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