]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Add no-asm argument to Configure if needed.
[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 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
8 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
9 outside the validated module in the FIPS capable OpenSSL.
10 [Steve Henson]
11
12 *) Minor change to DRBG entropy callback semantics. In some cases
13 there is no mutiple of the block length between min_len and
14 max_len. Allow the callback to return more than max_len bytes
15 of entropy but discard any extra: it is the callback's responsibility
16 to ensure that the extra data discarded does not impact the
17 requested amount of entropy.
18 [Steve Henson]
19
20 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
21 information in FIPS186-3, SP800-57 and SP800-131A.
22 [Steve Henson]
23
24 *) CCM support via EVP. Interface is very similar to GCM case except we
25 must supply all data in one chunk (i.e. no update, final) and the
26 message length must be supplied if AAD is used. Add algorithm test
27 support.
28 [Steve Henson]
29
30 *) Initial version of POST overhaul. Add POST callback to allow the status
31 of POST to be monitored and/or failures induced. Modify fips_test_suite
32 to use callback. Always run all selftests even if one fails.
33 [Steve Henson]
34
35 *) XTS support including algorithm test driver in the fips_gcmtest program.
36 Note: this does increase the maximum key length from 32 to 64 bytes but
37 there should be no binary compatibility issues as existing applications
38 will never use XTS mode.
39 [Steve Henson]
40
41 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
42 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
43 performs algorithm blocking for unapproved PRNG types. Also do not
44 set PRNG type in FIPS_mode_set(): leave this to the application.
45 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
46 the standard OpenSSL PRNG: set additional data to a date time vector.
47 [Steve Henson]
48
49 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
50 This shouldn't present any incompatibility problems because applications
51 shouldn't be using these directly and any that are will need to rethink
52 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
53 [Steve Henson]
54
55 *) Extensive self tests and health checking required by SP800-90 DRBG.
56 Remove strength parameter from FIPS_drbg_instantiate and always
57 instantiate at maximum supported strength.
58 [Steve Henson]
59
60 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
61 [Steve Henson]
62
63 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
64 [Steve Henson]
65
66 *) New function DH_compute_key_padded() to compute a DH key and pad with
67 leading zeroes if needed: this complies with SP800-56A et al.
68 [Steve Henson]
69
70 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
71 anything, incomplete, subject to change and largely untested at present.
72 [Steve Henson]
73
74 *) Modify fipscanisteronly build option to only build the necessary object
75 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
76 [Steve Henson]
77
78 *) Add experimental option FIPSSYMS to give all symbols in
79 fipscanister.o and FIPS or fips prefix. This will avoid
80 conflicts with future versions of OpenSSL. Add perl script
81 util/fipsas.pl to preprocess assembly language source files
82 and rename any affected symbols.
83 [Steve Henson]
84
85 *) Add selftest checks and algorithm block of non-fips algorithms in
86 FIPS mode. Remove DES2 from selftests.
87 [Steve Henson]
88
89 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
90 return internal method without any ENGINE dependencies. Add new
91 tiny fips sign and verify functions.
92 [Steve Henson]
93
94 *) New build option no-ec2m to disable characteristic 2 code.
95 [Steve Henson]
96
97 *) New build option "fipscanisteronly". This only builds fipscanister.o
98 and (currently) associated fips utilities. Uses the file Makefile.fips
99 instead of Makefile.org as the prototype.
100 [Steve Henson]
101
102 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
103 Update fips_gcmtest to use IV generator.
104 [Steve Henson]
105
106 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
107 setting output buffer to NULL. The *Final function must be
108 called although it will not retrieve any additional data. The tag
109 can be set or retrieved with a ctrl. The IV length is by default 12
110 bytes (96 bits) but can be set to an alternative value. If the IV
111 length exceeds the maximum IV length (currently 16 bytes) it cannot be
112 set before the key.
113 [Steve Henson]
114
115 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
116 underlying do_cipher function handles all cipher semantics itself
117 including padding and finalisation. This is useful if (for example)
118 an ENGINE cipher handles block padding itself. The behaviour of
119 do_cipher is subtly changed if this flag is set: the return value
120 is the number of characters written to the output buffer (zero is
121 no longer an error code) or a negative error code. Also if the
122 input buffer is NULL and length 0 finalisation should be performed.
123 [Steve Henson]
124
125 *) If a candidate issuer certificate is already part of the constructed
126 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
127 [Steve Henson]
128
129 *) Improve forward-security support: add functions
130
131 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
132 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
133
134 for use by SSL/TLS servers; the callback function will be called whenever a
135 new session is created, and gets to decide whether the session may be
136 cached to make it resumable (return 0) or not (return 1). (As by the
137 SSL/TLS protocol specifications, the session_id sent by the server will be
138 empty to indicate that the session is not resumable; also, the server will
139 not generate RFC 4507 (RFC 5077) session tickets.)
140
141 A simple reasonable callback implementation is to return is_forward_secure.
142 This parameter will be set to 1 or 0 depending on the ciphersuite selected
143 by the SSL/TLS server library, indicating whether it can provide forward
144 security.
145 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
146
147 *) Add Next Protocol Negotiation,
148 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
149 disabled with a no-npn flag to config or Configure. Code donated
150 by Google.
151 [Adam Langley <agl@google.com> and Ben Laurie]
152
153 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
154 all platforms. Move ssize_t definition from e_os.h to the public
155 header file e_os2.h as it now appears in public header file cms.h
156 [Steve Henson]
157
158 *) New function OPENSSL_gmtime_diff to find the difference in days
159 and seconds between two tm structures. This will be used to provide
160 additional functionality for ASN1_TIME.
161 [Steve Henson]
162
163 *) New -sigopt option to the ca, req and x509 utilities. Additional
164 signature parameters can be passed using this option and in
165 particular PSS.
166 [Steve Henson]
167
168 *) Add RSA PSS signing function. This will generate and set the
169 appropriate AlgorithmIdentifiers for PSS based on those in the
170 corresponding EVP_MD_CTX structure. No application support yet.
171 [Steve Henson]
172
173 *) Support for companion algorithm specific ASN1 signing routines.
174 New function ASN1_item_sign_ctx() signs a pre-initialised
175 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
176 the appropriate parameters.
177 [Steve Henson]
178
179 *) Add new algorithm specific ASN1 verification initialisation function
180 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
181 handling will be the same no matter what EVP_PKEY_METHOD is used.
182 Add a PSS handler to support verification of PSS signatures: checked
183 against a number of sample certificates.
184 [Steve Henson]
185
186 *) Add signature printing for PSS. Add PSS OIDs.
187 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
188
189 *) Add algorithm specific signature printing. An individual ASN1 method
190 can now print out signatures instead of the standard hex dump.
191
192 More complex signatures (e.g. PSS) can print out more meaningful
193 information. Include DSA version that prints out the signature
194 parameters r, s.
195 [Steve Henson]
196
197 *) Add -trusted_first option which attempts to find certificates in the
198 trusted store even if an untrusted chain is also supplied.
199 [Steve Henson]
200
201 *) Initial experimental support for explicitly trusted non-root CAs.
202 OpenSSL still tries to build a complete chain to a root but if an
203 intermediate CA has a trust setting included that is used. The first
204 setting is used: whether to trust or reject.
205 [Steve Henson]
206
207 *) New -verify_name option in command line utilities to set verification
208 parameters by name.
209 [Steve Henson]
210
211 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
212 Add CMAC pkey methods.
213 [Steve Henson]
214
215 *) Experiemental regnegotiation in s_server -www mode. If the client
216 browses /reneg connection is renegotiated. If /renegcert it is
217 renegotiated requesting a certificate.
218 [Steve Henson]
219
220 *) Add an "external" session cache for debugging purposes to s_server. This
221 should help trace issues which normally are only apparent in deployed
222 multi-process servers.
223 [Steve Henson]
224
225 *) Experiemental password based recipient info support for CMS library:
226 implementing RFC3211.
227 [Steve Henson]
228
229 *) Split password based encryption into PBES2 and PBKDF2 functions. This
230 neatly separates the code into cipher and PBE sections and is required
231 for some algorithms that split PBES2 into separate pieces (such as
232 password based CMS).
233 [Steve Henson]
234
235 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
236 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
237 BIO_set_cipher() and some obscure PEM functions were changed so they
238 can now return an error. The RAND changes required a change to the
239 RAND_METHOD structure.
240 [Steve Henson]
241
242 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
243 a gcc attribute to warn if the result of a function is ignored. This
244 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
245 whose return value is often ignored.
246 [Steve Henson]
247
248 Changes between 1.0.0e and 1.0.1 [xx XXX xxxx]
249
250 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
251 ENGINE is used then we cannot handle that in the FIPS module so we
252 keep original code iff non-FIPS operations are allowed.
253 [Steve Henson]
254
255 *) Add -attime option to openssl verify.
256 [Peter Eckersley <pde@eff.org> and Ben Laurie]
257
258 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
259 [Steve Henson]
260
261 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
262 FIPS EC methods unconditionally for now.
263 [Steve Henson]
264
265 *) New build option no-ec2m to disable characteristic 2 code.
266 [Steve Henson]
267
268 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
269 all cases can be covered as some introduce binary incompatibilities.
270 [Steve Henson]
271
272 *) Redirect RSA operations to FIPS module including keygen,
273 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
274 [Steve Henson]
275
276 *) Add similar low level API blocking to ciphers.
277 [Steve Henson]
278
279 *) Low level digest APIs are not approved in FIPS mode: any attempt
280 to use these will cause a fatal error. Applications that *really* want
281 to use them can use the private_* version instead.
282 [Steve Henson]
283
284 *) Redirect cipher operations to FIPS module for FIPS builds.
285 [Steve Henson]
286
287 *) Redirect digest operations to FIPS module for FIPS builds.
288 [Steve Henson]
289
290 *) Update build system to add "fips" flag which will link in fipscanister.o
291 for static and shared library builds embedding a signature if needed.
292 [Steve Henson]
293
294 *) Output TLS supported curves in preference order instead of numerical
295 order. This is currently hardcoded for the highest order curves first.
296 This should be configurable so applications can judge speed vs strength.
297 [Steve Henson]
298
299 *) Add TLS v1.2 server support for client authentication.
300 [Steve Henson]
301
302 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
303 and enable MD5.
304 [Steve Henson]
305
306 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
307 FIPS modules versions.
308 [Steve Henson]
309
310 *) Add TLS v1.2 client side support for client authentication. Keep cache
311 of handshake records longer as we don't know the hash algorithm to use
312 until after the certificate request message is received.
313 [Steve Henson]
314
315 *) Initial TLS v1.2 client support. Add a default signature algorithms
316 extension including all the algorithms we support. Parse new signature
317 format in client key exchange. Relax some ECC signing restrictions for
318 TLS v1.2 as indicated in RFC5246.
319 [Steve Henson]
320
321 *) Add server support for TLS v1.2 signature algorithms extension. Switch
322 to new signature format when needed using client digest preference.
323 All server ciphersuites should now work correctly in TLS v1.2. No client
324 support yet and no support for client certificates.
325 [Steve Henson]
326
327 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
328 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
329 ciphersuites. At present only RSA key exchange ciphersuites work with
330 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
331 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
332 and version checking.
333 [Steve Henson]
334
335 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
336 with this defined it will not be affected by any changes to ssl internal
337 structures. Add several utility functions to allow openssl application
338 to work with OPENSSL_NO_SSL_INTERN defined.
339 [Steve Henson]
340
341 *) Add SRP support.
342 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
343
344 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
345 [Steve Henson]
346
347 *) Add EC_GFp_nistp224_method(), a 64-bit optimized implementation for
348 elliptic curve NIST-P224 with constant-time single point multiplication on
349 typical inputs. EC_GROUP_new_by_curve_name() will automatically use this
350 (while EC_GROUP_new_curve_GFp() currently won't and prefers the more
351 flexible implementations).
352
353 The implementation requires support for the nonstandard type __uint128_t,
354 and so is disabled by default. To include this in your build of OpenSSL,
355 use -DEC_NISTP224_64_GCC_128 on the Configure (or config) command line,
356 and run "make depend" (or "make update").
