]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Add flags for DH FIPS method.
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 1.0.1 and 1.1.0 [xx XXX xxxx]
6
7 *) Output TLS supported curves in preference order instead of numerical
8 order. This is currently hardcoded for the highest order curves first.
9 This should be configurable so applications can judge speed vs strength.
10 [Steve Henson]
11
12 *) Add protection against ECDSA timing attacks as mentioned in the paper
13 by Billy Bob Brumley and Nicola Tuveri, see:
14
15 http://eprint.iacr.org/2011/232.pdf
16
17 [Billy Bob Brumley and Nicola Tuveri]
18
19 *) Add TLS v1.2 server support for client authentication.
20 [Steve Henson]
21
22 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
23 and enable MD5.
24 [Steve Henson]
25
26 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
27 FIPS modules versions.
28 [Steve Henson]
29
30 *) Add TLS v1.2 client side support for client authentication. Keep cache
31 of handshake records longer as we don't know the hash algorithm to use
32 until after the certificate request message is received.
33 [Steve Henson]
34
35 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
36 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
37 outside the validated module in the FIPS capable OpenSSL.
38 [Steve Henson]
39
40 *) Initial TLS v1.2 client support. Add a default signature algorithms
41 extension including all the algorithms we support. Parse new signature
42 format in client key exchange. Relax some ECC signing restrictions for
43 TLS v1.2 as indicated in RFC5246.
44 [Steve Henson]
45
46 *) Add server support for TLS v1.2 signature algorithms extension. Switch
47 to new signature format when needed using client digest preference.
48 All server ciphersuites should now work correctly in TLS v1.2. No client
49 support yet and no support for client certificates.
50 [Steve Henson]
51
52 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
53 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
54 ciphersuites. At present only RSA key exchange ciphersuites work with
55 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
56 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
57 and version checking.
58 [Steve Henson]
59
60 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
61 with this defined it will not be affected by any changes to ssl internal
62 structures. Add several utility functions to allow openssl application
63 to work with OPENSSL_NO_SSL_INTERN defined.
64 [Steve Henson]
65
66 *) Minor change to DRBG entropy callback semantics. In some cases
67 there is no mutiple of the block length between min_len and
68 max_len. Allow the callback to return more than max_len bytes
69 of entropy but discard any extra: it is the callback's responsibility
70 to ensure that the extra data discarded does not impact the
71 requested amount of entropy.
72 [Steve Henson]
73
74 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
75 information in FIPS186-3, SP800-57 and SP800-131A.
76 [Steve Henson]
77
78 *) CCM support via EVP. Interface is very similar to GCM case except we
79 must supply all data in one chunk (i.e. no update, final) and the
80 message length must be supplied if AAD is used. Add algorithm test
81 support.
82 [Steve Henson]
83
84 *) Initial version of POST overhaul. Add POST callback to allow the status
85 of POST to be monitored and/or failures induced. Modify fips_test_suite
86 to use callback. Always run all selftests even if one fails.
87 [Steve Henson]
88
89 *) XTS support including algorithm test driver in the fips_gcmtest program.
90 Note: this does increase the maximum key length from 32 to 64 bytes but
91 there should be no binary compatibility issues as existing applications
92 will never use XTS mode.
93 [Steve Henson]
94
95 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
96 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
97 performs algorithm blocking for unapproved PRNG types. Also do not
98 set PRNG type in FIPS_mode_set(): leave this to the application.
99 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
100 the standard OpenSSL PRNG: set additional data to a date time vector.
101 [Steve Henson]
102
103 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
104 This shouldn't present any incompatibility problems because applications
105 shouldn't be using these directly and any that are will need to rethink
106 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
107 [Steve Henson]
108
109 *) Extensive self tests and health checking required by SP800-90 DRBG.
110 Remove strength parameter from FIPS_drbg_instantiate and always
111 instantiate at maximum supported strength.
112 [Steve Henson]
113
114 *) Add SRP support.
115 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
116
117 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
118 [Steve Henson]
119
120 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
121 [Steve Henson]
122
123 *) New function DH_compute_key_padded() to compute a DH key and pad with
124 leading zeroes if needed: this complies with SP800-56A et al.
125 [Steve Henson]
126
127 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
128 anything, incomplete, subject to change and largely untested at present.
129 [Steve Henson]
130
131 *) Modify fipscanisteronly build option to only build the necessary object
132 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
133 [Steve Henson]
134
135 *) Add experimental option FIPSSYMS to give all symbols in
136 fipscanister.o and FIPS or fips prefix. This will avoid
137 conflicts with future versions of OpenSSL. Add perl script
138 util/fipsas.pl to preprocess assembly language source files
139 and rename any affected symbols.
140 [Steve Henson]
141
142 *) Add selftest checks and algorithm block of non-fips algorithms in
143 FIPS mode. Remove DES2 from selftests.
144 [Steve Henson]
145
146 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
147 return internal method without any ENGINE dependencies. Add new
148 tiny fips sign and verify functions.
149 [Steve Henson]
150
151 *) New build option no-ec2m to disable characteristic 2 code.
152 [Steve Henson]
153
154 *) New build option "fipscanisteronly". This only builds fipscanister.o
155 and (currently) associated fips utilities. Uses the file Makefile.fips
156 instead of Makefile.org as the prototype.
157 [Steve Henson]
158
159 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
160 Update fips_gcmtest to use IV generator.
161 [Steve Henson]
162
163 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
164 setting output buffer to NULL. The *Final function must be
165 called although it will not retrieve any additional data. The tag
166 can be set or retrieved with a ctrl. The IV length is by default 12
167 bytes (96 bits) but can be set to an alternative value. If the IV
168 length exceeds the maximum IV length (currently 16 bytes) it cannot be
169 set before the key.
170 [Steve Henson]
171
172 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
173 underlying do_cipher function handles all cipher semantics itself
174 including padding and finalisation. This is useful if (for example)
175 an ENGINE cipher handles block padding itself. The behaviour of
176 do_cipher is subtly changed if this flag is set: the return value
177 is the number of characters written to the output buffer (zero is
178 no longer an error code) or a negative error code. Also if the
179 input buffer is NULL and length 0 finalisation should be performed.
180 [Steve Henson]
181
182 *) If a candidate issuer certificate is already part of the constructed
183 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
184 [Steve Henson]
185
186 *) Improve forward-security support: add functions
187
188 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
189 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
190
191 for use by SSL/TLS servers; the callback function will be called whenever a
192 new session is created, and gets to decide whether the session may be
193 cached to make it resumable (return 0) or not (return 1). (As by the
194 SSL/TLS protocol specifications, the session_id sent by the server will be
195 empty to indicate that the session is not resumable; also, the server will
196 not generate RFC 4507 (RFC 5077) session tickets.)
197
198 A simple reasonable callback implementation is to return is_forward_secure.
199 This parameter will be set to 1 or 0 depending on the ciphersuite selected
200 by the SSL/TLS server library, indicating whether it can provide forward
201 security.
202 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
203
204 *) Add Next Protocol Negotiation,
205 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
206 disabled with a no-npn flag to config or Configure. Code donated
207 by Google.
208 [Adam Langley <agl@google.com> and Ben Laurie]
209
210 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
211 all platforms. Move ssize_t definition from e_os.h to the public
212 header file e_os2.h as it now appears in public header file cms.h
213 [Steve Henson]
214
215 *) New function OPENSSL_gmtime_diff to find the difference in days
216 and seconds between two tm structures. This will be used to provide
217 additional functionality for ASN1_TIME.
218 [Steve Henson]
219
220 *) New -sigopt option to the ca, req and x509 utilities. Additional
221 signature parameters can be passed using this option and in
222 particular PSS.
223 [Steve Henson]
224
225 *) Add RSA PSS signing function. This will generate and set the
226 appropriate AlgorithmIdentifiers for PSS based on those in the
227 corresponding EVP_MD_CTX structure. No application support yet.
228 [Steve Henson]
229
230 *) Support for companion algorithm specific ASN1 signing routines.
231 New function ASN1_item_sign_ctx() signs a pre-initialised
232 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
233 the appropriate parameters.
234 [Steve Henson]
235
236 *) Add new algorithm specific ASN1 verification initialisation function
237 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
238 handling will be the same no matter what EVP_PKEY_METHOD is used.
239 Add a PSS handler to support verification of PSS signatures: checked
240 against a number of sample certificates.
241 [Steve Henson]
242
243 *) Add signature printing for PSS. Add PSS OIDs.
244 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
245
246 *) Add algorithm specific signature printing. An individual ASN1 method
247 can now print out signatures instead of the standard hex dump.
248
249 More complex signatures (e.g. PSS) can print out more meaningful
250 information. Include DSA version that prints out the signature
251 parameters r, s.
252 [Steve Henson]
253
254 *) Add -trusted_first option which attempts to find certificates in the
255 trusted store even if an untrusted chain is also supplied.
256 [Steve Henson]
257
258 *) Initial experimental support for explicitly trusted non-root CAs.
259 OpenSSL still tries to build a complete chain to a root but if an
260 intermediate CA has a trust setting included that is used. The first
261 setting is used: whether to trust or reject.
262 [Steve Henson]
263
264 *) New -verify_name option in command line utilities to set verification
265 parameters by name.
266 [Steve Henson]
267
268 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
269 Add CMAC pkey methods.
270 [Steve Henson]
271
272 *) Experiemental regnegotiation in s_server -www mode. If the client
273 browses /reneg connection is renegotiated. If /renegcert it is
274 renegotiated requesting a certificate.
275 [Steve Henson]
276
277 *) Add an "external" session cache for debugging purposes to s_server. This
278 should help trace issues which normally are only apparent in deployed
279 multi-process servers.
280 [Steve Henson]
281
282 *) Experiemental password based recipient info support for CMS library:
283 implementing RFC3211.
284 [Steve Henson]
285
286 *) Split password based encryption into PBES2 and PBKDF2 functions. This
287 neatly separates the code into cipher and PBE sections and is required
288 for some algorithms that split PBES2 into separate pieces (such as
289 password based CMS).
290 [Steve Henson]
291
292 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
293 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
294 BIO_set_cipher() and some obscure PEM functions were changed so they
295 can now return an error. The RAND changes required a change to the
296 RAND_METHOD structure.
297 [Steve Henson]
298
299 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
300 a gcc attribute to warn if the result of a function is ignored. This
301 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
302 whose return value is often ignored.
303 [Steve Henson]
304
305 Changes between 1.0.0d and 1.0.1 [xx XXX xxxx]
306
307 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
308 [Steve Henson]
309
310 *) Add EC_GFp_nistp224_method(), a 64-bit optimized implementation for
311 elliptic curve NIST-P224 with constant-time single point multiplication on
312 typical inputs. EC_GROUP_new_by_curve_name() will automatically use this
313 (while EC_GROUP_new_curve_GFp() currently won't and prefers the more
314 flexible implementations).
315
316 The implementation requires support for the nonstandard type __uint128_t,
317 and so is disabled by default. To include this in your build of OpenSSL,
318 use -DEC_NISTP224_64_GCC_128 on the Configure (or config) command line,
319 and run "make depend" (or "make update").
320 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
321
322 *) Permit abbreviated handshakes when renegotiating using the function
323 SSL_renegotiate_abbreviated().
324 [Robin Seggelmann <seggelmann@fh-muenster.de>]
325
326 *) Add call to ENGINE_register_all_complete() to
327 ENGINE_load_builtin_engines(), so some implementations get used
328 automatically instead of needing explicit application support.
329 [Steve Henson]
330
331 *) Add support for TLS key exporter as described in RFC5705.
332 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
333
334 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
335 a few changes are required:
336
337 Add SSL_OP_NO_TLSv1_1 flag.
338 Add TLSv1_1 methods.
339 Update version checking logic to handle version 1.1.
340 Add explicit IV handling (ported from DTLS code).
341 Add command line options to s_client/s_server.
342 [Steve Henson]
343
344 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
345
346 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
347 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
348
349 *) Fix bug in string printing code: if *any* escaping is enabled we must
350 escape the escape character (backslash) or the resulting string is
351 ambiguous.
352 [Steve Henson]
353
354 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
355
356 *) Disable code workaround for ancient and obsolete Netscape browsers
357 and servers: an attacker can use it in a ciphersuite downgrade attack.
358 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
359 [Steve Henson]
360
361 *) Fixed J-PAKE implementation error, originally discovered by
362 Sebastien Martini, further info and confirmation from Stefan
363 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
364 [Ben Laurie]
365
366 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
367
368 *) Fix extension code to avoid race conditions which can result in a buffer
369 overrun vulnerability: resumed sessions must not be modified as they can
370 be shared by multiple threads. CVE-2010-3864
371 [Steve Henson]
372
373 *) Fix WIN32 build system to correctly link an ENGINE directory into
374 a DLL.
375 [Steve Henson]
376
377 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
378
379 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
380 (CVE-2010-1633)
381 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
382
383 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
384
385 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
386 context. The operation can be customised via the ctrl mechanism in
387 case ENGINEs want to include additional functionality.
388 [Steve Henson]
389
390 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
391 [Steve Henson]
392
393 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
394 output hashes compatible with older versions of OpenSSL.
395 [Willy Weisz <weisz@vcpc.univie.ac.at>]
396
397 *) Fix compression algorithm handling: if resuming a session use the
398 compression algorithm of the resumed session instead of determining
399 it from client hello again. Don't allow server to change algorithm.
400 [Steve Henson]
401
402 *) Add load_crls() function to apps tidying load_certs() too. Add option
403 to verify utility to allow additional CRLs to be included.
404 [Steve Henson]
405
406 *) Update OCSP request code to permit adding custom headers to the request:
407 some responders need this.
408 [Steve Henson]
409
410 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
411 correctly.
412 [Julia Lawall <julia@diku.dk>]
413
414 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
415 needlessly dereferenced structures, used obsolete functions and
416 didn't handle all updated verify codes correctly.
417 [Steve Henson]
418
419 *) Disable MD2 in the default configuration.
420 [Steve Henson]
421
422 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
423 indicate the initial BIO being pushed or popped. This makes it possible
424 to determine whether the BIO is the one explicitly called or as a result
425 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
426 it handles reference counts correctly and doesn't zero out the I/O bio
427 when it is not being explicitly popped. WARNING: applications which
428 included workarounds for the old buggy behaviour will need to be modified
429 or they could free up already freed BIOs.
430 [Steve Henson]
431
432 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
433 renaming to all platforms (within the 0.9.8 branch, this was
434 done conditionally on Netware platforms to avoid a name clash).
435 [Guenter <lists@gknw.net>]
436
437 *) Add ECDHE and PSK support to DTLS.
438 [Michael Tuexen <tuexen@fh-muenster.de>]
439
440 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
441 be used on C++.
442 [Steve Henson]
443
444 *) Add "missing" function EVP_MD_flags() (without this the only way to
445 retrieve a digest flags is by accessing the structure directly. Update
446 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
447 or cipher is registered as in the "from" argument. Print out all
448 registered digests in the dgst usage message instead of manually
449 attempting to work them out.
450 [Steve Henson]
451
452 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
453 this allows the use of compression and extensions. Change default cipher
454 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
455 by default unless an application cipher string requests it.
456 [Steve Henson]
457
458 *) Alter match criteria in PKCS12_parse(). It used to try to use local
459 key ids to find matching certificates and keys but some PKCS#12 files
460 don't follow the (somewhat unwritten) rules and this strategy fails.
461 Now just gather all certificates together and the first private key
462 then look for the first certificate that matches the key.
463 [Steve Henson]
464
465 *) Support use of registered digest and cipher names for dgst and cipher
466 commands instead of having to add each one as a special case. So now
467 you can do:
468
469 openssl sha256 foo
470
471 as well as:
472
473 openssl dgst -sha256 foo
474
475 and this works for ENGINE based algorithms too.
476
477 [Steve Henson]
478
479 *) Update Gost ENGINE to support parameter files.
480 [Victor B. Wagner <vitus@cryptocom.ru>]
481
482 *) Support GeneralizedTime in ca utility.
483 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
484
485 *) Enhance the hash format used for certificate directory links. The new
486 form uses the canonical encoding (meaning equivalent names will work
487 even if they aren't identical) and uses SHA1 instead of MD5. This form
488 is incompatible with the older format and as a result c_rehash should
489 be used to rebuild symbolic links.
490 [Steve Henson]
491
492 *) Make PKCS#8 the default write format for private keys, replacing the
493 traditional format. This form is standardised, more secure and doesn't
494 include an implicit MD5 dependency.
495 [Steve Henson]
496
497 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
498 committed to OpenSSL should pass this lot as a minimum.
499 [Steve Henson]
500
501 *) Add session ticket override functionality for use by EAP-FAST.
502 [Jouni Malinen <j@w1.fi>]
503
504 *) Modify HMAC functions to return a value. Since these can be implemented
505 in an ENGINE errors can occur.
506 [Steve Henson]
507
508 *) Type-checked OBJ_bsearch_ex.
509 [Ben Laurie]
510
511 *) Type-checked OBJ_bsearch. Also some constification necessitated
512 by type-checking. Still to come: TXT_DB, bsearch(?),
513 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
514 CONF_VALUE.
515 [Ben Laurie]
516
517 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
518 seconds to a tm structure directly, instead of going through OS
519 specific date routines. This avoids any issues with OS routines such
520 as the year 2038 bug. New *_adj() functions for ASN1 time structures
521 and X509_time_adj_ex() to cover the extended range. The existing
522 X509_time_adj() is still usable and will no longer have any date issues.
523 [Steve Henson]
524
525 *) Delta CRL support. New use deltas option which will attempt to locate
526 and search any appropriate delta CRLs available.
527
528 This work was sponsored by Google.
529 [Steve Henson]
530
531 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
532 code and add additional score elements. Validate alternate CRL paths
533 as part of the CRL checking and indicate a new error "CRL path validation
534 error" in this case. Applications wanting additional details can use
535 the verify callback and check the new "parent" field. If this is not
536 NULL CRL path validation is taking place. Existing applications wont
537 see this because it requires extended CRL support which is off by
538 default.
539
540 This work was sponsored by Google.
541 [Steve Henson]
542
543 *) Support for freshest CRL extension.
544
545 This work was sponsored by Google.
546 [Steve Henson]
547
548 *) Initial indirect CRL support. Currently only supported in the CRLs
549 passed directly and not via lookup. Process certificate issuer
550 CRL entry extension and lookup CRL entries by bother issuer name
551 and serial number. Check and process CRL issuer entry in IDP extension.
552
553 This work was sponsored by Google.
554 [Steve Henson]
555
556 *) Add support for distinct certificate and CRL paths. The CRL issuer
557 certificate is validated separately in this case. Only enabled if
558 an extended CRL support flag is set: this flag will enable additional
559 CRL functionality in future.
560
561 This work was sponsored by Google.
562 [Steve Henson]
563
564 *) Add support for policy mappings extension.
565
566 This work was sponsored by Google.
567 [Steve Henson]
568
569 *) Fixes to pathlength constraint, self issued certificate handling,
570 policy processing to align with RFC3280 and PKITS tests.
571
572 This work was sponsored by Google.
573 [Steve Henson]
574
575 *) Support for name constraints certificate extension. DN, email, DNS
576 and URI types are currently supported.
577
578 This work was sponsored by Google.
579 [Steve Henson]
580
581 *) To cater for systems that provide a pointer-based thread ID rather
582 than numeric, deprecate the current numeric thread ID mechanism and
583 replace it with a structure and associated callback type. This
584 mechanism allows a numeric "hash" to be extracted from a thread ID in
585 either case, and on platforms where pointers are larger than 'long',
586 mixing is done to help ensure the numeric 'hash' is usable even if it
587 can't be guaranteed unique. The default mechanism is to use "&errno"
588 as a pointer-based thread ID to distinguish between threads.
589
590 Applications that want to provide their own thread IDs should now use
591 CRYPTO_THREADID_set_callback() to register a callback that will call
592 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
593
594 Note that ERR_remove_state() is now deprecated, because it is tied
595 to the assumption that thread IDs are numeric. ERR_remove_state(0)
596 to free the current thread's error state should be replaced by
597 ERR_remove_thread_state(NULL).
598
599 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
600 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
601 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
602 application was previously providing a numeric thread callback that
603 was inappropriate for distinguishing threads, then uniqueness might
604 have been obtained with &errno that happened immediately in the
605 intermediate development versions of OpenSSL; this is no longer the
606 case, the numeric thread callback will now override the automatic use
607 of &errno.)
608 [Geoff Thorpe, with help from Bodo Moeller]
609
610 *) Initial support for different CRL issuing certificates. This covers a
611 simple case where the self issued certificates in the chain exist and
612 the real CRL issuer is higher in the existing chain.
613
614 This work was sponsored by Google.
615 [Steve Henson]
616
617 *) Removed effectively defunct crypto/store from the build.
618 [Ben Laurie]
619
620 *) Revamp of STACK to provide stronger type-checking. Still to come:
621 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
622 ASN1_STRING, CONF_VALUE.
623 [Ben Laurie]
624
625 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
626 RAM on SSL connections. This option can save about 34k per idle SSL.
627 [Nick Mathewson]
628
629 *) Revamp of LHASH to provide stronger type-checking. Still to come:
630 STACK, TXT_DB, bsearch, qsort.
631 [Ben Laurie]
632
633 *) Initial support for Cryptographic Message Syntax (aka CMS) based
634 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
635 support for data, signedData, compressedData, digestedData and
636 encryptedData, envelopedData types included. Scripts to check against
637 RFC4134 examples draft and interop and consistency checks of many
638 content types and variants.
639 [Steve Henson]
640
641 *) Add options to enc utility to support use of zlib compression BIO.
642 [Steve Henson]
643
644 *) Extend mk1mf to support importing of options and assembly language
645 files from Configure script, currently only included in VC-WIN32.
646 The assembly language rules can now optionally generate the source
647 files from the associated perl scripts.
648 [Steve Henson]
649
650 *) Implement remaining functionality needed to support GOST ciphersuites.
651 Interop testing has been performed using CryptoPro implementations.
652 [Victor B. Wagner <vitus@cryptocom.ru>]
653
654 *) s390x assembler pack.
655 [Andy Polyakov]
656
657 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
658 "family."
659 [Andy Polyakov]
660
661 *) Implement Opaque PRF Input TLS extension as specified in
662 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
663 official specification yet and no extension type assignment by
664 IANA exists, this extension (for now) will have to be explicitly
665 enabled when building OpenSSL by providing the extension number
666 to use. For example, specify an option
667
668 -DTLSEXT_TYPE_opaque_prf_input=0x9527
669
670 to the "config" or "Configure" script to enable the extension,
671 assuming extension number 0x9527 (which is a completely arbitrary
672 and unofficial assignment based on the MD5 hash of the Internet
673 Draft). Note that by doing so, you potentially lose
674 interoperability with other TLS implementations since these might
675 be using the same extension number for other purposes.
676
677 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
678 opaque PRF input value to use in the handshake. This will create
679 an interal copy of the length-'len' string at 'src', and will
680 return non-zero for success.
681
682 To get more control and flexibility, provide a callback function
683 by using
684
685 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
686 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
687
688 where
689
690 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
691 void *arg;
692
693 Callback function 'cb' will be called in handshakes, and is
694 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
695 Argument 'arg' is for application purposes (the value as given to
696 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
697 be provided to the callback function). The callback function
698 has to return non-zero to report success: usually 1 to use opaque
699 PRF input just if possible, or 2 to enforce use of the opaque PRF
700 input. In the latter case, the library will abort the handshake
701 if opaque PRF input is not successfully negotiated.
702
703 Arguments 'peerinput' and 'len' given to the callback function
704 will always be NULL and 0 in the case of a client. A server will
705 see the client's opaque PRF input through these variables if
706 available (NULL and 0 otherwise). Note that if the server
707 provides an opaque PRF input, the length must be the same as the
708 length of the client's opaque PRF input.
709
710 Note that the callback function will only be called when creating
711 a new session (session resumption can resume whatever was
712 previously negotiated), and will not be called in SSL 2.0
713 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
714 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
715 for applications that need to enforce opaque PRF input.
716
717 [Bodo Moeller]
718
719 *) Update ssl code to support digests other than SHA1+MD5 for handshake
720 MAC.
721
722 [Victor B. Wagner <vitus@cryptocom.ru>]
723
724 *) Add RFC4507 support to OpenSSL. This includes the corrections in
725 RFC4507bis. The encrypted ticket format is an encrypted encoded
726 SSL_SESSION structure, that way new session features are automatically
727 supported.
728
729 If a client application caches session in an SSL_SESSION structure
730 support is transparent because tickets are now stored in the encoded
731 SSL_SESSION.
732
733 The SSL_CTX structure automatically generates keys for ticket
734 protection in servers so again support should be possible
735 with no application modification.
736
737 If a client or server wishes to disable RFC4507 support then the option
738 SSL_OP_NO_TICKET can be set.
739
740 Add a TLS extension debugging callback to allow the contents of any client
741 or server extensions to be examined.
742
743 This work was sponsored by Google.
744 [Steve Henson]
745
746 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
747 OpenSSL should now compile cleanly on gcc 4.2
748 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
749
750 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
751 support including streaming MAC support: this is required for GOST
752 ciphersuite support.
753 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
754
755 *) Add option -stream to use PKCS#7 streaming in smime utility. New
756 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
757 to output in BER and PEM format.
758 [Steve Henson]
759
760 *) Experimental support for use of HMAC via EVP_PKEY interface. This
761 allows HMAC to be handled via the EVP_DigestSign*() interface. The
762 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
763 ENGINE support for HMAC keys which are unextractable. New -mac and
764 -macopt options to dgst utility.
765 [Steve Henson]
766
767 *) New option -sigopt to dgst utility. Update dgst to use
768 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
769 alternative signing paramaters such as X9.31 or PSS in the dgst
770 utility.
771 [Steve Henson]
772
773 *) Change ssl_cipher_apply_rule(), the internal function that does
774 the work each time a ciphersuite string requests enabling
775 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
776 removing ("!foo+bar") a class of ciphersuites: Now it maintains
777 the order of disabled ciphersuites such that those ciphersuites
778 that most recently went from enabled to disabled not only stay
779 in order with respect to each other, but also have higher priority
780 than other disabled ciphersuites the next time ciphersuites are
781 enabled again.
782
783 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
784 the same ciphersuites as with "HIGH" alone, but in a specific
785 order where the PSK ciphersuites come first (since they are the
786 most recently disabled ciphersuites when "HIGH" is parsed).
787
788 Also, change ssl_create_cipher_list() (using this new
789 funcionality) such that between otherwise identical
790 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
791 the default order.
792 [Bodo Moeller]
793
794 *) Change ssl_create_cipher_list() so that it automatically
795 arranges the ciphersuites in reasonable order before starting
796 to process the rule string. Thus, the definition for "DEFAULT"
797 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
798 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
799 This makes it much easier to arrive at a reasonable default order
800 in applications for which anonymous ciphers are OK (meaning
801 that you can't actually use DEFAULT).
802 [Bodo Moeller; suggested by Victor Duchovni]
803
804 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
805 processing) into multiple integers instead of setting
806 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
807 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
808 (These masks as well as the individual bit definitions are hidden
809 away into the non-exported interface ssl/ssl_locl.h, so this
810 change to the definition of the SSL_CIPHER structure shouldn't
811 affect applications.) This give us more bits for each of these
812 categories, so there is no longer a need to coagulate AES128 and
813 AES256 into a single algorithm bit, and to coagulate Camellia128
814 and Camellia256 into a single algorithm bit, which has led to all
815 kinds of kludges.
816
817 Thus, among other things, the kludge introduced in 0.9.7m and
818 0.9.8e for masking out AES256 independently of AES128 or masking
819 out Camellia256 independently of AES256 is not needed here in 0.9.9.
820
821 With the change, we also introduce new ciphersuite aliases that
822 so far were missing: "AES128", "AES256", "CAMELLIA128", and
823 "CAMELLIA256".
824 [Bodo Moeller]
825
826 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
827 Use the leftmost N bytes of the signature input if the input is
828 larger than the prime q (with N being the size in bytes of q).
829 [Nils Larsch]
830
831 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
832 it yet and it is largely untested.
833 [Steve Henson]
834
835 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
836 [Nils Larsch]
837
838 *) Initial incomplete changes to avoid need for function casts in OpenSSL
839 some compilers (gcc 4.2 and later) reject their use. Safestack is
840 reimplemented. Update ASN1 to avoid use of legacy functions.
841 [Steve Henson]
842
843 *) Win32/64 targets are linked with Winsock2.
844 [Andy Polyakov]
845
846 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
847 to external functions. This can be used to increase CRL handling
848 efficiency especially when CRLs are very large by (for example) storing
849 the CRL revoked certificates in a database.
850 [Steve Henson]
851
852 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
853 new CRLs added to a directory can be used. New command line option
854 -verify_return_error to s_client and s_server. This causes real errors
855 to be returned by the verify callback instead of carrying on no matter
856 what. This reflects the way a "real world" verify callback would behave.
857 [Steve Henson]
858
859 *) GOST engine, supporting several GOST algorithms and public key formats.
860 Kindly donated by Cryptocom.
861 [Cryptocom]
862
863 *) Partial support for Issuing Distribution Point CRL extension. CRLs
864 partitioned by DP are handled but no indirect CRL or reason partitioning
865 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
866 selected via a scoring technique which handles IDP and AKID in CRLs.
867 [Steve Henson]
868
869 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
870 will ultimately be used for all verify operations: this will remove the
871 X509_STORE dependency on certificate verification and allow alternative
872 lookup methods. X509_STORE based implementations of these two callbacks.
873 [Steve Henson]
874
875 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
876 Modify get_crl() to find a valid (unexpired) CRL if possible.
877 [Steve Henson]
878
879 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
880 this would be called X509_CRL_cmp() but that name is already used by
881 a function that just compares CRL issuer names. Cache several CRL
882 extensions in X509_CRL structure and cache CRLDP in X509.
883 [Steve Henson]
884
885 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
886 this maps equivalent X509_NAME structures into a consistent structure.
887 Name comparison can then be performed rapidly using memcmp().
888 [Steve Henson]
889
890 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
891 utility.
892 [Steve Henson]
893
894 *) Allow digests to supply their own micalg string for S/MIME type using
895 the ctrl EVP_MD_CTRL_MICALG.
896 [Steve Henson]
897
898 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
899 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
900 ctrl. It can then customise the structure before and/or after signing
901 if necessary.
902 [Steve Henson]
903
904 *) New function OBJ_add_sigid() to allow application defined signature OIDs
905 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
906 to free up any added signature OIDs.
907 [Steve Henson]
908
909 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
910 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
911 digest and cipher tables. New options added to openssl utility:
912 list-message-digest-algorithms and list-cipher-algorithms.
913 [Steve Henson]
914
915 *) Change the array representation of binary polynomials: the list
916 of degrees of non-zero coefficients is now terminated with -1.
917 Previously it was terminated with 0, which was also part of the
918 value; thus, the array representation was not applicable to
919 polynomials where t^0 has coefficient zero. This change makes
920 the array representation useful in a more general context.
921 [Douglas Stebila]
922
923 *) Various modifications and fixes to SSL/TLS cipher string
924 handling. For ECC, the code now distinguishes between fixed ECDH
925 with RSA certificates on the one hand and with ECDSA certificates
926 on the other hand, since these are separate ciphersuites. The
927 unused code for Fortezza ciphersuites has been removed.
928
929 For consistency with EDH, ephemeral ECDH is now called "EECDH"
930 (not "ECDHE"). For consistency with the code for DH
931 certificates, use of ECDH certificates is now considered ECDH
932 authentication, not RSA or ECDSA authentication (the latter is
933 merely the CA's signing algorithm and not actively used in the
934 protocol).
935
936 The temporary ciphersuite alias "ECCdraft" is no longer
937 available, and ECC ciphersuites are no longer excluded from "ALL"
938 and "DEFAULT". The following aliases now exist for RFC 4492
939 ciphersuites, most of these by analogy with the DH case:
940
941 kECDHr - ECDH cert, signed with RSA
942 kECDHe - ECDH cert, signed with ECDSA
943 kECDH - ECDH cert (signed with either RSA or ECDSA)
944 kEECDH - ephemeral ECDH
945 ECDH - ECDH cert or ephemeral ECDH
946
947 aECDH - ECDH cert
948 aECDSA - ECDSA cert
949 ECDSA - ECDSA cert
950
951 AECDH - anonymous ECDH
952 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
953
954 [Bodo Moeller]
955
956 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
957 Use correct micalg parameters depending on digest(s) in signed message.
958 [Steve Henson]
959
960 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
961 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
962 [Steve Henson]
963
964 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
965 an engine to register a method. Add ENGINE lookups for methods and
966 functional reference processing.
967 [Steve Henson]
968
969 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
970 EVP_{Sign,Verify}* which allow an application to customise the signature
971 process.
972 [Steve Henson]
973
974 *) New -resign option to smime utility. This adds one or more signers
975 to an existing PKCS#7 signedData structure. Also -md option to use an
976 alternative message digest algorithm for signing.
977 [Steve Henson]
978
979 *) Tidy up PKCS#7 routines and add new functions to make it easier to
980 create PKCS7 structures containing multiple signers. Update smime
981 application to support multiple signers.
982 [Steve Henson]
983
984 *) New -macalg option to pkcs12 utility to allow setting of an alternative
985 digest MAC.
986 [Steve Henson]
987
988 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
989 Reorganize PBE internals to lookup from a static table using NIDs,
990 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
991 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
992 PRF which will be automatically used with PBES2.
993 [Steve Henson]
994
995 *) Replace the algorithm specific calls to generate keys in "req" with the
996 new API.
997 [Steve Henson]
998
999 *) Update PKCS#7 enveloped data routines to use new API. This is now
1000 supported by any public key method supporting the encrypt operation. A
1001 ctrl is added to allow the public key algorithm to examine or modify
1002 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1003 a no op.
1004 [Steve Henson]
1005
1006 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1007 a default digest type to use. In most cases this will be SHA1 but some
1008 algorithms (such as GOST) need to specify an alternative digest. The
1009 return value indicates how strong the prefernce is 1 means optional and
1010 2 is mandatory (that is it is the only supported type). Modify
1011 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1012 use the default md. Update openssl utilities to use the default digest
1013 type for signing if it is not explicitly indicated.
1014 [Steve Henson]
1015
1016 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1017 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1018 signing method from the key type. This effectively removes the link
1019 between digests and public key types.
1020 [Steve Henson]
1021
1022 *) Add an OID cross reference table and utility functions. Its purpose is to
1023 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1024 rsaEncryption. This will allow some of the algorithm specific hackery
1025 needed to use the correct OID to be removed.
1026 [Steve Henson]
1027
1028 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1029 structures for PKCS7_sign(). They are now set up by the relevant public
1030 key ASN1 method.
1031 [Steve Henson]
1032
1033 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1034 [Steve Henson]
1035
1036 *) Add support for key derivation (agreement) in the API, DH method and
1037 pkeyutl.
1038 [Steve Henson]
1039
1040 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1041 public and private key formats. As a side effect these add additional
1042 command line functionality not previously available: DSA signatures can be
1043 generated and verified using pkeyutl and DH key support and generation in
1044 pkey, genpkey.
1045 [Steve Henson]
1046
1047 *) BeOS support.
1048 [Oliver Tappe <zooey@hirschkaefer.de>]
1049
1050 *) New make target "install_html_docs" installs HTML renditions of the
1051 manual pages.
1052 [Oliver Tappe <zooey@hirschkaefer.de>]
1053
1054 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1055 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1056 support key and parameter generation and add initial key generation
1057 functionality for RSA.
1058 [Steve Henson]
1059
1060 *) Add functions for main EVP_PKEY_method operations. The undocumented
1061 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1062 EVP_PKEY_{encrypt,decrypt}_old.
1063 [Steve Henson]
1064
1065 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1066 key API, doesn't do much yet.
1067 [Steve Henson]
1068
1069 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1070 public key algorithms. New option to openssl utility:
1071 "list-public-key-algorithms" to print out info.
1072 [Steve Henson]
1073
1074 *) Implement the Supported Elliptic Curves Extension for
1075 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1076 [Douglas Stebila]
1077
1078 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1079 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1080 [Steve Henson]
1081
1082 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1083 utilities such as rsa, dsa, dsaparam etc except they process any key
1084 type.
1085 [Steve Henson]
1086
1087 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1088 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1089 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1090 structure.
1091 [Steve Henson]
1092
1093 *) Initial support for pluggable public key ASN1.
1094 De-spaghettify the public key ASN1 handling. Move public and private
1095 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1096 algorithm specific handling to a single module within the relevant
1097 algorithm directory. Add functions to allow (near) opaque processing
1098 of public and private key structures.
1099 [Steve Henson]
1100
1101 *) Implement the Supported Point Formats Extension for
1102 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1103 [Douglas Stebila]
1104
1105 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1106 for the psk identity [hint] and the psk callback functions to the
1107 SSL_SESSION, SSL and SSL_CTX structure.
1108
1109 New ciphersuites:
1110 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1111 PSK-AES256-CBC-SHA
1112
1113 New functions:
1114 SSL_CTX_use_psk_identity_hint
1115 SSL_get_psk_identity_hint
1116 SSL_get_psk_identity
1117 SSL_use_psk_identity_hint
1118
1119 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1120
1121 *) Add RFC 3161 compliant time stamp request creation, response generation
1122 and response verification functionality.
1123 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1124
1125 *) Add initial support for TLS extensions, specifically for the server_name
1126 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1127 have new members for a host name. The SSL data structure has an
1128 additional member SSL_CTX *initial_ctx so that new sessions can be
1129 stored in that context to allow for session resumption, even after the
1130 SSL has been switched to a new SSL_CTX in reaction to a client's
1131 server_name extension.
1132
1133 New functions (subject to change):
1134
1135 SSL_get_servername()
1136 SSL_get_servername_type()
1137 SSL_set_SSL_CTX()
1138
1139 New CTRL codes and macros (subject to change):
1140
1141 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1142 - SSL_CTX_set_tlsext_servername_callback()
1143 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1144 - SSL_CTX_set_tlsext_servername_arg()
1145 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1146
1147 openssl s_client has a new '-servername ...' option.
1148
1149 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1150 '-key2 ...', '-servername_fatal' (subject to change). This allows
1151 testing the HostName extension for a specific single host name ('-cert'
1152 and '-key' remain fallbacks for handshakes without HostName
1153 negotiation). If the unrecogninzed_name alert has to be sent, this by
1154 default is a warning; it becomes fatal with the '-servername_fatal'
1155 option.
1156
1157 [Peter Sylvester, Remy Allais, Christophe Renou]
1158
1159 *) Whirlpool hash implementation is added.
1160 [Andy Polyakov]
1161
1162 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1163 bn(64,32). Because of instruction set limitations it doesn't have
1164 any negative impact on performance. This was done mostly in order
1165 to make it possible to share assembler modules, such as bn_mul_mont
1166 implementations, between 32- and 64-bit builds without hassle.
1167 [Andy Polyakov]
1168
1169 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1170 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1171 macro.
1172 [Bodo Moeller]
1173
1174 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1175 dedicated Montgomery multiplication procedure, is introduced.
1176 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1177 "64-bit" performance on certain 32-bit targets.
1178 [Andy Polyakov]
1179
1180 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1181 in SSL structures. New SSL ctrl to set maximum send fragment size.
1182 Save memory by seeting the I/O buffer sizes dynamically instead of
1183 using the maximum available value.
1184 [Steve Henson]
1185
1186 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1187 in addition to the text details.
1188 [Bodo Moeller]
1189
1190 *) Very, very preliminary EXPERIMENTAL support for printing of general
1191 ASN1 structures. This currently produces rather ugly output and doesn't
1192 handle several customised structures at all.
1193 [Steve Henson]
1194
1195 *) Integrated support for PVK file format and some related formats such
1196 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1197 these in the 'rsa' and 'dsa' utilities.
1198 [Steve Henson]
1199
1200 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1201 [Steve Henson]
1202
1203 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1204 place for the (very old) "NETSCAPE" format certificates which are now
1205 handled using new ASN1 code equivalents.
1206 [Steve Henson]
1207
1208 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1209 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1210 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1211 [Nils Larsch]
1212
1213 *) Modify CRL distribution points extension code to print out previously
1214 unsupported fields. Enhance extension setting code to allow setting of
1215 all fields.
1216 [Steve Henson]
1217
1218 *) Add print and set support for Issuing Distribution Point CRL extension.
1219 [Steve Henson]
1220
1221 *) Change 'Configure' script to enable Camellia by default.
1222 [NTT]
1223
1224 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1225
1226 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1227 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1228
1229 *) Fix bug in string printing code: if *any* escaping is enabled we must
1230 escape the escape character (backslash) or the resulting string is
1231 ambiguous.
1232 [Steve Henson]
1233
1234 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1235
1236 *) Disable code workaround for ancient and obsolete Netscape browsers
1237 and servers: an attacker can use it in a ciphersuite downgrade attack.
1238 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1239 [Steve Henson]
1240
1241 *) Fixed J-PAKE implementation error, originally discovered by
1242 Sebastien Martini, further info and confirmation from Stefan
1243 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1244 [Ben Laurie]
1245
1246 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1247
1248 *) Fix extension code to avoid race conditions which can result in a buffer
1249 overrun vulnerability: resumed sessions must not be modified as they can
1250 be shared by multiple threads. CVE-2010-3864
1251 [Steve Henson]
1252
1253 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1254 [Steve Henson]
1255
1256 *) Don't reencode certificate when calculating signature: cache and use
1257 the original encoding instead. This makes signature verification of
1258 some broken encodings work correctly.
1259 [Steve Henson]
1260
1261 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1262 is also one of the inputs.
