]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Change default curve (for compatibility with a
[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 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
8 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
9 [Steve Henson]
10
11 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
12 utilities such as rsa, dsa, dsaparam etc except they processes any key
13 type.
14 [Steve Henson]
15
16 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
17 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
18 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
19 structure.
20 [Steve Henson]
21
22 *) Initial support for pluggable public key ASN1.
23 De-spaghettify the public key ASN1 handling. Move public and private
24 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
25 algorithm specific handling to a single module within the relevant
26 algorithm directory. Add functions to allow (near) opaque processing
27 of public and private key structures.
28 [Steve Henson]
29
30 *) Implement the Supported Point Formats Extension for
31 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
32 [Douglas Stebila]
33
34 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
35 for the psk identity [hint] and the psk callback functions to the
36 SSL_SESSION, SSL and SSL_CTX structure.
37
38 New ciphersuites:
39 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
40 PSK-AES256-CBC-SHA
41
42 New functions:
43 SSL_CTX_use_psk_identity_hint
44 SSL_get_psk_identity_hint
45 SSL_get_psk_identity
46 SSL_use_psk_identity_hint
47
48 [Mika Kousa and Pasi Eronen of Nokia Corporation]
49
50 *) Add RFC 3161 compliant time stamp request creation, response generation
51 and response verification functionality.
52 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
53
54 *) Add initial support for TLS extensions, specifically for the server_name
55 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
56 have new members for a host name. The SSL data structure has an
57 additional member SSL_CTX *initial_ctx so that new sessions can be
58 stored in that context to allow for session resumption, even after the
59 SSL has been switched to a new SSL_CTX in reaction to a client's
60 server_name extension.
61
62 New functions (subject to change):
63
64 SSL_get_servername()
65 SSL_get_servername_type()
66 SSL_set_SSL_CTX()
67
68 New CTRL codes and macros (subject to change):
69
70 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
71 - SSL_CTX_set_tlsext_servername_callback()
72 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
73 - SSL_CTX_set_tlsext_servername_arg()
74 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_hostname()
75
76 openssl s_client has a new '-servername ...' option.
77
78 openssl s_server has new options '-servername_host ...', '-cert2 ...',
79 '-key2 ...', '-servername_fatal' (subject to change). This allows
80 testing the HostName extension for a specific single host name ('-cert'
81 and '-key' remain fallbacks for handshakes without HostName
82 negotiation). If the unrecogninzed_name alert has to be sent, this by
83 default is a warning; it becomes fatal with the '-servername_fatal'
84 option.
85
86 [Peter Sylvester, Remy Allais, Christophe Renou]
87
88 *) Whirlpool hash implementation is added.
89 [Andy Polyakov]
90
91 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
92 bn(64,32). Because of instruction set limitations it doesn't have
93 any negative impact on performance. This was done mostly in order
94 to make it possible to share assembler modules, such as bn_mul_mont
95 implementations, between 32- and 64-bit builds without hassle.
96 [Andy Polyakov]
97
98 *) Disable rogue ciphersuites:
99
100 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
101 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
102 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
103
104 The latter two were purportedly from
105 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
106 appear there.
107
108 Other ciphersuites from draft-ietf-tls-56-bit-ciphersuites-01.txt
109 remain enabled for now, but are just as unofficial, and the ID
110 has long expired; these will probably disappear soon.
111 [Bodo Moeller]
112
113 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
114 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
115 macro.
116 [Bodo Moeller]
117
118 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
119 dedicated Montgomery multiplication procedure, is introduced.
120 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
121 "64-bit" performance on certain 32-bit targets.
122 [Andy Polyakov]
123
124 *) New option SSL_OP_NO_COMP to disable use of compression selectively
125 in SSL structures. New SSL ctrl to set maximum send fragment size.
126 Save memory by seeting the I/O buffer sizes dynamically instead of
127 using the maximum available value.
128 [Steve Henson]
129
130 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
131 in addition to the text details.
132 [Bodo Moeller]
133
134 *) Very, very preliminary EXPERIMENTAL support for printing of general
135 ASN1 structures. This currently produces rather ugly output and doesn't
136 handle several customised structures at all.
137 [Steve Henson]
138
139 *) Integrated support for PVK file format and some related formats such
140 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
141 these in the 'rsa' and 'dsa' utilities.
142 [Steve Henson]
143
144 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
145 [Steve Henson]
146
147 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
148 place for the (very old) "NETSCAPE" format certificates which are now
149 handled using new ASN1 code equivalents.
150 [Steve Henson]
151
152 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
153 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
154 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
155 [Nils Larsch]
156
157 *) Modify CRL distribution points extension code to print out previously
158 unsupported fields. Enhance extension setting code to allow setting of
159 all fields.
160 [Steve Henson]
161
162 *) Add print and set support for Issuing Distribution Point CRL extension.
163 [Steve Henson]
164
165 Changes between 0.9.8a and 0.9.8b [XX xxx XXXX]
166
167 *) Link in manifests for VC++ if needed.
168 [Austin Ziegler <halostatue@gmail.com>]
169
170 *) Update support for ECC-based TLS ciphersuites according to
171 draft-ietf-tls-ecc-12.txt with proposed changes (but without
172 TLS extensions, which are supported starting with the 0.9.9
173 branch, not in the OpenSSL 0.9.8 branch).
174 [Douglas Stebila]
175
176 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
177 opaque EVP_CIPHER_CTX handling.
178 [Steve Henson]
179
180 *) Fixes and enhancements to zlib compression code. We now only use
181 "zlib1.dll" and use the default __cdecl calling convention on Win32
182 to conform with the standards mentioned here:
183 http://www.zlib.net/DLL_FAQ.txt
184 Static zlib linking now works on Windows and the new --with-zlib-include
185 --with-zlib-lib options to Configure can be used to supply the location
186 of the headers and library. Gracefully handle case where zlib library
187 can't be loaded.
188 [Steve Henson]
189
190 *) Several fixes and enhancements to the OID generation code. The old code
191 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
192 handle numbers larger than ULONG_MAX, truncated printing and had a
193 non standard OBJ_obj2txt() behaviour.
194 [Steve Henson]
195
196 *) Add support for building of engines under engine/ as shared libraries
197 under VC++ build system.
198 [Steve Henson]
199
200 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
201 Hopefully, we will not see any false combination of paths any more.
202 [Richard Levitte]
203
204 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
205
206 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
207 (part of SSL_OP_ALL). This option used to disable the
208 countermeasure against man-in-the-middle protocol-version
209 rollback in the SSL 2.0 server implementation, which is a bad
210 idea. (CVE-2005-2969)
211
212 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
213 for Information Security, National Institute of Advanced Industrial
214 Science and Technology [AIST], Japan)]
215
216 *) Add two function to clear and return the verify parameter flags.
217 [Steve Henson]
218
219 *) Keep cipherlists sorted in the source instead of sorting them at
220 runtime, thus removing the need for a lock.
221 [Nils Larsch]
222
223 *) Avoid some small subgroup attacks in Diffie-Hellman.
224 [Nick Mathewson and Ben Laurie]
225
226 *) Add functions for well-known primes.
227 [Nick Mathewson]
228
229 *) Extended Windows CE support.
230 [Satoshi Nakamura and Andy Polyakov]
231
232 *) Initialize SSL_METHOD structures at compile time instead of during
233 runtime, thus removing the need for a lock.
234 [Steve Henson]
235
236 *) Make PKCS7_decrypt() work even if no certificate is supplied by
237 attempting to decrypt each encrypted key in turn. Add support to
238 smime utility.
239 [Steve Henson]
240
241 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
242
243 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
244 [Richard Levitte]
245
246 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
247 key into the same file any more.
248 [Richard Levitte]
249
250 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
251 [Andy Polyakov]
252
253 *) Add -utf8 command line and config file option to 'ca'.
254 [Stefan <stf@udoma.org]
255
256 *) Removed the macro des_crypt(), as it seems to conflict with some
257 libraries. Use DES_crypt().
258 [Richard Levitte]
259
260 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
261 involves renaming the source and generated shared-libs for
262 both. The engines will accept the corrected or legacy ids
263 ('ncipher' and '4758_cca' respectively) when binding. NB,
264 this only applies when building 'shared'.
265 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
266
267 *) Add attribute functions to EVP_PKEY structure. Modify
268 PKCS12_create() to recognize a CSP name attribute and
269 use it. Make -CSP option work again in pkcs12 utility.
270 [Steve Henson]
271
272 *) Add new functionality to the bn blinding code:
273 - automatic re-creation of the BN_BLINDING parameters after
274 a fixed number of uses (currently 32)
275 - add new function for parameter creation
276 - introduce flags to control the update behaviour of the
277 BN_BLINDING parameters
278 - hide BN_BLINDING structure
279 Add a second BN_BLINDING slot to the RSA structure to improve
280 performance when a single RSA object is shared among several
281 threads.
282 [Nils Larsch]
283
284 *) Add support for DTLS.
285 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
286
287 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
288 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
289 [Walter Goulet]
290
291 *) Remove buggy and incompletet DH cert support from
292 ssl/ssl_rsa.c and ssl/s3_both.c
293 [Nils Larsch]
294
295 *) Use SHA-1 instead of MD5 as the default digest algorithm for
296 the apps/openssl applications.
297 [Nils Larsch]
298
299 *) Compile clean with "-Wall -Wmissing-prototypes
300 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
301 DEBUG_SAFESTACK must also be set.
302 [Ben Laurie]
303
304 *) Change ./Configure so that certain algorithms can be disabled by default.
305 The new counterpiece to "no-xxx" is "enable-xxx".
306
307 The patented RC5 and MDC2 algorithms will now be disabled unless
308 "enable-rc5" and "enable-mdc2", respectively, are specified.
309
310 (IDEA remains enabled despite being patented. This is because IDEA
311 is frequently required for interoperability, and there is no license
312 fee for non-commercial use. As before, "no-idea" can be used to
313 avoid this algorithm.)
314
315 [Bodo Moeller]
316
317 *) Add processing of proxy certificates (see RFC 3820). This work was
318 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
319 EGEE (Enabling Grids for E-science in Europe).
320 [Richard Levitte]
321
322 *) RC4 performance overhaul on modern architectures/implementations, such
323 as Intel P4, IA-64 and AMD64.
324 [Andy Polyakov]
325
326 *) New utility extract-section.pl. This can be used specify an alternative
327 section number in a pod file instead of having to treat each file as
328 a separate case in Makefile. This can be done by adding two lines to the
329 pod file:
330
331 =for comment openssl_section:XXX
332
333 The blank line is mandatory.
334
335 [Steve Henson]
336
337 *) New arguments -certform, -keyform and -pass for s_client and s_server
338 to allow alternative format key and certificate files and passphrase
339 sources.
340 [Steve Henson]
341
342 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
343 update associated structures and add various utility functions.
344
345 Add new policy related verify parameters, include policy checking in
346 standard verify code. Enhance 'smime' application with extra parameters
347 to support policy checking and print out.
348 [Steve Henson]
349
350 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
351 Nehemiah processors. These extensions support AES encryption in hardware
352 as well as RNG (though RNG support is currently disabled).
353 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
354
355 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
356 [Geoff Thorpe]
357
358 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
359 [Andy Polyakov and a number of other people]
360
361 *) Improved PowerPC platform support. Most notably BIGNUM assembler
362 implementation contributed by IBM.
363 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
364
365 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
366 exponent rather than 'unsigned long'. There is a corresponding change to
367 the new 'rsa_keygen' element of the RSA_METHOD structure.
368 [Jelte Jansen, Geoff Thorpe]
369
370 *) Functionality for creating the initial serial number file is now
371 moved from CA.pl to the 'ca' utility with a new option -create_serial.
372
373 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
374 number file to 1, which is bound to cause problems. To avoid
375 the problems while respecting compatibility between different 0.9.7
376 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
377 CA.pl for serial number initialization. With the new release 0.9.8,
378 we can fix the problem directly in the 'ca' utility.)
379 [Steve Henson]
380
381 *) Reduced header interdepencies by declaring more opaque objects in
382 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
383 give fewer recursive includes, which could break lazy source code - so
384 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
385 developers should define this symbol when building and using openssl to
386 ensure they track the recommended behaviour, interfaces, [etc], but
387 backwards-compatible behaviour prevails when this isn't defined.
388 [Geoff Thorpe]
389
390 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
391 [Steve Henson]
392
393 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
394 This will generate a random key of the appropriate length based on the
395 cipher context. The EVP_CIPHER can provide its own random key generation
396 routine to support keys of a specific form. This is used in the des and
397 3des routines to generate a key of the correct parity. Update S/MIME
398 code to use new functions and hence generate correct parity DES keys.
399 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
400 valid (weak or incorrect parity).
401 [Steve Henson]
402
403 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
404 as looking them up. This is useful when the verified structure may contain
405 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
406 present unless the new PKCS7_NO_CRL flag is asserted.
407 [Steve Henson]
408
409 *) Extend ASN1 oid configuration module. It now additionally accepts the
410 syntax:
411
412 shortName = some long name, 1.2.3.4
413 [Steve Henson]
414
415 *) Reimplemented the BN_CTX implementation. There is now no more static
416 limitation on the number of variables it can handle nor the depth of the
417 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
418 information can now expand as required, and rather than having a single
419 static array of bignums, BN_CTX now uses a linked-list of such arrays
420 allowing it to expand on demand whilst maintaining the usefulness of
421 BN_CTX's "bundling".
422 [Geoff Thorpe]
423
424 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
425 to allow all RSA operations to function using a single BN_CTX.
426 [Geoff Thorpe]
427
428 *) Preliminary support for certificate policy evaluation and checking. This
429 is initially intended to pass the tests outlined in "Conformance Testing
430 of Relying Party Client Certificate Path Processing Logic" v1.07.
431 [Steve Henson]
432
433 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
434 remained unused and not that useful. A variety of other little bignum
435 tweaks and fixes have also been made continuing on from the audit (see
436 below).
437 [Geoff Thorpe]
438
439 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
440 associated ASN1, EVP and SSL functions and old ASN1 macros.
441 [Richard Levitte]
442
443 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
444 and this should never fail. So the return value from the use of
445 BN_set_word() (which can fail due to needless expansion) is now deprecated;
446 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
447 [Geoff Thorpe]
448
449 *) BN_CTX_get() should return zero-valued bignums, providing the same
450 initialised value as BN_new().
451 [Geoff Thorpe, suggested by Ulf Möller]
452
453 *) Support for inhibitAnyPolicy certificate extension.
454 [Steve Henson]
455
456 *) An audit of the BIGNUM code is underway, for which debugging code is
457 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
458 is considered valid when processing BIGNUMs, and causes execution to
459 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
460 further steps are taken to deliberately pollute unused data in BIGNUM
461 structures to try and expose faulty code further on. For now, openssl will
462 (in its default mode of operation) continue to tolerate the inconsistent
463 forms that it has tolerated in the past, but authors and packagers should
464 consider trying openssl and their own applications when compiled with
465 these debugging symbols defined. It will help highlight potential bugs in
466 their own code, and will improve the test coverage for OpenSSL itself. At
467 some point, these tighter rules will become openssl's default to improve
468 maintainability, though the assert()s and other overheads will remain only
469 in debugging configurations. See bn.h for more details.
470 [Geoff Thorpe, Nils Larsch, Ulf Möller]
471
472 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
473 that can only be obtained through BN_CTX_new() (which implicitly
474 initialises it). The presence of this function only made it possible
475 to overwrite an existing structure (and cause memory leaks).
476 [Geoff Thorpe]
477
478 *) Because of the callback-based approach for implementing LHASH as a
479 template type, lh_insert() adds opaque objects to hash-tables and
480 lh_doall() or lh_doall_arg() are typically used with a destructor callback
481 to clean up those corresponding objects before destroying the hash table
482 (and losing the object pointers). So some over-zealous constifications in
483 LHASH have been relaxed so that lh_insert() does not take (nor store) the
484 objects as "const" and the lh_doall[_arg] callback wrappers are not
485 prototyped to have "const" restrictions on the object pointers they are
486 given (and so aren't required to cast them away any more).
487 [Geoff Thorpe]
488
489 *) The tmdiff.h API was so ugly and minimal that our own timing utility
490 (speed) prefers to use its own implementation. The two implementations
491 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
492 its object type properly exposed (MS_TM) instead of casting to/from "char
493 *". This may still change yet if someone realises MS_TM and "ms_time_***"
494 aren't necessarily the greatest nomenclatures - but this is what was used
495 internally to the implementation so I've used that for now.
496 [Geoff Thorpe]
497
498 *) Ensure that deprecated functions do not get compiled when
499 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
500 the self-tests were still using deprecated key-generation functions so
501 these have been updated also.
502 [Geoff Thorpe]
503
504 *) Reorganise PKCS#7 code to separate the digest location functionality
505 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
506 New function PKCS7_set_digest() to set the digest type for PKCS#7
507 digestedData type. Add additional code to correctly generate the
508 digestedData type and add support for this type in PKCS7 initialization
509 functions.
510 [Steve Henson]
511
512 *) New function PKCS7_set0_type_other() this initializes a PKCS7
513 structure of type "other".
514 [Steve Henson]
515
516 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
517 sure the loop does correctly stop and breaking ("division by zero")
518 modulus operations are not performed. The (pre-generated) prime
519 table crypto/bn/bn_prime.h was already correct, but it could not be
520 re-generated on some platforms because of the "division by zero"
521 situation in the script.
522 [Ralf S. Engelschall]
523
524 *) Update support for ECC-based TLS ciphersuites according to
525 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
526 SHA-1 now is only used for "small" curves (where the
527 representation of a field element takes up to 24 bytes); for
528 larger curves, the field element resulting from ECDH is directly
529 used as premaster secret.
530 [Douglas Stebila (Sun Microsystems Laboratories)]
531
532 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
533 curve secp160r1 to the tests.
534 [Douglas Stebila (Sun Microsystems Laboratories)]
535
536 *) Add the possibility to load symbols globally with DSO.
537 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
538
539 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
540 control of the error stack.
541 [Richard Levitte]
542
543 *) Add support for STORE in ENGINE.
544 [Richard Levitte]
545
546 *) Add the STORE type. The intention is to provide a common interface
547 to certificate and key stores, be they simple file-based stores, or
548 HSM-type store, or LDAP stores, or...
549 NOTE: The code is currently UNTESTED and isn't really used anywhere.
550 [Richard Levitte]
551
552 *) Add a generic structure called OPENSSL_ITEM. This can be used to
553 pass a list of arguments to any function as well as provide a way
554 for a function to pass data back to the caller.
555 [Richard Levitte]
556
557 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
558 works like BUF_strdup() but can be used to duplicate a portion of
559 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
560 a memory area.
561 [Richard Levitte]
562
563 *) Add the function sk_find_ex() which works like sk_find(), but will
564 return an index to an element even if an exact match couldn't be
565 found. The index is guaranteed to point at the element where the
566 searched-for key would be inserted to preserve sorting order.
567 [Richard Levitte]
568
569 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
570 takes an extra flags argument for optional functionality. Currently,
571 the following flags are defined:
572
573 OBJ_BSEARCH_VALUE_ON_NOMATCH
574 This one gets OBJ_bsearch_ex() to return a pointer to the first
575 element where the comparing function returns a negative or zero
576 number.
577
578 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
579 This one gets OBJ_bsearch_ex() to return a pointer to the first
580 element where the comparing function returns zero. This is useful
581 if there are more than one element where the comparing function
582 returns zero.
583 [Richard Levitte]
584
585 *) Make it possible to create self-signed certificates with 'openssl ca'
586 in such a way that the self-signed certificate becomes part of the
587 CA database and uses the same mechanisms for serial number generation
588 as all other certificate signing. The new flag '-selfsign' enables
589 this functionality. Adapt CA.sh and CA.pl.in.
590 [Richard Levitte]
591
592 *) Add functionality to check the public key of a certificate request
593 against a given private. This is useful to check that a certificate
594 request can be signed by that key (self-signing).
595 [Richard Levitte]
596
597 *) Make it possible to have multiple active certificates with the same
598 subject in the CA index file. This is done only if the keyword
599 'unique_subject' is set to 'no' in the main CA section (default
600 if 'CA_default') of the configuration file. The value is saved
601 with the database itself in a separate index attribute file,
602 named like the index file with '.attr' appended to the name.
603 [Richard Levitte]
604
605 *) Generate muti valued AVAs using '+' notation in config files for
606 req and dirName.
607 [Steve Henson]
608
609 *) Support for nameConstraints certificate extension.
610 [Steve Henson]
611
612 *) Support for policyConstraints certificate extension.
613 [Steve Henson]
614
615 *) Support for policyMappings certificate extension.
616 [Steve Henson]
617
618 *) Make sure the default DSA_METHOD implementation only uses its
619 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
620 and change its own handlers to be NULL so as to remove unnecessary
621 indirection. This lets alternative implementations fallback to the
622 default implementation more easily.
623 [Geoff Thorpe]
624
625 *) Support for directoryName in GeneralName related extensions
626 in config files.
627 [Steve Henson]
628
629 *) Make it possible to link applications using Makefile.shared.
630 Make that possible even when linking against static libraries!
631 [Richard Levitte]
632
633 *) Support for single pass processing for S/MIME signing. This now
634 means that S/MIME signing can be done from a pipe, in addition
635 cleartext signing (multipart/signed type) is effectively streaming
636 and the signed data does not need to be all held in memory.
637
638 This is done with a new flag PKCS7_STREAM. When this flag is set
639 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
640 is done after the data is output (and digests calculated) in
641 SMIME_write_PKCS7().
642 [Steve Henson]
643
644 *) Add full support for -rpath/-R, both in shared libraries and
645 applications, at least on the platforms where it's known how
646 to do it.
647 [Richard Levitte]
648
649 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
650 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
651 will now compute a table of multiples of the generator that
652 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
653 faster (notably in the case of a single point multiplication,
654 scalar * generator).
655 [Nils Larsch, Bodo Moeller]
656
657 *) IPv6 support for certificate extensions. The various extensions
658 which use the IP:a.b.c.d can now take IPv6 addresses using the
659 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
660 correctly.
661 [Steve Henson]
662
663 *) Added an ENGINE that implements RSA by performing private key
664 exponentiations with the GMP library. The conversions to and from
665 GMP's mpz_t format aren't optimised nor are any montgomery forms
666 cached, and on x86 it appears OpenSSL's own performance has caught up.
667 However there are likely to be other architectures where GMP could
668 provide a boost. This ENGINE is not built in by default, but it can be
669 specified at Configure time and should be accompanied by the necessary
670 linker additions, eg;
671 ./config -DOPENSSL_USE_GMP -lgmp
672 [Geoff Thorpe]
673
674 *) "openssl engine" will not display ENGINE/DSO load failure errors when
675 testing availability of engines with "-t" - the old behaviour is
676 produced by increasing the feature's verbosity with "-tt".
677 [Geoff Thorpe]
678
679 *) ECDSA routines: under certain error conditions uninitialized BN objects
680 could be freed. Solution: make sure initialization is performed early
681 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
682 via PR#459)
683 [Lutz Jaenicke]
684
685 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
686 and DH_METHOD (eg. by ENGINE implementations) to override the normal
687 software implementations. For DSA and DH, parameter generation can
688 also be overriden by providing the appropriate method callbacks.
689 [Geoff Thorpe]
690
691 *) Change the "progress" mechanism used in key-generation and
692 primality testing to functions that take a new BN_GENCB pointer in
693 place of callback/argument pairs. The new API functions have "_ex"
694 postfixes and the older functions are reimplemented as wrappers for
695 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
696 declarations of the old functions to help (graceful) attempts to
697 migrate to the new functions. Also, the new key-generation API
698 functions operate on a caller-supplied key-structure and return
699 success/failure rather than returning a key or NULL - this is to
700 help make "keygen" another member function of RSA_METHOD etc.
701
702 Example for using the new callback interface:
703
704 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
705 void *my_arg = ...;
706 BN_GENCB my_cb;
707
708 BN_GENCB_set(&my_cb, my_callback, my_arg);
709
710 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
711 /* For the meaning of a, b in calls to my_callback(), see the
712 * documentation of the function that calls the callback.
713 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
714 * my_callback should return 1 if it wants BN_is_prime_ex()
715 * to continue, or 0 to stop.
716 */
717
718 [Geoff Thorpe]
719
720 *) Change the ZLIB compression method to be stateful, and make it
721 available to TLS with the number defined in
722 draft-ietf-tls-compression-04.txt.
723 [Richard Levitte]
724
725 *) Add the ASN.1 structures and functions for CertificatePair, which
726 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
727
728 CertificatePair ::= SEQUENCE {
729 forward [0] Certificate OPTIONAL,
730 reverse [1] Certificate OPTIONAL,
731 -- at least one of the pair shall be present -- }
732
733 Also implement the PEM functions to read and write certificate
734 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
735
736 This needed to be defined, mostly for the sake of the LDAP
737 attribute crossCertificatePair, but may prove useful elsewhere as
738 well.
739 [Richard Levitte]
740
741 *) Make it possible to inhibit symlinking of shared libraries in
742 Makefile.shared, for Cygwin's sake.
743 [Richard Levitte]
744
745 *) Extend the BIGNUM API by creating a function
746 void BN_set_negative(BIGNUM *a, int neg);
747 and a macro that behave like
748 int BN_is_negative(const BIGNUM *a);
749
750 to avoid the need to access 'a->neg' directly in applications.
751 [Nils Larsch]
752
753 *) Implement fast modular reduction for pseudo-Mersenne primes
754 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
755 EC_GROUP_new_curve_GFp() will now automatically use this
756 if applicable.
757 [Nils Larsch <nla@trustcenter.de>]
758
759 *) Add new lock type (CRYPTO_LOCK_BN).
760 [Bodo Moeller]
761
762 *) Change the ENGINE framework to automatically load engines
763 dynamically from specific directories unless they could be
764 found to already be built in or loaded. Move all the
765 current engines except for the cryptodev one to a new
766 directory engines/.
767 The engines in engines/ are built as shared libraries if
768 the "shared" options was given to ./Configure or ./config.
769 Otherwise, they are inserted in libcrypto.a.
770 /usr/local/ssl/engines is the default directory for dynamic
771 engines, but that can be overriden at configure time through
772 the usual use of --prefix and/or --openssldir, and at run
773 time with the environment variable OPENSSL_ENGINES.
774 [Geoff Thorpe and Richard Levitte]
775
776 *) Add Makefile.shared, a helper makefile to build shared
777 libraries. Addapt Makefile.org.
778 [Richard Levitte]
779
780 *) Add version info to Win32 DLLs.
781 [Peter 'Luna' Runestig" <peter@runestig.com>]
782
783 *) Add new 'medium level' PKCS#12 API. Certificates and keys
784 can be added using this API to created arbitrary PKCS#12
785 files while avoiding the low level API.
786
787 New options to PKCS12_create(), key or cert can be NULL and
788 will then be omitted from the output file. The encryption
789 algorithm NIDs can be set to -1 for no encryption, the mac
790 iteration count can be set to 0 to omit the mac.
791
792 Enhance pkcs12 utility by making the -nokeys and -nocerts
793 options work when creating a PKCS#12 file. New option -nomac
794 to omit the mac, NONE can be set for an encryption algorithm.
795 New code is modified to use the enhanced PKCS12_create()
796 instead of the low level API.
797 [Steve Henson]
798
799 *) Extend ASN1 encoder to support indefinite length constructed
800 encoding. This can output sequences tags and octet strings in
801 this form. Modify pk7_asn1.c to support indefinite length
802 encoding. This is experimental and needs additional code to
803 be useful, such as an ASN1 bio and some enhanced streaming
804 PKCS#7 code.
805
806 Extend template encode functionality so that tagging is passed
807 down to the template encoder.
808 [Steve Henson]
809
810 *) Let 'openssl req' fail if an argument to '-newkey' is not
811 recognized instead of using RSA as a default.
812 [Bodo Moeller]
813
814 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
815 As these are not official, they are not included in "ALL";
816 the "ECCdraft" ciphersuite group alias can be used to select them.
817 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
818
819 *) Add ECDH engine support.
820 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
821
822 *) Add ECDH in new directory crypto/ecdh/.
823 [Douglas Stebila (Sun Microsystems Laboratories)]
824
825 *) Let BN_rand_range() abort with an error after 100 iterations
826 without success (which indicates a broken PRNG).
827 [Bodo Moeller]
828
829 *) Change BN_mod_sqrt() so that it verifies that the input value
830 is really the square of the return value. (Previously,
831 BN_mod_sqrt would show GIGO behaviour.)
832 [Bodo Moeller]
833
834 *) Add named elliptic curves over binary fields from X9.62, SECG,
835 and WAP/WTLS; add OIDs that were still missing.
836
837 [Sheueling Chang Shantz and Douglas Stebila
838 (Sun Microsystems Laboratories)]
839
840 *) Extend the EC library for elliptic curves over binary fields
841 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
842 New EC_METHOD:
843
844 EC_GF2m_simple_method
845
846 New API functions:
847
848 EC_GROUP_new_curve_GF2m
849 EC_GROUP_set_curve_GF2m
850 EC_GROUP_get_curve_GF2m
851 EC_POINT_set_affine_coordinates_GF2m
852 EC_POINT_get_affine_coordinates_GF2m
853 EC_POINT_set_compressed_coordinates_GF2m
854
855 Point compression for binary fields is disabled by default for
856 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
857 enable it).
858
859 As binary polynomials are represented as BIGNUMs, various members
860 of the EC_GROUP and EC_POINT data structures can be shared
861 between the implementations for prime fields and binary fields;
862 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
863 are essentially identical to their ..._GFp counterparts.
864 (For simplicity, the '..._GFp' prefix has been dropped from
865 various internal method names.)
866
867 An internal 'field_div' method (similar to 'field_mul' and
868 'field_sqr') has been added; this is used only for binary fields.
869
870 [Sheueling Chang Shantz and Douglas Stebila
871 (Sun Microsystems Laboratories)]
872
873 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
874 through methods ('mul', 'precompute_mult').
875
876 The generic implementations (now internally called 'ec_wNAF_mul'
877 and 'ec_wNAF_precomputed_mult') remain the default if these
878 methods are undefined.
879
880 [Sheueling Chang Shantz and Douglas Stebila
881 (Sun Microsystems Laboratories)]
882
883 *) New function EC_GROUP_get_degree, which is defined through
884 EC_METHOD. For curves over prime fields, this returns the bit
885 length of the modulus.
886
887 [Sheueling Chang Shantz and Douglas Stebila
888 (Sun Microsystems Laboratories)]
889
890 *) New functions EC_GROUP_dup, EC_POINT_dup.
891 (These simply call ..._new and ..._copy).
892
893 [Sheueling Chang Shantz and Douglas Stebila
894 (Sun Microsystems Laboratories)]
895
896 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
897 Polynomials are represented as BIGNUMs (where the sign bit is not
898 used) in the following functions [macros]:
899
900 BN_GF2m_add
901 BN_GF2m_sub [= BN_GF2m_add]
902 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
903 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
904 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
905 BN_GF2m_mod_inv
906 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
907 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
908 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
909 BN_GF2m_cmp [= BN_ucmp]
910
911 (Note that only the 'mod' functions are actually for fields GF(2^m).
912 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
913
914 For some functions, an the irreducible polynomial defining a
915 field can be given as an 'unsigned int[]' with strictly
916 decreasing elements giving the indices of those bits that are set;
917 i.e., p[] represents the polynomial
918 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
919 where
920 p[0] > p[1] > ... > p[k] = 0.
921 This applies to the following functions:
922
923 BN_GF2m_mod_arr
924 BN_GF2m_mod_mul_arr
925 BN_GF2m_mod_sqr_arr
926 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
927 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
928 BN_GF2m_mod_exp_arr
929 BN_GF2m_mod_sqrt_arr
930 BN_GF2m_mod_solve_quad_arr
931 BN_GF2m_poly2arr
932 BN_GF2m_arr2poly
933
934 Conversion can be performed by the following functions:
935
936 BN_GF2m_poly2arr
937 BN_GF2m_arr2poly
938
939 bntest.c has additional tests for binary polynomial arithmetic.
940
941 Two implementations for BN_GF2m_mod_div() are available.
942 The default algorithm simply uses BN_GF2m_mod_inv() and
943 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
944 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
945 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
946
947 [Sheueling Chang Shantz and Douglas Stebila
948 (Sun Microsystems Laboratories)]
949
950 *) Add new error code 'ERR_R_DISABLED' that can be used when some
951 functionality is disabled at compile-time.
952 [Douglas Stebila <douglas.stebila@sun.com>]
953
954 *) Change default behaviour of 'openssl asn1parse' so that more
955 information is visible when viewing, e.g., a certificate:
956
957 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
958 mode the content of non-printable OCTET STRINGs is output in a
959 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
960 avoid the appearance of a printable string.
961 [Nils Larsch <nla@trustcenter.de>]
962
963 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
964 functions
965 EC_GROUP_set_asn1_flag()
966 EC_GROUP_get_asn1_flag()
967 EC_GROUP_set_point_conversion_form()
968 EC_GROUP_get_point_conversion_form()
969 These control ASN1 encoding details:
970 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
971 has been set to OPENSSL_EC_NAMED_CURVE.
972 - Points are encoded in uncompressed form by default; options for
973 asn1_for are as for point2oct, namely
974 POINT_CONVERSION_COMPRESSED
975 POINT_CONVERSION_UNCOMPRESSED
976 POINT_CONVERSION_HYBRID
977
978 Also add 'seed' and 'seed_len' members to EC_GROUP with access
979 functions
980 EC_GROUP_set_seed()
981 EC_GROUP_get0_seed()
982 EC_GROUP_get_seed_len()
983 This is used only for ASN1 purposes (so far).
984 [Nils Larsch <nla@trustcenter.de>]
985
986 *) Add 'field_type' member to EC_METHOD, which holds the NID
987 of the appropriate field type OID. The new function
988 EC_METHOD_get_field_type() returns this value.
989 [Nils Larsch <nla@trustcenter.de>]
990
991 *) Add functions
992 EC_POINT_point2bn()
993 EC_POINT_bn2point()
994 EC_POINT_point2hex()
995 EC_POINT_hex2point()
996 providing useful interfaces to EC_POINT_point2oct() and
997 EC_POINT_oct2point().
998 [Nils Larsch <nla@trustcenter.de>]
999
1000 *) Change internals of the EC library so that the functions
1001 EC_GROUP_set_generator()
1002 EC_GROUP_get_generator()
1003 EC_GROUP_get_order()
1004 EC_GROUP_get_cofactor()
1005 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1006 to methods, which would lead to unnecessary code duplication when
1007 adding different types of curves.