357 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
358
359 *) Permit abbreviated handshakes when renegotiating using the function
360 SSL_renegotiate_abbreviated().
361 [Robin Seggelmann <seggelmann@fh-muenster.de>]
362
363 *) Add call to ENGINE_register_all_complete() to
364 ENGINE_load_builtin_engines(), so some implementations get used
365 automatically instead of needing explicit application support.
366 [Steve Henson]
367
368 *) Add support for TLS key exporter as described in RFC5705.
369 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
370
371 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
372 a few changes are required:
373
374 Add SSL_OP_NO_TLSv1_1 flag.
375 Add TLSv1_1 methods.
376 Update version checking logic to handle version 1.1.
377 Add explicit IV handling (ported from DTLS code).
378 Add command line options to s_client/s_server.
379 [Steve Henson]
380
381 Changes between 1.0.0d and 1.0.0e [xx XXX xxxx]
382
383 *) Add protection against ECDSA timing attacks as mentioned in the paper
384 by Billy Bob Brumley and Nicola Tuveri, see:
385
386 http://eprint.iacr.org/2011/232.pdf
387
388 [Billy Bob Brumley and Nicola Tuveri]
389
390 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
391
392 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
393 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
394
395 *) Fix bug in string printing code: if *any* escaping is enabled we must
396 escape the escape character (backslash) or the resulting string is
397 ambiguous.
398 [Steve Henson]
399
400 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
401
402 *) Disable code workaround for ancient and obsolete Netscape browsers
403 and servers: an attacker can use it in a ciphersuite downgrade attack.
404 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
405 [Steve Henson]
406
407 *) Fixed J-PAKE implementation error, originally discovered by
408 Sebastien Martini, further info and confirmation from Stefan
409 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
410 [Ben Laurie]
411
412 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
413
414 *) Fix extension code to avoid race conditions which can result in a buffer
415 overrun vulnerability: resumed sessions must not be modified as they can
416 be shared by multiple threads. CVE-2010-3864
417 [Steve Henson]
418
419 *) Fix WIN32 build system to correctly link an ENGINE directory into
420 a DLL.
421 [Steve Henson]
422
423 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
424
425 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
426 (CVE-2010-1633)
427 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
428
429 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
430
431 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
432 context. The operation can be customised via the ctrl mechanism in
433 case ENGINEs want to include additional functionality.
434 [Steve Henson]
435
436 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
437 [Steve Henson]
438
439 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
440 output hashes compatible with older versions of OpenSSL.
441 [Willy Weisz <weisz@vcpc.univie.ac.at>]
442
443 *) Fix compression algorithm handling: if resuming a session use the
444 compression algorithm of the resumed session instead of determining
445 it from client hello again. Don't allow server to change algorithm.
446 [Steve Henson]
447
448 *) Add load_crls() function to apps tidying load_certs() too. Add option
449 to verify utility to allow additional CRLs to be included.
450 [Steve Henson]
451
452 *) Update OCSP request code to permit adding custom headers to the request:
453 some responders need this.
454 [Steve Henson]
455
456 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
457 correctly.
458 [Julia Lawall <julia@diku.dk>]
459
460 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
461 needlessly dereferenced structures, used obsolete functions and
462 didn't handle all updated verify codes correctly.
463 [Steve Henson]
464
465 *) Disable MD2 in the default configuration.
466 [Steve Henson]
467
468 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
469 indicate the initial BIO being pushed or popped. This makes it possible
470 to determine whether the BIO is the one explicitly called or as a result
471 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
472 it handles reference counts correctly and doesn't zero out the I/O bio
473 when it is not being explicitly popped. WARNING: applications which
474 included workarounds for the old buggy behaviour will need to be modified
475 or they could free up already freed BIOs.
476 [Steve Henson]
477
478 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
479 renaming to all platforms (within the 0.9.8 branch, this was
480 done conditionally on Netware platforms to avoid a name clash).
481 [Guenter <lists@gknw.net>]
482
483 *) Add ECDHE and PSK support to DTLS.
484 [Michael Tuexen <tuexen@fh-muenster.de>]
485
486 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
487 be used on C++.
488 [Steve Henson]
489
490 *) Add "missing" function EVP_MD_flags() (without this the only way to
491 retrieve a digest flags is by accessing the structure directly. Update
492 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
493 or cipher is registered as in the "from" argument. Print out all
494 registered digests in the dgst usage message instead of manually
495 attempting to work them out.
496 [Steve Henson]
497
498 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
499 this allows the use of compression and extensions. Change default cipher
500 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
501 by default unless an application cipher string requests it.
502 [Steve Henson]
503
504 *) Alter match criteria in PKCS12_parse(). It used to try to use local
505 key ids to find matching certificates and keys but some PKCS#12 files
506 don't follow the (somewhat unwritten) rules and this strategy fails.
507 Now just gather all certificates together and the first private key
508 then look for the first certificate that matches the key.
509 [Steve Henson]
510
511 *) Support use of registered digest and cipher names for dgst and cipher
512 commands instead of having to add each one as a special case. So now
513 you can do:
514
515 openssl sha256 foo
516
517 as well as:
518
519 openssl dgst -sha256 foo
520
521 and this works for ENGINE based algorithms too.
522
523 [Steve Henson]
524
525 *) Update Gost ENGINE to support parameter files.
526 [Victor B. Wagner <vitus@cryptocom.ru>]
527
528 *) Support GeneralizedTime in ca utility.
529 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
530
531 *) Enhance the hash format used for certificate directory links. The new
532 form uses the canonical encoding (meaning equivalent names will work
533 even if they aren't identical) and uses SHA1 instead of MD5. This form
534 is incompatible with the older format and as a result c_rehash should
535 be used to rebuild symbolic links.
536 [Steve Henson]
537
538 *) Make PKCS#8 the default write format for private keys, replacing the
539 traditional format. This form is standardised, more secure and doesn't
540 include an implicit MD5 dependency.
541 [Steve Henson]
542
543 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
544 committed to OpenSSL should pass this lot as a minimum.
545 [Steve Henson]
546
547 *) Add session ticket override functionality for use by EAP-FAST.
548 [Jouni Malinen <j@w1.fi>]
549
550 *) Modify HMAC functions to return a value. Since these can be implemented
551 in an ENGINE errors can occur.
552 [Steve Henson]
553
554 *) Type-checked OBJ_bsearch_ex.
555 [Ben Laurie]
556
557 *) Type-checked OBJ_bsearch. Also some constification necessitated
558 by type-checking. Still to come: TXT_DB, bsearch(?),
559 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
560 CONF_VALUE.
561 [Ben Laurie]
562
563 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
564 seconds to a tm structure directly, instead of going through OS
565 specific date routines. This avoids any issues with OS routines such
566 as the year 2038 bug. New *_adj() functions for ASN1 time structures
567 and X509_time_adj_ex() to cover the extended range. The existing
568 X509_time_adj() is still usable and will no longer have any date issues.
569 [Steve Henson]
570
571 *) Delta CRL support. New use deltas option which will attempt to locate
572 and search any appropriate delta CRLs available.
573
574 This work was sponsored by Google.
575 [Steve Henson]
576
577 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
578 code and add additional score elements. Validate alternate CRL paths
579 as part of the CRL checking and indicate a new error "CRL path validation
580 error" in this case. Applications wanting additional details can use
581 the verify callback and check the new "parent" field. If this is not
582 NULL CRL path validation is taking place. Existing applications wont
583 see this because it requires extended CRL support which is off by
584 default.
585
586 This work was sponsored by Google.
587 [Steve Henson]
588
589 *) Support for freshest CRL extension.
590
591 This work was sponsored by Google.
592 [Steve Henson]
593
594 *) Initial indirect CRL support. Currently only supported in the CRLs
595 passed directly and not via lookup. Process certificate issuer
596 CRL entry extension and lookup CRL entries by bother issuer name
597 and serial number. Check and process CRL issuer entry in IDP extension.
598
599 This work was sponsored by Google.
600 [Steve Henson]
601
602 *) Add support for distinct certificate and CRL paths. The CRL issuer
603 certificate is validated separately in this case. Only enabled if
604 an extended CRL support flag is set: this flag will enable additional
605 CRL functionality in future.
606
607 This work was sponsored by Google.
608 [Steve Henson]
609
610 *) Add support for policy mappings extension.
611
612 This work was sponsored by Google.
613 [Steve Henson]
614
615 *) Fixes to pathlength constraint, self issued certificate handling,
616 policy processing to align with RFC3280 and PKITS tests.
617
618 This work was sponsored by Google.
619 [Steve Henson]
620
621 *) Support for name constraints certificate extension. DN, email, DNS
622 and URI types are currently supported.
623
624 This work was sponsored by Google.
625 [Steve Henson]
626
627 *) To cater for systems that provide a pointer-based thread ID rather
628 than numeric, deprecate the current numeric thread ID mechanism and
629 replace it with a structure and associated callback type. This
630 mechanism allows a numeric "hash" to be extracted from a thread ID in
631 either case, and on platforms where pointers are larger than 'long',
632 mixing is done to help ensure the numeric 'hash' is usable even if it
633 can't be guaranteed unique. The default mechanism is to use "&errno"
634 as a pointer-based thread ID to distinguish between threads.
635
636 Applications that want to provide their own thread IDs should now use
637 CRYPTO_THREADID_set_callback() to register a callback that will call
638 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
639
640 Note that ERR_remove_state() is now deprecated, because it is tied
641 to the assumption that thread IDs are numeric. ERR_remove_state(0)
642 to free the current thread's error state should be replaced by
643 ERR_remove_thread_state(NULL).
644
645 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
646 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
647 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
648 application was previously providing a numeric thread callback that
649 was inappropriate for distinguishing threads, then uniqueness might
650 have been obtained with &errno that happened immediately in the
651 intermediate development versions of OpenSSL; this is no longer the
652 case, the numeric thread callback will now override the automatic use
653 of &errno.)
654 [Geoff Thorpe, with help from Bodo Moeller]
655
656 *) Initial support for different CRL issuing certificates. This covers a
657 simple case where the self issued certificates in the chain exist and
658 the real CRL issuer is higher in the existing chain.
659
660 This work was sponsored by Google.
661 [Steve Henson]
662
663 *) Removed effectively defunct crypto/store from the build.
664 [Ben Laurie]
665
666 *) Revamp of STACK to provide stronger type-checking. Still to come:
667 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
668 ASN1_STRING, CONF_VALUE.
669 [Ben Laurie]
670
671 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
672 RAM on SSL connections. This option can save about 34k per idle SSL.
673 [Nick Mathewson]
674
675 *) Revamp of LHASH to provide stronger type-checking. Still to come:
676 STACK, TXT_DB, bsearch, qsort.
677 [Ben Laurie]
678
679 *) Initial support for Cryptographic Message Syntax (aka CMS) based
680 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
681 support for data, signedData, compressedData, digestedData and
682 encryptedData, envelopedData types included. Scripts to check against
683 RFC4134 examples draft and interop and consistency checks of many
684 content types and variants.
685 [Steve Henson]
686
687 *) Add options to enc utility to support use of zlib compression BIO.
688 [Steve Henson]
689
690 *) Extend mk1mf to support importing of options and assembly language
691 files from Configure script, currently only included in VC-WIN32.
692 The assembly language rules can now optionally generate the source
693 files from the associated perl scripts.
694 [Steve Henson]
695
696 *) Implement remaining functionality needed to support GOST ciphersuites.
697 Interop testing has been performed using CryptoPro implementations.
698 [Victor B. Wagner <vitus@cryptocom.ru>]
699
700 *) s390x assembler pack.
701 [Andy Polyakov]
702
703 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
704 "family."
705 [Andy Polyakov]
706
707 *) Implement Opaque PRF Input TLS extension as specified in
708 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
709 official specification yet and no extension type assignment by
710 IANA exists, this extension (for now) will have to be explicitly
711 enabled when building OpenSSL by providing the extension number
712 to use. For example, specify an option
713
714 -DTLSEXT_TYPE_opaque_prf_input=0x9527
715
716 to the "config" or "Configure" script to enable the extension,
717 assuming extension number 0x9527 (which is a completely arbitrary
718 and unofficial assignment based on the MD5 hash of the Internet
719 Draft). Note that by doing so, you potentially lose
720 interoperability with other TLS implementations since these might
721 be using the same extension number for other purposes.