1263 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1264
1265 *) Don't repeatedly append PBE algorithms to table if they already exist.
1266 Sort table on each new add. This effectively makes the table read only
1267 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1268 etc are non-op.
1269 [Steve Henson]
1270
1271 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1272
1273 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1274 OpenSSL 1.0.0.]
1275
1276 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1277 access or freeing data twice (CVE-2010-0742)
1278 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1279
1280 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1281 common in certificates and some applications which only call
1282 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1283 [Steve Henson]
1284
1285 *) VMS fixes:
1286 Reduce copying into .apps and .test in makevms.com
1287 Don't try to use blank CA certificate in CA.com
1288 Allow use of C files from original directories in maketests.com
1289 [Steven M. Schweda" <sms@antinode.info>]
1290
1291 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1292
1293 *) When rejecting SSL/TLS records due to an incorrect version number, never
1294 update s->server with a new major version number. As of
1295 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1296 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1297 the previous behavior could result in a read attempt at NULL when
1298 receiving specific incorrect SSL/TLS records once record payload
1299 protection is active. (CVE-2010-0740)
1300 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1301
1302 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1303 could be crashed if the relevant tables were not present (e.g. chrooted).
1304 [Tomas Hoger <thoger@redhat.com>]
1305
1306 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1307
1308 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1309 [Martin Olsson, Neel Mehta]
1310
1311 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1312 accommodate for stack sorting, always a write lock!).
1313 [Bodo Moeller]
1314
1315 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1316 excessive delays in the RAND_poll(): over a minute. As a workaround
1317 include a time check in the inner Heap32Next loop too.
1318 [Steve Henson]
1319
1320 *) The code that handled flushing of data in SSL/TLS originally used the
1321 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1322 the problem outlined in PR#1949. The fix suggested there however can
1323 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1324 of Apache). So instead simplify the code to flush unconditionally.
1325 This should be fine since flushing with no data to flush is a no op.
1326 [Steve Henson]
1327
1328 *) Handle TLS versions 2.0 and later properly and correctly use the
1329 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1330 off ancient servers have a habit of sticking around for a while...
1331 [Steve Henson]
1332
1333 *) Modify compression code so it frees up structures without using the
1334 ex_data callbacks. This works around a problem where some applications
1335 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1336 restarting) then use compression (e.g. SSL with compression) later.
1337 This results in significant per-connection memory leaks and
1338 has caused some security issues including CVE-2008-1678 and
1339 CVE-2009-4355.
1340 [Steve Henson]
1341
1342 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1343 change when encrypting or decrypting.
1344 [Bodo Moeller]
1345
1346 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1347 connect and renegotiate with servers which do not support RI.
1348 Until RI is more widely deployed this option is enabled by default.
1349 [Steve Henson]
1350
1351 *) Add "missing" ssl ctrls to clear options and mode.
1352 [Steve Henson]
1353
1354 *) If client attempts to renegotiate and doesn't support RI respond with
1355 a no_renegotiation alert as required by RFC5746. Some renegotiating
1356 TLS clients will continue a connection gracefully when they receive
1357 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1358 waiting for a server hello which it will never receive. Now we treat a
1359 received no_renegotiation alert as a fatal error. This is because
1360 applications requesting a renegotiation might well expect it to succeed
1361 and would have no code in place to handle the server denying it so the
1362 only safe thing to do is to terminate the connection.
1363 [Steve Henson]
1364
1365 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1366 peer supports secure renegotiation and 0 otherwise. Print out peer
1367 renegotiation support in s_client/s_server.
1368 [Steve Henson]
1369
1370 *) Replace the highly broken and deprecated SPKAC certification method with
1371 the updated NID creation version. This should correctly handle UTF8.
1372 [Steve Henson]
1373
1374 *) Implement RFC5746. Re-enable renegotiation but require the extension
1375 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1376 turns out to be a bad idea. It has been replaced by
1377 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1378 SSL_CTX_set_options(). This is really not recommended unless you
1379 know what you are doing.
1380 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1381
1382 *) Fixes to stateless session resumption handling. Use initial_ctx when
1383 issuing and attempting to decrypt tickets in case it has changed during
1384 servername handling. Use a non-zero length session ID when attempting
1385 stateless session resumption: this makes it possible to determine if
1386 a resumption has occurred immediately after receiving server hello
1387 (several places in OpenSSL subtly assume this) instead of later in
1388 the handshake.
1389 [Steve Henson]
1390
1391 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1392 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1393 fixes for a few places where the return code is not checked
1394 correctly.
1395 [Julia Lawall <julia@diku.dk>]
1396
1397 *) Add --strict-warnings option to Configure script to include devteam
1398 warnings in other configurations.
1399 [Steve Henson]
1400
1401 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1402 makes it possible to install openssl libraries in locations which
1403 have names other than "lib", for example "/usr/lib64" which some
1404 systems need.
1405 [Steve Henson, based on patch from Jeremy Utley]
1406
1407 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1408 X690 8.9.12 and can produce some misleading textual output of OIDs.
1409 [Steve Henson, reported by Dan Kaminsky]
1410
1411 *) Delete MD2 from algorithm tables. This follows the recommendation in
1412 several standards that it is not used in new applications due to
1413 several cryptographic weaknesses. For binary compatibility reasons
1414 the MD2 API is still compiled in by default.
1415 [Steve Henson]
1416
1417 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1418 and restored.
1419 [Steve Henson]
1420
1421 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1422 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1423 clash.
1424 [Guenter <lists@gknw.net>]
1425
1426 *) Fix the server certificate chain building code to use X509_verify_cert(),
1427 it used to have an ad-hoc builder which was unable to cope with anything
1428 other than a simple chain.
1429 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1430
1431 *) Don't check self signed certificate signatures in X509_verify_cert()
1432 by default (a flag can override this): it just wastes time without
1433 adding any security. As a useful side effect self signed root CAs
1434 with non-FIPS digests are now usable in FIPS mode.
1435 [Steve Henson]
1436
1437 *) In dtls1_process_out_of_seq_message() the check if the current message
1438 is already buffered was missing. For every new message was memory
1439 allocated, allowing an attacker to perform an denial of service attack
1440 with sending out of seq handshake messages until there is no memory
1441 left. Additionally every future messege was buffered, even if the
1442 sequence number made no sense and would be part of another handshake.
1443 So only messages with sequence numbers less than 10 in advance will be
1444 buffered. (CVE-2009-1378)
1445 [Robin Seggelmann, discovered by Daniel Mentz]
1446
1447 *) Records are buffered if they arrive with a future epoch to be
1448 processed after finishing the corresponding handshake. There is
1449 currently no limitation to this buffer allowing an attacker to perform
1450 a DOS attack with sending records with future epochs until there is no
1451 memory left. This patch adds the pqueue_size() function to detemine
1452 the size of a buffer and limits the record buffer to 100 entries.
1453 (CVE-2009-1377)
1454 [Robin Seggelmann, discovered by Daniel Mentz]
1455
1456 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1457 parent structure is freed. (CVE-2009-1379)
1458 [Daniel Mentz]
1459
1460 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1461 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1462
1463 *) Add 2.5.4.* OIDs
1464 [Ilya O. <vrghost@gmail.com>]
1465
1466 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1467
1468 *) Disable renegotiation completely - this fixes a severe security
1469 problem (CVE-2009-3555) at the cost of breaking all
1470 renegotiation. Renegotiation can be re-enabled by setting
1471 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1472 run-time. This is really not recommended unless you know what
1473 you're doing.
1474 [Ben Laurie]
1475
1476 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1477
1478 *) Don't set val to NULL when freeing up structures, it is freed up by
1479 underlying code. If sizeof(void *) > sizeof(long) this can result in
1480 zeroing past the valid field. (CVE-2009-0789)
1481 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1482
1483 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1484 checked correctly. This would allow some invalid signed attributes to
1485 appear to verify correctly. (CVE-2009-0591)
1486 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1487
1488 *) Reject UniversalString and BMPString types with invalid lengths. This
1489 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1490 a legal length. (CVE-2009-0590)
1491 [Steve Henson]
1492
1493 *) Set S/MIME signing as the default purpose rather than setting it
1494 unconditionally. This allows applications to override it at the store
1495 level.
1496 [Steve Henson]
1497
1498 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1499 to handle some structures.
1500 [Steve Henson]
1501
1502 *) Improve efficiency of mem_gets: don't search whole buffer each time
1503 for a '\n'
1504 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1505
1506 *) New -hex option for openssl rand.
1507 [Matthieu Herrb]
1508
1509 *) Print out UTF8String and NumericString when parsing ASN1.
1510 [Steve Henson]
1511
1512 *) Support NumericString type for name components.
1513 [Steve Henson]
1514
1515 *) Allow CC in the environment to override the automatically chosen
1516 compiler. Note that nothing is done to ensure flags work with the
1517 chosen compiler.
1518 [Ben Laurie]
1519
1520 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1521
1522 *) Properly check EVP_VerifyFinal() and similar return values
1523 (CVE-2008-5077).
1524 [Ben Laurie, Bodo Moeller, Google Security Team]
1525
1526 *) Enable TLS extensions by default.
1527 [Ben Laurie]
1528
1529 *) Allow the CHIL engine to be loaded, whether the application is
1530 multithreaded or not. (This does not release the developer from the
1531 obligation to set up the dynamic locking callbacks.)
1532 [Sander Temme <sander@temme.net>]
1533
1534 *) Use correct exit code if there is an error in dgst command.
1535 [Steve Henson; problem pointed out by Roland Dirlewanger]
1536
1537 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1538 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1539 [Bodo Moeller]
1540
1541 *) Add experimental JPAKE support, including demo authentication in
1542 s_client and s_server.
1543 [Ben Laurie]
1544
1545 *) Set the comparison function in v3_addr_canonize().
1546 [Rob Austein <sra@hactrn.net>]
1547
1548 *) Add support for XMPP STARTTLS in s_client.
1549 [Philip Paeps <philip@freebsd.org>]
1550
1551 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1552 to ensure that even with this option, only ciphersuites in the
1553 server's preference list will be accepted. (Note that the option
1554 applies only when resuming a session, so the earlier behavior was
1555 just about the algorithm choice for symmetric cryptography.)
1556 [Bodo Moeller]
1557
1558 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1559
1560 *) Fix NULL pointer dereference if a DTLS server received
1561 ChangeCipherSpec as first record (CVE-2009-1386).
1562 [PR #1679]
1563
1564 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1565 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1566 [Nagendra Modadugu]
1567
1568 *) The fix in 0.9.8c that supposedly got rid of unsafe
1569 double-checked locking was incomplete for RSA blinding,
1570 addressing just one layer of what turns out to have been
1571 doubly unsafe triple-checked locking.
1572
1573 So now fix this for real by retiring the MONT_HELPER macro
1574 in crypto/rsa/rsa_eay.c.
1575
1576 [Bodo Moeller; problem pointed out by Marius Schilder]
1577
1578 *) Various precautionary measures:
1579
1580 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1581
1582 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1583 (NB: This would require knowledge of the secret session ticket key
1584 to exploit, in which case you'd be SOL either way.)
1585
1586 - Change bn_nist.c so that it will properly handle input BIGNUMs
1587 outside the expected range.
1588
1589 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1590 builds.
1591
1592 [Neel Mehta, Bodo Moeller]
1593
1594 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1595 the load fails. Useful for distros.
1596 [Ben Laurie and the FreeBSD team]
1597
1598 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1599 [Steve Henson]
1600
1601 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1602 [Huang Ying]
1603
1604 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1605
1606 This work was sponsored by Logica.
1607 [Steve Henson]
1608
1609 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1610 keystores. Support for SSL/TLS client authentication too.
1611 Not compiled unless enable-capieng specified to Configure.
1612
1613 This work was sponsored by Logica.
1614 [Steve Henson]
1615
1616 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1617 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1618 attribute creation routines such as certifcate requests and PKCS#12
1619 files.
1620 [Steve Henson]
1621
1622 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1623
1624 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1625 handshake which could lead to a cilent crash as found using the
1626 Codenomicon TLS test suite (CVE-2008-1672)
1627 [Steve Henson, Mark Cox]
1628
1629 *) Fix double free in TLS server name extensions which could lead to
1630 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1631 [Joe Orton]
1632
1633 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1634
1635 Clear the error queue to ensure that error entries left from
1636 older function calls do not interfere with the correct operation.
1637 [Lutz Jaenicke, Erik de Castro Lopo]
1638
1639 *) Remove root CA certificates of commercial CAs:
1640
1641 The OpenSSL project does not recommend any specific CA and does not
1642 have any policy with respect to including or excluding any CA.
1643 Therefore it does not make any sense to ship an arbitrary selection
1644 of root CA certificates with the OpenSSL software.
1645 [Lutz Jaenicke]
1646
1647 *) RSA OAEP patches to fix two separate invalid memory reads.
1648 The first one involves inputs when 'lzero' is greater than
1649 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1650 before the beginning of from). The second one involves inputs where
1651 the 'db' section contains nothing but zeroes (there is a one-byte
1652 invalid read after the end of 'db').
1653 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1654
1655 *) Partial backport from 0.9.9-dev:
1656
1657 Introduce bn_mul_mont (dedicated Montgomery multiplication
1658 procedure) as a candidate for BIGNUM assembler implementation.
1659 While 0.9.9-dev uses assembler for various architectures, only
1660 x86_64 is available by default here in the 0.9.8 branch, and
1661 32-bit x86 is available through a compile-time setting.
1662
1663 To try the 32-bit x86 assembler implementation, use Configure
1664 option "enable-montasm" (which exists only for this backport).
1665
1666 As "enable-montasm" for 32-bit x86 disclaims code stability
1667 anyway, in this constellation we activate additional code
1668 backported from 0.9.9-dev for further performance improvements,
1669 namely BN_from_montgomery_word. (To enable this otherwise,
1670 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1671
1672 [Andy Polyakov (backport partially by Bodo Moeller)]
1673
1674 *) Add TLS session ticket callback. This allows an application to set
1675 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1676 values. This is useful for key rollover for example where several key
1677 sets may exist with different names.
1678 [Steve Henson]
1679
1680 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1681 This was broken until now in 0.9.8 releases, such that the only way
1682 a registered ENGINE could be used (assuming it initialises
1683 successfully on the host) was to explicitly set it as the default
1684 for the relevant algorithms. This is in contradiction with 0.9.7
1685 behaviour and the documentation. With this fix, when an ENGINE is
1686 registered into a given algorithm's table of implementations, the
1687 'uptodate' flag is reset so that auto-discovery will be used next
1688 time a new context for that algorithm attempts to select an
1689 implementation.
1690 [Ian Lister (tweaked by Geoff Thorpe)]
1691
1692 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1693 implemention in the following ways:
1694
1695 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1696 hard coded.
1697
1698 Lack of BER streaming support means one pass streaming processing is
1699 only supported if data is detached: setting the streaming flag is
1700 ignored for embedded content.
1701
1702 CMS support is disabled by default and must be explicitly enabled
1703 with the enable-cms configuration option.
1704 [Steve Henson]
1705
1706 *) Update the GMP engine glue to do direct copies between BIGNUM and
1707 mpz_t when openssl and GMP use the same limb size. Otherwise the
1708 existing "conversion via a text string export" trick is still used.
1709 [Paul Sheer <paulsheer@gmail.com>]
1710
1711 *) Zlib compression BIO. This is a filter BIO which compressed and
1712 uncompresses any data passed through it.
1713 [Steve Henson]
1714
1715 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1716 RFC3394 compatible AES key wrapping.
1717 [Steve Henson]
1718
1719 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1720 sets string data without copying. X509_ALGOR_set0() and
1721 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1722 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1723 from an X509_ATTRIBUTE structure optionally checking it occurs only
1724 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1725 data.
1726 [Steve Henson]
1727
1728 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1729 to get the expected BN_FLG_CONSTTIME behavior.
1730 [Bodo Moeller (Google)]
1731
1732 *) Netware support:
1733
1734 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1735 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1736 - added some more tests to do_tests.pl
1737 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1738 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1739 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1740 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1741 - various changes to netware.pl to enable gcc-cross builds on Win32
1742 platform
1743 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1744 - various changes to fix missing prototype warnings
1745 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1746 - added AES, WHIRLPOOL and CPUID assembler code to build files
1747 - added missing AES assembler make rules to mk1mf.pl
1748 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1749 [Guenter Knauf <eflash@gmx.net>]
1750
1751 *) Implement certificate status request TLS extension defined in RFC3546.
1752 A client can set the appropriate parameters and receive the encoded
1753 OCSP response via a callback. A server can query the supplied parameters
1754 and set the encoded OCSP response in the callback. Add simplified examples
1755 to s_client and s_server.
1756 [Steve Henson]
1757
1758 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1759
1760 *) Fix various bugs:
1761 + Binary incompatibility of ssl_ctx_st structure
1762 + DTLS interoperation with non-compliant servers
1763 + Don't call get_session_cb() without proposed session
1764 + Fix ia64 assembler code
1765 [Andy Polyakov, Steve Henson]
1766
1767 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1768
1769 *) DTLS Handshake overhaul. There were longstanding issues with
1770 OpenSSL DTLS implementation, which were making it impossible for
1771 RFC 4347 compliant client to communicate with OpenSSL server.
1772 Unfortunately just fixing these incompatibilities would "cut off"
1773 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1774 server keeps tolerating non RFC compliant syntax. The opposite is
1775 not true, 0.9.8f client can not communicate with earlier server.
1776 This update even addresses CVE-2007-4995.
1777 [Andy Polyakov]
1778
1779 *) Changes to avoid need for function casts in OpenSSL: some compilers
1780 (gcc 4.2 and later) reject their use.
1781 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1782 Steve Henson]
1783
1784 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1785 RFC4507bis. The encrypted ticket format is an encrypted encoded
1786 SSL_SESSION structure, that way new session features are automatically
1787 supported.
1788
1789 If a client application caches session in an SSL_SESSION structure
1790 support is transparent because tickets are now stored in the encoded
1791 SSL_SESSION.
1792
1793 The SSL_CTX structure automatically generates keys for ticket
1794 protection in servers so again support should be possible
1795 with no application modification.
1796
1797 If a client or server wishes to disable RFC4507 support then the option
1798 SSL_OP_NO_TICKET can be set.
1799
1800 Add a TLS extension debugging callback to allow the contents of any client
1801 or server extensions to be examined.
1802
1803 This work was sponsored by Google.
1804 [Steve Henson]
1805
1806 *) Add initial support for TLS extensions, specifically for the server_name
1807 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1808 have new members for a host name. The SSL data structure has an
1809 additional member SSL_CTX *initial_ctx so that new sessions can be
1810 stored in that context to allow for session resumption, even after the
1811 SSL has been switched to a new SSL_CTX in reaction to a client's
1812 server_name extension.
1813
1814 New functions (subject to change):
1815
1816 SSL_get_servername()
1817 SSL_get_servername_type()
1818 SSL_set_SSL_CTX()
1819
1820 New CTRL codes and macros (subject to change):
1821
1822 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1823 - SSL_CTX_set_tlsext_servername_callback()
1824 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1825 - SSL_CTX_set_tlsext_servername_arg()
1826 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1827
1828 openssl s_client has a new '-servername ...' option.
1829
1830 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1831 '-key2 ...', '-servername_fatal' (subject to change). This allows
1832 testing the HostName extension for a specific single host name ('-cert'
1833 and '-key' remain fallbacks for handshakes without HostName
1834 negotiation). If the unrecogninzed_name alert has to be sent, this by
1835 default is a warning; it becomes fatal with the '-servername_fatal'
1836 option.
1837
1838 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1839
1840 *) Add AES and SSE2 assembly language support to VC++ build.
1841 [Steve Henson]
1842
1843 *) Mitigate attack on final subtraction in Montgomery reduction.
1844 [Andy Polyakov]
1845
1846 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1847 (which previously caused an internal error).
1848 [Bodo Moeller]
1849
1850 *) Squeeze another 10% out of IGE mode when in != out.
1851 [Ben Laurie]
1852
1853 *) AES IGE mode speedup.
1854 [Dean Gaudet (Google)]
1855
1856 *) Add the Korean symmetric 128-bit cipher SEED (see
1857 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1858 add SEED ciphersuites from RFC 4162:
1859
1860 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1861 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1862 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1863 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1864
1865 To minimize changes between patchlevels in the OpenSSL 0.9.8
1866 series, SEED remains excluded from compilation unless OpenSSL
1867 is configured with 'enable-seed'.
1868 [KISA, Bodo Moeller]
1869
1870 *) Mitigate branch prediction attacks, which can be practical if a
1871 single processor is shared, allowing a spy process to extract
1872 information. For detailed background information, see
1873 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1874 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1875 and Necessary Software Countermeasures"). The core of the change
1876 are new versions BN_div_no_branch() and
1877 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1878 respectively, which are slower, but avoid the security-relevant
1879 conditional branches. These are automatically called by BN_div()
1880 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1881 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1882 remove a conditional branch.
1883
1884 BN_FLG_CONSTTIME is the new name for the previous
1885 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1886 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1887 in the exponent causes BN_mod_exp_mont() to use the alternative
1888 implementation in BN_mod_exp_mont_consttime().) The old name
1889 remains as a deprecated alias.
1890
1891 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1892 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1893 constant-time implementations for more than just exponentiation.
1894 Here too the old name is kept as a deprecated alias.
1895
1896 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1897 the BN_BLINDING structure gets an independent copy of the
1898 modulus. This means that the previous "BIGNUM *m" argument to
1899 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1900 essentially becomes "const BIGNUM *m", although we can't actually
1901 change this in the header file before 0.9.9. It allows
1902 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1903 enable BN_FLG_CONSTTIME.
1904
1905 [Matthew D Wood (Intel Corp)]
1906
1907 *) In the SSL/TLS server implementation, be strict about session ID
1908 context matching (which matters if an application uses a single
1909 external cache for different purposes). Previously,
1910 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1911 set. This did ensure strict client verification, but meant that,
1912 with applications using a single external cache for quite
1913 different requirements, clients could circumvent ciphersuite
1914 restrictions for a given session ID context by starting a session
1915 in a different context.
1916 [Bodo Moeller]
1917
1918 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1919 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1920 authentication-only ciphersuites.
1921 [Bodo Moeller]
1922
1923 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1924 not complete and could lead to a possible single byte overflow
1925 (CVE-2007-5135) [Ben Laurie]
1926
1927 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1928
1929 *) Since AES128 and AES256 (and similarly Camellia128 and
1930 Camellia256) share a single mask bit in the logic of
1931 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1932 kludge to work properly if AES128 is available and AES256 isn't
1933 (or if Camellia128 is available and Camellia256 isn't).
1934 [Victor Duchovni]
1935
1936 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1937 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1938 When a point or a seed is encoded in a BIT STRING, we need to
1939 prevent the removal of trailing zero bits to get the proper DER
1940 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1941 of a NamedBitList, for which trailing 0 bits need to be removed.)
1942 [Bodo Moeller]
1943
1944 *) Have SSL/TLS server implementation tolerate "mismatched" record
1945 protocol version while receiving ClientHello even if the
1946 ClientHello is fragmented. (The server can't insist on the
1947 particular protocol version it has chosen before the ServerHello
1948 message has informed the client about his choice.)
1949 [Bodo Moeller]
1950
1951 *) Add RFC 3779 support.
1952 [Rob Austein for ARIN, Ben Laurie]
1953
1954 *) Load error codes if they are not already present instead of using a
1955 static variable. This allows them to be cleanly unloaded and reloaded.
1956 Improve header file function name parsing.
1957 [Steve Henson]
1958
1959 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1960 or CAPABILITY handshake as required by RFCs.
1961 [Goetz Babin-Ebell]
1962
1963 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1964
1965 *) Introduce limits to prevent malicious keys being able to
1966 cause a denial of service. (CVE-2006-2940)
1967 [Steve Henson, Bodo Moeller]
1968
1969 *) Fix ASN.1 parsing of certain invalid structures that can result
1970 in a denial of service. (CVE-2006-2937) [Steve Henson]
1971
1972 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1973 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1974
1975 *) Fix SSL client code which could crash if connecting to a
1976 malicious SSLv2 server. (CVE-2006-4343)
1977 [Tavis Ormandy and Will Drewry, Google Security Team]
1978
1979 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1980 match only those. Before that, "AES256-SHA" would be interpreted
1981 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1982 the same strength classification in 0.9.7h) as we currently only
1983 have a single AES bit in the ciphersuite description bitmap.
1984 That change, however, also applied to ciphersuite strings such as
1985 "RC4-MD5" that intentionally matched multiple ciphersuites --
1986 namely, SSL 2.0 ciphersuites in addition to the more common ones
1987 from SSL 3.0/TLS 1.0.
1988
1989 So we change the selection algorithm again: Naming an explicit
1990 ciphersuite selects this one ciphersuite, and any other similar
1991 ciphersuite (same bitmap) from *other* protocol versions.
1992 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1993 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1994
1995 Since SSL 2.0 does not have any ciphersuites for which the
1996 128/256 bit distinction would be relevant, this works for now.
1997 The proper fix will be to use different bits for AES128 and
1998 AES256, which would have avoided the problems from the beginning;
1999 however, bits are scarce, so we can only do this in a new release
2000 (not just a patchlevel) when we can change the SSL_CIPHER
2001 definition to split the single 'unsigned long mask' bitmap into
2002 multiple values to extend the available space.
2003
2004 [Bodo Moeller]
2005
2006 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2007
2008 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2009 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2010
2011 *) Add AES IGE and biIGE modes.
2012 [Ben Laurie]
2013
2014 *) Change the Unix randomness entropy gathering to use poll() when
2015 possible instead of select(), since the latter has some
2016 undesirable limitations.
2017 [Darryl Miles via Richard Levitte and Bodo Moeller]
2018
2019 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2020 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2021 cannot be implicitly activated as part of, e.g., the "AES" alias.
2022 However, please upgrade to OpenSSL 0.9.9[-dev] for
2023 non-experimental use of the ECC ciphersuites to get TLS extension
2024 support, which is required for curve and point format negotiation
2025 to avoid potential handshake problems.
2026 [Bodo Moeller]
2027
2028 *) Disable rogue ciphersuites:
2029
2030 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2031 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2032 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2033
2034 The latter two were purportedly from
2035 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2036 appear there.
2037
2038 Also deactivate the remaining ciphersuites from
2039 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2040 unofficial, and the ID has long expired.
2041 [Bodo Moeller]
2042
2043 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2044 dual-core machines) and other potential thread-safety issues.
2045 [Bodo Moeller]
2046
2047 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2048 versions), which is now available for royalty-free use
2049 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2050 Also, add Camellia TLS ciphersuites from RFC 4132.
2051
2052 To minimize changes between patchlevels in the OpenSSL 0.9.8
2053 series, Camellia remains excluded from compilation unless OpenSSL
2054 is configured with 'enable-camellia'.
2055 [NTT]
2056
2057 *) Disable the padding bug check when compression is in use. The padding
2058 bug check assumes the first packet is of even length, this is not
2059 necessarily true if compresssion is enabled and can result in false
2060 positives causing handshake failure. The actual bug test is ancient
2061 code so it is hoped that implementations will either have fixed it by
2062 now or any which still have the bug do not support compression.
2063 [Steve Henson]
2064
2065 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2066
2067 *) When applying a cipher rule check to see if string match is an explicit
2068 cipher suite and only match that one cipher suite if it is.
2069 [Steve Henson]
2070
2071 *) Link in manifests for VC++ if needed.
2072 [Austin Ziegler <halostatue@gmail.com>]
2073
2074 *) Update support for ECC-based TLS ciphersuites according to
2075 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2076 TLS extensions, which are supported starting with the 0.9.9
2077 branch, not in the OpenSSL 0.9.8 branch).
2078 [Douglas Stebila]
2079
2080 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2081 opaque EVP_CIPHER_CTX handling.
2082 [Steve Henson]
2083
2084 *) Fixes and enhancements to zlib compression code. We now only use
2085 "zlib1.dll" and use the default __cdecl calling convention on Win32
2086 to conform with the standards mentioned here:
2087 http://www.zlib.net/DLL_FAQ.txt
2088 Static zlib linking now works on Windows and the new --with-zlib-include
2089 --with-zlib-lib options to Configure can be used to supply the location
2090 of the headers and library. Gracefully handle case where zlib library
2091 can't be loaded.
2092 [Steve Henson]
2093
2094 *) Several fixes and enhancements to the OID generation code. The old code
2095 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2096 handle numbers larger than ULONG_MAX, truncated printing and had a
2097 non standard OBJ_obj2txt() behaviour.
2098 [Steve Henson]
2099
2100 *) Add support for building of engines under engine/ as shared libraries
2101 under VC++ build system.
2102 [Steve Henson]
2103
2104 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2105 Hopefully, we will not see any false combination of paths any more.
2106 [Richard Levitte]
2107
2108 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2109
2110 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2111 (part of SSL_OP_ALL). This option used to disable the
2112 countermeasure against man-in-the-middle protocol-version
2113 rollback in the SSL 2.0 server implementation, which is a bad
2114 idea. (CVE-2005-2969)
2115
2116 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2117 for Information Security, National Institute of Advanced Industrial
2118 Science and Technology [AIST], Japan)]
2119
2120 *) Add two function to clear and return the verify parameter flags.
2121 [Steve Henson]
2122
2123 *) Keep cipherlists sorted in the source instead of sorting them at
2124 runtime, thus removing the need for a lock.
2125 [Nils Larsch]
2126
2127 *) Avoid some small subgroup attacks in Diffie-Hellman.
2128 [Nick Mathewson and Ben Laurie]
2129
2130 *) Add functions for well-known primes.
2131 [Nick Mathewson]
2132
2133 *) Extended Windows CE support.
2134 [Satoshi Nakamura and Andy Polyakov]
2135
2136 *) Initialize SSL_METHOD structures at compile time instead of during
2137 runtime, thus removing the need for a lock.
2138 [Steve Henson]
2139
2140 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2141 attempting to decrypt each encrypted key in turn. Add support to
2142 smime utility.
2143 [Steve Henson]
2144
2145 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2146
2147 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2148 OpenSSL 0.9.8.]
2149
2150 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2151 [Richard Levitte]
2152
2153 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2154 key into the same file any more.
2155 [Richard Levitte]
2156
2157 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2158 [Andy Polyakov]
2159
2160 *) Add -utf8 command line and config file option to 'ca'.
2161 [Stefan <stf@udoma.org]
2162
2163 *) Removed the macro des_crypt(), as it seems to conflict with some
2164 libraries. Use DES_crypt().
2165 [Richard Levitte]
2166
2167 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2168 involves renaming the source and generated shared-libs for
2169 both. The engines will accept the corrected or legacy ids
2170 ('ncipher' and '4758_cca' respectively) when binding. NB,
2171 this only applies when building 'shared'.
2172 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2173
2174 *) Add attribute functions to EVP_PKEY structure. Modify
2175 PKCS12_create() to recognize a CSP name attribute and
2176 use it. Make -CSP option work again in pkcs12 utility.
2177 [Steve Henson]
2178
2179 *) Add new functionality to the bn blinding code:
2180 - automatic re-creation of the BN_BLINDING parameters after
2181 a fixed number of uses (currently 32)
2182 - add new function for parameter creation
2183 - introduce flags to control the update behaviour of the
2184 BN_BLINDING parameters
2185 - hide BN_BLINDING structure
2186 Add a second BN_BLINDING slot to the RSA structure to improve
2187 performance when a single RSA object is shared among several
2188 threads.
2189 [Nils Larsch]
2190
2191 *) Add support for DTLS.
2192 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2193
2194 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2195 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2196 [Walter Goulet]
2197
2198 *) Remove buggy and incompletet DH cert support from
2199 ssl/ssl_rsa.c and ssl/s3_both.c
2200 [Nils Larsch]
2201
2202 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2203 the apps/openssl applications.
2204 [Nils Larsch]
2205
2206 *) Compile clean with "-Wall -Wmissing-prototypes
2207 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2208 DEBUG_SAFESTACK must also be set.
2209 [Ben Laurie]
2210
2211 *) Change ./Configure so that certain algorithms can be disabled by default.
2212 The new counterpiece to "no-xxx" is "enable-xxx".
2213
2214 The patented RC5 and MDC2 algorithms will now be disabled unless
2215 "enable-rc5" and "enable-mdc2", respectively, are specified.
2216
2217 (IDEA remains enabled despite being patented. This is because IDEA
2218 is frequently required for interoperability, and there is no license
2219 fee for non-commercial use. As before, "no-idea" can be used to
2220 avoid this algorithm.)
2221
2222 [Bodo Moeller]
2223
2224 *) Add processing of proxy certificates (see RFC 3820). This work was
2225 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2226 EGEE (Enabling Grids for E-science in Europe).
2227 [Richard Levitte]
2228
2229 *) RC4 performance overhaul on modern architectures/implementations, such
2230 as Intel P4, IA-64 and AMD64.
2231 [Andy Polyakov]
2232
2233 *) New utility extract-section.pl. This can be used specify an alternative
2234 section number in a pod file instead of having to treat each file as
2235 a separate case in Makefile. This can be done by adding two lines to the
2236 pod file:
2237
2238 =for comment openssl_section:XXX
2239
2240 The blank line is mandatory.
2241
2242 [Steve Henson]
2243
2244 *) New arguments -certform, -keyform and -pass for s_client and s_server
2245 to allow alternative format key and certificate files and passphrase
2246 sources.
2247 [Steve Henson]
2248
2249 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2250 update associated structures and add various utility functions.
2251
2252 Add new policy related verify parameters, include policy checking in
2253 standard verify code. Enhance 'smime' application with extra parameters
2254 to support policy checking and print out.
2255 [Steve Henson]
2256
2257 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2258 Nehemiah processors. These extensions support AES encryption in hardware
2259 as well as RNG (though RNG support is currently disabled).
2260 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2261
2262 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2263 [Geoff Thorpe]
2264
2265 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2266 [Andy Polyakov and a number of other people]
2267
2268 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2269 implementation contributed by IBM.
2270 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2271
2272 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2273 exponent rather than 'unsigned long'. There is a corresponding change to
2274 the new 'rsa_keygen' element of the RSA_METHOD structure.
2275 [Jelte Jansen, Geoff Thorpe]
2276
2277 *) Functionality for creating the initial serial number file is now
2278 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2279
2280 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2281 number file to 1, which is bound to cause problems. To avoid
2282 the problems while respecting compatibility between different 0.9.7
2283 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2284 CA.pl for serial number initialization. With the new release 0.9.8,
2285 we can fix the problem directly in the 'ca' utility.)
2286 [Steve Henson]
2287
2288 *) Reduced header interdepencies by declaring more opaque objects in
2289 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2290 give fewer recursive includes, which could break lazy source code - so
2291 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2292 developers should define this symbol when building and using openssl to
2293 ensure they track the recommended behaviour, interfaces, [etc], but
2294 backwards-compatible behaviour prevails when this isn't defined.
2295 [Geoff Thorpe]
2296
2297 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2298 [Steve Henson]
2299
2300 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2301 This will generate a random key of the appropriate length based on the
2302 cipher context. The EVP_CIPHER can provide its own random key generation
2303 routine to support keys of a specific form. This is used in the des and
2304 3des routines to generate a key of the correct parity. Update S/MIME
2305 code to use new functions and hence generate correct parity DES keys.
2306 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2307 valid (weak or incorrect parity).
2308 [Steve Henson]
2309
2310 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2311 as looking them up. This is useful when the verified structure may contain
2312 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2313 present unless the new PKCS7_NO_CRL flag is asserted.
2314 [Steve Henson]
2315
2316 *) Extend ASN1 oid configuration module. It now additionally accepts the
2317 syntax:
2318
2319 shortName = some long name, 1.2.3.4
2320 [Steve Henson]
2321
2322 *) Reimplemented the BN_CTX implementation. There is now no more static
2323 limitation on the number of variables it can handle nor the depth of the
2324 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2325 information can now expand as required, and rather than having a single
2326 static array of bignums, BN_CTX now uses a linked-list of such arrays
2327 allowing it to expand on demand whilst maintaining the usefulness of
2328 BN_CTX's "bundling".
2329 [Geoff Thorpe]
2330
2331 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2332 to allow all RSA operations to function using a single BN_CTX.
2333 [Geoff Thorpe]
2334
2335 *) Preliminary support for certificate policy evaluation and checking. This
2336 is initially intended to pass the tests outlined in "Conformance Testing
2337 of Relying Party Client Certificate Path Processing Logic" v1.07.
2338 [Steve Henson]
2339
2340 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2341 remained unused and not that useful. A variety of other little bignum
2342 tweaks and fixes have also been made continuing on from the audit (see
2343 below).
2344 [Geoff Thorpe]
2345
2346 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2347 associated ASN1, EVP and SSL functions and old ASN1 macros.
2348 [Richard Levitte]
2349
2350 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2351 and this should never fail. So the return value from the use of
2352 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2353 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2354 [Geoff Thorpe]
2355
2356 *) BN_CTX_get() should return zero-valued bignums, providing the same
2357 initialised value as BN_new().
2358 [Geoff Thorpe, suggested by Ulf Möller]
2359
2360 *) Support for inhibitAnyPolicy certificate extension.
2361 [Steve Henson]
2362
2363 *) An audit of the BIGNUM code is underway, for which debugging code is
2364 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2365 is considered valid when processing BIGNUMs, and causes execution to
2366 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2367 further steps are taken to deliberately pollute unused data in BIGNUM
2368 structures to try and expose faulty code further on. For now, openssl will
2369 (in its default mode of operation) continue to tolerate the inconsistent
2370 forms that it has tolerated in the past, but authors and packagers should
2371 consider trying openssl and their own applications when compiled with
2372 these debugging symbols defined. It will help highlight potential bugs in
2373 their own code, and will improve the test coverage for OpenSSL itself. At
2374 some point, these tighter rules will become openssl's default to improve
2375 maintainability, though the assert()s and other overheads will remain only
2376 in debugging configurations. See bn.h for more details.
2377 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2378
2379 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2380 that can only be obtained through BN_CTX_new() (which implicitly
2381 initialises it). The presence of this function only made it possible
2382 to overwrite an existing structure (and cause memory leaks).
2383 [Geoff Thorpe]
2384
2385 *) Because of the callback-based approach for implementing LHASH as a
2386 template type, lh_insert() adds opaque objects to hash-tables and
2387 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2388 to clean up those corresponding objects before destroying the hash table
2389 (and losing the object pointers). So some over-zealous constifications in
2390 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2391 objects as "const" and the lh_doall[_arg] callback wrappers are not
2392 prototyped to have "const" restrictions on the object pointers they are
2393 given (and so aren't required to cast them away any more).
2394 [Geoff Thorpe]
2395
2396 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2397 (speed) prefers to use its own implementation. The two implementations
2398 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2399 its object type properly exposed (MS_TM) instead of casting to/from "char
2400 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2401 aren't necessarily the greatest nomenclatures - but this is what was used
2402 internally to the implementation so I've used that for now.
2403 [Geoff Thorpe]
2404
2405 *) Ensure that deprecated functions do not get compiled when
2406 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2407 the self-tests were still using deprecated key-generation functions so
2408 these have been updated also.
2409 [Geoff Thorpe]
2410
2411 *) Reorganise PKCS#7 code to separate the digest location functionality
2412 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2413 New function PKCS7_set_digest() to set the digest type for PKCS#7
2414 digestedData type. Add additional code to correctly generate the
2415 digestedData type and add support for this type in PKCS7 initialization
2416 functions.
2417 [Steve Henson]
2418
2419 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2420 structure of type "other".
2421 [Steve Henson]
2422
2423 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2424 sure the loop does correctly stop and breaking ("division by zero")
2425 modulus operations are not performed. The (pre-generated) prime
2426 table crypto/bn/bn_prime.h was already correct, but it could not be
2427 re-generated on some platforms because of the "division by zero"
2428 situation in the script.
2429 [Ralf S. Engelschall]
2430
2431 *) Update support for ECC-based TLS ciphersuites according to
2432 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2433 SHA-1 now is only used for "small" curves (where the
2434 representation of a field element takes up to 24 bytes); for
2435 larger curves, the field element resulting from ECDH is directly
2436 used as premaster secret.
2437 [Douglas Stebila (Sun Microsystems Laboratories)]
2438
2439 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2440 curve secp160r1 to the tests.
2441 [Douglas Stebila (Sun Microsystems Laboratories)]
2442
2443 *) Add the possibility to load symbols globally with DSO.
2444 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2445
2446 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2447 control of the error stack.
2448 [Richard Levitte]
2449
2450 *) Add support for STORE in ENGINE.
2451 [Richard Levitte]
2452
2453 *) Add the STORE type. The intention is to provide a common interface
2454 to certificate and key stores, be they simple file-based stores, or
2455 HSM-type store, or LDAP stores, or...
2456 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2457 [Richard Levitte]
2458
2459 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2460 pass a list of arguments to any function as well as provide a way
2461 for a function to pass data back to the caller.
2462 [Richard Levitte]
2463
2464 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2465 works like BUF_strdup() but can be used to duplicate a portion of
2466 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2467 a memory area.
2468 [Richard Levitte]
2469
2470 *) Add the function sk_find_ex() which works like sk_find(), but will
2471 return an index to an element even if an exact match couldn't be
2472 found. The index is guaranteed to point at the element where the
2473 searched-for key would be inserted to preserve sorting order.
2474 [Richard Levitte]
2475
2476 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2477 takes an extra flags argument for optional functionality. Currently,
2478 the following flags are defined:
2479
2480 OBJ_BSEARCH_VALUE_ON_NOMATCH
2481 This one gets OBJ_bsearch_ex() to return a pointer to the first
2482 element where the comparing function returns a negative or zero
2483 number.