1008 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1009
1010 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1011 arithmetic, and such that modified wNAFs are generated
1012 (which avoid length expansion in many cases).
1013 [Bodo Moeller]
1014
1015 *) Add a function EC_GROUP_check_discriminant() (defined via
1016 EC_METHOD) that verifies that the curve discriminant is non-zero.
1017
1018 Add a function EC_GROUP_check() that makes some sanity tests
1019 on a EC_GROUP, its generator and order. This includes
1020 EC_GROUP_check_discriminant().
1021 [Nils Larsch <nla@trustcenter.de>]
1022
1023 *) Add ECDSA in new directory crypto/ecdsa/.
1024
1025 Add applications 'openssl ecparam' and 'openssl ecdsa'
1026 (these are based on 'openssl dsaparam' and 'openssl dsa').
1027
1028 ECDSA support is also included in various other files across the
1029 library. Most notably,
1030 - 'openssl req' now has a '-newkey ecdsa:file' option;
1031 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1032 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1033 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1034 them suitable for ECDSA where domain parameters must be
1035 extracted before the specific public key;
1036 - ECDSA engine support has been added.
1037 [Nils Larsch <nla@trustcenter.de>]
1038
1039 *) Include some named elliptic curves, and add OIDs from X9.62,
1040 SECG, and WAP/WTLS. Each curve can be obtained from the new
1041 function
1042 EC_GROUP_new_by_curve_name(),
1043 and the list of available named curves can be obtained with
1044 EC_get_builtin_curves().
1045 Also add a 'curve_name' member to EC_GROUP objects, which can be
1046 accessed via
1047 EC_GROUP_set_curve_name()
1048 EC_GROUP_get_curve_name()
1049 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1050
1051 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1052 was actually never needed) and in BN_mul(). The removal in BN_mul()
1053 required a small change in bn_mul_part_recursive() and the addition
1054 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1055 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1056 bn_sub_words() and bn_add_words() except they take arrays with
1057 differing sizes.
1058 [Richard Levitte]
1059
1060 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1061
1062 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1063 The value now differs depending on if you build for FIPS or not.
1064 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1065 safely run with a non-FIPSed libcrypto, as it may crash because of
1066 the difference induced by this change.
1067 [Andy Polyakov]
1068
1069 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1070
1071 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1072 (part of SSL_OP_ALL). This option used to disable the
1073 countermeasure against man-in-the-middle protocol-version
1074 rollback in the SSL 2.0 server implementation, which is a bad
1075 idea. (CVE-2005-2969)
1076
1077 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1078 for Information Security, National Institute of Advanced Industrial
1079 Science and Technology [AIST], Japan)]
1080
1081 *) Minimal support for X9.31 signatures and PSS padding modes. This is
1082 mainly for FIPS compliance and not fully integrated at this stage.
1083 [Steve Henson]
1084
1085 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
1086 the exponentiation using a fixed-length exponent. (Otherwise,
1087 the information leaked through timing could expose the secret key
1088 after many signatures; cf. Bleichenbacher's attack on DSA with
1089 biased k.)
1090 [Bodo Moeller]
1091
1092 *) Make a new fixed-window mod_exp implementation the default for
1093 RSA, DSA, and DH private-key operations so that the sequence of
1094 squares and multiplies and the memory access pattern are
1095 independent of the particular secret key. This will mitigate
1096 cache-timing and potential related attacks.
1097
1098 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
1099 and this is automatically used by BN_mod_exp_mont() if the new flag
1100 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
1101 will use this BN flag for private exponents unless the flag
1102 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
1103 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
1104
1105 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
1106
1107 *) Change the client implementation for SSLv23_method() and
1108 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
1109 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
1110 (Previously, the SSL 2.0 backwards compatible Client Hello
1111 message format would be used even with SSL_OP_NO_SSLv2.)
1112 [Bodo Moeller]
1113
1114 *) Add support for smime-type MIME parameter in S/MIME messages which some
1115 clients need.
1116 [Steve Henson]
1117
1118 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
1119 a threadsafe manner. Modify rsa code to use new function and add calls
1120 to dsa and dh code (which had race conditions before).
1121 [Steve Henson]
1122
1123 *) Include the fixed error library code in the C error file definitions
1124 instead of fixing them up at runtime. This keeps the error code
1125 structures constant.
1126 [Steve Henson]
1127
1128 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
1129
1130 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
1131 OpenSSL 0.9.8.]
1132
1133 *) Fixes for newer kerberos headers. NB: the casts are needed because
1134 the 'length' field is signed on one version and unsigned on another
1135 with no (?) obvious way to tell the difference, without these VC++
1136 complains. Also the "definition" of FAR (blank) is no longer included
1137 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
1138 some needed definitions.
1139 [Steve Henson]
1140
1141 *) Undo Cygwin change.
1142 [Ulf Möller]
1143
1144 *) Added support for proxy certificates according to RFC 3820.
1145 Because they may be a security thread to unaware applications,
1146 they must be explicitely allowed in run-time. See
1147 docs/HOWTO/proxy_certificates.txt for further information.
1148 [Richard Levitte]
1149
1150 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
1151
1152 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
1153 server and client random values. Previously
1154 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
1155 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
1156
1157 This change has negligible security impact because:
1158
1159 1. Server and client random values still have 24 bytes of pseudo random
1160 data.
1161
1162 2. Server and client random values are sent in the clear in the initial
1163 handshake.
1164
1165 3. The master secret is derived using the premaster secret (48 bytes in
1166 size for static RSA ciphersuites) as well as client server and random
1167 values.
1168
1169 The OpenSSL team would like to thank the UK NISCC for bringing this issue
1170 to our attention.
1171
1172 [Stephen Henson, reported by UK NISCC]
1173
1174 *) Use Windows randomness collection on Cygwin.
1175 [Ulf Möller]
1176
1177 *) Fix hang in EGD/PRNGD query when communication socket is closed
1178 prematurely by EGD/PRNGD.
1179 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
1180
1181 *) Prompt for pass phrases when appropriate for PKCS12 input format.
1182 [Steve Henson]
1183
1184 *) Back-port of selected performance improvements from development
1185 branch, as well as improved support for PowerPC platforms.
1186 [Andy Polyakov]
1187
1188 *) Add lots of checks for memory allocation failure, error codes to indicate
1189 failure and freeing up memory if a failure occurs.
1190 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
1191
1192 *) Add new -passin argument to dgst.
1193 [Steve Henson]
1194
1195 *) Perform some character comparisons of different types in X509_NAME_cmp:
1196 this is needed for some certificates that reencode DNs into UTF8Strings
1197 (in violation of RFC3280) and can't or wont issue name rollover
1198 certificates.
1199 [Steve Henson]
1200
1201 *) Make an explicit check during certificate validation to see that
1202 the CA setting in each certificate on the chain is correct. As a
1203 side effect always do the following basic checks on extensions,
1204 not just when there's an associated purpose to the check:
1205
1206 - if there is an unhandled critical extension (unless the user
1207 has chosen to ignore this fault)
1208 - if the path length has been exceeded (if one is set at all)
1209 - that certain extensions fit the associated purpose (if one has
1210 been given)
1211 [Richard Levitte]
1212
1213 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
1214
1215 *) Avoid a race condition when CRLs are checked in a multi threaded
1216 environment. This would happen due to the reordering of the revoked
1217 entries during signature checking and serial number lookup. Now the
1218 encoding is cached and the serial number sort performed under a lock.
1219 Add new STACK function sk_is_sorted().
1220 [Steve Henson]
1221
1222 *) Add Delta CRL to the extension code.
1223 [Steve Henson]
1224
1225 *) Various fixes to s3_pkt.c so alerts are sent properly.
1226 [David Holmes <d.holmes@f5.com>]
1227
1228 *) Reduce the chances of duplicate issuer name and serial numbers (in
1229 violation of RFC3280) using the OpenSSL certificate creation utilities.
1230 This is done by creating a random 64 bit value for the initial serial
1231 number when a serial number file is created or when a self signed
1232 certificate is created using 'openssl req -x509'. The initial serial
1233 number file is created using 'openssl x509 -next_serial' in CA.pl
1234 rather than being initialized to 1.
1235 [Steve Henson]
1236
1237 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
1238
1239 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
1240 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
1241 [Joe Orton, Steve Henson]
1242
1243 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
1244 (CVE-2004-0112)
1245 [Joe Orton, Steve Henson]
1246
1247 *) Make it possible to have multiple active certificates with the same
1248 subject in the CA index file. This is done only if the keyword
1249 'unique_subject' is set to 'no' in the main CA section (default
1250 if 'CA_default') of the configuration file. The value is saved
1251 with the database itself in a separate index attribute file,
1252 named like the index file with '.attr' appended to the name.
1253 [Richard Levitte]
1254
1255 *) X509 verify fixes. Disable broken certificate workarounds when
1256 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
1257 keyUsage extension present. Don't accept CRLs with unhandled critical
1258 extensions: since verify currently doesn't process CRL extensions this
1259 rejects a CRL with *any* critical extensions. Add new verify error codes
1260 for these cases.
1261 [Steve Henson]
1262
1263 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
1264 A clarification of RFC2560 will require the use of OCTET STRINGs and
1265 some implementations cannot handle the current raw format. Since OpenSSL
1266 copies and compares OCSP nonces as opaque blobs without any attempt at
1267 parsing them this should not create any compatibility issues.
1268 [Steve Henson]
1269
1270 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
1271 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
1272 this HMAC (and other) operations are several times slower than OpenSSL
1273 < 0.9.7.
1274 [Steve Henson]
1275
1276 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
1277 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1278
1279 *) Use the correct content when signing type "other".
1280 [Steve Henson]
1281
1282 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
1283
1284 *) Fix various bugs revealed by running the NISCC test suite:
1285
1286 Stop out of bounds reads in the ASN1 code when presented with
1287 invalid tags (CVE-2003-0543 and CVE-2003-0544).
1288
1289 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
1290
1291 If verify callback ignores invalid public key errors don't try to check
1292 certificate signature with the NULL public key.
1293
1294 [Steve Henson]
1295
1296 *) New -ignore_err option in ocsp application to stop the server
1297 exiting on the first error in a request.
1298 [Steve Henson]
1299
1300 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
1301 if the server requested one: as stated in TLS 1.0 and SSL 3.0
1302 specifications.
1303 [Steve Henson]
1304
1305 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
1306 extra data after the compression methods not only for TLS 1.0
1307 but also for SSL 3.0 (as required by the specification).
1308 [Bodo Moeller; problem pointed out by Matthias Loepfe]
1309
1310 *) Change X509_certificate_type() to mark the key as exported/exportable
1311 when it's 512 *bits* long, not 512 bytes.
1312 [Richard Levitte]
1313
1314 *) Change AES_cbc_encrypt() so it outputs exact multiple of
1315 blocks during encryption.
1316 [Richard Levitte]
1317
1318 *) Various fixes to base64 BIO and non blocking I/O. On write
1319 flushes were not handled properly if the BIO retried. On read
1320 data was not being buffered properly and had various logic bugs.
1321 This also affects blocking I/O when the data being decoded is a
1322 certain size.
1323 [Steve Henson]
1324
1325 *) Various S/MIME bugfixes and compatibility changes:
1326 output correct application/pkcs7 MIME type if
1327 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
1328 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
1329 of files as .eml work). Correctly handle very long lines in MIME
1330 parser.
1331 [Steve Henson]
1332
1333 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
1334
1335 *) Countermeasure against the Klima-Pokorny-Rosa extension of
1336 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
1337 a protocol version number mismatch like a decryption error
1338 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
1339 [Bodo Moeller]
1340
1341 *) Turn on RSA blinding by default in the default implementation
1342 to avoid a timing attack. Applications that don't want it can call
1343 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
1344 They would be ill-advised to do so in most cases.
1345 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
1346
1347 *) Change RSA blinding code so that it works when the PRNG is not
1348 seeded (in this case, the secret RSA exponent is abused as
1349 an unpredictable seed -- if it is not unpredictable, there
1350 is no point in blinding anyway). Make RSA blinding thread-safe
1351 by remembering the creator's thread ID in rsa->blinding and
1352 having all other threads use local one-time blinding factors
1353 (this requires more computation than sharing rsa->blinding, but
1354 avoids excessive locking; and if an RSA object is not shared
1355 between threads, blinding will still be very fast).
1356 [Bodo Moeller]
1357
1358 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
1359 ENGINE as defaults for all supported algorithms irrespective of
1360 the 'flags' parameter. 'flags' is now honoured, so applications
1361 should make sure they are passing it correctly.
1362 [Geoff Thorpe]
1363
1364 *) Target "mingw" now allows native Windows code to be generated in
1365 the Cygwin environment as well as with the MinGW compiler.
1366 [Ulf Moeller]
1367
1368 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
1369
1370 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
1371 via timing by performing a MAC computation even if incorrrect
1372 block cipher padding has been found. This is a countermeasure
1373 against active attacks where the attacker has to distinguish
1374 between bad padding and a MAC verification error. (CVE-2003-0078)
1375
1376 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
1377 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
1378 Martin Vuagnoux (EPFL, Ilion)]
1379
1380 *) Make the no-err option work as intended. The intention with no-err
1381 is not to have the whole error stack handling routines removed from
1382 libcrypto, it's only intended to remove all the function name and
1383 reason texts, thereby removing some of the footprint that may not
1384 be interesting if those errors aren't displayed anyway.
1385
1386 NOTE: it's still possible for any application or module to have it's
1387 own set of error texts inserted. The routines are there, just not
1388 used by default when no-err is given.
1389 [Richard Levitte]
1390
1391 *) Add support for FreeBSD on IA64.
1392 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
1393
1394 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
1395 Kerberos function mit_des_cbc_cksum(). Before this change,
1396 the value returned by DES_cbc_cksum() was like the one from
1397 mit_des_cbc_cksum(), except the bytes were swapped.
1398 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
1399
1400 *) Allow an application to disable the automatic SSL chain building.
1401 Before this a rather primitive chain build was always performed in
1402 ssl3_output_cert_chain(): an application had no way to send the
1403 correct chain if the automatic operation produced an incorrect result.
1404
1405 Now the chain builder is disabled if either:
1406
1407 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
1408
1409 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
1410
1411 The reasoning behind this is that an application would not want the
1412 auto chain building to take place if extra chain certificates are
1413 present and it might also want a means of sending no additional
1414 certificates (for example the chain has two certificates and the
1415 root is omitted).
1416 [Steve Henson]
1417
1418 *) Add the possibility to build without the ENGINE framework.
1419 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1420
1421 *) Under Win32 gmtime() can return NULL: check return value in
1422 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
1423 [Steve Henson]
1424
1425 *) DSA routines: under certain error conditions uninitialized BN objects
1426 could be freed. Solution: make sure initialization is performed early
1427 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
1428 Nils Larsch <nla@trustcenter.de> via PR#459)
1429 [Lutz Jaenicke]
1430
1431 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
1432 checked on reconnect on the client side, therefore session resumption
1433 could still fail with a "ssl session id is different" error. This
1434 behaviour is masked when SSL_OP_ALL is used due to
1435 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
1436 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1437 followup to PR #377.
1438 [Lutz Jaenicke]
1439
1440 *) IA-32 assembler support enhancements: unified ELF targets, support
1441 for SCO/Caldera platforms, fix for Cygwin shared build.
1442 [Andy Polyakov]
1443
1444 *) Add support for FreeBSD on sparc64. As a consequence, support for
1445 FreeBSD on non-x86 processors is separate from x86 processors on
1446 the config script, much like the NetBSD support.
1447 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
1448
1449 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
1450
1451 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
1452 OpenSSL 0.9.7.]
1453
1454 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
1455 code (06) was taken as the first octet of the session ID and the last
1456 octet was ignored consequently. As a result SSLv2 client side session
1457 caching could not have worked due to the session ID mismatch between
1458 client and server.
1459 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
1460 PR #377.
1461 [Lutz Jaenicke]
1462
1463 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
1464 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
1465 removed entirely.
1466 [Richard Levitte]
1467
1468 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
1469 seems that in spite of existing for more than a year, many application
1470 author have done nothing to provide the necessary callbacks, which
1471 means that this particular engine will not work properly anywhere.
1472 This is a very unfortunate situation which forces us, in the name
1473 of usability, to give the hw_ncipher.c a static lock, which is part
1474 of libcrypto.
1475 NOTE: This is for the 0.9.7 series ONLY. This hack will never
1476 appear in 0.9.8 or later. We EXPECT application authors to have
1477 dealt properly with this when 0.9.8 is released (unless we actually
1478 make such changes in the libcrypto locking code that changes will
1479 have to be made anyway).
1480 [Richard Levitte]
1481
1482 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
1483 octets have been read, EOF or an error occurs. Without this change
1484 some truncated ASN1 structures will not produce an error.
1485 [Steve Henson]
1486
1487 *) Disable Heimdal support, since it hasn't been fully implemented.
1488 Still give the possibility to force the use of Heimdal, but with
1489 warnings and a request that patches get sent to openssl-dev.
1490 [Richard Levitte]
1491
1492 *) Add the VC-CE target, introduce the WINCE sysname, and add
1493 INSTALL.WCE and appropriate conditionals to make it build.
1494 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
1495
1496 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
1497 cygssl-x.y.z.dll, where x, y and z are the major, minor and
1498 edit numbers of the version.
1499 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
1500
1501 *) Introduce safe string copy and catenation functions
1502 (BUF_strlcpy() and BUF_strlcat()).
1503 [Ben Laurie (CHATS) and Richard Levitte]
1504
1505 *) Avoid using fixed-size buffers for one-line DNs.
1506 [Ben Laurie (CHATS)]
1507
1508 *) Add BUF_MEM_grow_clean() to avoid information leakage when
1509 resizing buffers containing secrets, and use where appropriate.
1510 [Ben Laurie (CHATS)]
1511
1512 *) Avoid using fixed size buffers for configuration file location.
1513 [Ben Laurie (CHATS)]
1514
1515 *) Avoid filename truncation for various CA files.
1516 [Ben Laurie (CHATS)]
1517
1518 *) Use sizeof in preference to magic numbers.
1519 [Ben Laurie (CHATS)]
1520
1521 *) Avoid filename truncation in cert requests.
1522 [Ben Laurie (CHATS)]
1523
1524 *) Add assertions to check for (supposedly impossible) buffer
1525 overflows.
1526 [Ben Laurie (CHATS)]
1527
1528 *) Don't cache truncated DNS entries in the local cache (this could
1529 potentially lead to a spoofing attack).
1530 [Ben Laurie (CHATS)]
1531
1532 *) Fix various buffers to be large enough for hex/decimal
1533 representations in a platform independent manner.
1534 [Ben Laurie (CHATS)]
1535
1536 *) Add CRYPTO_realloc_clean() to avoid information leakage when
1537 resizing buffers containing secrets, and use where appropriate.
1538 [Ben Laurie (CHATS)]
1539
1540 *) Add BIO_indent() to avoid much slightly worrying code to do
1541 indents.
1542 [Ben Laurie (CHATS)]
1543
1544 *) Convert sprintf()/BIO_puts() to BIO_printf().
1545 [Ben Laurie (CHATS)]
1546
1547 *) buffer_gets() could terminate with the buffer only half
1548 full. Fixed.
1549 [Ben Laurie (CHATS)]
1550
1551 *) Add assertions to prevent user-supplied crypto functions from
1552 overflowing internal buffers by having large block sizes, etc.
1553 [Ben Laurie (CHATS)]
1554
1555 *) New OPENSSL_assert() macro (similar to assert(), but enabled
1556 unconditionally).
1557 [Ben Laurie (CHATS)]
1558
1559 *) Eliminate unused copy of key in RC4.
1560 [Ben Laurie (CHATS)]
1561
1562 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
1563 [Ben Laurie (CHATS)]
1564
1565 *) Fix off-by-one error in EGD path.
1566 [Ben Laurie (CHATS)]
1567
1568 *) If RANDFILE path is too long, ignore instead of truncating.
1569 [Ben Laurie (CHATS)]
1570
1571 *) Eliminate unused and incorrectly sized X.509 structure
1572 CBCParameter.
1573 [Ben Laurie (CHATS)]
1574
1575 *) Eliminate unused and dangerous function knumber().
1576 [Ben Laurie (CHATS)]
1577
1578 *) Eliminate unused and dangerous structure, KSSL_ERR.
1579 [Ben Laurie (CHATS)]
1580
1581 *) Protect against overlong session ID context length in an encoded
1582 session object. Since these are local, this does not appear to be
1583 exploitable.
1584 [Ben Laurie (CHATS)]
1585
1586 *) Change from security patch (see 0.9.6e below) that did not affect
1587 the 0.9.6 release series:
1588
1589 Remote buffer overflow in SSL3 protocol - an attacker could
1590 supply an oversized master key in Kerberos-enabled versions.
1591 (CVE-2002-0657)
1592 [Ben Laurie (CHATS)]
1593
1594 *) Change the SSL kerb5 codes to match RFC 2712.
1595 [Richard Levitte]
1596
1597 *) Make -nameopt work fully for req and add -reqopt switch.
1598 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
1599
1600 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
1601 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
1602
1603 *) Make sure tests can be performed even if the corresponding algorithms
1604 have been removed entirely. This was also the last step to make
1605 OpenSSL compilable with DJGPP under all reasonable conditions.
1606 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
1607
1608 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
1609 to allow version independent disabling of normally unselected ciphers,
1610 which may be activated as a side-effect of selecting a single cipher.
1611
1612 (E.g., cipher list string "RSA" enables ciphersuites that are left
1613 out of "ALL" because they do not provide symmetric encryption.
1614 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
1615 [Lutz Jaenicke, Bodo Moeller]
1616
1617 *) Add appropriate support for separate platform-dependent build
1618 directories. The recommended way to make a platform-dependent
1619 build directory is the following (tested on Linux), maybe with
1620 some local tweaks:
1621
1622 # Place yourself outside of the OpenSSL source tree. In
1623 # this example, the environment variable OPENSSL_SOURCE
1624 # is assumed to contain the absolute OpenSSL source directory.
1625 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
1626 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
1627 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
1628 mkdir -p `dirname $F`
1629 ln -s $OPENSSL_SOURCE/$F $F
1630 done
1631
1632 To be absolutely sure not to disturb the source tree, a "make clean"
1633 is a good thing. If it isn't successfull, don't worry about it,
1634 it probably means the source directory is very clean.
1635 [Richard Levitte]
1636
1637 *) Make sure any ENGINE control commands make local copies of string
1638 pointers passed to them whenever necessary. Otherwise it is possible
1639 the caller may have overwritten (or deallocated) the original string
1640 data when a later ENGINE operation tries to use the stored values.
1641 [Götz Babin-Ebell <babinebell@trustcenter.de>]
1642
1643 *) Improve diagnostics in file reading and command-line digests.
1644 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
1645
1646 *) Add AES modes CFB and OFB to the object database. Correct an
1647 error in AES-CFB decryption.
1648 [Richard Levitte]
1649
1650 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
1651 allows existing EVP_CIPHER_CTX structures to be reused after
1652 calling EVP_*Final(). This behaviour is used by encryption
1653 BIOs and some applications. This has the side effect that
1654 applications must explicitly clean up cipher contexts with
1655 EVP_CIPHER_CTX_cleanup() or they will leak memory.
1656 [Steve Henson]
1657
1658 *) Check the values of dna and dnb in bn_mul_recursive before calling
1659 bn_mul_comba (a non zero value means the a or b arrays do not contain
1660 n2 elements) and fallback to bn_mul_normal if either is not zero.
1661 [Steve Henson]
1662
1663 *) Fix escaping of non-ASCII characters when using the -subj option
1664 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
1665 [Lutz Jaenicke]
1666
1667 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
1668 form for "surname", serialNumber has no short form.
1669 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
1670 therefore remove "mail" short name for "internet 7".
1671 The OID for unique identifiers in X509 certificates is
1672 x500UniqueIdentifier, not uniqueIdentifier.
1673 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
1674 [Lutz Jaenicke]
1675
1676 *) Add an "init" command to the ENGINE config module and auto initialize
1677 ENGINEs. Without any "init" command the ENGINE will be initialized
1678 after all ctrl commands have been executed on it. If init=1 the
1679 ENGINE is initailized at that point (ctrls before that point are run
1680 on the uninitialized ENGINE and after on the initialized one). If
1681 init=0 then the ENGINE will not be iniatialized at all.
1682 [Steve Henson]
1683
1684 *) Fix the 'app_verify_callback' interface so that the user-defined
1685 argument is actually passed to the callback: In the
1686 SSL_CTX_set_cert_verify_callback() prototype, the callback
1687 declaration has been changed from
1688 int (*cb)()
1689 into
1690 int (*cb)(X509_STORE_CTX *,void *);
1691 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
1692 i=s->ctx->app_verify_callback(&ctx)
1693 has been changed into
1694 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
1695
1696 To update applications using SSL_CTX_set_cert_verify_callback(),
1697 a dummy argument can be added to their callback functions.
1698 [D. K. Smetters <smetters@parc.xerox.com>]
1699
1700 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
1701 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
1702
1703 *) Add and OPENSSL_LOAD_CONF define which will cause
1704 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
1705 This allows older applications to transparently support certain
1706 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
1707 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
1708 load the config file and OPENSSL_add_all_algorithms_conf() which will
1709 always load it have also been added.
1710 [Steve Henson]
1711
1712 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
1713 Adjust NIDs and EVP layer.
1714 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1715
1716 *) Config modules support in openssl utility.
1717
1718 Most commands now load modules from the config file,
1719 though in a few (such as version) this isn't done
1720 because it couldn't be used for anything.
1721
1722 In the case of ca and req the config file used is
1723 the same as the utility itself: that is the -config
1724 command line option can be used to specify an
1725 alternative file.
1726 [Steve Henson]
1727
1728 *) Move default behaviour from OPENSSL_config(). If appname is NULL
1729 use "openssl_conf" if filename is NULL use default openssl config file.
1730 [Steve Henson]
1731
1732 *) Add an argument to OPENSSL_config() to allow the use of an alternative
1733 config section name. Add a new flag to tolerate a missing config file
1734 and move code to CONF_modules_load_file().
1735 [Steve Henson]
1736
1737 *) Support for crypto accelerator cards from Accelerated Encryption
1738 Processing, www.aep.ie. (Use engine 'aep')
1739 The support was copied from 0.9.6c [engine] and adapted/corrected
1740 to work with the new engine framework.
1741 [AEP Inc. and Richard Levitte]
1742
1743 *) Support for SureWare crypto accelerator cards from Baltimore
1744 Technologies. (Use engine 'sureware')
1745 The support was copied from 0.9.6c [engine] and adapted
1746 to work with the new engine framework.
1747 [Richard Levitte]
1748
1749 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
1750 make the newer ENGINE framework commands for the CHIL engine work.
1751 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
1752
1753 *) Make it possible to produce shared libraries on ReliantUNIX.
1754 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
1755
1756 *) Add the configuration target debug-linux-ppro.
1757 Make 'openssl rsa' use the general key loading routines
1758 implemented in apps.c, and make those routines able to
1759 handle the key format FORMAT_NETSCAPE and the variant
1760 FORMAT_IISSGC.
1761 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1762
1763 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
1764 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
1765
1766 *) Add -keyform to rsautl, and document -engine.
1767 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
1768
1769 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
1770 BIO_R_NO_SUCH_FILE error code rather than the generic
1771 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
1772 [Ben Laurie]
1773
1774 *) Add new functions
1775 ERR_peek_last_error
1776 ERR_peek_last_error_line
1777 ERR_peek_last_error_line_data.
1778 These are similar to
1779 ERR_peek_error
1780 ERR_peek_error_line
1781 ERR_peek_error_line_data,
1782 but report on the latest error recorded rather than the first one
1783 still in the error queue.
1784 [Ben Laurie, Bodo Moeller]
1785
1786 *) default_algorithms option in ENGINE config module. This allows things
1787 like:
1788 default_algorithms = ALL
1789 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
1790 [Steve Henson]
1791
1792 *) Prelminary ENGINE config module.
1793 [Steve Henson]
1794
1795 *) New experimental application configuration code.
1796 [Steve Henson]
1797
1798 *) Change the AES code to follow the same name structure as all other
1799 symmetric ciphers, and behave the same way. Move everything to
1800 the directory crypto/aes, thereby obsoleting crypto/rijndael.
1801 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
1802
1803 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
1804 [Ben Laurie and Theo de Raadt]
1805
1806 *) Add option to output public keys in req command.
1807 [Massimiliano Pala madwolf@openca.org]
1808
1809 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
1810 (up to about 10% better than before for P-192 and P-224).
1811 [Bodo Moeller]
1812
1813 *) New functions/macros
1814
1815 SSL_CTX_set_msg_callback(ctx, cb)
1816 SSL_CTX_set_msg_callback_arg(ctx, arg)
1817 SSL_set_msg_callback(ssl, cb)
1818 SSL_set_msg_callback_arg(ssl, arg)
1819
1820 to request calling a callback function
1821
1822 void cb(int write_p, int version, int content_type,
1823 const void *buf, size_t len, SSL *ssl, void *arg)
1824
1825 whenever a protocol message has been completely received
1826 (write_p == 0) or sent (write_p == 1). Here 'version' is the
1827 protocol version according to which the SSL library interprets
1828 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
1829 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
1830 the content type as defined in the SSL 3.0/TLS 1.0 protocol
1831 specification (change_cipher_spec(20), alert(21), handshake(22)).
1832 'buf' and 'len' point to the actual message, 'ssl' to the
1833 SSL object, and 'arg' is the application-defined value set by
1834 SSL[_CTX]_set_msg_callback_arg().
1835
1836 'openssl s_client' and 'openssl s_server' have new '-msg' options
1837 to enable a callback that displays all protocol messages.
1838 [Bodo Moeller]
1839
1840 *) Change the shared library support so shared libraries are built as
1841 soon as the corresponding static library is finished, and thereby get
1842 openssl and the test programs linked against the shared library.
1843 This still only happens when the keyword "shard" has been given to
1844 the configuration scripts.
1845
1846 NOTE: shared library support is still an experimental thing, and
1847 backward binary compatibility is still not guaranteed.
1848 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
1849
1850 *) Add support for Subject Information Access extension.
1851 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
1852
1853 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
1854 additional bytes when new memory had to be allocated, not just
1855 when reusing an existing buffer.
1856 [Bodo Moeller]
1857
1858 *) New command line and configuration option 'utf8' for the req command.
1859 This allows field values to be specified as UTF8 strings.
1860 [Steve Henson]
1861
1862 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
1863 runs for the former and machine-readable output for the latter.
1864 [Ben Laurie]
1865
1866 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
1867 of the e-mail address in the DN (i.e., it will go into a certificate
1868 extension only). The new configuration file option 'email_in_dn = no'
1869 has the same effect.
1870 [Massimiliano Pala madwolf@openca.org]
1871
1872 *) Change all functions with names starting with des_ to be starting
1873 with DES_ instead. Add wrappers that are compatible with libdes,
1874 but are named _ossl_old_des_*. Finally, add macros that map the
1875 des_* symbols to the corresponding _ossl_old_des_* if libdes
1876 compatibility is desired. If OpenSSL 0.9.6c compatibility is
1877 desired, the des_* symbols will be mapped to DES_*, with one
1878 exception.
1879
1880 Since we provide two compatibility mappings, the user needs to
1881 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
1882 compatibility is desired. The default (i.e., when that macro
1883 isn't defined) is OpenSSL 0.9.6c compatibility.
1884
1885 There are also macros that enable and disable the support of old
1886 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
1887 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
1888 are defined, the default will apply: to support the old des routines.
1889
1890 In either case, one must include openssl/des.h to get the correct
1891 definitions. Do not try to just include openssl/des_old.h, that
1892 won't work.
1893
1894 NOTE: This is a major break of an old API into a new one. Software
1895 authors are encouraged to switch to the DES_ style functions. Some
1896 time in the future, des_old.h and the libdes compatibility functions
1897 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
1898 default), and then completely removed.
1899 [Richard Levitte]
1900
1901 *) Test for certificates which contain unsupported critical extensions.
1902 If such a certificate is found during a verify operation it is
1903 rejected by default: this behaviour can be overridden by either
1904 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
1905 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
1906 X509_supported_extension() has also been added which returns 1 if a
1907 particular extension is supported.
1908 [Steve Henson]
1909
1910 *) Modify the behaviour of EVP cipher functions in similar way to digests
1911 to retain compatibility with existing code.
1912 [Steve Henson]
1913
1914 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
1915 compatibility with existing code. In particular the 'ctx' parameter does
1916 not have to be to be initialized before the call to EVP_DigestInit() and
1917 it is tidied up after a call to EVP_DigestFinal(). New function
1918 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
1919 EVP_MD_CTX_copy() changed to not require the destination to be
1920 initialized valid and new function EVP_MD_CTX_copy_ex() added which
1921 requires the destination to be valid.
1922
1923 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
1924 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
1925 [Steve Henson]
1926
1927 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
1928 so that complete 'Handshake' protocol structures are kept in memory
1929 instead of overwriting 'msg_type' and 'length' with 'body' data.
1930 [Bodo Moeller]
1931
1932 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
1933 [Massimo Santin via Richard Levitte]
1934
1935 *) Major restructuring to the underlying ENGINE code. This includes
1936 reduction of linker bloat, separation of pure "ENGINE" manipulation
1937 (initialisation, etc) from functionality dealing with implementations
1938 of specific crypto iterfaces. This change also introduces integrated
1939 support for symmetric ciphers and digest implementations - so ENGINEs
1940 can now accelerate these by providing EVP_CIPHER and EVP_MD
1941 implementations of their own. This is detailed in crypto/engine/README
1942 as it couldn't be adequately described here. However, there are a few
1943 API changes worth noting - some RSA, DSA, DH, and RAND functions that
1944 were changed in the original introduction of ENGINE code have now
1945 reverted back - the hooking from this code to ENGINE is now a good
1946 deal more passive and at run-time, operations deal directly with
1947 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
1948 dereferencing through an ENGINE pointer any more. Also, the ENGINE
1949 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
1950 they were not being used by the framework as there is no concept of a
1951 BIGNUM_METHOD and they could not be generalised to the new
1952 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
1953 ENGINE_cpy() has been removed as it cannot be consistently defined in
1954 the new code.
1955 [Geoff Thorpe]
1956
1957 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
1958 [Steve Henson]
1959
1960 *) Change mkdef.pl to sort symbols that get the same entry number,
1961 and make sure the automatically generated functions ERR_load_*
1962 become part of libeay.num as well.
1963 [Richard Levitte]
1964
1965 *) New function SSL_renegotiate_pending(). This returns true once
1966 renegotiation has been requested (either SSL_renegotiate() call
1967 or HelloRequest/ClientHello receveived from the peer) and becomes
1968 false once a handshake has been completed.