722
723 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
724 opaque PRF input value to use in the handshake. This will create
725 an interal copy of the length-'len' string at 'src', and will
726 return non-zero for success.
727
728 To get more control and flexibility, provide a callback function
729 by using
730
731 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
732 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
733
734 where
735
736 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
737 void *arg;
738
739 Callback function 'cb' will be called in handshakes, and is
740 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
741 Argument 'arg' is for application purposes (the value as given to
742 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
743 be provided to the callback function). The callback function
744 has to return non-zero to report success: usually 1 to use opaque
745 PRF input just if possible, or 2 to enforce use of the opaque PRF
746 input. In the latter case, the library will abort the handshake
747 if opaque PRF input is not successfully negotiated.
748
749 Arguments 'peerinput' and 'len' given to the callback function
750 will always be NULL and 0 in the case of a client. A server will
751 see the client's opaque PRF input through these variables if
752 available (NULL and 0 otherwise). Note that if the server
753 provides an opaque PRF input, the length must be the same as the
754 length of the client's opaque PRF input.
755
756 Note that the callback function will only be called when creating
757 a new session (session resumption can resume whatever was
758 previously negotiated), and will not be called in SSL 2.0
759 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
760 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
761 for applications that need to enforce opaque PRF input.
762
763 [Bodo Moeller]
764
765 *) Update ssl code to support digests other than SHA1+MD5 for handshake
766 MAC.
767
768 [Victor B. Wagner <vitus@cryptocom.ru>]
769
770 *) Add RFC4507 support to OpenSSL. This includes the corrections in
771 RFC4507bis. The encrypted ticket format is an encrypted encoded
772 SSL_SESSION structure, that way new session features are automatically
773 supported.
774
775 If a client application caches session in an SSL_SESSION structure
776 support is transparent because tickets are now stored in the encoded
777 SSL_SESSION.
778
779 The SSL_CTX structure automatically generates keys for ticket
780 protection in servers so again support should be possible
781 with no application modification.
782
783 If a client or server wishes to disable RFC4507 support then the option
784 SSL_OP_NO_TICKET can be set.
785
786 Add a TLS extension debugging callback to allow the contents of any client
787 or server extensions to be examined.
788
789 This work was sponsored by Google.
790 [Steve Henson]
791
792 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
793 OpenSSL should now compile cleanly on gcc 4.2
794 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
795
796 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
797 support including streaming MAC support: this is required for GOST
798 ciphersuite support.
799 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
800
801 *) Add option -stream to use PKCS#7 streaming in smime utility. New
802 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
803 to output in BER and PEM format.
804 [Steve Henson]
805
806 *) Experimental support for use of HMAC via EVP_PKEY interface. This
807 allows HMAC to be handled via the EVP_DigestSign*() interface. The
808 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
809 ENGINE support for HMAC keys which are unextractable. New -mac and
810 -macopt options to dgst utility.
811 [Steve Henson]
812
813 *) New option -sigopt to dgst utility. Update dgst to use
814 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
815 alternative signing paramaters such as X9.31 or PSS in the dgst
816 utility.
817 [Steve Henson]
818
819 *) Change ssl_cipher_apply_rule(), the internal function that does
820 the work each time a ciphersuite string requests enabling
821 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
822 removing ("!foo+bar") a class of ciphersuites: Now it maintains
823 the order of disabled ciphersuites such that those ciphersuites
824 that most recently went from enabled to disabled not only stay
825 in order with respect to each other, but also have higher priority
826 than other disabled ciphersuites the next time ciphersuites are
827 enabled again.
828
829 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
830 the same ciphersuites as with "HIGH" alone, but in a specific
831 order where the PSK ciphersuites come first (since they are the
832 most recently disabled ciphersuites when "HIGH" is parsed).
833
834 Also, change ssl_create_cipher_list() (using this new
835 funcionality) such that between otherwise identical
836 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
837 the default order.
838 [Bodo Moeller]
839
840 *) Change ssl_create_cipher_list() so that it automatically
841 arranges the ciphersuites in reasonable order before starting
842 to process the rule string. Thus, the definition for "DEFAULT"
843 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
844 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
845 This makes it much easier to arrive at a reasonable default order
846 in applications for which anonymous ciphers are OK (meaning
847 that you can't actually use DEFAULT).
848 [Bodo Moeller; suggested by Victor Duchovni]
849
850 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
851 processing) into multiple integers instead of setting
852 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
853 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
854 (These masks as well as the individual bit definitions are hidden
855 away into the non-exported interface ssl/ssl_locl.h, so this
856 change to the definition of the SSL_CIPHER structure shouldn't
857 affect applications.) This give us more bits for each of these
858 categories, so there is no longer a need to coagulate AES128 and
859 AES256 into a single algorithm bit, and to coagulate Camellia128
860 and Camellia256 into a single algorithm bit, which has led to all
861 kinds of kludges.
862
863 Thus, among other things, the kludge introduced in 0.9.7m and
864 0.9.8e for masking out AES256 independently of AES128 or masking
865 out Camellia256 independently of AES256 is not needed here in 0.9.9.
866
867 With the change, we also introduce new ciphersuite aliases that
868 so far were missing: "AES128", "AES256", "CAMELLIA128", and
869 "CAMELLIA256".
870 [Bodo Moeller]
871
872 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
873 Use the leftmost N bytes of the signature input if the input is
874 larger than the prime q (with N being the size in bytes of q).
875 [Nils Larsch]
876
877 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
878 it yet and it is largely untested.
879 [Steve Henson]
880
881 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
882 [Nils Larsch]
883
884 *) Initial incomplete changes to avoid need for function casts in OpenSSL
885 some compilers (gcc 4.2 and later) reject their use. Safestack is
886 reimplemented. Update ASN1 to avoid use of legacy functions.
887 [Steve Henson]
888
889 *) Win32/64 targets are linked with Winsock2.
890 [Andy Polyakov]
891
892 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
893 to external functions. This can be used to increase CRL handling
894 efficiency especially when CRLs are very large by (for example) storing
895 the CRL revoked certificates in a database.
896 [Steve Henson]
897
898 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
899 new CRLs added to a directory can be used. New command line option
900 -verify_return_error to s_client and s_server. This causes real errors
901 to be returned by the verify callback instead of carrying on no matter
902 what. This reflects the way a "real world" verify callback would behave.
903 [Steve Henson]
904
905 *) GOST engine, supporting several GOST algorithms and public key formats.
906 Kindly donated by Cryptocom.
907 [Cryptocom]
908
909 *) Partial support for Issuing Distribution Point CRL extension. CRLs
910 partitioned by DP are handled but no indirect CRL or reason partitioning
911 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
912 selected via a scoring technique which handles IDP and AKID in CRLs.
913 [Steve Henson]
914
915 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
916 will ultimately be used for all verify operations: this will remove the
917 X509_STORE dependency on certificate verification and allow alternative
918 lookup methods. X509_STORE based implementations of these two callbacks.
919 [Steve Henson]
920
921 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
922 Modify get_crl() to find a valid (unexpired) CRL if possible.
923 [Steve Henson]
924
925 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
926 this would be called X509_CRL_cmp() but that name is already used by
927 a function that just compares CRL issuer names. Cache several CRL
928 extensions in X509_CRL structure and cache CRLDP in X509.
929 [Steve Henson]
930
931 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
932 this maps equivalent X509_NAME structures into a consistent structure.
933 Name comparison can then be performed rapidly using memcmp().
934 [Steve Henson]
935
936 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
937 utility.
938 [Steve Henson]
939
940 *) Allow digests to supply their own micalg string for S/MIME type using
941 the ctrl EVP_MD_CTRL_MICALG.
942 [Steve Henson]
943
944 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
945 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
946 ctrl. It can then customise the structure before and/or after signing
947 if necessary.
948 [Steve Henson]
949
950 *) New function OBJ_add_sigid() to allow application defined signature OIDs
951 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
952 to free up any added signature OIDs.
953 [Steve Henson]
954
955 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
956 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
957 digest and cipher tables. New options added to openssl utility:
958 list-message-digest-algorithms and list-cipher-algorithms.
959 [Steve Henson]
960
961 *) Change the array representation of binary polynomials: the list
962 of degrees of non-zero coefficients is now terminated with -1.
963 Previously it was terminated with 0, which was also part of the
964 value; thus, the array representation was not applicable to
965 polynomials where t^0 has coefficient zero. This change makes
966 the array representation useful in a more general context.
967 [Douglas Stebila]
968
969 *) Various modifications and fixes to SSL/TLS cipher string
970 handling. For ECC, the code now distinguishes between fixed ECDH
971 with RSA certificates on the one hand and with ECDSA certificates
972 on the other hand, since these are separate ciphersuites. The
973 unused code for Fortezza ciphersuites has been removed.
974
975 For consistency with EDH, ephemeral ECDH is now called "EECDH"
976 (not "ECDHE"). For consistency with the code for DH
977 certificates, use of ECDH certificates is now considered ECDH
978 authentication, not RSA or ECDSA authentication (the latter is
979 merely the CA's signing algorithm and not actively used in the
980 protocol).
981
982 The temporary ciphersuite alias "ECCdraft" is no longer
983 available, and ECC ciphersuites are no longer excluded from "ALL"
984 and "DEFAULT". The following aliases now exist for RFC 4492
985 ciphersuites, most of these by analogy with the DH case:
986
987 kECDHr - ECDH cert, signed with RSA
988 kECDHe - ECDH cert, signed with ECDSA
989 kECDH - ECDH cert (signed with either RSA or ECDSA)
990 kEECDH - ephemeral ECDH
991 ECDH - ECDH cert or ephemeral ECDH
992
993 aECDH - ECDH cert
994 aECDSA - ECDSA cert
995 ECDSA - ECDSA cert
996
997 AECDH - anonymous ECDH
998 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
999
1000 [Bodo Moeller]
1001
1002 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1003 Use correct micalg parameters depending on digest(s) in signed message.
1004 [Steve Henson]
1005
1006 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1007 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1008 [Steve Henson]
1009
1010 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1011 an engine to register a method. Add ENGINE lookups for methods and
1012 functional reference processing.
1013 [Steve Henson]
1014
1015 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1016 EVP_{Sign,Verify}* which allow an application to customise the signature
1017 process.
1018 [Steve Henson]
1019
1020 *) New -resign option to smime utility. This adds one or more signers
1021 to an existing PKCS#7 signedData structure. Also -md option to use an
1022 alternative message digest algorithm for signing.
1023 [Steve Henson]
1024
1025 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1026 create PKCS7 structures containing multiple signers. Update smime
1027 application to support multiple signers.
1028 [Steve Henson]
1029
1030 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1031 digest MAC.
1032 [Steve Henson]
1033
1034 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1035 Reorganize PBE internals to lookup from a static table using NIDs,
1036 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1037 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1038 PRF which will be automatically used with PBES2.
1039 [Steve Henson]
1040
1041 *) Replace the algorithm specific calls to generate keys in "req" with the
1042 new API.
1043 [Steve Henson]
1044
1045 *) Update PKCS#7 enveloped data routines to use new API. This is now
1046 supported by any public key method supporting the encrypt operation. A
1047 ctrl is added to allow the public key algorithm to examine or modify
1048 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1049 a no op.
1050 [Steve Henson]
1051
1052 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1053 a default digest type to use. In most cases this will be SHA1 but some
1054 algorithms (such as GOST) need to specify an alternative digest. The
1055 return value indicates how strong the prefernce is 1 means optional and
1056 2 is mandatory (that is it is the only supported type). Modify
1057 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1058 use the default md. Update openssl utilities to use the default digest
1059 type for signing if it is not explicitly indicated.
1060 [Steve Henson]
1061
1062 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1063 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1064 signing method from the key type. This effectively removes the link
1065 between digests and public key types.
1066 [Steve Henson]
1067
1068 *) Add an OID cross reference table and utility functions. Its purpose is to
1069 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1070 rsaEncryption. This will allow some of the algorithm specific hackery
1071 needed to use the correct OID to be removed.
1072 [Steve Henson]
1073
1074 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1075 structures for PKCS7_sign(). They are now set up by the relevant public
1076 key ASN1 method.
1077 [Steve Henson]
1078
1079 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1080 [Steve Henson]
1081
1082 *) Add support for key derivation (agreement) in the API, DH method and
1083 pkeyutl.