2484
2485 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2486 This one gets OBJ_bsearch_ex() to return a pointer to the first
2487 element where the comparing function returns zero. This is useful
2488 if there are more than one element where the comparing function
2489 returns zero.
2490 [Richard Levitte]
2491
2492 *) Make it possible to create self-signed certificates with 'openssl ca'
2493 in such a way that the self-signed certificate becomes part of the
2494 CA database and uses the same mechanisms for serial number generation
2495 as all other certificate signing. The new flag '-selfsign' enables
2496 this functionality. Adapt CA.sh and CA.pl.in.
2497 [Richard Levitte]
2498
2499 *) Add functionality to check the public key of a certificate request
2500 against a given private. This is useful to check that a certificate
2501 request can be signed by that key (self-signing).
2502 [Richard Levitte]
2503
2504 *) Make it possible to have multiple active certificates with the same
2505 subject in the CA index file. This is done only if the keyword
2506 'unique_subject' is set to 'no' in the main CA section (default
2507 if 'CA_default') of the configuration file. The value is saved
2508 with the database itself in a separate index attribute file,
2509 named like the index file with '.attr' appended to the name.
2510 [Richard Levitte]
2511
2512 *) Generate muti valued AVAs using '+' notation in config files for
2513 req and dirName.
2514 [Steve Henson]
2515
2516 *) Support for nameConstraints certificate extension.
2517 [Steve Henson]
2518
2519 *) Support for policyConstraints certificate extension.
2520 [Steve Henson]
2521
2522 *) Support for policyMappings certificate extension.
2523 [Steve Henson]
2524
2525 *) Make sure the default DSA_METHOD implementation only uses its
2526 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2527 and change its own handlers to be NULL so as to remove unnecessary
2528 indirection. This lets alternative implementations fallback to the
2529 default implementation more easily.
2530 [Geoff Thorpe]
2531
2532 *) Support for directoryName in GeneralName related extensions
2533 in config files.
2534 [Steve Henson]
2535
2536 *) Make it possible to link applications using Makefile.shared.
2537 Make that possible even when linking against static libraries!
2538 [Richard Levitte]
2539
2540 *) Support for single pass processing for S/MIME signing. This now
2541 means that S/MIME signing can be done from a pipe, in addition
2542 cleartext signing (multipart/signed type) is effectively streaming
2543 and the signed data does not need to be all held in memory.
2544
2545 This is done with a new flag PKCS7_STREAM. When this flag is set
2546 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2547 is done after the data is output (and digests calculated) in
2548 SMIME_write_PKCS7().
2549 [Steve Henson]
2550
2551 *) Add full support for -rpath/-R, both in shared libraries and
2552 applications, at least on the platforms where it's known how
2553 to do it.
2554 [Richard Levitte]
2555
2556 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2557 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2558 will now compute a table of multiples of the generator that
2559 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2560 faster (notably in the case of a single point multiplication,
2561 scalar * generator).
2562 [Nils Larsch, Bodo Moeller]
2563
2564 *) IPv6 support for certificate extensions. The various extensions
2565 which use the IP:a.b.c.d can now take IPv6 addresses using the
2566 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2567 correctly.
2568 [Steve Henson]
2569
2570 *) Added an ENGINE that implements RSA by performing private key
2571 exponentiations with the GMP library. The conversions to and from
2572 GMP's mpz_t format aren't optimised nor are any montgomery forms
2573 cached, and on x86 it appears OpenSSL's own performance has caught up.
2574 However there are likely to be other architectures where GMP could
2575 provide a boost. This ENGINE is not built in by default, but it can be
2576 specified at Configure time and should be accompanied by the necessary
2577 linker additions, eg;
2578 ./config -DOPENSSL_USE_GMP -lgmp
2579 [Geoff Thorpe]
2580
2581 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2582 testing availability of engines with "-t" - the old behaviour is
2583 produced by increasing the feature's verbosity with "-tt".
2584 [Geoff Thorpe]
2585
2586 *) ECDSA routines: under certain error conditions uninitialized BN objects
2587 could be freed. Solution: make sure initialization is performed early
2588 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2589 via PR#459)
2590 [Lutz Jaenicke]
2591
2592 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2593 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2594 software implementations. For DSA and DH, parameter generation can
2595 also be overriden by providing the appropriate method callbacks.
2596 [Geoff Thorpe]
2597
2598 *) Change the "progress" mechanism used in key-generation and
2599 primality testing to functions that take a new BN_GENCB pointer in
2600 place of callback/argument pairs. The new API functions have "_ex"
2601 postfixes and the older functions are reimplemented as wrappers for
2602 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2603 declarations of the old functions to help (graceful) attempts to
2604 migrate to the new functions. Also, the new key-generation API
2605 functions operate on a caller-supplied key-structure and return
2606 success/failure rather than returning a key or NULL - this is to
2607 help make "keygen" another member function of RSA_METHOD etc.
2608
2609 Example for using the new callback interface:
2610
2611 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2612 void *my_arg = ...;
2613 BN_GENCB my_cb;
2614
2615 BN_GENCB_set(&my_cb, my_callback, my_arg);
2616
2617 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2618 /* For the meaning of a, b in calls to my_callback(), see the
2619 * documentation of the function that calls the callback.
2620 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2621 * my_callback should return 1 if it wants BN_is_prime_ex()
2622 * to continue, or 0 to stop.
2623 */
2624
2625 [Geoff Thorpe]
2626
2627 *) Change the ZLIB compression method to be stateful, and make it
2628 available to TLS with the number defined in
2629 draft-ietf-tls-compression-04.txt.
2630 [Richard Levitte]
2631
2632 *) Add the ASN.1 structures and functions for CertificatePair, which
2633 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2634
2635 CertificatePair ::= SEQUENCE {
2636 forward [0] Certificate OPTIONAL,
2637 reverse [1] Certificate OPTIONAL,
2638 -- at least one of the pair shall be present -- }
2639
2640 Also implement the PEM functions to read and write certificate
2641 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2642
2643 This needed to be defined, mostly for the sake of the LDAP
2644 attribute crossCertificatePair, but may prove useful elsewhere as
2645 well.
2646 [Richard Levitte]
2647
2648 *) Make it possible to inhibit symlinking of shared libraries in
2649 Makefile.shared, for Cygwin's sake.
2650 [Richard Levitte]
2651
2652 *) Extend the BIGNUM API by creating a function
2653 void BN_set_negative(BIGNUM *a, int neg);
2654 and a macro that behave like
2655 int BN_is_negative(const BIGNUM *a);
2656
2657 to avoid the need to access 'a->neg' directly in applications.
2658 [Nils Larsch]
2659
2660 *) Implement fast modular reduction for pseudo-Mersenne primes
2661 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2662 EC_GROUP_new_curve_GFp() will now automatically use this
2663 if applicable.
2664 [Nils Larsch <nla@trustcenter.de>]
2665
2666 *) Add new lock type (CRYPTO_LOCK_BN).
2667 [Bodo Moeller]
2668
2669 *) Change the ENGINE framework to automatically load engines
2670 dynamically from specific directories unless they could be
2671 found to already be built in or loaded. Move all the
2672 current engines except for the cryptodev one to a new
2673 directory engines/.
2674 The engines in engines/ are built as shared libraries if
2675 the "shared" options was given to ./Configure or ./config.
2676 Otherwise, they are inserted in libcrypto.a.
2677 /usr/local/ssl/engines is the default directory for dynamic
2678 engines, but that can be overriden at configure time through
2679 the usual use of --prefix and/or --openssldir, and at run
2680 time with the environment variable OPENSSL_ENGINES.
2681 [Geoff Thorpe and Richard Levitte]
2682
2683 *) Add Makefile.shared, a helper makefile to build shared
2684 libraries. Addapt Makefile.org.
2685 [Richard Levitte]
2686
2687 *) Add version info to Win32 DLLs.
2688 [Peter 'Luna' Runestig" <peter@runestig.com>]
2689
2690 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2691 can be added using this API to created arbitrary PKCS#12
2692 files while avoiding the low level API.
2693
2694 New options to PKCS12_create(), key or cert can be NULL and
2695 will then be omitted from the output file. The encryption
2696 algorithm NIDs can be set to -1 for no encryption, the mac
2697 iteration count can be set to 0 to omit the mac.
2698
2699 Enhance pkcs12 utility by making the -nokeys and -nocerts
2700 options work when creating a PKCS#12 file. New option -nomac
2701 to omit the mac, NONE can be set for an encryption algorithm.
2702 New code is modified to use the enhanced PKCS12_create()
2703 instead of the low level API.
2704 [Steve Henson]
2705
2706 *) Extend ASN1 encoder to support indefinite length constructed
2707 encoding. This can output sequences tags and octet strings in
2708 this form. Modify pk7_asn1.c to support indefinite length
2709 encoding. This is experimental and needs additional code to
2710 be useful, such as an ASN1 bio and some enhanced streaming
2711 PKCS#7 code.
2712
2713 Extend template encode functionality so that tagging is passed
2714 down to the template encoder.
2715 [Steve Henson]
2716
2717 *) Let 'openssl req' fail if an argument to '-newkey' is not
2718 recognized instead of using RSA as a default.
2719 [Bodo Moeller]
2720
2721 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2722 As these are not official, they are not included in "ALL";
2723 the "ECCdraft" ciphersuite group alias can be used to select them.
2724 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2725
2726 *) Add ECDH engine support.
2727 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2728
2729 *) Add ECDH in new directory crypto/ecdh/.
2730 [Douglas Stebila (Sun Microsystems Laboratories)]
2731
2732 *) Let BN_rand_range() abort with an error after 100 iterations
2733 without success (which indicates a broken PRNG).
2734 [Bodo Moeller]
2735
2736 *) Change BN_mod_sqrt() so that it verifies that the input value
2737 is really the square of the return value. (Previously,
2738 BN_mod_sqrt would show GIGO behaviour.)
2739 [Bodo Moeller]
2740
2741 *) Add named elliptic curves over binary fields from X9.62, SECG,
2742 and WAP/WTLS; add OIDs that were still missing.
2743
2744 [Sheueling Chang Shantz and Douglas Stebila
2745 (Sun Microsystems Laboratories)]
2746
2747 *) Extend the EC library for elliptic curves over binary fields
2748 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2749 New EC_METHOD:
2750
2751 EC_GF2m_simple_method
2752
2753 New API functions:
2754
2755 EC_GROUP_new_curve_GF2m
2756 EC_GROUP_set_curve_GF2m
2757 EC_GROUP_get_curve_GF2m
2758 EC_POINT_set_affine_coordinates_GF2m
2759 EC_POINT_get_affine_coordinates_GF2m
2760 EC_POINT_set_compressed_coordinates_GF2m
2761
2762 Point compression for binary fields is disabled by default for
2763 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2764 enable it).
2765
2766 As binary polynomials are represented as BIGNUMs, various members
2767 of the EC_GROUP and EC_POINT data structures can be shared
2768 between the implementations for prime fields and binary fields;
2769 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2770 are essentially identical to their ..._GFp counterparts.
2771 (For simplicity, the '..._GFp' prefix has been dropped from
2772 various internal method names.)
2773
2774 An internal 'field_div' method (similar to 'field_mul' and
2775 'field_sqr') has been added; this is used only for binary fields.
2776
2777 [Sheueling Chang Shantz and Douglas Stebila
2778 (Sun Microsystems Laboratories)]
2779
2780 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2781 through methods ('mul', 'precompute_mult').
2782
2783 The generic implementations (now internally called 'ec_wNAF_mul'
2784 and 'ec_wNAF_precomputed_mult') remain the default if these
2785 methods are undefined.
2786
2787 [Sheueling Chang Shantz and Douglas Stebila
2788 (Sun Microsystems Laboratories)]
2789
2790 *) New function EC_GROUP_get_degree, which is defined through
2791 EC_METHOD. For curves over prime fields, this returns the bit
2792 length of the modulus.
2793
2794 [Sheueling Chang Shantz and Douglas Stebila
2795 (Sun Microsystems Laboratories)]
2796
2797 *) New functions EC_GROUP_dup, EC_POINT_dup.
2798 (These simply call ..._new and ..._copy).
2799
2800 [Sheueling Chang Shantz and Douglas Stebila
2801 (Sun Microsystems Laboratories)]
2802
2803 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2804 Polynomials are represented as BIGNUMs (where the sign bit is not
2805 used) in the following functions [macros]:
2806
2807 BN_GF2m_add
2808 BN_GF2m_sub [= BN_GF2m_add]
2809 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2810 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2811 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2812 BN_GF2m_mod_inv
2813 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2814 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2815 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2816 BN_GF2m_cmp [= BN_ucmp]
2817
2818 (Note that only the 'mod' functions are actually for fields GF(2^m).
2819 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2820
2821 For some functions, an the irreducible polynomial defining a
2822 field can be given as an 'unsigned int[]' with strictly
2823 decreasing elements giving the indices of those bits that are set;
2824 i.e., p[] represents the polynomial
2825 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2826 where
2827 p[0] > p[1] > ... > p[k] = 0.
2828 This applies to the following functions:
2829
2830 BN_GF2m_mod_arr
2831 BN_GF2m_mod_mul_arr
2832 BN_GF2m_mod_sqr_arr
2833 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2834 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2835 BN_GF2m_mod_exp_arr
2836 BN_GF2m_mod_sqrt_arr
2837 BN_GF2m_mod_solve_quad_arr
2838 BN_GF2m_poly2arr
2839 BN_GF2m_arr2poly
2840
2841 Conversion can be performed by the following functions:
2842
2843 BN_GF2m_poly2arr
2844 BN_GF2m_arr2poly
2845
2846 bntest.c has additional tests for binary polynomial arithmetic.
2847
2848 Two implementations for BN_GF2m_mod_div() are available.
2849 The default algorithm simply uses BN_GF2m_mod_inv() and
2850 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2851 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2852 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2853
2854 [Sheueling Chang Shantz and Douglas Stebila
2855 (Sun Microsystems Laboratories)]
2856
2857 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2858 functionality is disabled at compile-time.
2859 [Douglas Stebila <douglas.stebila@sun.com>]
2860
2861 *) Change default behaviour of 'openssl asn1parse' so that more
2862 information is visible when viewing, e.g., a certificate:
2863
2864 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2865 mode the content of non-printable OCTET STRINGs is output in a
2866 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2867 avoid the appearance of a printable string.
2868 [Nils Larsch <nla@trustcenter.de>]
2869
2870 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2871 functions
2872 EC_GROUP_set_asn1_flag()
2873 EC_GROUP_get_asn1_flag()
2874 EC_GROUP_set_point_conversion_form()
2875 EC_GROUP_get_point_conversion_form()
2876 These control ASN1 encoding details:
2877 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2878 has been set to OPENSSL_EC_NAMED_CURVE.
2879 - Points are encoded in uncompressed form by default; options for
2880 asn1_for are as for point2oct, namely
2881 POINT_CONVERSION_COMPRESSED
2882 POINT_CONVERSION_UNCOMPRESSED
2883 POINT_CONVERSION_HYBRID
2884
2885 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2886 functions
2887 EC_GROUP_set_seed()
2888 EC_GROUP_get0_seed()
2889 EC_GROUP_get_seed_len()
2890 This is used only for ASN1 purposes (so far).
2891 [Nils Larsch <nla@trustcenter.de>]
2892
2893 *) Add 'field_type' member to EC_METHOD, which holds the NID
2894 of the appropriate field type OID. The new function
2895 EC_METHOD_get_field_type() returns this value.
2896 [Nils Larsch <nla@trustcenter.de>]
2897
2898 *) Add functions
2899 EC_POINT_point2bn()
2900 EC_POINT_bn2point()
2901 EC_POINT_point2hex()
2902 EC_POINT_hex2point()
2903 providing useful interfaces to EC_POINT_point2oct() and
2904 EC_POINT_oct2point().
2905 [Nils Larsch <nla@trustcenter.de>]
2906
2907 *) Change internals of the EC library so that the functions
2908 EC_GROUP_set_generator()
2909 EC_GROUP_get_generator()
2910 EC_GROUP_get_order()
2911 EC_GROUP_get_cofactor()
2912 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2913 to methods, which would lead to unnecessary code duplication when
2914 adding different types of curves.
2915 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2916
2917 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2918 arithmetic, and such that modified wNAFs are generated
2919 (which avoid length expansion in many cases).
2920 [Bodo Moeller]
2921
2922 *) Add a function EC_GROUP_check_discriminant() (defined via
2923 EC_METHOD) that verifies that the curve discriminant is non-zero.
2924
2925 Add a function EC_GROUP_check() that makes some sanity tests
2926 on a EC_GROUP, its generator and order. This includes
2927 EC_GROUP_check_discriminant().
2928 [Nils Larsch <nla@trustcenter.de>]
2929
2930 *) Add ECDSA in new directory crypto/ecdsa/.
2931
2932 Add applications 'openssl ecparam' and 'openssl ecdsa'
2933 (these are based on 'openssl dsaparam' and 'openssl dsa').
2934
2935 ECDSA support is also included in various other files across the
2936 library. Most notably,
2937 - 'openssl req' now has a '-newkey ecdsa:file' option;
2938 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2939 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2940 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2941 them suitable for ECDSA where domain parameters must be
2942 extracted before the specific public key;
2943 - ECDSA engine support has been added.
2944 [Nils Larsch <nla@trustcenter.de>]
2945
2946 *) Include some named elliptic curves, and add OIDs from X9.62,
2947 SECG, and WAP/WTLS. Each curve can be obtained from the new
2948 function
2949 EC_GROUP_new_by_curve_name(),
2950 and the list of available named curves can be obtained with
2951 EC_get_builtin_curves().
2952 Also add a 'curve_name' member to EC_GROUP objects, which can be
2953 accessed via
2954 EC_GROUP_set_curve_name()
2955 EC_GROUP_get_curve_name()
2956 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2957
2958 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2959 was actually never needed) and in BN_mul(). The removal in BN_mul()
2960 required a small change in bn_mul_part_recursive() and the addition
2961 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2962 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2963 bn_sub_words() and bn_add_words() except they take arrays with
2964 differing sizes.
2965 [Richard Levitte]
2966
2967 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2968
2969 *) Cleanse PEM buffers before freeing them since they may contain
2970 sensitive data.
2971 [Benjamin Bennett <ben@psc.edu>]
2972
2973 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2974 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2975 authentication-only ciphersuites.
2976 [Bodo Moeller]
2977
2978 *) Since AES128 and AES256 share a single mask bit in the logic of
2979 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2980 kludge to work properly if AES128 is available and AES256 isn't.
2981 [Victor Duchovni]
2982
2983 *) Expand security boundary to match 1.1.1 module.
2984 [Steve Henson]
2985
2986 *) Remove redundant features: hash file source, editing of test vectors
2987 modify fipsld to use external fips_premain.c signature.
2988 [Steve Henson]
2989
2990 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2991 run algorithm test programs.
2992 [Steve Henson]
2993
2994 *) Make algorithm test programs more tolerant of whitespace.
2995 [Steve Henson]
2996
2997 *) Have SSL/TLS server implementation tolerate "mismatched" record
2998 protocol version while receiving ClientHello even if the
2999 ClientHello is fragmented. (The server can't insist on the
3000 particular protocol version it has chosen before the ServerHello
3001 message has informed the client about his choice.)
3002 [Bodo Moeller]
3003
3004 *) Load error codes if they are not already present instead of using a
3005 static variable. This allows them to be cleanly unloaded and reloaded.
3006 [Steve Henson]
3007
3008 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3009
3010 *) Introduce limits to prevent malicious keys being able to
3011 cause a denial of service. (CVE-2006-2940)
3012 [Steve Henson, Bodo Moeller]
3013
3014 *) Fix ASN.1 parsing of certain invalid structures that can result
3015 in a denial of service. (CVE-2006-2937) [Steve Henson]
3016
3017 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3018 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3019
3020 *) Fix SSL client code which could crash if connecting to a
3021 malicious SSLv2 server. (CVE-2006-4343)
3022 [Tavis Ormandy and Will Drewry, Google Security Team]
3023
3024 *) Change ciphersuite string processing so that an explicit
3025 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3026 will no longer include "AES128-SHA"), and any other similar
3027 ciphersuite (same bitmap) from *other* protocol versions (so that
3028 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3029 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3030 changes from 0.9.8b and 0.9.8d.
3031 [Bodo Moeller]
3032
3033 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3034
3035 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3036 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3037
3038 *) Change the Unix randomness entropy gathering to use poll() when
3039 possible instead of select(), since the latter has some
3040 undesirable limitations.
3041 [Darryl Miles via Richard Levitte and Bodo Moeller]
3042
3043 *) Disable rogue ciphersuites:
3044
3045 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3046 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3047 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3048
3049 The latter two were purportedly from
3050 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3051 appear there.
3052
3053 Also deactive the remaining ciphersuites from
3054 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3055 unofficial, and the ID has long expired.
3056 [Bodo Moeller]
3057
3058 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3059 dual-core machines) and other potential thread-safety issues.
3060 [Bodo Moeller]
3061
3062 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3063
3064 *) Adapt fipsld and the build system to link against the validated FIPS
3065 module in FIPS mode.
3066 [Steve Henson]
3067
3068 *) Fixes for VC++ 2005 build under Windows.
3069 [Steve Henson]
3070
3071 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3072 from a Windows bash shell such as MSYS. It is autodetected from the
3073 "config" script when run from a VC++ environment. Modify standard VC++
3074 build to use fipscanister.o from the GNU make build.
3075 [Steve Henson]
3076
3077 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3078
3079 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3080 The value now differs depending on if you build for FIPS or not.
3081 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3082 safely run with a non-FIPSed libcrypto, as it may crash because of
3083 the difference induced by this change.
3084 [Andy Polyakov]
3085
3086 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3087
3088 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3089 (part of SSL_OP_ALL). This option used to disable the
3090 countermeasure against man-in-the-middle protocol-version
3091 rollback in the SSL 2.0 server implementation, which is a bad
3092 idea. (CVE-2005-2969)
3093
3094 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3095 for Information Security, National Institute of Advanced Industrial
3096 Science and Technology [AIST], Japan)]
3097
3098 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3099 mainly for FIPS compliance and not fully integrated at this stage.
3100 [Steve Henson]
3101
3102 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3103 the exponentiation using a fixed-length exponent. (Otherwise,
3104 the information leaked through timing could expose the secret key
3105 after many signatures; cf. Bleichenbacher's attack on DSA with
3106 biased k.)
3107 [Bodo Moeller]
3108
3109 *) Make a new fixed-window mod_exp implementation the default for
3110 RSA, DSA, and DH private-key operations so that the sequence of
3111 squares and multiplies and the memory access pattern are
3112 independent of the particular secret key. This will mitigate
3113 cache-timing and potential related attacks.
3114
3115 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3116 and this is automatically used by BN_mod_exp_mont() if the new flag
3117 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3118 will use this BN flag for private exponents unless the flag
3119 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3120 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3121
3122 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3123
3124 *) Change the client implementation for SSLv23_method() and
3125 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3126 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3127 (Previously, the SSL 2.0 backwards compatible Client Hello
3128 message format would be used even with SSL_OP_NO_SSLv2.)
3129 [Bodo Moeller]
3130
3131 *) Add support for smime-type MIME parameter in S/MIME messages which some
3132 clients need.
3133 [Steve Henson]
3134
3135 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3136 a threadsafe manner. Modify rsa code to use new function and add calls
3137 to dsa and dh code (which had race conditions before).
3138 [Steve Henson]
3139
3140 *) Include the fixed error library code in the C error file definitions
3141 instead of fixing them up at runtime. This keeps the error code
3142 structures constant.
3143 [Steve Henson]
3144
3145 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3146
3147 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3148 OpenSSL 0.9.8.]
3149
3150 *) Fixes for newer kerberos headers. NB: the casts are needed because
3151 the 'length' field is signed on one version and unsigned on another
3152 with no (?) obvious way to tell the difference, without these VC++
3153 complains. Also the "definition" of FAR (blank) is no longer included
3154 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3155 some needed definitions.
3156 [Steve Henson]
3157
3158 *) Undo Cygwin change.
3159 [Ulf Möller]
3160
3161 *) Added support for proxy certificates according to RFC 3820.
3162 Because they may be a security thread to unaware applications,
3163 they must be explicitely allowed in run-time. See
3164 docs/HOWTO/proxy_certificates.txt for further information.
3165 [Richard Levitte]
3166
3167 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3168
3169 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3170 server and client random values. Previously
3171 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3172 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3173
3174 This change has negligible security impact because:
3175
3176 1. Server and client random values still have 24 bytes of pseudo random
3177 data.
3178
3179 2. Server and client random values are sent in the clear in the initial
3180 handshake.
3181
3182 3. The master secret is derived using the premaster secret (48 bytes in
3183 size for static RSA ciphersuites) as well as client server and random
3184 values.
3185
3186 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3187 to our attention.
3188
3189 [Stephen Henson, reported by UK NISCC]
3190
3191 *) Use Windows randomness collection on Cygwin.
3192 [Ulf Möller]
3193
3194 *) Fix hang in EGD/PRNGD query when communication socket is closed
3195 prematurely by EGD/PRNGD.
3196 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3197
3198 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3199 [Steve Henson]
3200
3201 *) Back-port of selected performance improvements from development
3202 branch, as well as improved support for PowerPC platforms.
3203 [Andy Polyakov]
3204
3205 *) Add lots of checks for memory allocation failure, error codes to indicate
3206 failure and freeing up memory if a failure occurs.
3207 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3208
3209 *) Add new -passin argument to dgst.
3210 [Steve Henson]
3211
3212 *) Perform some character comparisons of different types in X509_NAME_cmp:
3213 this is needed for some certificates that reencode DNs into UTF8Strings
3214 (in violation of RFC3280) and can't or wont issue name rollover
3215 certificates.
3216 [Steve Henson]
3217
3218 *) Make an explicit check during certificate validation to see that
3219 the CA setting in each certificate on the chain is correct. As a
3220 side effect always do the following basic checks on extensions,
3221 not just when there's an associated purpose to the check:
3222
3223 - if there is an unhandled critical extension (unless the user
3224 has chosen to ignore this fault)
3225 - if the path length has been exceeded (if one is set at all)
3226 - that certain extensions fit the associated purpose (if one has
3227 been given)
3228 [Richard Levitte]
3229
3230 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3231
3232 *) Avoid a race condition when CRLs are checked in a multi threaded
3233 environment. This would happen due to the reordering of the revoked
3234 entries during signature checking and serial number lookup. Now the
3235 encoding is cached and the serial number sort performed under a lock.
3236 Add new STACK function sk_is_sorted().
3237 [Steve Henson]
3238
3239 *) Add Delta CRL to the extension code.
3240 [Steve Henson]
3241
3242 *) Various fixes to s3_pkt.c so alerts are sent properly.
3243 [David Holmes <d.holmes@f5.com>]
3244
3245 *) Reduce the chances of duplicate issuer name and serial numbers (in
3246 violation of RFC3280) using the OpenSSL certificate creation utilities.
3247 This is done by creating a random 64 bit value for the initial serial
3248 number when a serial number file is created or when a self signed
3249 certificate is created using 'openssl req -x509'. The initial serial
3250 number file is created using 'openssl x509 -next_serial' in CA.pl
3251 rather than being initialized to 1.
3252 [Steve Henson]
3253
3254 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3255
3256 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3257 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3258 [Joe Orton, Steve Henson]
3259
3260 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3261 (CVE-2004-0112)
3262 [Joe Orton, Steve Henson]
3263
3264 *) Make it possible to have multiple active certificates with the same
3265 subject in the CA index file. This is done only if the keyword
3266 'unique_subject' is set to 'no' in the main CA section (default
3267 if 'CA_default') of the configuration file. The value is saved
3268 with the database itself in a separate index attribute file,
3269 named like the index file with '.attr' appended to the name.
3270 [Richard Levitte]
3271
3272 *) X509 verify fixes. Disable broken certificate workarounds when
3273 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3274 keyUsage extension present. Don't accept CRLs with unhandled critical
3275 extensions: since verify currently doesn't process CRL extensions this
3276 rejects a CRL with *any* critical extensions. Add new verify error codes
3277 for these cases.
3278 [Steve Henson]
3279
3280 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3281 A clarification of RFC2560 will require the use of OCTET STRINGs and
3282 some implementations cannot handle the current raw format. Since OpenSSL
3283 copies and compares OCSP nonces as opaque blobs without any attempt at
3284 parsing them this should not create any compatibility issues.
3285 [Steve Henson]
3286
3287 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3288 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3289 this HMAC (and other) operations are several times slower than OpenSSL
3290 < 0.9.7.
3291 [Steve Henson]
3292
3293 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3294 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3295
3296 *) Use the correct content when signing type "other".
3297 [Steve Henson]
3298
3299 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3300
3301 *) Fix various bugs revealed by running the NISCC test suite:
3302
3303 Stop out of bounds reads in the ASN1 code when presented with
3304 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3305
3306 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3307
3308 If verify callback ignores invalid public key errors don't try to check
3309 certificate signature with the NULL public key.
3310
3311 [Steve Henson]
3312
3313 *) New -ignore_err option in ocsp application to stop the server
3314 exiting on the first error in a request.
3315 [Steve Henson]
3316
3317 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3318 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3319 specifications.
3320 [Steve Henson]
3321
3322 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3323 extra data after the compression methods not only for TLS 1.0
3324 but also for SSL 3.0 (as required by the specification).
3325 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3326
3327 *) Change X509_certificate_type() to mark the key as exported/exportable
3328 when it's 512 *bits* long, not 512 bytes.
3329 [Richard Levitte]
3330
3331 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3332 blocks during encryption.
3333 [Richard Levitte]
3334
3335 *) Various fixes to base64 BIO and non blocking I/O. On write
3336 flushes were not handled properly if the BIO retried. On read
3337 data was not being buffered properly and had various logic bugs.
3338 This also affects blocking I/O when the data being decoded is a
3339 certain size.
3340 [Steve Henson]
3341
3342 *) Various S/MIME bugfixes and compatibility changes:
3343 output correct application/pkcs7 MIME type if
3344 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3345 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3346 of files as .eml work). Correctly handle very long lines in MIME
3347 parser.
3348 [Steve Henson]
3349
3350 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3351
3352 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3353 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3354 a protocol version number mismatch like a decryption error
3355 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3356 [Bodo Moeller]
3357
3358 *) Turn on RSA blinding by default in the default implementation
3359 to avoid a timing attack. Applications that don't want it can call
3360 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3361 They would be ill-advised to do so in most cases.
3362 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3363
3364 *) Change RSA blinding code so that it works when the PRNG is not
3365 seeded (in this case, the secret RSA exponent is abused as
3366 an unpredictable seed -- if it is not unpredictable, there
3367 is no point in blinding anyway). Make RSA blinding thread-safe
3368 by remembering the creator's thread ID in rsa->blinding and
3369 having all other threads use local one-time blinding factors
3370 (this requires more computation than sharing rsa->blinding, but
3371 avoids excessive locking; and if an RSA object is not shared
3372 between threads, blinding will still be very fast).
3373 [Bodo Moeller]
3374
3375 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3376 ENGINE as defaults for all supported algorithms irrespective of
3377 the 'flags' parameter. 'flags' is now honoured, so applications
3378 should make sure they are passing it correctly.
3379 [Geoff Thorpe]
3380
3381 *) Target "mingw" now allows native Windows code to be generated in
3382 the Cygwin environment as well as with the MinGW compiler.
3383 [Ulf Moeller]
3384
3385 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3386
3387 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3388 via timing by performing a MAC computation even if incorrrect
3389 block cipher padding has been found. This is a countermeasure
3390 against active attacks where the attacker has to distinguish
3391 between bad padding and a MAC verification error. (CVE-2003-0078)
3392
3393 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3394 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3395 Martin Vuagnoux (EPFL, Ilion)]
3396
3397 *) Make the no-err option work as intended. The intention with no-err
3398 is not to have the whole error stack handling routines removed from
3399 libcrypto, it's only intended to remove all the function name and
3400 reason texts, thereby removing some of the footprint that may not
3401 be interesting if those errors aren't displayed anyway.
3402
3403 NOTE: it's still possible for any application or module to have it's
3404 own set of error texts inserted. The routines are there, just not
3405 used by default when no-err is given.
3406 [Richard Levitte]
3407
3408 *) Add support for FreeBSD on IA64.
3409 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3410
3411 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3412 Kerberos function mit_des_cbc_cksum(). Before this change,
3413 the value returned by DES_cbc_cksum() was like the one from
3414 mit_des_cbc_cksum(), except the bytes were swapped.
3415 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3416
3417 *) Allow an application to disable the automatic SSL chain building.
3418 Before this a rather primitive chain build was always performed in
3419 ssl3_output_cert_chain(): an application had no way to send the
3420 correct chain if the automatic operation produced an incorrect result.
3421
3422 Now the chain builder is disabled if either:
3423
3424 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3425
3426 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3427
3428 The reasoning behind this is that an application would not want the
3429 auto chain building to take place if extra chain certificates are
3430 present and it might also want a means of sending no additional
3431 certificates (for example the chain has two certificates and the
3432 root is omitted).
3433 [Steve Henson]
3434
3435 *) Add the possibility to build without the ENGINE framework.
3436 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3437
3438 *) Under Win32 gmtime() can return NULL: check return value in
3439 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3440 [Steve Henson]
3441
3442 *) DSA routines: under certain error conditions uninitialized BN objects
3443 could be freed. Solution: make sure initialization is performed early
3444 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3445 Nils Larsch <nla@trustcenter.de> via PR#459)
3446 [Lutz Jaenicke]
3447
3448 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3449 checked on reconnect on the client side, therefore session resumption
3450 could still fail with a "ssl session id is different" error. This
3451 behaviour is masked when SSL_OP_ALL is used due to
3452 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3453 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3454 followup to PR #377.
3455 [Lutz Jaenicke]
3456
3457 *) IA-32 assembler support enhancements: unified ELF targets, support
3458 for SCO/Caldera platforms, fix for Cygwin shared build.
3459 [Andy Polyakov]
3460
3461 *) Add support for FreeBSD on sparc64. As a consequence, support for
3462 FreeBSD on non-x86 processors is separate from x86 processors on
3463 the config script, much like the NetBSD support.
3464 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3465
3466 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3467
3468 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3469 OpenSSL 0.9.7.]
3470
3471 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3472 code (06) was taken as the first octet of the session ID and the last
3473 octet was ignored consequently. As a result SSLv2 client side session
3474 caching could not have worked due to the session ID mismatch between
3475 client and server.
3476 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3477 PR #377.
3478 [Lutz Jaenicke]
3479
3480 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3481 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3482 removed entirely.
3483 [Richard Levitte]
3484
3485 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3486 seems that in spite of existing for more than a year, many application
3487 author have done nothing to provide the necessary callbacks, which
3488 means that this particular engine will not work properly anywhere.
3489 This is a very unfortunate situation which forces us, in the name
3490 of usability, to give the hw_ncipher.c a static lock, which is part
3491 of libcrypto.
3492 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3493 appear in 0.9.8 or later. We EXPECT application authors to have
3494 dealt properly with this when 0.9.8 is released (unless we actually
3495 make such changes in the libcrypto locking code that changes will
3496 have to be made anyway).
3497 [Richard Levitte]
3498
3499 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3500 octets have been read, EOF or an error occurs. Without this change
3501 some truncated ASN1 structures will not produce an error.
3502 [Steve Henson]
3503
3504 *) Disable Heimdal support, since it hasn't been fully implemented.
3505 Still give the possibility to force the use of Heimdal, but with
3506 warnings and a request that patches get sent to openssl-dev.
3507 [Richard Levitte]
3508
3509 *) Add the VC-CE target, introduce the WINCE sysname, and add
3510 INSTALL.WCE and appropriate conditionals to make it build.
3511 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3512
3513 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3514 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3515 edit numbers of the version.
3516 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3517
3518 *) Introduce safe string copy and catenation functions
3519 (BUF_strlcpy() and BUF_strlcat()).
3520 [Ben Laurie (CHATS) and Richard Levitte]
3521
3522 *) Avoid using fixed-size buffers for one-line DNs.
3523 [Ben Laurie (CHATS)]
3524
3525 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3526 resizing buffers containing secrets, and use where appropriate.
3527 [Ben Laurie (CHATS)]
3528
3529 *) Avoid using fixed size buffers for configuration file location.
3530 [Ben Laurie (CHATS)]
3531
3532 *) Avoid filename truncation for various CA files.
3533 [Ben Laurie (CHATS)]
3534
3535 *) Use sizeof in preference to magic numbers.
3536 [Ben Laurie (CHATS)]
3537
3538 *) Avoid filename truncation in cert requests.
3539 [Ben Laurie (CHATS)]
3540
3541 *) Add assertions to check for (supposedly impossible) buffer
3542 overflows.
3543 [Ben Laurie (CHATS)]
3544
3545 *) Don't cache truncated DNS entries in the local cache (this could
3546 potentially lead to a spoofing attack).
3547 [Ben Laurie (CHATS)]
3548
3549 *) Fix various buffers to be large enough for hex/decimal
3550 representations in a platform independent manner.
3551 [Ben Laurie (CHATS)]
3552
3553 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3554 resizing buffers containing secrets, and use where appropriate.
3555 [Ben Laurie (CHATS)]
3556
3557 *) Add BIO_indent() to avoid much slightly worrying code to do
3558 indents.
3559 [Ben Laurie (CHATS)]
3560
3561 *) Convert sprintf()/BIO_puts() to BIO_printf().
3562 [Ben Laurie (CHATS)]
3563
3564 *) buffer_gets() could terminate with the buffer only half
3565 full. Fixed.
3566 [Ben Laurie (CHATS)]
3567
3568 *) Add assertions to prevent user-supplied crypto functions from
3569 overflowing internal buffers by having large block sizes, etc.
3570 [Ben Laurie (CHATS)]
3571
3572 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3573 unconditionally).
3574 [Ben Laurie (CHATS)]
3575
3576 *) Eliminate unused copy of key in RC4.
3577 [Ben Laurie (CHATS)]
3578
3579 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3580 [Ben Laurie (CHATS)]
3581
3582 *) Fix off-by-one error in EGD path.
3583 [Ben Laurie (CHATS)]
3584
3585 *) If RANDFILE path is too long, ignore instead of truncating.
3586 [Ben Laurie (CHATS)]
3587
3588 *) Eliminate unused and incorrectly sized X.509 structure
3589 CBCParameter.
3590 [Ben Laurie (CHATS)]
3591
3592 *) Eliminate unused and dangerous function knumber().
3593 [Ben Laurie (CHATS)]
3594
3595 *) Eliminate unused and dangerous structure, KSSL_ERR.
3596 [Ben Laurie (CHATS)]
3597
3598 *) Protect against overlong session ID context length in an encoded
3599 session object. Since these are local, this does not appear to be
3600 exploitable.
3601 [Ben Laurie (CHATS)]
3602
3603 *) Change from security patch (see 0.9.6e below) that did not affect
3604 the 0.9.6 release series:
3605
3606 Remote buffer overflow in SSL3 protocol - an attacker could
3607 supply an oversized master key in Kerberos-enabled versions.
3608 (CVE-2002-0657)
3609 [Ben Laurie (CHATS)]
3610
3611 *) Change the SSL kerb5 codes to match RFC 2712.
3612 [Richard Levitte]
3613
3614 *) Make -nameopt work fully for req and add -reqopt switch.
3615 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3616
3617 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3618 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3619
3620 *) Make sure tests can be performed even if the corresponding algorithms
3621 have been removed entirely. This was also the last step to make
3622 OpenSSL compilable with DJGPP under all reasonable conditions.
3623 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3624
3625 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3626 to allow version independent disabling of normally unselected ciphers,
3627 which may be activated as a side-effect of selecting a single cipher.
3628
3629 (E.g., cipher list string "RSA" enables ciphersuites that are left
3630 out of "ALL" because they do not provide symmetric encryption.
3631 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3632 [Lutz Jaenicke, Bodo Moeller]
3633
3634 *) Add appropriate support for separate platform-dependent build
3635 directories. The recommended way to make a platform-dependent
3636 build directory is the following (tested on Linux), maybe with
3637 some local tweaks:
3638
3639 # Place yourself outside of the OpenSSL source tree. In
3640 # this example, the environment variable OPENSSL_SOURCE
3641 # is assumed to contain the absolute OpenSSL source directory.
3642 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3643 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3644 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3645 mkdir -p `dirname $F`
3646 ln -s $OPENSSL_SOURCE/$F $F
3647 done
3648
3649 To be absolutely sure not to disturb the source tree, a "make clean"
3650 is a good thing. If it isn't successfull, don't worry about it,
3651 it probably means the source directory is very clean.
3652 [Richard Levitte]
3653
3654 *) Make sure any ENGINE control commands make local copies of string
3655 pointers passed to them whenever necessary. Otherwise it is possible
3656 the caller may have overwritten (or deallocated) the original string
3657 data when a later ENGINE operation tries to use the stored values.
3658 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3659
3660 *) Improve diagnostics in file reading and command-line digests.
3661 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3662
3663 *) Add AES modes CFB and OFB to the object database. Correct an
3664 error in AES-CFB decryption.
3665 [Richard Levitte]
3666
3667 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3668 allows existing EVP_CIPHER_CTX structures to be reused after
3669 calling EVP_*Final(). This behaviour is used by encryption
3670 BIOs and some applications. This has the side effect that
3671 applications must explicitly clean up cipher contexts with
3672 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3673 [Steve Henson]
3674
3675 *) Check the values of dna and dnb in bn_mul_recursive before calling
3676 bn_mul_comba (a non zero value means the a or b arrays do not contain
3677 n2 elements) and fallback to bn_mul_normal if either is not zero.