1969 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
1970 sends a HelloRequest, but does not ensure that a handshake takes
1971 place. SSL_renegotiate_pending() is useful for checking if the
1972 client has followed the request.)
1973 [Bodo Moeller]
1974
1975 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
1976 By default, clients may request session resumption even during
1977 renegotiation (if session ID contexts permit); with this option,
1978 session resumption is possible only in the first handshake.
1979
1980 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
1981 more bits available for options that should not be part of
1982 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
1983 [Bodo Moeller]
1984
1985 *) Add some demos for certificate and certificate request creation.
1986 [Steve Henson]
1987
1988 *) Make maximum certificate chain size accepted from the peer application
1989 settable (SSL*_get/set_max_cert_list()), as proposed by
1990 "Douglas E. Engert" <deengert@anl.gov>.
1991 [Lutz Jaenicke]
1992
1993 *) Add support for shared libraries for Unixware-7
1994 (Boyd Lynn Gerber <gerberb@zenez.com>).
1995 [Lutz Jaenicke]
1996
1997 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
1998 be done prior to destruction. Use this to unload error strings from
1999 ENGINEs that load their own error strings. NB: This adds two new API
2000 functions to "get" and "set" this destroy handler in an ENGINE.
2001 [Geoff Thorpe]
2002
2003 *) Alter all existing ENGINE implementations (except "openssl" and
2004 "openbsd") to dynamically instantiate their own error strings. This
2005 makes them more flexible to be built both as statically-linked ENGINEs
2006 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2007 Also, add stub code to each that makes building them as self-contained
2008 shared-libraries easier (see README.ENGINE).
2009 [Geoff Thorpe]
2010
2011 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2012 implementations into applications that are completely implemented in
2013 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2014 commands that can be used to configure what shared-library to load and
2015 to control aspects of the way it is handled. Also, made an update to
2016 the README.ENGINE file that brings its information up-to-date and
2017 provides some information and instructions on the "dynamic" ENGINE
2018 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2019 [Geoff Thorpe]
2020
2021 *) Make it possible to unload ranges of ERR strings with a new
2022 "ERR_unload_strings" function.
2023 [Geoff Thorpe]
2024
2025 *) Add a copy() function to EVP_MD.
2026 [Ben Laurie]
2027
2028 *) Make EVP_MD routines take a context pointer instead of just the
2029 md_data void pointer.
2030 [Ben Laurie]
2031
2032 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2033 that the digest can only process a single chunk of data
2034 (typically because it is provided by a piece of
2035 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2036 is only going to provide a single chunk of data, and hence the
2037 framework needn't accumulate the data for oneshot drivers.
2038 [Ben Laurie]
2039
2040 *) As with "ERR", make it possible to replace the underlying "ex_data"
2041 functions. This change also alters the storage and management of global
2042 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2043 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2044 index counters. The API functions that use this state have been changed
2045 to take a "class_index" rather than pointers to the class's local STACK
2046 and counter, and there is now an API function to dynamically create new
2047 classes. This centralisation allows us to (a) plug a lot of the
2048 thread-safety problems that existed, and (b) makes it possible to clean
2049 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2050 such data would previously have always leaked in application code and
2051 workarounds were in place to make the memory debugging turn a blind eye
2052 to it. Application code that doesn't use this new function will still
2053 leak as before, but their memory debugging output will announce it now
2054 rather than letting it slide.
2055
2056 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2057 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2058 has a return value to indicate success or failure.
2059 [Geoff Thorpe]
2060
2061 *) Make it possible to replace the underlying "ERR" functions such that the
2062 global state (2 LHASH tables and 2 locks) is only used by the "default"
2063 implementation. This change also adds two functions to "get" and "set"
2064 the implementation prior to it being automatically set the first time
2065 any other ERR function takes place. Ie. an application can call "get",
2066 pass the return value to a module it has just loaded, and that module
2067 can call its own "set" function using that value. This means the
2068 module's "ERR" operations will use (and modify) the error state in the
2069 application and not in its own statically linked copy of OpenSSL code.
2070 [Geoff Thorpe]
2071
2072 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
2073 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
2074 the operation, and provides a more encapsulated way for external code
2075 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
2076 to use these functions rather than manually incrementing the counts.
2077
2078 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
2079 [Geoff Thorpe]
2080
2081 *) Add EVP test program.
2082 [Ben Laurie]
2083
2084 *) Add symmetric cipher support to ENGINE. Expect the API to change!
2085 [Ben Laurie]
2086
2087 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
2088 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
2089 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
2090 These allow a CRL to be built without having to access X509_CRL fields
2091 directly. Modify 'ca' application to use new functions.
2092 [Steve Henson]
2093
2094 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
2095 bug workarounds. Rollback attack detection is a security feature.
2096 The problem will only arise on OpenSSL servers when TLSv1 is not
2097 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
2098 Software authors not wanting to support TLSv1 will have special reasons
2099 for their choice and can explicitly enable this option.
2100 [Bodo Moeller, Lutz Jaenicke]
2101
2102 *) Rationalise EVP so it can be extended: don't include a union of
2103 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
2104 (similar to those existing for EVP_CIPHER_CTX).
2105 Usage example:
2106
2107 EVP_MD_CTX md;
2108
2109 EVP_MD_CTX_init(&md); /* new function call */
2110 EVP_DigestInit(&md, EVP_sha1());
2111 EVP_DigestUpdate(&md, in, len);
2112 EVP_DigestFinal(&md, out, NULL);
2113 EVP_MD_CTX_cleanup(&md); /* new function call */
2114
2115 [Ben Laurie]
2116
2117 *) Make DES key schedule conform to the usual scheme, as well as
2118 correcting its structure. This means that calls to DES functions
2119 now have to pass a pointer to a des_key_schedule instead of a
2120 plain des_key_schedule (which was actually always a pointer
2121 anyway): E.g.,
2122
2123 des_key_schedule ks;
2124
2125 des_set_key_checked(..., &ks);
2126 des_ncbc_encrypt(..., &ks, ...);
2127
2128 (Note that a later change renames 'des_...' into 'DES_...'.)
2129 [Ben Laurie]
2130
2131 *) Initial reduction of linker bloat: the use of some functions, such as
2132 PEM causes large amounts of unused functions to be linked in due to
2133 poor organisation. For example pem_all.c contains every PEM function
2134 which has a knock on effect of linking in large amounts of (unused)
2135 ASN1 code. Grouping together similar functions and splitting unrelated
2136 functions prevents this.
2137 [Steve Henson]
2138
2139 *) Cleanup of EVP macros.
2140 [Ben Laurie]
2141
2142 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
2143 correct _ecb suffix.
2144 [Ben Laurie]
2145
2146 *) Add initial OCSP responder support to ocsp application. The
2147 revocation information is handled using the text based index
2148 use by the ca application. The responder can either handle
2149 requests generated internally, supplied in files (for example
2150 via a CGI script) or using an internal minimal server.
2151 [Steve Henson]
2152
2153 *) Add configuration choices to get zlib compression for TLS.
2154 [Richard Levitte]
2155
2156 *) Changes to Kerberos SSL for RFC 2712 compliance:
2157 1. Implemented real KerberosWrapper, instead of just using
2158 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
2159 2. Implemented optional authenticator field of KerberosWrapper.
2160
2161 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
2162 and authenticator structs; see crypto/krb5/.
2163
2164 Generalized Kerberos calls to support multiple Kerberos libraries.
2165 [Vern Staats <staatsvr@asc.hpc.mil>,
2166 Jeffrey Altman <jaltman@columbia.edu>
2167 via Richard Levitte]
2168
2169 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
2170 already does with RSA. testdsa.h now has 'priv_key/pub_key'
2171 values for each of the key sizes rather than having just
2172 parameters (and 'speed' generating keys each time).
2173 [Geoff Thorpe]
2174
2175 *) Speed up EVP routines.
2176 Before:
2177 encrypt
2178 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
2179 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
2180 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
2181 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
2182 decrypt
2183 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
2184 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
2185 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
2186 After:
2187 encrypt
2188 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
2189 decrypt
2190 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
2191 [Ben Laurie]
2192
2193 *) Added the OS2-EMX target.
2194 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
2195
2196 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
2197 to support NCONF routines in extension code. New function CONF_set_nconf()
2198 to allow functions which take an NCONF to also handle the old LHASH
2199 structure: this means that the old CONF compatible routines can be
2200 retained (in particular wrt extensions) without having to duplicate the
2201 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
2202 [Steve Henson]
2203
2204 *) Enhance the general user interface with mechanisms for inner control
2205 and with possibilities to have yes/no kind of prompts.
2206 [Richard Levitte]
2207
2208 *) Change all calls to low level digest routines in the library and
2209 applications to use EVP. Add missing calls to HMAC_cleanup() and
2210 don't assume HMAC_CTX can be copied using memcpy().
2211 [Verdon Walker <VWalker@novell.com>, Steve Henson]
2212
2213 *) Add the possibility to control engines through control names but with
2214 arbitrary arguments instead of just a string.
2215 Change the key loaders to take a UI_METHOD instead of a callback
2216 function pointer. NOTE: this breaks binary compatibility with earlier
2217 versions of OpenSSL [engine].
2218 Adapt the nCipher code for these new conditions and add a card insertion
2219 callback.
2220 [Richard Levitte]
2221
2222 *) Enhance the general user interface with mechanisms to better support
2223 dialog box interfaces, application-defined prompts, the possibility
2224 to use defaults (for example default passwords from somewhere else)
2225 and interrupts/cancellations.
2226 [Richard Levitte]
2227
2228 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
2229 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
2230 [Steve Henson]
2231
2232 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
2233 tidy up some unnecessarily weird code in 'sk_new()').
2234 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
2235
2236 *) Change the key loading routines for ENGINEs to use the same kind
2237 callback (pem_password_cb) as all other routines that need this
2238 kind of callback.
2239 [Richard Levitte]
2240
2241 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
2242 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
2243 than this minimum value is recommended.
2244 [Lutz Jaenicke]
2245
2246 *) New random seeder for OpenVMS, using the system process statistics
2247 that are easily reachable.
2248 [Richard Levitte]
2249
2250 *) Windows apparently can't transparently handle global
2251 variables defined in DLLs. Initialisations such as:
2252
2253 const ASN1_ITEM *it = &ASN1_INTEGER_it;
2254
2255 wont compile. This is used by the any applications that need to
2256 declare their own ASN1 modules. This was fixed by adding the option
2257 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
2258 needed for static libraries under Win32.
2259 [Steve Henson]
2260
2261 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
2262 setting of purpose and trust fields. New X509_STORE trust and
2263 purpose functions and tidy up setting in other SSL functions.
2264 [Steve Henson]
2265
2266 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
2267 structure. These are inherited by X509_STORE_CTX when it is
2268 initialised. This allows various defaults to be set in the
2269 X509_STORE structure (such as flags for CRL checking and custom
2270 purpose or trust settings) for functions which only use X509_STORE_CTX
2271 internally such as S/MIME.
2272
2273 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
2274 trust settings if they are not set in X509_STORE. This allows X509_STORE
2275 purposes and trust (in S/MIME for example) to override any set by default.
2276
2277 Add command line options for CRL checking to smime, s_client and s_server
2278 applications.
2279 [Steve Henson]
2280
2281 *) Initial CRL based revocation checking. If the CRL checking flag(s)
2282 are set then the CRL is looked up in the X509_STORE structure and
2283 its validity and signature checked, then if the certificate is found
2284 in the CRL the verify fails with a revoked error.
2285
2286 Various new CRL related callbacks added to X509_STORE_CTX structure.
2287
2288 Command line options added to 'verify' application to support this.
2289
2290 This needs some additional work, such as being able to handle multiple
2291 CRLs with different times, extension based lookup (rather than just
2292 by subject name) and ultimately more complete V2 CRL extension
2293 handling.
2294 [Steve Henson]
2295
2296 *) Add a general user interface API (crypto/ui/). This is designed
2297 to replace things like des_read_password and friends (backward
2298 compatibility functions using this new API are provided).
2299 The purpose is to remove prompting functions from the DES code
2300 section as well as provide for prompting through dialog boxes in
2301 a window system and the like.
2302 [Richard Levitte]
2303
2304 *) Add "ex_data" support to ENGINE so implementations can add state at a
2305 per-structure level rather than having to store it globally.
2306 [Geoff]
2307
2308 *) Make it possible for ENGINE structures to be copied when retrieved by
2309 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
2310 This causes the "original" ENGINE structure to act like a template,
2311 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
2312 operational state can be localised to each ENGINE structure, despite the
2313 fact they all share the same "methods". New ENGINE structures returned in
2314 this case have no functional references and the return value is the single
2315 structural reference. This matches the single structural reference returned
2316 by ENGINE_by_id() normally, when it is incremented on the pre-existing
2317 ENGINE structure.
2318 [Geoff]
2319
2320 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
2321 needs to match any other type at all we need to manually clear the
2322 tag cache.
2323 [Steve Henson]
2324
2325 *) Changes to the "openssl engine" utility to include;
2326 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
2327 about an ENGINE's available control commands.
2328 - executing control commands from command line arguments using the
2329 '-pre' and '-post' switches. '-post' is only used if '-t' is
2330 specified and the ENGINE is successfully initialised. The syntax for
2331 the individual commands are colon-separated, for example;
2332 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
2333 [Geoff]
2334
2335 *) New dynamic control command support for ENGINEs. ENGINEs can now
2336 declare their own commands (numbers), names (strings), descriptions,
2337 and input types for run-time discovery by calling applications. A
2338 subset of these commands are implicitly classed as "executable"
2339 depending on their input type, and only these can be invoked through
2340 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
2341 can be based on user input, config files, etc). The distinction is
2342 that "executable" commands cannot return anything other than a boolean
2343 result and can only support numeric or string input, whereas some
2344 discoverable commands may only be for direct use through
2345 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
2346 pointers, or other custom uses. The "executable" commands are to
2347 support parameterisations of ENGINE behaviour that can be
2348 unambiguously defined by ENGINEs and used consistently across any
2349 OpenSSL-based application. Commands have been added to all the
2350 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
2351 control over shared-library paths without source code alterations.
2352 [Geoff]
2353
2354 *) Changed all ENGINE implementations to dynamically allocate their
2355 ENGINEs rather than declaring them statically. Apart from this being
2356 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
2357 this also allows the implementations to compile without using the
2358 internal engine_int.h header.
2359 [Geoff]
2360
2361 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
2362 'const' value. Any code that should be able to modify a RAND_METHOD
2363 should already have non-const pointers to it (ie. they should only
2364 modify their own ones).
2365 [Geoff]
2366
2367 *) Made a variety of little tweaks to the ENGINE code.
2368 - "atalla" and "ubsec" string definitions were moved from header files
2369 to C code. "nuron" string definitions were placed in variables
2370 rather than hard-coded - allowing parameterisation of these values
2371 later on via ctrl() commands.
2372 - Removed unused "#if 0"'d code.
2373 - Fixed engine list iteration code so it uses ENGINE_free() to release
2374 structural references.
2375 - Constified the RAND_METHOD element of ENGINE structures.
2376 - Constified various get/set functions as appropriate and added
2377 missing functions (including a catch-all ENGINE_cpy that duplicates
2378 all ENGINE values onto a new ENGINE except reference counts/state).
2379 - Removed NULL parameter checks in get/set functions. Setting a method
2380 or function to NULL is a way of cancelling out a previously set
2381 value. Passing a NULL ENGINE parameter is just plain stupid anyway
2382 and doesn't justify the extra error symbols and code.
2383 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
2384 flags from engine_int.h to engine.h.
2385 - Changed prototypes for ENGINE handler functions (init(), finish(),
2386 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
2387 [Geoff]
2388
2389 *) Implement binary inversion algorithm for BN_mod_inverse in addition
2390 to the algorithm using long division. The binary algorithm can be
2391 used only if the modulus is odd. On 32-bit systems, it is faster
2392 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
2393 roughly 5-15% for 256-bit moduli), so we use it only for moduli
2394 up to 450 bits. In 64-bit environments, the binary algorithm
2395 appears to be advantageous for much longer moduli; here we use it
2396 for moduli up to 2048 bits.
2397 [Bodo Moeller]
2398
2399 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
2400 could not support the combine flag in choice fields.
2401 [Steve Henson]
2402
2403 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
2404 extensions from a certificate request to the certificate.
2405 [Steve Henson]
2406
2407 *) Allow multiple 'certopt' and 'nameopt' options to be separated
2408 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
2409 file: this allows the display of the certificate about to be
2410 signed to be customised, to allow certain fields to be included
2411 or excluded and extension details. The old system didn't display
2412 multicharacter strings properly, omitted fields not in the policy
2413 and couldn't display additional details such as extensions.
2414 [Steve Henson]
2415
2416 *) Function EC_POINTs_mul for multiple scalar multiplication
2417 of an arbitrary number of elliptic curve points
2418 \sum scalars[i]*points[i],
2419 optionally including the generator defined for the EC_GROUP:
2420 scalar*generator + \sum scalars[i]*points[i].
2421
2422 EC_POINT_mul is a simple wrapper function for the typical case
2423 that the point list has just one item (besides the optional
2424 generator).
2425 [Bodo Moeller]
2426
2427 *) First EC_METHODs for curves over GF(p):
2428
2429 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
2430 operations and provides various method functions that can also
2431 operate with faster implementations of modular arithmetic.
2432
2433 EC_GFp_mont_method() reuses most functions that are part of
2434 EC_GFp_simple_method, but uses Montgomery arithmetic.
2435
2436 [Bodo Moeller; point addition and point doubling
2437 implementation directly derived from source code provided by
2438 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
2439
2440 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
2441 crypto/ec/ec_lib.c):
2442
2443 Curves are EC_GROUP objects (with an optional group generator)
2444 based on EC_METHODs that are built into the library.
2445
2446 Points are EC_POINT objects based on EC_GROUP objects.
2447
2448 Most of the framework would be able to handle curves over arbitrary
2449 finite fields, but as there are no obvious types for fields other
2450 than GF(p), some functions are limited to that for now.
2451 [Bodo Moeller]
2452
2453 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
2454 that the file contains a complete HTTP response.
2455 [Richard Levitte]
2456
2457 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
2458 change the def and num file printf format specifier from "%-40sXXX"
2459 to "%-39s XXX". The latter will always guarantee a space after the
2460 field while the former will cause them to run together if the field
2461 is 40 of more characters long.
2462 [Steve Henson]
2463
2464 *) Constify the cipher and digest 'method' functions and structures
2465 and modify related functions to take constant EVP_MD and EVP_CIPHER
2466 pointers.
2467 [Steve Henson]
2468
2469 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
2470 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
2471 [Bodo Moeller]
2472
2473 *) Modify EVP_Digest*() routines so they now return values. Although the
2474 internal software routines can never fail additional hardware versions
2475 might.
2476 [Steve Henson]
2477
2478 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
2479
2480 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
2481 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
2482
2483 ASN1 error codes
2484 ERR_R_NESTED_ASN1_ERROR
2485 ...
2486 ERR_R_MISSING_ASN1_EOS
2487 were 4 .. 9, conflicting with
2488 ERR_LIB_RSA (= ERR_R_RSA_LIB)
2489 ...
2490 ERR_LIB_PEM (= ERR_R_PEM_LIB).
2491 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
2492
2493 Add new error code 'ERR_R_INTERNAL_ERROR'.
2494 [Bodo Moeller]
2495
2496 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
2497 suffices.
2498 [Bodo Moeller]
2499
2500 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
2501 sets the subject name for a new request or supersedes the
2502 subject name in a given request. Formats that can be parsed are
2503 'CN=Some Name, OU=myOU, C=IT'
2504 and
2505 'CN=Some Name/OU=myOU/C=IT'.
2506
2507 Add options '-batch' and '-verbose' to 'openssl req'.
2508 [Massimiliano Pala <madwolf@hackmasters.net>]
2509
2510 *) Introduce the possibility to access global variables through
2511 functions on platform were that's the best way to handle exporting
2512 global variables in shared libraries. To enable this functionality,
2513 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
2514 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
2515 is normally done by Configure or something similar).
2516
2517 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
2518 in the source file (foo.c) like this:
2519
2520 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
2521 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
2522
2523 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
2524 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
2525
2526 OPENSSL_DECLARE_GLOBAL(int,foo);
2527 #define foo OPENSSL_GLOBAL_REF(foo)
2528 OPENSSL_DECLARE_GLOBAL(double,bar);
2529 #define bar OPENSSL_GLOBAL_REF(bar)
2530
2531 The #defines are very important, and therefore so is including the
2532 header file everywhere where the defined globals are used.
2533
2534 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
2535 of ASN.1 items, but that structure is a bit different.
2536
2537 The largest change is in util/mkdef.pl which has been enhanced with
2538 better and easier to understand logic to choose which symbols should
2539 go into the Windows .def files as well as a number of fixes and code
2540 cleanup (among others, algorithm keywords are now sorted
2541 lexicographically to avoid constant rewrites).
2542 [Richard Levitte]
2543
2544 *) In BN_div() keep a copy of the sign of 'num' before writing the
2545 result to 'rm' because if rm==num the value will be overwritten
2546 and produce the wrong result if 'num' is negative: this caused
2547 problems with BN_mod() and BN_nnmod().
2548 [Steve Henson]
2549
2550 *) Function OCSP_request_verify(). This checks the signature on an
2551 OCSP request and verifies the signer certificate. The signer
2552 certificate is just checked for a generic purpose and OCSP request
2553 trust settings.
2554 [Steve Henson]
2555
2556 *) Add OCSP_check_validity() function to check the validity of OCSP
2557 responses. OCSP responses are prepared in real time and may only
2558 be a few seconds old. Simply checking that the current time lies
2559 between thisUpdate and nextUpdate max reject otherwise valid responses
2560 caused by either OCSP responder or client clock inaccuracy. Instead
2561 we allow thisUpdate and nextUpdate to fall within a certain period of
2562 the current time. The age of the response can also optionally be
2563 checked. Two new options -validity_period and -status_age added to
2564 ocsp utility.
2565 [Steve Henson]
2566
2567 *) If signature or public key algorithm is unrecognized print out its
2568 OID rather that just UNKNOWN.
2569 [Steve Henson]
2570
2571 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
2572 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
2573 ID to be generated from the issuer certificate alone which can then be
2574 passed to OCSP_id_issuer_cmp().
2575 [Steve Henson]
2576
2577 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
2578 ASN1 modules to export functions returning ASN1_ITEM pointers
2579 instead of the ASN1_ITEM structures themselves. This adds several
2580 new macros which allow the underlying ASN1 function/structure to
2581 be accessed transparently. As a result code should not use ASN1_ITEM
2582 references directly (such as &X509_it) but instead use the relevant
2583 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
2584 use of the new ASN1 code on platforms where exporting structures
2585 is problematical (for example in shared libraries) but exporting
2586 functions returning pointers to structures is not.
2587 [Steve Henson]
2588
2589 *) Add support for overriding the generation of SSL/TLS session IDs.
2590 These callbacks can be registered either in an SSL_CTX or per SSL.
2591 The purpose of this is to allow applications to control, if they wish,
2592 the arbitrary values chosen for use as session IDs, particularly as it
2593 can be useful for session caching in multiple-server environments. A
2594 command-line switch for testing this (and any client code that wishes
2595 to use such a feature) has been added to "s_server".
2596 [Geoff Thorpe, Lutz Jaenicke]
2597
2598 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
2599 of the form '#if defined(...) || defined(...) || ...' and
2600 '#if !defined(...) && !defined(...) && ...'. This also avoids
2601 the growing number of special cases it was previously handling.
2602 [Richard Levitte]
2603
2604 *) Make all configuration macros available for application by making
2605 sure they are available in opensslconf.h, by giving them names starting
2606 with "OPENSSL_" to avoid conflicts with other packages and by making
2607 sure e_os2.h will cover all platform-specific cases together with
2608 opensslconf.h.
2609 Additionally, it is now possible to define configuration/platform-
2610 specific names (called "system identities"). In the C code, these
2611 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
2612 macro with the name beginning with "OPENSSL_SYS_", which is determined
2613 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
2614 what is available.
2615 [Richard Levitte]
2616
2617 *) New option -set_serial to 'req' and 'x509' this allows the serial
2618 number to use to be specified on the command line. Previously self
2619 signed certificates were hard coded with serial number 0 and the
2620 CA options of 'x509' had to use a serial number in a file which was
2621 auto incremented.
2622 [Steve Henson]
2623
2624 *) New options to 'ca' utility to support V2 CRL entry extensions.
2625 Currently CRL reason, invalidity date and hold instruction are
2626 supported. Add new CRL extensions to V3 code and some new objects.
2627 [Steve Henson]
2628
2629 *) New function EVP_CIPHER_CTX_set_padding() this is used to
2630 disable standard block padding (aka PKCS#5 padding) in the EVP
2631 API, which was previously mandatory. This means that the data is
2632 not padded in any way and so the total length much be a multiple
2633 of the block size, otherwise an error occurs.
2634 [Steve Henson]
2635
2636 *) Initial (incomplete) OCSP SSL support.
2637 [Steve Henson]
2638
2639 *) New function OCSP_parse_url(). This splits up a URL into its host,
2640 port and path components: primarily to parse OCSP URLs. New -url
2641 option to ocsp utility.
2642 [Steve Henson]
2643
2644 *) New nonce behavior. The return value of OCSP_check_nonce() now
2645 reflects the various checks performed. Applications can decide
2646 whether to tolerate certain situations such as an absent nonce
2647 in a response when one was present in a request: the ocsp application
2648 just prints out a warning. New function OCSP_add1_basic_nonce()
2649 this is to allow responders to include a nonce in a response even if
2650 the request is nonce-less.
2651 [Steve Henson]
2652
2653 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
2654 skipped when using openssl x509 multiple times on a single input file,
2655 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
2656 [Bodo Moeller]
2657
2658 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
2659 set string type: to handle setting ASN1_TIME structures. Fix ca
2660 utility to correctly initialize revocation date of CRLs.
2661 [Steve Henson]
2662
2663 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
2664 the clients preferred ciphersuites and rather use its own preferences.
2665 Should help to work around M$ SGC (Server Gated Cryptography) bug in
2666 Internet Explorer by ensuring unchanged hash method during stepup.
2667 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
2668 [Lutz Jaenicke]
2669
2670 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
2671 to aes and add a new 'exist' option to print out symbols that don't
2672 appear to exist.
2673 [Steve Henson]
2674
2675 *) Additional options to ocsp utility to allow flags to be set and
2676 additional certificates supplied.
2677 [Steve Henson]
2678
2679 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
2680 OCSP client a number of certificate to only verify the response
2681 signature against.
2682 [Richard Levitte]
2683
2684 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
2685 handle the new API. Currently only ECB, CBC modes supported. Add new
2686 AES OIDs.
2687
2688 Add TLS AES ciphersuites as described in RFC3268, "Advanced
2689 Encryption Standard (AES) Ciphersuites for Transport Layer
2690 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
2691 not enabled by default and were not part of the "ALL" ciphersuite
2692 alias because they were not yet official; they could be
2693 explicitly requested by specifying the "AESdraft" ciphersuite
2694 group alias. In the final release of OpenSSL 0.9.7, the group
2695 alias is called "AES" and is part of "ALL".)
2696 [Ben Laurie, Steve Henson, Bodo Moeller]
2697
2698 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
2699 request to response.
2700 [Steve Henson]
2701
2702 *) Functions for OCSP responders. OCSP_request_onereq_count(),
2703 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
2704 extract information from a certificate request. OCSP_response_create()
2705 creates a response and optionally adds a basic response structure.
2706 OCSP_basic_add1_status() adds a complete single response to a basic
2707 response and returns the OCSP_SINGLERESP structure just added (to allow
2708 extensions to be included for example). OCSP_basic_add1_cert() adds a
2709 certificate to a basic response and OCSP_basic_sign() signs a basic
2710 response with various flags. New helper functions ASN1_TIME_check()
2711 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
2712 (converts ASN1_TIME to GeneralizedTime).
2713 [Steve Henson]
2714
2715 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
2716 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
2717 structure from a certificate. X509_pubkey_digest() digests the public_key
2718 contents: this is used in various key identifiers.
2719 [Steve Henson]
2720
2721 *) Make sk_sort() tolerate a NULL argument.
2722 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
2723
2724 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
2725 passed by the function are trusted implicitly. If any of them signed the
2726 response then it is assumed to be valid and is not verified.
2727 [Steve Henson]
2728
2729 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
2730 to data. This was previously part of the PKCS7 ASN1 code. This
2731 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
2732 [Steve Henson, reported by Kenneth R. Robinette
2733 <support@securenetterm.com>]
2734
2735 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
2736 routines: without these tracing memory leaks is very painful.
2737 Fix leaks in PKCS12 and PKCS7 routines.
2738 [Steve Henson]
2739
2740 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
2741 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
2742 effectively meant GeneralizedTime would never be used. Now it
2743 is initialised to -1 but X509_time_adj() now has to check the value
2744 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
2745 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
2746 [Steve Henson, reported by Kenneth R. Robinette
2747 <support@securenetterm.com>]
2748
2749 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
2750 result in a zero length in the ASN1_INTEGER structure which was
2751 not consistent with the structure when d2i_ASN1_INTEGER() was used
2752 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
2753 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
2754 where it did not print out a minus for negative ASN1_INTEGER.
2755 [Steve Henson]
2756
2757 *) Add summary printout to ocsp utility. The various functions which
2758 convert status values to strings have been renamed to:
2759 OCSP_response_status_str(), OCSP_cert_status_str() and
2760 OCSP_crl_reason_str() and are no longer static. New options
2761 to verify nonce values and to disable verification. OCSP response
2762 printout format cleaned up.
2763 [Steve Henson]
2764
2765 *) Add additional OCSP certificate checks. These are those specified
2766 in RFC2560. This consists of two separate checks: the CA of the
2767 certificate being checked must either be the OCSP signer certificate
2768 or the issuer of the OCSP signer certificate. In the latter case the
2769 OCSP signer certificate must contain the OCSP signing extended key
2770 usage. This check is performed by attempting to match the OCSP
2771 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
2772 in the OCSP_CERTID structures of the response.
2773 [Steve Henson]
2774
2775 *) Initial OCSP certificate verification added to OCSP_basic_verify()
2776 and related routines. This uses the standard OpenSSL certificate
2777 verify routines to perform initial checks (just CA validity) and
2778 to obtain the certificate chain. Then additional checks will be
2779 performed on the chain. Currently the root CA is checked to see
2780 if it is explicitly trusted for OCSP signing. This is used to set
2781 a root CA as a global signing root: that is any certificate that
2782 chains to that CA is an acceptable OCSP signing certificate.
2783 [Steve Henson]
2784
2785 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
2786 extensions from a separate configuration file.
2787 As when reading extensions from the main configuration file,
2788 the '-extensions ...' option may be used for specifying the
2789 section to use.
2790 [Massimiliano Pala <madwolf@comune.modena.it>]
2791
2792 *) New OCSP utility. Allows OCSP requests to be generated or
2793 read. The request can be sent to a responder and the output
2794 parsed, outputed or printed in text form. Not complete yet:
2795 still needs to check the OCSP response validity.
2796 [Steve Henson]
2797
2798 *) New subcommands for 'openssl ca':
2799 'openssl ca -status <serial>' prints the status of the cert with
2800 the given serial number (according to the index file).
2801 'openssl ca -updatedb' updates the expiry status of certificates
2802 in the index file.
2803 [Massimiliano Pala <madwolf@comune.modena.it>]
2804
2805 *) New '-newreq-nodes' command option to CA.pl. This is like
2806 '-newreq', but calls 'openssl req' with the '-nodes' option
2807 so that the resulting key is not encrypted.
2808 [Damien Miller <djm@mindrot.org>]
2809
2810 *) New configuration for the GNU Hurd.
2811 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
2812
2813 *) Initial code to implement OCSP basic response verify. This
2814 is currently incomplete. Currently just finds the signer's
2815 certificate and verifies the signature on the response.
2816 [Steve Henson]
2817
2818 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
2819 value of OPENSSLDIR. This is available via the new '-d' option
2820 to 'openssl version', and is also included in 'openssl version -a'.
2821 [Bodo Moeller]
2822
2823 *) Allowing defining memory allocation callbacks that will be given
2824 file name and line number information in additional arguments
2825 (a const char* and an int). The basic functionality remains, as
2826 well as the original possibility to just replace malloc(),
2827 realloc() and free() by functions that do not know about these
2828 additional arguments. To register and find out the current
2829 settings for extended allocation functions, the following
2830 functions are provided:
2831
2832 CRYPTO_set_mem_ex_functions
2833 CRYPTO_set_locked_mem_ex_functions
2834 CRYPTO_get_mem_ex_functions
2835 CRYPTO_get_locked_mem_ex_functions
2836
2837 These work the same way as CRYPTO_set_mem_functions and friends.
2838 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
2839 extended allocation function is enabled.
2840 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
2841 a conventional allocation function is enabled.
2842 [Richard Levitte, Bodo Moeller]
2843
2844 *) Finish off removing the remaining LHASH function pointer casts.
2845 There should no longer be any prototype-casting required when using
2846 the LHASH abstraction, and any casts that remain are "bugs". See
2847 the callback types and macros at the head of lhash.h for details
2848 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
2849 [Geoff Thorpe]
2850
2851 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
2852 If /dev/[u]random devices are not available or do not return enough
2853 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
2854 be queried.
2855 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
2856 /etc/entropy will be queried once each in this sequence, quering stops
2857 when enough entropy was collected without querying more sockets.
2858 [Lutz Jaenicke]
2859
2860 *) Change the Unix RAND_poll() variant to be able to poll several
2861 random devices, as specified by DEVRANDOM, until a sufficient amount
2862 of data has been collected. We spend at most 10 ms on each file
2863 (select timeout) and read in non-blocking mode. DEVRANDOM now
2864 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
2865 (previously it was just the string "/dev/urandom"), so on typical
2866 platforms the 10 ms delay will never occur.
2867 Also separate out the Unix variant to its own file, rand_unix.c.
2868 For VMS, there's a currently-empty rand_vms.c.
2869 [Richard Levitte]
2870
2871 *) Move OCSP client related routines to ocsp_cl.c. These
2872 provide utility functions which an application needing
2873 to issue a request to an OCSP responder and analyse the
2874 response will typically need: as opposed to those which an
2875 OCSP responder itself would need which will be added later.
2876
2877 OCSP_request_sign() signs an OCSP request with an API similar
2878 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
2879 response. OCSP_response_get1_basic() extracts basic response
2880 from response. OCSP_resp_find_status(): finds and extracts status
2881 information from an OCSP_CERTID structure (which will be created
2882 when the request structure is built). These are built from lower
2883 level functions which work on OCSP_SINGLERESP structures but
2884 wont normally be used unless the application wishes to examine
2885 extensions in the OCSP response for example.
