't do a
7232 'make files': perl util/mkfiles.pl >MINFO should work.
7233 [Steve Henson]
7234
7235 *) Major rework of DES function declarations, in the pursuit of correctness
7236 and purity. As a result, many evil casts evaporated, and some weirdness,
7237 too. You may find this causes warnings in your code. Zapping your evil
7238 casts will probably fix them. Mostly.
7239 [Ben Laurie]
7240
7241 *) Fix for a typo in asn1.h. Bug fix to object creation script
7242 obj_dat.pl. It considered a zero in an object definition to mean
7243 "end of object": none of the objects in objects.h have any zeros
7244 so it wasn't spotted.
7245 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7246
7247 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7248 Masking (CBCM). In the absence of test vectors, the best I have been able
7249 to do is check that the decrypt undoes the encrypt, so far. Send me test
7250 vectors if you have them.
7251 [Ben Laurie]
7252
7253 *) Correct calculation of key length for export ciphers (too much space was
7254 allocated for null ciphers). This has not been tested!
7255 [Ben Laurie]
7256
7257 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7258 message is now correct (it understands "crypto" and "ssl" on its
7259 command line). There is also now an "update" option. This will update
7260 the util/ssleay.num and util/libeay.num files with any new functions.
7261 If you do a:
7262 perl util/mkdef.pl crypto ssl update
7263 it will update them.
7264 [Steve Henson]
7265
7266 *) Overhauled the Perl interface (perl/*):
7267 - ported BN stuff to OpenSSL's different BN library
7268 - made the perl/ source tree CVS-aware
7269 - renamed the package from SSLeay to OpenSSL (the files still contain
7270 their history because I've copied them in the repository)
7271 - removed obsolete files (the test scripts will be replaced
7272 by better Test::Harness variants in the future)
7273 [Ralf S. Engelschall]
7274
7275 *) First cut for a very conservative source tree cleanup:
7276 1. merge various obsolete readme texts into doc/ssleay.txt
7277 where we collect the old documents and readme texts.
7278 2. remove the first part of files where I'm already sure that we no
7279 longer need them because of three reasons: either they are just temporary
7280 files which were left by Eric or they are preserved original files where
7281 I've verified that the diff is also available in the CVS via "cvs diff
7282 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7283 the crypto/md/ stuff).
7284 [Ralf S. Engelschall]
7285
7286 *) More extension code. Incomplete support for subject and issuer alt
7287 name, issuer and authority key id. Change the i2v function parameters
7288 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7289 what that's for :-) Fix to ASN1 macro which messed up
7290 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7291 [Steve Henson]
7292
7293 *) Preliminary support for ENUMERATED type. This is largely copied from the
7294 INTEGER code.
7295 [Steve Henson]
7296
7297 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7298 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7299
7300 *) Make sure `make rehash' target really finds the `openssl' program.
7301 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7302
7303 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7304 like to hear about it if this slows down other processors.
7305 [Ben Laurie]
7306
7307 *) Add CygWin32 platform information to Configure script.
7308 [Alan Batie <batie@aahz.jf.intel.com>]
7309
7310 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7311 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7312
7313 *) New program nseq to manipulate netscape certificate sequences
7314 [Steve Henson]
7315
7316 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7317 few typos.
7318 [Steve Henson]
7319
7320 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7321 but the BN code had some problems that would cause failures when
7322 doing certificate verification and some other functions.
7323 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7324
7325 *) Add ASN1 and PEM code to support netscape certificate sequences.
7326 [Steve Henson]
7327
7328 *) Add ASN1 and PEM code to support netscape certificate sequences.
7329 [Steve Henson]
7330
7331 *) Add several PKIX and private extended key usage OIDs.
7332 [Steve Henson]
7333
7334 *) Modify the 'ca' program to handle the new extension code. Modify
7335 openssl.cnf for new extension format, add comments.
7336 [Steve Henson]
7337
7338 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7339 and add a sample to openssl.cnf so req -x509 now adds appropriate
7340 CA extensions.
7341 [Steve Henson]
7342
7343 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7344 error code, add initial support to X509_print() and x509 application.