3678 [Steve Henson]
3679
3680 *) Fix escaping of non-ASCII characters when using the -subj option
3681 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3682 [Lutz Jaenicke]
3683
3684 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3685 form for "surname", serialNumber has no short form.
3686 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3687 therefore remove "mail" short name for "internet 7".
3688 The OID for unique identifiers in X509 certificates is
3689 x500UniqueIdentifier, not uniqueIdentifier.
3690 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3691 [Lutz Jaenicke]
3692
3693 *) Add an "init" command to the ENGINE config module and auto initialize
3694 ENGINEs. Without any "init" command the ENGINE will be initialized
3695 after all ctrl commands have been executed on it. If init=1 the
3696 ENGINE is initailized at that point (ctrls before that point are run
3697 on the uninitialized ENGINE and after on the initialized one). If
3698 init=0 then the ENGINE will not be iniatialized at all.
3699 [Steve Henson]
3700
3701 *) Fix the 'app_verify_callback' interface so that the user-defined
3702 argument is actually passed to the callback: In the
3703 SSL_CTX_set_cert_verify_callback() prototype, the callback
3704 declaration has been changed from
3705 int (*cb)()
3706 into
3707 int (*cb)(X509_STORE_CTX *,void *);
3708 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3709 i=s->ctx->app_verify_callback(&ctx)
3710 has been changed into
3711 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3712
3713 To update applications using SSL_CTX_set_cert_verify_callback(),
3714 a dummy argument can be added to their callback functions.
3715 [D. K. Smetters <smetters@parc.xerox.com>]
3716
3717 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3718 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3719
3720 *) Add and OPENSSL_LOAD_CONF define which will cause
3721 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3722 This allows older applications to transparently support certain
3723 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3724 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3725 load the config file and OPENSSL_add_all_algorithms_conf() which will
3726 always load it have also been added.
3727 [Steve Henson]
3728
3729 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3730 Adjust NIDs and EVP layer.
3731 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3732
3733 *) Config modules support in openssl utility.
3734
3735 Most commands now load modules from the config file,
3736 though in a few (such as version) this isn't done
3737 because it couldn't be used for anything.
3738
3739 In the case of ca and req the config file used is
3740 the same as the utility itself: that is the -config
3741 command line option can be used to specify an
3742 alternative file.
3743 [Steve Henson]
3744
3745 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3746 use "openssl_conf" if filename is NULL use default openssl config file.
3747 [Steve Henson]
3748
3749 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3750 config section name. Add a new flag to tolerate a missing config file
3751 and move code to CONF_modules_load_file().
3752 [Steve Henson]
3753
3754 *) Support for crypto accelerator cards from Accelerated Encryption
3755 Processing, www.aep.ie. (Use engine 'aep')
3756 The support was copied from 0.9.6c [engine] and adapted/corrected
3757 to work with the new engine framework.
3758 [AEP Inc. and Richard Levitte]
3759
3760 *) Support for SureWare crypto accelerator cards from Baltimore
3761 Technologies. (Use engine 'sureware')
3762 The support was copied from 0.9.6c [engine] and adapted
3763 to work with the new engine framework.
3764 [Richard Levitte]
3765
3766 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3767 make the newer ENGINE framework commands for the CHIL engine work.
3768 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3769
3770 *) Make it possible to produce shared libraries on ReliantUNIX.
3771 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3772
3773 *) Add the configuration target debug-linux-ppro.
3774 Make 'openssl rsa' use the general key loading routines
3775 implemented in apps.c, and make those routines able to
3776 handle the key format FORMAT_NETSCAPE and the variant
3777 FORMAT_IISSGC.
3778 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3779
3780 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3781 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3782
3783 *) Add -keyform to rsautl, and document -engine.
3784 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3785
3786 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3787 BIO_R_NO_SUCH_FILE error code rather than the generic
3788 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3789 [Ben Laurie]
3790
3791 *) Add new functions
3792 ERR_peek_last_error
3793 ERR_peek_last_error_line
3794 ERR_peek_last_error_line_data.
3795 These are similar to
3796 ERR_peek_error
3797 ERR_peek_error_line
3798 ERR_peek_error_line_data,
3799 but report on the latest error recorded rather than the first one
3800 still in the error queue.
3801 [Ben Laurie, Bodo Moeller]
3802
3803 *) default_algorithms option in ENGINE config module. This allows things
3804 like:
3805 default_algorithms = ALL
3806 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3807 [Steve Henson]
3808
3809 *) Prelminary ENGINE config module.
3810 [Steve Henson]
3811
3812 *) New experimental application configuration code.
3813 [Steve Henson]
3814
3815 *) Change the AES code to follow the same name structure as all other
3816 symmetric ciphers, and behave the same way. Move everything to
3817 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3818 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3819
3820 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3821 [Ben Laurie and Theo de Raadt]
3822
3823 *) Add option to output public keys in req command.
3824 [Massimiliano Pala madwolf@openca.org]
3825
3826 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3827 (up to about 10% better than before for P-192 and P-224).
3828 [Bodo Moeller]
3829
3830 *) New functions/macros
3831
3832 SSL_CTX_set_msg_callback(ctx, cb)
3833 SSL_CTX_set_msg_callback_arg(ctx, arg)
3834 SSL_set_msg_callback(ssl, cb)
3835 SSL_set_msg_callback_arg(ssl, arg)
3836
3837 to request calling a callback function
3838
3839 void cb(int write_p, int version, int content_type,
3840 const void *buf, size_t len, SSL *ssl, void *arg)
3841
3842 whenever a protocol message has been completely received
3843 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3844 protocol version according to which the SSL library interprets
3845 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3846 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3847 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3848 specification (change_cipher_spec(20), alert(21), handshake(22)).
3849 'buf' and 'len' point to the actual message, 'ssl' to the
3850 SSL object, and 'arg' is the application-defined value set by
3851 SSL[_CTX]_set_msg_callback_arg().
3852
3853 'openssl s_client' and 'openssl s_server' have new '-msg' options
3854 to enable a callback that displays all protocol messages.
3855 [Bodo Moeller]
3856
3857 *) Change the shared library support so shared libraries are built as
3858 soon as the corresponding static library is finished, and thereby get
3859 openssl and the test programs linked against the shared library.
3860 This still only happens when the keyword "shard" has been given to
3861 the configuration scripts.
3862
3863 NOTE: shared library support is still an experimental thing, and
3864 backward binary compatibility is still not guaranteed.
3865 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3866
3867 *) Add support for Subject Information Access extension.
3868 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3869
3870 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3871 additional bytes when new memory had to be allocated, not just
3872 when reusing an existing buffer.
3873 [Bodo Moeller]
3874
3875 *) New command line and configuration option 'utf8' for the req command.
3876 This allows field values to be specified as UTF8 strings.
3877 [Steve Henson]
3878
3879 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3880 runs for the former and machine-readable output for the latter.
3881 [Ben Laurie]
3882
3883 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3884 of the e-mail address in the DN (i.e., it will go into a certificate
3885 extension only). The new configuration file option 'email_in_dn = no'
3886 has the same effect.
3887 [Massimiliano Pala madwolf@openca.org]
3888
3889 *) Change all functions with names starting with des_ to be starting
3890 with DES_ instead. Add wrappers that are compatible with libdes,
3891 but are named _ossl_old_des_*. Finally, add macros that map the
3892 des_* symbols to the corresponding _ossl_old_des_* if libdes
3893 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3894 desired, the des_* symbols will be mapped to DES_*, with one
3895 exception.
3896
3897 Since we provide two compatibility mappings, the user needs to
3898 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3899 compatibility is desired. The default (i.e., when that macro
3900 isn't defined) is OpenSSL 0.9.6c compatibility.
3901
3902 There are also macros that enable and disable the support of old
3903 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3904 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3905 are defined, the default will apply: to support the old des routines.
3906
3907 In either case, one must include openssl/des.h to get the correct
3908 definitions. Do not try to just include openssl/des_old.h, that
3909 won't work.
3910
3911 NOTE: This is a major break of an old API into a new one. Software
3912 authors are encouraged to switch to the DES_ style functions. Some
3913 time in the future, des_old.h and the libdes compatibility functions
3914 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3915 default), and then completely removed.
3916 [Richard Levitte]
3917
3918 *) Test for certificates which contain unsupported critical extensions.
3919 If such a certificate is found during a verify operation it is
3920 rejected by default: this behaviour can be overridden by either
3921 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3922 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3923 X509_supported_extension() has also been added which returns 1 if a
3924 particular extension is supported.
3925 [Steve Henson]
3926
3927 *) Modify the behaviour of EVP cipher functions in similar way to digests
3928 to retain compatibility with existing code.
3929 [Steve Henson]
3930
3931 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3932 compatibility with existing code. In particular the 'ctx' parameter does
3933 not have to be to be initialized before the call to EVP_DigestInit() and
3934 it is tidied up after a call to EVP_DigestFinal(). New function
3935 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3936 EVP_MD_CTX_copy() changed to not require the destination to be
3937 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3938 requires the destination to be valid.
3939
3940 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3941 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3942 [Steve Henson]
3943
3944 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3945 so that complete 'Handshake' protocol structures are kept in memory
3946 instead of overwriting 'msg_type' and 'length' with 'body' data.
3947 [Bodo Moeller]
3948
3949 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3950 [Massimo Santin via Richard Levitte]
3951
3952 *) Major restructuring to the underlying ENGINE code. This includes
3953 reduction of linker bloat, separation of pure "ENGINE" manipulation
3954 (initialisation, etc) from functionality dealing with implementations
3955 of specific crypto iterfaces. This change also introduces integrated
3956 support for symmetric ciphers and digest implementations - so ENGINEs
3957 can now accelerate these by providing EVP_CIPHER and EVP_MD
3958 implementations of their own. This is detailed in crypto/engine/README
3959 as it couldn't be adequately described here. However, there are a few
3960 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3961 were changed in the original introduction of ENGINE code have now
3962 reverted back - the hooking from this code to ENGINE is now a good
3963 deal more passive and at run-time, operations deal directly with
3964 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3965 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3966 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3967 they were not being used by the framework as there is no concept of a
3968 BIGNUM_METHOD and they could not be generalised to the new
3969 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3970 ENGINE_cpy() has been removed as it cannot be consistently defined in
3971 the new code.
3972 [Geoff Thorpe]
3973
3974 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3975 [Steve Henson]
3976
3977 *) Change mkdef.pl to sort symbols that get the same entry number,
3978 and make sure the automatically generated functions ERR_load_*
3979 become part of libeay.num as well.
3980 [Richard Levitte]
3981
3982 *) New function SSL_renegotiate_pending(). This returns true once
3983 renegotiation has been requested (either SSL_renegotiate() call
3984 or HelloRequest/ClientHello receveived from the peer) and becomes
3985 false once a handshake has been completed.
3986 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3987 sends a HelloRequest, but does not ensure that a handshake takes
3988 place. SSL_renegotiate_pending() is useful for checking if the
3989 client has followed the request.)
3990 [Bodo Moeller]
3991
3992 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3993 By default, clients may request session resumption even during
3994 renegotiation (if session ID contexts permit); with this option,
3995 session resumption is possible only in the first handshake.
3996
3997 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3998 more bits available for options that should not be part of
3999 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4000 [Bodo Moeller]
4001
4002 *) Add some demos for certificate and certificate request creation.
4003 [Steve Henson]
4004
4005 *) Make maximum certificate chain size accepted from the peer application
4006 settable (SSL*_get/set_max_cert_list()), as proposed by
4007 "Douglas E. Engert" <deengert@anl.gov>.
4008 [Lutz Jaenicke]
4009
4010 *) Add support for shared libraries for Unixware-7
4011 (Boyd Lynn Gerber <gerberb@zenez.com>).
4012 [Lutz Jaenicke]
4013
4014 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4015 be done prior to destruction. Use this to unload error strings from
4016 ENGINEs that load their own error strings. NB: This adds two new API
4017 functions to "get" and "set" this destroy handler in an ENGINE.
4018 [Geoff Thorpe]
4019
4020 *) Alter all existing ENGINE implementations (except "openssl" and
4021 "openbsd") to dynamically instantiate their own error strings. This
4022 makes them more flexible to be built both as statically-linked ENGINEs
4023 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4024 Also, add stub code to each that makes building them as self-contained
4025 shared-libraries easier (see README.ENGINE).
4026 [Geoff Thorpe]
4027
4028 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4029 implementations into applications that are completely implemented in
4030 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4031 commands that can be used to configure what shared-library to load and
4032 to control aspects of the way it is handled. Also, made an update to
4033 the README.ENGINE file that brings its information up-to-date and
4034 provides some information and instructions on the "dynamic" ENGINE
4035 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4036 [Geoff Thorpe]
4037
4038 *) Make it possible to unload ranges of ERR strings with a new
4039 "ERR_unload_strings" function.
4040 [Geoff Thorpe]
4041
4042 *) Add a copy() function to EVP_MD.
4043 [Ben Laurie]
4044
4045 *) Make EVP_MD routines take a context pointer instead of just the
4046 md_data void pointer.
4047 [Ben Laurie]
4048
4049 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4050 that the digest can only process a single chunk of data
4051 (typically because it is provided by a piece of
4052 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4053 is only going to provide a single chunk of data, and hence the
4054 framework needn't accumulate the data for oneshot drivers.
4055 [Ben Laurie]
4056
4057 *) As with "ERR", make it possible to replace the underlying "ex_data"
4058 functions. This change also alters the storage and management of global
4059 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4060 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4061 index counters. The API functions that use this state have been changed
4062 to take a "class_index" rather than pointers to the class's local STACK
4063 and counter, and there is now an API function to dynamically create new
4064 classes. This centralisation allows us to (a) plug a lot of the
4065 thread-safety problems that existed, and (b) makes it possible to clean
4066 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4067 such data would previously have always leaked in application code and
4068 workarounds were in place to make the memory debugging turn a blind eye
4069 to it. Application code that doesn't use this new function will still
4070 leak as before, but their memory debugging output will announce it now
4071 rather than letting it slide.
4072
4073 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4074 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4075 has a return value to indicate success or failure.
4076 [Geoff Thorpe]
4077
4078 *) Make it possible to replace the underlying "ERR" functions such that the
4079 global state (2 LHASH tables and 2 locks) is only used by the "default"
4080 implementation. This change also adds two functions to "get" and "set"
4081 the implementation prior to it being automatically set the first time
4082 any other ERR function takes place. Ie. an application can call "get",
4083 pass the return value to a module it has just loaded, and that module
4084 can call its own "set" function using that value. This means the
4085 module's "ERR" operations will use (and modify) the error state in the
4086 application and not in its own statically linked copy of OpenSSL code.
4087 [Geoff Thorpe]
4088
4089 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4090 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4091 the operation, and provides a more encapsulated way for external code
4092 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4093 to use these functions rather than manually incrementing the counts.
4094
4095 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4096 [Geoff Thorpe]
4097
4098 *) Add EVP test program.
4099 [Ben Laurie]
4100
4101 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4102 [Ben Laurie]
4103
4104 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4105 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4106 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4107 These allow a CRL to be built without having to access X509_CRL fields
4108 directly. Modify 'ca' application to use new functions.
4109 [Steve Henson]
4110
4111 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4112 bug workarounds. Rollback attack detection is a security feature.
4113 The problem will only arise on OpenSSL servers when TLSv1 is not
4114 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4115 Software authors not wanting to support TLSv1 will have special reasons
4116 for their choice and can explicitly enable this option.
4117 [Bodo Moeller, Lutz Jaenicke]
4118
4119 *) Rationalise EVP so it can be extended: don't include a union of
4120 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4121 (similar to those existing for EVP_CIPHER_CTX).
4122 Usage example:
4123
4124 EVP_MD_CTX md;
4125
4126 EVP_MD_CTX_init(&md); /* new function call */
4127 EVP_DigestInit(&md, EVP_sha1());
4128 EVP_DigestUpdate(&md, in, len);
4129 EVP_DigestFinal(&md, out, NULL);
4130 EVP_MD_CTX_cleanup(&md); /* new function call */
4131
4132 [Ben Laurie]
4133
4134 *) Make DES key schedule conform to the usual scheme, as well as
4135 correcting its structure. This means that calls to DES functions
4136 now have to pass a pointer to a des_key_schedule instead of a
4137 plain des_key_schedule (which was actually always a pointer
4138 anyway): E.g.,
4139
4140 des_key_schedule ks;
4141
4142 des_set_key_checked(..., &ks);
4143 des_ncbc_encrypt(..., &ks, ...);
4144
4145 (Note that a later change renames 'des_...' into 'DES_...'.)
4146 [Ben Laurie]
4147
4148 *) Initial reduction of linker bloat: the use of some functions, such as
4149 PEM causes large amounts of unused functions to be linked in due to
4150 poor organisation. For example pem_all.c contains every PEM function
4151 which has a knock on effect of linking in large amounts of (unused)
4152 ASN1 code. Grouping together similar functions and splitting unrelated
4153 functions prevents this.
4154 [Steve Henson]
4155
4156 *) Cleanup of EVP macros.
4157 [Ben Laurie]
4158
4159 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4160 correct _ecb suffix.
4161 [Ben Laurie]
4162
4163 *) Add initial OCSP responder support to ocsp application. The
4164 revocation information is handled using the text based index
4165 use by the ca application. The responder can either handle
4166 requests generated internally, supplied in files (for example
4167 via a CGI script) or using an internal minimal server.
4168 [Steve Henson]
4169
4170 *) Add configuration choices to get zlib compression for TLS.
4171 [Richard Levitte]
4172
4173 *) Changes to Kerberos SSL for RFC 2712 compliance:
4174 1. Implemented real KerberosWrapper, instead of just using
4175 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4176 2. Implemented optional authenticator field of KerberosWrapper.
4177
4178 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4179 and authenticator structs; see crypto/krb5/.
4180
4181 Generalized Kerberos calls to support multiple Kerberos libraries.
4182 [Vern Staats <staatsvr@asc.hpc.mil>,
4183 Jeffrey Altman <jaltman@columbia.edu>
4184 via Richard Levitte]
4185
4186 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4187 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4188 values for each of the key sizes rather than having just
4189 parameters (and 'speed' generating keys each time).
4190 [Geoff Thorpe]
4191
4192 *) Speed up EVP routines.
4193 Before:
4194 encrypt
4195 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4196 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4197 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4198 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4199 decrypt
4200 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4201 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4202 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4203 After:
4204 encrypt
4205 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4206 decrypt
4207 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4208 [Ben Laurie]
4209
4210 *) Added the OS2-EMX target.
4211 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4212
4213 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4214 to support NCONF routines in extension code. New function CONF_set_nconf()
4215 to allow functions which take an NCONF to also handle the old LHASH
4216 structure: this means that the old CONF compatible routines can be
4217 retained (in particular wrt extensions) without having to duplicate the
4218 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4219 [Steve Henson]
4220
4221 *) Enhance the general user interface with mechanisms for inner control
4222 and with possibilities to have yes/no kind of prompts.
4223 [Richard Levitte]
4224
4225 *) Change all calls to low level digest routines in the library and
4226 applications to use EVP. Add missing calls to HMAC_cleanup() and
4227 don't assume HMAC_CTX can be copied using memcpy().
4228 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4229
4230 *) Add the possibility to control engines through control names but with
4231 arbitrary arguments instead of just a string.
4232 Change the key loaders to take a UI_METHOD instead of a callback
4233 function pointer. NOTE: this breaks binary compatibility with earlier
4234 versions of OpenSSL [engine].
4235 Adapt the nCipher code for these new conditions and add a card insertion
4236 callback.
4237 [Richard Levitte]
4238
4239 *) Enhance the general user interface with mechanisms to better support
4240 dialog box interfaces, application-defined prompts, the possibility
4241 to use defaults (for example default passwords from somewhere else)
4242 and interrupts/cancellations.
4243 [Richard Levitte]
4244
4245 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4246 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4247 [Steve Henson]
4248
4249 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4250 tidy up some unnecessarily weird code in 'sk_new()').
4251 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4252
4253 *) Change the key loading routines for ENGINEs to use the same kind
4254 callback (pem_password_cb) as all other routines that need this
4255 kind of callback.
4256 [Richard Levitte]
4257
4258 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4259 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4260 than this minimum value is recommended.
4261 [Lutz Jaenicke]
4262
4263 *) New random seeder for OpenVMS, using the system process statistics
4264 that are easily reachable.
4265 [Richard Levitte]
4266
4267 *) Windows apparently can't transparently handle global
4268 variables defined in DLLs. Initialisations such as:
4269
4270 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4271
4272 wont compile. This is used by the any applications that need to
4273 declare their own ASN1 modules. This was fixed by adding the option
4274 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4275 needed for static libraries under Win32.
4276 [Steve Henson]
4277
4278 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4279 setting of purpose and trust fields. New X509_STORE trust and
4280 purpose functions and tidy up setting in other SSL functions.
4281 [Steve Henson]
4282
4283 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4284 structure. These are inherited by X509_STORE_CTX when it is
4285 initialised. This allows various defaults to be set in the
4286 X509_STORE structure (such as flags for CRL checking and custom
4287 purpose or trust settings) for functions which only use X509_STORE_CTX
4288 internally such as S/MIME.
4289
4290 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4291 trust settings if they are not set in X509_STORE. This allows X509_STORE
4292 purposes and trust (in S/MIME for example) to override any set by default.
4293
4294 Add command line options for CRL checking to smime, s_client and s_server
4295 applications.
4296 [Steve Henson]
4297
4298 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4299 are set then the CRL is looked up in the X509_STORE structure and
4300 its validity and signature checked, then if the certificate is found
4301 in the CRL the verify fails with a revoked error.
4302
4303 Various new CRL related callbacks added to X509_STORE_CTX structure.
4304
4305 Command line options added to 'verify' application to support this.
4306
4307 This needs some additional work, such as being able to handle multiple
4308 CRLs with different times, extension based lookup (rather than just
4309 by subject name) and ultimately more complete V2 CRL extension
4310 handling.
4311 [Steve Henson]
4312
4313 *) Add a general user interface API (crypto/ui/). This is designed
4314 to replace things like des_read_password and friends (backward
4315 compatibility functions using this new API are provided).
4316 The purpose is to remove prompting functions from the DES code
4317 section as well as provide for prompting through dialog boxes in
4318 a window system and the like.
4319 [Richard Levitte]
4320
4321 *) Add "ex_data" support to ENGINE so implementations can add state at a
4322 per-structure level rather than having to store it globally.
4323 [Geoff]
4324
4325 *) Make it possible for ENGINE structures to be copied when retrieved by
4326 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4327 This causes the "original" ENGINE structure to act like a template,
4328 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4329 operational state can be localised to each ENGINE structure, despite the
4330 fact they all share the same "methods". New ENGINE structures returned in
4331 this case have no functional references and the return value is the single
4332 structural reference. This matches the single structural reference returned
4333 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4334 ENGINE structure.
4335 [Geoff]
4336
4337 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4338 needs to match any other type at all we need to manually clear the
4339 tag cache.
4340 [Steve Henson]
4341
4342 *) Changes to the "openssl engine" utility to include;
4343 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4344 about an ENGINE's available control commands.
4345 - executing control commands from command line arguments using the
4346 '-pre' and '-post' switches. '-post' is only used if '-t' is
4347 specified and the ENGINE is successfully initialised. The syntax for
4348 the individual commands are colon-separated, for example;
4349 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4350 [Geoff]
4351
4352 *) New dynamic control command support for ENGINEs. ENGINEs can now
4353 declare their own commands (numbers), names (strings), descriptions,
4354 and input types for run-time discovery by calling applications. A
4355 subset of these commands are implicitly classed as "executable"
4356 depending on their input type, and only these can be invoked through
4357 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4358 can be based on user input, config files, etc). The distinction is
4359 that "executable" commands cannot return anything other than a boolean
4360 result and can only support numeric or string input, whereas some
4361 discoverable commands may only be for direct use through
4362 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4363 pointers, or other custom uses. The "executable" commands are to
4364 support parameterisations of ENGINE behaviour that can be
4365 unambiguously defined by ENGINEs and used consistently across any
4366 OpenSSL-based application. Commands have been added to all the
4367 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4368 control over shared-library paths without source code alterations.
4369 [Geoff]
4370
4371 *) Changed all ENGINE implementations to dynamically allocate their
4372 ENGINEs rather than declaring them statically. Apart from this being
4373 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4374 this also allows the implementations to compile without using the
4375 internal engine_int.h header.
4376 [Geoff]
4377
4378 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4379 'const' value. Any code that should be able to modify a RAND_METHOD
4380 should already have non-const pointers to it (ie. they should only
4381 modify their own ones).
4382 [Geoff]
4383
4384 *) Made a variety of little tweaks to the ENGINE code.
4385 - "atalla" and "ubsec" string definitions were moved from header files
4386 to C code. "nuron" string definitions were placed in variables
4387 rather than hard-coded - allowing parameterisation of these values
4388 later on via ctrl() commands.
4389 - Removed unused "#if 0"'d code.
4390 - Fixed engine list iteration code so it uses ENGINE_free() to release
4391 structural references.
4392 - Constified the RAND_METHOD element of ENGINE structures.
4393 - Constified various get/set functions as appropriate and added
4394 missing functions (including a catch-all ENGINE_cpy that duplicates
4395 all ENGINE values onto a new ENGINE except reference counts/state).
4396 - Removed NULL parameter checks in get/set functions. Setting a method
4397 or function to NULL is a way of cancelling out a previously set
4398 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4399 and doesn't justify the extra error symbols and code.
4400 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4401 flags from engine_int.h to engine.h.
4402 - Changed prototypes for ENGINE handler functions (init(), finish(),
4403 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4404 [Geoff]
4405
4406 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4407 to the algorithm using long division. The binary algorithm can be
4408 used only if the modulus is odd. On 32-bit systems, it is faster
4409 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4410 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4411 up to 450 bits. In 64-bit environments, the binary algorithm
4412 appears to be advantageous for much longer moduli; here we use it
4413 for moduli up to 2048 bits.
4414 [Bodo Moeller]
4415
4416 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4417 could not support the combine flag in choice fields.
4418 [Steve Henson]
4419
4420 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4421 extensions from a certificate request to the certificate.
4422 [Steve Henson]
4423
4424 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4425 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4426 file: this allows the display of the certificate about to be
4427 signed to be customised, to allow certain fields to be included
4428 or excluded and extension details. The old system didn't display
4429 multicharacter strings properly, omitted fields not in the policy
4430 and couldn't display additional details such as extensions.
4431 [Steve Henson]
4432
4433 *) Function EC_POINTs_mul for multiple scalar multiplication
4434 of an arbitrary number of elliptic curve points
4435 \sum scalars[i]*points[i],
4436 optionally including the generator defined for the EC_GROUP:
4437 scalar*generator + \sum scalars[i]*points[i].
4438
4439 EC_POINT_mul is a simple wrapper function for the typical case
4440 that the point list has just one item (besides the optional
4441 generator).
4442 [Bodo Moeller]
4443
4444 *) First EC_METHODs for curves over GF(p):
4445
4446 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4447 operations and provides various method functions that can also
4448 operate with faster implementations of modular arithmetic.
4449
4450 EC_GFp_mont_method() reuses most functions that are part of
4451 EC_GFp_simple_method, but uses Montgomery arithmetic.
4452
4453 [Bodo Moeller; point addition and point doubling
4454 implementation directly derived from source code provided by
4455 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4456
4457 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4458 crypto/ec/ec_lib.c):
4459
4460 Curves are EC_GROUP objects (with an optional group generator)
4461 based on EC_METHODs that are built into the library.
4462
4463 Points are EC_POINT objects based on EC_GROUP objects.
4464
4465 Most of the framework would be able to handle curves over arbitrary
4466 finite fields, but as there are no obvious types for fields other
4467 than GF(p), some functions are limited to that for now.
4468 [Bodo Moeller]
4469
4470 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4471 that the file contains a complete HTTP response.
4472 [Richard Levitte]
4473
4474 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4475 change the def and num file printf format specifier from "%-40sXXX"
4476 to "%-39s XXX". The latter will always guarantee a space after the
4477 field while the former will cause them to run together if the field
4478 is 40 of more characters long.
4479 [Steve Henson]
4480
4481 *) Constify the cipher and digest 'method' functions and structures
4482 and modify related functions to take constant EVP_MD and EVP_CIPHER
4483 pointers.
4484 [Steve Henson]
4485
4486 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4487 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4488 [Bodo Moeller]
4489
4490 *) Modify EVP_Digest*() routines so they now return values. Although the
4491 internal software routines can never fail additional hardware versions
4492 might.
4493 [Steve Henson]
4494
4495 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4496
4497 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4498 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4499
4500 ASN1 error codes
4501 ERR_R_NESTED_ASN1_ERROR
4502 ...
4503 ERR_R_MISSING_ASN1_EOS
4504 were 4 .. 9, conflicting with
4505 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4506 ...
4507 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4508 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4509
4510 Add new error code 'ERR_R_INTERNAL_ERROR'.
4511 [Bodo Moeller]
4512
4513 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4514 suffices.
4515 [Bodo Moeller]
4516
4517 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4518 sets the subject name for a new request or supersedes the
4519 subject name in a given request. Formats that can be parsed are
4520 'CN=Some Name, OU=myOU, C=IT'
4521 and
4522 'CN=Some Name/OU=myOU/C=IT'.
4523
4524 Add options '-batch' and '-verbose' to 'openssl req'.
4525 [Massimiliano Pala <madwolf@hackmasters.net>]
4526
4527 *) Introduce the possibility to access global variables through
4528 functions on platform were that's the best way to handle exporting
4529 global variables in shared libraries. To enable this functionality,
4530 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4531 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4532 is normally done by Configure or something similar).
4533
4534 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4535 in the source file (foo.c) like this:
4536
4537 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4538 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4539
4540 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4541 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4542
4543 OPENSSL_DECLARE_GLOBAL(int,foo);
4544 #define foo OPENSSL_GLOBAL_REF(foo)
4545 OPENSSL_DECLARE_GLOBAL(double,bar);
4546 #define bar OPENSSL_GLOBAL_REF(bar)
4547
4548 The #defines are very important, and therefore so is including the
4549 header file everywhere where the defined globals are used.
4550
4551 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4552 of ASN.1 items, but that structure is a bit different.
4553
4554 The largest change is in util/mkdef.pl which has been enhanced with
4555 better and easier to understand logic to choose which symbols should
4556 go into the Windows .def files as well as a number of fixes and code
4557 cleanup (among others, algorithm keywords are now sorted
4558 lexicographically to avoid constant rewrites).
4559 [Richard Levitte]
4560
4561 *) In BN_div() keep a copy of the sign of 'num' before writing the
4562 result to 'rm' because if rm==num the value will be overwritten
4563 and produce the wrong result if 'num' is negative: this caused
4564 problems with BN_mod() and BN_nnmod().
4565 [Steve Henson]
4566
4567 *) Function OCSP_request_verify(). This checks the signature on an
4568 OCSP request and verifies the signer certificate. The signer
4569 certificate is just checked for a generic purpose and OCSP request
4570 trust settings.
4571 [Steve Henson]
4572
4573 *) Add OCSP_check_validity() function to check the validity of OCSP
4574 responses. OCSP responses are prepared in real time and may only
4575 be a few seconds old. Simply checking that the current time lies
4576 between thisUpdate and nextUpdate max reject otherwise valid responses
4577 caused by either OCSP responder or client clock inaccuracy. Instead
4578 we allow thisUpdate and nextUpdate to fall within a certain period of
4579 the current time. The age of the response can also optionally be
4580 checked. Two new options -validity_period and -status_age added to
4581 ocsp utility.
4582 [Steve Henson]
4583
4584 *) If signature or public key algorithm is unrecognized print out its
4585 OID rather that just UNKNOWN.
4586 [Steve Henson]
4587
4588 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4589 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4590 ID to be generated from the issuer certificate alone which can then be
4591 passed to OCSP_id_issuer_cmp().
4592 [Steve Henson]
4593
4594 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4595 ASN1 modules to export functions returning ASN1_ITEM pointers
4596 instead of the ASN1_ITEM structures themselves. This adds several
4597 new macros which allow the underlying ASN1 function/structure to
4598 be accessed transparently. As a result code should not use ASN1_ITEM
4599 references directly (such as &X509_it) but instead use the relevant
4600 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4601 use of the new ASN1 code on platforms where exporting structures
4602 is problematical (for example in shared libraries) but exporting
4603 functions returning pointers to structures is not.
4604 [Steve Henson]
4605
4606 *) Add support for overriding the generation of SSL/TLS session IDs.
4607 These callbacks can be registered either in an SSL_CTX or per SSL.
4608 The purpose of this is to allow applications to control, if they wish,
4609 the arbitrary values chosen for use as session IDs, particularly as it
4610 can be useful for session caching in multiple-server environments. A
4611 command-line switch for testing this (and any client code that wishes
4612 to use such a feature) has been added to "s_server".
4613 [Geoff Thorpe, Lutz Jaenicke]
4614
4615 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4616 of the form '#if defined(...) || defined(...) || ...' and
4617 '#if !defined(...) && !defined(...) && ...'. This also avoids
4618 the growing number of special cases it was previously handling.
4619 [Richard Levitte]
4620
4621 *) Make all configuration macros available for application by making
4622 sure they are available in opensslconf.h, by giving them names starting
4623 with "OPENSSL_" to avoid conflicts with other packages and by making
4624 sure e_os2.h will cover all platform-specific cases together with
4625 opensslconf.h.
4626 Additionally, it is now possible to define configuration/platform-
4627 specific names (called "system identities"). In the C code, these
4628 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4629 macro with the name beginning with "OPENSSL_SYS_", which is determined
4630 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4631 what is available.
4632 [Richard Levitte]
4633
4634 *) New option -set_serial to 'req' and 'x509' this allows the serial
4635 number to use to be specified on the command line. Previously self
4636 signed certificates were hard coded with serial number 0 and the
4637 CA options of 'x509' had to use a serial number in a file which was
4638 auto incremented.
4639 [Steve Henson]
4640
4641 *) New options to 'ca' utility to support V2 CRL entry extensions.
4642 Currently CRL reason, invalidity date and hold instruction are
4643 supported. Add new CRL extensions to V3 code and some new objects.
4644 [Steve Henson]
4645
4646 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4647 disable standard block padding (aka PKCS#5 padding) in the EVP
4648 API, which was previously mandatory. This means that the data is
4649 not padded in any way and so the total length much be a multiple
4650 of the block size, otherwise an error occurs.
4651 [Steve Henson]
4652
4653 *) Initial (incomplete) OCSP SSL support.
4654 [Steve Henson]
4655
4656 *) New function OCSP_parse_url(). This splits up a URL into its host,
4657 port and path components: primarily to parse OCSP URLs. New -url
4658 option to ocsp utility.
4659 [Steve Henson]
4660
4661 *) New nonce behavior. The return value of OCSP_check_nonce() now
4662 reflects the various checks performed. Applications can decide
4663 whether to tolerate certain situations such as an absent nonce
4664 in a response when one was present in a request: the ocsp application
4665 just prints out a warning. New function OCSP_add1_basic_nonce()
4666 this is to allow responders to include a nonce in a response even if
4667 the request is nonce-less.
4668 [Steve Henson]
4669
4670 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4671 skipped when using openssl x509 multiple times on a single input file,
4672 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4673 [Bodo Moeller]
4674
4675 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4676 set string type: to handle setting ASN1_TIME structures. Fix ca
4677 utility to correctly initialize revocation date of CRLs.
4678 [Steve Henson]
4679
4680 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4681 the clients preferred ciphersuites and rather use its own preferences.
4682 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4683 Internet Explorer by ensuring unchanged hash method during stepup.
4684 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4685 [Lutz Jaenicke]
4686
4687 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4688 to aes and add a new 'exist' option to print out symbols that don't
4689 appear to exist.
4690 [Steve Henson]
4691
4692 *) Additional options to ocsp utility to allow flags to be set and
4693 additional certificates supplied.
4694 [Steve Henson]
4695
4696 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4697 OCSP client a number of certificate to only verify the response
4698 signature against.
4699 [Richard Levitte]
4700
4701 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4702 handle the new API. Currently only ECB, CBC modes supported. Add new
4703 AES OIDs.
4704
4705 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4706 Encryption Standard (AES) Ciphersuites for Transport Layer
4707 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4708 not enabled by default and were not part of the "ALL" ciphersuite
4709 alias because they were not yet official; they could be
4710 explicitly requested by specifying the "AESdraft" ciphersuite
4711 group alias. In the final release of OpenSSL 0.9.7, the group
4712 alias is called "AES" and is part of "ALL".)
4713 [Ben Laurie, Steve Henson, Bodo Moeller]
4714
4715 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4716 request to response.
4717 [Steve Henson]
4718
4719 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4720 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4721 extract information from a certificate request. OCSP_response_create()
4722 creates a response and optionally adds a basic response structure.
4723 OCSP_basic_add1_status() adds a complete single response to a basic
4724 response and returns the OCSP_SINGLERESP structure just added (to allow
4725 extensions to be included for example). OCSP_basic_add1_cert() adds a
4726 certificate to a basic response and OCSP_basic_sign() signs a basic
4727 response with various flags. New helper functions ASN1_TIME_check()
4728 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4729 (converts ASN1_TIME to GeneralizedTime).
4730 [Steve Henson]
4731
4732 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4733 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4734 structure from a certificate. X509_pubkey_digest() digests the public_key
4735 contents: this is used in various key identifiers.
4736 [Steve Henson]
4737
4738 *) Make sk_sort() tolerate a NULL argument.
4739 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4740
4741 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4742 passed by the function are trusted implicitly. If any of them signed the
4743 response then it is assumed to be valid and is not verified.
4744 [Steve Henson]
4745
4746 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4747 to data. This was previously part of the PKCS7 ASN1 code. This
4748 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4749 [Steve Henson, reported by Kenneth R. Robinette
4750 <support@securenetterm.com>]
4751
4752 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4753 routines: without these tracing memory leaks is very painful.
4754 Fix leaks in PKCS12 and PKCS7 routines.
4755 [Steve Henson]
4756
4757 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4758 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4759 effectively meant GeneralizedTime would never be used. Now it
4760 is initialised to -1 but X509_time_adj() now has to check the value
4761 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4762 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4763 [Steve Henson, reported by Kenneth R. Robinette
4764 <support@securenetterm.com>]
4765
4766 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4767 result in a zero length in the ASN1_INTEGER structure which was
4768 not consistent with the structure when d2i_ASN1_INTEGER() was used
4769 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4770 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4771 where it did not print out a minus for negative ASN1_INTEGER.
4772 [Steve Henson]
4773
4774 *) Add summary printout to ocsp utility. The various functions which
4775 convert status values to strings have been renamed to:
4776 OCSP_response_status_str(), OCSP_cert_status_str() and
4777 OCSP_crl_reason_str() and are no longer static. New options
4778 to verify nonce values and to disable verification. OCSP response
4779 printout format cleaned up.
4780 [Steve Henson]
4781
4782 *) Add additional OCSP certificate checks. These are those specified
4783 in RFC2560. This consists of two separate checks: the CA of the
4784 certificate being checked must either be the OCSP signer certificate
4785 or the issuer of the OCSP signer certificate. In the latter case the
4786 OCSP signer certificate must contain the OCSP signing extended key
4787 usage. This check is performed by attempting to match the OCSP
4788 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4789 in the OCSP_CERTID structures of the response.
4790 [Steve Henson]
4791
4792 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4793 and related routines. This uses the standard OpenSSL certificate
4794 verify routines to perform initial checks (just CA validity) and
4795 to obtain the certificate chain. Then additional checks will be
4796 performed on the chain. Currently the root CA is checked to see
4797 if it is explicitly trusted for OCSP signing. This is used to set
4798 a root CA as a global signing root: that is any certificate that
4799 chains to that CA is an acceptable OCSP signing certificate.
4800 [Steve Henson]
4801
4802 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4803 extensions from a separate configuration file.
4804 As when reading extensions from the main configuration file,
4805 the '-extensions ...' option may be used for specifying the
4806 section to use.
4807 [Massimiliano Pala <madwolf@comune.modena.it>]
4808
4809 *) New OCSP utility. Allows OCSP requests to be generated or
4810 read. The request can be sent to a responder and the output
4811 parsed, outputed or printed in text form. Not complete yet:
4812 still needs to check the OCSP response validity.
4813 [Steve Henson]
4814
4815 *) New subcommands for 'openssl ca':
4816 'openssl ca -status <serial>' prints the status of the cert with
4817 the given serial number (according to the index file).
4818 'openssl ca -updatedb' updates the expiry status of certificates
4819 in the index file.
4820 [Massimiliano Pala <madwolf@comune.modena.it>]
4821
4822 *) New '-newreq-nodes' command option to CA.pl. This is like
4823 '-newreq', but calls 'openssl req' with the '-nodes' option
4824 so that the resulting key is not encrypted.
4825 [Damien Miller <djm@mindrot.org>]
4826
4827 *) New configuration for the GNU Hurd.
4828 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4829
4830 *) Initial code to implement OCSP basic response verify. This
4831 is currently incomplete. Currently just finds the signer's
4832 certificate and verifies the signature on the response.
4833 [Steve Henson]
4834
4835 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4836 value of OPENSSLDIR. This is available via the new '-d' option
4837 to 'openssl version', and is also included in 'openssl version -a'.
4838 [Bodo Moeller]
4839
4840 *) Allowing defining memory allocation callbacks that will be given
4841 file name and line number information in additional arguments
4842 (a const char* and an int). The basic functionality remains, as
4843 well as the original possibility to just replace malloc(),
4844 realloc() and free() by functions that do not know about these
4845 additional arguments. To register and find out the current
4846 settings for extended allocation functions, the following
4847 functions are provided:
4848
4849 CRYPTO_set_mem_ex_functions
4850 CRYPTO_set_locked_mem_ex_functions
4851 CRYPTO_get_mem_ex_functions
4852 CRYPTO_get_locked_mem_ex_functions
4853
4854 These work the same way as CRYPTO_set_mem_functions and friends.
4855 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4856 extended allocation function is enabled.