2886
2887 Replace nonce routines with a pair of functions.
2888 OCSP_request_add1_nonce() adds a nonce value and optionally
2889 generates a random value. OCSP_check_nonce() checks the
2890 validity of the nonce in an OCSP response.
2891 [Steve Henson]
2892
2893 *) Change function OCSP_request_add() to OCSP_request_add0_id().
2894 This doesn't copy the supplied OCSP_CERTID and avoids the
2895 need to free up the newly created id. Change return type
2896 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
2897 This can then be used to add extensions to the request.
2898 Deleted OCSP_request_new(), since most of its functionality
2899 is now in OCSP_REQUEST_new() (and the case insensitive name
2900 clash) apart from the ability to set the request name which
2901 will be added elsewhere.
2902 [Steve Henson]
2903
2904 *) Update OCSP API. Remove obsolete extensions argument from
2905 various functions. Extensions are now handled using the new
2906 OCSP extension code. New simple OCSP HTTP function which
2907 can be used to send requests and parse the response.
2908 [Steve Henson]
2909
2910 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
2911 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
2912 uses the special reorder version of SET OF to sort the attributes
2913 and reorder them to match the encoded order. This resolves a long
2914 standing problem: a verify on a PKCS7 structure just after signing
2915 it used to fail because the attribute order did not match the
2916 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
2917 it uses the received order. This is necessary to tolerate some broken
2918 software that does not order SET OF. This is handled by encoding
2919 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
2920 to produce the required SET OF.
2921 [Steve Henson]
2922
2923 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
2924 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
2925 files to get correct declarations of the ASN.1 item variables.
2926 [Richard Levitte]
2927
2928 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
2929 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
2930 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
2931 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
2932 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
2933 ASN1_ITEM and no wrapper functions.
2934 [Steve Henson]
2935
2936 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
2937 replace the old function pointer based I/O routines. Change most of
2938 the *_d2i_bio() and *_d2i_fp() functions to use these.
2939 [Steve Henson]
2940
2941 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
2942 lines, recognice more "algorithms" that can be deselected, and make
2943 it complain about algorithm deselection that isn't recognised.
2944 [Richard Levitte]
2945
2946 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
2947 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
2948 to use new functions. Add NO_ASN1_OLD which can be set to remove
2949 some old style ASN1 functions: this can be used to determine if old
2950 code will still work when these eventually go away.
2951 [Steve Henson]
2952
2953 *) New extension functions for OCSP structures, these follow the
2954 same conventions as certificates and CRLs.
2955 [Steve Henson]
2956
2957 *) New function X509V3_add1_i2d(). This automatically encodes and
2958 adds an extension. Its behaviour can be customised with various
2959 flags to append, replace or delete. Various wrappers added for
2960 certifcates and CRLs.
2961 [Steve Henson]
2962
2963 *) Fix to avoid calling the underlying ASN1 print routine when
2964 an extension cannot be parsed. Correct a typo in the
2965 OCSP_SERVICELOC extension. Tidy up print OCSP format.
2966 [Steve Henson]
2967
2968 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
2969 entries for variables.
2970 [Steve Henson]
2971
2972 *) Add functionality to apps/openssl.c for detecting locking
2973 problems: As the program is single-threaded, all we have
2974 to do is register a locking callback using an array for
2975 storing which locks are currently held by the program.
2976 [Bodo Moeller]
2977
2978 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
2979 SSL_get_ex_data_X509_STORE_idx(), which is used in
2980 ssl_verify_cert_chain() and thus can be called at any time
2981 during TLS/SSL handshakes so that thread-safety is essential.
2982 Unfortunately, the ex_data design is not at all suited
2983 for multi-threaded use, so it probably should be abolished.
2984 [Bodo Moeller]
2985
2986 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
2987 [Broadcom, tweaked and integrated by Geoff Thorpe]
2988
2989 *) Move common extension printing code to new function
2990 X509V3_print_extensions(). Reorganise OCSP print routines and
2991 implement some needed OCSP ASN1 functions. Add OCSP extensions.
2992 [Steve Henson]
2993
2994 *) New function X509_signature_print() to remove duplication in some
2995 print routines.
2996 [Steve Henson]
2997
2998 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
2999 set (this was treated exactly the same as SET OF previously). This
3000 is used to reorder the STACK representing the structure to match the
3001 encoding. This will be used to get round a problem where a PKCS7
3002 structure which was signed could not be verified because the STACK
3003 order did not reflect the encoded order.
3004 [Steve Henson]
3005
3006 *) Reimplement the OCSP ASN1 module using the new code.
3007 [Steve Henson]
3008
3009 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3010 for its ASN1 operations. The old style function pointers still exist
3011 for now but they will eventually go away.
3012 [Steve Henson]
3013
3014 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3015 completely replaces the old ASN1 functionality with a table driven
3016 encoder and decoder which interprets an ASN1_ITEM structure describing
3017 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3018 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3019 has also been converted to the new form.
3020 [Steve Henson]
3021
3022 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3023 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3024 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3025 for negative moduli.
3026 [Bodo Moeller]
3027
3028 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3029 of not touching the result's sign bit.
3030 [Bodo Moeller]
3031
3032 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3033 set.
3034 [Bodo Moeller]
3035
3036 *) Changed the LHASH code to use prototypes for callbacks, and created
3037 macros to declare and implement thin (optionally static) functions
3038 that provide type-safety and avoid function pointer casting for the
3039 type-specific callbacks.
3040 [Geoff Thorpe]
3041
3042 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3043 RFC 2712.
3044 [Veers Staats <staatsvr@asc.hpc.mil>,
3045 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3046
3047 *) Reformat the FAQ so the different questions and answers can be divided
3048 in sections depending on the subject.
3049 [Richard Levitte]
3050
3051 *) Have the zlib compression code load ZLIB.DLL dynamically under
3052 Windows.
3053 [Richard Levitte]
3054
3055 *) New function BN_mod_sqrt for computing square roots modulo a prime
3056 (using the probabilistic Tonelli-Shanks algorithm unless
3057 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3058 be handled deterministically).
3059 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3060
3061 *) Make BN_mod_inverse faster by explicitly handling small quotients
3062 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3063 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3064 [Bodo Moeller]
3065
3066 *) New function BN_kronecker.
3067 [Bodo Moeller]
3068
3069 *) Fix BN_gcd so that it works on negative inputs; the result is
3070 positive unless both parameters are zero.
3071 Previously something reasonably close to an infinite loop was
3072 possible because numbers could be growing instead of shrinking
3073 in the implementation of Euclid's algorithm.
3074 [Bodo Moeller]
3075
3076 *) Fix BN_is_word() and BN_is_one() macros to take into account the
3077 sign of the number in question.
3078
3079 Fix BN_is_word(a,w) to work correctly for w == 0.
3080
3081 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
3082 because its test if the absolute value of 'a' equals 'w'.
3083 Note that BN_abs_is_word does *not* handle w == 0 reliably;
3084 it exists mostly for use in the implementations of BN_is_zero(),
3085 BN_is_one(), and BN_is_word().
3086 [Bodo Moeller]
3087
3088 *) New function BN_swap.
3089 [Bodo Moeller]
3090
3091 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
3092 the exponentiation functions are more likely to produce reasonable
3093 results on negative inputs.
3094 [Bodo Moeller]
3095
3096 *) Change BN_mod_mul so that the result is always non-negative.
3097 Previously, it could be negative if one of the factors was negative;
3098 I don't think anyone really wanted that behaviour.
3099 [Bodo Moeller]
3100
3101 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
3102 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
3103 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
3104 and add new functions:
3105
3106 BN_nnmod
3107 BN_mod_sqr
3108 BN_mod_add
3109 BN_mod_add_quick
3110 BN_mod_sub
3111 BN_mod_sub_quick
3112 BN_mod_lshift1
3113 BN_mod_lshift1_quick
3114 BN_mod_lshift
3115 BN_mod_lshift_quick
3116
3117 These functions always generate non-negative results.
3118
3119 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
3120 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
3121
3122 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
3123 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
3124 be reduced modulo m.
3125 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3126
3127 #if 0
3128 The following entry accidentily appeared in the CHANGES file
3129 distributed with OpenSSL 0.9.7. The modifications described in
3130 it do *not* apply to OpenSSL 0.9.7.
3131
3132 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3133 was actually never needed) and in BN_mul(). The removal in BN_mul()
3134 required a small change in bn_mul_part_recursive() and the addition
3135 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3136 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3137 bn_sub_words() and bn_add_words() except they take arrays with
3138 differing sizes.
3139 [Richard Levitte]
3140 #endif
3141
3142 *) In 'openssl passwd', verify passwords read from the terminal
3143 unless the '-salt' option is used (which usually means that
3144 verification would just waste user's time since the resulting
3145 hash is going to be compared with some given password hash)
3146 or the new '-noverify' option is used.
3147
3148 This is an incompatible change, but it does not affect
3149 non-interactive use of 'openssl passwd' (passwords on the command
3150 line, '-stdin' option, '-in ...' option) and thus should not
3151 cause any problems.
3152 [Bodo Moeller]
3153
3154 *) Remove all references to RSAref, since there's no more need for it.
3155 [Richard Levitte]
3156
3157 *) Make DSO load along a path given through an environment variable
3158 (SHLIB_PATH) with shl_load().
3159 [Richard Levitte]
3160
3161 *) Constify the ENGINE code as a result of BIGNUM constification.
3162 Also constify the RSA code and most things related to it. In a
3163 few places, most notable in the depth of the ASN.1 code, ugly
3164 casts back to non-const were required (to be solved at a later
3165 time)
3166 [Richard Levitte]
3167
3168 *) Make it so the openssl application has all engines loaded by default.
3169 [Richard Levitte]
3170
3171 *) Constify the BIGNUM routines a little more.
3172 [Richard Levitte]
3173
3174 *) Add the following functions:
3175
3176 ENGINE_load_cswift()
3177 ENGINE_load_chil()
3178 ENGINE_load_atalla()
3179 ENGINE_load_nuron()
3180 ENGINE_load_builtin_engines()
3181
3182 That way, an application can itself choose if external engines that
3183 are built-in in OpenSSL shall ever be used or not. The benefit is
3184 that applications won't have to be linked with libdl or other dso
3185 libraries unless it's really needed.
3186
3187 Changed 'openssl engine' to load all engines on demand.
3188 Changed the engine header files to avoid the duplication of some
3189 declarations (they differed!).
3190 [Richard Levitte]
3191
3192 *) 'openssl engine' can now list capabilities.
3193 [Richard Levitte]
3194
3195 *) Better error reporting in 'openssl engine'.
3196 [Richard Levitte]
3197
3198 *) Never call load_dh_param(NULL) in s_server.
3199 [Bodo Moeller]
3200
3201 *) Add engine application. It can currently list engines by name and
3202 identity, and test if they are actually available.
3203 [Richard Levitte]
3204
3205 *) Improve RPM specification file by forcing symbolic linking and making
3206 sure the installed documentation is also owned by root.root.
3207 [Damien Miller <djm@mindrot.org>]
3208
3209 *) Give the OpenSSL applications more possibilities to make use of
3210 keys (public as well as private) handled by engines.
3211 [Richard Levitte]
3212
3213 *) Add OCSP code that comes from CertCo.
3214 [Richard Levitte]
3215
3216 *) Add VMS support for the Rijndael code.
3217 [Richard Levitte]
3218
3219 *) Added untested support for Nuron crypto accelerator.
3220 [Ben Laurie]
3221
3222 *) Add support for external cryptographic devices. This code was
3223 previously distributed separately as the "engine" branch.
3224 [Geoff Thorpe, Richard Levitte]
3225
3226 *) Rework the filename-translation in the DSO code. It is now possible to
3227 have far greater control over how a "name" is turned into a filename
3228 depending on the operating environment and any oddities about the
3229 different shared library filenames on each system.
3230 [Geoff Thorpe]
3231
3232 *) Support threads on FreeBSD-elf in Configure.
3233 [Richard Levitte]
3234
3235 *) Fix for SHA1 assembly problem with MASM: it produces
3236 warnings about corrupt line number information when assembling
3237 with debugging information. This is caused by the overlapping
3238 of two sections.
3239 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
3240
3241 *) NCONF changes.
3242 NCONF_get_number() has no error checking at all. As a replacement,
3243 NCONF_get_number_e() is defined (_e for "error checking") and is
3244 promoted strongly. The old NCONF_get_number is kept around for
3245 binary backward compatibility.
3246 Make it possible for methods to load from something other than a BIO,
3247 by providing a function pointer that is given a name instead of a BIO.
3248 For example, this could be used to load configuration data from an
3249 LDAP server.
3250 [Richard Levitte]
3251
3252 *) Fix for non blocking accept BIOs. Added new I/O special reason
3253 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
3254 with non blocking I/O was not possible because no retry code was
3255 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
3256 this case.
3257 [Steve Henson]
3258
3259 *) Added the beginnings of Rijndael support.
3260 [Ben Laurie]
3261
3262 *) Fix for bug in DirectoryString mask setting. Add support for
3263 X509_NAME_print_ex() in 'req' and X509_print_ex() function
3264 to allow certificate printing to more controllable, additional
3265 'certopt' option to 'x509' to allow new printing options to be
3266 set.
3267 [Steve Henson]
3268
3269 *) Clean old EAY MD5 hack from e_os.h.
3270 [Richard Levitte]
3271
3272 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
3273
3274 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3275 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3276 [Joe Orton, Steve Henson]
3277
3278 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
3279
3280 *) Fix additional bug revealed by the NISCC test suite:
3281
3282 Stop bug triggering large recursion when presented with
3283 certain ASN.1 tags (CVE-2003-0851)
3284 [Steve Henson]
3285
3286 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
3287
3288 *) Fix various bugs revealed by running the NISCC test suite:
3289
3290 Stop out of bounds reads in the ASN1 code when presented with
3291 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3292
3293 If verify callback ignores invalid public key errors don't try to check
3294 certificate signature with the NULL public key.
3295
3296 [Steve Henson]
3297
3298 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3299 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3300 specifications.
3301 [Steve Henson]
3302
3303 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3304 extra data after the compression methods not only for TLS 1.0
3305 but also for SSL 3.0 (as required by the specification).
3306 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3307
3308 *) Change X509_certificate_type() to mark the key as exported/exportable
3309 when it's 512 *bits* long, not 512 bytes.
3310 [Richard Levitte]
3311
3312 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
3313
3314 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3315 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3316 a protocol version number mismatch like a decryption error
3317 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3318 [Bodo Moeller]
3319
3320 *) Turn on RSA blinding by default in the default implementation
3321 to avoid a timing attack. Applications that don't want it can call
3322 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3323 They would be ill-advised to do so in most cases.
3324 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3325
3326 *) Change RSA blinding code so that it works when the PRNG is not
3327 seeded (in this case, the secret RSA exponent is abused as
3328 an unpredictable seed -- if it is not unpredictable, there
3329 is no point in blinding anyway). Make RSA blinding thread-safe
3330 by remembering the creator's thread ID in rsa->blinding and
3331 having all other threads use local one-time blinding factors
3332 (this requires more computation than sharing rsa->blinding, but
3333 avoids excessive locking; and if an RSA object is not shared
3334 between threads, blinding will still be very fast).
3335 [Bodo Moeller]
3336
3337 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
3338
3339 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3340 via timing by performing a MAC computation even if incorrrect
3341 block cipher padding has been found. This is a countermeasure
3342 against active attacks where the attacker has to distinguish
3343 between bad padding and a MAC verification error. (CVE-2003-0078)
3344
3345 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3346 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3347 Martin Vuagnoux (EPFL, Ilion)]
3348
3349 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
3350
3351 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
3352 memory from it's contents. This is done with a counter that will
3353 place alternating values in each byte. This can be used to solve
3354 two issues: 1) the removal of calls to memset() by highly optimizing
3355 compilers, and 2) cleansing with other values than 0, since those can
3356 be read through on certain media, for example a swap space on disk.
3357 [Geoff Thorpe]
3358
3359 *) Bugfix: client side session caching did not work with external caching,
3360 because the session->cipher setting was not restored when reloading
3361 from the external cache. This problem was masked, when
3362 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
3363 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
3364 [Lutz Jaenicke]
3365
3366 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
3367 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
3368 [Zeev Lieber <zeev-l@yahoo.com>]
3369
3370 *) Undo an undocumented change introduced in 0.9.6e which caused
3371 repeated calls to OpenSSL_add_all_ciphers() and
3372 OpenSSL_add_all_digests() to be ignored, even after calling
3373 EVP_cleanup().
3374 [Richard Levitte]
3375
3376 *) Change the default configuration reader to deal with last line not
3377 being properly terminated.
3378 [Richard Levitte]
3379
3380 *) Change X509_NAME_cmp() so it applies the special rules on handling
3381 DN values that are of type PrintableString, as well as RDNs of type
3382 emailAddress where the value has the type ia5String.
3383 [stefank@valicert.com via Richard Levitte]
3384
3385 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
3386 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
3387 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
3388 the bitwise-OR of the two for use by the majority of applications
3389 wanting this behaviour, and update the docs. The documented
3390 behaviour and actual behaviour were inconsistent and had been
3391 changing anyway, so this is more a bug-fix than a behavioural
3392 change.
3393 [Geoff Thorpe, diagnosed by Nadav Har'El]
3394
3395 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
3396 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
3397 [Bodo Moeller]
3398
3399 *) Fix initialization code race conditions in
3400 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
3401 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
3402 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
3403 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
3404 ssl2_get_cipher_by_char(),
3405 ssl3_get_cipher_by_char().
3406 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
3407
3408 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
3409 the cached sessions are flushed, as the remove_cb() might use ex_data
3410 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
3411 (see [openssl.org #212]).
3412 [Geoff Thorpe, Lutz Jaenicke]
3413
3414 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
3415 length, instead of the encoding length to d2i_ASN1_OBJECT.
3416 [Steve Henson]
3417
3418 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
3419
3420 *) [In 0.9.6g-engine release:]
3421 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
3422 [Lynn Gazis <lgazis@rainbow.com>]
3423
3424 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
3425
3426 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
3427 and get fix the header length calculation.
3428 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
3429 Alon Kantor <alonk@checkpoint.com> (and others),
3430 Steve Henson]
3431
3432 *) Use proper error handling instead of 'assertions' in buffer
3433 overflow checks added in 0.9.6e. This prevents DoS (the
3434 assertions could call abort()).
3435 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
3436
3437 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
3438
3439 *) Add various sanity checks to asn1_get_length() to reject
3440 the ASN1 length bytes if they exceed sizeof(long), will appear
3441 negative or the content length exceeds the length of the
3442 supplied buffer.
3443 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3444
3445 *) Fix cipher selection routines: ciphers without encryption had no flags
3446 for the cipher strength set and where therefore not handled correctly
3447 by the selection routines (PR #130).
3448 [Lutz Jaenicke]
3449
3450 *) Fix EVP_dsa_sha macro.
3451 [Nils Larsch]
3452
3453 *) New option
3454 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
3455 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
3456 that was added in OpenSSL 0.9.6d.
3457
3458 As the countermeasure turned out to be incompatible with some
3459 broken SSL implementations, the new option is part of SSL_OP_ALL.
3460 SSL_OP_ALL is usually employed when compatibility with weird SSL
3461 implementations is desired (e.g. '-bugs' option to 's_client' and
3462 's_server'), so the new option is automatically set in many
3463 applications.
3464 [Bodo Moeller]
3465
3466 *) Changes in security patch:
3467
3468 Changes marked "(CHATS)" were sponsored by the Defense Advanced
3469 Research Projects Agency (DARPA) and Air Force Research Laboratory,
3470 Air Force Materiel Command, USAF, under agreement number
3471 F30602-01-2-0537.
3472
3473 *) Add various sanity checks to asn1_get_length() to reject
3474 the ASN1 length bytes if they exceed sizeof(long), will appear
3475 negative or the content length exceeds the length of the
3476 supplied buffer. (CVE-2002-0659)
3477 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
3478
3479 *) Assertions for various potential buffer overflows, not known to
3480 happen in practice.
3481 [Ben Laurie (CHATS)]
3482
3483 *) Various temporary buffers to hold ASCII versions of integers were
3484 too small for 64 bit platforms. (CVE-2002-0655)
3485 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
3486
3487 *) Remote buffer overflow in SSL3 protocol - an attacker could
3488 supply an oversized session ID to a client. (CVE-2002-0656)
3489 [Ben Laurie (CHATS)]
3490
3491 *) Remote buffer overflow in SSL2 protocol - an attacker could
3492 supply an oversized client master key. (CVE-2002-0656)
3493 [Ben Laurie (CHATS)]
3494
3495 Changes between 0.9.6c and 0.9.6d [9 May 2002]
3496
3497 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
3498 encoded as NULL) with id-dsa-with-sha1.
3499 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
3500
3501 *) Check various X509_...() return values in apps/req.c.
3502 [Nils Larsch <nla@trustcenter.de>]
3503
3504 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
3505 an end-of-file condition would erronously be flagged, when the CRLF
3506 was just at the end of a processed block. The bug was discovered when
3507 processing data through a buffering memory BIO handing the data to a
3508 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
3509 <ptsekov@syntrex.com> and Nedelcho Stanev.
3510 [Lutz Jaenicke]
3511
3512 *) Implement a countermeasure against a vulnerability recently found
3513 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
3514 before application data chunks to avoid the use of known IVs
3515 with data potentially chosen by the attacker.
3516 [Bodo Moeller]
3517
3518 *) Fix length checks in ssl3_get_client_hello().
3519 [Bodo Moeller]
3520
3521 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
3522 to prevent ssl3_read_internal() from incorrectly assuming that
3523 ssl3_read_bytes() found application data while handshake
3524 processing was enabled when in fact s->s3->in_read_app_data was
3525 merely automatically cleared during the initial handshake.
3526 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
3527
3528 *) Fix object definitions for Private and Enterprise: they were not
3529 recognized in their shortname (=lowercase) representation. Extend
3530 obj_dat.pl to issue an error when using undefined keywords instead
3531 of silently ignoring the problem (Svenning Sorensen
3532 <sss@sss.dnsalias.net>).
3533 [Lutz Jaenicke]
3534
3535 *) Fix DH_generate_parameters() so that it works for 'non-standard'
3536 generators, i.e. generators other than 2 and 5. (Previously, the
3537 code did not properly initialise the 'add' and 'rem' values to
3538 BN_generate_prime().)
3539
3540 In the new general case, we do not insist that 'generator' is
3541 actually a primitive root: This requirement is rather pointless;
3542 a generator of the order-q subgroup is just as good, if not
3543 better.
3544 [Bodo Moeller]
3545
3546 *) Map new X509 verification errors to alerts. Discovered and submitted by
3547 Tom Wu <tom@arcot.com>.
3548 [Lutz Jaenicke]
3549
3550 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
3551 returning non-zero before the data has been completely received
3552 when using non-blocking I/O.
3553 [Bodo Moeller; problem pointed out by John Hughes]
3554
3555 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
3556 [Ben Laurie, Lutz Jaenicke]
3557
3558 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
3559 Yoram Zahavi <YoramZ@gilian.com>).
3560 [Lutz Jaenicke]
3561
3562 *) Add information about CygWin 1.3 and on, and preserve proper
3563 configuration for the versions before that.
3564 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3565
3566 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
3567 check whether we deal with a copy of a session and do not delete from
3568 the cache in this case. Problem reported by "Izhar Shoshani Levi"
3569 <izhar@checkpoint.com>.
3570 [Lutz Jaenicke]
3571
3572 *) Do not store session data into the internal session cache, if it
3573 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
3574 flag is set). Proposed by Aslam <aslam@funk.com>.
3575 [Lutz Jaenicke]
3576
3577 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
3578 value is 0.
3579 [Richard Levitte]
3580
3581 *) [In 0.9.6d-engine release:]
3582 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3583 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3584
3585 *) Add the configuration target linux-s390x.
3586 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
3587
3588 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
3589 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
3590 variable as an indication that a ClientHello message has been
3591 received. As the flag value will be lost between multiple
3592 invocations of ssl3_accept when using non-blocking I/O, the
3593 function may not be aware that a handshake has actually taken
3594 place, thus preventing a new session from being added to the
3595 session cache.
3596
3597 To avoid this problem, we now set s->new_session to 2 instead of
3598 using a local variable.
3599 [Lutz Jaenicke, Bodo Moeller]
3600
3601 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
3602 if the SSL_R_LENGTH_MISMATCH error is detected.
3603 [Geoff Thorpe, Bodo Moeller]
3604
3605 *) New 'shared_ldflag' column in Configure platform table.
3606 [Richard Levitte]
3607
3608 *) Fix EVP_CIPHER_mode macro.
3609 ["Dan S. Camper" <dan@bti.net>]
3610
3611 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
3612 type, we must throw them away by setting rr->length to 0.
3613 [D P Chang <dpc@qualys.com>]
3614
3615 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
3616
3617 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
3618 <Dominikus.Scherkl@biodata.com>. (The previous implementation
3619 worked incorrectly for those cases where range = 10..._2 and
3620 3*range is two bits longer than range.)
3621 [Bodo Moeller]
3622
3623 *) Only add signing time to PKCS7 structures if it is not already
3624 present.
3625 [Steve Henson]
3626
3627 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
3628 OBJ_ld_ce should be OBJ_id_ce.
3629 Also some ip-pda OIDs in crypto/objects/objects.txt were
3630 incorrect (cf. RFC 3039).
3631 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
3632
3633 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
3634 returns early because it has nothing to do.
3635 [Andy Schneider <andy.schneider@bjss.co.uk>]
3636
3637 *) [In 0.9.6c-engine release:]
3638 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
3639 [Andy Schneider <andy.schneider@bjss.co.uk>]
3640
3641 *) [In 0.9.6c-engine release:]
3642 Add support for Cryptographic Appliance's keyserver technology.
3643 (Use engine 'keyclient')
3644 [Cryptographic Appliances and Geoff Thorpe]
3645
3646 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
3647 is called via tools/c89.sh because arguments have to be
3648 rearranged (all '-L' options must appear before the first object
3649 modules).
3650 [Richard Shapiro <rshapiro@abinitio.com>]
3651
3652 *) [In 0.9.6c-engine release:]
3653 Add support for Broadcom crypto accelerator cards, backported
3654 from 0.9.7.
3655 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
3656
3657 *) [In 0.9.6c-engine release:]
3658 Add support for SureWare crypto accelerator cards from
3659 Baltimore Technologies. (Use engine 'sureware')
3660 [Baltimore Technologies and Mark Cox]
3661
3662 *) [In 0.9.6c-engine release:]
3663 Add support for crypto accelerator cards from Accelerated
3664 Encryption Processing, www.aep.ie. (Use engine 'aep')
3665 [AEP Inc. and Mark Cox]
3666
3667 *) Add a configuration entry for gcc on UnixWare.
3668 [Gary Benson <gbenson@redhat.com>]
3669
3670 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
3671 messages are stored in a single piece (fixed-length part and
3672 variable-length part combined) and fix various bugs found on the way.
3673 [Bodo Moeller]
3674
3675 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
3676 instead. BIO_gethostbyname() does not know what timeouts are
3677 appropriate, so entries would stay in cache even when they have
3678 become invalid.
3679 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
3680
3681 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
3682 faced with a pathologically small ClientHello fragment that does
3683 not contain client_version: Instead of aborting with an error,
3684 simply choose the highest available protocol version (i.e.,
3685 TLS 1.0 unless it is disabled). In practice, ClientHello
3686 messages are never sent like this, but this change gives us
3687 strictly correct behaviour at least for TLS.
3688 [Bodo Moeller]
3689
3690 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
3691 never resets s->method to s->ctx->method when called from within
3692 one of the SSL handshake functions.
3693 [Bodo Moeller; problem pointed out by Niko Baric]
3694
3695 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
3696 (sent using the client's version number) if client_version is
3697 smaller than the protocol version in use. Also change
3698 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
3699 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
3700 the client will at least see that alert.
3701 [Bodo Moeller]
3702
3703 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
3704 correctly.
3705 [Bodo Moeller]
3706
3707 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
3708 client receives HelloRequest while in a handshake.
3709 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
3710
3711 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
3712 should end in 'break', not 'goto end' which circuments various
3713 cleanups done in state SSL_ST_OK. But session related stuff
3714 must be disabled for SSL_ST_OK in the case that we just sent a
3715 HelloRequest.
3716
3717 Also avoid some overhead by not calling ssl_init_wbio_buffer()
3718 before just sending a HelloRequest.
3719 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
3720
3721 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
3722 reveal whether illegal block cipher padding was found or a MAC
3723 verification error occured. (Neither SSLerr() codes nor alerts
3724 are directly visible to potential attackers, but the information
3725 may leak via logfiles.)
3726
3727 Similar changes are not required for the SSL 2.0 implementation
3728 because the number of padding bytes is sent in clear for SSL 2.0,
3729 and the extra bytes are just ignored. However ssl/s2_pkt.c
3730 failed to verify that the purported number of padding bytes is in
3731 the legal range.
3732 [Bodo Moeller]
3733
3734 *) Add OpenUNIX-8 support including shared libraries
3735 (Boyd Lynn Gerber <gerberb@zenez.com>).
3736 [Lutz Jaenicke]
3737
3738 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
3739 'wristwatch attack' using huge encoding parameters (cf.
3740 James H. Manger's CRYPTO 2001 paper). Note that the
3741 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
3742 encoding parameters and hence was not vulnerable.
3743 [Bodo Moeller]
3744
3745 *) BN_sqr() bug fix.
3746 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
3747
3748 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
3749 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
3750 followed by modular reduction.
3751 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
3752
3753 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
3754 equivalent based on BN_pseudo_rand() instead of BN_rand().
3755 [Bodo Moeller]
3756
3757 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
3758 This function was broken, as the check for a new client hello message
3759 to handle SGC did not allow these large messages.
3760 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
3761 [Lutz Jaenicke]
3762
3763 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
3764 [Lutz Jaenicke]
3765
3766 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
3767 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
3768 [Lutz Jaenicke]
3769
3770 *) Rework the configuration and shared library support for Tru64 Unix.
3771 The configuration part makes use of modern compiler features and
3772 still retains old compiler behavior for those that run older versions
3773 of the OS. The shared library support part includes a variant that
3774 uses the RPATH feature, and is available through the special
3775 configuration target "alpha-cc-rpath", which will never be selected
3776 automatically.
3777 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
3778
3779 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
3780 with the same message size as in ssl3_get_certificate_request().
3781 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
3782 messages might inadvertently be reject as too long.
3783 [Petr Lampa <lampa@fee.vutbr.cz>]
3784
3785 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
3786 [Andy Polyakov]
3787
3788 *) Modified SSL library such that the verify_callback that has been set
3789 specificly for an SSL object with SSL_set_verify() is actually being
3790 used. Before the change, a verify_callback set with this function was
3791 ignored and the verify_callback() set in the SSL_CTX at the time of
3792 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
3793 to allow the necessary settings.
3794 [Lutz Jaenicke]
3795
3796 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
3797 explicitly to NULL, as at least on Solaris 8 this seems not always to be
3798 done automatically (in contradiction to the requirements of the C
3799 standard). This made problems when used from OpenSSH.
3800 [Lutz Jaenicke]
3801
3802 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
3803 dh->length and always used
3804
3805 BN_rand_range(priv_key, dh->p).
3806
3807 BN_rand_range() is not necessary for Diffie-Hellman, and this
3808 specific range makes Diffie-Hellman unnecessarily inefficient if
3809 dh->length (recommended exponent length) is much smaller than the
3810 length of dh->p. We could use BN_rand_range() if the order of
3811 the subgroup was stored in the DH structure, but we only have
3812 dh->length.
3813
3814 So switch back to
3815
3816 BN_rand(priv_key, l, ...)
3817
3818 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
3819 otherwise.
3820 [Bodo Moeller]
3821
3822 *) In
3823
3824 RSA_eay_public_encrypt
3825 RSA_eay_private_decrypt
3826 RSA_eay_private_encrypt (signing)
3827 RSA_eay_public_decrypt (signature verification)
3828
3829 (default implementations for RSA_public_encrypt,
3830 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
3831 always reject numbers >= n.
3832 [Bodo Moeller]
3833
3834 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
3835 to synchronize access to 'locking_thread'. This is necessary on
3836 systems where access to 'locking_thread' (an 'unsigned long'
3837 variable) is not atomic.
3838 [Bodo Moeller]
3839
3840 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
3841 *before* setting the 'crypto_lock_rand' flag. The previous code had
3842 a race condition if 0 is a valid thread ID.
3843 [Travis Vitek <vitek@roguewave.com>]
3844
3845 *) Add support for shared libraries under Irix.
3846 [Albert Chin-A-Young <china@thewrittenword.com>]
3847
3848 *) Add configuration option to build on Linux on both big-endian and
3849 little-endian MIPS.
3850 [Ralf Baechle <ralf@uni-koblenz.de>]
3851
3852 *) Add the possibility to create shared libraries on HP-UX.
3853 [Richard Levitte]
3854
3855 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
3856
3857 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
3858 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
3859 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
3860 PRNG state recovery was possible based on the output of
3861 one PRNG request appropriately sized to gain knowledge on
3862 'md' followed by enough consecutive 1-byte PRNG requests
3863 to traverse all of 'state'.
3864
3865 1. When updating 'md_local' (the current thread's copy of 'md')
3866 during PRNG output generation, hash all of the previous
3867 'md_local' value, not just the half used for PRNG output.
3868
3869 2. Make the number of bytes from 'state' included into the hash
3870 independent from the number of PRNG bytes requested.
3871
3872 The first measure alone would be sufficient to avoid
3873 Markku-Juhani's attack. (Actually it had never occurred
3874 to me that the half of 'md_local' used for chaining was the
3875 half from which PRNG output bytes were taken -- I had always
3876 assumed that the secret half would be used.) The second
3877 measure makes sure that additional data from 'state' is never
3878 mixed into 'md_local' in small portions; this heuristically
3879 further strengthens the PRNG.
3880 [Bodo Moeller]
3881
3882 *) Fix crypto/bn/asm/mips3.s.
3883 [Andy Polyakov]
3884
3885 *) When only the key is given to "enc", the IV is undefined. Print out
3886 an error message in this case.
3887 [Lutz Jaenicke]
3888
3889 *) Handle special case when X509_NAME is empty in X509 printing routines.
3890 [Steve Henson]
3891
3892 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
3893 positive and less than q.
3894 [Bodo Moeller]
3895
3896 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
3897 used: it isn't thread safe and the add_lock_callback should handle
3898 that itself.