7345 [Steve Henson]
7346
7347 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7348 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7349 stuff is currently isolated and isn't even compiled yet.
7350 [Steve Henson]
7351
7352 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7353 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7354 Removed the versions check from X509 routines when loading extensions:
7355 this allows certain broken certificates that don't set the version
7356 properly to be processed.
7357 [Steve Henson]
7358
7359 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7360 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7361 can still be regenerated with "make depend".
7362 [Ben Laurie]
7363
7364 *) Spelling mistake in C version of CAST-128.
7365 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7366
7367 *) Changes to the error generation code. The perl script err-code.pl
7368 now reads in the old error codes and retains the old numbers, only
7369 adding new ones if necessary. It also only changes the .err files if new
7370 codes are added. The makefiles have been modified to only insert errors
7371 when needed (to avoid needlessly modifying header files). This is done
7372 by only inserting errors if the .err file is newer than the auto generated
7373 C file. To rebuild all the error codes from scratch (the old behaviour)
7374 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7375 or delete all the .err files.
7376 [Steve Henson]
7377
7378 *) CAST-128 was incorrectly implemented for short keys. The C version has
7379 been fixed, but is untested. The assembler versions are also fixed, but
7380 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7381 to regenerate it if needed.
7382 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7383 Hagino <itojun@kame.net>]
7384
7385 *) File was opened incorrectly in randfile.c.
7386 [Ulf Möller <ulf@fitug.de>]
7387
7388 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7389 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7390 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7391 al: it's just almost always a UTCTime. Note this patch adds new error
7392 codes so do a "make errors" if there are problems.
7393 [Steve Henson]
7394
7395 *) Correct Linux 1 recognition in config.
7396 [Ulf Möller <ulf@fitug.de>]
7397
7398 *) Remove pointless MD5 hash when using DSA keys in ca.
7399 [Anonymous <nobody@replay.com>]
7400
7401 *) Generate an error if given an empty string as a cert directory. Also
7402 generate an error if handed NULL (previously returned 0 to indicate an
7403 error, but didn't set one).
7404 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7405
7406 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7407 [Ben Laurie]
7408
7409 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7410 parameters. This was causing a warning which killed off the Win32 compile.
7411 [Steve Henson]
7412
7413 *) Remove C++ style comments from crypto/bn/bn_local.h.
7414 [Neil Costigan <neil.costigan@celocom.com>]
7415
7416 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7417 based on a text string, looking up short and long names and finally
7418 "dot" format. The "dot" format stuff didn't work. Added new function
7419 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7420 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7421 OID is not part of the table.
7422 [Steve Henson]
7423
7424 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7425 X509_LOOKUP_by_alias().
7426 [Ben Laurie]
7427
7428 *) Sort openssl functions by name.
7429 [Ben Laurie]
7430
7431 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7432 encryption from sample DSA keys (in case anyone is interested the password
7433 was "1234").
7434 [Steve Henson]
7435
7436 *) Make _all_ *_free functions accept a NULL pointer.
7437 [Frans Heymans <fheymans@isaserver.be>]
7438
7439 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7440 NULL pointers.
7441 [Anonymous <nobody@replay.com>]
7442
7443 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7444 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7445
7446 *) Don't blow it for numeric -newkey arguments to apps/req.
7447 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7448
7449 *) Temp key "for export" tests were wrong in s3_srvr.c.
7450 [Anonymous <nobody@replay.com>]
7451
7452 *) Add prototype for temp key callback functions
7453 SSL_CTX_set_tmp_{rsa,dh}_callback().
7454 [Ben Laurie]
7455
7456 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7457 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7458 [Steve Henson]
7459
7460 *) X509_name_add_entry() freed the wrong thing after an error.
7461 [Arne Ansper <arne@ats.cyber.ee>]
7462
7463 *) rsa_eay.c would attempt to free a NULL context.
7464 [Arne Ansper <arne@ats.cyber.ee>]
7465
7466 *) BIO_s_socket() had a broken should_retry() on Windoze.