4857 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4858 a conventional allocation function is enabled.
4859 [Richard Levitte, Bodo Moeller]
4860
4861 *) Finish off removing the remaining LHASH function pointer casts.
4862 There should no longer be any prototype-casting required when using
4863 the LHASH abstraction, and any casts that remain are "bugs". See
4864 the callback types and macros at the head of lhash.h for details
4865 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4866 [Geoff Thorpe]
4867
4868 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4869 If /dev/[u]random devices are not available or do not return enough
4870 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4871 be queried.
4872 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4873 /etc/entropy will be queried once each in this sequence, quering stops
4874 when enough entropy was collected without querying more sockets.
4875 [Lutz Jaenicke]
4876
4877 *) Change the Unix RAND_poll() variant to be able to poll several
4878 random devices, as specified by DEVRANDOM, until a sufficient amount
4879 of data has been collected. We spend at most 10 ms on each file
4880 (select timeout) and read in non-blocking mode. DEVRANDOM now
4881 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4882 (previously it was just the string "/dev/urandom"), so on typical
4883 platforms the 10 ms delay will never occur.
4884 Also separate out the Unix variant to its own file, rand_unix.c.
4885 For VMS, there's a currently-empty rand_vms.c.
4886 [Richard Levitte]
4887
4888 *) Move OCSP client related routines to ocsp_cl.c. These
4889 provide utility functions which an application needing
4890 to issue a request to an OCSP responder and analyse the
4891 response will typically need: as opposed to those which an
4892 OCSP responder itself would need which will be added later.
4893
4894 OCSP_request_sign() signs an OCSP request with an API similar
4895 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4896 response. OCSP_response_get1_basic() extracts basic response
4897 from response. OCSP_resp_find_status(): finds and extracts status
4898 information from an OCSP_CERTID structure (which will be created
4899 when the request structure is built). These are built from lower
4900 level functions which work on OCSP_SINGLERESP structures but
4901 wont normally be used unless the application wishes to examine
4902 extensions in the OCSP response for example.
4903
4904 Replace nonce routines with a pair of functions.
4905 OCSP_request_add1_nonce() adds a nonce value and optionally
4906 generates a random value. OCSP_check_nonce() checks the
4907 validity of the nonce in an OCSP response.
4908 [Steve Henson]
4909
4910 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4911 This doesn't copy the supplied OCSP_CERTID and avoids the
4912 need to free up the newly created id. Change return type
4913 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4914 This can then be used to add extensions to the request.
4915 Deleted OCSP_request_new(), since most of its functionality
4916 is now in OCSP_REQUEST_new() (and the case insensitive name
4917 clash) apart from the ability to set the request name which
4918 will be added elsewhere.
4919 [Steve Henson]
4920
4921 *) Update OCSP API. Remove obsolete extensions argument from
4922 various functions. Extensions are now handled using the new
4923 OCSP extension code. New simple OCSP HTTP function which
4924 can be used to send requests and parse the response.
4925 [Steve Henson]
4926
4927 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4928 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4929 uses the special reorder version of SET OF to sort the attributes
4930 and reorder them to match the encoded order. This resolves a long
4931 standing problem: a verify on a PKCS7 structure just after signing
4932 it used to fail because the attribute order did not match the
4933 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4934 it uses the received order. This is necessary to tolerate some broken
4935 software that does not order SET OF. This is handled by encoding
4936 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4937 to produce the required SET OF.
4938 [Steve Henson]
4939
4940 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4941 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4942 files to get correct declarations of the ASN.1 item variables.
4943 [Richard Levitte]
4944
4945 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4946 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4947 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4948 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4949 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4950 ASN1_ITEM and no wrapper functions.
4951 [Steve Henson]
4952
4953 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4954 replace the old function pointer based I/O routines. Change most of
4955 the *_d2i_bio() and *_d2i_fp() functions to use these.
4956 [Steve Henson]
4957
4958 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4959 lines, recognice more "algorithms" that can be deselected, and make
4960 it complain about algorithm deselection that isn't recognised.
4961 [Richard Levitte]
4962
4963 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4964 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4965 to use new functions. Add NO_ASN1_OLD which can be set to remove
4966 some old style ASN1 functions: this can be used to determine if old
4967 code will still work when these eventually go away.
4968 [Steve Henson]
4969
4970 *) New extension functions for OCSP structures, these follow the
4971 same conventions as certificates and CRLs.
4972 [Steve Henson]
4973
4974 *) New function X509V3_add1_i2d(). This automatically encodes and
4975 adds an extension. Its behaviour can be customised with various
4976 flags to append, replace or delete. Various wrappers added for
4977 certifcates and CRLs.
4978 [Steve Henson]
4979
4980 *) Fix to avoid calling the underlying ASN1 print routine when
4981 an extension cannot be parsed. Correct a typo in the
4982 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4983 [Steve Henson]
4984
4985 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4986 entries for variables.
4987 [Steve Henson]
4988
4989 *) Add functionality to apps/openssl.c for detecting locking
4990 problems: As the program is single-threaded, all we have
4991 to do is register a locking callback using an array for
4992 storing which locks are currently held by the program.
4993 [Bodo Moeller]
4994
4995 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4996 SSL_get_ex_data_X509_STORE_idx(), which is used in
4997 ssl_verify_cert_chain() and thus can be called at any time
4998 during TLS/SSL handshakes so that thread-safety is essential.
4999 Unfortunately, the ex_data design is not at all suited
5000 for multi-threaded use, so it probably should be abolished.
5001 [Bodo Moeller]
5002
5003 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5004 [Broadcom, tweaked and integrated by Geoff Thorpe]
5005
5006 *) Move common extension printing code to new function
5007 X509V3_print_extensions(). Reorganise OCSP print routines and
5008 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5009 [Steve Henson]
5010
5011 *) New function X509_signature_print() to remove duplication in some
5012 print routines.
5013 [Steve Henson]
5014
5015 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5016 set (this was treated exactly the same as SET OF previously). This
5017 is used to reorder the STACK representing the structure to match the
5018 encoding. This will be used to get round a problem where a PKCS7
5019 structure which was signed could not be verified because the STACK
5020 order did not reflect the encoded order.
5021 [Steve Henson]
5022
5023 *) Reimplement the OCSP ASN1 module using the new code.
5024 [Steve Henson]
5025
5026 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5027 for its ASN1 operations. The old style function pointers still exist
5028 for now but they will eventually go away.
5029 [Steve Henson]
5030
5031 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5032 completely replaces the old ASN1 functionality with a table driven
5033 encoder and decoder which interprets an ASN1_ITEM structure describing
5034 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5035 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5036 has also been converted to the new form.
5037 [Steve Henson]
5038
5039 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5040 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5041 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5042 for negative moduli.
5043 [Bodo Moeller]
5044
5045 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5046 of not touching the result's sign bit.
5047 [Bodo Moeller]
5048
5049 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5050 set.
5051 [Bodo Moeller]
5052
5053 *) Changed the LHASH code to use prototypes for callbacks, and created
5054 macros to declare and implement thin (optionally static) functions
5055 that provide type-safety and avoid function pointer casting for the
5056 type-specific callbacks.
5057 [Geoff Thorpe]
5058
5059 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5060 RFC 2712.
5061 [Veers Staats <staatsvr@asc.hpc.mil>,
5062 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5063
5064 *) Reformat the FAQ so the different questions and answers can be divided
5065 in sections depending on the subject.
5066 [Richard Levitte]
5067
5068 *) Have the zlib compression code load ZLIB.DLL dynamically under
5069 Windows.
5070 [Richard Levitte]
5071
5072 *) New function BN_mod_sqrt for computing square roots modulo a prime
5073 (using the probabilistic Tonelli-Shanks algorithm unless
5074 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5075 be handled deterministically).
5076 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5077
5078 *) Make BN_mod_inverse faster by explicitly handling small quotients
5079 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5080 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5081 [Bodo Moeller]
5082
5083 *) New function BN_kronecker.
5084 [Bodo Moeller]
5085
5086 *) Fix BN_gcd so that it works on negative inputs; the result is
5087 positive unless both parameters are zero.
5088 Previously something reasonably close to an infinite loop was
5089 possible because numbers could be growing instead of shrinking
5090 in the implementation of Euclid's algorithm.
5091 [Bodo Moeller]
5092
5093 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5094 sign of the number in question.
5095
5096 Fix BN_is_word(a,w) to work correctly for w == 0.
5097
5098 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5099 because its test if the absolute value of 'a' equals 'w'.
5100 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5101 it exists mostly for use in the implementations of BN_is_zero(),
5102 BN_is_one(), and BN_is_word().
5103 [Bodo Moeller]
5104
5105 *) New function BN_swap.
5106 [Bodo Moeller]
5107
5108 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5109 the exponentiation functions are more likely to produce reasonable
5110 results on negative inputs.
5111 [Bodo Moeller]
5112
5113 *) Change BN_mod_mul so that the result is always non-negative.
5114 Previously, it could be negative if one of the factors was negative;
5115 I don't think anyone really wanted that behaviour.
5116 [Bodo Moeller]
5117
5118 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5119 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5120 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5121 and add new functions:
5122
5123 BN_nnmod
5124 BN_mod_sqr
5125 BN_mod_add
5126 BN_mod_add_quick
5127 BN_mod_sub
5128 BN_mod_sub_quick
5129 BN_mod_lshift1
5130 BN_mod_lshift1_quick
5131 BN_mod_lshift
5132 BN_mod_lshift_quick
5133
5134 These functions always generate non-negative results.
5135
5136 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5137 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5138
5139 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5140 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5141 be reduced modulo m.
5142 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5143
5144 #if 0
5145 The following entry accidentily appeared in the CHANGES file
5146 distributed with OpenSSL 0.9.7. The modifications described in
5147 it do *not* apply to OpenSSL 0.9.7.
5148
5149 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5150 was actually never needed) and in BN_mul(). The removal in BN_mul()
5151 required a small change in bn_mul_part_recursive() and the addition
5152 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5153 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5154 bn_sub_words() and bn_add_words() except they take arrays with
5155 differing sizes.
5156 [Richard Levitte]
5157 #endif
5158
5159 *) In 'openssl passwd', verify passwords read from the terminal
5160 unless the '-salt' option is used (which usually means that
5161 verification would just waste user's time since the resulting
5162 hash is going to be compared with some given password hash)
5163 or the new '-noverify' option is used.
5164
5165 This is an incompatible change, but it does not affect
5166 non-interactive use of 'openssl passwd' (passwords on the command
5167 line, '-stdin' option, '-in ...' option) and thus should not
5168 cause any problems.
5169 [Bodo Moeller]
5170
5171 *) Remove all references to RSAref, since there's no more need for it.
5172 [Richard Levitte]
5173
5174 *) Make DSO load along a path given through an environment variable
5175 (SHLIB_PATH) with shl_load().
5176 [Richard Levitte]
5177
5178 *) Constify the ENGINE code as a result of BIGNUM constification.
5179 Also constify the RSA code and most things related to it. In a
5180 few places, most notable in the depth of the ASN.1 code, ugly
5181 casts back to non-const were required (to be solved at a later
5182 time)
5183 [Richard Levitte]
5184
5185 *) Make it so the openssl application has all engines loaded by default.
5186 [Richard Levitte]
5187
5188 *) Constify the BIGNUM routines a little more.
5189 [Richard Levitte]
5190
5191 *) Add the following functions:
5192
5193 ENGINE_load_cswift()
5194 ENGINE_load_chil()
5195 ENGINE_load_atalla()
5196 ENGINE_load_nuron()
5197 ENGINE_load_builtin_engines()
5198
5199 That way, an application can itself choose if external engines that
5200 are built-in in OpenSSL shall ever be used or not. The benefit is
5201 that applications won't have to be linked with libdl or other dso
5202 libraries unless it's really needed.
5203
5204 Changed 'openssl engine' to load all engines on demand.
5205 Changed the engine header files to avoid the duplication of some
5206 declarations (they differed!).
5207 [Richard Levitte]
5208
5209 *) 'openssl engine' can now list capabilities.
5210 [Richard Levitte]
5211
5212 *) Better error reporting in 'openssl engine'.
5213 [Richard Levitte]
5214
5215 *) Never call load_dh_param(NULL) in s_server.
5216 [Bodo Moeller]
5217
5218 *) Add engine application. It can currently list engines by name and
5219 identity, and test if they are actually available.
5220 [Richard Levitte]
5221
5222 *) Improve RPM specification file by forcing symbolic linking and making
5223 sure the installed documentation is also owned by root.root.
5224 [Damien Miller <djm@mindrot.org>]
5225
5226 *) Give the OpenSSL applications more possibilities to make use of
5227 keys (public as well as private) handled by engines.
5228 [Richard Levitte]
5229
5230 *) Add OCSP code that comes from CertCo.
5231 [Richard Levitte]
5232
5233 *) Add VMS support for the Rijndael code.
5234 [Richard Levitte]
5235
5236 *) Added untested support for Nuron crypto accelerator.
5237 [Ben Laurie]
5238
5239 *) Add support for external cryptographic devices. This code was
5240 previously distributed separately as the "engine" branch.
5241 [Geoff Thorpe, Richard Levitte]
5242
5243 *) Rework the filename-translation in the DSO code. It is now possible to
5244 have far greater control over how a "name" is turned into a filename
5245 depending on the operating environment and any oddities about the
5246 different shared library filenames on each system.
5247 [Geoff Thorpe]
5248
5249 *) Support threads on FreeBSD-elf in Configure.
5250 [Richard Levitte]
5251
5252 *) Fix for SHA1 assembly problem with MASM: it produces
5253 warnings about corrupt line number information when assembling
5254 with debugging information. This is caused by the overlapping
5255 of two sections.
5256 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5257
5258 *) NCONF changes.
5259 NCONF_get_number() has no error checking at all. As a replacement,
5260 NCONF_get_number_e() is defined (_e for "error checking") and is
5261 promoted strongly. The old NCONF_get_number is kept around for
5262 binary backward compatibility.
5263 Make it possible for methods to load from something other than a BIO,
5264 by providing a function pointer that is given a name instead of a BIO.
5265 For example, this could be used to load configuration data from an
5266 LDAP server.
5267 [Richard Levitte]
5268
5269 *) Fix for non blocking accept BIOs. Added new I/O special reason
5270 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5271 with non blocking I/O was not possible because no retry code was
5272 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5273 this case.
5274 [Steve Henson]
5275
5276 *) Added the beginnings of Rijndael support.
5277 [Ben Laurie]
5278
5279 *) Fix for bug in DirectoryString mask setting. Add support for
5280 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5281 to allow certificate printing to more controllable, additional
5282 'certopt' option to 'x509' to allow new printing options to be
5283 set.
5284 [Steve Henson]
5285
5286 *) Clean old EAY MD5 hack from e_os.h.
5287 [Richard Levitte]
5288
5289 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5290
5291 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5292 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5293 [Joe Orton, Steve Henson]
5294
5295 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5296
5297 *) Fix additional bug revealed by the NISCC test suite:
5298
5299 Stop bug triggering large recursion when presented with
5300 certain ASN.1 tags (CVE-2003-0851)
5301 [Steve Henson]
5302
5303 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5304
5305 *) Fix various bugs revealed by running the NISCC test suite:
5306
5307 Stop out of bounds reads in the ASN1 code when presented with
5308 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5309
5310 If verify callback ignores invalid public key errors don't try to check
5311 certificate signature with the NULL public key.
5312
5313 [Steve Henson]
5314
5315 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5316 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5317 specifications.
5318 [Steve Henson]
5319
5320 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5321 extra data after the compression methods not only for TLS 1.0
5322 but also for SSL 3.0 (as required by the specification).
5323 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5324
5325 *) Change X509_certificate_type() to mark the key as exported/exportable
5326 when it's 512 *bits* long, not 512 bytes.
5327 [Richard Levitte]
5328
5329 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5330
5331 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5332 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5333 a protocol version number mismatch like a decryption error
5334 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5335 [Bodo Moeller]
5336
5337 *) Turn on RSA blinding by default in the default implementation
5338 to avoid a timing attack. Applications that don't want it can call
5339 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5340 They would be ill-advised to do so in most cases.
5341 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5342
5343 *) Change RSA blinding code so that it works when the PRNG is not
5344 seeded (in this case, the secret RSA exponent is abused as
5345 an unpredictable seed -- if it is not unpredictable, there
5346 is no point in blinding anyway). Make RSA blinding thread-safe
5347 by remembering the creator's thread ID in rsa->blinding and
5348 having all other threads use local one-time blinding factors
5349 (this requires more computation than sharing rsa->blinding, but
5350 avoids excessive locking; and if an RSA object is not shared
5351 between threads, blinding will still be very fast).
5352 [Bodo Moeller]
5353
5354 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5355
5356 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5357 via timing by performing a MAC computation even if incorrrect
5358 block cipher padding has been found. This is a countermeasure
5359 against active attacks where the attacker has to distinguish
5360 between bad padding and a MAC verification error. (CVE-2003-0078)
5361
5362 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5363 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5364 Martin Vuagnoux (EPFL, Ilion)]
5365
5366 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5367
5368 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5369 memory from it's contents. This is done with a counter that will
5370 place alternating values in each byte. This can be used to solve
5371 two issues: 1) the removal of calls to memset() by highly optimizing
5372 compilers, and 2) cleansing with other values than 0, since those can
5373 be read through on certain media, for example a swap space on disk.
5374 [Geoff Thorpe]
5375
5376 *) Bugfix: client side session caching did not work with external caching,
5377 because the session->cipher setting was not restored when reloading
5378 from the external cache. This problem was masked, when
5379 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5380 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5381 [Lutz Jaenicke]
5382
5383 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5384 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5385 [Zeev Lieber <zeev-l@yahoo.com>]
5386
5387 *) Undo an undocumented change introduced in 0.9.6e which caused
5388 repeated calls to OpenSSL_add_all_ciphers() and
5389 OpenSSL_add_all_digests() to be ignored, even after calling
5390 EVP_cleanup().
5391 [Richard Levitte]
5392
5393 *) Change the default configuration reader to deal with last line not
5394 being properly terminated.
5395 [Richard Levitte]
5396
5397 *) Change X509_NAME_cmp() so it applies the special rules on handling
5398 DN values that are of type PrintableString, as well as RDNs of type
5399 emailAddress where the value has the type ia5String.
5400 [stefank@valicert.com via Richard Levitte]
5401
5402 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5403 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5404 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5405 the bitwise-OR of the two for use by the majority of applications
5406 wanting this behaviour, and update the docs. The documented
5407 behaviour and actual behaviour were inconsistent and had been
5408 changing anyway, so this is more a bug-fix than a behavioural
5409 change.
5410 [Geoff Thorpe, diagnosed by Nadav Har'El]
5411
5412 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5413 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5414 [Bodo Moeller]
5415
5416 *) Fix initialization code race conditions in
5417 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5418 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5419 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5420 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5421 ssl2_get_cipher_by_char(),
5422 ssl3_get_cipher_by_char().
5423 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5424
5425 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5426 the cached sessions are flushed, as the remove_cb() might use ex_data
5427 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5428 (see [openssl.org #212]).
5429 [Geoff Thorpe, Lutz Jaenicke]
5430
5431 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5432 length, instead of the encoding length to d2i_ASN1_OBJECT.
5433 [Steve Henson]
5434
5435 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5436
5437 *) [In 0.9.6g-engine release:]
5438 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5439 [Lynn Gazis <lgazis@rainbow.com>]
5440
5441 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5442
5443 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5444 and get fix the header length calculation.
5445 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5446 Alon Kantor <alonk@checkpoint.com> (and others),
5447 Steve Henson]
5448
5449 *) Use proper error handling instead of 'assertions' in buffer
5450 overflow checks added in 0.9.6e. This prevents DoS (the
5451 assertions could call abort()).
5452 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5453
5454 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5455
5456 *) Add various sanity checks to asn1_get_length() to reject
5457 the ASN1 length bytes if they exceed sizeof(long), will appear
5458 negative or the content length exceeds the length of the
5459 supplied buffer.
5460 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5461
5462 *) Fix cipher selection routines: ciphers without encryption had no flags
5463 for the cipher strength set and where therefore not handled correctly
5464 by the selection routines (PR #130).
5465 [Lutz Jaenicke]
5466
5467 *) Fix EVP_dsa_sha macro.
5468 [Nils Larsch]
5469
5470 *) New option
5471 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5472 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5473 that was added in OpenSSL 0.9.6d.
5474
5475 As the countermeasure turned out to be incompatible with some
5476 broken SSL implementations, the new option is part of SSL_OP_ALL.
5477 SSL_OP_ALL is usually employed when compatibility with weird SSL
5478 implementations is desired (e.g. '-bugs' option to 's_client' and
5479 's_server'), so the new option is automatically set in many
5480 applications.
5481 [Bodo Moeller]
5482
5483 *) Changes in security patch:
5484
5485 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5486 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5487 Air Force Materiel Command, USAF, under agreement number
5488 F30602-01-2-0537.
5489
5490 *) Add various sanity checks to asn1_get_length() to reject
5491 the ASN1 length bytes if they exceed sizeof(long), will appear
5492 negative or the content length exceeds the length of the
5493 supplied buffer. (CVE-2002-0659)
5494 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5495
5496 *) Assertions for various potential buffer overflows, not known to
5497 happen in practice.
5498 [Ben Laurie (CHATS)]
5499
5500 *) Various temporary buffers to hold ASCII versions of integers were
5501 too small for 64 bit platforms. (CVE-2002-0655)
5502 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5503
5504 *) Remote buffer overflow in SSL3 protocol - an attacker could
5505 supply an oversized session ID to a client. (CVE-2002-0656)
5506 [Ben Laurie (CHATS)]
5507
5508 *) Remote buffer overflow in SSL2 protocol - an attacker could
5509 supply an oversized client master key. (CVE-2002-0656)
5510 [Ben Laurie (CHATS)]
5511
5512 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5513
5514 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5515 encoded as NULL) with id-dsa-with-sha1.
5516 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5517
5518 *) Check various X509_...() return values in apps/req.c.
5519 [Nils Larsch <nla@trustcenter.de>]
5520
5521 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5522 an end-of-file condition would erronously be flagged, when the CRLF
5523 was just at the end of a processed block. The bug was discovered when
5524 processing data through a buffering memory BIO handing the data to a
5525 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5526 <ptsekov@syntrex.com> and Nedelcho Stanev.
5527 [Lutz Jaenicke]
5528
5529 *) Implement a countermeasure against a vulnerability recently found
5530 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5531 before application data chunks to avoid the use of known IVs
5532 with data potentially chosen by the attacker.
5533 [Bodo Moeller]
5534
5535 *) Fix length checks in ssl3_get_client_hello().
5536 [Bodo Moeller]
5537
5538 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5539 to prevent ssl3_read_internal() from incorrectly assuming that
5540 ssl3_read_bytes() found application data while handshake
5541 processing was enabled when in fact s->s3->in_read_app_data was
5542 merely automatically cleared during the initial handshake.
5543 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5544
5545 *) Fix object definitions for Private and Enterprise: they were not
5546 recognized in their shortname (=lowercase) representation. Extend
5547 obj_dat.pl to issue an error when using undefined keywords instead
5548 of silently ignoring the problem (Svenning Sorensen
5549 <sss@sss.dnsalias.net>).
5550 [Lutz Jaenicke]
5551
5552 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5553 generators, i.e. generators other than 2 and 5. (Previously, the
5554 code did not properly initialise the 'add' and 'rem' values to
5555 BN_generate_prime().)
5556
5557 In the new general case, we do not insist that 'generator' is
5558 actually a primitive root: This requirement is rather pointless;
5559 a generator of the order-q subgroup is just as good, if not
5560 better.
5561 [Bodo Moeller]
5562
5563 *) Map new X509 verification errors to alerts. Discovered and submitted by
5564 Tom Wu <tom@arcot.com>.
5565 [Lutz Jaenicke]
5566
5567 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5568 returning non-zero before the data has been completely received
5569 when using non-blocking I/O.
5570 [Bodo Moeller; problem pointed out by John Hughes]
5571
5572 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5573 [Ben Laurie, Lutz Jaenicke]
5574
5575 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5576 Yoram Zahavi <YoramZ@gilian.com>).
5577 [Lutz Jaenicke]
5578
5579 *) Add information about CygWin 1.3 and on, and preserve proper
5580 configuration for the versions before that.
5581 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5582
5583 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5584 check whether we deal with a copy of a session and do not delete from
5585 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5586 <izhar@checkpoint.com>.
5587 [Lutz Jaenicke]
5588
5589 *) Do not store session data into the internal session cache, if it
5590 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5591 flag is set). Proposed by Aslam <aslam@funk.com>.
5592 [Lutz Jaenicke]
5593
5594 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5595 value is 0.
5596 [Richard Levitte]
5597
5598 *) [In 0.9.6d-engine release:]
5599 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5600 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5601
5602 *) Add the configuration target linux-s390x.
5603 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5604
5605 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5606 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5607 variable as an indication that a ClientHello message has been
5608 received. As the flag value will be lost between multiple
5609 invocations of ssl3_accept when using non-blocking I/O, the
5610 function may not be aware that a handshake has actually taken
5611 place, thus preventing a new session from being added to the
5612 session cache.
5613
5614 To avoid this problem, we now set s->new_session to 2 instead of
5615 using a local variable.
5616 [Lutz Jaenicke, Bodo Moeller]
5617
5618 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5619 if the SSL_R_LENGTH_MISMATCH error is detected.
5620 [Geoff Thorpe, Bodo Moeller]
5621
5622 *) New 'shared_ldflag' column in Configure platform table.
5623 [Richard Levitte]
5624
5625 *) Fix EVP_CIPHER_mode macro.
5626 ["Dan S. Camper" <dan@bti.net>]
5627
5628 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5629 type, we must throw them away by setting rr->length to 0.
5630 [D P Chang <dpc@qualys.com>]
5631
5632 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5633
5634 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5635 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5636 worked incorrectly for those cases where range = 10..._2 and
5637 3*range is two bits longer than range.)
5638 [Bodo Moeller]
5639
5640 *) Only add signing time to PKCS7 structures if it is not already
5641 present.
5642 [Steve Henson]
5643
5644 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5645 OBJ_ld_ce should be OBJ_id_ce.
5646 Also some ip-pda OIDs in crypto/objects/objects.txt were
5647 incorrect (cf. RFC 3039).
5648 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5649
5650 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5651 returns early because it has nothing to do.
5652 [Andy Schneider <andy.schneider@bjss.co.uk>]
5653
5654 *) [In 0.9.6c-engine release:]
5655 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5656 [Andy Schneider <andy.schneider@bjss.co.uk>]
5657
5658 *) [In 0.9.6c-engine release:]
5659 Add support for Cryptographic Appliance's keyserver technology.
5660 (Use engine 'keyclient')
5661 [Cryptographic Appliances and Geoff Thorpe]
5662
5663 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5664 is called via tools/c89.sh because arguments have to be
5665 rearranged (all '-L' options must appear before the first object
5666 modules).
5667 [Richard Shapiro <rshapiro@abinitio.com>]
5668
5669 *) [In 0.9.6c-engine release:]
5670 Add support for Broadcom crypto accelerator cards, backported
5671 from 0.9.7.
5672 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5673
5674 *) [In 0.9.6c-engine release:]
5675 Add support for SureWare crypto accelerator cards from
5676 Baltimore Technologies. (Use engine 'sureware')
5677 [Baltimore Technologies and Mark Cox]
5678
5679 *) [In 0.9.6c-engine release:]
5680 Add support for crypto accelerator cards from Accelerated
5681 Encryption Processing, www.aep.ie. (Use engine 'aep')
5682 [AEP Inc. and Mark Cox]
5683
5684 *) Add a configuration entry for gcc on UnixWare.
5685 [Gary Benson <gbenson@redhat.com>]
5686
5687 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5688 messages are stored in a single piece (fixed-length part and
5689 variable-length part combined) and fix various bugs found on the way.
5690 [Bodo Moeller]
5691
5692 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5693 instead. BIO_gethostbyname() does not know what timeouts are
5694 appropriate, so entries would stay in cache even when they have
5695 become invalid.
5696 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5697
5698 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5699 faced with a pathologically small ClientHello fragment that does
5700 not contain client_version: Instead of aborting with an error,
5701 simply choose the highest available protocol version (i.e.,
5702 TLS 1.0 unless it is disabled). In practice, ClientHello
5703 messages are never sent like this, but this change gives us
5704 strictly correct behaviour at least for TLS.
5705 [Bodo Moeller]
5706
5707 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5708 never resets s->method to s->ctx->method when called from within
5709 one of the SSL handshake functions.
5710 [Bodo Moeller; problem pointed out by Niko Baric]
5711
5712 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5713 (sent using the client's version number) if client_version is
5714 smaller than the protocol version in use. Also change
5715 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5716 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5717 the client will at least see that alert.
5718 [Bodo Moeller]
5719
5720 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5721 correctly.
5722 [Bodo Moeller]
5723
5724 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5725 client receives HelloRequest while in a handshake.
5726 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5727
5728 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5729 should end in 'break', not 'goto end' which circuments various
5730 cleanups done in state SSL_ST_OK. But session related stuff
5731 must be disabled for SSL_ST_OK in the case that we just sent a
5732 HelloRequest.
5733
5734 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5735 before just sending a HelloRequest.
5736 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5737
5738 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5739 reveal whether illegal block cipher padding was found or a MAC
5740 verification error occured. (Neither SSLerr() codes nor alerts
5741 are directly visible to potential attackers, but the information
5742 may leak via logfiles.)
5743
5744 Similar changes are not required for the SSL 2.0 implementation
5745 because the number of padding bytes is sent in clear for SSL 2.0,
5746 and the extra bytes are just ignored. However ssl/s2_pkt.c
5747 failed to verify that the purported number of padding bytes is in
5748 the legal range.
5749 [Bodo Moeller]
5750
5751 *) Add OpenUNIX-8 support including shared libraries
5752 (Boyd Lynn Gerber <gerberb@zenez.com>).
5753 [Lutz Jaenicke]
5754
5755 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5756 'wristwatch attack' using huge encoding parameters (cf.
5757 James H. Manger's CRYPTO 2001 paper). Note that the
5758 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5759 encoding parameters and hence was not vulnerable.
5760 [Bodo Moeller]
5761
5762 *) BN_sqr() bug fix.
5763 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5764
5765 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5766 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5767 followed by modular reduction.
5768 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5769
5770 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5771 equivalent based on BN_pseudo_rand() instead of BN_rand().
5772 [Bodo Moeller]
5773
5774 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5775 This function was broken, as the check for a new client hello message
5776 to handle SGC did not allow these large messages.
5777 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5778 [Lutz Jaenicke]
5779
5780 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5781 [Lutz Jaenicke]
5782
5783 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5784 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5785 [Lutz Jaenicke]
5786
5787 *) Rework the configuration and shared library support for Tru64 Unix.
5788 The configuration part makes use of modern compiler features and
5789 still retains old compiler behavior for those that run older versions
5790 of the OS. The shared library support part includes a variant that
5791 uses the RPATH feature, and is available through the special
5792 configuration target "alpha-cc-rpath", which will never be selected
5793 automatically.
5794 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5795
5796 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5797 with the same message size as in ssl3_get_certificate_request().
5798 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5799 messages might inadvertently be reject as too long.
5800 [Petr Lampa <lampa@fee.vutbr.cz>]
5801
5802 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5803 [Andy Polyakov]
5804
5805 *) Modified SSL library such that the verify_callback that has been set
5806 specificly for an SSL object with SSL_set_verify() is actually being
5807 used. Before the change, a verify_callback set with this function was
5808 ignored and the verify_callback() set in the SSL_CTX at the time of
5809 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5810 to allow the necessary settings.
5811 [Lutz Jaenicke]
5812
5813 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5814 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5815 done automatically (in contradiction to the requirements of the C
5816 standard). This made problems when used from OpenSSH.
5817 [Lutz Jaenicke]
5818
5819 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5820 dh->length and always used
5821
5822 BN_rand_range(priv_key, dh->p).
5823
5824 BN_rand_range() is not necessary for Diffie-Hellman, and this
5825 specific range makes Diffie-Hellman unnecessarily inefficient if
5826 dh->length (recommended exponent length) is much smaller than the
5827 length of dh->p. We could use BN_rand_range() if the order of
5828 the subgroup was stored in the DH structure, but we only have
5829 dh->length.
5830
5831 So switch back to
5832
5833 BN_rand(priv_key, l, ...)
5834
5835 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5836 otherwise.
5837 [Bodo Moeller]
5838
5839 *) In
5840
5841 RSA_eay_public_encrypt
5842 RSA_eay_private_decrypt
5843 RSA_eay_private_encrypt (signing)
5844 RSA_eay_public_decrypt (signature verification)
5845
5846 (default implementations for RSA_public_encrypt,
5847 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5848 always reject numbers >= n.
5849 [Bodo Moeller]
5850
5851 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5852 to synchronize access to 'locking_thread'. This is necessary on
5853 systems where access to 'locking_thread' (an 'unsigned long'
5854 variable) is not atomic.
5855 [Bodo Moeller]
5856
5857 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5858 *before* setting the 'crypto_lock_rand' flag. The previous code had
5859 a race condition if 0 is a valid thread ID.
5860 [Travis Vitek <vitek@roguewave.com>]
5861
5862 *) Add support for shared libraries under Irix.
5863 [Albert Chin-A-Young <china@thewrittenword.com>]
5864
5865 *) Add configuration option to build on Linux on both big-endian and
5866 little-endian MIPS.
5867 [Ralf Baechle <ralf@uni-koblenz.de>]
5868
5869 *) Add the possibility to create shared libraries on HP-UX.
5870 [Richard Levitte]
5871
5872 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5873
5874 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5875 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5876 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5877 PRNG state recovery was possible based on the output of
5878 one PRNG request appropriately sized to gain knowledge on
5879 'md' followed by enough consecutive 1-byte PRNG requests
5880 to traverse all of 'state'.
5881
5882 1. When updating 'md_local' (the current thread's copy of 'md')
5883 during PRNG output generation, hash all of the previous
5884 'md_local' value, not just the half used for PRNG output.
5885
5886 2. Make the number of bytes from 'state' included into the hash
5887 independent from the number of PRNG bytes requested.
5888
5889 The first measure alone would be sufficient to avoid
5890 Markku-Juhani's attack. (Actually it had never occurred
5891 to me that the half of 'md_local' used for chaining was the
5892 half from which PRNG output bytes were taken -- I had always
5893 assumed that the secret half would be used.) The second
5894 measure makes sure that additional data from 'state' is never
5895 mixed into 'md_local' in small portions; this heuristically
5896 further strengthens the PRNG.
5897 [Bodo Moeller]
5898
5899 *) Fix crypto/bn/asm/mips3.s.
5900 [Andy Polyakov]
5901
5902 *) When only the key is given to "enc", the IV is undefined. Print out
5903 an error message in this case.
5904 [Lutz Jaenicke]
5905
5906 *) Handle special case when X509_NAME is empty in X509 printing routines.
5907 [Steve Henson]
5908
5909 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5910 positive and less than q.
5911 [Bodo Moeller]
5912
5913 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5914 used: it isn't thread safe and the add_lock_callback should handle
5915 that itself.
5916 [Paul Rose <Paul.Rose@bridge.com>]
5917
5918 *) Verify that incoming data obeys the block size in
5919 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5920 [Bodo Moeller]
5921
5922 *) Fix OAEP check.
5923 [Ulf Möller, Bodo Möller]
5924
5925 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5926 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5927 when fixing the server behaviour for backwards-compatible 'client
5928 hello' messages. (Note that the attack is impractical against
5929 SSL 3.0 and TLS 1.0 anyway because length and version checking
5930 means that the probability of guessing a valid ciphertext is
5931 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5932 paper.)
5933
5934 Before 0.9.5, the countermeasure (hide the error by generating a
5935 random 'decryption result') did not work properly because
5936 ERR_clear_error() was missing, meaning that SSL_get_error() would
5937 detect the supposedly ignored error.
5938
5939 Both problems are now fixed.
5940 [Bodo Moeller]
5941
5942 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5943 (previously it was 1024).
5944 [Bodo Moeller]
5945
5946 *) Fix for compatibility mode trust settings: ignore trust settings
5947 unless some valid trust or reject settings are present.
5948 [Steve Henson]
5949
5950 *) Fix for blowfish EVP: its a variable length cipher.
5951 [Steve Henson]
5952
5953 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5954 parameters in DSA public key structures and return an error in the
5955 DSA routines if parameters are absent.
5956 [Steve Henson]
5957
5958 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5959 in the current directory if neither $RANDFILE nor $HOME was set.
5960 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5961 caused some confusion to Windows users who haven't defined $HOME.
5962 Thus RAND_file_name() is changed again: e_os.h can define a
5963 DEFAULT_HOME, which will be used if $HOME is not set.
5964 For Windows, we use "C:"; on other platforms, we still require
5965 environment variables.
5966
5967 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5968 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5969 having multiple threads call RAND_poll() concurrently.
5970 [Bodo Moeller]
5971
5972 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5973 combination of a flag and a thread ID variable.
5974 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5975 flag), *other* threads can enter ssleay_add_bytes without obeying
5976 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5977 that they do not hold after the first thread unsets add_do_not_lock).
5978 [Bodo Moeller]
5979
5980 *) Change bctest again: '-x' expressions are not available in all
5981 versions of 'test'.
5982 [Bodo Moeller]
5983
5984 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5985
5986 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5987 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5988
5989 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5990 the default extension for executables, if any. Also, make the perl
5991 scripts that use symlink() to test if it really exists and use "cp"
5992 if it doesn't. All this made OpenSSL compilable and installable in
5993 CygWin.
5994 [Richard Levitte]
5995
5996 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5997 If SEQUENCE is length is indefinite just set c->slen to the total
5998 amount of data available.
5999 [Steve Henson, reported by shige@FreeBSD.org]
6000 [This change does not apply to 0.9.7.]
6001
6002 *) Change bctest to avoid here-documents inside command substitution
6003 (workaround for FreeBSD /bin/sh bug).
6004 For compatibility with Ultrix, avoid shell functions (introduced
6005 in the bctest version that searches along $PATH).
6006 [Bodo Moeller]
6007
6008 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6009 with des_encrypt() defined on some operating systems, like Solaris
6010 and UnixWare.
6011 [Richard Levitte]
6012
6013 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6014 On the Importance of Eliminating Errors in Cryptographic
6015 Computations, J. Cryptology 14 (2001) 2, 101-119,
6016 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6017 [Ulf Moeller]
6018
6019 *) MIPS assembler BIGNUM division bug fix.
6020 [Andy Polyakov]
6021
6022 *) Disabled incorrect Alpha assembler code.
6023 [Richard Levitte]
6024
6025 *) Fix PKCS#7 decode routines so they correctly update the length
6026 after reading an EOC for the EXPLICIT tag.
6027 [Steve Henson]
6028 [This change does not apply to 0.9.7.]
6029
6030 *) Fix bug in PKCS#12 key generation routines. This was triggered
6031 if a 3DES key was generated with a 0 initial byte. Include
6032 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6033 (but broken) behaviour.
6034 [Steve Henson]
6035
6036 *) Enhance bctest to search for a working bc along $PATH and print
6037 it when found.
6038 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6039
6040 *) Fix memory leaks in err.c: free err_data string if necessary;
6041 don't write to the wrong index in ERR_set_error_data.
6042 [Bodo Moeller]
6043
6044 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6045 did not exist.
6046 [Bodo Moeller]
6047
6048 *) Replace rdtsc with _emit statements for VC++ version 5.
6049 [Jeremy Cooper <jeremy@baymoo.org>]
6050
6051 *) Make it possible to reuse SSLv2 sessions.
6052 [Richard Levitte]
6053
6054 *) In copy_email() check for >= 0 as a return value for
6055 X509_NAME_get_index_by_NID() since 0 is a valid index.
6056 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6057
6058 *) Avoid coredump with unsupported or invalid public keys by checking if
6059 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6060 PKCS7_verify() fails with non detached data.
6061 [Steve Henson]
6062
6063 *) Don't use getenv in library functions when run as setuid/setgid.
6064 New function OPENSSL_issetugid().
6065 [Ulf Moeller]
6066
6067 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6068 due to incorrect handling of multi-threading:
6069
6070 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6071
6072 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6073
6074 3. Count how many times MemCheck_off() has been called so that
6075 nested use can be treated correctly. This also avoids
6076 inband-signalling in the previous code (which relied on the
6077 assumption that thread ID 0 is impossible).
6078 [Bodo Moeller]
6079
6080 *) Add "-rand" option also to s_client and s_server.
6081 [Lutz Jaenicke]
6082
6083 *) Fix CPU detection on Irix 6.x.
6084 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6085 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6086
6087 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6088 was empty.
6089 [Steve Henson]
6090 [This change does not apply to 0.9.7.]
6091
6092 *) Use the cached encoding of an X509_NAME structure rather than
6093 copying it. This is apparently the reason for the libsafe "errors"
6094 but the code is actually correct.
6095 [Steve Henson]
6096
6097 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6098 Bleichenbacher's DSA attack.
6099 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6100 to be set and top=0 forces the highest bit to be set; top=-1 is new
6101 and leaves the highest bit random.
6102 [Ulf Moeller, Bodo Moeller]
6103
6104 *) In the NCONF_...-based implementations for CONF_... queries
6105 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6106 a temporary CONF structure with the data component set to NULL
6107 (which gives segmentation faults in lh_retrieve).