1084 [Steve Henson]
1085
1086 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1087 public and private key formats. As a side effect these add additional
1088 command line functionality not previously available: DSA signatures can be
1089 generated and verified using pkeyutl and DH key support and generation in
1090 pkey, genpkey.
1091 [Steve Henson]
1092
1093 *) BeOS support.
1094 [Oliver Tappe <zooey@hirschkaefer.de>]
1095
1096 *) New make target "install_html_docs" installs HTML renditions of the
1097 manual pages.
1098 [Oliver Tappe <zooey@hirschkaefer.de>]
1099
1100 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1101 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1102 support key and parameter generation and add initial key generation
1103 functionality for RSA.
1104 [Steve Henson]
1105
1106 *) Add functions for main EVP_PKEY_method operations. The undocumented
1107 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1108 EVP_PKEY_{encrypt,decrypt}_old.
1109 [Steve Henson]
1110
1111 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1112 key API, doesn't do much yet.
1113 [Steve Henson]
1114
1115 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1116 public key algorithms. New option to openssl utility:
1117 "list-public-key-algorithms" to print out info.
1118 [Steve Henson]
1119
1120 *) Implement the Supported Elliptic Curves Extension for
1121 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1122 [Douglas Stebila]
1123
1124 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1125 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1126 [Steve Henson]
1127
1128 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1129 utilities such as rsa, dsa, dsaparam etc except they process any key
1130 type.
1131 [Steve Henson]
1132
1133 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1134 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1135 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1136 structure.
1137 [Steve Henson]
1138
1139 *) Initial support for pluggable public key ASN1.
1140 De-spaghettify the public key ASN1 handling. Move public and private
1141 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1142 algorithm specific handling to a single module within the relevant
1143 algorithm directory. Add functions to allow (near) opaque processing
1144 of public and private key structures.
1145 [Steve Henson]
1146
1147 *) Implement the Supported Point Formats Extension for
1148 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1149 [Douglas Stebila]
1150
1151 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1152 for the psk identity [hint] and the psk callback functions to the
1153 SSL_SESSION, SSL and SSL_CTX structure.
1154
1155 New ciphersuites:
1156 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1157 PSK-AES256-CBC-SHA
1158
1159 New functions:
1160 SSL_CTX_use_psk_identity_hint
1161 SSL_get_psk_identity_hint
1162 SSL_get_psk_identity
1163 SSL_use_psk_identity_hint
1164
1165 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1166
1167 *) Add RFC 3161 compliant time stamp request creation, response generation
1168 and response verification functionality.
1169 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1170
1171 *) Add initial support for TLS extensions, specifically for the server_name
1172 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1173 have new members for a host name. The SSL data structure has an
1174 additional member SSL_CTX *initial_ctx so that new sessions can be
1175 stored in that context to allow for session resumption, even after the
1176 SSL has been switched to a new SSL_CTX in reaction to a client's
1177 server_name extension.
1178
1179 New functions (subject to change):
1180
1181 SSL_get_servername()
1182 SSL_get_servername_type()
1183 SSL_set_SSL_CTX()
1184
1185 New CTRL codes and macros (subject to change):
1186
1187 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1188 - SSL_CTX_set_tlsext_servername_callback()
1189 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1190 - SSL_CTX_set_tlsext_servername_arg()
1191 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1192
1193 openssl s_client has a new '-servername ...' option.
1194
1195 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1196 '-key2 ...', '-servername_fatal' (subject to change). This allows
1197 testing the HostName extension for a specific single host name ('-cert'
1198 and '-key' remain fallbacks for handshakes without HostName
1199 negotiation). If the unrecogninzed_name alert has to be sent, this by
1200 default is a warning; it becomes fatal with the '-servername_fatal'
1201 option.
1202
1203 [Peter Sylvester, Remy Allais, Christophe Renou]
1204
1205 *) Whirlpool hash implementation is added.
1206 [Andy Polyakov]
1207
1208 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1209 bn(64,32). Because of instruction set limitations it doesn't have
1210 any negative impact on performance. This was done mostly in order
1211 to make it possible to share assembler modules, such as bn_mul_mont
1212 implementations, between 32- and 64-bit builds without hassle.
1213 [Andy Polyakov]
1214
1215 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1216 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1217 macro.
1218 [Bodo Moeller]
1219
1220 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1221 dedicated Montgomery multiplication procedure, is introduced.
1222 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1223 "64-bit" performance on certain 32-bit targets.
1224 [Andy Polyakov]
1225
1226 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1227 in SSL structures. New SSL ctrl to set maximum send fragment size.
1228 Save memory by seeting the I/O buffer sizes dynamically instead of
1229 using the maximum available value.
1230 [Steve Henson]
1231
1232 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1233 in addition to the text details.
1234 [Bodo Moeller]
1235
1236 *) Very, very preliminary EXPERIMENTAL support for printing of general
1237 ASN1 structures. This currently produces rather ugly output and doesn't
1238 handle several customised structures at all.
1239 [Steve Henson]
1240
1241 *) Integrated support for PVK file format and some related formats such
1242 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1243 these in the 'rsa' and 'dsa' utilities.
1244 [Steve Henson]
1245
1246 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1247 [Steve Henson]
1248
1249 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1250 place for the (very old) "NETSCAPE" format certificates which are now
1251 handled using new ASN1 code equivalents.
1252 [Steve Henson]
1253
1254 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1255 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1256 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1257 [Nils Larsch]
1258
1259 *) Modify CRL distribution points extension code to print out previously
1260 unsupported fields. Enhance extension setting code to allow setting of
1261 all fields.
1262 [Steve Henson]
1263
1264 *) Add print and set support for Issuing Distribution Point CRL extension.
1265 [Steve Henson]
1266
1267 *) Change 'Configure' script to enable Camellia by default.
1268 [NTT]
1269
1270 Changes between 0.9.8r and 0.9.8s [xx XXX xxxx]
1271
1272 *) Add protection against ECDSA timing attacks as mentioned in the paper
1273 by Billy Bob Brumley and Nicola Tuveri, see:
1274
1275 http://eprint.iacr.org/2011/232.pdf
1276
1277 [Billy Bob Brumley and Nicola Tuveri]
1278
1279 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1280
1281 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1282 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1283
1284 *) Fix bug in string printing code: if *any* escaping is enabled we must
1285 escape the escape character (backslash) or the resulting string is
1286 ambiguous.
1287 [Steve Henson]
1288
1289 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1290
1291 *) Disable code workaround for ancient and obsolete Netscape browsers
1292 and servers: an attacker can use it in a ciphersuite downgrade attack.
1293 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1294 [Steve Henson]
1295
1296 *) Fixed J-PAKE implementation error, originally discovered by
1297 Sebastien Martini, further info and confirmation from Stefan
1298 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1299 [Ben Laurie]
1300
1301 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1302
1303 *) Fix extension code to avoid race conditions which can result in a buffer
1304 overrun vulnerability: resumed sessions must not be modified as they can
1305 be shared by multiple threads. CVE-2010-3864
1306 [Steve Henson]
1307
1308 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1309 [Steve Henson]
1310
1311 *) Don't reencode certificate when calculating signature: cache and use
1312 the original encoding instead. This makes signature verification of
1313 some broken encodings work correctly.
1314 [Steve Henson]
1315
1316 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1317 is also one of the inputs.
1318 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1319
1320 *) Don't repeatedly append PBE algorithms to table if they already exist.
1321 Sort table on each new add. This effectively makes the table read only
1322 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1323 etc are non-op.
1324 [Steve Henson]
1325
1326 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1327
1328 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1329 OpenSSL 1.0.0.]
1330
1331 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1332 access or freeing data twice (CVE-2010-0742)
1333 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1334
1335 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1336 common in certificates and some applications which only call
1337 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1338 [Steve Henson]
1339
1340 *) VMS fixes:
1341 Reduce copying into .apps and .test in makevms.com
1342 Don't try to use blank CA certificate in CA.com
1343 Allow use of C files from original directories in maketests.com
1344 [Steven M. Schweda" <sms@antinode.info>]
1345
1346 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1347
1348 *) When rejecting SSL/TLS records due to an incorrect version number, never
1349 update s->server with a new major version number. As of
1350 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1351 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1352 the previous behavior could result in a read attempt at NULL when
1353 receiving specific incorrect SSL/TLS records once record payload
1354 protection is active. (CVE-2010-0740)
1355 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1356
1357 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1358 could be crashed if the relevant tables were not present (e.g. chrooted).
1359 [Tomas Hoger <thoger@redhat.com>]
1360
1361 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1362
1363 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1364 [Martin Olsson, Neel Mehta]
1365
1366 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1367 accommodate for stack sorting, always a write lock!).
1368 [Bodo Moeller]
1369
1370 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1371 excessive delays in the RAND_poll(): over a minute. As a workaround
1372 include a time check in the inner Heap32Next loop too.
1373 [Steve Henson]
1374
1375 *) The code that handled flushing of data in SSL/TLS originally used the
1376 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1377 the problem outlined in PR#1949. The fix suggested there however can
1378 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1379 of Apache). So instead simplify the code to flush unconditionally.
1380 This should be fine since flushing with no data to flush is a no op.
1381 [Steve Henson]
1382
1383 *) Handle TLS versions 2.0 and later properly and correctly use the
1384 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1385 off ancient servers have a habit of sticking around for a while...
1386 [Steve Henson]
1387
1388 *) Modify compression code so it frees up structures without using the
1389 ex_data callbacks. This works around a problem where some applications
1390 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1391 restarting) then use compression (e.g. SSL with compression) later.
1392 This results in significant per-connection memory leaks and
1393 has caused some security issues including CVE-2008-1678 and
1394 CVE-2009-4355.
1395 [Steve Henson]
1396
1397 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1398 change when encrypting or decrypting.
1399 [Bodo Moeller]
1400
1401 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1402 connect and renegotiate with servers which do not support RI.
1403 Until RI is more widely deployed this option is enabled by default.
1404 [Steve Henson]
1405
1406 *) Add "missing" ssl ctrls to clear options and mode.
1407 [Steve Henson]
1408
1409 *) If client attempts to renegotiate and doesn't support RI respond with
1410 a no_renegotiation alert as required by RFC5746. Some renegotiating
1411 TLS clients will continue a connection gracefully when they receive
1412 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1413 waiting for a server hello which it will never receive. Now we treat a
1414 received no_renegotiation alert as a fatal error. This is because
1415 applications requesting a renegotiation might well expect it to succeed
1416 and would have no code in place to handle the server denying it so the
1417 only safe thing to do is to terminate the connection.
1418 [Steve Henson]
1419
1420 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1421 peer supports secure renegotiation and 0 otherwise. Print out peer
1422 renegotiation support in s_client/s_server.
1423 [Steve Henson]
1424
1425 *) Replace the highly broken and deprecated SPKAC certification method with
1426 the updated NID creation version. This should correctly handle UTF8.
1427 [Steve Henson]
1428
1429 *) Implement RFC5746. Re-enable renegotiation but require the extension
1430 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1431 turns out to be a bad idea. It has been replaced by
1432 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1433 SSL_CTX_set_options(). This is really not recommended unless you
1434 know what you are doing.
1435 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1436
1437 *) Fixes to stateless session resumption handling. Use initial_ctx when
1438 issuing and attempting to decrypt tickets in case it has changed during
1439 servername handling. Use a non-zero length session ID when attempting
1440 stateless session resumption: this makes it possible to determine if
1441 a resumption has occurred immediately after receiving server hello
1442 (several places in OpenSSL subtly assume this) instead of later in
1443 the handshake.
1444 [Steve Henson]
1445
1446 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1447 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1448 fixes for a few places where the return code is not checked
1449 correctly.
1450 [Julia Lawall <julia@diku.dk>]
1451
1452 *) Add --strict-warnings option to Configure script to include devteam
1453 warnings in other configurations.
1454 [Steve Henson]
1455
1456 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1457 makes it possible to install openssl libraries in locations which
1458 have names other than "lib", for example "/usr/lib64" which some
1459 systems need.
1460 [Steve Henson, based on patch from Jeremy Utley]
1461
1462 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1463 X690 8.9.12 and can produce some misleading textual output of OIDs.