7467 [Arne Ansper <arne@ats.cyber.ee>]
7468
7469 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7470 [Arne Ansper <arne@ats.cyber.ee>]
7471
7472 *) Make sure the already existing X509_STORE->depth variable is initialized
7473 in X509_STORE_new(), but document the fact that this variable is still
7474 unused in the certificate verification process.
7475 [Ralf S. Engelschall]
7476
7477 *) Fix the various library and apps files to free up pkeys obtained from
7478 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7479 [Steve Henson]
7480
7481 *) Fix reference counting in X509_PUBKEY_get(). This makes
7482 demos/maurice/example2.c work, amongst others, probably.
7483 [Steve Henson and Ben Laurie]
7484
7485 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7486 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7487 are no longer created. This way we have a single and consistent command
7488 line interface `openssl <command>', similar to `cvs <command>'.
7489 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7490
7491 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7492 BIT STRING wrapper always have zero unused bits.
7493 [Steve Henson]
7494
7495 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7496 [Steve Henson]
7497
7498 *) Make the top-level INSTALL documentation easier to understand.
7499 [Paul Sutton]
7500
7501 *) Makefiles updated to exit if an error occurs in a sub-directory
7502 make (including if user presses ^C) [Paul Sutton]
7503
7504 *) Make Montgomery context stuff explicit in RSA data structure.
7505 [Ben Laurie]
7506
7507 *) Fix build order of pem and err to allow for generated pem.h.
7508 [Ben Laurie]
7509
7510 *) Fix renumbering bug in X509_NAME_delete_entry().
7511 [Ben Laurie]
7512
7513 *) Enhanced the err-ins.pl script so it makes the error library number
7514 global and can add a library name. This is needed for external ASN1 and
7515 other error libraries.
7516 [Steve Henson]
7517
7518 *) Fixed sk_insert which never worked properly.
7519 [Steve Henson]
7520
7521 *) Fix ASN1 macros so they can handle indefinite length construted
7522 EXPLICIT tags. Some non standard certificates use these: they can now
7523 be read in.
7524 [Steve Henson]
7525
7526 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7527 into a single doc/ssleay.txt bundle. This way the information is still
7528 preserved but no longer messes up this directory. Now it's new room for
7529 the new set of documenation files.
7530 [Ralf S. Engelschall]
7531
7532 *) SETs were incorrectly DER encoded. This was a major pain, because they
7533 shared code with SEQUENCEs, which aren't coded the same. This means that
7534 almost everything to do with SETs or SEQUENCEs has either changed name or
7535 number of arguments.
7536 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7537
7538 *) Fix test data to work with the above.
7539 [Ben Laurie]
7540
7541 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7542 was already fixed by Eric for 0.9.1 it seems.
7543 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7544
7545 *) Autodetect FreeBSD3.
7546 [Ben Laurie]
7547
7548 *) Fix various bugs in Configure. This affects the following platforms:
7549 nextstep
7550 ncr-scde
7551 unixware-2.0
7552 unixware-2.0-pentium
7553 sco5-cc.
7554 [Ben Laurie]
7555
7556 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7557 before they are needed.
7558 [Ben Laurie]
7559
7560 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7561 [Ben Laurie]
7562
7563
7564 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7565
7566 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7567 changed SSLeay to OpenSSL in version strings.
7568 [Ralf S. Engelschall]
7569
7570 *) Some fixups to the top-level documents.
7571 [Paul Sutton]
7572
7573 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7574 because the symlink to include/ was missing.
7575 [Ralf S. Engelschall]
7576
7577 *) Incorporated the popular no-RSA/DSA-only patches
7578 which allow to compile a RSA-free SSLeay.
7579 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7580
7581 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7582 when "ssleay" is still not found.
7583 [Ralf S. Engelschall]
7584
7585 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7586 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7587
7588 *) Updated the README file.
7589 [Ralf S. Engelschall]
7590
7591 *) Added various .cvsignore files in the CVS repository subdirs
7592 to make a "cvs update" really silent.