6108 Instead, use NULL for the CONF pointer in CONF_get_string and
6109 CONF_get_number (which may use environment variables) and directly
6110 return NULL from CONF_get_section.
6111 [Bodo Moeller]
6112
6113 *) Fix potential buffer overrun for EBCDIC.
6114 [Ulf Moeller]
6115
6116 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6117 keyUsage if basicConstraints absent for a CA.
6118 [Steve Henson]
6119
6120 *) Make SMIME_write_PKCS7() write mail header values with a format that
6121 is more generally accepted (no spaces before the semicolon), since
6122 some programs can't parse those values properly otherwise. Also make
6123 sure BIO's that break lines after each write do not create invalid
6124 headers.
6125 [Richard Levitte]
6126
6127 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6128 macros previously used would not encode an empty SEQUENCE OF
6129 and break the signature.
6130 [Steve Henson]
6131 [This change does not apply to 0.9.7.]
6132
6133 *) Zero the premaster secret after deriving the master secret in
6134 DH ciphersuites.
6135 [Steve Henson]
6136
6137 *) Add some EVP_add_digest_alias registrations (as found in
6138 OpenSSL_add_all_digests()) to SSL_library_init()
6139 aka OpenSSL_add_ssl_algorithms(). This provides improved
6140 compatibility with peers using X.509 certificates
6141 with unconventional AlgorithmIdentifier OIDs.
6142 [Bodo Moeller]
6143
6144 *) Fix for Irix with NO_ASM.
6145 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6146
6147 *) ./config script fixes.
6148 [Ulf Moeller, Richard Levitte]
6149
6150 *) Fix 'openssl passwd -1'.
6151 [Bodo Moeller]
6152
6153 *) Change PKCS12_key_gen_asc() so it can cope with non null
6154 terminated strings whose length is passed in the passlen
6155 parameter, for example from PEM callbacks. This was done
6156 by adding an extra length parameter to asc2uni().
6157 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6158
6159 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6160 call failed, free the DSA structure.
6161 [Bodo Moeller]
6162
6163 *) Fix to uni2asc() to cope with zero length Unicode strings.
6164 These are present in some PKCS#12 files.
6165 [Steve Henson]
6166
6167 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6168 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6169 when writing a 32767 byte record.
6170 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6171
6172 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6173 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6174
6175 (RSA objects have a reference count access to which is protected
6176 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6177 so they are meant to be shared between threads.)
6178 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6179 "Reddie, Steven" <Steven.Reddie@ca.com>]
6180
6181 *) Fix a deadlock in CRYPTO_mem_leaks().
6182 [Bodo Moeller]
6183
6184 *) Use better test patterns in bntest.
6185 [Ulf Möller]
6186
6187 *) rand_win.c fix for Borland C.
6188 [Ulf Möller]
6189
6190 *) BN_rshift bugfix for n == 0.
6191 [Bodo Moeller]
6192
6193 *) Add a 'bctest' script that checks for some known 'bc' bugs
6194 so that 'make test' does not abort just because 'bc' is broken.
6195 [Bodo Moeller]
6196
6197 *) Store verify_result within SSL_SESSION also for client side to
6198 avoid potential security hole. (Re-used sessions on the client side
6199 always resulted in verify_result==X509_V_OK, not using the original
6200 result of the server certificate verification.)
6201 [Lutz Jaenicke]
6202
6203 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6204 SSL3_RT_APPLICATION_DATA, return 0.
6205 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6206 [Bodo Moeller]
6207
6208 *) Fix SSL_peek:
6209 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6210 releases, have been re-implemented by renaming the previous
6211 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6212 and ssl3_read_internal, respectively, and adding 'peek' parameters
6213 to them. The new ssl[23]_{read,peek} functions are calls to
6214 ssl[23]_read_internal with the 'peek' flag set appropriately.
6215 A 'peek' parameter has also been added to ssl3_read_bytes, which
6216 does the actual work for ssl3_read_internal.
6217 [Bodo Moeller]
6218
6219 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6220 the method-specific "init()" handler. Also clean up ex_data after
6221 calling the method-specific "finish()" handler. Previously, this was
6222 happening the other way round.
6223 [Geoff Thorpe]
6224
6225 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6226 The previous value, 12, was not always sufficient for BN_mod_exp().
6227 [Bodo Moeller]
6228
6229 *) Make sure that shared libraries get the internal name engine with
6230 the full version number and not just 0. This should mark the
6231 shared libraries as not backward compatible. Of course, this should
6232 be changed again when we can guarantee backward binary compatibility.
6233 [Richard Levitte]
6234
6235 *) Fix typo in get_cert_by_subject() in by_dir.c
6236 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6237
6238 *) Rework the system to generate shared libraries:
6239
6240 - Make note of the expected extension for the shared libraries and
6241 if there is a need for symbolic links from for example libcrypto.so.0
6242 to libcrypto.so.0.9.7. There is extended info in Configure for
6243 that.
6244
6245 - Make as few rebuilds of the shared libraries as possible.
6246
6247 - Still avoid linking the OpenSSL programs with the shared libraries.
6248
6249 - When installing, install the shared libraries separately from the
6250 static ones.
6251 [Richard Levitte]
6252
6253 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6254
6255 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6256 and not in SSL_clear because the latter is also used by the
6257 accept/connect functions; previously, the settings made by
6258 SSL_set_read_ahead would be lost during the handshake.
6259 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6260
6261 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6262 Previously, it would create entries for disableed algorithms no
6263 matter what.
6264 [Richard Levitte]
6265
6266 *) Added several new manual pages for SSL_* function.
6267 [Lutz Jaenicke]
6268
6269 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6270
6271 *) In ssl23_get_client_hello, generate an error message when faced
6272 with an initial SSL 3.0/TLS record that is too small to contain the
6273 first two bytes of the ClientHello message, i.e. client_version.
6274 (Note that this is a pathologic case that probably has never happened
6275 in real life.) The previous approach was to use the version number
6276 from the record header as a substitute; but our protocol choice
6277 should not depend on that one because it is not authenticated
6278 by the Finished messages.
6279 [Bodo Moeller]
6280
6281 *) More robust randomness gathering functions for Windows.
6282 [Jeffrey Altman <jaltman@columbia.edu>]
6283
6284 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6285 not set then we don't setup the error code for issuer check errors
6286 to avoid possibly overwriting other errors which the callback does
6287 handle. If an application does set the flag then we assume it knows
6288 what it is doing and can handle the new informational codes
6289 appropriately.
6290 [Steve Henson]
6291
6292 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6293 a general "ANY" type, as such it should be able to decode anything
6294 including tagged types. However it didn't check the class so it would
6295 wrongly interpret tagged types in the same way as their universal
6296 counterpart and unknown types were just rejected. Changed so that the
6297 tagged and unknown types are handled in the same way as a SEQUENCE:
6298 that is the encoding is stored intact. There is also a new type
6299 "V_ASN1_OTHER" which is used when the class is not universal, in this
6300 case we have no idea what the actual type is so we just lump them all
6301 together.
6302 [Steve Henson]
6303
6304 *) On VMS, stdout may very well lead to a file that is written to
6305 in a record-oriented fashion. That means that every write() will
6306 write a separate record, which will be read separately by the
6307 programs trying to read from it. This can be very confusing.
6308
6309 The solution is to put a BIO filter in the way that will buffer
6310 text until a linefeed is reached, and then write everything a
6311 line at a time, so every record written will be an actual line,
6312 not chunks of lines and not (usually doesn't happen, but I've
6313 seen it once) several lines in one record. BIO_f_linebuffer() is
6314 the answer.
6315
6316 Currently, it's a VMS-only method, because that's where it has
6317 been tested well enough.
6318 [Richard Levitte]
6319
6320 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6321 it can return incorrect results.
6322 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6323 but it was in 0.9.6-beta[12].)
6324 [Bodo Moeller]
6325
6326 *) Disable the check for content being present when verifying detached
6327 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6328 include zero length content when signing messages.
6329 [Steve Henson]
6330
6331 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6332 BIO_ctrl (for BIO pairs).
6333 [Bodo Möller]
6334
6335 *) Add DSO method for VMS.
6336 [Richard Levitte]
6337
6338 *) Bug fix: Montgomery multiplication could produce results with the
6339 wrong sign.
6340 [Ulf Möller]
6341
6342 *) Add RPM specification openssl.spec and modify it to build three
6343 packages. The default package contains applications, application
6344 documentation and run-time libraries. The devel package contains
6345 include files, static libraries and function documentation. The
6346 doc package contains the contents of the doc directory. The original
6347 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6348 [Richard Levitte]
6349
6350 *) Add a large number of documentation files for many SSL routines.
6351 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6352
6353 *) Add a configuration entry for Sony News 4.
6354 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6355
6356 *) Don't set the two most significant bits to one when generating a
6357 random number < q in the DSA library.
6358 [Ulf Möller]
6359
6360 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6361 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6362 the underlying transport is blocking) if a handshake took place.
6363 (The default behaviour is needed by applications such as s_client
6364 and s_server that use select() to determine when to use SSL_read;
6365 but for applications that know in advance when to expect data, it
6366 just makes things more complicated.)
6367 [Bodo Moeller]
6368
6369 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6370 from EGD.
6371 [Ben Laurie]
6372
6373 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6374 work better on such systems.
6375 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6376
6377 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6378 Update PKCS12_parse() so it copies the friendlyName and the
6379 keyid to the certificates aux info.
6380 [Steve Henson]
6381
6382 *) Fix bug in PKCS7_verify() which caused an infinite loop
6383 if there was more than one signature.
6384 [Sven Uszpelkat <su@celocom.de>]
6385
6386 *) Major change in util/mkdef.pl to include extra information
6387 about each symbol, as well as presentig variables as well
6388 as functions. This change means that there's n more need
6389 to rebuild the .num files when some algorithms are excluded.
6390 [Richard Levitte]
6391
6392 *) Allow the verify time to be set by an application,
6393 rather than always using the current time.
6394 [Steve Henson]
6395
6396 *) Phase 2 verify code reorganisation. The certificate
6397 verify code now looks up an issuer certificate by a
6398 number of criteria: subject name, authority key id
6399 and key usage. It also verifies self signed certificates
6400 by the same criteria. The main comparison function is
6401 X509_check_issued() which performs these checks.
6402
6403 Lot of changes were necessary in order to support this
6404 without completely rewriting the lookup code.
6405
6406 Authority and subject key identifier are now cached.
6407
6408 The LHASH 'certs' is X509_STORE has now been replaced
6409 by a STACK_OF(X509_OBJECT). This is mainly because an
6410 LHASH can't store or retrieve multiple objects with
6411 the same hash value.
6412
6413 As a result various functions (which were all internal
6414 use only) have changed to handle the new X509_STORE
6415 structure. This will break anything that messed round
6416 with X509_STORE internally.
6417
6418 The functions X509_STORE_add_cert() now checks for an
6419 exact match, rather than just subject name.
6420
6421 The X509_STORE API doesn't directly support the retrieval
6422 of multiple certificates matching a given criteria, however
6423 this can be worked round by performing a lookup first
6424 (which will fill the cache with candidate certificates)
6425 and then examining the cache for matches. This is probably
6426 the best we can do without throwing out X509_LOOKUP
6427 entirely (maybe later...).
6428
6429 The X509_VERIFY_CTX structure has been enhanced considerably.
6430
6431 All certificate lookup operations now go via a get_issuer()
6432 callback. Although this currently uses an X509_STORE it
6433 can be replaced by custom lookups. This is a simple way
6434 to bypass the X509_STORE hackery necessary to make this
6435 work and makes it possible to use more efficient techniques
6436 in future. A very simple version which uses a simple
6437 STACK for its trusted certificate store is also provided
6438 using X509_STORE_CTX_trusted_stack().
6439
6440 The verify_cb() and verify() callbacks now have equivalents
6441 in the X509_STORE_CTX structure.
6442
6443 X509_STORE_CTX also has a 'flags' field which can be used
6444 to customise the verify behaviour.
6445 [Steve Henson]
6446
6447 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6448 excludes S/MIME capabilities.
6449 [Steve Henson]
6450
6451 *) When a certificate request is read in keep a copy of the
6452 original encoding of the signed data and use it when outputing
6453 again. Signatures then use the original encoding rather than
6454 a decoded, encoded version which may cause problems if the
6455 request is improperly encoded.
6456 [Steve Henson]
6457
6458 *) For consistency with other BIO_puts implementations, call
6459 buffer_write(b, ...) directly in buffer_puts instead of calling
6460 BIO_write(b, ...).
6461
6462 In BIO_puts, increment b->num_write as in BIO_write.
6463 [Peter.Sylvester@EdelWeb.fr]
6464
6465 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6466 BN_zero, we may not return a BIGNUM with an array consisting of
6467 words set to zero.)
6468 [Bodo Moeller]
6469
6470 *) Avoid calling abort() from within the library when problems are
6471 detected, except if preprocessor symbols have been defined
6472 (such as REF_CHECK, BN_DEBUG etc.).
6473 [Bodo Moeller]
6474
6475 *) New openssl application 'rsautl'. This utility can be
6476 used for low level RSA operations. DER public key
6477 BIO/fp routines also added.
6478 [Steve Henson]
6479
6480 *) New Configure entry and patches for compiling on QNX 4.
6481 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6482
6483 *) A demo state-machine implementation was sponsored by
6484 Nuron (http://www.nuron.com/) and is now available in
6485 demos/state_machine.
6486 [Ben Laurie]
6487
6488 *) New options added to the 'dgst' utility for signature
6489 generation and verification.
6490 [Steve Henson]
6491
6492 *) Unrecognized PKCS#7 content types are now handled via a
6493 catch all ASN1_TYPE structure. This allows unsupported
6494 types to be stored as a "blob" and an application can
6495 encode and decode it manually.
6496 [Steve Henson]
6497
6498 *) Fix various signed/unsigned issues to make a_strex.c
6499 compile under VC++.
6500 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6501
6502 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6503 length if passed a buffer. ASN1_INTEGER_to_BN failed
6504 if passed a NULL BN and its argument was negative.
6505 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6506
6507 *) Modification to PKCS#7 encoding routines to output definite
6508 length encoding. Since currently the whole structures are in
6509 memory there's not real point in using indefinite length
6510 constructed encoding. However if OpenSSL is compiled with
6511 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6512 [Steve Henson]
6513
6514 *) Added BIO_vprintf() and BIO_vsnprintf().
6515 [Richard Levitte]
6516
6517 *) Added more prefixes to parse for in the the strings written
6518 through a logging bio, to cover all the levels that are available
6519 through syslog. The prefixes are now:
6520
6521 PANIC, EMERG, EMR => LOG_EMERG
6522 ALERT, ALR => LOG_ALERT
6523 CRIT, CRI => LOG_CRIT
6524 ERROR, ERR => LOG_ERR
6525 WARNING, WARN, WAR => LOG_WARNING
6526 NOTICE, NOTE, NOT => LOG_NOTICE
6527 INFO, INF => LOG_INFO
6528 DEBUG, DBG => LOG_DEBUG
6529
6530 and as before, if none of those prefixes are present at the
6531 beginning of the string, LOG_ERR is chosen.
6532
6533 On Win32, the LOG_* levels are mapped according to this:
6534
6535 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6536 LOG_WARNING => EVENTLOG_WARNING_TYPE
6537 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6538
6539 [Richard Levitte]
6540
6541 *) Made it possible to reconfigure with just the configuration
6542 argument "reconf" or "reconfigure". The command line arguments
6543 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6544 and are retrieved from there when reconfiguring.
6545 [Richard Levitte]
6546
6547 *) MD4 implemented.
6548 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6549
6550 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6551 [Richard Levitte]
6552
6553 *) The obj_dat.pl script was messing up the sorting of object
6554 names. The reason was that it compared the quoted version
6555 of strings as a result "OCSP" > "OCSP Signing" because
6556 " > SPACE. Changed script to store unquoted versions of
6557 names and add quotes on output. It was also omitting some
6558 names from the lookup table if they were given a default
6559 value (that is if SN is missing it is given the same
6560 value as LN and vice versa), these are now added on the
6561 grounds that if an object has a name we should be able to
6562 look it up. Finally added warning output when duplicate
6563 short or long names are found.
6564 [Steve Henson]
6565
6566 *) Changes needed for Tandem NSK.
6567 [Scott Uroff <scott@xypro.com>]
6568
6569 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6570 RSA_padding_check_SSLv23(), special padding was never detected
6571 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6572 version rollback attacks was not effective.
6573
6574 In s23_clnt.c, don't use special rollback-attack detection padding
6575 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6576 client; similarly, in s23_srvr.c, don't do the rollback check if
6577 SSL 2.0 is the only protocol enabled in the server.
6578 [Bodo Moeller]
6579
6580 *) Make it possible to get hexdumps of unprintable data with 'openssl
6581 asn1parse'. By implication, the functions ASN1_parse_dump() and
6582 BIO_dump_indent() are added.
6583 [Richard Levitte]
6584
6585 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6586 these print out strings and name structures based on various
6587 flags including RFC2253 support and proper handling of
6588 multibyte characters. Added options to the 'x509' utility
6589 to allow the various flags to be set.
6590 [Steve Henson]
6591
6592 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6593 Also change the functions X509_cmp_current_time() and
6594 X509_gmtime_adj() work with an ASN1_TIME structure,
6595 this will enable certificates using GeneralizedTime in validity
6596 dates to be checked.
6597 [Steve Henson]
6598
6599 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6600 negative public key encodings) on by default,
6601 NO_NEG_PUBKEY_BUG can be set to disable it.
6602 [Steve Henson]
6603
6604 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6605 content octets. An i2c_ASN1_OBJECT is unnecessary because
6606 the encoding can be trivially obtained from the structure.
6607 [Steve Henson]
6608
6609 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6610 not read locks (CRYPTO_r_[un]lock).
6611 [Bodo Moeller]
6612
6613 *) A first attempt at creating official support for shared
6614 libraries through configuration. I've kept it so the
6615 default is static libraries only, and the OpenSSL programs
6616 are always statically linked for now, but there are
6617 preparations for dynamic linking in place.
6618 This has been tested on Linux and Tru64.
6619 [Richard Levitte]
6620
6621 *) Randomness polling function for Win9x, as described in:
6622 Peter Gutmann, Software Generation of Practically Strong
6623 Random Numbers.
6624 [Ulf Möller]
6625
6626 *) Fix so PRNG is seeded in req if using an already existing
6627 DSA key.
6628 [Steve Henson]
6629
6630 *) New options to smime application. -inform and -outform
6631 allow alternative formats for the S/MIME message including
6632 PEM and DER. The -content option allows the content to be
6633 specified separately. This should allow things like Netscape
6634 form signing output easier to verify.
6635 [Steve Henson]
6636
6637 *) Fix the ASN1 encoding of tags using the 'long form'.
6638 [Steve Henson]
6639
6640 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6641 STRING types. These convert content octets to and from the
6642 underlying type. The actual tag and length octets are
6643 already assumed to have been read in and checked. These
6644 are needed because all other string types have virtually
6645 identical handling apart from the tag. By having versions
6646 of the ASN1 functions that just operate on content octets
6647 IMPLICIT tagging can be handled properly. It also allows
6648 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6649 and ASN1_INTEGER are identical apart from the tag.
6650 [Steve Henson]
6651
6652 *) Change the handling of OID objects as follows:
6653
6654 - New object identifiers are inserted in objects.txt, following
6655 the syntax given in objects.README.
6656 - objects.pl is used to process obj_mac.num and create a new
6657 obj_mac.h.
6658 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6659 obj_mac.h.
6660
6661 This is currently kind of a hack, and the perl code in objects.pl
6662 isn't very elegant, but it works as I intended. The simplest way
6663 to check that it worked correctly is to look in obj_dat.h and
6664 check the array nid_objs and make sure the objects haven't moved
6665 around (this is important!). Additions are OK, as well as
6666 consistent name changes.
6667 [Richard Levitte]
6668
6669 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6670 [Bodo Moeller]
6671
6672 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6673 The given file adds to whatever has already been seeded into the
6674 random pool through the RANDFILE configuration file option or
6675 environment variable, or the default random state file.
6676 [Richard Levitte]
6677
6678 *) mkstack.pl now sorts each macro group into lexical order.
6679 Previously the output order depended on the order the files
6680 appeared in the directory, resulting in needless rewriting
6681 of safestack.h .
6682 [Steve Henson]
6683
6684 *) Patches to make OpenSSL compile under Win32 again. Mostly
6685 work arounds for the VC++ problem that it treats func() as
6686 func(void). Also stripped out the parts of mkdef.pl that
6687 added extra typesafe functions: these no longer exist.
6688 [Steve Henson]
6689
6690 *) Reorganisation of the stack code. The macros are now all
6691 collected in safestack.h . Each macro is defined in terms of
6692 a "stack macro" of the form SKM_<name>(type, a, b). The
6693 DEBUG_SAFESTACK is now handled in terms of function casts,
6694 this has the advantage of retaining type safety without the
6695 use of additional functions. If DEBUG_SAFESTACK is not defined
6696 then the non typesafe macros are used instead. Also modified the
6697 mkstack.pl script to handle the new form. Needs testing to see
6698 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6699 the default if no major problems. Similar behaviour for ASN1_SET_OF
6700 and PKCS12_STACK_OF.
6701 [Steve Henson]
6702
6703 *) When some versions of IIS use the 'NET' form of private key the
6704 key derivation algorithm is different. Normally MD5(password) is
6705 used as a 128 bit RC4 key. In the modified case
6706 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6707 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6708 as the old Netscape_RSA functions except they have an additional
6709 'sgckey' parameter which uses the modified algorithm. Also added
6710 an -sgckey command line option to the rsa utility. Thanks to
6711 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6712 algorithm to openssl-dev.
6713 [Steve Henson]
6714
6715 *) The evp_local.h macros were using 'c.##kname' which resulted in
6716 invalid expansion on some systems (SCO 5.0.5 for example).
6717 Corrected to 'c.kname'.
6718 [Phillip Porch <root@theporch.com>]
6719
6720 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6721 a STACK of email addresses from a certificate or request, these look
6722 in the subject name and the subject alternative name extensions and
6723 omit any duplicate addresses.
6724 [Steve Henson]
6725
6726 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6727 This makes DSA verification about 2 % faster.
6728 [Bodo Moeller]
6729
6730 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6731 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6732 plus overhead for 1024 bit moduli).
6733 This makes exponentiations about 0.5 % faster for 1024 bit
6734 exponents (as measured by "openssl speed rsa2048").
6735 [Bodo Moeller]
6736
6737 *) Rename memory handling macros to avoid conflicts with other
6738 software:
6739 Malloc => OPENSSL_malloc
6740 Malloc_locked => OPENSSL_malloc_locked
6741 Realloc => OPENSSL_realloc
6742 Free => OPENSSL_free
6743 [Richard Levitte]
6744
6745 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6746 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6747 [Bodo Moeller]
6748
6749 *) CygWin32 support.
6750 [John Jarvie <jjarvie@newsguy.com>]
6751
6752 *) The type-safe stack code has been rejigged. It is now only compiled
6753 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6754 by default all type-specific stack functions are "#define"d back to
6755 standard stack functions. This results in more streamlined output
6756 but retains the type-safety checking possibilities of the original
6757 approach.
6758 [Geoff Thorpe]
6759
6760 *) The STACK code has been cleaned up, and certain type declarations
6761 that didn't make a lot of sense have been brought in line. This has
6762 also involved a cleanup of sorts in safestack.h to more correctly
6763 map type-safe stack functions onto their plain stack counterparts.
6764 This work has also resulted in a variety of "const"ifications of
6765 lots of the code, especially "_cmp" operations which should normally
6766 be prototyped with "const" parameters anyway.
6767 [Geoff Thorpe]
6768
6769 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6770 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6771 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6772 where all of 'md' is used each time the PRNG is used, but 'state'
6773 is used only indexed by a cyclic counter. As entropy may not be
6774 well distributed from the beginning, 'md' is important as a
6775 chaining variable. However, the output function chains only half
6776 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6777 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6778 in all of 'state' being rewritten, with the new values depending
6779 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6780 [Bodo Moeller]
6781
6782 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6783 the handshake is continued after ssl_verify_cert_chain();
6784 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6785 can lead to 'unexplainable' connection aborts later.
6786 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6787
6788 *) Major EVP API cipher revision.
6789 Add hooks for extra EVP features. This allows various cipher
6790 parameters to be set in the EVP interface. Support added for variable
6791 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6792 setting of RC2 and RC5 parameters.
6793
6794 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6795 ciphers.
6796
6797 Remove lots of duplicated code from the EVP library. For example *every*
6798 cipher init() function handles the 'iv' in the same way according to the
6799 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6800 for CFB and OFB modes they zero ctx->num.
6801
6802 New functionality allows removal of S/MIME code RC2 hack.
6803
6804 Most of the routines have the same form and so can be declared in terms
6805 of macros.
6806
6807 By shifting this to the top level EVP_CipherInit() it can be removed from
6808 all individual ciphers. If the cipher wants to handle IVs or keys
6809 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6810 flags.
6811
6812 Change lots of functions like EVP_EncryptUpdate() to now return a
6813 value: although software versions of the algorithms cannot fail
6814 any installed hardware versions can.
6815 [Steve Henson]
6816
6817 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6818 this option is set, tolerate broken clients that send the negotiated
6819 protocol version number instead of the requested protocol version
6820 number.
6821 [Bodo Moeller]
6822
6823 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6824 i.e. non-zero for export ciphersuites, zero otherwise.
6825 Previous versions had this flag inverted, inconsistent with
6826 rsa_tmp_cb (..._TMP_RSA_CB).
6827 [Bodo Moeller; problem reported by Amit Chopra]
6828
6829 *) Add missing DSA library text string. Work around for some IIS
6830 key files with invalid SEQUENCE encoding.
6831 [Steve Henson]
6832
6833 *) Add a document (doc/standards.txt) that list all kinds of standards
6834 and so on that are implemented in OpenSSL.
6835 [Richard Levitte]
6836
6837 *) Enhance c_rehash script. Old version would mishandle certificates
6838 with the same subject name hash and wouldn't handle CRLs at all.
6839 Added -fingerprint option to crl utility, to support new c_rehash
6840 features.
6841 [Steve Henson]
6842
6843 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6844 [Ulf Möller]
6845
6846 *) Fix for SSL server purpose checking. Server checking was
6847 rejecting certificates which had extended key usage present
6848 but no ssl client purpose.
6849 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6850
6851 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6852 is a little unclear about how a blank password is handled.
6853 Since the password in encoded as a BMPString with terminating
6854 double NULL a zero length password would end up as just the
6855 double NULL. However no password at all is different and is
6856 handled differently in the PKCS#12 key generation code. NS
6857 treats a blank password as zero length. MSIE treats it as no
6858 password on export: but it will try both on import. We now do
6859 the same: PKCS12_parse() tries zero length and no password if
6860 the password is set to "" or NULL (NULL is now a valid password:
6861 it wasn't before) as does the pkcs12 application.
6862 [Steve Henson]
6863
6864 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6865 perror when PEM_read_bio_X509_REQ fails, the error message must
6866 be obtained from the error queue.
6867 [Bodo Moeller]
6868
6869 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6870 it in ERR_remove_state if appropriate, and change ERR_get_state
6871 accordingly to avoid race conditions (this is necessary because
6872 thread_hash is no longer constant once set).
6873 [Bodo Moeller]
6874
6875 *) Bugfix for linux-elf makefile.one.
6876 [Ulf Möller]
6877
6878 *) RSA_get_default_method() will now cause a default
6879 RSA_METHOD to be chosen if one doesn't exist already.
6880 Previously this was only set during a call to RSA_new()
6881 or RSA_new_method(NULL) meaning it was possible for
6882 RSA_get_default_method() to return NULL.
6883 [Geoff Thorpe]
6884
6885 *) Added native name translation to the existing DSO code
6886 that will convert (if the flag to do so is set) filenames
6887 that are sufficiently small and have no path information
6888 into a canonical native form. Eg. "blah" converted to
6889 "libblah.so" or "blah.dll" etc.
6890 [Geoff Thorpe]
6891
6892 *) New function ERR_error_string_n(e, buf, len) which is like
6893 ERR_error_string(e, buf), but writes at most 'len' bytes
6894 including the 0 terminator. For ERR_error_string_n, 'buf'
6895 may not be NULL.
6896 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6897
6898 *) CONF library reworked to become more general. A new CONF
6899 configuration file reader "class" is implemented as well as a
6900 new functions (NCONF_*, for "New CONF") to handle it. The now
6901 old CONF_* functions are still there, but are reimplemented to
6902 work in terms of the new functions. Also, a set of functions
6903 to handle the internal storage of the configuration data is
6904 provided to make it easier to write new configuration file
6905 reader "classes" (I can definitely see something reading a
6906 configuration file in XML format, for example), called _CONF_*,
6907 or "the configuration storage API"...
6908
6909 The new configuration file reading functions are:
6910
6911 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6912 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6913
6914 NCONF_default, NCONF_WIN32
6915
6916 NCONF_dump_fp, NCONF_dump_bio
6917
6918 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6919 NCONF_new creates a new CONF object. This works in the same way
6920 as other interfaces in OpenSSL, like the BIO interface.
6921 NCONF_dump_* dump the internal storage of the configuration file,
6922 which is useful for debugging. All other functions take the same
6923 arguments as the old CONF_* functions wth the exception of the
6924 first that must be a `CONF *' instead of a `LHASH *'.
6925
6926 To make it easer to use the new classes with the old CONF_* functions,
6927 the function CONF_set_default_method is provided.
6928 [Richard Levitte]
6929
6930 *) Add '-tls1' option to 'openssl ciphers', which was already
6931 mentioned in the documentation but had not been implemented.
6932 (This option is not yet really useful because even the additional
6933 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6934 [Bodo Moeller]
6935
6936 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6937 OpenSSL-based applications) load shared libraries and bind to
6938 them in a portable way.
6939 [Geoff Thorpe, with contributions from Richard Levitte]
6940
6941 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6942
6943 *) Make sure _lrotl and _lrotr are only used with MSVC.
6944
6945 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6946 (the default implementation of RAND_status).
6947
6948 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6949 to '-clrext' (= clear extensions), as intended and documented.
6950 [Bodo Moeller; inconsistency pointed out by Michael Attili
6951 <attili@amaxo.com>]
6952
6953 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6954 was larger than the MD block size.
6955 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6956
6957 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6958 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6959 using the passed key: if the passed key was a private key the result
6960 of X509_print(), for example, would be to print out all the private key
6961 components.
6962 [Steve Henson]
6963
6964 *) des_quad_cksum() byte order bug fix.
6965 [Ulf Möller, using the problem description in krb4-0.9.7, where
6966 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6967
6968 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6969 discouraged.
6970 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6971
6972 *) For easily testing in shell scripts whether some command
6973 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6974 returns with exit code 0 iff no command of the given name is available.
6975 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6976 the output goes to stdout and nothing is printed to stderr.
6977 Additional arguments are always ignored.
6978
6979 Since for each cipher there is a command of the same name,
6980 the 'no-cipher' compilation switches can be tested this way.
6981
6982 ('openssl no-XXX' is not able to detect pseudo-commands such
6983 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6984 [Bodo Moeller]
6985
6986 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6987 [Bodo Moeller]
6988
6989 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6990 is set; it will be thrown away anyway because each handshake creates
6991 its own key.
6992 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6993 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6994 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6995 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6996 [Bodo Moeller]
6997
6998 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6999 'Q' and 'R' lose their special meanings (quit/renegotiate).
7000 This is part of what -quiet does; unlike -quiet, -ign_eof
7001 does not suppress any output.
7002 [Richard Levitte]
7003
7004 *) Add compatibility options to the purpose and trust code. The
7005 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7006 accepts a certificate or CA, this was the previous behaviour,
7007 with all the associated security issues.
7008
7009 X509_TRUST_COMPAT is the old trust behaviour: only and
7010 automatically trust self signed roots in certificate store. A
7011 new trust setting X509_TRUST_DEFAULT is used to specify that
7012 a purpose has no associated trust setting and it should instead
7013 use the value in the default purpose.
7014 [Steve Henson]
7015
7016 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7017 and fix a memory leak.
7018 [Steve Henson]
7019
7020 *) In util/mkerr.pl (which implements 'make errors'), preserve
7021 reason strings from the previous version of the .c file, as
7022 the default to have only downcase letters (and digits) in
7023 automatically generated reasons codes is not always appropriate.
7024 [Bodo Moeller]
7025
7026 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7027 using strerror. Previously, ERR_reason_error_string() returned
7028 library names as reason strings for SYSerr; but SYSerr is a special
7029 case where small numbers are errno values, not library numbers.
7030 [Bodo Moeller]
7031
7032 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7033 converts DSA parameters into DH parameters. (When creating parameters,
7034 DSA_generate_parameters is used.)
7035 [Bodo Moeller]
7036
7037 *) Include 'length' (recommended exponent length) in C code generated
7038 by 'openssl dhparam -C'.
7039 [Bodo Moeller]
7040
7041 *) The second argument to set_label in perlasm was already being used
7042 so couldn't be used as a "file scope" flag. Moved to third argument
7043 which was free.
7044 [Steve Henson]
7045
7046 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7047 instead of RAND_bytes for encryption IVs and salts.
7048 [Bodo Moeller]
7049
7050 *) Include RAND_status() into RAND_METHOD instead of implementing
7051 it only for md_rand.c Otherwise replacing the PRNG by calling
7052 RAND_set_rand_method would be impossible.
7053 [Bodo Moeller]
7054
7055 *) Don't let DSA_generate_key() enter an infinite loop if the random
7056 number generation fails.
7057 [Bodo Moeller]
7058
7059 *) New 'rand' application for creating pseudo-random output.
7060 [Bodo Moeller]
7061
7062 *) Added configuration support for Linux/IA64
7063 [Rolf Haberrecker <rolf@suse.de>]
7064
7065 *) Assembler module support for Mingw32.
7066 [Ulf Möller]
7067
7068 *) Shared library support for HPUX (in shlib/).
7069 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7070
7071 *) Shared library support for Solaris gcc.
7072 [Lutz Behnke <behnke@trustcenter.de>]
7073
7074 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7075
7076 *) PKCS7_encrypt() was adding text MIME headers twice because they
7077 were added manually and by SMIME_crlf_copy().
7078 [Steve Henson]
7079
7080 *) In bntest.c don't call BN_rand with zero bits argument.
7081 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7082
7083 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7084 case was implemented. This caused BN_div_recp() to fail occasionally.
7085 [Ulf Möller]
7086
7087 *) Add an optional second argument to the set_label() in the perl
7088 assembly language builder. If this argument exists and is set
7089 to 1 it signals that the assembler should use a symbol whose
7090 scope is the entire file, not just the current function. This
7091 is needed with MASM which uses the format label:: for this scope.
7092 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7093
7094 *) Change the ASN1 types so they are typedefs by default. Before
7095 almost all types were #define'd to ASN1_STRING which was causing
7096 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7097 for example.
7098 [Steve Henson]
7099
7100 *) Change names of new functions to the new get1/get0 naming
7101 convention: After 'get1', the caller owns a reference count
7102 and has to call ..._free; 'get0' returns a pointer to some
7103 data structure without incrementing reference counters.
7104 (Some of the existing 'get' functions increment a reference
7105 counter, some don't.)
7106 Similarly, 'set1' and 'add1' functions increase reference
7107 counters or duplicate objects.
7108 [Steve Henson]
7109
7110 *) Allow for the possibility of temp RSA key generation failure:
7111 the code used to assume it always worked and crashed on failure.
7112 [Steve Henson]
7113
7114 *) Fix potential buffer overrun problem in BIO_printf().
7115 [Ulf Möller, using public domain code by Patrick Powell; problem
7116 pointed out by David Sacerdote <das33@cornell.edu>]
7117
7118 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7119 RAND_egd() and RAND_status(). In the command line application,
7120 the EGD socket can be specified like a seed file using RANDFILE
7121 or -rand.
7122 [Ulf Möller]
7123
7124 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7125 Some CAs (e.g. Verisign) distribute certificates in this form.
7126 [Steve Henson]
7127
7128 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7129 list to exclude them. This means that no special compilation option
7130 is needed to use anonymous DH: it just needs to be included in the
7131 cipher list.
7132 [Steve Henson]
7133
7134 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7135 EVP_MD_type. The old functionality is available in a new macro called
7136 EVP_MD_md(). Change code that uses it and update docs.
7137 [Steve Henson]
7138
7139 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7140 where the 'void *' argument is replaced by a function pointer argument.
7141 Previously 'void *' was abused to point to functions, which works on
7142 many platforms, but is not correct. As these functions are usually
7143 called by macros defined in OpenSSL header files, most source code
7144 should work without changes.
7145 [Richard Levitte]
7146
7147 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7148 sections with information on -D... compiler switches used for
7149 compiling the library so that applications can see them. To enable
7150 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7151 must be defined. E.g.,
7152 #define OPENSSL_ALGORITHM_DEFINES
7153 #include <openssl/opensslconf.h>
7154 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7155 [Richard Levitte, Ulf and Bodo Möller]
7156
7157 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7158 record layer.
7159 [Bodo Moeller]
7160
7161 *) Change the 'other' type in certificate aux info to a STACK_OF
7162 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7163 the required ASN1 format: arbitrary types determined by an OID.
7164 [Steve Henson]
7165
7166 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7167 argument to 'req'. This is not because the function is newer or
7168 better than others it just uses the work 'NEW' in the certificate
7169 request header lines. Some software needs this.
7170 [Steve Henson]
7171
7172 *) Reorganise password command line arguments: now passwords can be
7173 obtained from various sources. Delete the PEM_cb function and make
7174 it the default behaviour: i.e. if the callback is NULL and the
7175 usrdata argument is not NULL interpret it as a null terminated pass
7176 phrase. If usrdata and the callback are NULL then the pass phrase
7177 is prompted for as usual.
7178 [Steve Henson]
7179
7180 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7181 the support is automatically enabled. The resulting binaries will
7182 autodetect the card and use it if present.
7183 [Ben Laurie and Compaq Inc.]
7184
7185 *) Work around for Netscape hang bug. This sends certificate request
7186 and server done in one record. Since this is perfectly legal in the
7187 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7188 the bugs/SSLv3 entry for more info.
7189 [Steve Henson]
7190
7191 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7192 [Andy Polyakov]
7193
7194 *) Add -rand argument to smime and pkcs12 applications and read/write
7195 of seed file.
7196 [Steve Henson]
7197
7198 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7199 [Bodo Moeller]
7200
7201 *) Add command line password options to the remaining applications.
7202 [Steve Henson]
7203
7204 *) Bug fix for BN_div_recp() for numerators with an even number of
7205 bits.
7206 [Ulf Möller]
7207
7208 *) More tests in bntest.c, and changed test_bn output.
7209 [Ulf Möller]
7210
7211 *) ./config recognizes MacOS X now.
7212 [Andy Polyakov]
7213
7214 *) Bug fix for BN_div() when the first words of num and divsor are
7215 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7216 [Ulf Möller]
7217
7218 *) Add support for various broken PKCS#8 formats, and command line
7219 options to produce them.
7220 [Steve Henson]
7221
7222 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7223 get temporary BIGNUMs from a BN_CTX.
7224 [Ulf Möller]
7225
7226 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7227 for p == 0.
7228 [Ulf Möller]
7229
7230 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7231 include a #define from the old name to the new. The original intent
7232 was that statically linked binaries could for example just call
7233 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7234 link with digests. This never worked becayse SSLeay_add_all_digests()
7235 and SSLeay_add_all_ciphers() were in the same source file so calling
7236 one would link with the other. They are now in separate source files.
7237 [Steve Henson]
7238
7239 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7240 [Steve Henson]
7241
7242 *) Use a less unusual form of the Miller-Rabin primality test (it used
7243 a binary algorithm for exponentiation integrated into the Miller-Rabin
7244 loop, our standard modexp algorithms are faster).
7245 [Bodo Moeller]
7246
7247 *) Support for the EBCDIC character set completed.
7248 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7249
7250 *) Source code cleanups: use const where appropriate, eliminate casts,
7251 use void * instead of char * in lhash.
7252 [Ulf Möller]
7253
7254 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7255 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7256 this the server could overwrite ephemeral keys that the client
7257 has already seen).
7258 [Bodo Moeller]
7259
7260 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7261 using 50 iterations of the Rabin-Miller test.
7262
7263 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7264 iterations of the Rabin-Miller test as required by the appendix
7265 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7266 As BN_is_prime_fasttest includes trial division, DSA parameter
7267 generation becomes much faster.
7268
7269 This implies a change for the callback functions in DSA_is_prime
7270 and DSA_generate_parameters: The callback function is called once
7271 for each positive witness in the Rabin-Miller test, not just
7272 occasionally in the inner loop; and the parameters to the
7273 callback function now provide an iteration count for the outer
7274 loop rather than for the current invocation of the inner loop.
7275 DSA_generate_parameters additionally can call the callback
7276 function with an 'iteration count' of -1, meaning that a
7277 candidate has passed the trial division test (when q is generated
7278 from an application-provided seed, trial division is skipped).
7279 [Bodo Moeller]
7280
7281 *) New function BN_is_prime_fasttest that optionally does trial
7282 division before starting the Rabin-Miller test and has
7283 an additional BN_CTX * argument (whereas BN_is_prime always
7284 has to allocate at least one BN_CTX).
7285 'callback(1, -1, cb_arg)' is called when a number has passed the
7286 trial division stage.
7287 [Bodo Moeller]
7288
7289 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7290 as ASN1_TIME.