3899 [Paul Rose <Paul.Rose@bridge.com>]
3900
3901 *) Verify that incoming data obeys the block size in
3902 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
3903 [Bodo Moeller]
3904
3905 *) Fix OAEP check.
3906 [Ulf Möller, Bodo Möller]
3907
3908 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
3909 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
3910 when fixing the server behaviour for backwards-compatible 'client
3911 hello' messages. (Note that the attack is impractical against
3912 SSL 3.0 and TLS 1.0 anyway because length and version checking
3913 means that the probability of guessing a valid ciphertext is
3914 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
3915 paper.)
3916
3917 Before 0.9.5, the countermeasure (hide the error by generating a
3918 random 'decryption result') did not work properly because
3919 ERR_clear_error() was missing, meaning that SSL_get_error() would
3920 detect the supposedly ignored error.
3921
3922 Both problems are now fixed.
3923 [Bodo Moeller]
3924
3925 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
3926 (previously it was 1024).
3927 [Bodo Moeller]
3928
3929 *) Fix for compatibility mode trust settings: ignore trust settings
3930 unless some valid trust or reject settings are present.
3931 [Steve Henson]
3932
3933 *) Fix for blowfish EVP: its a variable length cipher.
3934 [Steve Henson]
3935
3936 *) Fix various bugs related to DSA S/MIME verification. Handle missing
3937 parameters in DSA public key structures and return an error in the
3938 DSA routines if parameters are absent.
3939 [Steve Henson]
3940
3941 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
3942 in the current directory if neither $RANDFILE nor $HOME was set.
3943 RAND_file_name() in 0.9.6a returned NULL in this case. This has
3944 caused some confusion to Windows users who haven't defined $HOME.
3945 Thus RAND_file_name() is changed again: e_os.h can define a
3946 DEFAULT_HOME, which will be used if $HOME is not set.
3947 For Windows, we use "C:"; on other platforms, we still require
3948 environment variables.
3949
3950 *) Move 'if (!initialized) RAND_poll()' into regions protected by
3951 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
3952 having multiple threads call RAND_poll() concurrently.
3953 [Bodo Moeller]
3954
3955 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
3956 combination of a flag and a thread ID variable.
3957 Otherwise while one thread is in ssleay_rand_bytes (which sets the
3958 flag), *other* threads can enter ssleay_add_bytes without obeying
3959 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
3960 that they do not hold after the first thread unsets add_do_not_lock).
3961 [Bodo Moeller]
3962
3963 *) Change bctest again: '-x' expressions are not available in all
3964 versions of 'test'.
3965 [Bodo Moeller]
3966
3967 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
3968
3969 *) Fix a couple of memory leaks in PKCS7_dataDecode()
3970 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
3971
3972 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
3973 the default extension for executables, if any. Also, make the perl
3974 scripts that use symlink() to test if it really exists and use "cp"
3975 if it doesn't. All this made OpenSSL compilable and installable in
3976 CygWin.
3977 [Richard Levitte]
3978
3979 *) Fix for asn1_GetSequence() for indefinite length constructed data.
3980 If SEQUENCE is length is indefinite just set c->slen to the total
3981 amount of data available.
3982 [Steve Henson, reported by shige@FreeBSD.org]
3983 [This change does not apply to 0.9.7.]
3984
3985 *) Change bctest to avoid here-documents inside command substitution
3986 (workaround for FreeBSD /bin/sh bug).
3987 For compatibility with Ultrix, avoid shell functions (introduced
3988 in the bctest version that searches along $PATH).
3989 [Bodo Moeller]
3990
3991 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
3992 with des_encrypt() defined on some operating systems, like Solaris
3993 and UnixWare.
3994 [Richard Levitte]
3995
3996 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
3997 On the Importance of Eliminating Errors in Cryptographic
3998 Computations, J. Cryptology 14 (2001) 2, 101-119,
3999 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4000 [Ulf Moeller]
4001
4002 *) MIPS assembler BIGNUM division bug fix.
4003 [Andy Polyakov]
4004
4005 *) Disabled incorrect Alpha assembler code.
4006 [Richard Levitte]
4007
4008 *) Fix PKCS#7 decode routines so they correctly update the length
4009 after reading an EOC for the EXPLICIT tag.
4010 [Steve Henson]
4011 [This change does not apply to 0.9.7.]
4012
4013 *) Fix bug in PKCS#12 key generation routines. This was triggered
4014 if a 3DES key was generated with a 0 initial byte. Include
4015 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4016 (but broken) behaviour.
4017 [Steve Henson]
4018
4019 *) Enhance bctest to search for a working bc along $PATH and print
4020 it when found.
4021 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4022
4023 *) Fix memory leaks in err.c: free err_data string if necessary;
4024 don't write to the wrong index in ERR_set_error_data.
4025 [Bodo Moeller]
4026
4027 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4028 did not exist.
4029 [Bodo Moeller]
4030
4031 *) Replace rdtsc with _emit statements for VC++ version 5.
4032 [Jeremy Cooper <jeremy@baymoo.org>]
4033
4034 *) Make it possible to reuse SSLv2 sessions.
4035 [Richard Levitte]
4036
4037 *) In copy_email() check for >= 0 as a return value for
4038 X509_NAME_get_index_by_NID() since 0 is a valid index.
4039 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4040
4041 *) Avoid coredump with unsupported or invalid public keys by checking if
4042 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4043 PKCS7_verify() fails with non detached data.
4044 [Steve Henson]
4045
4046 *) Don't use getenv in library functions when run as setuid/setgid.
4047 New function OPENSSL_issetugid().
4048 [Ulf Moeller]
4049
4050 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4051 due to incorrect handling of multi-threading:
4052
4053 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4054
4055 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4056
4057 3. Count how many times MemCheck_off() has been called so that
4058 nested use can be treated correctly. This also avoids
4059 inband-signalling in the previous code (which relied on the
4060 assumption that thread ID 0 is impossible).
4061 [Bodo Moeller]
4062
4063 *) Add "-rand" option also to s_client and s_server.
4064 [Lutz Jaenicke]
4065
4066 *) Fix CPU detection on Irix 6.x.
4067 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4068 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4069
4070 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
4071 was empty.
4072 [Steve Henson]
4073 [This change does not apply to 0.9.7.]
4074
4075 *) Use the cached encoding of an X509_NAME structure rather than
4076 copying it. This is apparently the reason for the libsafe "errors"
4077 but the code is actually correct.
4078 [Steve Henson]
4079
4080 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
4081 Bleichenbacher's DSA attack.
4082 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
4083 to be set and top=0 forces the highest bit to be set; top=-1 is new
4084 and leaves the highest bit random.
4085 [Ulf Moeller, Bodo Moeller]
4086
4087 *) In the NCONF_...-based implementations for CONF_... queries
4088 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
4089 a temporary CONF structure with the data component set to NULL
4090 (which gives segmentation faults in lh_retrieve).
4091 Instead, use NULL for the CONF pointer in CONF_get_string and
4092 CONF_get_number (which may use environment variables) and directly
4093 return NULL from CONF_get_section.
4094 [Bodo Moeller]
4095
4096 *) Fix potential buffer overrun for EBCDIC.
4097 [Ulf Moeller]
4098
4099 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
4100 keyUsage if basicConstraints absent for a CA.
4101 [Steve Henson]
4102
4103 *) Make SMIME_write_PKCS7() write mail header values with a format that
4104 is more generally accepted (no spaces before the semicolon), since
4105 some programs can't parse those values properly otherwise. Also make
4106 sure BIO's that break lines after each write do not create invalid
4107 headers.
4108 [Richard Levitte]
4109
4110 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
4111 macros previously used would not encode an empty SEQUENCE OF
4112 and break the signature.
4113 [Steve Henson]
4114 [This change does not apply to 0.9.7.]
4115
4116 *) Zero the premaster secret after deriving the master secret in
4117 DH ciphersuites.
4118 [Steve Henson]
4119
4120 *) Add some EVP_add_digest_alias registrations (as found in
4121 OpenSSL_add_all_digests()) to SSL_library_init()
4122 aka OpenSSL_add_ssl_algorithms(). This provides improved
4123 compatibility with peers using X.509 certificates
4124 with unconventional AlgorithmIdentifier OIDs.
4125 [Bodo Moeller]
4126
4127 *) Fix for Irix with NO_ASM.
4128 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4129
4130 *) ./config script fixes.
4131 [Ulf Moeller, Richard Levitte]
4132
4133 *) Fix 'openssl passwd -1'.
4134 [Bodo Moeller]
4135
4136 *) Change PKCS12_key_gen_asc() so it can cope with non null
4137 terminated strings whose length is passed in the passlen
4138 parameter, for example from PEM callbacks. This was done
4139 by adding an extra length parameter to asc2uni().
4140 [Steve Henson, reported by <oddissey@samsung.co.kr>]
4141
4142 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
4143 call failed, free the DSA structure.
4144 [Bodo Moeller]
4145
4146 *) Fix to uni2asc() to cope with zero length Unicode strings.
4147 These are present in some PKCS#12 files.
4148 [Steve Henson]
4149
4150 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
4151 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
4152 when writing a 32767 byte record.
4153 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
4154
4155 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
4156 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
4157
4158 (RSA objects have a reference count access to which is protected
4159 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
4160 so they are meant to be shared between threads.)
4161 [Bodo Moeller, Geoff Thorpe; original patch submitted by
4162 "Reddie, Steven" <Steven.Reddie@ca.com>]
4163
4164 *) Fix a deadlock in CRYPTO_mem_leaks().
4165 [Bodo Moeller]
4166
4167 *) Use better test patterns in bntest.
4168 [Ulf Möller]
4169
4170 *) rand_win.c fix for Borland C.
4171 [Ulf Möller]
4172
4173 *) BN_rshift bugfix for n == 0.
4174 [Bodo Moeller]
4175
4176 *) Add a 'bctest' script that checks for some known 'bc' bugs
4177 so that 'make test' does not abort just because 'bc' is broken.
4178 [Bodo Moeller]
4179
4180 *) Store verify_result within SSL_SESSION also for client side to
4181 avoid potential security hole. (Re-used sessions on the client side
4182 always resulted in verify_result==X509_V_OK, not using the original
4183 result of the server certificate verification.)
4184 [Lutz Jaenicke]
4185
4186 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
4187 SSL3_RT_APPLICATION_DATA, return 0.
4188 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
4189 [Bodo Moeller]
4190
4191 *) Fix SSL_peek:
4192 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
4193 releases, have been re-implemented by renaming the previous
4194 implementations of ssl2_read and ssl3_read to ssl2_read_internal
4195 and ssl3_read_internal, respectively, and adding 'peek' parameters
4196 to them. The new ssl[23]_{read,peek} functions are calls to
4197 ssl[23]_read_internal with the 'peek' flag set appropriately.
4198 A 'peek' parameter has also been added to ssl3_read_bytes, which
4199 does the actual work for ssl3_read_internal.
4200 [Bodo Moeller]
4201
4202 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
4203 the method-specific "init()" handler. Also clean up ex_data after
4204 calling the method-specific "finish()" handler. Previously, this was
4205 happening the other way round.
4206 [Geoff Thorpe]
4207
4208 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
4209 The previous value, 12, was not always sufficient for BN_mod_exp().
4210 [Bodo Moeller]
4211
4212 *) Make sure that shared libraries get the internal name engine with
4213 the full version number and not just 0. This should mark the
4214 shared libraries as not backward compatible. Of course, this should
4215 be changed again when we can guarantee backward binary compatibility.
4216 [Richard Levitte]
4217
4218 *) Fix typo in get_cert_by_subject() in by_dir.c
4219 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
4220
4221 *) Rework the system to generate shared libraries:
4222
4223 - Make note of the expected extension for the shared libraries and
4224 if there is a need for symbolic links from for example libcrypto.so.0
4225 to libcrypto.so.0.9.7. There is extended info in Configure for
4226 that.
4227
4228 - Make as few rebuilds of the shared libraries as possible.
4229
4230 - Still avoid linking the OpenSSL programs with the shared libraries.
4231
4232 - When installing, install the shared libraries separately from the
4233 static ones.
4234 [Richard Levitte]
4235
4236 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
4237
4238 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
4239 and not in SSL_clear because the latter is also used by the
4240 accept/connect functions; previously, the settings made by
4241 SSL_set_read_ahead would be lost during the handshake.
4242 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
4243
4244 *) Correct util/mkdef.pl to be selective about disabled algorithms.
4245 Previously, it would create entries for disableed algorithms no
4246 matter what.
4247 [Richard Levitte]
4248
4249 *) Added several new manual pages for SSL_* function.
4250 [Lutz Jaenicke]
4251
4252 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
4253
4254 *) In ssl23_get_client_hello, generate an error message when faced
4255 with an initial SSL 3.0/TLS record that is too small to contain the
4256 first two bytes of the ClientHello message, i.e. client_version.
4257 (Note that this is a pathologic case that probably has never happened
4258 in real life.) The previous approach was to use the version number
4259 from the record header as a substitute; but our protocol choice
4260 should not depend on that one because it is not authenticated
4261 by the Finished messages.
4262 [Bodo Moeller]
4263
4264 *) More robust randomness gathering functions for Windows.
4265 [Jeffrey Altman <jaltman@columbia.edu>]
4266
4267 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
4268 not set then we don't setup the error code for issuer check errors
4269 to avoid possibly overwriting other errors which the callback does
4270 handle. If an application does set the flag then we assume it knows
4271 what it is doing and can handle the new informational codes
4272 appropriately.
4273 [Steve Henson]
4274
4275 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
4276 a general "ANY" type, as such it should be able to decode anything
4277 including tagged types. However it didn't check the class so it would
4278 wrongly interpret tagged types in the same way as their universal
4279 counterpart and unknown types were just rejected. Changed so that the
4280 tagged and unknown types are handled in the same way as a SEQUENCE:
4281 that is the encoding is stored intact. There is also a new type
4282 "V_ASN1_OTHER" which is used when the class is not universal, in this
4283 case we have no idea what the actual type is so we just lump them all
4284 together.
4285 [Steve Henson]
4286
4287 *) On VMS, stdout may very well lead to a file that is written to
4288 in a record-oriented fashion. That means that every write() will
4289 write a separate record, which will be read separately by the
4290 programs trying to read from it. This can be very confusing.
4291
4292 The solution is to put a BIO filter in the way that will buffer
4293 text until a linefeed is reached, and then write everything a
4294 line at a time, so every record written will be an actual line,
4295 not chunks of lines and not (usually doesn't happen, but I've
4296 seen it once) several lines in one record. BIO_f_linebuffer() is
4297 the answer.
4298
4299 Currently, it's a VMS-only method, because that's where it has
4300 been tested well enough.
4301 [Richard Levitte]
4302
4303 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
4304 it can return incorrect results.
4305 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
4306 but it was in 0.9.6-beta[12].)
4307 [Bodo Moeller]
4308
4309 *) Disable the check for content being present when verifying detached
4310 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
4311 include zero length content when signing messages.
4312 [Steve Henson]
4313
4314 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
4315 BIO_ctrl (for BIO pairs).
4316 [Bodo Möller]
4317
4318 *) Add DSO method for VMS.
4319 [Richard Levitte]
4320
4321 *) Bug fix: Montgomery multiplication could produce results with the
4322 wrong sign.
4323 [Ulf Möller]
4324
4325 *) Add RPM specification openssl.spec and modify it to build three
4326 packages. The default package contains applications, application
4327 documentation and run-time libraries. The devel package contains
4328 include files, static libraries and function documentation. The
4329 doc package contains the contents of the doc directory. The original
4330 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
4331 [Richard Levitte]
4332
4333 *) Add a large number of documentation files for many SSL routines.
4334 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
4335
4336 *) Add a configuration entry for Sony News 4.
4337 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
4338
4339 *) Don't set the two most significant bits to one when generating a
4340 random number < q in the DSA library.
4341 [Ulf Möller]
4342
4343 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
4344 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
4345 the underlying transport is blocking) if a handshake took place.
4346 (The default behaviour is needed by applications such as s_client
4347 and s_server that use select() to determine when to use SSL_read;
4348 but for applications that know in advance when to expect data, it
4349 just makes things more complicated.)
4350 [Bodo Moeller]
4351
4352 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
4353 from EGD.
4354 [Ben Laurie]
4355
4356 *) Add a few more EBCDIC conditionals that make `req' and `x509'
4357 work better on such systems.
4358 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4359
4360 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
4361 Update PKCS12_parse() so it copies the friendlyName and the
4362 keyid to the certificates aux info.
4363 [Steve Henson]
4364
4365 *) Fix bug in PKCS7_verify() which caused an infinite loop
4366 if there was more than one signature.
4367 [Sven Uszpelkat <su@celocom.de>]
4368
4369 *) Major change in util/mkdef.pl to include extra information
4370 about each symbol, as well as presentig variables as well
4371 as functions. This change means that there's n more need
4372 to rebuild the .num files when some algorithms are excluded.
4373 [Richard Levitte]
4374
4375 *) Allow the verify time to be set by an application,
4376 rather than always using the current time.
4377 [Steve Henson]
4378
4379 *) Phase 2 verify code reorganisation. The certificate
4380 verify code now looks up an issuer certificate by a
4381 number of criteria: subject name, authority key id
4382 and key usage. It also verifies self signed certificates
4383 by the same criteria. The main comparison function is
4384 X509_check_issued() which performs these checks.
4385
4386 Lot of changes were necessary in order to support this
4387 without completely rewriting the lookup code.
4388
4389 Authority and subject key identifier are now cached.
4390
4391 The LHASH 'certs' is X509_STORE has now been replaced
4392 by a STACK_OF(X509_OBJECT). This is mainly because an
4393 LHASH can't store or retrieve multiple objects with
4394 the same hash value.
4395
4396 As a result various functions (which were all internal
4397 use only) have changed to handle the new X509_STORE
4398 structure. This will break anything that messed round
4399 with X509_STORE internally.
4400
4401 The functions X509_STORE_add_cert() now checks for an
4402 exact match, rather than just subject name.
4403
4404 The X509_STORE API doesn't directly support the retrieval
4405 of multiple certificates matching a given criteria, however
4406 this can be worked round by performing a lookup first
4407 (which will fill the cache with candidate certificates)
4408 and then examining the cache for matches. This is probably
4409 the best we can do without throwing out X509_LOOKUP
4410 entirely (maybe later...).
4411
4412 The X509_VERIFY_CTX structure has been enhanced considerably.
4413
4414 All certificate lookup operations now go via a get_issuer()
4415 callback. Although this currently uses an X509_STORE it
4416 can be replaced by custom lookups. This is a simple way
4417 to bypass the X509_STORE hackery necessary to make this
4418 work and makes it possible to use more efficient techniques
4419 in future. A very simple version which uses a simple
4420 STACK for its trusted certificate store is also provided
4421 using X509_STORE_CTX_trusted_stack().
4422
4423 The verify_cb() and verify() callbacks now have equivalents
4424 in the X509_STORE_CTX structure.
4425
4426 X509_STORE_CTX also has a 'flags' field which can be used
4427 to customise the verify behaviour.
4428 [Steve Henson]
4429
4430 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
4431 excludes S/MIME capabilities.
4432 [Steve Henson]
4433
4434 *) When a certificate request is read in keep a copy of the
4435 original encoding of the signed data and use it when outputing
4436 again. Signatures then use the original encoding rather than
4437 a decoded, encoded version which may cause problems if the
4438 request is improperly encoded.
4439 [Steve Henson]
4440
4441 *) For consistency with other BIO_puts implementations, call
4442 buffer_write(b, ...) directly in buffer_puts instead of calling
4443 BIO_write(b, ...).
4444
4445 In BIO_puts, increment b->num_write as in BIO_write.
4446 [Peter.Sylvester@EdelWeb.fr]
4447
4448 *) Fix BN_mul_word for the case where the word is 0. (We have to use
4449 BN_zero, we may not return a BIGNUM with an array consisting of
4450 words set to zero.)
4451 [Bodo Moeller]
4452
4453 *) Avoid calling abort() from within the library when problems are
4454 detected, except if preprocessor symbols have been defined
4455 (such as REF_CHECK, BN_DEBUG etc.).
4456 [Bodo Moeller]
4457
4458 *) New openssl application 'rsautl'. This utility can be
4459 used for low level RSA operations. DER public key
4460 BIO/fp routines also added.
4461 [Steve Henson]
4462
4463 *) New Configure entry and patches for compiling on QNX 4.
4464 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
4465
4466 *) A demo state-machine implementation was sponsored by
4467 Nuron (http://www.nuron.com/) and is now available in
4468 demos/state_machine.
4469 [Ben Laurie]
4470
4471 *) New options added to the 'dgst' utility for signature
4472 generation and verification.
4473 [Steve Henson]
4474
4475 *) Unrecognized PKCS#7 content types are now handled via a
4476 catch all ASN1_TYPE structure. This allows unsupported
4477 types to be stored as a "blob" and an application can
4478 encode and decode it manually.
4479 [Steve Henson]
4480
4481 *) Fix various signed/unsigned issues to make a_strex.c
4482 compile under VC++.
4483 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
4484
4485 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
4486 length if passed a buffer. ASN1_INTEGER_to_BN failed
4487 if passed a NULL BN and its argument was negative.
4488 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
4489
4490 *) Modification to PKCS#7 encoding routines to output definite
4491 length encoding. Since currently the whole structures are in
4492 memory there's not real point in using indefinite length
4493 constructed encoding. However if OpenSSL is compiled with
4494 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
4495 [Steve Henson]
4496
4497 *) Added BIO_vprintf() and BIO_vsnprintf().
4498 [Richard Levitte]
4499
4500 *) Added more prefixes to parse for in the the strings written
4501 through a logging bio, to cover all the levels that are available
4502 through syslog. The prefixes are now:
4503
4504 PANIC, EMERG, EMR => LOG_EMERG
4505 ALERT, ALR => LOG_ALERT
4506 CRIT, CRI => LOG_CRIT
4507 ERROR, ERR => LOG_ERR
4508 WARNING, WARN, WAR => LOG_WARNING
4509 NOTICE, NOTE, NOT => LOG_NOTICE
4510 INFO, INF => LOG_INFO
4511 DEBUG, DBG => LOG_DEBUG
4512
4513 and as before, if none of those prefixes are present at the
4514 beginning of the string, LOG_ERR is chosen.
4515
4516 On Win32, the LOG_* levels are mapped according to this:
4517
4518 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
4519 LOG_WARNING => EVENTLOG_WARNING_TYPE
4520 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
4521
4522 [Richard Levitte]
4523
4524 *) Made it possible to reconfigure with just the configuration
4525 argument "reconf" or "reconfigure". The command line arguments
4526 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
4527 and are retrieved from there when reconfiguring.
4528 [Richard Levitte]
4529
4530 *) MD4 implemented.
4531 [Assar Westerlund <assar@sics.se>, Richard Levitte]
4532
4533 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
4534 [Richard Levitte]
4535
4536 *) The obj_dat.pl script was messing up the sorting of object
4537 names. The reason was that it compared the quoted version
4538 of strings as a result "OCSP" > "OCSP Signing" because
4539 " > SPACE. Changed script to store unquoted versions of
4540 names and add quotes on output. It was also omitting some
4541 names from the lookup table if they were given a default
4542 value (that is if SN is missing it is given the same
4543 value as LN and vice versa), these are now added on the
4544 grounds that if an object has a name we should be able to
4545 look it up. Finally added warning output when duplicate
4546 short or long names are found.
4547 [Steve Henson]
4548
4549 *) Changes needed for Tandem NSK.
4550 [Scott Uroff <scott@xypro.com>]
4551
4552 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
4553 RSA_padding_check_SSLv23(), special padding was never detected
4554 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
4555 version rollback attacks was not effective.
4556
4557 In s23_clnt.c, don't use special rollback-attack detection padding
4558 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
4559 client; similarly, in s23_srvr.c, don't do the rollback check if
4560 SSL 2.0 is the only protocol enabled in the server.
4561 [Bodo Moeller]
4562
4563 *) Make it possible to get hexdumps of unprintable data with 'openssl
4564 asn1parse'. By implication, the functions ASN1_parse_dump() and
4565 BIO_dump_indent() are added.
4566 [Richard Levitte]
4567
4568 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
4569 these print out strings and name structures based on various
4570 flags including RFC2253 support and proper handling of
4571 multibyte characters. Added options to the 'x509' utility
4572 to allow the various flags to be set.
4573 [Steve Henson]
4574
4575 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
4576 Also change the functions X509_cmp_current_time() and
4577 X509_gmtime_adj() work with an ASN1_TIME structure,
4578 this will enable certificates using GeneralizedTime in validity
4579 dates to be checked.
4580 [Steve Henson]
4581
4582 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
4583 negative public key encodings) on by default,
4584 NO_NEG_PUBKEY_BUG can be set to disable it.
4585 [Steve Henson]
4586
4587 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
4588 content octets. An i2c_ASN1_OBJECT is unnecessary because
4589 the encoding can be trivially obtained from the structure.
4590 [Steve Henson]
4591
4592 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
4593 not read locks (CRYPTO_r_[un]lock).
4594 [Bodo Moeller]
4595
4596 *) A first attempt at creating official support for shared
4597 libraries through configuration. I've kept it so the
4598 default is static libraries only, and the OpenSSL programs
4599 are always statically linked for now, but there are
4600 preparations for dynamic linking in place.
4601 This has been tested on Linux and Tru64.
4602 [Richard Levitte]
4603
4604 *) Randomness polling function for Win9x, as described in:
4605 Peter Gutmann, Software Generation of Practically Strong
4606 Random Numbers.
4607 [Ulf Möller]
4608
4609 *) Fix so PRNG is seeded in req if using an already existing
4610 DSA key.
4611 [Steve Henson]
4612
4613 *) New options to smime application. -inform and -outform
4614 allow alternative formats for the S/MIME message including
4615 PEM and DER. The -content option allows the content to be
4616 specified separately. This should allow things like Netscape
4617 form signing output easier to verify.
4618 [Steve Henson]
4619
4620 *) Fix the ASN1 encoding of tags using the 'long form'.
4621 [Steve Henson]
4622
4623 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
4624 STRING types. These convert content octets to and from the
4625 underlying type. The actual tag and length octets are
4626 already assumed to have been read in and checked. These
4627 are needed because all other string types have virtually
4628 identical handling apart from the tag. By having versions
4629 of the ASN1 functions that just operate on content octets
4630 IMPLICIT tagging can be handled properly. It also allows
4631 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
4632 and ASN1_INTEGER are identical apart from the tag.
4633 [Steve Henson]
4634
4635 *) Change the handling of OID objects as follows:
4636
4637 - New object identifiers are inserted in objects.txt, following
4638 the syntax given in objects.README.
4639 - objects.pl is used to process obj_mac.num and create a new
4640 obj_mac.h.
4641 - obj_dat.pl is used to create a new obj_dat.h, using the data in
4642 obj_mac.h.
4643
4644 This is currently kind of a hack, and the perl code in objects.pl
4645 isn't very elegant, but it works as I intended. The simplest way
4646 to check that it worked correctly is to look in obj_dat.h and
4647 check the array nid_objs and make sure the objects haven't moved
4648 around (this is important!). Additions are OK, as well as
4649 consistent name changes.
4650 [Richard Levitte]
4651
4652 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
4653 [Bodo Moeller]
4654
4655 *) Addition of the command line parameter '-rand file' to 'openssl req'.
4656 The given file adds to whatever has already been seeded into the
4657 random pool through the RANDFILE configuration file option or
4658 environment variable, or the default random state file.
4659 [Richard Levitte]
4660
4661 *) mkstack.pl now sorts each macro group into lexical order.
4662 Previously the output order depended on the order the files
4663 appeared in the directory, resulting in needless rewriting
4664 of safestack.h .
4665 [Steve Henson]
4666
4667 *) Patches to make OpenSSL compile under Win32 again. Mostly
4668 work arounds for the VC++ problem that it treats func() as
4669 func(void). Also stripped out the parts of mkdef.pl that
4670 added extra typesafe functions: these no longer exist.
4671 [Steve Henson]
4672
4673 *) Reorganisation of the stack code. The macros are now all
4674 collected in safestack.h . Each macro is defined in terms of
4675 a "stack macro" of the form SKM_<name>(type, a, b). The
4676 DEBUG_SAFESTACK is now handled in terms of function casts,
4677 this has the advantage of retaining type safety without the
4678 use of additional functions. If DEBUG_SAFESTACK is not defined
4679 then the non typesafe macros are used instead. Also modified the
4680 mkstack.pl script to handle the new form. Needs testing to see
4681 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
4682 the default if no major problems. Similar behaviour for ASN1_SET_OF
4683 and PKCS12_STACK_OF.
4684 [Steve Henson]
4685
4686 *) When some versions of IIS use the 'NET' form of private key the
4687 key derivation algorithm is different. Normally MD5(password) is
4688 used as a 128 bit RC4 key. In the modified case
4689 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
4690 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
4691 as the old Netscape_RSA functions except they have an additional
4692 'sgckey' parameter which uses the modified algorithm. Also added
4693 an -sgckey command line option to the rsa utility. Thanks to
4694 Adrian Peck <bertie@ncipher.com> for posting details of the modified
4695 algorithm to openssl-dev.
4696 [Steve Henson]
4697
4698 *) The evp_local.h macros were using 'c.##kname' which resulted in
4699 invalid expansion on some systems (SCO 5.0.5 for example).
4700 Corrected to 'c.kname'.
4701 [Phillip Porch <root@theporch.com>]
4702
4703 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
4704 a STACK of email addresses from a certificate or request, these look
4705 in the subject name and the subject alternative name extensions and
4706 omit any duplicate addresses.
4707 [Steve Henson]
4708
4709 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
4710 This makes DSA verification about 2 % faster.
4711 [Bodo Moeller]
4712
4713 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
4714 (meaning that now 2^5 values will be precomputed, which is only 4 KB
4715 plus overhead for 1024 bit moduli).
4716 This makes exponentiations about 0.5 % faster for 1024 bit
4717 exponents (as measured by "openssl speed rsa2048").
4718 [Bodo Moeller]
4719
4720 *) Rename memory handling macros to avoid conflicts with other
4721 software:
4722 Malloc => OPENSSL_malloc
4723 Malloc_locked => OPENSSL_malloc_locked
4724 Realloc => OPENSSL_realloc
4725 Free => OPENSSL_free
4726 [Richard Levitte]
4727
4728 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
4729 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
4730 [Bodo Moeller]
4731
4732 *) CygWin32 support.
4733 [John Jarvie <jjarvie@newsguy.com>]
4734
4735 *) The type-safe stack code has been rejigged. It is now only compiled
4736 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
4737 by default all type-specific stack functions are "#define"d back to
4738 standard stack functions. This results in more streamlined output
4739 but retains the type-safety checking possibilities of the original
4740 approach.
4741 [Geoff Thorpe]
4742
4743 *) The STACK code has been cleaned up, and certain type declarations
4744 that didn't make a lot of sense have been brought in line. This has
4745 also involved a cleanup of sorts in safestack.h to more correctly
4746 map type-safe stack functions onto their plain stack counterparts.
4747 This work has also resulted in a variety of "const"ifications of
4748 lots of the code, especially "_cmp" operations which should normally
4749 be prototyped with "const" parameters anyway.
4750 [Geoff Thorpe]
4751
4752 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
4753 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
4754 (The PRNG state consists of two parts, the large pool 'state' and 'md',
4755 where all of 'md' is used each time the PRNG is used, but 'state'
4756 is used only indexed by a cyclic counter. As entropy may not be
4757 well distributed from the beginning, 'md' is important as a
4758 chaining variable. However, the output function chains only half
4759 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
4760 all of 'md', and seeding with STATE_SIZE dummy bytes will result
4761 in all of 'state' being rewritten, with the new values depending
4762 on virtually all of 'md'. This overcomes the 80 bit limitation.)
4763 [Bodo Moeller]
4764
4765 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
4766 the handshake is continued after ssl_verify_cert_chain();
4767 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
4768 can lead to 'unexplainable' connection aborts later.
4769 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
4770
4771 *) Major EVP API cipher revision.
4772 Add hooks for extra EVP features. This allows various cipher
4773 parameters to be set in the EVP interface. Support added for variable
4774 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
4775 setting of RC2 and RC5 parameters.
4776
4777 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
4778 ciphers.
4779
4780 Remove lots of duplicated code from the EVP library. For example *every*
4781 cipher init() function handles the 'iv' in the same way according to the
4782 cipher mode. They also all do nothing if the 'key' parameter is NULL and
4783 for CFB and OFB modes they zero ctx->num.
4784
4785 New functionality allows removal of S/MIME code RC2 hack.
4786
4787 Most of the routines have the same form and so can be declared in terms
4788 of macros.
4789
4790 By shifting this to the top level EVP_CipherInit() it can be removed from
4791 all individual ciphers. If the cipher wants to handle IVs or keys
4792 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
4793 flags.
4794
4795 Change lots of functions like EVP_EncryptUpdate() to now return a
4796 value: although software versions of the algorithms cannot fail
4797 any installed hardware versions can.
4798 [Steve Henson]
4799
4800 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
4801 this option is set, tolerate broken clients that send the negotiated
4802 protocol version number instead of the requested protocol version
4803 number.
4804 [Bodo Moeller]
4805
4806 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
4807 i.e. non-zero for export ciphersuites, zero otherwise.
4808 Previous versions had this flag inverted, inconsistent with
4809 rsa_tmp_cb (..._TMP_RSA_CB).
4810 [Bodo Moeller; problem reported by Amit Chopra]
4811
4812 *) Add missing DSA library text string. Work around for some IIS
4813 key files with invalid SEQUENCE encoding.
4814 [Steve Henson]
4815
4816 *) Add a document (doc/standards.txt) that list all kinds of standards
4817 and so on that are implemented in OpenSSL.
4818 [Richard Levitte]
4819
4820 *) Enhance c_rehash script. Old version would mishandle certificates
4821 with the same subject name hash and wouldn't handle CRLs at all.
4822 Added -fingerprint option to crl utility, to support new c_rehash
4823 features.
4824 [Steve Henson]
4825
4826 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
4827 [Ulf Möller]
4828
4829 *) Fix for SSL server purpose checking. Server checking was
4830 rejecting certificates which had extended key usage present
4831 but no ssl client purpose.
4832 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
4833
4834 *) Make PKCS#12 code work with no password. The PKCS#12 spec
4835 is a little unclear about how a blank password is handled.