1464 [Steve Henson, reported by Dan Kaminsky]
1465
1466 *) Delete MD2 from algorithm tables. This follows the recommendation in
1467 several standards that it is not used in new applications due to
1468 several cryptographic weaknesses. For binary compatibility reasons
1469 the MD2 API is still compiled in by default.
1470 [Steve Henson]
1471
1472 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1473 and restored.
1474 [Steve Henson]
1475
1476 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1477 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1478 clash.
1479 [Guenter <lists@gknw.net>]
1480
1481 *) Fix the server certificate chain building code to use X509_verify_cert(),
1482 it used to have an ad-hoc builder which was unable to cope with anything
1483 other than a simple chain.
1484 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1485
1486 *) Don't check self signed certificate signatures in X509_verify_cert()
1487 by default (a flag can override this): it just wastes time without
1488 adding any security. As a useful side effect self signed root CAs
1489 with non-FIPS digests are now usable in FIPS mode.
1490 [Steve Henson]
1491
1492 *) In dtls1_process_out_of_seq_message() the check if the current message
1493 is already buffered was missing. For every new message was memory
1494 allocated, allowing an attacker to perform an denial of service attack
1495 with sending out of seq handshake messages until there is no memory
1496 left. Additionally every future messege was buffered, even if the
1497 sequence number made no sense and would be part of another handshake.
1498 So only messages with sequence numbers less than 10 in advance will be
1499 buffered. (CVE-2009-1378)
1500 [Robin Seggelmann, discovered by Daniel Mentz]
1501
1502 *) Records are buffered if they arrive with a future epoch to be
1503 processed after finishing the corresponding handshake. There is
1504 currently no limitation to this buffer allowing an attacker to perform
1505 a DOS attack with sending records with future epochs until there is no
1506 memory left. This patch adds the pqueue_size() function to detemine
1507 the size of a buffer and limits the record buffer to 100 entries.
1508 (CVE-2009-1377)
1509 [Robin Seggelmann, discovered by Daniel Mentz]
1510
1511 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1512 parent structure is freed. (CVE-2009-1379)
1513 [Daniel Mentz]
1514
1515 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1516 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1517
1518 *) Add 2.5.4.* OIDs
1519 [Ilya O. <vrghost@gmail.com>]
1520
1521 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1522
1523 *) Disable renegotiation completely - this fixes a severe security
1524 problem (CVE-2009-3555) at the cost of breaking all
1525 renegotiation. Renegotiation can be re-enabled by setting
1526 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1527 run-time. This is really not recommended unless you know what
1528 you're doing.
1529 [Ben Laurie]
1530
1531 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1532
1533 *) Don't set val to NULL when freeing up structures, it is freed up by
1534 underlying code. If sizeof(void *) > sizeof(long) this can result in
1535 zeroing past the valid field. (CVE-2009-0789)
1536 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1537
1538 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1539 checked correctly. This would allow some invalid signed attributes to
1540 appear to verify correctly. (CVE-2009-0591)
1541 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1542
1543 *) Reject UniversalString and BMPString types with invalid lengths. This
1544 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1545 a legal length. (CVE-2009-0590)
1546 [Steve Henson]
1547
1548 *) Set S/MIME signing as the default purpose rather than setting it
1549 unconditionally. This allows applications to override it at the store
1550 level.
1551 [Steve Henson]
1552
1553 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1554 to handle some structures.
1555 [Steve Henson]
1556
1557 *) Improve efficiency of mem_gets: don't search whole buffer each time
1558 for a '\n'
1559 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1560
1561 *) New -hex option for openssl rand.
1562 [Matthieu Herrb]
1563
1564 *) Print out UTF8String and NumericString when parsing ASN1.
1565 [Steve Henson]
1566
1567 *) Support NumericString type for name components.
1568 [Steve Henson]
1569
1570 *) Allow CC in the environment to override the automatically chosen
1571 compiler. Note that nothing is done to ensure flags work with the
1572 chosen compiler.
1573 [Ben Laurie]
1574
1575 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1576
1577 *) Properly check EVP_VerifyFinal() and similar return values
1578 (CVE-2008-5077).
1579 [Ben Laurie, Bodo Moeller, Google Security Team]
1580
1581 *) Enable TLS extensions by default.
1582 [Ben Laurie]
1583
1584 *) Allow the CHIL engine to be loaded, whether the application is
1585 multithreaded or not. (This does not release the developer from the
1586 obligation to set up the dynamic locking callbacks.)
1587 [Sander Temme <sander@temme.net>]
1588
1589 *) Use correct exit code if there is an error in dgst command.
1590 [Steve Henson; problem pointed out by Roland Dirlewanger]
1591
1592 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1593 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1594 [Bodo Moeller]
1595
1596 *) Add experimental JPAKE support, including demo authentication in
1597 s_client and s_server.
1598 [Ben Laurie]
1599
1600 *) Set the comparison function in v3_addr_canonize().
1601 [Rob Austein <sra@hactrn.net>]
1602
1603 *) Add support for XMPP STARTTLS in s_client.
1604 [Philip Paeps <philip@freebsd.org>]
1605
1606 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1607 to ensure that even with this option, only ciphersuites in the
1608 server's preference list will be accepted. (Note that the option
1609 applies only when resuming a session, so the earlier behavior was
1610 just about the algorithm choice for symmetric cryptography.)
1611 [Bodo Moeller]
1612
1613 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1614
1615 *) Fix NULL pointer dereference if a DTLS server received
1616 ChangeCipherSpec as first record (CVE-2009-1386).
1617 [PR #1679]
1618
1619 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1620 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1621 [Nagendra Modadugu]
1622
1623 *) The fix in 0.9.8c that supposedly got rid of unsafe
1624 double-checked locking was incomplete for RSA blinding,
1625 addressing just one layer of what turns out to have been
1626 doubly unsafe triple-checked locking.
1627
1628 So now fix this for real by retiring the MONT_HELPER macro
1629 in crypto/rsa/rsa_eay.c.
1630
1631 [Bodo Moeller; problem pointed out by Marius Schilder]
1632
1633 *) Various precautionary measures:
1634
1635 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1636
1637 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1638 (NB: This would require knowledge of the secret session ticket key
1639 to exploit, in which case you'd be SOL either way.)
1640
1641 - Change bn_nist.c so that it will properly handle input BIGNUMs
1642 outside the expected range.
1643
1644 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1645 builds.
1646
1647 [Neel Mehta, Bodo Moeller]
1648
1649 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1650 the load fails. Useful for distros.
1651 [Ben Laurie and the FreeBSD team]
1652
1653 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1654 [Steve Henson]
1655
1656 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1657 [Huang Ying]
1658
1659 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1660
1661 This work was sponsored by Logica.
1662 [Steve Henson]
1663
1664 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1665 keystores. Support for SSL/TLS client authentication too.
1666 Not compiled unless enable-capieng specified to Configure.
1667
1668 This work was sponsored by Logica.
1669 [Steve Henson]
1670
1671 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1672 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1673 attribute creation routines such as certifcate requests and PKCS#12
1674 files.
1675 [Steve Henson]
1676
1677 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1678
1679 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1680 handshake which could lead to a cilent crash as found using the
1681 Codenomicon TLS test suite (CVE-2008-1672)
1682 [Steve Henson, Mark Cox]
1683
1684 *) Fix double free in TLS server name extensions which could lead to
1685 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1686 [Joe Orton]
1687
1688 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1689
1690 Clear the error queue to ensure that error entries left from
1691 older function calls do not interfere with the correct operation.
1692 [Lutz Jaenicke, Erik de Castro Lopo]
1693
1694 *) Remove root CA certificates of commercial CAs:
1695
1696 The OpenSSL project does not recommend any specific CA and does not
1697 have any policy with respect to including or excluding any CA.
1698 Therefore it does not make any sense to ship an arbitrary selection
1699 of root CA certificates with the OpenSSL software.
1700 [Lutz Jaenicke]
1701
1702 *) RSA OAEP patches to fix two separate invalid memory reads.
1703 The first one involves inputs when 'lzero' is greater than
1704 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1705 before the beginning of from). The second one involves inputs where
1706 the 'db' section contains nothing but zeroes (there is a one-byte
1707 invalid read after the end of 'db').
1708 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1709
1710 *) Partial backport from 0.9.9-dev:
1711
1712 Introduce bn_mul_mont (dedicated Montgomery multiplication
1713 procedure) as a candidate for BIGNUM assembler implementation.
1714 While 0.9.9-dev uses assembler for various architectures, only
1715 x86_64 is available by default here in the 0.9.8 branch, and
1716 32-bit x86 is available through a compile-time setting.
1717
1718 To try the 32-bit x86 assembler implementation, use Configure
1719 option "enable-montasm" (which exists only for this backport).
1720
1721 As "enable-montasm" for 32-bit x86 disclaims code stability
1722 anyway, in this constellation we activate additional code
1723 backported from 0.9.9-dev for further performance improvements,
1724 namely BN_from_montgomery_word. (To enable this otherwise,
1725 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1726
1727 [Andy Polyakov (backport partially by Bodo Moeller)]
1728
1729 *) Add TLS session ticket callback. This allows an application to set
1730 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1731 values. This is useful for key rollover for example where several key
1732 sets may exist with different names.
1733 [Steve Henson]
1734
1735 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1736 This was broken until now in 0.9.8 releases, such that the only way
1737 a registered ENGINE could be used (assuming it initialises
1738 successfully on the host) was to explicitly set it as the default
1739 for the relevant algorithms. This is in contradiction with 0.9.7
1740 behaviour and the documentation. With this fix, when an ENGINE is
1741 registered into a given algorithm's table of implementations, the
1742 'uptodate' flag is reset so that auto-discovery will be used next
1743 time a new context for that algorithm attempts to select an
1744 implementation.
1745 [Ian Lister (tweaked by Geoff Thorpe)]
1746
1747 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1748 implemention in the following ways:
1749
1750 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1751 hard coded.
1752
1753 Lack of BER streaming support means one pass streaming processing is
1754 only supported if data is detached: setting the streaming flag is
1755 ignored for embedded content.
1756
1757 CMS support is disabled by default and must be explicitly enabled
1758 with the enable-cms configuration option.
1759 [Steve Henson]
1760
1761 *) Update the GMP engine glue to do direct copies between BIGNUM and
1762 mpz_t when openssl and GMP use the same limb size. Otherwise the
1763 existing "conversion via a text string export" trick is still used.
1764 [Paul Sheer <paulsheer@gmail.com>]
1765
1766 *) Zlib compression BIO. This is a filter BIO which compressed and
1767 uncompresses any data passed through it.
1768 [Steve Henson]
1769
1770 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1771 RFC3394 compatible AES key wrapping.
1772 [Steve Henson]
1773
1774 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1775 sets string data without copying. X509_ALGOR_set0() and
1776 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1777 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1778 from an X509_ATTRIBUTE structure optionally checking it occurs only
1779 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1780 data.
1781 [Steve Henson]
1782
1783 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1784 to get the expected BN_FLG_CONSTTIME behavior.
1785 [Bodo Moeller (Google)]
1786
1787 *) Netware support:
1788
1789 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1790 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1791 - added some more tests to do_tests.pl
1792 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1793 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1794 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1795 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1796 - various changes to netware.pl to enable gcc-cross builds on Win32
1797 platform
1798 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1799 - various changes to fix missing prototype warnings
1800 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1801 - added AES, WHIRLPOOL and CPUID assembler code to build files
1802 - added missing AES assembler make rules to mk1mf.pl
1803 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1804 [Guenter Knauf <eflash@gmx.net>]
1805
1806 *) Implement certificate status request TLS extension defined in RFC3546.
1807 A client can set the appropriate parameters and receive the encoded
1808 OCSP response via a callback. A server can query the supplied parameters
1809 and set the encoded OCSP response in the callback. Add simplified examples
1810 to s_client and s_server.
1811 [Steve Henson]
1812
1813 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1814
1815 *) Fix various bugs:
1816 + Binary incompatibility of ssl_ctx_st structure
1817 + DTLS interoperation with non-compliant servers
1818 + Don't call get_session_cb() without proposed session
1819 + Fix ia64 assembler code
1820 [Andy Polyakov, Steve Henson]
1821
1822 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1823
1824 *) DTLS Handshake overhaul. There were longstanding issues with
1825 OpenSSL DTLS implementation, which were making it impossible for
1826 RFC 4347 compliant client to communicate with OpenSSL server.
1827 Unfortunately just fixing these incompatibilities would "cut off"
1828 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1829 server keeps tolerating non RFC compliant syntax. The opposite is
1830 not true, 0.9.8f client can not communicate with earlier server.