7291 [Steve Henson]
7292
7293 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7294 [Steve Henson]
7295
7296 *) New function BN_pseudo_rand().
7297 [Ulf Möller]
7298
7299 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7300 bignum version of BN_from_montgomery() with the working code from
7301 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7302 the comments.
7303 [Ulf Möller]
7304
7305 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7306 made it impossible to use the same SSL_SESSION data structure in
7307 SSL2 clients in multiple threads.
7308 [Bodo Moeller]
7309
7310 *) The return value of RAND_load_file() no longer counts bytes obtained
7311 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7312 to seed the PRNG (previously an explicit byte count was required).
7313 [Ulf Möller, Bodo Möller]
7314
7315 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7316 used (char *) instead of (void *) and had casts all over the place.
7317 [Steve Henson]
7318
7319 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7320 [Ulf Möller]
7321
7322 *) Retain source code compatibility for BN_prime_checks macro:
7323 BN_is_prime(..., BN_prime_checks, ...) now uses
7324 BN_prime_checks_for_size to determine the appropriate number of
7325 Rabin-Miller iterations.
7326 [Ulf Möller]
7327
7328 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7329 DH_CHECK_P_NOT_SAFE_PRIME.
7330 (Check if this is true? OpenPGP calls them "strong".)
7331 [Ulf Möller]
7332
7333 *) Merge the functionality of "dh" and "gendh" programs into a new program
7334 "dhparam". The old programs are retained for now but will handle DH keys
7335 (instead of parameters) in future.
7336 [Steve Henson]
7337
7338 *) Make the ciphers, s_server and s_client programs check the return values
7339 when a new cipher list is set.
7340 [Steve Henson]
7341
7342 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7343 ciphers. Before when the 56bit ciphers were enabled the sorting was
7344 wrong.
7345
7346 The syntax for the cipher sorting has been extended to support sorting by
7347 cipher-strength (using the strength_bits hard coded in the tables).
7348 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7349
7350 Fix a bug in the cipher-command parser: when supplying a cipher command
7351 string with an "undefined" symbol (neither command nor alphanumeric
7352 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7353 an error is flagged.
7354
7355 Due to the strength-sorting extension, the code of the
7356 ssl_create_cipher_list() function was completely rearranged. I hope that
7357 the readability was also increased :-)
7358 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7359
7360 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7361 for the first serial number and places 2 in the serial number file. This
7362 avoids problems when the root CA is created with serial number zero and
7363 the first user certificate has the same issuer name and serial number
7364 as the root CA.
7365 [Steve Henson]
7366
7367 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7368 the new code. Add documentation for this stuff.
7369 [Steve Henson]
7370
7371 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7372 X509_*() to X509at_*() on the grounds that they don't handle X509
7373 structures and behave in an analagous way to the X509v3 functions:
7374 they shouldn't be called directly but wrapper functions should be used
7375 instead.
7376
7377 So we also now have some wrapper functions that call the X509at functions
7378 when passed certificate requests. (TO DO: similar things can be done with
7379 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7380 things. Some of these need some d2i or i2d and print functionality
7381 because they handle more complex structures.)
7382 [Steve Henson]
7383
7384 *) Add missing #ifndefs that caused missing symbols when building libssl
7385 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7386 NO_RSA in ssl/s2*.c.
7387 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7388
7389 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7390 has a return value which indicates the quality of the random data
7391 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7392 error queue. New function RAND_pseudo_bytes() generates output that is
7393 guaranteed to be unique but not unpredictable. RAND_add is like
7394 RAND_seed, but takes an extra argument for an entropy estimate
7395 (RAND_seed always assumes full entropy).
7396 [Ulf Möller]
7397
7398 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7399 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7400 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7401 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7402 false-positive rate of at most 2^-80 for random input.
7403 [Bodo Moeller]
7404
7405 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7406 [Bodo Moeller]
7407
7408 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7409 in the 0.9.5 release), this returns the chain
7410 from an X509_CTX structure with a dup of the stack and all
7411 the X509 reference counts upped: so the stack will exist
7412 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7413 to use this.
7414
7415 Also make SSL_SESSION_print() print out the verify return
7416 code.
7417 [Steve Henson]
7418
7419 *) Add manpage for the pkcs12 command. Also change the default
7420 behaviour so MAC iteration counts are used unless the new
7421 -nomaciter option is used. This improves file security and
7422 only older versions of MSIE (4.0 for example) need it.
7423 [Steve Henson]
7424
7425 *) Honor the no-xxx Configure options when creating .DEF files.
7426 [Ulf Möller]
7427
7428 *) Add PKCS#10 attributes to field table: challengePassword,
7429 unstructuredName and unstructuredAddress. These are taken from
7430 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7431 international characters are used.
7432
7433 More changes to X509_ATTRIBUTE code: allow the setting of types
7434 based on strings. Remove the 'loc' parameter when adding
7435 attributes because these will be a SET OF encoding which is sorted
7436 in ASN1 order.
7437 [Steve Henson]
7438
7439 *) Initial changes to the 'req' utility to allow request generation
7440 automation. This will allow an application to just generate a template
7441 file containing all the field values and have req construct the
7442 request.
7443
7444 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7445 used all over the place including certificate requests and PKCS#7
7446 structures. They are currently handled manually where necessary with
7447 some primitive wrappers for PKCS#7. The new functions behave in a
7448 manner analogous to the X509 extension functions: they allow
7449 attributes to be looked up by NID and added.
7450
7451 Later something similar to the X509V3 code would be desirable to
7452 automatically handle the encoding, decoding and printing of the
7453 more complex types. The string types like challengePassword can
7454 be handled by the string table functions.
7455
7456 Also modified the multi byte string table handling. Now there is
7457 a 'global mask' which masks out certain types. The table itself
7458 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7459 is useful when for example there is only one permissible type
7460 (as in countryName) and using the mask might result in no valid
7461 types at all.
7462 [Steve Henson]
7463
7464 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7465 SSL_get_peer_finished to allow applications to obtain the latest
7466 Finished messages sent to the peer or expected from the peer,
7467 respectively. (SSL_get_peer_finished is usually the Finished message
7468 actually received from the peer, otherwise the protocol will be aborted.)
7469
7470 As the Finished message are message digests of the complete handshake
7471 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7472 be used for external authentication procedures when the authentication
7473 provided by SSL/TLS is not desired or is not enough.
7474 [Bodo Moeller]
7475
7476 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7477 the host supports BWX extension and if Compaq C is present on the
7478 $PATH. Just exploiting of the BWX extension results in 20-30%
7479 performance kick for some algorithms, e.g. DES and RC4 to mention
7480 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7481 SHA1.
7482 [Andy Polyakov]
7483
7484 *) Add support for MS "fast SGC". This is arguably a violation of the
7485 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7486 weak crypto and after checking the certificate is SGC a second one
7487 with strong crypto. MS SGC stops the first handshake after receiving
7488 the server certificate message and sends a second client hello. Since
7489 a server will typically do all the time consuming operations before
7490 expecting any further messages from the client (server key exchange
7491 is the most expensive) there is little difference between the two.
7492
7493 To get OpenSSL to support MS SGC we have to permit a second client
7494 hello message after we have sent server done. In addition we have to
7495 reset the MAC if we do get this second client hello.
7496 [Steve Henson]
7497
7498 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7499 if a DER encoded private key is RSA or DSA traditional format. Changed
7500 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7501 format DER encoded private key. Newer code should use PKCS#8 format which
7502 has the key type encoded in the ASN1 structure. Added DER private key
7503 support to pkcs8 application.
7504 [Steve Henson]
7505
7506 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7507 ciphersuites has been selected (as required by the SSL 3/TLS 1
7508 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7509 is set, we interpret this as a request to violate the specification
7510 (the worst that can happen is a handshake failure, and 'correct'
7511 behaviour would result in a handshake failure anyway).
7512 [Bodo Moeller]
7513
7514 *) In SSL_CTX_add_session, take into account that there might be multiple
7515 SSL_SESSION structures with the same session ID (e.g. when two threads
7516 concurrently obtain them from an external cache).
7517 The internal cache can handle only one SSL_SESSION with a given ID,
7518 so if there's a conflict, we now throw out the old one to achieve
7519 consistency.
7520 [Bodo Moeller]
7521
7522 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7523 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7524 some routines that use cipher OIDs: some ciphers do not have OIDs
7525 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7526 example.
7527 [Steve Henson]
7528
7529 *) Simplify the trust setting structure and code. Now we just have
7530 two sequences of OIDs for trusted and rejected settings. These will
7531 typically have values the same as the extended key usage extension
7532 and any application specific purposes.
7533
7534 The trust checking code now has a default behaviour: it will just
7535 check for an object with the same NID as the passed id. Functions can
7536 be provided to override either the default behaviour or the behaviour
7537 for a given id. SSL client, server and email already have functions
7538 in place for compatibility: they check the NID and also return "trusted"
7539 if the certificate is self signed.
7540 [Steve Henson]
7541
7542 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7543 traditional format into an EVP_PKEY structure.
7544 [Steve Henson]
7545
7546 *) Add a password callback function PEM_cb() which either prompts for
7547 a password if usr_data is NULL or otherwise assumes it is a null
7548 terminated password. Allow passwords to be passed on command line
7549 environment or config files in a few more utilities.
7550 [Steve Henson]
7551
7552 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7553 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7554 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7555 Update documentation.
7556 [Steve Henson]
7557
7558 *) Support for ASN1 "NULL" type. This could be handled before by using
7559 ASN1_TYPE but there wasn't any function that would try to read a NULL
7560 and produce an error if it couldn't. For compatibility we also have
7561 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7562 don't allocate anything because they don't need to.
7563 [Steve Henson]
7564
7565 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7566 for details.
7567 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7568
7569 *) Rebuild of the memory allocation routines used by OpenSSL code and
7570 possibly others as well. The purpose is to make an interface that
7571 provide hooks so anyone can build a separate set of allocation and
7572 deallocation routines to be used by OpenSSL, for example memory
7573 pool implementations, or something else, which was previously hard
7574 since Malloc(), Realloc() and Free() were defined as macros having
7575 the values malloc, realloc and free, respectively (except for Win32
7576 compilations). The same is provided for memory debugging code.
7577 OpenSSL already comes with functionality to find memory leaks, but
7578 this gives people a chance to debug other memory problems.
7579
7580 With these changes, a new set of functions and macros have appeared:
7581
7582 CRYPTO_set_mem_debug_functions() [F]
7583 CRYPTO_get_mem_debug_functions() [F]
7584 CRYPTO_dbg_set_options() [F]
7585 CRYPTO_dbg_get_options() [F]
7586 CRYPTO_malloc_debug_init() [M]
7587
7588 The memory debug functions are NULL by default, unless the library
7589 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7590 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7591 gives the standard debugging functions that come with OpenSSL) or
7592 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7593 provided by the library user) must be used. When the standard
7594 debugging functions are used, CRYPTO_dbg_set_options can be used to
7595 request additional information:
7596 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7597 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7598
7599 Also, things like CRYPTO_set_mem_functions will always give the
7600 expected result (the new set of functions is used for allocation
7601 and deallocation) at all times, regardless of platform and compiler
7602 options.
7603
7604 To finish it up, some functions that were never use in any other
7605 way than through macros have a new API and new semantic:
7606
7607 CRYPTO_dbg_malloc()
7608 CRYPTO_dbg_realloc()
7609 CRYPTO_dbg_free()
7610
7611 All macros of value have retained their old syntax.
7612 [Richard Levitte and Bodo Moeller]
7613
7614 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7615 ordering of SMIMECapabilities wasn't in "strength order" and there
7616 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7617 algorithm.
7618 [Steve Henson]
7619
7620 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7621 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7622 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7623
7624 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7625 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7626 functionality to handle multipart/signed properly) and a utility
7627 called 'smime' to call all this stuff. This is based on code I
7628 originally wrote for Celo who have kindly allowed it to be
7629 included in OpenSSL.
7630 [Steve Henson]
7631
7632 *) Add variants des_set_key_checked and des_set_key_unchecked of
7633 des_set_key (aka des_key_sched). Global variable des_check_key
7634 decides which of these is called by des_set_key; this way
7635 des_check_key behaves as it always did, but applications and
7636 the library itself, which was buggy for des_check_key == 1,
7637 have a cleaner way to pick the version they need.
7638 [Bodo Moeller]
7639
7640 *) New function PKCS12_newpass() which changes the password of a
7641 PKCS12 structure.
7642 [Steve Henson]
7643
7644 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7645 dynamic mix. In both cases the ids can be used as an index into the
7646 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7647 functions so they accept a list of the field values and the
7648 application doesn't need to directly manipulate the X509_TRUST
7649 structure.
7650 [Steve Henson]
7651
7652 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7653 need initialising.
7654 [Steve Henson]
7655
7656 *) Modify the way the V3 extension code looks up extensions. This now
7657 works in a similar way to the object code: we have some "standard"
7658 extensions in a static table which is searched with OBJ_bsearch()
7659 and the application can add dynamic ones if needed. The file
7660 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7661 updated whenever a new extension is added to the core code and kept
7662 in ext_nid order. There is a simple program 'tabtest.c' which checks
7663 this. New extensions are not added too often so this file can readily
7664 be maintained manually.
7665
7666 There are two big advantages in doing things this way. The extensions
7667 can be looked up immediately and no longer need to be "added" using
7668 X509V3_add_standard_extensions(): this function now does nothing.
7669 [Side note: I get *lots* of email saying the extension code doesn't
7670 work because people forget to call this function]
7671 Also no dynamic allocation is done unless new extensions are added:
7672 so if we don't add custom extensions there is no need to call
7673 X509V3_EXT_cleanup().
7674 [Steve Henson]
7675
7676 *) Modify enc utility's salting as follows: make salting the default. Add a
7677 magic header, so unsalted files fail gracefully instead of just decrypting
7678 to garbage. This is because not salting is a big security hole, so people
7679 should be discouraged from doing it.
7680 [Ben Laurie]
7681
7682 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7683 digest to be passed on the command line but it only used this
7684 parameter when signing a certificate. Modified so all relevant
7685 operations are affected by the digest parameter including the
7686 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7687 DSA key was used because it didn't fix the digest.
7688 [Steve Henson]
7689
7690 *) Initial certificate chain verify code. Currently tests the untrusted
7691 certificates for consistency with the verify purpose (which is set
7692 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7693
7694 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7695 this is because it will reject chains with invalid extensions whereas
7696 every previous version of OpenSSL and SSLeay made no checks at all.
7697
7698 Trust code: checks the root CA for the relevant trust settings. Trust
7699 settings have an initial value consistent with the verify purpose: e.g.
7700 if the verify purpose is for SSL client use it expects the CA to be
7701 trusted for SSL client use. However the default value can be changed to
7702 permit custom trust settings: one example of this would be to only trust
7703 certificates from a specific "secure" set of CAs.
7704
7705 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7706 which should be used for version portability: especially since the
7707 verify structure is likely to change more often now.
7708
7709 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7710 to set them. If not set then assume SSL clients will verify SSL servers
7711 and vice versa.
7712
7713 Two new options to the verify program: -untrusted allows a set of
7714 untrusted certificates to be passed in and -purpose which sets the
7715 intended purpose of the certificate. If a purpose is set then the
7716 new chain verify code is used to check extension consistency.
7717 [Steve Henson]
7718
7719 *) Support for the authority information access extension.
7720 [Steve Henson]
7721
7722 *) Modify RSA and DSA PEM read routines to transparently handle
7723 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7724 public keys in a format compatible with certificate
7725 SubjectPublicKeyInfo structures. Unfortunately there were already
7726 functions called *_PublicKey_* which used various odd formats so
7727 these are retained for compatibility: however the DSA variants were
7728 never in a public release so they have been deleted. Changed dsa/rsa
7729 utilities to handle the new format: note no releases ever handled public
7730 keys so we should be OK.
7731
7732 The primary motivation for this change is to avoid the same fiasco
7733 that dogs private keys: there are several incompatible private key
7734 formats some of which are standard and some OpenSSL specific and
7735 require various evil hacks to allow partial transparent handling and
7736 even then it doesn't work with DER formats. Given the option anything
7737 other than PKCS#8 should be dumped: but the other formats have to
7738 stay in the name of compatibility.
7739
7740 With public keys and the benefit of hindsight one standard format
7741 is used which works with EVP_PKEY, RSA or DSA structures: though
7742 it clearly returns an error if you try to read the wrong kind of key.
7743
7744 Added a -pubkey option to the 'x509' utility to output the public key.
7745 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7746 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7747 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7748 that do the same as the EVP_PKEY_assign_*() except they up the
7749 reference count of the added key (they don't "swallow" the
7750 supplied key).
7751 [Steve Henson]
7752
7753 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7754 CRLs would fail if the file contained no certificates or no CRLs:
7755 added a new function to read in both types and return the number
7756 read: this means that if none are read it will be an error. The
7757 DER versions of the certificate and CRL reader would always fail
7758 because it isn't possible to mix certificates and CRLs in DER format
7759 without choking one or the other routine. Changed this to just read
7760 a certificate: this is the best we can do. Also modified the code
7761 in apps/verify.c to take notice of return codes: it was previously
7762 attempting to read in certificates from NULL pointers and ignoring
7763 any errors: this is one reason why the cert and CRL reader seemed
7764 to work. It doesn't check return codes from the default certificate
7765 routines: these may well fail if the certificates aren't installed.
7766 [Steve Henson]
7767
7768 *) Code to support otherName option in GeneralName.
7769 [Steve Henson]
7770
7771 *) First update to verify code. Change the verify utility
7772 so it warns if it is passed a self signed certificate:
7773 for consistency with the normal behaviour. X509_verify
7774 has been modified to it will now verify a self signed
7775 certificate if *exactly* the same certificate appears
7776 in the store: it was previously impossible to trust a
7777 single self signed certificate. This means that:
7778 openssl verify ss.pem
7779 now gives a warning about a self signed certificate but
7780 openssl verify -CAfile ss.pem ss.pem
7781 is OK.
7782 [Steve Henson]
7783
7784 *) For servers, store verify_result in SSL_SESSION data structure
7785 (and add it to external session representation).
7786 This is needed when client certificate verifications fails,
7787 but an application-provided verification callback (set by
7788 SSL_CTX_set_cert_verify_callback) allows accepting the session
7789 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7790 but returns 1): When the session is reused, we have to set
7791 ssl->verify_result to the appropriate error code to avoid
7792 security holes.
7793 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7794
7795 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7796 case in PKCS7_dataInit() where the signed PKCS7 structure
7797 didn't contain any existing data because it was being created.
7798 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7799
7800 *) Add a salt to the key derivation routines in enc.c. This
7801 forms the first 8 bytes of the encrypted file. Also add a
7802 -S option to allow a salt to be input on the command line.
7803 [Steve Henson]
7804
7805 *) New function X509_cmp(). Oddly enough there wasn't a function
7806 to compare two certificates. We do this by working out the SHA1
7807 hash and comparing that. X509_cmp() will be needed by the trust
7808 code.
7809 [Steve Henson]
7810
7811 *) SSL_get1_session() is like SSL_get_session(), but increments
7812 the reference count in the SSL_SESSION returned.
7813 [Geoff Thorpe <geoff@eu.c2.net>]
7814
7815 *) Fix for 'req': it was adding a null to request attributes.
7816 Also change the X509_LOOKUP and X509_INFO code to handle
7817 certificate auxiliary information.
7818 [Steve Henson]
7819
7820 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7821 the 'enc' command.
7822 [Steve Henson]
7823
7824 *) Add the possibility to add extra information to the memory leak
7825 detecting output, to form tracebacks, showing from where each
7826 allocation was originated: CRYPTO_push_info("constant string") adds
7827 the string plus current file name and line number to a per-thread
7828 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7829 is like calling CYRPTO_pop_info() until the stack is empty.
7830 Also updated memory leak detection code to be multi-thread-safe.
7831 [Richard Levitte]
7832
7833 *) Add options -text and -noout to pkcs7 utility and delete the
7834 encryption options which never did anything. Update docs.
7835 [Steve Henson]
7836
7837 *) Add options to some of the utilities to allow the pass phrase
7838 to be included on either the command line (not recommended on
7839 OSes like Unix) or read from the environment. Update the
7840 manpages and fix a few bugs.
7841 [Steve Henson]
7842
7843 *) Add a few manpages for some of the openssl commands.
7844 [Steve Henson]
7845
7846 *) Fix the -revoke option in ca. It was freeing up memory twice,
7847 leaking and not finding already revoked certificates.
7848 [Steve Henson]
7849
7850 *) Extensive changes to support certificate auxiliary information.
7851 This involves the use of X509_CERT_AUX structure and X509_AUX
7852 functions. An X509_AUX function such as PEM_read_X509_AUX()
7853 can still read in a certificate file in the usual way but it
7854 will also read in any additional "auxiliary information". By
7855 doing things this way a fair degree of compatibility can be
7856 retained: existing certificates can have this information added
7857 using the new 'x509' options.
7858
7859 Current auxiliary information includes an "alias" and some trust
7860 settings. The trust settings will ultimately be used in enhanced
7861 certificate chain verification routines: currently a certificate
7862 can only be trusted if it is self signed and then it is trusted
7863 for all purposes.
7864 [Steve Henson]
7865
7866 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7867 The problem was that one of the replacement routines had not been working
7868 since SSLeay releases. For now the offending routine has been replaced
7869 with non-optimised assembler. Even so, this now gives around 95%
7870 performance improvement for 1024 bit RSA signs.
7871 [Mark Cox]
7872
7873 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7874 handling. Most clients have the effective key size in bits equal to
7875 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7876 A few however don't do this and instead use the size of the decrypted key
7877 to determine the RC2 key length and the AlgorithmIdentifier to determine
7878 the effective key length. In this case the effective key length can still
7879 be 40 bits but the key length can be 168 bits for example. This is fixed
7880 by manually forcing an RC2 key into the EVP_PKEY structure because the
7881 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7882 the key length and effective key length are equal.
7883 [Steve Henson]
7884
7885 *) Add a bunch of functions that should simplify the creation of
7886 X509_NAME structures. Now you should be able to do:
7887 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7888 and have it automatically work out the correct field type and fill in
7889 the structures. The more adventurous can try:
7890 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7891 and it will (hopefully) work out the correct multibyte encoding.
7892 [Steve Henson]
7893
7894 *) Change the 'req' utility to use the new field handling and multibyte
7895 copy routines. Before the DN field creation was handled in an ad hoc
7896 way in req, ca, and x509 which was rather broken and didn't support
7897 BMPStrings or UTF8Strings. Since some software doesn't implement
7898 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7899 using the dirstring_type option. See the new comment in the default
7900 openssl.cnf for more info.
7901 [Steve Henson]
7902
7903 *) Make crypto/rand/md_rand.c more robust:
7904 - Assure unique random numbers after fork().
7905 - Make sure that concurrent threads access the global counter and
7906 md serializably so that we never lose entropy in them
7907 or use exactly the same state in multiple threads.
7908 Access to the large state is not always serializable because
7909 the additional locking could be a performance killer, and
7910 md should be large enough anyway.
7911 [Bodo Moeller]
7912
7913 *) New file apps/app_rand.c with commonly needed functionality
7914 for handling the random seed file.
7915
7916 Use the random seed file in some applications that previously did not:
7917 ca,
7918 dsaparam -genkey (which also ignored its '-rand' option),
7919 s_client,
7920 s_server,
7921 x509 (when signing).
7922 Except on systems with /dev/urandom, it is crucial to have a random
7923 seed file at least for key creation, DSA signing, and for DH exchanges;
7924 for RSA signatures we could do without one.
7925
7926 gendh and gendsa (unlike genrsa) used to read only the first byte
7927 of each file listed in the '-rand' option. The function as previously
7928 found in genrsa is now in app_rand.c and is used by all programs
7929 that support '-rand'.
7930 [Bodo Moeller]
7931
7932 *) In RAND_write_file, use mode 0600 for creating files;
7933 don't just chmod when it may be too late.
7934 [Bodo Moeller]
7935
7936 *) Report an error from X509_STORE_load_locations
7937 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7938 [Bill Perry]
7939
7940 *) New function ASN1_mbstring_copy() this copies a string in either
7941 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7942 into an ASN1_STRING type. A mask of permissible types is passed
7943 and it chooses the "minimal" type to use or an error if not type
7944 is suitable.
7945 [Steve Henson]
7946
7947 *) Add function equivalents to the various macros in asn1.h. The old
7948 macros are retained with an M_ prefix. Code inside the library can
7949 use the M_ macros. External code (including the openssl utility)
7950 should *NOT* in order to be "shared library friendly".
7951 [Steve Henson]
7952
7953 *) Add various functions that can check a certificate's extensions
7954 to see if it usable for various purposes such as SSL client,
7955 server or S/MIME and CAs of these types. This is currently
7956 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7957 verification. Also added a -purpose flag to x509 utility to
7958 print out all the purposes.
7959 [Steve Henson]
7960
7961 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7962 functions.
7963 [Steve Henson]
7964
7965 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7966 for, obtain and decode and extension and obtain its critical flag.
7967 This allows all the necessary extension code to be handled in a
7968 single function call.
7969 [Steve Henson]
7970
7971 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7972 platforms. See crypto/rc4/rc4_enc.c for further details.
7973 [Andy Polyakov]
7974
7975 *) New -noout option to asn1parse. This causes no output to be produced
7976 its main use is when combined with -strparse and -out to extract data
7977 from a file (which may not be in ASN.1 format).
7978 [Steve Henson]
7979
7980 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7981 when producing the local key id.
7982 [Richard Levitte <levitte@stacken.kth.se>]
7983
7984 *) New option -dhparam in s_server. This allows a DH parameter file to be
7985 stated explicitly. If it is not stated then it tries the first server
7986 certificate file. The previous behaviour hard coded the filename
7987 "server.pem".
7988 [Steve Henson]
7989
7990 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7991 a public key to be input or output. For example:
7992 openssl rsa -in key.pem -pubout -out pubkey.pem
7993 Also added necessary DSA public key functions to handle this.
7994 [Steve Henson]
7995
7996 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7997 in the message. This was handled by allowing
7998 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7999 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8000
8001 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8002 to the end of the strings whereas this didn't. This would cause problems
8003 if strings read with d2i_ASN1_bytes() were later modified.
8004 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8005
8006 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8007 data and it contains EOF it will end up returning an error. This is
8008 caused by input 46 bytes long. The cause is due to the way base64
8009 BIOs find the start of base64 encoded data. They do this by trying a
8010 trial decode on each line until they find one that works. When they
8011 do a flag is set and it starts again knowing it can pass all the
8012 data directly through the decoder. Unfortunately it doesn't reset
8013 the context it uses. This means that if EOF is reached an attempt
8014 is made to pass two EOFs through the context and this causes the
8015 resulting error. This can also cause other problems as well. As is
8016 usual with these problems it takes *ages* to find and the fix is
8017 trivial: move one line.
8018 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8019
8020 *) Ugly workaround to get s_client and s_server working under Windows. The
8021 old code wouldn't work because it needed to select() on sockets and the
8022 tty (for keypresses and to see if data could be written). Win32 only
8023 supports select() on sockets so we select() with a 1s timeout on the
8024 sockets and then see if any characters are waiting to be read, if none
8025 are present then we retry, we also assume we can always write data to
8026 the tty. This isn't nice because the code then blocks until we've
8027 received a complete line of data and it is effectively polling the
8028 keyboard at 1s intervals: however it's quite a bit better than not
8029 working at all :-) A dedicated Windows application might handle this
8030 with an event loop for example.
8031 [Steve Henson]
8032
8033 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8034 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8035 will be called when RSA_sign() and RSA_verify() are used. This is useful
8036 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8037 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8038 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8039 This necessitated the support of an extra signature type NID_md5_sha1
8040 for SSL signatures and modifications to the SSL library to use it instead
8041 of calling RSA_public_decrypt() and RSA_private_encrypt().
8042 [Steve Henson]
8043
8044 *) Add new -verify -CAfile and -CApath options to the crl program, these
8045 will lookup a CRL issuers certificate and verify the signature in a
8046 similar way to the verify program. Tidy up the crl program so it
8047 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8048 less strict. It will now permit CRL extensions even if it is not
8049 a V2 CRL: this will allow it to tolerate some broken CRLs.
8050 [Steve Henson]
8051
8052 *) Initialize all non-automatic variables each time one of the openssl
8053 sub-programs is started (this is necessary as they may be started
8054 multiple times from the "OpenSSL>" prompt).
8055 [Lennart Bang, Bodo Moeller]
8056
8057 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8058 removing all other RSA functionality (this is what NO_RSA does). This
8059 is so (for example) those in the US can disable those operations covered
8060 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8061 key generation.
8062 [Steve Henson]
8063
8064 *) Non-copying interface to BIO pairs.
8065 (still largely untested)
8066 [Bodo Moeller]
8067
8068 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8069 ASCII string. This was handled independently in various places before.
8070 [Steve Henson]
8071
8072 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8073 UTF8 strings a character at a time.
8074 [Steve Henson]
8075
8076 *) Use client_version from client hello to select the protocol
8077 (s23_srvr.c) and for RSA client key exchange verification
8078 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8079 [Bodo Moeller]
8080
8081 *) Add various utility functions to handle SPKACs, these were previously
8082 handled by poking round in the structure internals. Added new function
8083 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8084 print, verify and generate SPKACs. Based on an original idea from
8085 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8086 [Steve Henson]
8087
8088 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8089 [Andy Polyakov]
8090
8091 *) Allow the config file extension section to be overwritten on the
8092 command line. Based on an original idea from Massimiliano Pala
8093 <madwolf@comune.modena.it>. The new option is called -extensions
8094 and can be applied to ca, req and x509. Also -reqexts to override
8095 the request extensions in req and -crlexts to override the crl extensions
8096 in ca.
8097 [Steve Henson]
8098
8099 *) Add new feature to the SPKAC handling in ca. Now you can include
8100 the same field multiple times by preceding it by "XXXX." for example:
8101 1.OU="Unit name 1"
8102 2.OU="Unit name 2"
8103 this is the same syntax as used in the req config file.
8104 [Steve Henson]
8105
8106 *) Allow certificate extensions to be added to certificate requests. These
8107 are specified in a 'req_extensions' option of the req section of the
8108 config file. They can be printed out with the -text option to req but
8109 are otherwise ignored at present.
8110 [Steve Henson]
8111
8112 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8113 data read consists of only the final block it would not decrypted because
8114 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8115 A misplaced 'break' also meant the decrypted final block might not be
8116 copied until the next read.
8117 [Steve Henson]
8118
8119 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8120 a few extra parameters to the DH structure: these will be useful if
8121 for example we want the value of 'q' or implement X9.42 DH.
8122 [Steve Henson]
8123
8124 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8125 provides hooks that allow the default DSA functions or functions on a
8126 "per key" basis to be replaced. This allows hardware acceleration and
8127 hardware key storage to be handled without major modification to the
8128 library. Also added low level modexp hooks and CRYPTO_EX structure and
8129 associated functions.
8130 [Steve Henson]
8131
8132 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8133 as "read only": it can't be written to and the buffer it points to will
8134 not be freed. Reading from a read only BIO is much more efficient than
8135 a normal memory BIO. This was added because there are several times when
8136 an area of memory needs to be read from a BIO. The previous method was
8137 to create a memory BIO and write the data to it, this results in two
8138 copies of the data and an O(n^2) reading algorithm. There is a new
8139 function BIO_new_mem_buf() which creates a read only memory BIO from
8140 an area of memory. Also modified the PKCS#7 routines to use read only
8141 memory BIOs.
8142 [Steve Henson]
8143
8144 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8145 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8146 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8147 but a retry condition occured while trying to read the rest.
8148 [Bodo Moeller]
8149
8150 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8151 NID_pkcs7_encrypted by default: this was wrong since this should almost
8152 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8153 the encrypted data type: this is a more sensible place to put it and it
8154 allows the PKCS#12 code to be tidied up that duplicated this
8155 functionality.
8156 [Steve Henson]
8157
8158 *) Changed obj_dat.pl script so it takes its input and output files on
8159 the command line. This should avoid shell escape redirection problems
8160 under Win32.
8161 [Steve Henson]
8162
8163 *) Initial support for certificate extension requests, these are included
8164 in things like Xenroll certificate requests. Included functions to allow
8165 extensions to be obtained and added.
8166 [Steve Henson]
8167
8168 *) -crlf option to s_client and s_server for sending newlines as
8169 CRLF (as required by many protocols).
8170 [Bodo Moeller]
8171
8172 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8173
8174 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8175 [Ralf S. Engelschall]
8176
8177 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8178 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8179
8180 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8181 program.
8182 [Steve Henson]
8183
8184 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8185 DH parameters/keys (q is lost during that conversion, but the resulting
8186 DH parameters contain its length).
8187
8188 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8189 much faster than DH_generate_parameters (which creates parameters
8190 where p = 2*q + 1), and also the smaller q makes DH computations
8191 much more efficient (160-bit exponentiation instead of 1024-bit
8192 exponentiation); so this provides a convenient way to support DHE
8193 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8194 utter importance to use
8195 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8196 or
8197 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8198 when such DH parameters are used, because otherwise small subgroup
8199 attacks may become possible!
8200 [Bodo Moeller]
8201
8202 *) Avoid memory leak in i2d_DHparams.
8203 [Bodo Moeller]
8204
8205 *) Allow the -k option to be used more than once in the enc program:
8206 this allows the same encrypted message to be read by multiple recipients.
8207 [Steve Henson]
8208
8209 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8210 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8211 it will always use the numerical form of the OID, even if it has a short
8212 or long name.
8213 [Steve Henson]
8214
8215 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8216 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8217 otherwise bn_mod_exp was called. In the case of hardware keys for example
8218 no private key components need be present and it might store extra data
8219 in the RSA structure, which cannot be accessed from bn_mod_exp.
8220 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8221 private key operations.
8222 [Steve Henson]
8223
8224 *) Added support for SPARC Linux.
8225 [Andy Polyakov]
8226
8227 *) pem_password_cb function type incompatibly changed from
8228 typedef int pem_password_cb(char *buf, int size, int rwflag);
8229 to
8230 ....(char *buf, int size, int rwflag, void *userdata);
8231 so that applications can pass data to their callbacks:
8232 The PEM[_ASN1]_{read,write}... functions and macros now take an
8233 additional void * argument, which is just handed through whenever
8234 the password callback is called.
8235 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8236
8237 New function SSL_CTX_set_default_passwd_cb_userdata.
8238
8239 Compatibility note: As many C implementations push function arguments
8240 onto the stack in reverse order, the new library version is likely to
8241 interoperate with programs that have been compiled with the old
8242 pem_password_cb definition (PEM_whatever takes some data that
8243 happens to be on the stack as its last argument, and the callback
8244 just ignores this garbage); but there is no guarantee whatsoever that
8245 this will work.
8246
8247 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8248 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8249 problems not only on Windows, but also on some Unix platforms.
8250 To avoid problematic command lines, these definitions are now in an
8251 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8252 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8253 [Bodo Moeller]
8254
8255 *) MIPS III/IV assembler module is reimplemented.
8256 [Andy Polyakov]
8257
8258 *) More DES library cleanups: remove references to srand/rand and
8259 delete an unused file.
8260 [Ulf Möller]
8261
8262 *) Add support for the the free Netwide assembler (NASM) under Win32,
8263 since not many people have MASM (ml) and it can be hard to obtain.
8264 This is currently experimental but it seems to work OK and pass all
8265 the tests. Check out INSTALL.W32 for info.
8266 [Steve Henson]
8267
8268 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8269 without temporary keys kept an extra copy of the server key,
8270 and connections with temporary keys did not free everything in case
8271 of an error.
8272 [Bodo Moeller]
8273
8274 *) New function RSA_check_key and new openssl rsa option -check
8275 for verifying the consistency of RSA keys.
8276 [Ulf Moeller, Bodo Moeller]
8277
8278 *) Various changes to make Win32 compile work:
8279 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8280 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8281 comparison" warnings.
8282 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8283 [Steve Henson]
8284
8285 *) Add a debugging option to PKCS#5 v2 key generation function: when
8286 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8287 derived keys are printed to stderr.
8288 [Steve Henson]
8289
8290 *) Copy the flags in ASN1_STRING_dup().
8291 [Roman E. Pavlov <pre@mo.msk.ru>]
8292
8293 *) The x509 application mishandled signing requests containing DSA
8294 keys when the signing key was also DSA and the parameters didn't match.
8295
8296 It was supposed to omit the parameters when they matched the signing key:
8297 the verifying software was then supposed to automatically use the CA's
8298 parameters if they were absent from the end user certificate.
8299
8300 Omitting parameters is no longer recommended. The test was also
8301 the wrong way round! This was probably due to unusual behaviour in
8302 EVP_cmp_parameters() which returns 1 if the parameters match.
8303 This meant that parameters were omitted when they *didn't* match and
8304 the certificate was useless. Certificates signed with 'ca' didn't have
8305 this bug.
8306 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8307
8308 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8309 The interface is as follows:
8310 Applications can use
8311 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8312 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8313 "off" is now the default.
8314 The library internally uses
8315 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8316 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8317 to disable memory-checking temporarily.
8318
8319 Some inconsistent states that previously were possible (and were
8320 even the default) are now avoided.
8321
8322 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8323 with each memory chunk allocated; this is occasionally more helpful
8324 than just having a counter.
8325
8326 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8327
8328 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8329 extensions.
8330 [Bodo Moeller]
8331
8332 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8333 which largely parallels "options", but is for changing API behaviour,
8334 whereas "options" are about protocol behaviour.
8335 Initial "mode" flags are:
8336
8337 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8338 a single record has been written.
8339 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8340 retries use the same buffer location.
8341 (But all of the contents must be
8342 copied!)
8343 [Bodo Moeller]
8344
8345 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8346 worked.
8347
8348 *) Fix problems with no-hmac etc.
8349 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8350
8351 *) New functions RSA_get_default_method(), RSA_set_method() and
8352 RSA_get_method(). These allows replacement of RSA_METHODs without having
8353 to mess around with the internals of an RSA structure.
8354 [Steve Henson]
8355
8356 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8357 Also really enable memory leak checks in openssl.c and in some
8358 test programs.
8359 [Chad C. Mulligan, Bodo Moeller]
8360
8361 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8362 up the length of negative integers. This has now been simplified to just
8363 store the length when it is first determined and use it later, rather
8364 than trying to keep track of where data is copied and updating it to
8365 point to the end.
8366 [Steve Henson, reported by Brien Wheeler
8367 <bwheeler@authentica-security.com>]
8368
8369 *) Add a new function PKCS7_signatureVerify. This allows the verification
8370 of a PKCS#7 signature but with the signing certificate passed to the
8371 function itself. This contrasts with PKCS7_dataVerify which assumes the
8372 certificate is present in the PKCS#7 structure. This isn't always the
8373 case: certificates can be omitted from a PKCS#7 structure and be
8374 distributed by "out of band" means (such as a certificate database).
8375 [Steve Henson]
8376
8377 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8378 function prototypes in pem.h, also change util/mkdef.pl to add the
8379 necessary function names.
8380 [Steve Henson]
8381
8382 *) mk1mf.pl (used by Windows builds) did not properly read the
8383 options set by Configure in the top level Makefile, and Configure
8384 was not even able to write more than one option correctly.
8385 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8386 [Bodo Moeller]
8387
8388 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8389 file to be loaded from a BIO or FILE pointer. The BIO version will
8390 for example allow memory BIOs to contain config info.
8391 [Steve Henson]
8392
8393 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8394 Whoever hopes to achieve shared-library compatibility across versions
8395 must use this, not the compile-time macro.
8396 (Exercise 0.9.4: Which is the minimum library version required by
8397 such programs?)
8398 Note: All this applies only to multi-threaded programs, others don't
8399 need locks.
8400 [Bodo Moeller]
8401
8402 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8403 through a BIO pair triggered the default case, i.e.
8404 SSLerr(...,SSL_R_UNKNOWN_STATE).
8405 [Bodo Moeller]
8406
8407 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8408 can use the SSL library even if none of the specific BIOs is
8409 appropriate.
8410 [Bodo Moeller]
8411
8412 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8413 for the encoded length.
8414 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8415
8416 *) Add initial documentation of the X509V3 functions.
8417 [Steve Henson]
8418
8419 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8420 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8421 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8422 secure PKCS#8 private key format with a high iteration count.
8423 [Steve Henson]
8424
8425 *) Fix determination of Perl interpreter: A perl or perl5
8426 _directory_ in $PATH was also accepted as the interpreter.
8427 [Ralf S. Engelschall]
8428
8429 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8430 wrong with it but it was very old and did things like calling
8431 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8432 unusual formatting.
8433 [Steve Henson]
8434
8435 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8436 to use the new extension code.
8437 [Steve Henson]
8438
8439 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8440 with macros. This should make it easier to change their form, add extra
8441 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8442 constant.
8443 [Steve Henson]
8444
8445 *) Add to configuration table a new entry that can specify an alternative
8446 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8447 according to Mark Crispin <MRC@Panda.COM>.
8448 [Bodo Moeller]
8449
8450 #if 0
8451 *) DES CBC did not update the IV. Weird.
8452 [Ben Laurie]
8453 #else
8454 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8455 Changing the behaviour of the former might break existing programs --
8456 where IV updating is needed, des_ncbc_encrypt can be used.
8457 #endif
8458
8459 *) When bntest is run from "make test" it drives bc to check its
8460 calculations, as well as internally checking them. If an internal check
8461 fails, it needs to cause bc to give a non-zero result or make test carries
8462 on without noticing the failure. Fixed.
8463 [Ben Laurie]
8464
8465 *) DES library cleanups.