4836 Since the password in encoded as a BMPString with terminating
4837 double NULL a zero length password would end up as just the
4838 double NULL. However no password at all is different and is
4839 handled differently in the PKCS#12 key generation code. NS
4840 treats a blank password as zero length. MSIE treats it as no
4841 password on export: but it will try both on import. We now do
4842 the same: PKCS12_parse() tries zero length and no password if
4843 the password is set to "" or NULL (NULL is now a valid password:
4844 it wasn't before) as does the pkcs12 application.
4845 [Steve Henson]
4846
4847 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
4848 perror when PEM_read_bio_X509_REQ fails, the error message must
4849 be obtained from the error queue.
4850 [Bodo Moeller]
4851
4852 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
4853 it in ERR_remove_state if appropriate, and change ERR_get_state
4854 accordingly to avoid race conditions (this is necessary because
4855 thread_hash is no longer constant once set).
4856 [Bodo Moeller]
4857
4858 *) Bugfix for linux-elf makefile.one.
4859 [Ulf Möller]
4860
4861 *) RSA_get_default_method() will now cause a default
4862 RSA_METHOD to be chosen if one doesn't exist already.
4863 Previously this was only set during a call to RSA_new()
4864 or RSA_new_method(NULL) meaning it was possible for
4865 RSA_get_default_method() to return NULL.
4866 [Geoff Thorpe]
4867
4868 *) Added native name translation to the existing DSO code
4869 that will convert (if the flag to do so is set) filenames
4870 that are sufficiently small and have no path information
4871 into a canonical native form. Eg. "blah" converted to
4872 "libblah.so" or "blah.dll" etc.
4873 [Geoff Thorpe]
4874
4875 *) New function ERR_error_string_n(e, buf, len) which is like
4876 ERR_error_string(e, buf), but writes at most 'len' bytes
4877 including the 0 terminator. For ERR_error_string_n, 'buf'
4878 may not be NULL.
4879 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
4880
4881 *) CONF library reworked to become more general. A new CONF
4882 configuration file reader "class" is implemented as well as a
4883 new functions (NCONF_*, for "New CONF") to handle it. The now
4884 old CONF_* functions are still there, but are reimplemented to
4885 work in terms of the new functions. Also, a set of functions
4886 to handle the internal storage of the configuration data is
4887 provided to make it easier to write new configuration file
4888 reader "classes" (I can definitely see something reading a
4889 configuration file in XML format, for example), called _CONF_*,
4890 or "the configuration storage API"...
4891
4892 The new configuration file reading functions are:
4893
4894 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
4895 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
4896
4897 NCONF_default, NCONF_WIN32
4898
4899 NCONF_dump_fp, NCONF_dump_bio
4900
4901 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
4902 NCONF_new creates a new CONF object. This works in the same way
4903 as other interfaces in OpenSSL, like the BIO interface.
4904 NCONF_dump_* dump the internal storage of the configuration file,
4905 which is useful for debugging. All other functions take the same
4906 arguments as the old CONF_* functions wth the exception of the
4907 first that must be a `CONF *' instead of a `LHASH *'.
4908
4909 To make it easer to use the new classes with the old CONF_* functions,
4910 the function CONF_set_default_method is provided.
4911 [Richard Levitte]
4912
4913 *) Add '-tls1' option to 'openssl ciphers', which was already
4914 mentioned in the documentation but had not been implemented.
4915 (This option is not yet really useful because even the additional
4916 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
4917 [Bodo Moeller]
4918
4919 *) Initial DSO code added into libcrypto for letting OpenSSL (and
4920 OpenSSL-based applications) load shared libraries and bind to
4921 them in a portable way.
4922 [Geoff Thorpe, with contributions from Richard Levitte]
4923
4924 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
4925
4926 *) Make sure _lrotl and _lrotr are only used with MSVC.
4927
4928 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
4929 (the default implementation of RAND_status).
4930
4931 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
4932 to '-clrext' (= clear extensions), as intended and documented.
4933 [Bodo Moeller; inconsistency pointed out by Michael Attili
4934 <attili@amaxo.com>]
4935
4936 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
4937 was larger than the MD block size.
4938 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
4939
4940 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
4941 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
4942 using the passed key: if the passed key was a private key the result
4943 of X509_print(), for example, would be to print out all the private key
4944 components.
4945 [Steve Henson]
4946
4947 *) des_quad_cksum() byte order bug fix.
4948 [Ulf Möller, using the problem description in krb4-0.9.7, where
4949 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
4950
4951 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
4952 discouraged.
4953 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
4954
4955 *) For easily testing in shell scripts whether some command
4956 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
4957 returns with exit code 0 iff no command of the given name is available.
4958 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
4959 the output goes to stdout and nothing is printed to stderr.
4960 Additional arguments are always ignored.
4961
4962 Since for each cipher there is a command of the same name,
4963 the 'no-cipher' compilation switches can be tested this way.
4964
4965 ('openssl no-XXX' is not able to detect pseudo-commands such
4966 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
4967 [Bodo Moeller]
4968
4969 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
4970 [Bodo Moeller]
4971
4972 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
4973 is set; it will be thrown away anyway because each handshake creates
4974 its own key.
4975 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
4976 to parameters -- in previous versions (since OpenSSL 0.9.3) the
4977 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
4978 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
4979 [Bodo Moeller]
4980
4981 *) New s_client option -ign_eof: EOF at stdin is ignored, and
4982 'Q' and 'R' lose their special meanings (quit/renegotiate).
4983 This is part of what -quiet does; unlike -quiet, -ign_eof
4984 does not suppress any output.
4985 [Richard Levitte]
4986
4987 *) Add compatibility options to the purpose and trust code. The
4988 purpose X509_PURPOSE_ANY is "any purpose" which automatically
4989 accepts a certificate or CA, this was the previous behaviour,
4990 with all the associated security issues.
4991
4992 X509_TRUST_COMPAT is the old trust behaviour: only and
4993 automatically trust self signed roots in certificate store. A
4994 new trust setting X509_TRUST_DEFAULT is used to specify that
4995 a purpose has no associated trust setting and it should instead
4996 use the value in the default purpose.
4997 [Steve Henson]
4998
4999 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5000 and fix a memory leak.
5001 [Steve Henson]
5002
5003 *) In util/mkerr.pl (which implements 'make errors'), preserve
5004 reason strings from the previous version of the .c file, as
5005 the default to have only downcase letters (and digits) in
5006 automatically generated reasons codes is not always appropriate.
5007 [Bodo Moeller]
5008
5009 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5010 using strerror. Previously, ERR_reason_error_string() returned
5011 library names as reason strings for SYSerr; but SYSerr is a special
5012 case where small numbers are errno values, not library numbers.
5013 [Bodo Moeller]
5014
5015 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5016 converts DSA parameters into DH parameters. (When creating parameters,
5017 DSA_generate_parameters is used.)
5018 [Bodo Moeller]
5019
5020 *) Include 'length' (recommended exponent length) in C code generated
5021 by 'openssl dhparam -C'.
5022 [Bodo Moeller]
5023
5024 *) The second argument to set_label in perlasm was already being used
5025 so couldn't be used as a "file scope" flag. Moved to third argument
5026 which was free.
5027 [Steve Henson]
5028
5029 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5030 instead of RAND_bytes for encryption IVs and salts.
5031 [Bodo Moeller]
5032
5033 *) Include RAND_status() into RAND_METHOD instead of implementing
5034 it only for md_rand.c Otherwise replacing the PRNG by calling
5035 RAND_set_rand_method would be impossible.
5036 [Bodo Moeller]
5037
5038 *) Don't let DSA_generate_key() enter an infinite loop if the random
5039 number generation fails.
5040 [Bodo Moeller]
5041
5042 *) New 'rand' application for creating pseudo-random output.
5043 [Bodo Moeller]
5044
5045 *) Added configuration support for Linux/IA64
5046 [Rolf Haberrecker <rolf@suse.de>]
5047
5048 *) Assembler module support for Mingw32.
5049 [Ulf Möller]
5050
5051 *) Shared library support for HPUX (in shlib/).
5052 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5053
5054 *) Shared library support for Solaris gcc.
5055 [Lutz Behnke <behnke@trustcenter.de>]
5056
5057 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5058
5059 *) PKCS7_encrypt() was adding text MIME headers twice because they
5060 were added manually and by SMIME_crlf_copy().
5061 [Steve Henson]
5062
5063 *) In bntest.c don't call BN_rand with zero bits argument.
5064 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5065
5066 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5067 case was implemented. This caused BN_div_recp() to fail occasionally.
5068 [Ulf Möller]
5069
5070 *) Add an optional second argument to the set_label() in the perl
5071 assembly language builder. If this argument exists and is set
5072 to 1 it signals that the assembler should use a symbol whose
5073 scope is the entire file, not just the current function. This
5074 is needed with MASM which uses the format label:: for this scope.
5075 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
5076
5077 *) Change the ASN1 types so they are typedefs by default. Before
5078 almost all types were #define'd to ASN1_STRING which was causing
5079 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
5080 for example.
5081 [Steve Henson]
5082
5083 *) Change names of new functions to the new get1/get0 naming
5084 convention: After 'get1', the caller owns a reference count
5085 and has to call ..._free; 'get0' returns a pointer to some
5086 data structure without incrementing reference counters.
5087 (Some of the existing 'get' functions increment a reference
5088 counter, some don't.)
5089 Similarly, 'set1' and 'add1' functions increase reference
5090 counters or duplicate objects.
5091 [Steve Henson]
5092
5093 *) Allow for the possibility of temp RSA key generation failure:
5094 the code used to assume it always worked and crashed on failure.
5095 [Steve Henson]
5096
5097 *) Fix potential buffer overrun problem in BIO_printf().
5098 [Ulf Möller, using public domain code by Patrick Powell; problem
5099 pointed out by David Sacerdote <das33@cornell.edu>]
5100
5101 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
5102 RAND_egd() and RAND_status(). In the command line application,
5103 the EGD socket can be specified like a seed file using RANDFILE
5104 or -rand.
5105 [Ulf Möller]
5106
5107 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
5108 Some CAs (e.g. Verisign) distribute certificates in this form.
5109 [Steve Henson]
5110
5111 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
5112 list to exclude them. This means that no special compilation option
5113 is needed to use anonymous DH: it just needs to be included in the
5114 cipher list.
5115 [Steve Henson]
5116
5117 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
5118 EVP_MD_type. The old functionality is available in a new macro called
5119 EVP_MD_md(). Change code that uses it and update docs.
5120 [Steve Henson]
5121
5122 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
5123 where the 'void *' argument is replaced by a function pointer argument.
5124 Previously 'void *' was abused to point to functions, which works on
5125 many platforms, but is not correct. As these functions are usually
5126 called by macros defined in OpenSSL header files, most source code
5127 should work without changes.
5128 [Richard Levitte]
5129
5130 *) <openssl/opensslconf.h> (which is created by Configure) now contains
5131 sections with information on -D... compiler switches used for
5132 compiling the library so that applications can see them. To enable
5133 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
5134 must be defined. E.g.,
5135 #define OPENSSL_ALGORITHM_DEFINES
5136 #include <openssl/opensslconf.h>
5137 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
5138 [Richard Levitte, Ulf and Bodo Möller]
5139
5140 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
5141 record layer.
5142 [Bodo Moeller]
5143
5144 *) Change the 'other' type in certificate aux info to a STACK_OF
5145 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
5146 the required ASN1 format: arbitrary types determined by an OID.
5147 [Steve Henson]
5148
5149 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
5150 argument to 'req'. This is not because the function is newer or
5151 better than others it just uses the work 'NEW' in the certificate
5152 request header lines. Some software needs this.
5153 [Steve Henson]
5154
5155 *) Reorganise password command line arguments: now passwords can be
5156 obtained from various sources. Delete the PEM_cb function and make
5157 it the default behaviour: i.e. if the callback is NULL and the
5158 usrdata argument is not NULL interpret it as a null terminated pass
5159 phrase. If usrdata and the callback are NULL then the pass phrase
5160 is prompted for as usual.
5161 [Steve Henson]
5162
5163 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
5164 the support is automatically enabled. The resulting binaries will
5165 autodetect the card and use it if present.
5166 [Ben Laurie and Compaq Inc.]
5167
5168 *) Work around for Netscape hang bug. This sends certificate request
5169 and server done in one record. Since this is perfectly legal in the
5170 SSL/TLS protocol it isn't a "bug" option and is on by default. See
5171 the bugs/SSLv3 entry for more info.
5172 [Steve Henson]
5173
5174 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
5175 [Andy Polyakov]
5176
5177 *) Add -rand argument to smime and pkcs12 applications and read/write
5178 of seed file.
5179 [Steve Henson]
5180
5181 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
5182 [Bodo Moeller]
5183
5184 *) Add command line password options to the remaining applications.
5185 [Steve Henson]
5186
5187 *) Bug fix for BN_div_recp() for numerators with an even number of
5188 bits.
5189 [Ulf Möller]
5190
5191 *) More tests in bntest.c, and changed test_bn output.
5192 [Ulf Möller]
5193
5194 *) ./config recognizes MacOS X now.
5195 [Andy Polyakov]
5196
5197 *) Bug fix for BN_div() when the first words of num and divsor are
5198 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
5199 [Ulf Möller]
5200
5201 *) Add support for various broken PKCS#8 formats, and command line
5202 options to produce them.
5203 [Steve Henson]
5204
5205 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
5206 get temporary BIGNUMs from a BN_CTX.
5207 [Ulf Möller]
5208
5209 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
5210 for p == 0.
5211 [Ulf Möller]
5212
5213 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
5214 include a #define from the old name to the new. The original intent
5215 was that statically linked binaries could for example just call
5216 SSLeay_add_all_ciphers() to just add ciphers to the table and not
5217 link with digests. This never worked becayse SSLeay_add_all_digests()
5218 and SSLeay_add_all_ciphers() were in the same source file so calling
5219 one would link with the other. They are now in separate source files.
5220 [Steve Henson]
5221
5222 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
5223 [Steve Henson]
5224
5225 *) Use a less unusual form of the Miller-Rabin primality test (it used
5226 a binary algorithm for exponentiation integrated into the Miller-Rabin
5227 loop, our standard modexp algorithms are faster).
5228 [Bodo Moeller]
5229
5230 *) Support for the EBCDIC character set completed.
5231 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
5232
5233 *) Source code cleanups: use const where appropriate, eliminate casts,
5234 use void * instead of char * in lhash.
5235 [Ulf Möller]
5236
5237 *) Bugfix: ssl3_send_server_key_exchange was not restartable
5238 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
5239 this the server could overwrite ephemeral keys that the client
5240 has already seen).
5241 [Bodo Moeller]
5242
5243 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
5244 using 50 iterations of the Rabin-Miller test.
5245
5246 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
5247 iterations of the Rabin-Miller test as required by the appendix
5248 to FIPS PUB 186[-1]) instead of DSA_is_prime.
5249 As BN_is_prime_fasttest includes trial division, DSA parameter
5250 generation becomes much faster.
5251
5252 This implies a change for the callback functions in DSA_is_prime
5253 and DSA_generate_parameters: The callback function is called once
5254 for each positive witness in the Rabin-Miller test, not just
5255 occasionally in the inner loop; and the parameters to the
5256 callback function now provide an iteration count for the outer
5257 loop rather than for the current invocation of the inner loop.
5258 DSA_generate_parameters additionally can call the callback
5259 function with an 'iteration count' of -1, meaning that a
5260 candidate has passed the trial division test (when q is generated
5261 from an application-provided seed, trial division is skipped).
5262 [Bodo Moeller]
5263
5264 *) New function BN_is_prime_fasttest that optionally does trial
5265 division before starting the Rabin-Miller test and has
5266 an additional BN_CTX * argument (whereas BN_is_prime always
5267 has to allocate at least one BN_CTX).
5268 'callback(1, -1, cb_arg)' is called when a number has passed the
5269 trial division stage.
5270 [Bodo Moeller]
5271
5272 *) Fix for bug in CRL encoding. The validity dates weren't being handled
5273 as ASN1_TIME.
5274 [Steve Henson]
5275
5276 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
5277 [Steve Henson]
5278
5279 *) New function BN_pseudo_rand().
5280 [Ulf Möller]
5281
5282 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
5283 bignum version of BN_from_montgomery() with the working code from
5284 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
5285 the comments.
5286 [Ulf Möller]
5287
5288 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
5289 made it impossible to use the same SSL_SESSION data structure in
5290 SSL2 clients in multiple threads.
5291 [Bodo Moeller]
5292
5293 *) The return value of RAND_load_file() no longer counts bytes obtained
5294 by stat(). RAND_load_file(..., -1) is new and uses the complete file
5295 to seed the PRNG (previously an explicit byte count was required).
5296 [Ulf Möller, Bodo Möller]
5297
5298 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
5299 used (char *) instead of (void *) and had casts all over the place.
5300 [Steve Henson]
5301
5302 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
5303 [Ulf Möller]
5304
5305 *) Retain source code compatibility for BN_prime_checks macro:
5306 BN_is_prime(..., BN_prime_checks, ...) now uses
5307 BN_prime_checks_for_size to determine the appropriate number of
5308 Rabin-Miller iterations.
5309 [Ulf Möller]
5310
5311 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
5312 DH_CHECK_P_NOT_SAFE_PRIME.
5313 (Check if this is true? OpenPGP calls them "strong".)
5314 [Ulf Möller]
5315
5316 *) Merge the functionality of "dh" and "gendh" programs into a new program
5317 "dhparam". The old programs are retained for now but will handle DH keys
5318 (instead of parameters) in future.
5319 [Steve Henson]
5320
5321 *) Make the ciphers, s_server and s_client programs check the return values
5322 when a new cipher list is set.
5323 [Steve Henson]
5324
5325 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
5326 ciphers. Before when the 56bit ciphers were enabled the sorting was
5327 wrong.
5328
5329 The syntax for the cipher sorting has been extended to support sorting by
5330 cipher-strength (using the strength_bits hard coded in the tables).
5331 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
5332
5333 Fix a bug in the cipher-command parser: when supplying a cipher command
5334 string with an "undefined" symbol (neither command nor alphanumeric
5335 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
5336 an error is flagged.
5337
5338 Due to the strength-sorting extension, the code of the
5339 ssl_create_cipher_list() function was completely rearranged. I hope that
5340 the readability was also increased :-)
5341 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5342
5343 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
5344 for the first serial number and places 2 in the serial number file. This
5345 avoids problems when the root CA is created with serial number zero and
5346 the first user certificate has the same issuer name and serial number
5347 as the root CA.
5348 [Steve Henson]
5349
5350 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
5351 the new code. Add documentation for this stuff.
5352 [Steve Henson]
5353
5354 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
5355 X509_*() to X509at_*() on the grounds that they don't handle X509
5356 structures and behave in an analagous way to the X509v3 functions:
5357 they shouldn't be called directly but wrapper functions should be used
5358 instead.
5359
5360 So we also now have some wrapper functions that call the X509at functions
5361 when passed certificate requests. (TO DO: similar things can be done with
5362 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
5363 things. Some of these need some d2i or i2d and print functionality
5364 because they handle more complex structures.)
5365 [Steve Henson]
5366
5367 *) Add missing #ifndefs that caused missing symbols when building libssl
5368 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
5369 NO_RSA in ssl/s2*.c.
5370 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
5371
5372 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
5373 has a return value which indicates the quality of the random data
5374 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
5375 error queue. New function RAND_pseudo_bytes() generates output that is
5376 guaranteed to be unique but not unpredictable. RAND_add is like
5377 RAND_seed, but takes an extra argument for an entropy estimate
5378 (RAND_seed always assumes full entropy).
5379 [Ulf Möller]
5380
5381 *) Do more iterations of Rabin-Miller probable prime test (specifically,
5382 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
5383 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
5384 in crypto/bn/bn_prime.c for the complete table). This guarantees a
5385 false-positive rate of at most 2^-80 for random input.
5386 [Bodo Moeller]
5387
5388 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
5389 [Bodo Moeller]
5390
5391 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
5392 in the 0.9.5 release), this returns the chain
5393 from an X509_CTX structure with a dup of the stack and all
5394 the X509 reference counts upped: so the stack will exist
5395 after X509_CTX_cleanup() has been called. Modify pkcs12.c
5396 to use this.
5397
5398 Also make SSL_SESSION_print() print out the verify return
5399 code.
5400 [Steve Henson]
5401
5402 *) Add manpage for the pkcs12 command. Also change the default
5403 behaviour so MAC iteration counts are used unless the new
5404 -nomaciter option is used. This improves file security and
5405 only older versions of MSIE (4.0 for example) need it.
5406 [Steve Henson]
5407
5408 *) Honor the no-xxx Configure options when creating .DEF files.
5409 [Ulf Möller]
5410
5411 *) Add PKCS#10 attributes to field table: challengePassword,
5412 unstructuredName and unstructuredAddress. These are taken from
5413 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
5414 international characters are used.
5415
5416 More changes to X509_ATTRIBUTE code: allow the setting of types
5417 based on strings. Remove the 'loc' parameter when adding
5418 attributes because these will be a SET OF encoding which is sorted
5419 in ASN1 order.
5420 [Steve Henson]
5421
5422 *) Initial changes to the 'req' utility to allow request generation
5423 automation. This will allow an application to just generate a template
5424 file containing all the field values and have req construct the
5425 request.
5426
5427 Initial support for X509_ATTRIBUTE handling. Stacks of these are
5428 used all over the place including certificate requests and PKCS#7
5429 structures. They are currently handled manually where necessary with
5430 some primitive wrappers for PKCS#7. The new functions behave in a
5431 manner analogous to the X509 extension functions: they allow
5432 attributes to be looked up by NID and added.
5433
5434 Later something similar to the X509V3 code would be desirable to
5435 automatically handle the encoding, decoding and printing of the
5436 more complex types. The string types like challengePassword can
5437 be handled by the string table functions.
5438
5439 Also modified the multi byte string table handling. Now there is
5440 a 'global mask' which masks out certain types. The table itself
5441 can use the flag STABLE_NO_MASK to ignore the mask setting: this
5442 is useful when for example there is only one permissible type
5443 (as in countryName) and using the mask might result in no valid
5444 types at all.
5445 [Steve Henson]
5446
5447 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
5448 SSL_get_peer_finished to allow applications to obtain the latest
5449 Finished messages sent to the peer or expected from the peer,
5450 respectively. (SSL_get_peer_finished is usually the Finished message
5451 actually received from the peer, otherwise the protocol will be aborted.)
5452
5453 As the Finished message are message digests of the complete handshake
5454 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
5455 be used for external authentication procedures when the authentication
5456 provided by SSL/TLS is not desired or is not enough.
5457 [Bodo Moeller]
5458
5459 *) Enhanced support for Alpha Linux is added. Now ./config checks if
5460 the host supports BWX extension and if Compaq C is present on the
5461 $PATH. Just exploiting of the BWX extension results in 20-30%
5462 performance kick for some algorithms, e.g. DES and RC4 to mention
5463 a couple. Compaq C in turn generates ~20% faster code for MD5 and
5464 SHA1.
5465 [Andy Polyakov]
5466
5467 *) Add support for MS "fast SGC". This is arguably a violation of the
5468 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
5469 weak crypto and after checking the certificate is SGC a second one
5470 with strong crypto. MS SGC stops the first handshake after receiving
5471 the server certificate message and sends a second client hello. Since
5472 a server will typically do all the time consuming operations before
5473 expecting any further messages from the client (server key exchange
5474 is the most expensive) there is little difference between the two.
5475
5476 To get OpenSSL to support MS SGC we have to permit a second client
5477 hello message after we have sent server done. In addition we have to
5478 reset the MAC if we do get this second client hello.
5479 [Steve Henson]
5480
5481 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
5482 if a DER encoded private key is RSA or DSA traditional format. Changed
5483 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
5484 format DER encoded private key. Newer code should use PKCS#8 format which
5485 has the key type encoded in the ASN1 structure. Added DER private key
5486 support to pkcs8 application.
5487 [Steve Henson]
5488
5489 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
5490 ciphersuites has been selected (as required by the SSL 3/TLS 1
5491 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
5492 is set, we interpret this as a request to violate the specification
5493 (the worst that can happen is a handshake failure, and 'correct'
5494 behaviour would result in a handshake failure anyway).
5495 [Bodo Moeller]
5496
5497 *) In SSL_CTX_add_session, take into account that there might be multiple
5498 SSL_SESSION structures with the same session ID (e.g. when two threads
5499 concurrently obtain them from an external cache).
5500 The internal cache can handle only one SSL_SESSION with a given ID,
5501 so if there's a conflict, we now throw out the old one to achieve
5502 consistency.
5503 [Bodo Moeller]
5504
5505 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
5506 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
5507 some routines that use cipher OIDs: some ciphers do not have OIDs
5508 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
5509 example.
5510 [Steve Henson]
5511
5512 *) Simplify the trust setting structure and code. Now we just have
5513 two sequences of OIDs for trusted and rejected settings. These will
5514 typically have values the same as the extended key usage extension
5515 and any application specific purposes.
5516
5517 The trust checking code now has a default behaviour: it will just
5518 check for an object with the same NID as the passed id. Functions can
5519 be provided to override either the default behaviour or the behaviour
5520 for a given id. SSL client, server and email already have functions
5521 in place for compatibility: they check the NID and also return "trusted"
5522 if the certificate is self signed.
5523 [Steve Henson]
5524
5525 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
5526 traditional format into an EVP_PKEY structure.
5527 [Steve Henson]
5528
5529 *) Add a password callback function PEM_cb() which either prompts for
5530 a password if usr_data is NULL or otherwise assumes it is a null
5531 terminated password. Allow passwords to be passed on command line
5532 environment or config files in a few more utilities.
5533 [Steve Henson]
5534
5535 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
5536 keys. Add some short names for PKCS#8 PBE algorithms and allow them
5537 to be specified on the command line for the pkcs8 and pkcs12 utilities.
5538 Update documentation.
5539 [Steve Henson]
5540
5541 *) Support for ASN1 "NULL" type. This could be handled before by using
5542 ASN1_TYPE but there wasn't any function that would try to read a NULL
5543 and produce an error if it couldn't. For compatibility we also have
5544 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
5545 don't allocate anything because they don't need to.
5546 [Steve Henson]
5547
5548 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
5549 for details.
5550 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
5551
5552 *) Rebuild of the memory allocation routines used by OpenSSL code and
5553 possibly others as well. The purpose is to make an interface that
5554 provide hooks so anyone can build a separate set of allocation and
5555 deallocation routines to be used by OpenSSL, for example memory
5556 pool implementations, or something else, which was previously hard
5557 since Malloc(), Realloc() and Free() were defined as macros having
5558 the values malloc, realloc and free, respectively (except for Win32
5559 compilations). The same is provided for memory debugging code.
5560 OpenSSL already comes with functionality to find memory leaks, but
5561 this gives people a chance to debug other memory problems.
5562
5563 With these changes, a new set of functions and macros have appeared:
5564
5565 CRYPTO_set_mem_debug_functions() [F]
5566 CRYPTO_get_mem_debug_functions() [F]
5567 CRYPTO_dbg_set_options() [F]
5568 CRYPTO_dbg_get_options() [F]
5569 CRYPTO_malloc_debug_init() [M]
5570
5571 The memory debug functions are NULL by default, unless the library
5572 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
5573 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
5574 gives the standard debugging functions that come with OpenSSL) or
5575 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
5576 provided by the library user) must be used. When the standard
5577 debugging functions are used, CRYPTO_dbg_set_options can be used to
5578 request additional information:
5579 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
5580 the CRYPTO_MDEBUG_xxx macro when compiling the library.
5581
5582 Also, things like CRYPTO_set_mem_functions will always give the
5583 expected result (the new set of functions is used for allocation
5584 and deallocation) at all times, regardless of platform and compiler
5585 options.
5586
5587 To finish it up, some functions that were never use in any other
5588 way than through macros have a new API and new semantic:
5589
5590 CRYPTO_dbg_malloc()
5591 CRYPTO_dbg_realloc()
5592 CRYPTO_dbg_free()
5593
5594 All macros of value have retained their old syntax.
5595 [Richard Levitte and Bodo Moeller]
5596
5597 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
5598 ordering of SMIMECapabilities wasn't in "strength order" and there
5599 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
5600 algorithm.
5601 [Steve Henson]
5602
5603 *) Some ASN1 types with illegal zero length encoding (INTEGER,
5604 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
5605 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
5606
5607 *) Merge in my S/MIME library for OpenSSL. This provides a simple
5608 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
5609 functionality to handle multipart/signed properly) and a utility
5610 called 'smime' to call all this stuff. This is based on code I
5611 originally wrote for Celo who have kindly allowed it to be
5612 included in OpenSSL.
5613 [Steve Henson]
5614
5615 *) Add variants des_set_key_checked and des_set_key_unchecked of
5616 des_set_key (aka des_key_sched). Global variable des_check_key
5617 decides which of these is called by des_set_key; this way
5618 des_check_key behaves as it always did, but applications and
5619 the library itself, which was buggy for des_check_key == 1,
5620 have a cleaner way to pick the version they need.
5621 [Bodo Moeller]
5622
5623 *) New function PKCS12_newpass() which changes the password of a
5624 PKCS12 structure.
5625 [Steve Henson]
5626
5627 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
5628 dynamic mix. In both cases the ids can be used as an index into the
5629 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
5630 functions so they accept a list of the field values and the
5631 application doesn't need to directly manipulate the X509_TRUST
5632 structure.
5633 [Steve Henson]
5634
5635 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
5636 need initialising.
5637 [Steve Henson]
5638
5639 *) Modify the way the V3 extension code looks up extensions. This now
5640 works in a similar way to the object code: we have some "standard"
5641 extensions in a static table which is searched with OBJ_bsearch()
5642 and the application can add dynamic ones if needed. The file
5643 crypto/x509v3/ext_dat.h now has the info: this file needs to be
5644 updated whenever a new extension is added to the core code and kept
5645 in ext_nid order. There is a simple program 'tabtest.c' which checks
5646 this. New extensions are not added too often so this file can readily
5647 be maintained manually.
5648
5649 There are two big advantages in doing things this way. The extensions
5650 can be looked up immediately and no longer need to be "added" using
5651 X509V3_add_standard_extensions(): this function now does nothing.
5652 [Side note: I get *lots* of email saying the extension code doesn't
5653 work because people forget to call this function]
5654 Also no dynamic allocation is done unless new extensions are added:
5655 so if we don't add custom extensions there is no need to call
5656 X509V3_EXT_cleanup().
5657 [Steve Henson]
5658
5659 *) Modify enc utility's salting as follows: make salting the default. Add a
5660 magic header, so unsalted files fail gracefully instead of just decrypting
5661 to garbage. This is because not salting is a big security hole, so people
5662 should be discouraged from doing it.
5663 [Ben Laurie]
5664
5665 *) Fixes and enhancements to the 'x509' utility. It allowed a message
5666 digest to be passed on the command line but it only used this
5667 parameter when signing a certificate. Modified so all relevant
5668 operations are affected by the digest parameter including the
5669 -fingerprint and -x509toreq options. Also -x509toreq choked if a
5670 DSA key was used because it didn't fix the digest.
5671 [Steve Henson]
5672
5673 *) Initial certificate chain verify code. Currently tests the untrusted
5674 certificates for consistency with the verify purpose (which is set
5675 when the X509_STORE_CTX structure is set up) and checks the pathlength.
5676
5677 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
5678 this is because it will reject chains with invalid extensions whereas
5679 every previous version of OpenSSL and SSLeay made no checks at all.
5680
5681 Trust code: checks the root CA for the relevant trust settings. Trust
5682 settings have an initial value consistent with the verify purpose: e.g.
5683 if the verify purpose is for SSL client use it expects the CA to be
5684 trusted for SSL client use. However the default value can be changed to
5685 permit custom trust settings: one example of this would be to only trust
5686 certificates from a specific "secure" set of CAs.
5687
5688 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
5689 which should be used for version portability: especially since the
5690 verify structure is likely to change more often now.
5691
5692 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
5693 to set them. If not set then assume SSL clients will verify SSL servers
5694 and vice versa.
5695
5696 Two new options to the verify program: -untrusted allows a set of
5697 untrusted certificates to be passed in and -purpose which sets the
5698 intended purpose of the certificate. If a purpose is set then the
5699 new chain verify code is used to check extension consistency.
5700 [Steve Henson]
5701
5702 *) Support for the authority information access extension.
5703 [Steve Henson]
5704
5705 *) Modify RSA and DSA PEM read routines to transparently handle
5706 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
5707 public keys in a format compatible with certificate
5708 SubjectPublicKeyInfo structures. Unfortunately there were already
5709 functions called *_PublicKey_* which used various odd formats so
5710 these are retained for compatibility: however the DSA variants were
5711 never in a public release so they have been deleted. Changed dsa/rsa
5712 utilities to handle the new format: note no releases ever handled public
5713 keys so we should be OK.
5714
5715 The primary motivation for this change is to avoid the same fiasco
5716 that dogs private keys: there are several incompatible private key
5717 formats some of which are standard and some OpenSSL specific and
5718 require various evil hacks to allow partial transparent handling and
5719 even then it doesn't work with DER formats. Given the option anything
5720 other than PKCS#8 should be dumped: but the other formats have to
5721 stay in the name of compatibility.
5722
5723 With public keys and the benefit of hindsight one standard format
5724 is used which works with EVP_PKEY, RSA or DSA structures: though
5725 it clearly returns an error if you try to read the wrong kind of key.
5726
5727 Added a -pubkey option to the 'x509' utility to output the public key.
5728 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
5729 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
5730 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
5731 that do the same as the EVP_PKEY_assign_*() except they up the
5732 reference count of the added key (they don't "swallow" the
5733 supplied key).
5734 [Steve Henson]
5735
5736 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
5737 CRLs would fail if the file contained no certificates or no CRLs:
5738 added a new function to read in both types and return the number
5739 read: this means that if none are read it will be an error. The
5740 DER versions of the certificate and CRL reader would always fail
5741 because it isn't possible to mix certificates and CRLs in DER format
5742 without choking one or the other routine. Changed this to just read
5743 a certificate: this is the best we can do. Also modified the code
5744 in apps/verify.c to take notice of return codes: it was previously
5745 attempting to read in certificates from NULL pointers and ignoring
5746 any errors: this is one reason why the cert and CRL reader seemed
5747 to work. It doesn't check return codes from the default certificate
5748 routines: these may well fail if the certificates aren't installed.
5749 [Steve Henson]
5750
5751 *) Code to support otherName option in GeneralName.
5752 [Steve Henson]
5753
5754 *) First update to verify code. Change the verify utility
5755 so it warns if it is passed a self signed certificate:
5756 for consistency with the normal behaviour. X509_verify
5757 has been modified to it will now verify a self signed
5758 certificate if *exactly* the same certificate appears
5759 in the store: it was previously impossible to trust a
5760 single self signed certificate. This means that:
5761 openssl verify ss.pem
5762 now gives a warning about a self signed certificate but
5763 openssl verify -CAfile ss.pem ss.pem
5764 is OK.