7593 [Ralf S. Engelschall]
7594
7595 *) Recompiled the error-definition header files and added
7596 missing symbols to the Win32 linker tables.
7597 [Ralf S. Engelschall]
7598
7599 *) Cleaned up the top-level documents;
7600 o new files: CHANGES and LICENSE
7601 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7602 o merged COPYRIGHT into LICENSE
7603 o removed obsolete TODO file
7604 o renamed MICROSOFT to INSTALL.W32
7605 [Ralf S. Engelschall]
7606
7607 *) Removed dummy files from the 0.9.1b source tree:
7608 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7609 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7610 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7611 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7612 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7613 [Ralf S. Engelschall]
7614
7615 *) Added various platform portability fixes.
7616 [Mark J. Cox]
7617
7618 *) The Genesis of the OpenSSL rpject:
7619 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7620 Young and Tim J. Hudson created while they were working for C2Net until
7621 summer 1998.
7622 [The OpenSSL Project]
7623
7624
7625 Changes between 0.9.0b and 0.9.1b [not released]
7626
7627 *) Updated a few CA certificates under certs/
7628 [Eric A. Young]
7629
7630 *) Changed some BIGNUM api stuff.
7631 [Eric A. Young]
7632
7633 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7634 DGUX x86, Linux Alpha, etc.
7635 [Eric A. Young]
7636
7637 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7638 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7639 available).
7640 [Eric A. Young]
7641
7642 *) Add -strparse option to asn1pars program which parses nested
7643 binary structures
7644 [Dr Stephen Henson <shenson@bigfoot.com>]
7645
7646 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7647 [Eric A. Young]
7648
7649 *) DSA fix for "ca" program.
7650 [Eric A. Young]
7651
7652 *) Added "-genkey" option to "dsaparam" program.
7653 [Eric A. Young]
7654
7655 *) Added RIPE MD160 (rmd160) message digest.
7656 [Eric A. Young]
7657
7658 *) Added -a (all) option to "ssleay version" command.
7659 [Eric A. Young]
7660
7661 *) Added PLATFORM define which is the id given to Configure.
7662 [Eric A. Young]
7663
7664 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7665 [Eric A. Young]
7666
7667 *) Extended the ASN.1 parser routines.
7668 [Eric A. Young]
7669
7670 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7671 [Eric A. Young]
7672
7673 *) Added a BN_CTX to the BN library.
7674 [Eric A. Young]
7675
7676 *) Fixed the weak key values in DES library
7677 [Eric A. Young]
7678
7679 *) Changed API in EVP library for cipher aliases.
7680 [Eric A. Young]
7681
7682 *) Added support for RC2/64bit cipher.
7683 [Eric A. Young]
7684
7685 *) Converted the lhash library to the crypto/mem.c functions.
7686 [Eric A. Young]
7687
7688 *) Added more recognized ASN.1 object ids.
7689 [Eric A. Young]
7690
7691 *) Added more RSA padding checks for SSL/TLS.
7692 [Eric A. Young]
7693
7694 *) Added BIO proxy/filter functionality.
7695 [Eric A. Young]
7696
7697 *) Added extra_certs to SSL_CTX which can be used
7698 send extra CA certificates to the client in the CA cert chain sending
7699 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7700 [Eric A. Young]
7701
7702 *) Now Fortezza is denied in the authentication phase because
7703 this is key exchange mechanism is not supported by SSLeay at all.
7704 [Eric A. Young]
7705
7706 *) Additional PKCS1 checks.
7707 [Eric A. Young]
7708
7709 *) Support the string "TLSv1" for all TLS v1 ciphers.
7710 [Eric A. Young]
7711
7712 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7713 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7714 [Eric A. Young]
7715
7716 *) Fixed a few memory leaks.
7717 [Eric A. Young]
7718
7719 *) Fixed various code and comment typos.
7720 [Eric A. Young]
7721
7722 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7723 bytes sent in the client random.
7724 [Edward Bishop <ebishop@spyglass.com>]
7725