1831 This update even addresses CVE-2007-4995.
1832 [Andy Polyakov]
1833
1834 *) Changes to avoid need for function casts in OpenSSL: some compilers
1835 (gcc 4.2 and later) reject their use.
1836 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1837 Steve Henson]
1838
1839 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1840 RFC4507bis. The encrypted ticket format is an encrypted encoded
1841 SSL_SESSION structure, that way new session features are automatically
1842 supported.
1843
1844 If a client application caches session in an SSL_SESSION structure
1845 support is transparent because tickets are now stored in the encoded
1846 SSL_SESSION.
1847
1848 The SSL_CTX structure automatically generates keys for ticket
1849 protection in servers so again support should be possible
1850 with no application modification.
1851
1852 If a client or server wishes to disable RFC4507 support then the option
1853 SSL_OP_NO_TICKET can be set.
1854
1855 Add a TLS extension debugging callback to allow the contents of any client
1856 or server extensions to be examined.
1857
1858 This work was sponsored by Google.
1859 [Steve Henson]
1860
1861 *) Add initial support for TLS extensions, specifically for the server_name
1862 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1863 have new members for a host name. The SSL data structure has an
1864 additional member SSL_CTX *initial_ctx so that new sessions can be
1865 stored in that context to allow for session resumption, even after the
1866 SSL has been switched to a new SSL_CTX in reaction to a client's
1867 server_name extension.
1868
1869 New functions (subject to change):
1870
1871 SSL_get_servername()
1872 SSL_get_servername_type()
1873 SSL_set_SSL_CTX()
1874
1875 New CTRL codes and macros (subject to change):
1876
1877 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1878 - SSL_CTX_set_tlsext_servername_callback()
1879 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1880 - SSL_CTX_set_tlsext_servername_arg()
1881 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1882
1883 openssl s_client has a new '-servername ...' option.
1884
1885 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1886 '-key2 ...', '-servername_fatal' (subject to change). This allows
1887 testing the HostName extension for a specific single host name ('-cert'
1888 and '-key' remain fallbacks for handshakes without HostName
1889 negotiation). If the unrecogninzed_name alert has to be sent, this by
1890 default is a warning; it becomes fatal with the '-servername_fatal'
1891 option.
1892
1893 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1894
1895 *) Add AES and SSE2 assembly language support to VC++ build.
1896 [Steve Henson]
1897
1898 *) Mitigate attack on final subtraction in Montgomery reduction.
1899 [Andy Polyakov]
1900
1901 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1902 (which previously caused an internal error).
1903 [Bodo Moeller]
1904
1905 *) Squeeze another 10% out of IGE mode when in != out.
1906 [Ben Laurie]
1907
1908 *) AES IGE mode speedup.
1909 [Dean Gaudet (Google)]
1910
1911 *) Add the Korean symmetric 128-bit cipher SEED (see
1912 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1913 add SEED ciphersuites from RFC 4162:
1914
1915 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1916 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1917 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1918 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1919
1920 To minimize changes between patchlevels in the OpenSSL 0.9.8
1921 series, SEED remains excluded from compilation unless OpenSSL
1922 is configured with 'enable-seed'.
1923 [KISA, Bodo Moeller]
1924
1925 *) Mitigate branch prediction attacks, which can be practical if a
1926 single processor is shared, allowing a spy process to extract
1927 information. For detailed background information, see
1928 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1929 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1930 and Necessary Software Countermeasures"). The core of the change
1931 are new versions BN_div_no_branch() and
1932 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1933 respectively, which are slower, but avoid the security-relevant
1934 conditional branches. These are automatically called by BN_div()
1935 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1936 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1937 remove a conditional branch.
1938
1939 BN_FLG_CONSTTIME is the new name for the previous
1940 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1941 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1942 in the exponent causes BN_mod_exp_mont() to use the alternative
1943 implementation in BN_mod_exp_mont_consttime().) The old name
1944 remains as a deprecated alias.
1945
1946 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1947 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1948 constant-time implementations for more than just exponentiation.
1949 Here too the old name is kept as a deprecated alias.
1950
1951 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1952 the BN_BLINDING structure gets an independent copy of the
1953 modulus. This means that the previous "BIGNUM *m" argument to
1954 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1955 essentially becomes "const BIGNUM *m", although we can't actually
1956 change this in the header file before 0.9.9. It allows
1957 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1958 enable BN_FLG_CONSTTIME.
1959
1960 [Matthew D Wood (Intel Corp)]
1961
1962 *) In the SSL/TLS server implementation, be strict about session ID
1963 context matching (which matters if an application uses a single
1964 external cache for different purposes). Previously,
1965 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1966 set. This did ensure strict client verification, but meant that,
1967 with applications using a single external cache for quite
1968 different requirements, clients could circumvent ciphersuite
1969 restrictions for a given session ID context by starting a session
1970 in a different context.
1971 [Bodo Moeller]
1972
1973 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1974 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1975 authentication-only ciphersuites.
1976 [Bodo Moeller]
1977
1978 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1979 not complete and could lead to a possible single byte overflow
1980 (CVE-2007-5135) [Ben Laurie]
1981
1982 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1983
1984 *) Since AES128 and AES256 (and similarly Camellia128 and
1985 Camellia256) share a single mask bit in the logic of
1986 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1987 kludge to work properly if AES128 is available and AES256 isn't
1988 (or if Camellia128 is available and Camellia256 isn't).
1989 [Victor Duchovni]
1990
1991 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1992 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1993 When a point or a seed is encoded in a BIT STRING, we need to
1994 prevent the removal of trailing zero bits to get the proper DER
1995 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1996 of a NamedBitList, for which trailing 0 bits need to be removed.)
1997 [Bodo Moeller]
1998
1999 *) Have SSL/TLS server implementation tolerate "mismatched" record
2000 protocol version while receiving ClientHello even if the
2001 ClientHello is fragmented. (The server can't insist on the
2002 particular protocol version it has chosen before the ServerHello
2003 message has informed the client about his choice.)
2004 [Bodo Moeller]
2005
2006 *) Add RFC 3779 support.
2007 [Rob Austein for ARIN, Ben Laurie]
2008
2009 *) Load error codes if they are not already present instead of using a
2010 static variable. This allows them to be cleanly unloaded and reloaded.
2011 Improve header file function name parsing.
2012 [Steve Henson]
2013
2014 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2015 or CAPABILITY handshake as required by RFCs.
2016 [Goetz Babin-Ebell]
2017
2018 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2019
2020 *) Introduce limits to prevent malicious keys being able to
2021 cause a denial of service. (CVE-2006-2940)
2022 [Steve Henson, Bodo Moeller]
2023
2024 *) Fix ASN.1 parsing of certain invalid structures that can result
2025 in a denial of service. (CVE-2006-2937) [Steve Henson]
2026
2027 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2028 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2029
2030 *) Fix SSL client code which could crash if connecting to a
2031 malicious SSLv2 server. (CVE-2006-4343)
2032 [Tavis Ormandy and Will Drewry, Google Security Team]
2033
2034 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2035 match only those. Before that, "AES256-SHA" would be interpreted
2036 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2037 the same strength classification in 0.9.7h) as we currently only
2038 have a single AES bit in the ciphersuite description bitmap.
2039 That change, however, also applied to ciphersuite strings such as
2040 "RC4-MD5" that intentionally matched multiple ciphersuites --
2041 namely, SSL 2.0 ciphersuites in addition to the more common ones
2042 from SSL 3.0/TLS 1.0.
2043
2044 So we change the selection algorithm again: Naming an explicit
2045 ciphersuite selects this one ciphersuite, and any other similar
2046 ciphersuite (same bitmap) from *other* protocol versions.
2047 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2048 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2049
2050 Since SSL 2.0 does not have any ciphersuites for which the
2051 128/256 bit distinction would be relevant, this works for now.
2052 The proper fix will be to use different bits for AES128 and
2053 AES256, which would have avoided the problems from the beginning;
2054 however, bits are scarce, so we can only do this in a new release
2055 (not just a patchlevel) when we can change the SSL_CIPHER
2056 definition to split the single 'unsigned long mask' bitmap into
2057 multiple values to extend the available space.
2058
2059 [Bodo Moeller]
2060
2061 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2062
2063 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2064 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2065
2066 *) Add AES IGE and biIGE modes.
2067 [Ben Laurie]
2068
2069 *) Change the Unix randomness entropy gathering to use poll() when
2070 possible instead of select(), since the latter has some
2071 undesirable limitations.
2072 [Darryl Miles via Richard Levitte and Bodo Moeller]
2073
2074 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2075 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2076 cannot be implicitly activated as part of, e.g., the "AES" alias.
2077 However, please upgrade to OpenSSL 0.9.9[-dev] for
2078 non-experimental use of the ECC ciphersuites to get TLS extension
2079 support, which is required for curve and point format negotiation
2080 to avoid potential handshake problems.
2081 [Bodo Moeller]
2082
2083 *) Disable rogue ciphersuites:
2084
2085 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2086 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2087 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2088
2089 The latter two were purportedly from
2090 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2091 appear there.
2092
2093 Also deactivate the remaining ciphersuites from
2094 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2095 unofficial, and the ID has long expired.
2096 [Bodo Moeller]
2097
2098 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2099 dual-core machines) and other potential thread-safety issues.
2100 [Bodo Moeller]
2101
2102 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2103 versions), which is now available for royalty-free use
2104 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2105 Also, add Camellia TLS ciphersuites from RFC 4132.
2106
2107 To minimize changes between patchlevels in the OpenSSL 0.9.8
2108 series, Camellia remains excluded from compilation unless OpenSSL
2109 is configured with 'enable-camellia'.
2110 [NTT]
2111
2112 *) Disable the padding bug check when compression is in use. The padding
2113 bug check assumes the first packet is of even length, this is not
2114 necessarily true if compresssion is enabled and can result in false
2115 positives causing handshake failure. The actual bug test is ancient
2116 code so it is hoped that implementations will either have fixed it by
2117 now or any which still have the bug do not support compression.
2118 [Steve Henson]
2119
2120 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2121
2122 *) When applying a cipher rule check to see if string match is an explicit
2123 cipher suite and only match that one cipher suite if it is.
2124 [Steve Henson]
2125
2126 *) Link in manifests for VC++ if needed.
2127 [Austin Ziegler <halostatue@gmail.com>]
2128
2129 *) Update support for ECC-based TLS ciphersuites according to
2130 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2131 TLS extensions, which are supported starting with the 0.9.9
2132 branch, not in the OpenSSL 0.9.8 branch).
2133 [Douglas Stebila]
2134
2135 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2136 opaque EVP_CIPHER_CTX handling.
2137 [Steve Henson]
2138
2139 *) Fixes and enhancements to zlib compression code. We now only use
2140 "zlib1.dll" and use the default __cdecl calling convention on Win32
2141 to conform with the standards mentioned here:
2142 http://www.zlib.net/DLL_FAQ.txt
2143 Static zlib linking now works on Windows and the new --with-zlib-include
2144 --with-zlib-lib options to Configure can be used to supply the location
2145 of the headers and library. Gracefully handle case where zlib library
2146 can't be loaded.
2147 [Steve Henson]
2148
2149 *) Several fixes and enhancements to the OID generation code. The old code
2150 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2151 handle numbers larger than ULONG_MAX, truncated printing and had a
2152 non standard OBJ_obj2txt() behaviour.
2153 [Steve Henson]
2154
2155 *) Add support for building of engines under engine/ as shared libraries
2156 under VC++ build system.
2157 [Steve Henson]
2158
2159 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2160 Hopefully, we will not see any false combination of paths any more.
2161 [Richard Levitte]
2162
2163 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2164
2165 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2166 (part of SSL_OP_ALL). This option used to disable the
2167 countermeasure against man-in-the-middle protocol-version
2168 rollback in the SSL 2.0 server implementation, which is a bad
2169 idea. (CVE-2005-2969)
2170
2171 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2172 for Information Security, National Institute of Advanced Industrial
2173 Science and Technology [AIST], Japan)]
2174
2175 *) Add two function to clear and return the verify parameter flags.