5765 [Steve Henson]
5766
5767 *) For servers, store verify_result in SSL_SESSION data structure
5768 (and add it to external session representation).
5769 This is needed when client certificate verifications fails,
5770 but an application-provided verification callback (set by
5771 SSL_CTX_set_cert_verify_callback) allows accepting the session
5772 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
5773 but returns 1): When the session is reused, we have to set
5774 ssl->verify_result to the appropriate error code to avoid
5775 security holes.
5776 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
5777
5778 *) Fix a bug in the new PKCS#7 code: it didn't consider the
5779 case in PKCS7_dataInit() where the signed PKCS7 structure
5780 didn't contain any existing data because it was being created.
5781 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
5782
5783 *) Add a salt to the key derivation routines in enc.c. This
5784 forms the first 8 bytes of the encrypted file. Also add a
5785 -S option to allow a salt to be input on the command line.
5786 [Steve Henson]
5787
5788 *) New function X509_cmp(). Oddly enough there wasn't a function
5789 to compare two certificates. We do this by working out the SHA1
5790 hash and comparing that. X509_cmp() will be needed by the trust
5791 code.
5792 [Steve Henson]
5793
5794 *) SSL_get1_session() is like SSL_get_session(), but increments
5795 the reference count in the SSL_SESSION returned.
5796 [Geoff Thorpe <geoff@eu.c2.net>]
5797
5798 *) Fix for 'req': it was adding a null to request attributes.
5799 Also change the X509_LOOKUP and X509_INFO code to handle
5800 certificate auxiliary information.
5801 [Steve Henson]
5802
5803 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
5804 the 'enc' command.
5805 [Steve Henson]
5806
5807 *) Add the possibility to add extra information to the memory leak
5808 detecting output, to form tracebacks, showing from where each
5809 allocation was originated: CRYPTO_push_info("constant string") adds
5810 the string plus current file name and line number to a per-thread
5811 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
5812 is like calling CYRPTO_pop_info() until the stack is empty.
5813 Also updated memory leak detection code to be multi-thread-safe.
5814 [Richard Levitte]
5815
5816 *) Add options -text and -noout to pkcs7 utility and delete the
5817 encryption options which never did anything. Update docs.
5818 [Steve Henson]
5819
5820 *) Add options to some of the utilities to allow the pass phrase
5821 to be included on either the command line (not recommended on
5822 OSes like Unix) or read from the environment. Update the
5823 manpages and fix a few bugs.
5824 [Steve Henson]
5825
5826 *) Add a few manpages for some of the openssl commands.
5827 [Steve Henson]
5828
5829 *) Fix the -revoke option in ca. It was freeing up memory twice,
5830 leaking and not finding already revoked certificates.
5831 [Steve Henson]
5832
5833 *) Extensive changes to support certificate auxiliary information.
5834 This involves the use of X509_CERT_AUX structure and X509_AUX
5835 functions. An X509_AUX function such as PEM_read_X509_AUX()
5836 can still read in a certificate file in the usual way but it
5837 will also read in any additional "auxiliary information". By
5838 doing things this way a fair degree of compatibility can be
5839 retained: existing certificates can have this information added
5840 using the new 'x509' options.
5841
5842 Current auxiliary information includes an "alias" and some trust
5843 settings. The trust settings will ultimately be used in enhanced
5844 certificate chain verification routines: currently a certificate
5845 can only be trusted if it is self signed and then it is trusted
5846 for all purposes.
5847 [Steve Henson]
5848
5849 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
5850 The problem was that one of the replacement routines had not been working
5851 since SSLeay releases. For now the offending routine has been replaced
5852 with non-optimised assembler. Even so, this now gives around 95%
5853 performance improvement for 1024 bit RSA signs.
5854 [Mark Cox]
5855
5856 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
5857 handling. Most clients have the effective key size in bits equal to
5858 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
5859 A few however don't do this and instead use the size of the decrypted key
5860 to determine the RC2 key length and the AlgorithmIdentifier to determine
5861 the effective key length. In this case the effective key length can still
5862 be 40 bits but the key length can be 168 bits for example. This is fixed
5863 by manually forcing an RC2 key into the EVP_PKEY structure because the
5864 EVP code can't currently handle unusual RC2 key sizes: it always assumes
5865 the key length and effective key length are equal.
5866 [Steve Henson]
5867
5868 *) Add a bunch of functions that should simplify the creation of
5869 X509_NAME structures. Now you should be able to do:
5870 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
5871 and have it automatically work out the correct field type and fill in
5872 the structures. The more adventurous can try:
5873 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
5874 and it will (hopefully) work out the correct multibyte encoding.
5875 [Steve Henson]
5876
5877 *) Change the 'req' utility to use the new field handling and multibyte
5878 copy routines. Before the DN field creation was handled in an ad hoc
5879 way in req, ca, and x509 which was rather broken and didn't support
5880 BMPStrings or UTF8Strings. Since some software doesn't implement
5881 BMPStrings or UTF8Strings yet, they can be enabled using the config file
5882 using the dirstring_type option. See the new comment in the default
5883 openssl.cnf for more info.
5884 [Steve Henson]
5885
5886 *) Make crypto/rand/md_rand.c more robust:
5887 - Assure unique random numbers after fork().
5888 - Make sure that concurrent threads access the global counter and
5889 md serializably so that we never lose entropy in them
5890 or use exactly the same state in multiple threads.
5891 Access to the large state is not always serializable because
5892 the additional locking could be a performance killer, and
5893 md should be large enough anyway.
5894 [Bodo Moeller]
5895
5896 *) New file apps/app_rand.c with commonly needed functionality
5897 for handling the random seed file.
5898
5899 Use the random seed file in some applications that previously did not:
5900 ca,
5901 dsaparam -genkey (which also ignored its '-rand' option),
5902 s_client,
5903 s_server,
5904 x509 (when signing).
5905 Except on systems with /dev/urandom, it is crucial to have a random
5906 seed file at least for key creation, DSA signing, and for DH exchanges;
5907 for RSA signatures we could do without one.
5908
5909 gendh and gendsa (unlike genrsa) used to read only the first byte
5910 of each file listed in the '-rand' option. The function as previously
5911 found in genrsa is now in app_rand.c and is used by all programs
5912 that support '-rand'.
5913 [Bodo Moeller]
5914
5915 *) In RAND_write_file, use mode 0600 for creating files;
5916 don't just chmod when it may be too late.
5917 [Bodo Moeller]
5918
5919 *) Report an error from X509_STORE_load_locations
5920 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
5921 [Bill Perry]
5922
5923 *) New function ASN1_mbstring_copy() this copies a string in either
5924 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
5925 into an ASN1_STRING type. A mask of permissible types is passed
5926 and it chooses the "minimal" type to use or an error if not type
5927 is suitable.
5928 [Steve Henson]
5929
5930 *) Add function equivalents to the various macros in asn1.h. The old
5931 macros are retained with an M_ prefix. Code inside the library can
5932 use the M_ macros. External code (including the openssl utility)
5933 should *NOT* in order to be "shared library friendly".
5934 [Steve Henson]
5935
5936 *) Add various functions that can check a certificate's extensions
5937 to see if it usable for various purposes such as SSL client,
5938 server or S/MIME and CAs of these types. This is currently
5939 VERY EXPERIMENTAL but will ultimately be used for certificate chain
5940 verification. Also added a -purpose flag to x509 utility to
5941 print out all the purposes.
5942 [Steve Henson]
5943
5944 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
5945 functions.
5946 [Steve Henson]
5947
5948 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
5949 for, obtain and decode and extension and obtain its critical flag.
5950 This allows all the necessary extension code to be handled in a
5951 single function call.
5952 [Steve Henson]
5953
5954 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
5955 platforms. See crypto/rc4/rc4_enc.c for further details.
5956 [Andy Polyakov]
5957
5958 *) New -noout option to asn1parse. This causes no output to be produced
5959 its main use is when combined with -strparse and -out to extract data
5960 from a file (which may not be in ASN.1 format).
5961 [Steve Henson]
5962
5963 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
5964 when producing the local key id.
5965 [Richard Levitte <levitte@stacken.kth.se>]
5966
5967 *) New option -dhparam in s_server. This allows a DH parameter file to be
5968 stated explicitly. If it is not stated then it tries the first server
5969 certificate file. The previous behaviour hard coded the filename
5970 "server.pem".
5971 [Steve Henson]
5972
5973 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
5974 a public key to be input or output. For example:
5975 openssl rsa -in key.pem -pubout -out pubkey.pem
5976 Also added necessary DSA public key functions to handle this.
5977 [Steve Henson]
5978
5979 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
5980 in the message. This was handled by allowing
5981 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
5982 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
5983
5984 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
5985 to the end of the strings whereas this didn't. This would cause problems
5986 if strings read with d2i_ASN1_bytes() were later modified.
5987 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
5988
5989 *) Fix for base64 decode bug. When a base64 bio reads only one line of
5990 data and it contains EOF it will end up returning an error. This is
5991 caused by input 46 bytes long. The cause is due to the way base64
5992 BIOs find the start of base64 encoded data. They do this by trying a
5993 trial decode on each line until they find one that works. When they
5994 do a flag is set and it starts again knowing it can pass all the
5995 data directly through the decoder. Unfortunately it doesn't reset
5996 the context it uses. This means that if EOF is reached an attempt
5997 is made to pass two EOFs through the context and this causes the
5998 resulting error. This can also cause other problems as well. As is
5999 usual with these problems it takes *ages* to find and the fix is
6000 trivial: move one line.
6001 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6002
6003 *) Ugly workaround to get s_client and s_server working under Windows. The
6004 old code wouldn't work because it needed to select() on sockets and the
6005 tty (for keypresses and to see if data could be written). Win32 only
6006 supports select() on sockets so we select() with a 1s timeout on the
6007 sockets and then see if any characters are waiting to be read, if none
6008 are present then we retry, we also assume we can always write data to
6009 the tty. This isn't nice because the code then blocks until we've
6010 received a complete line of data and it is effectively polling the
6011 keyboard at 1s intervals: however it's quite a bit better than not
6012 working at all :-) A dedicated Windows application might handle this
6013 with an event loop for example.
6014 [Steve Henson]
6015
6016 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6017 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6018 will be called when RSA_sign() and RSA_verify() are used. This is useful
6019 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6020 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6021 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6022 This necessitated the support of an extra signature type NID_md5_sha1
6023 for SSL signatures and modifications to the SSL library to use it instead
6024 of calling RSA_public_decrypt() and RSA_private_encrypt().
6025 [Steve Henson]
6026
6027 *) Add new -verify -CAfile and -CApath options to the crl program, these
6028 will lookup a CRL issuers certificate and verify the signature in a
6029 similar way to the verify program. Tidy up the crl program so it
6030 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6031 less strict. It will now permit CRL extensions even if it is not
6032 a V2 CRL: this will allow it to tolerate some broken CRLs.
6033 [Steve Henson]
6034
6035 *) Initialize all non-automatic variables each time one of the openssl
6036 sub-programs is started (this is necessary as they may be started
6037 multiple times from the "OpenSSL>" prompt).
6038 [Lennart Bang, Bodo Moeller]
6039
6040 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6041 removing all other RSA functionality (this is what NO_RSA does). This
6042 is so (for example) those in the US can disable those operations covered
6043 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6044 key generation.
6045 [Steve Henson]
6046
6047 *) Non-copying interface to BIO pairs.
6048 (still largely untested)
6049 [Bodo Moeller]
6050
6051 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6052 ASCII string. This was handled independently in various places before.
6053 [Steve Henson]
6054
6055 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6056 UTF8 strings a character at a time.
6057 [Steve Henson]
6058
6059 *) Use client_version from client hello to select the protocol
6060 (s23_srvr.c) and for RSA client key exchange verification
6061 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6062 [Bodo Moeller]
6063
6064 *) Add various utility functions to handle SPKACs, these were previously
6065 handled by poking round in the structure internals. Added new function
6066 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6067 print, verify and generate SPKACs. Based on an original idea from
6068 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6069 [Steve Henson]
6070
6071 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
6072 [Andy Polyakov]
6073
6074 *) Allow the config file extension section to be overwritten on the
6075 command line. Based on an original idea from Massimiliano Pala
6076 <madwolf@comune.modena.it>. The new option is called -extensions
6077 and can be applied to ca, req and x509. Also -reqexts to override
6078 the request extensions in req and -crlexts to override the crl extensions
6079 in ca.
6080 [Steve Henson]
6081
6082 *) Add new feature to the SPKAC handling in ca. Now you can include
6083 the same field multiple times by preceding it by "XXXX." for example:
6084 1.OU="Unit name 1"
6085 2.OU="Unit name 2"
6086 this is the same syntax as used in the req config file.
6087 [Steve Henson]
6088
6089 *) Allow certificate extensions to be added to certificate requests. These
6090 are specified in a 'req_extensions' option of the req section of the
6091 config file. They can be printed out with the -text option to req but
6092 are otherwise ignored at present.
6093 [Steve Henson]
6094
6095 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
6096 data read consists of only the final block it would not decrypted because
6097 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
6098 A misplaced 'break' also meant the decrypted final block might not be
6099 copied until the next read.
6100 [Steve Henson]
6101
6102 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
6103 a few extra parameters to the DH structure: these will be useful if
6104 for example we want the value of 'q' or implement X9.42 DH.
6105 [Steve Henson]
6106
6107 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
6108 provides hooks that allow the default DSA functions or functions on a
6109 "per key" basis to be replaced. This allows hardware acceleration and
6110 hardware key storage to be handled without major modification to the
6111 library. Also added low level modexp hooks and CRYPTO_EX structure and
6112 associated functions.
6113 [Steve Henson]
6114
6115 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
6116 as "read only": it can't be written to and the buffer it points to will
6117 not be freed. Reading from a read only BIO is much more efficient than
6118 a normal memory BIO. This was added because there are several times when
6119 an area of memory needs to be read from a BIO. The previous method was
6120 to create a memory BIO and write the data to it, this results in two
6121 copies of the data and an O(n^2) reading algorithm. There is a new
6122 function BIO_new_mem_buf() which creates a read only memory BIO from
6123 an area of memory. Also modified the PKCS#7 routines to use read only
6124 memory BIOs.
6125 [Steve Henson]
6126
6127 *) Bugfix: ssl23_get_client_hello did not work properly when called in
6128 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
6129 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
6130 but a retry condition occured while trying to read the rest.
6131 [Bodo Moeller]
6132
6133 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
6134 NID_pkcs7_encrypted by default: this was wrong since this should almost
6135 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
6136 the encrypted data type: this is a more sensible place to put it and it
6137 allows the PKCS#12 code to be tidied up that duplicated this
6138 functionality.
6139 [Steve Henson]
6140
6141 *) Changed obj_dat.pl script so it takes its input and output files on
6142 the command line. This should avoid shell escape redirection problems
6143 under Win32.
6144 [Steve Henson]
6145
6146 *) Initial support for certificate extension requests, these are included
6147 in things like Xenroll certificate requests. Included functions to allow
6148 extensions to be obtained and added.
6149 [Steve Henson]
6150
6151 *) -crlf option to s_client and s_server for sending newlines as
6152 CRLF (as required by many protocols).
6153 [Bodo Moeller]
6154
6155 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
6156
6157 *) Install libRSAglue.a when OpenSSL is built with RSAref.
6158 [Ralf S. Engelschall]
6159
6160 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
6161 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
6162
6163 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
6164 program.
6165 [Steve Henson]
6166
6167 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
6168 DH parameters/keys (q is lost during that conversion, but the resulting
6169 DH parameters contain its length).
6170
6171 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
6172 much faster than DH_generate_parameters (which creates parameters
6173 where p = 2*q + 1), and also the smaller q makes DH computations
6174 much more efficient (160-bit exponentiation instead of 1024-bit
6175 exponentiation); so this provides a convenient way to support DHE
6176 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
6177 utter importance to use
6178 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6179 or
6180 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
6181 when such DH parameters are used, because otherwise small subgroup
6182 attacks may become possible!
6183 [Bodo Moeller]
6184
6185 *) Avoid memory leak in i2d_DHparams.
6186 [Bodo Moeller]
6187
6188 *) Allow the -k option to be used more than once in the enc program:
6189 this allows the same encrypted message to be read by multiple recipients.
6190 [Steve Henson]
6191
6192 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
6193 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
6194 it will always use the numerical form of the OID, even if it has a short
6195 or long name.
6196 [Steve Henson]
6197
6198 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
6199 method only got called if p,q,dmp1,dmq1,iqmp components were present,
6200 otherwise bn_mod_exp was called. In the case of hardware keys for example
6201 no private key components need be present and it might store extra data
6202 in the RSA structure, which cannot be accessed from bn_mod_exp.
6203 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
6204 private key operations.
6205 [Steve Henson]
6206
6207 *) Added support for SPARC Linux.
6208 [Andy Polyakov]
6209
6210 *) pem_password_cb function type incompatibly changed from
6211 typedef int pem_password_cb(char *buf, int size, int rwflag);
6212 to
6213 ....(char *buf, int size, int rwflag, void *userdata);
6214 so that applications can pass data to their callbacks:
6215 The PEM[_ASN1]_{read,write}... functions and macros now take an
6216 additional void * argument, which is just handed through whenever
6217 the password callback is called.
6218 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
6219
6220 New function SSL_CTX_set_default_passwd_cb_userdata.
6221
6222 Compatibility note: As many C implementations push function arguments
6223 onto the stack in reverse order, the new library version is likely to
6224 interoperate with programs that have been compiled with the old
6225 pem_password_cb definition (PEM_whatever takes some data that
6226 happens to be on the stack as its last argument, and the callback
6227 just ignores this garbage); but there is no guarantee whatsoever that
6228 this will work.
6229
6230 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
6231 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
6232 problems not only on Windows, but also on some Unix platforms.
6233 To avoid problematic command lines, these definitions are now in an
6234 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
6235 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
6236 [Bodo Moeller]
6237
6238 *) MIPS III/IV assembler module is reimplemented.
6239 [Andy Polyakov]
6240
6241 *) More DES library cleanups: remove references to srand/rand and
6242 delete an unused file.
6243 [Ulf Möller]
6244
6245 *) Add support for the the free Netwide assembler (NASM) under Win32,
6246 since not many people have MASM (ml) and it can be hard to obtain.
6247 This is currently experimental but it seems to work OK and pass all
6248 the tests. Check out INSTALL.W32 for info.
6249 [Steve Henson]
6250
6251 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
6252 without temporary keys kept an extra copy of the server key,
6253 and connections with temporary keys did not free everything in case
6254 of an error.
6255 [Bodo Moeller]
6256
6257 *) New function RSA_check_key and new openssl rsa option -check
6258 for verifying the consistency of RSA keys.
6259 [Ulf Moeller, Bodo Moeller]
6260
6261 *) Various changes to make Win32 compile work:
6262 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
6263 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
6264 comparison" warnings.
6265 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
6266 [Steve Henson]
6267
6268 *) Add a debugging option to PKCS#5 v2 key generation function: when
6269 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
6270 derived keys are printed to stderr.
6271 [Steve Henson]
6272
6273 *) Copy the flags in ASN1_STRING_dup().
6274 [Roman E. Pavlov <pre@mo.msk.ru>]
6275
6276 *) The x509 application mishandled signing requests containing DSA
6277 keys when the signing key was also DSA and the parameters didn't match.
6278
6279 It was supposed to omit the parameters when they matched the signing key:
6280 the verifying software was then supposed to automatically use the CA's
6281 parameters if they were absent from the end user certificate.
6282
6283 Omitting parameters is no longer recommended. The test was also
6284 the wrong way round! This was probably due to unusual behaviour in
6285 EVP_cmp_parameters() which returns 1 if the parameters match.
6286 This meant that parameters were omitted when they *didn't* match and
6287 the certificate was useless. Certificates signed with 'ca' didn't have
6288 this bug.
6289 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
6290
6291 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
6292 The interface is as follows:
6293 Applications can use
6294 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
6295 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
6296 "off" is now the default.
6297 The library internally uses
6298 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
6299 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
6300 to disable memory-checking temporarily.
6301
6302 Some inconsistent states that previously were possible (and were
6303 even the default) are now avoided.
6304
6305 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
6306 with each memory chunk allocated; this is occasionally more helpful
6307 than just having a counter.
6308
6309 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
6310
6311 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
6312 extensions.
6313 [Bodo Moeller]
6314
6315 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
6316 which largely parallels "options", but is for changing API behaviour,
6317 whereas "options" are about protocol behaviour.
6318 Initial "mode" flags are:
6319
6320 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
6321 a single record has been written.
6322 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
6323 retries use the same buffer location.
6324 (But all of the contents must be
6325 copied!)
6326 [Bodo Moeller]
6327
6328 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
6329 worked.
6330
6331 *) Fix problems with no-hmac etc.
6332 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
6333
6334 *) New functions RSA_get_default_method(), RSA_set_method() and
6335 RSA_get_method(). These allows replacement of RSA_METHODs without having
6336 to mess around with the internals of an RSA structure.
6337 [Steve Henson]
6338
6339 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
6340 Also really enable memory leak checks in openssl.c and in some
6341 test programs.
6342 [Chad C. Mulligan, Bodo Moeller]
6343
6344 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
6345 up the length of negative integers. This has now been simplified to just
6346 store the length when it is first determined and use it later, rather
6347 than trying to keep track of where data is copied and updating it to
6348 point to the end.
6349 [Steve Henson, reported by Brien Wheeler
6350 <bwheeler@authentica-security.com>]
6351
6352 *) Add a new function PKCS7_signatureVerify. This allows the verification
6353 of a PKCS#7 signature but with the signing certificate passed to the
6354 function itself. This contrasts with PKCS7_dataVerify which assumes the
6355 certificate is present in the PKCS#7 structure. This isn't always the
6356 case: certificates can be omitted from a PKCS#7 structure and be
6357 distributed by "out of band" means (such as a certificate database).
6358 [Steve Henson]
6359
6360 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
6361 function prototypes in pem.h, also change util/mkdef.pl to add the
6362 necessary function names.
6363 [Steve Henson]
6364
6365 *) mk1mf.pl (used by Windows builds) did not properly read the
6366 options set by Configure in the top level Makefile, and Configure
6367 was not even able to write more than one option correctly.
6368 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
6369 [Bodo Moeller]
6370
6371 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
6372 file to be loaded from a BIO or FILE pointer. The BIO version will
6373 for example allow memory BIOs to contain config info.
6374 [Steve Henson]
6375
6376 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
6377 Whoever hopes to achieve shared-library compatibility across versions
6378 must use this, not the compile-time macro.
6379 (Exercise 0.9.4: Which is the minimum library version required by
6380 such programs?)
6381 Note: All this applies only to multi-threaded programs, others don't
6382 need locks.
6383 [Bodo Moeller]
6384
6385 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
6386 through a BIO pair triggered the default case, i.e.
6387 SSLerr(...,SSL_R_UNKNOWN_STATE).
6388 [Bodo Moeller]
6389
6390 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
6391 can use the SSL library even if none of the specific BIOs is
6392 appropriate.
6393 [Bodo Moeller]
6394
6395 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
6396 for the encoded length.
6397 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
6398
6399 *) Add initial documentation of the X509V3 functions.
6400 [Steve Henson]
6401
6402 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
6403 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
6404 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
6405 secure PKCS#8 private key format with a high iteration count.
6406 [Steve Henson]
6407
6408 *) Fix determination of Perl interpreter: A perl or perl5
6409 _directory_ in $PATH was also accepted as the interpreter.
6410 [Ralf S. Engelschall]
6411
6412 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
6413 wrong with it but it was very old and did things like calling
6414 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
6415 unusual formatting.
6416 [Steve Henson]
6417
6418 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
6419 to use the new extension code.
6420 [Steve Henson]
6421
6422 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
6423 with macros. This should make it easier to change their form, add extra
6424 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
6425 constant.
6426 [Steve Henson]
6427
6428 *) Add to configuration table a new entry that can specify an alternative
6429 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
6430 according to Mark Crispin <MRC@Panda.COM>.
6431 [Bodo Moeller]
6432
6433 #if 0
6434 *) DES CBC did not update the IV. Weird.
6435 [Ben Laurie]
6436 #else
6437 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
6438 Changing the behaviour of the former might break existing programs --
6439 where IV updating is needed, des_ncbc_encrypt can be used.
6440 #endif
6441
6442 *) When bntest is run from "make test" it drives bc to check its
6443 calculations, as well as internally checking them. If an internal check
6444 fails, it needs to cause bc to give a non-zero result or make test carries
6445 on without noticing the failure. Fixed.
6446 [Ben Laurie]
6447
6448 *) DES library cleanups.
6449 [Ulf Möller]
6450
6451 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
6452 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
6453 ciphers. NOTE: although the key derivation function has been verified
6454 against some published test vectors it has not been extensively tested
6455 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
6456 of v2.0.
6457 [Steve Henson]
6458
6459 *) Instead of "mkdir -p", which is not fully portable, use new
6460 Perl script "util/mkdir-p.pl".
6461 [Bodo Moeller]
6462
6463 *) Rewrite the way password based encryption (PBE) is handled. It used to
6464 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
6465 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
6466 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
6467 the 'parameter' field of the AlgorithmIdentifier is passed to the
6468 underlying key generation function so it must do its own ASN1 parsing.
6469 This has also changed the EVP_PBE_CipherInit() function which now has a
6470 'parameter' argument instead of literal salt and iteration count values
6471 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
6472 [Steve Henson]
6473
6474 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
6475 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
6476 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
6477 KEY" because this clashed with PKCS#8 unencrypted string. Since this
6478 value was just used as a "magic string" and not used directly its
6479 value doesn't matter.
6480 [Steve Henson]
6481
6482 *) Introduce some semblance of const correctness to BN. Shame C doesn't
6483 support mutable.
6484 [Ben Laurie]
6485
6486 *) "linux-sparc64" configuration (ultrapenguin).
6487 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
6488 "linux-sparc" configuration.
6489 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
6490
6491 *) config now generates no-xxx options for missing ciphers.
6492 [Ulf Möller]
6493
6494 *) Support the EBCDIC character set (work in progress).
6495 File ebcdic.c not yet included because it has a different license.
6496 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6497
6498 *) Support BS2000/OSD-POSIX.
6499 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6500
6501 *) Make callbacks for key generation use void * instead of char *.
6502 [Ben Laurie]
6503
6504 *) Make S/MIME samples compile (not yet tested).
6505 [Ben Laurie]
6506
6507 *) Additional typesafe stacks.
6508 [Ben Laurie]
6509
6510 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
6511 [Bodo Moeller]
6512
6513
6514 Changes between 0.9.3 and 0.9.3a [29 May 1999]
6515
6516 *) New configuration variant "sco5-gcc".
6517
6518 *) Updated some demos.
6519 [Sean O Riordain, Wade Scholine]
6520
6521 *) Add missing BIO_free at exit of pkcs12 application.
6522 [Wu Zhigang]
6523
6524 *) Fix memory leak in conf.c.
6525 [Steve Henson]
6526
6527 *) Updates for Win32 to assembler version of MD5.
6528 [Steve Henson]
6529
6530 *) Set #! path to perl in apps/der_chop to where we found it
6531 instead of using a fixed path.
6532 [Bodo Moeller]
6533
6534 *) SHA library changes for irix64-mips4-cc.
6535 [Andy Polyakov]
6536
6537 *) Improvements for VMS support.
6538 [Richard Levitte]
6539
6540
6541 Changes between 0.9.2b and 0.9.3 [24 May 1999]
6542
6543 *) Bignum library bug fix. IRIX 6 passes "make test" now!
6544 This also avoids the problems with SC4.2 and unpatched SC5.
6545 [Andy Polyakov <appro@fy.chalmers.se>]
6546
6547 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
6548 These are required because of the typesafe stack would otherwise break
6549 existing code. If old code used a structure member which used to be STACK
6550 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
6551 sk_num or sk_value it would produce an error because the num, data members
6552 are not present in STACK_OF. Now it just produces a warning. sk_set
6553 replaces the old method of assigning a value to sk_value
6554 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
6555 that does this will no longer work (and should use sk_set instead) but
6556 this could be regarded as a "questionable" behaviour anyway.
6557 [Steve Henson]
6558
6559 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
6560 correctly handle encrypted S/MIME data.
6561 [Steve Henson]
6562
6563 *) Change type of various DES function arguments from des_cblock
6564 (which means, in function argument declarations, pointer to char)
6565 to des_cblock * (meaning pointer to array with 8 char elements),
6566 which allows the compiler to do more typechecking; it was like
6567 that back in SSLeay, but with lots of ugly casts.
6568
6569 Introduce new type const_des_cblock.
6570 [Bodo Moeller]
6571
6572 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
6573 problems: find RecipientInfo structure that matches recipient certificate
6574 and initialise the ASN1 structures properly based on passed cipher.
6575 [Steve Henson]
6576
6577 *) Belatedly make the BN tests actually check the results.
6578 [Ben Laurie]
6579
6580 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
6581 to and from BNs: it was completely broken. New compilation option
6582 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
6583 key elements as negative integers.
6584 [Steve Henson]
6585
6586 *) Reorganize and speed up MD5.
6587 [Andy Polyakov <appro@fy.chalmers.se>]
6588
6589 *) VMS support.
6590 [Richard Levitte <richard@levitte.org>]
6591
6592 *) New option -out to asn1parse to allow the parsed structure to be
6593 output to a file. This is most useful when combined with the -strparse
6594 option to examine the output of things like OCTET STRINGS.
6595 [Steve Henson]
6596
6597 *) Make SSL library a little more fool-proof by not requiring any longer
6598 that SSL_set_{accept,connect}_state be called before
6599 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
6600 in many applications because usually everything *appeared* to work as
6601 intended anyway -- now it really works as intended).
6602 [Bodo Moeller]
6603
6604 *) Move openssl.cnf out of lib/.
6605 [Ulf Möller]
6606
6607 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
6608 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
6609 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
6610 [Ralf S. Engelschall]
6611
6612 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
6613 handle PKCS#7 enveloped data properly.
6614 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
6615
6616 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
6617 copying pointers. The cert_st handling is changed by this in
6618 various ways (and thus what used to be known as ctx->default_cert
6619 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
6620 any longer when s->cert does not give us what we need).
6621 ssl_cert_instantiate becomes obsolete by this change.
6622 As soon as we've got the new code right (possibly it already is?),
6623 we have solved a couple of bugs of the earlier code where s->cert
6624 was used as if it could not have been shared with other SSL structures.
6625
6626 Note that using the SSL API in certain dirty ways now will result
6627 in different behaviour than observed with earlier library versions:
6628 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
6629 does not influence s as it used to.
6630
6631 In order to clean up things more thoroughly, inside SSL_SESSION
6632 we don't use CERT any longer, but a new structure SESS_CERT
6633 that holds per-session data (if available); currently, this is
6634 the peer's certificate chain and, for clients, the server's certificate
6635 and temporary key. CERT holds only those values that can have
6636 meaningful defaults in an SSL_CTX.
6637 [Bodo Moeller]
6638
6639 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
6640 from the internal representation. Various PKCS#7 fixes: remove some
6641 evil casts and set the enc_dig_alg field properly based on the signing
6642 key type.
6643 [Steve Henson]
6644
6645 *) Allow PKCS#12 password to be set from the command line or the
6646 environment. Let 'ca' get its config file name from the environment
6647 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
6648 and 'x509').
6649 [Steve Henson]
6650
6651 *) Allow certificate policies extension to use an IA5STRING for the
6652 organization field. This is contrary to the PKIX definition but
6653 VeriSign uses it and IE5 only recognises this form. Document 'x509'
6654 extension option.
6655 [Steve Henson]
6656
6657 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
6658 without disallowing inline assembler and the like for non-pedantic builds.
6659 [Ben Laurie]
6660
6661 *) Support Borland C++ builder.
6662 [Janez Jere <jj@void.si>, modified by Ulf Möller]
6663
6664 *) Support Mingw32.
6665 [Ulf Möller]
6666
6667 *) SHA-1 cleanups and performance enhancements.
6668 [Andy Polyakov <appro@fy.chalmers.se>]
6669
6670 *) Sparc v8plus assembler for the bignum library.
6671 [Andy Polyakov <appro@fy.chalmers.se>]
6672
6673 *) Accept any -xxx and +xxx compiler options in Configure.
6674 [Ulf Möller]
6675
6676 *) Update HPUX configuration.
6677 [Anonymous]
6678
6679 *) Add missing sk_<type>_unshift() function to safestack.h
6680 [Ralf S. Engelschall]
6681
6682 *) New function SSL_CTX_use_certificate_chain_file that sets the
6683 "extra_cert"s in addition to the certificate. (This makes sense
6684 only for "PEM" format files, as chains as a whole are not
6685 DER-encoded.)
6686 [Bodo Moeller]
6687
6688 *) Support verify_depth from the SSL API.
6689 x509_vfy.c had what can be considered an off-by-one-error:
6690 Its depth (which was not part of the external interface)
6691 was actually counting the number of certificates in a chain;
6692 now it really counts the depth.
6693 [Bodo Moeller]
6694
6695 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
6696 instead of X509err, which often resulted in confusing error
6697 messages since the error codes are not globally unique
6698 (e.g. an alleged error in ssl3_accept when a certificate
6699 didn't match the private key).
6700
6701 *) New function SSL_CTX_set_session_id_context that allows to set a default
6702 value (so that you don't need SSL_set_session_id_context for each
6703 connection using the SSL_CTX).
6704 [Bodo Moeller]
6705
6706 *) OAEP decoding bug fix.
6707 [Ulf Möller]
6708
6709 *) Support INSTALL_PREFIX for package builders, as proposed by
6710 David Harris.
6711 [Bodo Moeller]
6712
6713 *) New Configure options "threads" and "no-threads". For systems
6714 where the proper compiler options are known (currently Solaris
6715 and Linux), "threads" is the default.
6716 [Bodo Moeller]
6717
6718 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
6719 [Bodo Moeller]
6720
6721 *) Install various scripts to $(OPENSSLDIR)/misc, not to
6722 $(INSTALLTOP)/bin -- they shouldn't clutter directories
6723 such as /usr/local/bin.
6724 [Bodo Moeller]
6725
6726 *) "make linux-shared" to build shared libraries.
6727 [Niels Poppe <niels@netbox.org>]
6728
6729 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
6730 [Ulf Möller]
6731
6732 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
6733 extension adding in x509 utility.
6734 [Steve Henson]
6735
6736 *) Remove NOPROTO sections and error code comments.
6737 [Ulf Möller]
6738
6739 *) Partial rewrite of the DEF file generator to now parse the ANSI
6740 prototypes.