8466 [Ulf Möller]
8467
8468 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8469 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8470 ciphers. NOTE: although the key derivation function has been verified
8471 against some published test vectors it has not been extensively tested
8472 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8473 of v2.0.
8474 [Steve Henson]
8475
8476 *) Instead of "mkdir -p", which is not fully portable, use new
8477 Perl script "util/mkdir-p.pl".
8478 [Bodo Moeller]
8479
8480 *) Rewrite the way password based encryption (PBE) is handled. It used to
8481 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8482 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8483 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8484 the 'parameter' field of the AlgorithmIdentifier is passed to the
8485 underlying key generation function so it must do its own ASN1 parsing.
8486 This has also changed the EVP_PBE_CipherInit() function which now has a
8487 'parameter' argument instead of literal salt and iteration count values
8488 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8489 [Steve Henson]
8490
8491 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8492 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8493 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8494 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8495 value was just used as a "magic string" and not used directly its
8496 value doesn't matter.
8497 [Steve Henson]
8498
8499 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8500 support mutable.
8501 [Ben Laurie]
8502
8503 *) "linux-sparc64" configuration (ultrapenguin).
8504 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8505 "linux-sparc" configuration.
8506 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8507
8508 *) config now generates no-xxx options for missing ciphers.
8509 [Ulf Möller]
8510
8511 *) Support the EBCDIC character set (work in progress).
8512 File ebcdic.c not yet included because it has a different license.
8513 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8514
8515 *) Support BS2000/OSD-POSIX.
8516 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8517
8518 *) Make callbacks for key generation use void * instead of char *.
8519 [Ben Laurie]
8520
8521 *) Make S/MIME samples compile (not yet tested).
8522 [Ben Laurie]
8523
8524 *) Additional typesafe stacks.
8525 [Ben Laurie]
8526
8527 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8528 [Bodo Moeller]
8529
8530
8531 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8532
8533 *) New configuration variant "sco5-gcc".
8534
8535 *) Updated some demos.
8536 [Sean O Riordain, Wade Scholine]
8537
8538 *) Add missing BIO_free at exit of pkcs12 application.
8539 [Wu Zhigang]
8540
8541 *) Fix memory leak in conf.c.
8542 [Steve Henson]
8543
8544 *) Updates for Win32 to assembler version of MD5.
8545 [Steve Henson]
8546
8547 *) Set #! path to perl in apps/der_chop to where we found it
8548 instead of using a fixed path.
8549 [Bodo Moeller]
8550
8551 *) SHA library changes for irix64-mips4-cc.
8552 [Andy Polyakov]
8553
8554 *) Improvements for VMS support.
8555 [Richard Levitte]
8556
8557
8558 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8559
8560 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8561 This also avoids the problems with SC4.2 and unpatched SC5.
8562 [Andy Polyakov <appro@fy.chalmers.se>]
8563
8564 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8565 These are required because of the typesafe stack would otherwise break
8566 existing code. If old code used a structure member which used to be STACK
8567 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8568 sk_num or sk_value it would produce an error because the num, data members
8569 are not present in STACK_OF. Now it just produces a warning. sk_set
8570 replaces the old method of assigning a value to sk_value
8571 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8572 that does this will no longer work (and should use sk_set instead) but
8573 this could be regarded as a "questionable" behaviour anyway.
8574 [Steve Henson]
8575
8576 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8577 correctly handle encrypted S/MIME data.
8578 [Steve Henson]
8579
8580 *) Change type of various DES function arguments from des_cblock
8581 (which means, in function argument declarations, pointer to char)
8582 to des_cblock * (meaning pointer to array with 8 char elements),
8583 which allows the compiler to do more typechecking; it was like
8584 that back in SSLeay, but with lots of ugly casts.
8585
8586 Introduce new type const_des_cblock.
8587 [Bodo Moeller]
8588
8589 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8590 problems: find RecipientInfo structure that matches recipient certificate
8591 and initialise the ASN1 structures properly based on passed cipher.
8592 [Steve Henson]
8593
8594 *) Belatedly make the BN tests actually check the results.
8595 [Ben Laurie]
8596
8597 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8598 to and from BNs: it was completely broken. New compilation option
8599 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8600 key elements as negative integers.
8601 [Steve Henson]
8602
8603 *) Reorganize and speed up MD5.
8604 [Andy Polyakov <appro@fy.chalmers.se>]
8605
8606 *) VMS support.
8607 [Richard Levitte <richard@levitte.org>]
8608
8609 *) New option -out to asn1parse to allow the parsed structure to be
8610 output to a file. This is most useful when combined with the -strparse
8611 option to examine the output of things like OCTET STRINGS.
8612 [Steve Henson]
8613
8614 *) Make SSL library a little more fool-proof by not requiring any longer
8615 that SSL_set_{accept,connect}_state be called before
8616 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8617 in many applications because usually everything *appeared* to work as
8618 intended anyway -- now it really works as intended).
8619 [Bodo Moeller]
8620
8621 *) Move openssl.cnf out of lib/.
8622 [Ulf Möller]
8623
8624 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8625 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8626 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8627 [Ralf S. Engelschall]
8628
8629 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8630 handle PKCS#7 enveloped data properly.
8631 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8632
8633 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8634 copying pointers. The cert_st handling is changed by this in
8635 various ways (and thus what used to be known as ctx->default_cert
8636 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8637 any longer when s->cert does not give us what we need).
8638 ssl_cert_instantiate becomes obsolete by this change.
8639 As soon as we've got the new code right (possibly it already is?),
8640 we have solved a couple of bugs of the earlier code where s->cert
8641 was used as if it could not have been shared with other SSL structures.
8642
8643 Note that using the SSL API in certain dirty ways now will result
8644 in different behaviour than observed with earlier library versions:
8645 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8646 does not influence s as it used to.
8647
8648 In order to clean up things more thoroughly, inside SSL_SESSION
8649 we don't use CERT any longer, but a new structure SESS_CERT
8650 that holds per-session data (if available); currently, this is
8651 the peer's certificate chain and, for clients, the server's certificate
8652 and temporary key. CERT holds only those values that can have
8653 meaningful defaults in an SSL_CTX.
8654 [Bodo Moeller]
8655
8656 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8657 from the internal representation. Various PKCS#7 fixes: remove some
8658 evil casts and set the enc_dig_alg field properly based on the signing
8659 key type.
8660 [Steve Henson]
8661
8662 *) Allow PKCS#12 password to be set from the command line or the
8663 environment. Let 'ca' get its config file name from the environment
8664 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8665 and 'x509').
8666 [Steve Henson]
8667
8668 *) Allow certificate policies extension to use an IA5STRING for the
8669 organization field. This is contrary to the PKIX definition but
8670 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8671 extension option.
8672 [Steve Henson]
8673
8674 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8675 without disallowing inline assembler and the like for non-pedantic builds.
8676 [Ben Laurie]
8677
8678 *) Support Borland C++ builder.
8679 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8680
8681 *) Support Mingw32.
8682 [Ulf Möller]
8683
8684 *) SHA-1 cleanups and performance enhancements.
8685 [Andy Polyakov <appro@fy.chalmers.se>]
8686
8687 *) Sparc v8plus assembler for the bignum library.
8688 [Andy Polyakov <appro@fy.chalmers.se>]
8689
8690 *) Accept any -xxx and +xxx compiler options in Configure.
8691 [Ulf Möller]
8692
8693 *) Update HPUX configuration.
8694 [Anonymous]
8695
8696 *) Add missing sk_<type>_unshift() function to safestack.h
8697 [Ralf S. Engelschall]
8698
8699 *) New function SSL_CTX_use_certificate_chain_file that sets the
8700 "extra_cert"s in addition to the certificate. (This makes sense
8701 only for "PEM" format files, as chains as a whole are not
8702 DER-encoded.)
8703 [Bodo Moeller]
8704
8705 *) Support verify_depth from the SSL API.
8706 x509_vfy.c had what can be considered an off-by-one-error:
8707 Its depth (which was not part of the external interface)
8708 was actually counting the number of certificates in a chain;
8709 now it really counts the depth.
8710 [Bodo Moeller]
8711
8712 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8713 instead of X509err, which often resulted in confusing error
8714 messages since the error codes are not globally unique
8715 (e.g. an alleged error in ssl3_accept when a certificate
8716 didn't match the private key).
8717
8718 *) New function SSL_CTX_set_session_id_context that allows to set a default
8719 value (so that you don't need SSL_set_session_id_context for each
8720 connection using the SSL_CTX).
8721 [Bodo Moeller]
8722
8723 *) OAEP decoding bug fix.
8724 [Ulf Möller]
8725
8726 *) Support INSTALL_PREFIX for package builders, as proposed by
8727 David Harris.
8728 [Bodo Moeller]
8729
8730 *) New Configure options "threads" and "no-threads". For systems
8731 where the proper compiler options are known (currently Solaris
8732 and Linux), "threads" is the default.
8733 [Bodo Moeller]
8734
8735 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8736 [Bodo Moeller]
8737
8738 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8739 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8740 such as /usr/local/bin.
8741 [Bodo Moeller]
8742
8743 *) "make linux-shared" to build shared libraries.
8744 [Niels Poppe <niels@netbox.org>]
8745
8746 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8747 [Ulf Möller]
8748
8749 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8750 extension adding in x509 utility.
8751 [Steve Henson]
8752
8753 *) Remove NOPROTO sections and error code comments.
8754 [Ulf Möller]
8755
8756 *) Partial rewrite of the DEF file generator to now parse the ANSI
8757 prototypes.
8758 [Steve Henson]
8759
8760 *) New Configure options --prefix=DIR and --openssldir=DIR.
8761 [Ulf Möller]
8762
8763 *) Complete rewrite of the error code script(s). It is all now handled
8764 by one script at the top level which handles error code gathering,
8765 header rewriting and C source file generation. It should be much better
8766 than the old method: it now uses a modified version of Ulf's parser to
8767 read the ANSI prototypes in all header files (thus the old K&R definitions
8768 aren't needed for error creation any more) and do a better job of
8769 translating function codes into names. The old 'ASN1 error code imbedded
8770 in a comment' is no longer necessary and it doesn't use .err files which
8771 have now been deleted. Also the error code call doesn't have to appear all
8772 on one line (which resulted in some large lines...).
8773 [Steve Henson]
8774
8775 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8776 [Bodo Moeller]
8777
8778 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8779 0 (which usually indicates a closed connection), but continue reading.
8780 [Bodo Moeller]
8781
8782 *) Fix some race conditions.
8783 [Bodo Moeller]
8784
8785 *) Add support for CRL distribution points extension. Add Certificate
8786 Policies and CRL distribution points documentation.
8787 [Steve Henson]
8788
8789 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8790 [Ulf Möller]
8791
8792 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8793 8 of keying material. Merlin has also confirmed interop with this fix
8794 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8795 [Merlin Hughes <merlin@baltimore.ie>]
8796
8797 *) Fix lots of warnings.
8798 [Richard Levitte <levitte@stacken.kth.se>]
8799
8800 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8801 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8802 [Richard Levitte <levitte@stacken.kth.se>]
8803
8804 *) Fix problems with sizeof(long) == 8.
8805 [Andy Polyakov <appro@fy.chalmers.se>]
8806
8807 *) Change functions to ANSI C.
8808 [Ulf Möller]
8809
8810 *) Fix typos in error codes.
8811 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8812
8813 *) Remove defunct assembler files from Configure.
8814 [Ulf Möller]
8815
8816 *) SPARC v8 assembler BIGNUM implementation.
8817 [Andy Polyakov <appro@fy.chalmers.se>]
8818
8819 *) Support for Certificate Policies extension: both print and set.
8820 Various additions to support the r2i method this uses.
8821 [Steve Henson]
8822
8823 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8824 return a const string when you are expecting an allocated buffer.
8825 [Ben Laurie]
8826
8827 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8828 types DirectoryString and DisplayText.
8829 [Steve Henson]
8830
8831 *) Add code to allow r2i extensions to access the configuration database,
8832 add an LHASH database driver and add several ctx helper functions.
8833 [Steve Henson]
8834
8835 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8836 fail when they extended the size of a BIGNUM.
8837 [Steve Henson]
8838
8839 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8840 support typesafe stack.
8841 [Steve Henson]
8842
8843 *) Fix typo in SSL_[gs]et_options().
8844 [Nils Frostberg <nils@medcom.se>]
8845
8846 *) Delete various functions and files that belonged to the (now obsolete)
8847 old X509V3 handling code.
8848 [Steve Henson]
8849
8850 *) New Configure option "rsaref".
8851 [Ulf Möller]
8852
8853 *) Don't auto-generate pem.h.
8854 [Bodo Moeller]
8855
8856 *) Introduce type-safe ASN.1 SETs.
8857 [Ben Laurie]
8858
8859 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8860 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8861
8862 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8863 that links with OpenSSL (well at least cause lots of warnings), but fear
8864 not: the conversion is trivial, and it eliminates loads of evil casts. A
8865 few STACKed things have been converted already. Feel free to convert more.
8866 In the fullness of time, I'll do away with the STACK type altogether.
8867 [Ben Laurie]
8868
8869 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8870 specified in <certfile> by updating the entry in the index.txt file.
8871 This way one no longer has to edit the index.txt file manually for
8872 revoking a certificate. The -revoke option does the gory details now.
8873 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8874
8875 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8876 `-text' option at all and this way the `-noout -text' combination was
8877 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8878 [Ralf S. Engelschall]
8879
8880 *) Make sure a corresponding plain text error message exists for the
8881 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8882 verify callback function determined that a certificate was revoked.
8883 [Ralf S. Engelschall]
8884
8885 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8886 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8887 all available cipers including rc5, which was forgotten until now.
8888 In order to let the testing shell script know which algorithms
8889 are available, a new (up to now undocumented) command
8890 "openssl list-cipher-commands" is used.
8891 [Bodo Moeller]
8892
8893 *) Bugfix: s_client occasionally would sleep in select() when
8894 it should have checked SSL_pending() first.
8895 [Bodo Moeller]
8896
8897 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8898 the raw DSA values prior to ASN.1 encoding.
8899 [Ulf Möller]
8900
8901 *) Tweaks to Configure
8902 [Niels Poppe <niels@netbox.org>]
8903
8904 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8905 yet...
8906 [Steve Henson]
8907
8908 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8909 [Ulf Möller]
8910
8911 *) New config option to avoid instructions that are illegal on the 80386.
8912 The default code is faster, but requires at least a 486.
8913 [Ulf Möller]
8914
8915 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8916 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8917 same as SSL2_VERSION anyway.
8918 [Bodo Moeller]
8919
8920 *) New "-showcerts" option for s_client.
8921 [Bodo Moeller]
8922
8923 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8924 application. Various cleanups and fixes.
8925 [Steve Henson]
8926
8927 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8928 modify error routines to work internally. Add error codes and PBE init
8929 to library startup routines.
8930 [Steve Henson]
8931
8932 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8933 packing functions to asn1 and evp. Changed function names and error
8934 codes along the way.
8935 [Steve Henson]
8936
8937 *) PKCS12 integration: and so it begins... First of several patches to
8938 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8939 objects to objects.h
8940 [Steve Henson]
8941
8942 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8943 and display support for Thawte strong extranet extension.
8944 [Steve Henson]
8945
8946 *) Add LinuxPPC support.
8947 [Jeff Dubrule <igor@pobox.org>]
8948
8949 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8950 bn_div_words in alpha.s.
8951 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8952
8953 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8954 OAEP isn't supported when OpenSSL is built with RSAref.
8955 [Ulf Moeller <ulf@fitug.de>]
8956
8957 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8958 so they no longer are missing under -DNOPROTO.
8959 [Soren S. Jorvang <soren@t.dk>]
8960
8961
8962 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8963
8964 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8965 doesn't work when the session is reused. Coming soon!
8966 [Ben Laurie]
8967
8968 *) Fix a security hole, that allows sessions to be reused in the wrong
8969 context thus bypassing client cert protection! All software that uses
8970 client certs and session caches in multiple contexts NEEDS PATCHING to
8971 allow session reuse! A fuller solution is in the works.
8972 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8973
8974 *) Some more source tree cleanups (removed obsolete files
8975 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8976 permission on "config" script to be executable) and a fix for the INSTALL
8977 document.
8978 [Ulf Moeller <ulf@fitug.de>]
8979
8980 *) Remove some legacy and erroneous uses of malloc, free instead of
8981 Malloc, Free.
8982 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8983
8984 *) Make rsa_oaep_test return non-zero on error.
8985 [Ulf Moeller <ulf@fitug.de>]
8986
8987 *) Add support for native Solaris shared libraries. Configure
8988 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8989 if someone would make that last step automatic.
8990 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8991
8992 *) ctx_size was not built with the right compiler during "make links". Fixed.
8993 [Ben Laurie]
8994
8995 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8996 except NULL ciphers". This means the default cipher list will no longer
8997 enable NULL ciphers. They need to be specifically enabled e.g. with
8998 the string "DEFAULT:eNULL".
8999 [Steve Henson]
9000
9001 *) Fix to RSA private encryption routines: if p < q then it would
9002 occasionally produce an invalid result. This will only happen with
9003 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9004 [Steve Henson]
9005
9006 *) Be less restrictive and allow also `perl util/perlpath.pl
9007 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9008 because this way one can also use an interpreter named `perl5' (which is
9009 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9010 installed as `perl').
9011 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9012
9013 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9014 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9015
9016 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9017 advapi32.lib to Win32 build and change the pem test comparision
9018 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9019 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9020 and crypto/des/ede_cbcm_enc.c.
9021 [Steve Henson]
9022
9023 *) DES quad checksum was broken on big-endian architectures. Fixed.
9024 [Ben Laurie]
9025
9026 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9027 Win32 test batch file so it (might) work again. The Win32 test batch file
9028 is horrible: I feel ill....
9029 [Steve Henson]
9030
9031 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9032 in e_os.h. Audit of header files to check ANSI and non ANSI
9033 sections: 10 functions were absent from non ANSI section and not exported
9034 from Windows DLLs. Fixed up libeay.num for new functions.
9035 [Steve Henson]
9036
9037 *) Make `openssl version' output lines consistent.
9038 [Ralf S. Engelschall]
9039
9040 *) Fix Win32 symbol export lists for BIO functions: Added
9041 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9042 to ms/libeay{16,32}.def.
9043 [Ralf S. Engelschall]
9044
9045 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9046 fine under Unix and passes some trivial tests I've now added. But the
9047 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9048 added to make sure no one expects that this stuff really works in the
9049 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9050 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9051 openssl_bio.xs.
9052 [Ralf S. Engelschall]
9053
9054 *) Fix the generation of two part addresses in perl.
9055 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9056
9057 *) Add config entry for Linux on MIPS.
9058 [John Tobey <jtobey@channel1.com>]
9059
9060 *) Make links whenever Configure is run, unless we are on Windoze.
9061 [Ben Laurie]
9062
9063 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9064 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9065 in CRLs.
9066 [Steve Henson]
9067
9068 *) Add a useful kludge to allow package maintainers to specify compiler and
9069 other platforms details on the command line without having to patch the
9070 Configure script everytime: One now can use ``perl Configure
9071 <id>:<details>'', i.e. platform ids are allowed to have details appended
9072 to them (seperated by colons). This is treated as there would be a static
9073 pre-configured entry in Configure's %table under key <id> with value
9074 <details> and ``perl Configure <id>'' is called. So, when you want to
9075 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9076 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9077 now, which overrides the FreeBSD-elf entry on-the-fly.
9078 [Ralf S. Engelschall]
9079
9080 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9081 [Ben Laurie]
9082
9083 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9084 on the `perl Configure ...' command line. This way one can compile
9085 OpenSSL libraries with Position Independent Code (PIC) which is needed
9086 for linking it into DSOs.
9087 [Ralf S. Engelschall]
9088
9089 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9090 Fixed.
9091 [Ben Laurie]
9092
9093 *) Cleaned up the LICENSE document: The official contact for any license
9094 questions now is the OpenSSL core team under openssl-core@openssl.org.
9095 And add a paragraph about the dual-license situation to make sure people
9096 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9097 to the OpenSSL toolkit.
9098 [Ralf S. Engelschall]
9099
9100 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9101 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9102 Additonally cleaned up the `make links' target: Remove unnecessary
9103 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9104 to speed processing and no longer clutter the display with confusing
9105 stuff. Instead only the actually done links are displayed.
9106 [Ralf S. Engelschall]
9107
9108 *) Permit null encryption ciphersuites, used for authentication only. It used
9109 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9110 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9111 encryption.
9112 [Ben Laurie]
9113
9114 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9115 signed attributes when verifying signatures (this would break them),
9116 the detached data encoding was wrong and public keys obtained using
9117 X509_get_pubkey() weren't freed.
9118 [Steve Henson]
9119
9120 *) Add text documentation for the BUFFER functions. Also added a work around
9121 to a Win95 console bug. This was triggered by the password read stuff: the
9122 last character typed gets carried over to the next fread(). If you were
9123 generating a new cert request using 'req' for example then the last
9124 character of the passphrase would be CR which would then enter the first
9125 field as blank.
9126 [Steve Henson]
9127
9128 *) Added the new `Includes OpenSSL Cryptography Software' button as
9129 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9130 button and can be used by applications based on OpenSSL to show the
9131 relationship to the OpenSSL project.
9132 [Ralf S. Engelschall]
9133
9134 *) Remove confusing variables in function signatures in files
9135 ssl/ssl_lib.c and ssl/ssl.h.
9136 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9137
9138 *) Don't install bss_file.c under PREFIX/include/
9139 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9140
9141 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9142 functions that return function pointers and has support for NT specific
9143 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9144 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9145 unsigned to signed types: this was killing the Win32 compile.
9146 [Steve Henson]
9147
9148 *) Add new certificate file to stack functions,
9149 SSL_add_dir_cert_subjects_to_stack() and
9150 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9151 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9152 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9153 This means that Apache-SSL and similar packages don't have to mess around
9154 to add as many CAs as they want to the preferred list.
9155 [Ben Laurie]
9156
9157 *) Experiment with doxygen documentation. Currently only partially applied to
9158 ssl/ssl_lib.c.
9159 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9160 openssl.doxy as the configuration file.
9161 [Ben Laurie]
9162
9163 *) Get rid of remaining C++-style comments which strict C compilers hate.
9164 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9165
9166 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9167 compiled in by default: it has problems with large keys.
9168 [Steve Henson]
9169
9170 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9171 DH private keys and/or callback functions which directly correspond to
9172 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9173 is needed for applications which have to configure certificates on a
9174 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9175 (e.g. s_server).
9176 For the RSA certificate situation is makes no difference, but
9177 for the DSA certificate situation this fixes the "no shared cipher"
9178 problem where the OpenSSL cipher selection procedure failed because the
9179 temporary keys were not overtaken from the context and the API provided
9180 no way to reconfigure them.
9181 The new functions now let applications reconfigure the stuff and they
9182 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9183 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9184 non-public-API function ssl_cert_instantiate() is used as a helper
9185 function and also to reduce code redundancy inside ssl_rsa.c.
9186 [Ralf S. Engelschall]
9187
9188 *) Move s_server -dcert and -dkey options out of the undocumented feature
9189 area because they are useful for the DSA situation and should be
9190 recognized by the users.
9191 [Ralf S. Engelschall]
9192
9193 *) Fix the cipher decision scheme for export ciphers: the export bits are
9194 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9195 SSL_EXP_MASK. So, the original variable has to be used instead of the
9196 already masked variable.
9197 [Richard Levitte <levitte@stacken.kth.se>]
9198
9199 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9200 [Richard Levitte <levitte@stacken.kth.se>]
9201
9202 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9203 from `int' to `unsigned int' because it's a length and initialized by
9204 EVP_DigestFinal() which expects an `unsigned int *'.
9205 [Richard Levitte <levitte@stacken.kth.se>]
9206
9207 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9208 script. Instead use the usual Shell->Perl transition trick.
9209 [Ralf S. Engelschall]
9210
9211 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9212 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9213 -noout -modulus' as it's already the case for `openssl rsa -noout
9214 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9215 currently the public key is printed (a decision which was already done by
9216 `openssl dsa -modulus' in the past) which serves a similar purpose.
9217 Additionally the NO_RSA no longer completely removes the whole -modulus
9218 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9219 now, too.
9220 [Ralf S. Engelschall]
9221
9222 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9223 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9224 [Arne Ansper <arne@ats.cyber.ee>]
9225
9226 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9227 to be added. Now both 'req' and 'ca' can use new objects defined in the
9228 config file.
9229 [Steve Henson]
9230
9231 *) Add cool BIO that does syslog (or event log on NT).
9232 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9233
9234 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9235 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9236 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9237 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9238 [Ben Laurie]
9239
9240 *) Add preliminary config info for new extension code.
9241 [Steve Henson]
9242
9243 *) Make RSA_NO_PADDING really use no padding.
9244 [Ulf Moeller <ulf@fitug.de>]
9245
9246 *) Generate errors when private/public key check is done.
9247 [Ben Laurie]
9248
9249 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9250 for some CRL extensions and new objects added.
9251 [Steve Henson]
9252
9253 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9254 key usage extension and fuller support for authority key id.
9255 [Steve Henson]
9256
9257 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9258 padding method for RSA, which is recommended for new applications in PKCS
9259 #1 v2.0 (RFC 2437, October 1998).
9260 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9261 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9262 against Bleichbacher's attack on RSA.
9263 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9264 Ben Laurie]
9265
9266 *) Updates to the new SSL compression code
9267 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9268
9269 *) Fix so that the version number in the master secret, when passed
9270 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9271 (because the server will not accept higher), that the version number
9272 is 0x03,0x01, not 0x03,0x00
9273 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9274
9275 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9276 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9277 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9278 [Steve Henson]
9279
9280 *) Support for RAW extensions where an arbitrary extension can be
9281 created by including its DER encoding. See apps/openssl.cnf for
9282 an example.
9283 [Steve Henson]
9284
9285 *) Make sure latest Perl versions don't interpret some generated C array
9286 code as Perl array code in the crypto/err/err_genc.pl script.
9287 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9288
9289 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9290 not many people have the assembler. Various Win32 compilation fixes and
9291 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9292 build instructions.
9293 [Steve Henson]
9294
9295 *) Modify configure script 'Configure' to automatically create crypto/date.h
9296 file under Win32 and also build pem.h from pem.org. New script
9297 util/mkfiles.pl to create the MINFO file on environments that can't do a
9298 'make files': perl util/mkfiles.pl >MINFO should work.
9299 [Steve Henson]
9300
9301 *) Major rework of DES function declarations, in the pursuit of correctness
9302 and purity. As a result, many evil casts evaporated, and some weirdness,
9303 too. You may find this causes warnings in your code. Zapping your evil
9304 casts will probably fix them. Mostly.
9305 [Ben Laurie]
9306
9307 *) Fix for a typo in asn1.h. Bug fix to object creation script
9308 obj_dat.pl. It considered a zero in an object definition to mean
9309 "end of object": none of the objects in objects.h have any zeros
9310 so it wasn't spotted.
9311 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9312
9313 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9314 Masking (CBCM). In the absence of test vectors, the best I have been able
9315 to do is check that the decrypt undoes the encrypt, so far. Send me test
9316 vectors if you have them.
9317 [Ben Laurie]
9318
9319 *) Correct calculation of key length for export ciphers (too much space was
9320 allocated for null ciphers). This has not been tested!
9321 [Ben Laurie]
9322
9323 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9324 message is now correct (it understands "crypto" and "ssl" on its
9325 command line). There is also now an "update" option. This will update
9326 the util/ssleay.num and util/libeay.num files with any new functions.
9327 If you do a:
9328 perl util/mkdef.pl crypto ssl update
9329 it will update them.
9330 [Steve Henson]
9331
9332 *) Overhauled the Perl interface (perl/*):
9333 - ported BN stuff to OpenSSL's different BN library
9334 - made the perl/ source tree CVS-aware
9335 - renamed the package from SSLeay to OpenSSL (the files still contain
9336 their history because I've copied them in the repository)
9337 - removed obsolete files (the test scripts will be replaced
9338 by better Test::Harness variants in the future)
9339 [Ralf S. Engelschall]
9340
9341 *) First cut for a very conservative source tree cleanup:
9342 1. merge various obsolete readme texts into doc/ssleay.txt
9343 where we collect the old documents and readme texts.
9344 2. remove the first part of files where I'm already sure that we no
9345 longer need them because of three reasons: either they are just temporary
9346 files which were left by Eric or they are preserved original files where
9347 I've verified that the diff is also available in the CVS via "cvs diff
9348 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9349 the crypto/md/ stuff).
9350 [Ralf S. Engelschall]
9351
9352 *) More extension code. Incomplete support for subject and issuer alt
9353 name, issuer and authority key id. Change the i2v function parameters
9354 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9355 what that's for :-) Fix to ASN1 macro which messed up
9356 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9357 [Steve Henson]
9358
9359 *) Preliminary support for ENUMERATED type. This is largely copied from the
9360 INTEGER code.
9361 [Steve Henson]
9362
9363 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9364 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9365
9366 *) Make sure `make rehash' target really finds the `openssl' program.
9367 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9368
9369 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9370 like to hear about it if this slows down other processors.
9371 [Ben Laurie]
9372
9373 *) Add CygWin32 platform information to Configure script.
9374 [Alan Batie <batie@aahz.jf.intel.com>]
9375
9376 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9377 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9378
9379 *) New program nseq to manipulate netscape certificate sequences
9380 [Steve Henson]
9381
9382 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9383 few typos.
9384 [Steve Henson]
9385
9386 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9387 but the BN code had some problems that would cause failures when
9388 doing certificate verification and some other functions.
9389 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9390
9391 *) Add ASN1 and PEM code to support netscape certificate sequences.
9392 [Steve Henson]
9393
9394 *) Add ASN1 and PEM code to support netscape certificate sequences.
9395 [Steve Henson]
9396
9397 *) Add several PKIX and private extended key usage OIDs.
9398 [Steve Henson]
9399
9400 *) Modify the 'ca' program to handle the new extension code. Modify
9401 openssl.cnf for new extension format, add comments.
9402 [Steve Henson]
9403
9404 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9405 and add a sample to openssl.cnf so req -x509 now adds appropriate
9406 CA extensions.
9407 [Steve Henson]
9408
9409 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9410 error code, add initial support to X509_print() and x509 application.
9411 [Steve Henson]
9412
9413 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9414 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9415 stuff is currently isolated and isn't even compiled yet.
9416 [Steve Henson]
9417
9418 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9419 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9420 Removed the versions check from X509 routines when loading extensions:
9421 this allows certain broken certificates that don't set the version
9422 properly to be processed.
9423 [Steve Henson]
9424
9425 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9426 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9427 can still be regenerated with "make depend".
9428 [Ben Laurie]
9429
9430 *) Spelling mistake in C version of CAST-128.
9431 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9432
9433 *) Changes to the error generation code. The perl script err-code.pl
9434 now reads in the old error codes and retains the old numbers, only
9435 adding new ones if necessary. It also only changes the .err files if new
9436 codes are added. The makefiles have been modified to only insert errors
9437 when needed (to avoid needlessly modifying header files). This is done
9438 by only inserting errors if the .err file is newer than the auto generated
9439 C file. To rebuild all the error codes from scratch (the old behaviour)
9440 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9441 or delete all the .err files.
9442 [Steve Henson]
9443
9444 *) CAST-128 was incorrectly implemented for short keys. The C version has
9445 been fixed, but is untested. The assembler versions are also fixed, but
9446 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9447 to regenerate it if needed.
9448 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9449 Hagino <itojun@kame.net>]
9450
9451 *) File was opened incorrectly in randfile.c.
9452 [Ulf Möller <ulf@fitug.de>]
9453
9454 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9455 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9456 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9457 al: it's just almost always a UTCTime. Note this patch adds new error
9458 codes so do a "make errors" if there are problems.
9459 [Steve Henson]
9460
9461 *) Correct Linux 1 recognition in config.
9462 [Ulf Möller <ulf@fitug.de>]
9463
9464 *) Remove pointless MD5 hash when using DSA keys in ca.
9465 [Anonymous <nobody@replay.com>]
9466
9467 *) Generate an error if given an empty string as a cert directory. Also
9468 generate an error if handed NULL (previously returned 0 to indicate an
9469 error, but didn't set one).
9470 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9471
9472 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9473 [Ben Laurie]
9474
9475 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9476 parameters. This was causing a warning which killed off the Win32 compile.
9477 [Steve Henson]
9478
9479 *) Remove C++ style comments from crypto/bn/bn_local.h.
9480 [Neil Costigan <neil.costigan@celocom.com>]
9481
9482 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9483 based on a text string, looking up short and long names and finally
9484 "dot" format. The "dot" format stuff didn't work. Added new function
9485 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9486 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9487 OID is not part of the table.
9488 [Steve Henson]
9489
9490 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9491 X509_LOOKUP_by_alias().
9492 [Ben Laurie]
9493
9494 *) Sort openssl functions by name.
9495 [Ben Laurie]
9496
9497 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9498 encryption from sample DSA keys (in case anyone is interested the password
9499 was "1234").
9500 [Steve Henson]
9501
9502 *) Make _all_ *_free functions accept a NULL pointer.
9503 [Frans Heymans <fheymans@isaserver.be>]
9504
9505 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9506 NULL pointers.
9507 [Anonymous <nobody@replay.com>]
9508
9509 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9510 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9511
9512 *) Don't blow it for numeric -newkey arguments to apps/req.
9513 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9514
9515 *) Temp key "for export" tests were wrong in s3_srvr.c.
9516 [Anonymous <nobody@replay.com>]
9517
9518 *) Add prototype for temp key callback functions
9519 SSL_CTX_set_tmp_{rsa,dh}_callback().
9520 [Ben Laurie]
9521
9522 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9523 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9524 [Steve Henson]
9525
9526 *) X509_name_add_entry() freed the wrong thing after an error.
9527 [Arne Ansper <arne@ats.cyber.ee>]
9528
9529 *) rsa_eay.c would attempt to free a NULL context.
9530 [Arne Ansper <arne@ats.cyber.ee>]
9531
9532 *) BIO_s_socket() had a broken should_retry() on Windoze.
9533 [Arne Ansper <arne@ats.cyber.ee>]
9534
9535 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9536 [Arne Ansper <arne@ats.cyber.ee>]
9537
9538 *) Make sure the already existing X509_STORE->depth variable is initialized
9539 in X509_STORE_new(), but document the fact that this variable is still
9540 unused in the certificate verification process.
9541 [Ralf S. Engelschall]
9542
9543 *) Fix the various library and apps files to free up pkeys obtained from
9544 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9545 [Steve Henson]
9546
9547 *) Fix reference counting in X509_PUBKEY_get(). This makes
9548 demos/maurice/example2.c work, amongst others, probably.
9549 [Steve Henson and Ben Laurie]
9550
9551 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9552 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9553 are no longer created. This way we have a single and consistent command
9554 line interface `openssl <command>', similar to `cvs <command>'.
9555 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9556
9557 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9558 BIT STRING wrapper always have zero unused bits.
9559 [Steve Henson]
9560
9561 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9562 [Steve Henson]
9563
9564 *) Make the top-level INSTALL documentation easier to understand.
9565 [Paul Sutton]
9566
9567 *) Makefiles updated to exit if an error occurs in a sub-directory
9568 make (including if user presses ^C) [Paul Sutton]
9569
9570 *) Make Montgomery context stuff explicit in RSA data structure.
9571 [Ben Laurie]
9572
9573 *) Fix build order of pem and err to allow for generated pem.h.
9574 [Ben Laurie]
9575
9576 *) Fix renumbering bug in X509_NAME_delete_entry().
9577 [Ben Laurie]
9578
9579 *) Enhanced the err-ins.pl script so it makes the error library number
9580 global and can add a library name. This is needed for external ASN1 and
9581 other error libraries.
9582 [Steve Henson]
9583
9584 *) Fixed sk_insert which never worked properly.
9585 [Steve Henson]
9586
9587 *) Fix ASN1 macros so they can handle indefinite length construted
9588 EXPLICIT tags. Some non standard certificates use these: they can now
9589 be read in.
9590 [Steve Henson]
9591
9592 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9593 into a single doc/ssleay.txt bundle. This way the information is still
9594 preserved but no longer messes up this directory. Now it's new room for
9595 the new set of documenation files.
9596 [Ralf S. Engelschall]
9597
9598 *) SETs were incorrectly DER encoded. This was a major pain, because they
9599 shared code with SEQUENCEs, which aren't coded the same. This means that
9600 almost everything to do with SETs or SEQUENCEs has either changed name or
9601 number of arguments.
9602 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9603
9604 *) Fix test data to work with the above.
9605 [Ben Laurie]
9606
9607 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9608 was already fixed by Eric for 0.9.1 it seems.
9609 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9610
9611 *) Autodetect FreeBSD3.
9612 [Ben Laurie]
9613
9614 *) Fix various bugs in Configure. This affects the following platforms:
9615 nextstep
9616 ncr-scde
9617 unixware-2.0
9618 unixware-2.0-pentium
9619 sco5-cc.
9620 [Ben Laurie]
9621
9622 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9623 before they are needed.
9624 [Ben Laurie]
9625
9626 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9627 [Ben Laurie]
9628
9629
9630 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9631
9632 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9633 changed SSLeay to OpenSSL in version strings.
9634 [Ralf S. Engelschall]
9635
9636 *) Some fixups to the top-level documents.
9637 [Paul Sutton]
9638
9639 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9640 because the symlink to include/ was missing.
9641 [Ralf S. Engelschall]
9642
9643 *) Incorporated the popular no-RSA/DSA-only patches
9644 which allow to compile a RSA-free SSLeay.
9645 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9646
9647 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9648 when "ssleay" is still not found.
9649 [Ralf S. Engelschall]
9650
9651 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9652 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9653
9654 *) Updated the README file.
9655 [Ralf S. Engelschall]
9656
9657 *) Added various .cvsignore files in the CVS repository subdirs
9658 to make a "cvs update" really silent.
9659 [Ralf S. Engelschall]
9660
9661 *) Recompiled the error-definition header files and added
9662 missing symbols to the Win32 linker tables.
9663 [Ralf S. Engelschall]
9664
9665 *) Cleaned up the top-level documents;
9666 o new files: CHANGES and LICENSE
9667 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9668 o merged COPYRIGHT into LICENSE
9669 o removed obsolete TODO file
9670 o renamed MICROSOFT to INSTALL.W32
9671 [Ralf S. Engelschall]
9672
9673 *) Removed dummy files from the 0.9.1b source tree:
9674 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9675 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9676 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9677 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9678 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9679 [Ralf S. Engelschall]
9680
9681 *) Added various platform portability fixes.
9682 [Mark J. Cox]
9683
9684 *) The Genesis of the OpenSSL rpject:
9685 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9686 Young and Tim J. Hudson created while they were working for C2Net until
9687 summer 1998.
9688 [The OpenSSL Project]
9689
9690
9691 Changes between 0.9.0b and 0.9.1b [not released]
9692
9693 *) Updated a few CA certificates under certs/
9694 [Eric A. Young]
9695
9696 *) Changed some BIGNUM api stuff.
9697 [Eric A. Young]
9698
9699 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9700 DGUX x86, Linux Alpha, etc.
9701 [Eric A. Young]
9702
9703 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9704 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9705 available).
9706 [Eric A. Young]
9707
9708 *) Add -strparse option to asn1pars program which parses nested
9709 binary structures
9710 [Dr Stephen Henson <shenson@bigfoot.com>]
9711
9712 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9713 [Eric A. Young]
9714
9715 *) DSA fix for "ca" program.
9716 [Eric A. Young]
9717
9718 *) Added "-genkey" option to "dsaparam" program.
9719 [Eric A. Young]
9720
9721 *) Added RIPE MD160 (rmd160) message digest.
9722 [Eric A. Young]
9723
9724 *) Added -a (all) option to "ssleay version" command.
9725 [Eric A. Young]
9726
9727 *) Added PLATFORM define which is the id given to Configure.
9728 [Eric A. Young]
9729
9730 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9731 [Eric A. Young]
9732
9733 *) Extended the ASN.1 parser routines.
9734 [Eric A. Young]
9735
9736 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9737 [Eric A. Young]
9738
9739 *) Added a BN_CTX to the BN library.
9740 [Eric A. Young]
9741
9742 *) Fixed the weak key values in DES library
9743 [Eric A. Young]
9744
9745 *) Changed API in EVP library for cipher aliases.
9746 [Eric A. Young]
9747
9748 *) Added support for RC2/64bit cipher.
9749 [Eric A. Young]
9750
9751 *) Converted the lhash library to the crypto/mem.c functions.
9752 [Eric A. Young]
9753
9754 *) Added more recognized ASN.1 object ids.
9755 [Eric A. Young]
9756
9757 *) Added more RSA padding checks for SSL/TLS.
9758 [Eric A. Young]
9759
9760 *) Added BIO proxy/filter functionality.
9761 [Eric A. Young]
9762
9763 *) Added extra_certs to SSL_CTX which can be used
9764 send extra CA certificates to the client in the CA cert chain sending
9765 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9766 [Eric A. Young]
9767
9768 *) Now Fortezza is denied in the authentication phase because
9769 this is key exchange mechanism is not supported by SSLeay at all.
9770 [Eric A. Young]
9771
9772 *) Additional PKCS1 checks.
9773 [Eric A. Young]
9774
9775 *) Support the string "TLSv1" for all TLS v1 ciphers.
9776 [Eric A. Young]
9777
9778 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9779 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9780 [Eric A. Young]
9781
9782 *) Fixed a few memory leaks.
9783 [Eric A. Young]
9784
9785 *) Fixed various code and comment typos.
9786 [Eric A. Young]
9787
9788 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9789 bytes sent in the client random.
9790 [Edward Bishop <ebishop@spyglass.com>]
9791