2176 [Steve Henson]
2177
2178 *) Keep cipherlists sorted in the source instead of sorting them at
2179 runtime, thus removing the need for a lock.
2180 [Nils Larsch]
2181
2182 *) Avoid some small subgroup attacks in Diffie-Hellman.
2183 [Nick Mathewson and Ben Laurie]
2184
2185 *) Add functions for well-known primes.
2186 [Nick Mathewson]
2187
2188 *) Extended Windows CE support.
2189 [Satoshi Nakamura and Andy Polyakov]
2190
2191 *) Initialize SSL_METHOD structures at compile time instead of during
2192 runtime, thus removing the need for a lock.
2193 [Steve Henson]
2194
2195 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2196 attempting to decrypt each encrypted key in turn. Add support to
2197 smime utility.
2198 [Steve Henson]
2199
2200 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2201
2202 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2203 OpenSSL 0.9.8.]
2204
2205 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2206 [Richard Levitte]
2207
2208 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2209 key into the same file any more.
2210 [Richard Levitte]
2211
2212 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2213 [Andy Polyakov]
2214
2215 *) Add -utf8 command line and config file option to 'ca'.
2216 [Stefan <stf@udoma.org]
2217
2218 *) Removed the macro des_crypt(), as it seems to conflict with some
2219 libraries. Use DES_crypt().
2220 [Richard Levitte]
2221
2222 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2223 involves renaming the source and generated shared-libs for
2224 both. The engines will accept the corrected or legacy ids
2225 ('ncipher' and '4758_cca' respectively) when binding. NB,
2226 this only applies when building 'shared'.
2227 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2228
2229 *) Add attribute functions to EVP_PKEY structure. Modify
2230 PKCS12_create() to recognize a CSP name attribute and
2231 use it. Make -CSP option work again in pkcs12 utility.
2232 [Steve Henson]
2233
2234 *) Add new functionality to the bn blinding code:
2235 - automatic re-creation of the BN_BLINDING parameters after
2236 a fixed number of uses (currently 32)
2237 - add new function for parameter creation
2238 - introduce flags to control the update behaviour of the
2239 BN_BLINDING parameters
2240 - hide BN_BLINDING structure
2241 Add a second BN_BLINDING slot to the RSA structure to improve
2242 performance when a single RSA object is shared among several
2243 threads.
2244 [Nils Larsch]
2245
2246 *) Add support for DTLS.
2247 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2248
2249 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2250 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2251 [Walter Goulet]
2252
2253 *) Remove buggy and incompletet DH cert support from
2254 ssl/ssl_rsa.c and ssl/s3_both.c
2255 [Nils Larsch]
2256
2257 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2258 the apps/openssl applications.
2259 [Nils Larsch]
2260
2261 *) Compile clean with "-Wall -Wmissing-prototypes
2262 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2263 DEBUG_SAFESTACK must also be set.
2264 [Ben Laurie]
2265
2266 *) Change ./Configure so that certain algorithms can be disabled by default.
2267 The new counterpiece to "no-xxx" is "enable-xxx".
2268
2269 The patented RC5 and MDC2 algorithms will now be disabled unless
2270 "enable-rc5" and "enable-mdc2", respectively, are specified.
2271
2272 (IDEA remains enabled despite being patented. This is because IDEA
2273 is frequently required for interoperability, and there is no license
2274 fee for non-commercial use. As before, "no-idea" can be used to
2275 avoid this algorithm.)
2276
2277 [Bodo Moeller]
2278
2279 *) Add processing of proxy certificates (see RFC 3820). This work was
2280 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2281 EGEE (Enabling Grids for E-science in Europe).
2282 [Richard Levitte]
2283
2284 *) RC4 performance overhaul on modern architectures/implementations, such
2285 as Intel P4, IA-64 and AMD64.
2286 [Andy Polyakov]
2287
2288 *) New utility extract-section.pl. This can be used specify an alternative
2289 section number in a pod file instead of having to treat each file as
2290 a separate case in Makefile. This can be done by adding two lines to the
2291 pod file:
2292
2293 =for comment openssl_section:XXX
2294
2295 The blank line is mandatory.
2296
2297 [Steve Henson]
2298
2299 *) New arguments -certform, -keyform and -pass for s_client and s_server
2300 to allow alternative format key and certificate files and passphrase
2301 sources.
2302 [Steve Henson]
2303
2304 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2305 update associated structures and add various utility functions.
2306
2307 Add new policy related verify parameters, include policy checking in
2308 standard verify code. Enhance 'smime' application with extra parameters
2309 to support policy checking and print out.
2310 [Steve Henson]
2311
2312 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2313 Nehemiah processors. These extensions support AES encryption in hardware
2314 as well as RNG (though RNG support is currently disabled).
2315 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2316
2317 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2318 [Geoff Thorpe]
2319
2320 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2321 [Andy Polyakov and a number of other people]
2322
2323 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2324 implementation contributed by IBM.
2325 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2326
2327 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2328 exponent rather than 'unsigned long'. There is a corresponding change to
2329 the new 'rsa_keygen' element of the RSA_METHOD structure.
2330 [Jelte Jansen, Geoff Thorpe]
2331
2332 *) Functionality for creating the initial serial number file is now
2333 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2334
2335 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2336 number file to 1, which is bound to cause problems. To avoid
2337 the problems while respecting compatibility between different 0.9.7
2338 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2339 CA.pl for serial number initialization. With the new release 0.9.8,
2340 we can fix the problem directly in the 'ca' utility.)
2341 [Steve Henson]
2342
2343 *) Reduced header interdepencies by declaring more opaque objects in
2344 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2345 give fewer recursive includes, which could break lazy source code - so
2346 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2347 developers should define this symbol when building and using openssl to
2348 ensure they track the recommended behaviour, interfaces, [etc], but
2349 backwards-compatible behaviour prevails when this isn't defined.
2350 [Geoff Thorpe]
2351
2352 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2353 [Steve Henson]
2354
2355 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2356 This will generate a random key of the appropriate length based on the
2357 cipher context. The EVP_CIPHER can provide its own random key generation
2358 routine to support keys of a specific form. This is used in the des and
2359 3des routines to generate a key of the correct parity. Update S/MIME
2360 code to use new functions and hence generate correct parity DES keys.
2361 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2362 valid (weak or incorrect parity).
2363 [Steve Henson]
2364
2365 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2366 as looking them up. This is useful when the verified structure may contain
2367 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2368 present unless the new PKCS7_NO_CRL flag is asserted.
2369 [Steve Henson]
2370
2371 *) Extend ASN1 oid configuration module. It now additionally accepts the
2372 syntax:
2373
2374 shortName = some long name, 1.2.3.4
2375 [Steve Henson]
2376
2377 *) Reimplemented the BN_CTX implementation. There is now no more static
2378 limitation on the number of variables it can handle nor the depth of the
2379 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2380 information can now expand as required, and rather than having a single
2381 static array of bignums, BN_CTX now uses a linked-list of such arrays
2382 allowing it to expand on demand whilst maintaining the usefulness of
2383 BN_CTX's "bundling".
2384 [Geoff Thorpe]
2385
2386 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2387 to allow all RSA operations to function using a single BN_CTX.
2388 [Geoff Thorpe]
2389
2390 *) Preliminary support for certificate policy evaluation and checking. This
2391 is initially intended to pass the tests outlined in "Conformance Testing
2392 of Relying Party Client Certificate Path Processing Logic" v1.07.
2393 [Steve Henson]
2394
2395 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2396 remained unused and not that useful. A variety of other little bignum
2397 tweaks and fixes have also been made continuing on from the audit (see
2398 below).
2399 [Geoff Thorpe]
2400
2401 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2402 associated ASN1, EVP and SSL functions and old ASN1 macros.
2403 [Richard Levitte]
2404
2405 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2406 and this should never fail. So the return value from the use of
2407 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2408 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2409 [Geoff Thorpe]
2410
2411 *) BN_CTX_get() should return zero-valued bignums, providing the same
2412 initialised value as BN_new().
2413 [Geoff Thorpe, suggested by Ulf Möller]
2414
2415 *) Support for inhibitAnyPolicy certificate extension.
2416 [Steve Henson]
2417
2418 *) An audit of the BIGNUM code is underway, for which debugging code is
2419 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2420 is considered valid when processing BIGNUMs, and causes execution to
2421 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2422 further steps are taken to deliberately pollute unused data in BIGNUM
2423 structures to try and expose faulty code further on. For now, openssl will
2424 (in its default mode of operation) continue to tolerate the inconsistent
2425 forms that it has tolerated in the past, but authors and packagers should
2426 consider trying openssl and their own applications when compiled with
2427 these debugging symbols defined. It will help highlight potential bugs in
2428 their own code, and will improve the test coverage for OpenSSL itself. At
2429 some point, these tighter rules will become openssl's default to improve
2430 maintainability, though the assert()s and other overheads will remain only
2431 in debugging configurations. See bn.h for more details.
2432 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2433
2434 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2435 that can only be obtained through BN_CTX_new() (which implicitly
2436 initialises it). The presence of this function only made it possible
2437 to overwrite an existing structure (and cause memory leaks).
2438 [Geoff Thorpe]
2439
2440 *) Because of the callback-based approach for implementing LHASH as a
2441 template type, lh_insert() adds opaque objects to hash-tables and
2442 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2443 to clean up those corresponding objects before destroying the hash table
2444 (and losing the object pointers). So some over-zealous constifications in
2445 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2446 objects as "const" and the lh_doall[_arg] callback wrappers are not
2447 prototyped to have "const" restrictions on the object pointers they are
2448 given (and so aren't required to cast them away any more).
2449 [Geoff Thorpe]
2450
2451 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2452 (speed) prefers to use its own implementation. The two implementations
2453 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2454 its object type properly exposed (MS_TM) instead of casting to/from "char
2455 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2456 aren't necessarily the greatest nomenclatures - but this is what was used
2457 internally to the implementation so I've used that for now.
2458 [Geoff Thorpe]
2459
2460 *) Ensure that deprecated functions do not get compiled when
2461 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2462 the self-tests were still using deprecated key-generation functions so
2463 these have been updated also.
2464 [Geoff Thorpe]
2465
2466 *) Reorganise PKCS#7 code to separate the digest location functionality
2467 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2468 New function PKCS7_set_digest() to set the digest type for PKCS#7
2469 digestedData type. Add additional code to correctly generate the
2470 digestedData type and add support for this type in PKCS7 initialization
2471 functions.
2472 [Steve Henson]
2473
2474 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2475 structure of type "other".
2476 [Steve Henson]
2477
2478 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2479 sure the loop does correctly stop and breaking ("division by zero")
2480 modulus operations are not performed. The (pre-generated) prime
2481 table crypto/bn/bn_prime.h was already correct, but it could not be
2482 re-generated on some platforms because of the "division by zero"
2483 situation in the script.
2484 [Ralf S. Engelschall]
2485
2486 *) Update support for ECC-based TLS ciphersuites according to
2487 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2488 SHA-1 now is only used for "small" curves (where the
2489 representation of a field element takes up to 24 bytes); for
2490 larger curves, the field element resulting from ECDH is directly
2491 used as premaster secret.
2492 [Douglas Stebila (Sun Microsystems Laboratories)]
2493
2494 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2495 curve secp160r1 to the tests.
2496 [Douglas Stebila (Sun Microsystems Laboratories)]
2497
2498 *) Add the possibility to load symbols globally with DSO.
2499 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2500
2501 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2502 control of the error stack.
2503 [Richard Levitte]
2504
2505 *) Add support for STORE in ENGINE.
2506 [Richard Levitte]
2507
2508 *) Add the STORE type. The intention is to provide a common interface
2509 to certificate and key stores, be they simple file-based stores, or
2510 HSM-type store, or LDAP stores, or...
2511 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2512 [Richard Levitte]
2513
2514 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2515 pass a list of arguments to any function as well as provide a way
2516 for a function to pass data back to the caller.
2517 [Richard Levitte]
2518
2519 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2520 works like BUF_strdup() but can be used to duplicate a portion of
2521 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2522 a memory area.
2523 [Richard Levitte]
2524
2525 *) Add the function sk_find_ex() which works like sk_find(), but will
2526 return an index to an element even if an exact match couldn't be
2527 found. The index is guaranteed to point at the element where the
2528 searched-for key would be inserted to preserve sorting order.