6741 [Steve Henson]
6742
6743 *) New Configure options --prefix=DIR and --openssldir=DIR.
6744 [Ulf Möller]
6745
6746 *) Complete rewrite of the error code script(s). It is all now handled
6747 by one script at the top level which handles error code gathering,
6748 header rewriting and C source file generation. It should be much better
6749 than the old method: it now uses a modified version of Ulf's parser to
6750 read the ANSI prototypes in all header files (thus the old K&R definitions
6751 aren't needed for error creation any more) and do a better job of
6752 translating function codes into names. The old 'ASN1 error code imbedded
6753 in a comment' is no longer necessary and it doesn't use .err files which
6754 have now been deleted. Also the error code call doesn't have to appear all
6755 on one line (which resulted in some large lines...).
6756 [Steve Henson]
6757
6758 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
6759 [Bodo Moeller]
6760
6761 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
6762 0 (which usually indicates a closed connection), but continue reading.
6763 [Bodo Moeller]
6764
6765 *) Fix some race conditions.
6766 [Bodo Moeller]
6767
6768 *) Add support for CRL distribution points extension. Add Certificate
6769 Policies and CRL distribution points documentation.
6770 [Steve Henson]
6771
6772 *) Move the autogenerated header file parts to crypto/opensslconf.h.
6773 [Ulf Möller]
6774
6775 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
6776 8 of keying material. Merlin has also confirmed interop with this fix
6777 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
6778 [Merlin Hughes <merlin@baltimore.ie>]
6779
6780 *) Fix lots of warnings.
6781 [Richard Levitte <levitte@stacken.kth.se>]
6782
6783 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
6784 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
6785 [Richard Levitte <levitte@stacken.kth.se>]
6786
6787 *) Fix problems with sizeof(long) == 8.
6788 [Andy Polyakov <appro@fy.chalmers.se>]
6789
6790 *) Change functions to ANSI C.
6791 [Ulf Möller]
6792
6793 *) Fix typos in error codes.
6794 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
6795
6796 *) Remove defunct assembler files from Configure.
6797 [Ulf Möller]
6798
6799 *) SPARC v8 assembler BIGNUM implementation.
6800 [Andy Polyakov <appro@fy.chalmers.se>]
6801
6802 *) Support for Certificate Policies extension: both print and set.
6803 Various additions to support the r2i method this uses.
6804 [Steve Henson]
6805
6806 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
6807 return a const string when you are expecting an allocated buffer.
6808 [Ben Laurie]
6809
6810 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
6811 types DirectoryString and DisplayText.
6812 [Steve Henson]
6813
6814 *) Add code to allow r2i extensions to access the configuration database,
6815 add an LHASH database driver and add several ctx helper functions.
6816 [Steve Henson]
6817
6818 *) Fix an evil bug in bn_expand2() which caused various BN functions to
6819 fail when they extended the size of a BIGNUM.
6820 [Steve Henson]
6821
6822 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
6823 support typesafe stack.
6824 [Steve Henson]
6825
6826 *) Fix typo in SSL_[gs]et_options().
6827 [Nils Frostberg <nils@medcom.se>]
6828
6829 *) Delete various functions and files that belonged to the (now obsolete)
6830 old X509V3 handling code.
6831 [Steve Henson]
6832
6833 *) New Configure option "rsaref".
6834 [Ulf Möller]
6835
6836 *) Don't auto-generate pem.h.
6837 [Bodo Moeller]
6838
6839 *) Introduce type-safe ASN.1 SETs.
6840 [Ben Laurie]
6841
6842 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
6843 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
6844
6845 *) Introduce type-safe STACKs. This will almost certainly break lots of code
6846 that links with OpenSSL (well at least cause lots of warnings), but fear
6847 not: the conversion is trivial, and it eliminates loads of evil casts. A
6848 few STACKed things have been converted already. Feel free to convert more.
6849 In the fullness of time, I'll do away with the STACK type altogether.
6850 [Ben Laurie]
6851
6852 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
6853 specified in <certfile> by updating the entry in the index.txt file.
6854 This way one no longer has to edit the index.txt file manually for
6855 revoking a certificate. The -revoke option does the gory details now.
6856 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
6857
6858 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
6859 `-text' option at all and this way the `-noout -text' combination was
6860 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
6861 [Ralf S. Engelschall]
6862
6863 *) Make sure a corresponding plain text error message exists for the
6864 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
6865 verify callback function determined that a certificate was revoked.
6866 [Ralf S. Engelschall]
6867
6868 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
6869 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
6870 all available cipers including rc5, which was forgotten until now.
6871 In order to let the testing shell script know which algorithms
6872 are available, a new (up to now undocumented) command
6873 "openssl list-cipher-commands" is used.
6874 [Bodo Moeller]
6875
6876 *) Bugfix: s_client occasionally would sleep in select() when
6877 it should have checked SSL_pending() first.
6878 [Bodo Moeller]
6879
6880 *) New functions DSA_do_sign and DSA_do_verify to provide access to
6881 the raw DSA values prior to ASN.1 encoding.
6882 [Ulf Möller]
6883
6884 *) Tweaks to Configure
6885 [Niels Poppe <niels@netbox.org>]
6886
6887 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
6888 yet...
6889 [Steve Henson]
6890
6891 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
6892 [Ulf Möller]
6893
6894 *) New config option to avoid instructions that are illegal on the 80386.
6895 The default code is faster, but requires at least a 486.
6896 [Ulf Möller]
6897
6898 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
6899 SSL2_SERVER_VERSION (not used at all) macros, which are now the
6900 same as SSL2_VERSION anyway.
6901 [Bodo Moeller]
6902
6903 *) New "-showcerts" option for s_client.
6904 [Bodo Moeller]
6905
6906 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
6907 application. Various cleanups and fixes.
6908 [Steve Henson]
6909
6910 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
6911 modify error routines to work internally. Add error codes and PBE init
6912 to library startup routines.
6913 [Steve Henson]
6914
6915 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
6916 packing functions to asn1 and evp. Changed function names and error
6917 codes along the way.
6918 [Steve Henson]
6919
6920 *) PKCS12 integration: and so it begins... First of several patches to
6921 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
6922 objects to objects.h
6923 [Steve Henson]
6924
6925 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
6926 and display support for Thawte strong extranet extension.
6927 [Steve Henson]
6928
6929 *) Add LinuxPPC support.
6930 [Jeff Dubrule <igor@pobox.org>]
6931
6932 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
6933 bn_div_words in alpha.s.
6934 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
6935
6936 *) Make sure the RSA OAEP test is skipped under -DRSAref because
6937 OAEP isn't supported when OpenSSL is built with RSAref.
6938 [Ulf Moeller <ulf@fitug.de>]
6939
6940 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
6941 so they no longer are missing under -DNOPROTO.
6942 [Soren S. Jorvang <soren@t.dk>]
6943
6944
6945 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
6946
6947 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
6948 doesn't work when the session is reused. Coming soon!
6949 [Ben Laurie]
6950
6951 *) Fix a security hole, that allows sessions to be reused in the wrong
6952 context thus bypassing client cert protection! All software that uses
6953 client certs and session caches in multiple contexts NEEDS PATCHING to
6954 allow session reuse! A fuller solution is in the works.
6955 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
6956
6957 *) Some more source tree cleanups (removed obsolete files
6958 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
6959 permission on "config" script to be executable) and a fix for the INSTALL
6960 document.
6961 [Ulf Moeller <ulf@fitug.de>]
6962
6963 *) Remove some legacy and erroneous uses of malloc, free instead of
6964 Malloc, Free.
6965 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
6966
6967 *) Make rsa_oaep_test return non-zero on error.
6968 [Ulf Moeller <ulf@fitug.de>]
6969
6970 *) Add support for native Solaris shared libraries. Configure
6971 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
6972 if someone would make that last step automatic.
6973 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
6974
6975 *) ctx_size was not built with the right compiler during "make links". Fixed.
6976 [Ben Laurie]
6977
6978 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
6979 except NULL ciphers". This means the default cipher list will no longer
6980 enable NULL ciphers. They need to be specifically enabled e.g. with
6981 the string "DEFAULT:eNULL".
6982 [Steve Henson]
6983
6984 *) Fix to RSA private encryption routines: if p < q then it would
6985 occasionally produce an invalid result. This will only happen with
6986 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
6987 [Steve Henson]
6988
6989 *) Be less restrictive and allow also `perl util/perlpath.pl
6990 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
6991 because this way one can also use an interpreter named `perl5' (which is
6992 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
6993 installed as `perl').
6994 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6995
6996 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
6997 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
6998
6999 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7000 advapi32.lib to Win32 build and change the pem test comparision
7001 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7002 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7003 and crypto/des/ede_cbcm_enc.c.
7004 [Steve Henson]
7005
7006 *) DES quad checksum was broken on big-endian architectures. Fixed.
7007 [Ben Laurie]
7008
7009 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7010 Win32 test batch file so it (might) work again. The Win32 test batch file
7011 is horrible: I feel ill....
7012 [Steve Henson]
7013
7014 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7015 in e_os.h. Audit of header files to check ANSI and non ANSI
7016 sections: 10 functions were absent from non ANSI section and not exported
7017 from Windows DLLs. Fixed up libeay.num for new functions.
7018 [Steve Henson]
7019
7020 *) Make `openssl version' output lines consistent.
7021 [Ralf S. Engelschall]
7022
7023 *) Fix Win32 symbol export lists for BIO functions: Added
7024 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7025 to ms/libeay{16,32}.def.
7026 [Ralf S. Engelschall]
7027
7028 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7029 fine under Unix and passes some trivial tests I've now added. But the
7030 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7031 added to make sure no one expects that this stuff really works in the
7032 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7033 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7034 openssl_bio.xs.
7035 [Ralf S. Engelschall]
7036
7037 *) Fix the generation of two part addresses in perl.
7038 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7039
7040 *) Add config entry for Linux on MIPS.
7041 [John Tobey <jtobey@channel1.com>]
7042
7043 *) Make links whenever Configure is run, unless we are on Windoze.
7044 [Ben Laurie]
7045
7046 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7047 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7048 in CRLs.
7049 [Steve Henson]
7050
7051 *) Add a useful kludge to allow package maintainers to specify compiler and
7052 other platforms details on the command line without having to patch the
7053 Configure script everytime: One now can use ``perl Configure
7054 <id>:<details>'', i.e. platform ids are allowed to have details appended
7055 to them (seperated by colons). This is treated as there would be a static
7056 pre-configured entry in Configure's %table under key <id> with value
7057 <details> and ``perl Configure <id>'' is called. So, when you want to
7058 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7059 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7060 now, which overrides the FreeBSD-elf entry on-the-fly.
7061 [Ralf S. Engelschall]
7062
7063 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7064 [Ben Laurie]
7065
7066 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7067 on the `perl Configure ...' command line. This way one can compile
7068 OpenSSL libraries with Position Independent Code (PIC) which is needed
7069 for linking it into DSOs.
7070 [Ralf S. Engelschall]
7071
7072 *) Remarkably, export ciphers were totally broken and no-one had noticed!
7073 Fixed.
7074 [Ben Laurie]
7075
7076 *) Cleaned up the LICENSE document: The official contact for any license
7077 questions now is the OpenSSL core team under openssl-core@openssl.org.
7078 And add a paragraph about the dual-license situation to make sure people
7079 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
7080 to the OpenSSL toolkit.
7081 [Ralf S. Engelschall]
7082
7083 *) General source tree makefile cleanups: Made `making xxx in yyy...'
7084 display consistent in the source tree and replaced `/bin/rm' by `rm'.
7085 Additonally cleaned up the `make links' target: Remove unnecessary
7086 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
7087 to speed processing and no longer clutter the display with confusing
7088 stuff. Instead only the actually done links are displayed.
7089 [Ralf S. Engelschall]
7090
7091 *) Permit null encryption ciphersuites, used for authentication only. It used
7092 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
7093 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
7094 encryption.
7095 [Ben Laurie]
7096
7097 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
7098 signed attributes when verifying signatures (this would break them),
7099 the detached data encoding was wrong and public keys obtained using
7100 X509_get_pubkey() weren't freed.
7101 [Steve Henson]
7102
7103 *) Add text documentation for the BUFFER functions. Also added a work around
7104 to a Win95 console bug. This was triggered by the password read stuff: the
7105 last character typed gets carried over to the next fread(). If you were
7106 generating a new cert request using 'req' for example then the last
7107 character of the passphrase would be CR which would then enter the first
7108 field as blank.
7109 [Steve Henson]
7110
7111 *) Added the new `Includes OpenSSL Cryptography Software' button as
7112 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
7113 button and can be used by applications based on OpenSSL to show the
7114 relationship to the OpenSSL project.
7115 [Ralf S. Engelschall]
7116
7117 *) Remove confusing variables in function signatures in files
7118 ssl/ssl_lib.c and ssl/ssl.h.
7119 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7120
7121 *) Don't install bss_file.c under PREFIX/include/
7122 [Lennart Bong <lob@kulthea.stacken.kth.se>]
7123
7124 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
7125 functions that return function pointers and has support for NT specific
7126 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
7127 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
7128 unsigned to signed types: this was killing the Win32 compile.
7129 [Steve Henson]
7130
7131 *) Add new certificate file to stack functions,
7132 SSL_add_dir_cert_subjects_to_stack() and
7133 SSL_add_file_cert_subjects_to_stack(). These largely supplant
7134 SSL_load_client_CA_file(), and can be used to add multiple certs easily
7135 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
7136 This means that Apache-SSL and similar packages don't have to mess around
7137 to add as many CAs as they want to the preferred list.
7138 [Ben Laurie]
7139
7140 *) Experiment with doxygen documentation. Currently only partially applied to
7141 ssl/ssl_lib.c.
7142 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
7143 openssl.doxy as the configuration file.
7144 [Ben Laurie]
7145
7146 *) Get rid of remaining C++-style comments which strict C compilers hate.
7147 [Ralf S. Engelschall, pointed out by Carlos Amengual]
7148
7149 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
7150 compiled in by default: it has problems with large keys.
7151 [Steve Henson]
7152
7153 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
7154 DH private keys and/or callback functions which directly correspond to
7155 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
7156 is needed for applications which have to configure certificates on a
7157 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
7158 (e.g. s_server).
7159 For the RSA certificate situation is makes no difference, but
7160 for the DSA certificate situation this fixes the "no shared cipher"
7161 problem where the OpenSSL cipher selection procedure failed because the
7162 temporary keys were not overtaken from the context and the API provided
7163 no way to reconfigure them.
7164 The new functions now let applications reconfigure the stuff and they
7165 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
7166 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
7167 non-public-API function ssl_cert_instantiate() is used as a helper
7168 function and also to reduce code redundancy inside ssl_rsa.c.
7169 [Ralf S. Engelschall]
7170
7171 *) Move s_server -dcert and -dkey options out of the undocumented feature
7172 area because they are useful for the DSA situation and should be
7173 recognized by the users.
7174 [Ralf S. Engelschall]
7175
7176 *) Fix the cipher decision scheme for export ciphers: the export bits are
7177 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
7178 SSL_EXP_MASK. So, the original variable has to be used instead of the
7179 already masked variable.
7180 [Richard Levitte <levitte@stacken.kth.se>]
7181
7182 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
7183 [Richard Levitte <levitte@stacken.kth.se>]
7184
7185 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
7186 from `int' to `unsigned int' because it's a length and initialized by
7187 EVP_DigestFinal() which expects an `unsigned int *'.
7188 [Richard Levitte <levitte@stacken.kth.se>]
7189
7190 *) Don't hard-code path to Perl interpreter on shebang line of Configure
7191 script. Instead use the usual Shell->Perl transition trick.
7192 [Ralf S. Engelschall]
7193
7194 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
7195 (in addition to RSA certificates) to match the behaviour of `openssl dsa
7196 -noout -modulus' as it's already the case for `openssl rsa -noout
7197 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
7198 currently the public key is printed (a decision which was already done by
7199 `openssl dsa -modulus' in the past) which serves a similar purpose.
7200 Additionally the NO_RSA no longer completely removes the whole -modulus
7201 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
7202 now, too.
7203 [Ralf S. Engelschall]
7204
7205 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
7206 BIO. See the source (crypto/evp/bio_ok.c) for more info.
7207 [Arne Ansper <arne@ats.cyber.ee>]
7208
7209 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
7210 to be added. Now both 'req' and 'ca' can use new objects defined in the
7211 config file.
7212 [Steve Henson]
7213
7214 *) Add cool BIO that does syslog (or event log on NT).
7215 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
7216
7217 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
7218 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
7219 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
7220 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
7221 [Ben Laurie]
7222
7223 *) Add preliminary config info for new extension code.
7224 [Steve Henson]
7225
7226 *) Make RSA_NO_PADDING really use no padding.
7227 [Ulf Moeller <ulf@fitug.de>]
7228
7229 *) Generate errors when private/public key check is done.
7230 [Ben Laurie]
7231
7232 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
7233 for some CRL extensions and new objects added.
7234 [Steve Henson]
7235
7236 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
7237 key usage extension and fuller support for authority key id.
7238 [Steve Henson]
7239
7240 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
7241 padding method for RSA, which is recommended for new applications in PKCS
7242 #1 v2.0 (RFC 2437, October 1998).
7243 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
7244 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
7245 against Bleichbacher's attack on RSA.
7246 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
7247 Ben Laurie]
7248
7249 *) Updates to the new SSL compression code
7250 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7251
7252 *) Fix so that the version number in the master secret, when passed
7253 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
7254 (because the server will not accept higher), that the version number
7255 is 0x03,0x01, not 0x03,0x00
7256 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7257
7258 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
7259 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
7260 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
7261 [Steve Henson]
7262
7263 *) Support for RAW extensions where an arbitrary extension can be
7264 created by including its DER encoding. See apps/openssl.cnf for
7265 an example.
7266 [Steve Henson]
7267
7268 *) Make sure latest Perl versions don't interpret some generated C array
7269 code as Perl array code in the crypto/err/err_genc.pl script.
7270 [Lars Weber <3weber@informatik.uni-hamburg.de>]
7271
7272 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
7273 not many people have the assembler. Various Win32 compilation fixes and
7274 update to the INSTALL.W32 file with (hopefully) more accurate Win32
7275 build instructions.
7276 [Steve Henson]
7277
7278 *) Modify configure script 'Configure' to automatically create crypto/date.h
7279 file under Win32 and also build pem.h from pem.org. New script
7280 util/mkfiles.pl to create the MINFO file on environments that can't do a
7281 'make files': perl util/mkfiles.pl >MINFO should work.
7282 [Steve Henson]
7283
7284 *) Major rework of DES function declarations, in the pursuit of correctness
7285 and purity. As a result, many evil casts evaporated, and some weirdness,
7286 too. You may find this causes warnings in your code. Zapping your evil
7287 casts will probably fix them. Mostly.
7288 [Ben Laurie]
7289
7290 *) Fix for a typo in asn1.h. Bug fix to object creation script
7291 obj_dat.pl. It considered a zero in an object definition to mean
7292 "end of object": none of the objects in objects.h have any zeros
7293 so it wasn't spotted.
7294 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
7295
7296 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
7297 Masking (CBCM). In the absence of test vectors, the best I have been able
7298 to do is check that the decrypt undoes the encrypt, so far. Send me test
7299 vectors if you have them.
7300 [Ben Laurie]
7301
7302 *) Correct calculation of key length for export ciphers (too much space was
7303 allocated for null ciphers). This has not been tested!
7304 [Ben Laurie]
7305
7306 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
7307 message is now correct (it understands "crypto" and "ssl" on its
7308 command line). There is also now an "update" option. This will update
7309 the util/ssleay.num and util/libeay.num files with any new functions.
7310 If you do a:
7311 perl util/mkdef.pl crypto ssl update
7312 it will update them.
7313 [Steve Henson]
7314
7315 *) Overhauled the Perl interface (perl/*):
7316 - ported BN stuff to OpenSSL's different BN library
7317 - made the perl/ source tree CVS-aware
7318 - renamed the package from SSLeay to OpenSSL (the files still contain
7319 their history because I've copied them in the repository)
7320 - removed obsolete files (the test scripts will be replaced
7321 by better Test::Harness variants in the future)
7322 [Ralf S. Engelschall]
7323
7324 *) First cut for a very conservative source tree cleanup:
7325 1. merge various obsolete readme texts into doc/ssleay.txt
7326 where we collect the old documents and readme texts.
7327 2. remove the first part of files where I'm already sure that we no
7328 longer need them because of three reasons: either they are just temporary
7329 files which were left by Eric or they are preserved original files where
7330 I've verified that the diff is also available in the CVS via "cvs diff
7331 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
7332 the crypto/md/ stuff).
7333 [Ralf S. Engelschall]
7334
7335 *) More extension code. Incomplete support for subject and issuer alt
7336 name, issuer and authority key id. Change the i2v function parameters
7337 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
7338 what that's for :-) Fix to ASN1 macro which messed up
7339 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
7340 [Steve Henson]
7341
7342 *) Preliminary support for ENUMERATED type. This is largely copied from the
7343 INTEGER code.
7344 [Steve Henson]
7345
7346 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
7347 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7348
7349 *) Make sure `make rehash' target really finds the `openssl' program.
7350 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7351
7352 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
7353 like to hear about it if this slows down other processors.
7354 [Ben Laurie]
7355
7356 *) Add CygWin32 platform information to Configure script.
7357 [Alan Batie <batie@aahz.jf.intel.com>]
7358
7359 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
7360 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
7361
7362 *) New program nseq to manipulate netscape certificate sequences
7363 [Steve Henson]
7364
7365 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
7366 few typos.
7367 [Steve Henson]
7368
7369 *) Fixes to BN code. Previously the default was to define BN_RECURSION
7370 but the BN code had some problems that would cause failures when
7371 doing certificate verification and some other functions.
7372 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
7373
7374 *) Add ASN1 and PEM code to support netscape certificate sequences.
7375 [Steve Henson]
7376
7377 *) Add ASN1 and PEM code to support netscape certificate sequences.
7378 [Steve Henson]
7379
7380 *) Add several PKIX and private extended key usage OIDs.
7381 [Steve Henson]
7382
7383 *) Modify the 'ca' program to handle the new extension code. Modify
7384 openssl.cnf for new extension format, add comments.
7385 [Steve Henson]
7386
7387 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
7388 and add a sample to openssl.cnf so req -x509 now adds appropriate
7389 CA extensions.
7390 [Steve Henson]
7391
7392 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
7393 error code, add initial support to X509_print() and x509 application.
7394 [Steve Henson]
7395
7396 *) Takes a deep breath and start addding X509 V3 extension support code. Add
7397 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
7398 stuff is currently isolated and isn't even compiled yet.
7399 [Steve Henson]
7400
7401 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
7402 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
7403 Removed the versions check from X509 routines when loading extensions:
7404 this allows certain broken certificates that don't set the version
7405 properly to be processed.
7406 [Steve Henson]
7407
7408 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
7409 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
7410 can still be regenerated with "make depend".
7411 [Ben Laurie]
7412
7413 *) Spelling mistake in C version of CAST-128.
7414 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
7415
7416 *) Changes to the error generation code. The perl script err-code.pl
7417 now reads in the old error codes and retains the old numbers, only
7418 adding new ones if necessary. It also only changes the .err files if new
7419 codes are added. The makefiles have been modified to only insert errors
7420 when needed (to avoid needlessly modifying header files). This is done
7421 by only inserting errors if the .err file is newer than the auto generated
7422 C file. To rebuild all the error codes from scratch (the old behaviour)
7423 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
7424 or delete all the .err files.
7425 [Steve Henson]
7426
7427 *) CAST-128 was incorrectly implemented for short keys. The C version has
7428 been fixed, but is untested. The assembler versions are also fixed, but
7429 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
7430 to regenerate it if needed.
7431 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
7432 Hagino <itojun@kame.net>]
7433
7434 *) File was opened incorrectly in randfile.c.
7435 [Ulf Möller <ulf@fitug.de>]
7436
7437 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
7438 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
7439 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
7440 al: it's just almost always a UTCTime. Note this patch adds new error
7441 codes so do a "make errors" if there are problems.
7442 [Steve Henson]
7443
7444 *) Correct Linux 1 recognition in config.
7445 [Ulf Möller <ulf@fitug.de>]
7446
7447 *) Remove pointless MD5 hash when using DSA keys in ca.
7448 [Anonymous <nobody@replay.com>]
7449
7450 *) Generate an error if given an empty string as a cert directory. Also
7451 generate an error if handed NULL (previously returned 0 to indicate an
7452 error, but didn't set one).
7453 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
7454
7455 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
7456 [Ben Laurie]
7457
7458 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
7459 parameters. This was causing a warning which killed off the Win32 compile.
7460 [Steve Henson]
7461
7462 *) Remove C++ style comments from crypto/bn/bn_local.h.
7463 [Neil Costigan <neil.costigan@celocom.com>]
7464
7465 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
7466 based on a text string, looking up short and long names and finally
7467 "dot" format. The "dot" format stuff didn't work. Added new function
7468 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
7469 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
7470 OID is not part of the table.
7471 [Steve Henson]
7472
7473 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
7474 X509_LOOKUP_by_alias().
7475 [Ben Laurie]
7476
7477 *) Sort openssl functions by name.
7478 [Ben Laurie]
7479
7480 *) Get the gendsa program working (hopefully) and add it to app list. Remove
7481 encryption from sample DSA keys (in case anyone is interested the password
7482 was "1234").
7483 [Steve Henson]
7484
7485 *) Make _all_ *_free functions accept a NULL pointer.
7486 [Frans Heymans <fheymans@isaserver.be>]
7487
7488 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
7489 NULL pointers.
7490 [Anonymous <nobody@replay.com>]
7491
7492 *) s_server should send the CAfile as acceptable CAs, not its own cert.
7493 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7494
7495 *) Don't blow it for numeric -newkey arguments to apps/req.
7496 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
7497
7498 *) Temp key "for export" tests were wrong in s3_srvr.c.
7499 [Anonymous <nobody@replay.com>]
7500
7501 *) Add prototype for temp key callback functions
7502 SSL_CTX_set_tmp_{rsa,dh}_callback().
7503 [Ben Laurie]
7504
7505 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
7506 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
7507 [Steve Henson]
7508
7509 *) X509_name_add_entry() freed the wrong thing after an error.
7510 [Arne Ansper <arne@ats.cyber.ee>]
7511
7512 *) rsa_eay.c would attempt to free a NULL context.
7513 [Arne Ansper <arne@ats.cyber.ee>]
7514
7515 *) BIO_s_socket() had a broken should_retry() on Windoze.
7516 [Arne Ansper <arne@ats.cyber.ee>]
7517
7518 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
7519 [Arne Ansper <arne@ats.cyber.ee>]
7520
7521 *) Make sure the already existing X509_STORE->depth variable is initialized
7522 in X509_STORE_new(), but document the fact that this variable is still
7523 unused in the certificate verification process.
7524 [Ralf S. Engelschall]
7525
7526 *) Fix the various library and apps files to free up pkeys obtained from
7527 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
7528 [Steve Henson]
7529
7530 *) Fix reference counting in X509_PUBKEY_get(). This makes
7531 demos/maurice/example2.c work, amongst others, probably.
7532 [Steve Henson and Ben Laurie]
7533
7534 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
7535 `openssl' and second, the shortcut symlinks for the `openssl <command>'
7536 are no longer created. This way we have a single and consistent command
7537 line interface `openssl <command>', similar to `cvs <command>'.
7538 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
7539
7540 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
7541 BIT STRING wrapper always have zero unused bits.
7542 [Steve Henson]
7543
7544 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
7545 [Steve Henson]
7546
7547 *) Make the top-level INSTALL documentation easier to understand.
7548 [Paul Sutton]
7549
7550 *) Makefiles updated to exit if an error occurs in a sub-directory
7551 make (including if user presses ^C) [Paul Sutton]
7552
7553 *) Make Montgomery context stuff explicit in RSA data structure.
7554 [Ben Laurie]
7555
7556 *) Fix build order of pem and err to allow for generated pem.h.
7557 [Ben Laurie]
7558
7559 *) Fix renumbering bug in X509_NAME_delete_entry().
7560 [Ben Laurie]
7561
7562 *) Enhanced the err-ins.pl script so it makes the error library number
7563 global and can add a library name. This is needed for external ASN1 and
7564 other error libraries.
7565 [Steve Henson]
7566
7567 *) Fixed sk_insert which never worked properly.
7568 [Steve Henson]
7569
7570 *) Fix ASN1 macros so they can handle indefinite length construted
7571 EXPLICIT tags. Some non standard certificates use these: they can now
7572 be read in.
7573 [Steve Henson]
7574
7575 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
7576 into a single doc/ssleay.txt bundle. This way the information is still
7577 preserved but no longer messes up this directory. Now it's new room for
7578 the new set of documenation files.
7579 [Ralf S. Engelschall]
7580
7581 *) SETs were incorrectly DER encoded. This was a major pain, because they
7582 shared code with SEQUENCEs, which aren't coded the same. This means that
7583 almost everything to do with SETs or SEQUENCEs has either changed name or
7584 number of arguments.
7585 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
7586
7587 *) Fix test data to work with the above.
7588 [Ben Laurie]
7589
7590 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
7591 was already fixed by Eric for 0.9.1 it seems.
7592 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
7593
7594 *) Autodetect FreeBSD3.
7595 [Ben Laurie]
7596
7597 *) Fix various bugs in Configure. This affects the following platforms:
7598 nextstep
7599 ncr-scde
7600 unixware-2.0
7601 unixware-2.0-pentium
7602 sco5-cc.
7603 [Ben Laurie]
7604
7605 *) Eliminate generated files from CVS. Reorder tests to regenerate files
7606 before they are needed.
7607 [Ben Laurie]
7608
7609 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
7610 [Ben Laurie]
7611
7612
7613 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
7614
7615 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
7616 changed SSLeay to OpenSSL in version strings.
7617 [Ralf S. Engelschall]
7618
7619 *) Some fixups to the top-level documents.
7620 [Paul Sutton]
7621
7622 *) Fixed the nasty bug where rsaref.h was not found under compile-time
7623 because the symlink to include/ was missing.
7624 [Ralf S. Engelschall]
7625
7626 *) Incorporated the popular no-RSA/DSA-only patches
7627 which allow to compile a RSA-free SSLeay.
7628 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
7629
7630 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
7631 when "ssleay" is still not found.
7632 [Ralf S. Engelschall]
7633
7634 *) Added more platforms to Configure: Cray T3E, HPUX 11,
7635 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
7636
7637 *) Updated the README file.
7638 [Ralf S. Engelschall]
7639
7640 *) Added various .cvsignore files in the CVS repository subdirs
7641 to make a "cvs update" really silent.
7642 [Ralf S. Engelschall]
7643
7644 *) Recompiled the error-definition header files and added
7645 missing symbols to the Win32 linker tables.
7646 [Ralf S. Engelschall]
7647
7648 *) Cleaned up the top-level documents;
7649 o new files: CHANGES and LICENSE
7650 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
7651 o merged COPYRIGHT into LICENSE
7652 o removed obsolete TODO file
7653 o renamed MICROSOFT to INSTALL.W32
7654 [Ralf S. Engelschall]
7655
7656 *) Removed dummy files from the 0.9.1b source tree:
7657 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
7658 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
7659 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
7660 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
7661 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
7662 [Ralf S. Engelschall]
7663
7664 *) Added various platform portability fixes.
7665 [Mark J. Cox]
7666
7667 *) The Genesis of the OpenSSL rpject:
7668 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
7669 Young and Tim J. Hudson created while they were working for C2Net until
7670 summer 1998.
7671 [The OpenSSL Project]
7672
7673
7674 Changes between 0.9.0b and 0.9.1b [not released]
7675
7676 *) Updated a few CA certificates under certs/
7677 [Eric A. Young]
7678
7679 *) Changed some BIGNUM api stuff.
7680 [Eric A. Young]
7681
7682 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
7683 DGUX x86, Linux Alpha, etc.
7684 [Eric A. Young]
7685
7686 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
7687 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
7688 available).
7689 [Eric A. Young]
7690
7691 *) Add -strparse option to asn1pars program which parses nested
7692 binary structures
7693 [Dr Stephen Henson <shenson@bigfoot.com>]
7694
7695 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
7696 [Eric A. Young]
7697
7698 *) DSA fix for "ca" program.
7699 [Eric A. Young]
7700
7701 *) Added "-genkey" option to "dsaparam" program.
7702 [Eric A. Young]
7703
7704 *) Added RIPE MD160 (rmd160) message digest.
7705 [Eric A. Young]
7706
7707 *) Added -a (all) option to "ssleay version" command.
7708 [Eric A. Young]
7709
7710 *) Added PLATFORM define which is the id given to Configure.
7711 [Eric A. Young]
7712
7713 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
7714 [Eric A. Young]
7715
7716 *) Extended the ASN.1 parser routines.
7717 [Eric A. Young]
7718
7719 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
7720 [Eric A. Young]
7721
7722 *) Added a BN_CTX to the BN library.
7723 [Eric A. Young]
7724
7725 *) Fixed the weak key values in DES library
7726 [Eric A. Young]
7727
7728 *) Changed API in EVP library for cipher aliases.
7729 [Eric A. Young]
7730
7731 *) Added support for RC2/64bit cipher.
7732 [Eric A. Young]
7733
7734 *) Converted the lhash library to the crypto/mem.c functions.
7735 [Eric A. Young]
7736
7737 *) Added more recognized ASN.1 object ids.
7738 [Eric A. Young]
7739
7740 *) Added more RSA padding checks for SSL/TLS.
7741 [Eric A. Young]
7742
7743 *) Added BIO proxy/filter functionality.
7744 [Eric A. Young]
7745
7746 *) Added extra_certs to SSL_CTX which can be used
7747 send extra CA certificates to the client in the CA cert chain sending
7748 process. It can be configured with SSL_CTX_add_extra_chain_cert().
7749 [Eric A. Young]
7750
7751 *) Now Fortezza is denied in the authentication phase because
7752 this is key exchange mechanism is not supported by SSLeay at all.
7753 [Eric A. Young]
7754
7755 *) Additional PKCS1 checks.
7756 [Eric A. Young]
7757
7758 *) Support the string "TLSv1" for all TLS v1 ciphers.
7759 [Eric A. Young]
7760
7761 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
7762 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
7763 [Eric A. Young]
7764
7765 *) Fixed a few memory leaks.
7766 [Eric A. Young]
7767
7768 *) Fixed various code and comment typos.
7769 [Eric A. Young]
7770
7771 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
7772 bytes sent in the client random.
7773 [Edward Bishop <ebishop@spyglass.com>]
7774