2529 [Richard Levitte]
2530
2531 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2532 takes an extra flags argument for optional functionality. Currently,
2533 the following flags are defined:
2534
2535 OBJ_BSEARCH_VALUE_ON_NOMATCH
2536 This one gets OBJ_bsearch_ex() to return a pointer to the first
2537 element where the comparing function returns a negative or zero
2538 number.
2539
2540 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2541 This one gets OBJ_bsearch_ex() to return a pointer to the first
2542 element where the comparing function returns zero. This is useful
2543 if there are more than one element where the comparing function
2544 returns zero.
2545 [Richard Levitte]
2546
2547 *) Make it possible to create self-signed certificates with 'openssl ca'
2548 in such a way that the self-signed certificate becomes part of the
2549 CA database and uses the same mechanisms for serial number generation
2550 as all other certificate signing. The new flag '-selfsign' enables
2551 this functionality. Adapt CA.sh and CA.pl.in.
2552 [Richard Levitte]
2553
2554 *) Add functionality to check the public key of a certificate request
2555 against a given private. This is useful to check that a certificate
2556 request can be signed by that key (self-signing).
2557 [Richard Levitte]
2558
2559 *) Make it possible to have multiple active certificates with the same
2560 subject in the CA index file. This is done only if the keyword
2561 'unique_subject' is set to 'no' in the main CA section (default
2562 if 'CA_default') of the configuration file. The value is saved
2563 with the database itself in a separate index attribute file,
2564 named like the index file with '.attr' appended to the name.
2565 [Richard Levitte]
2566
2567 *) Generate muti valued AVAs using '+' notation in config files for
2568 req and dirName.
2569 [Steve Henson]
2570
2571 *) Support for nameConstraints certificate extension.
2572 [Steve Henson]
2573
2574 *) Support for policyConstraints certificate extension.
2575 [Steve Henson]
2576
2577 *) Support for policyMappings certificate extension.
2578 [Steve Henson]
2579
2580 *) Make sure the default DSA_METHOD implementation only uses its
2581 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2582 and change its own handlers to be NULL so as to remove unnecessary
2583 indirection. This lets alternative implementations fallback to the
2584 default implementation more easily.
2585 [Geoff Thorpe]
2586
2587 *) Support for directoryName in GeneralName related extensions
2588 in config files.
2589 [Steve Henson]
2590
2591 *) Make it possible to link applications using Makefile.shared.
2592 Make that possible even when linking against static libraries!
2593 [Richard Levitte]
2594
2595 *) Support for single pass processing for S/MIME signing. This now
2596 means that S/MIME signing can be done from a pipe, in addition
2597 cleartext signing (multipart/signed type) is effectively streaming
2598 and the signed data does not need to be all held in memory.
2599
2600 This is done with a new flag PKCS7_STREAM. When this flag is set
2601 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2602 is done after the data is output (and digests calculated) in
2603 SMIME_write_PKCS7().
2604 [Steve Henson]
2605
2606 *) Add full support for -rpath/-R, both in shared libraries and
2607 applications, at least on the platforms where it's known how
2608 to do it.
2609 [Richard Levitte]
2610
2611 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2612 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2613 will now compute a table of multiples of the generator that
2614 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2615 faster (notably in the case of a single point multiplication,
2616 scalar * generator).
2617 [Nils Larsch, Bodo Moeller]
2618
2619 *) IPv6 support for certificate extensions. The various extensions
2620 which use the IP:a.b.c.d can now take IPv6 addresses using the
2621 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2622 correctly.
2623 [Steve Henson]
2624
2625 *) Added an ENGINE that implements RSA by performing private key
2626 exponentiations with the GMP library. The conversions to and from
2627 GMP's mpz_t format aren't optimised nor are any montgomery forms
2628 cached, and on x86 it appears OpenSSL's own performance has caught up.
2629 However there are likely to be other architectures where GMP could
2630 provide a boost. This ENGINE is not built in by default, but it can be
2631 specified at Configure time and should be accompanied by the necessary
2632 linker additions, eg;
2633 ./config -DOPENSSL_USE_GMP -lgmp
2634 [Geoff Thorpe]
2635
2636 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2637 testing availability of engines with "-t" - the old behaviour is
2638 produced by increasing the feature's verbosity with "-tt".
2639 [Geoff Thorpe]
2640
2641 *) ECDSA routines: under certain error conditions uninitialized BN objects
2642 could be freed. Solution: make sure initialization is performed early
2643 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2644 via PR#459)
2645 [Lutz Jaenicke]
2646
2647 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2648 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2649 software implementations. For DSA and DH, parameter generation can
2650 also be overriden by providing the appropriate method callbacks.
2651 [Geoff Thorpe]
2652
2653 *) Change the "progress" mechanism used in key-generation and
2654 primality testing to functions that take a new BN_GENCB pointer in
2655 place of callback/argument pairs. The new API functions have "_ex"
2656 postfixes and the older functions are reimplemented as wrappers for
2657 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2658 declarations of the old functions to help (graceful) attempts to
2659 migrate to the new functions. Also, the new key-generation API
2660 functions operate on a caller-supplied key-structure and return
2661 success/failure rather than returning a key or NULL - this is to
2662 help make "keygen" another member function of RSA_METHOD etc.
2663
2664 Example for using the new callback interface:
2665
2666 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2667 void *my_arg = ...;
2668 BN_GENCB my_cb;
2669
2670 BN_GENCB_set(&my_cb, my_callback, my_arg);
2671
2672 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2673 /* For the meaning of a, b in calls to my_callback(), see the
2674 * documentation of the function that calls the callback.
2675 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2676 * my_callback should return 1 if it wants BN_is_prime_ex()
2677 * to continue, or 0 to stop.
2678 */
2679
2680 [Geoff Thorpe]
2681
2682 *) Change the ZLIB compression method to be stateful, and make it
2683 available to TLS with the number defined in
2684 draft-ietf-tls-compression-04.txt.
2685 [Richard Levitte]
2686
2687 *) Add the ASN.1 structures and functions for CertificatePair, which
2688 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2689
2690 CertificatePair ::= SEQUENCE {
2691 forward [0] Certificate OPTIONAL,
2692 reverse [1] Certificate OPTIONAL,
2693 -- at least one of the pair shall be present -- }
2694
2695 Also implement the PEM functions to read and write certificate
2696 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2697
2698 This needed to be defined, mostly for the sake of the LDAP
2699 attribute crossCertificatePair, but may prove useful elsewhere as
2700 well.
2701 [Richard Levitte]
2702
2703 *) Make it possible to inhibit symlinking of shared libraries in
2704 Makefile.shared, for Cygwin's sake.
2705 [Richard Levitte]
2706
2707 *) Extend the BIGNUM API by creating a function
2708 void BN_set_negative(BIGNUM *a, int neg);
2709 and a macro that behave like
2710 int BN_is_negative(const BIGNUM *a);
2711
2712 to avoid the need to access 'a->neg' directly in applications.
2713 [Nils Larsch]
2714
2715 *) Implement fast modular reduction for pseudo-Mersenne primes
2716 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2717 EC_GROUP_new_curve_GFp() will now automatically use this
2718 if applicable.
2719 [Nils Larsch <nla@trustcenter.de>]
2720
2721 *) Add new lock type (CRYPTO_LOCK_BN).
2722 [Bodo Moeller]
2723
2724 *) Change the ENGINE framework to automatically load engines
2725 dynamically from specific directories unless they could be
2726 found to already be built in or loaded. Move all the
2727 current engines except for the cryptodev one to a new
2728 directory engines/.
2729 The engines in engines/ are built as shared libraries if
2730 the "shared" options was given to ./Configure or ./config.
2731 Otherwise, they are inserted in libcrypto.a.
2732 /usr/local/ssl/engines is the default directory for dynamic
2733 engines, but that can be overriden at configure time through
2734 the usual use of --prefix and/or --openssldir, and at run
2735 time with the environment variable OPENSSL_ENGINES.
2736 [Geoff Thorpe and Richard Levitte]
2737
2738 *) Add Makefile.shared, a helper makefile to build shared
2739 libraries. Addapt Makefile.org.
2740 [Richard Levitte]
2741
2742 *) Add version info to Win32 DLLs.
2743 [Peter 'Luna' Runestig" <peter@runestig.com>]
2744
2745 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2746 can be added using this API to created arbitrary PKCS#12
2747 files while avoiding the low level API.
2748
2749 New options to PKCS12_create(), key or cert can be NULL and
2750 will then be omitted from the output file. The encryption
2751 algorithm NIDs can be set to -1 for no encryption, the mac
2752 iteration count can be set to 0 to omit the mac.
2753
2754 Enhance pkcs12 utility by making the -nokeys and -nocerts
2755 options work when creating a PKCS#12 file. New option -nomac
2756 to omit the mac, NONE can be set for an encryption algorithm.
2757 New code is modified to use the enhanced PKCS12_create()
2758 instead of the low level API.
2759 [Steve Henson]
2760
2761 *) Extend ASN1 encoder to support indefinite length constructed
2762 encoding. This can output sequences tags and octet strings in
2763 this form. Modify pk7_asn1.c to support indefinite length
2764 encoding. This is experimental and needs additional code to
2765 be useful, such as an ASN1 bio and some enhanced streaming
2766 PKCS#7 code.
2767
2768 Extend template encode functionality so that tagging is passed
2769 down to the template encoder.
2770 [Steve Henson]
2771
2772 *) Let 'openssl req' fail if an argument to '-newkey' is not
2773 recognized instead of using RSA as a default.
2774 [Bodo Moeller]
2775
2776 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2777 As these are not official, they are not included in "ALL";
2778 the "ECCdraft" ciphersuite group alias can be used to select them.
2779 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2780
2781 *) Add ECDH engine support.
2782 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2783
2784 *) Add ECDH in new directory crypto/ecdh/.
2785 [Douglas Stebila (Sun Microsystems Laboratories)]
2786
2787 *) Let BN_rand_range() abort with an error after 100 iterations
2788 without success (which indicates a broken PRNG).
2789 [Bodo Moeller]
2790
2791 *) Change BN_mod_sqrt() so that it verifies that the input value
2792 is really the square of the return value. (Previously,
2793 BN_mod_sqrt would show GIGO behaviour.)
2794 [Bodo Moeller]
2795
2796 *) Add named elliptic curves over binary fields from X9.62, SECG,
2797 and WAP/WTLS; add OIDs that were still missing.
2798
2799 [Sheueling Chang Shantz and Douglas Stebila
2800 (Sun Microsystems Laboratories)]
2801
2802 *) Extend the EC library for elliptic curves over binary fields
2803 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2804 New EC_METHOD:
2805
2806 EC_GF2m_simple_method
2807
2808 New API functions:
2809
2810 EC_GROUP_new_curve_GF2m
2811 EC_GROUP_set_curve_GF2m
2812 EC_GROUP_get_curve_GF2m
2813 EC_POINT_set_affine_coordinates_GF2m
2814 EC_POINT_get_affine_coordinates_GF2m
2815 EC_POINT_set_compressed_coordinates_GF2m
2816
2817 Point compression for binary fields is disabled by default for
2818 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2819 enable it).
2820
2821 As binary polynomials are represented as BIGNUMs, various members
2822 of the EC_GROUP and EC_POINT data structures can be shared
2823 between the implementations for prime fields and binary fields;
2824 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2825 are essentially identical to their ..._GFp counterparts.
2826 (For simplicity, the '..._GFp' prefix has been dropped from
2827 various internal method names.)
2828
2829 An internal 'field_div' method (similar to 'field_mul' and
2830 'field_sqr') has been added; this is used only for binary fields.
2831
2832 [Sheueling Chang Shantz and Douglas Stebila
2833 (Sun Microsystems Laboratories)]
2834
2835 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2836 through methods ('mul', 'precompute_mult').
2837
2838 The generic implementations (now internally called 'ec_wNAF_mul'
2839 and 'ec_wNAF_precomputed_mult') remain the default if these
2840 methods are undefined.
2841
2842 [Sheueling Chang Shantz and Douglas Stebila
2843 (Sun Microsystems Laboratories)]
2844
2845 *) New function EC_GROUP_get_degree, which is defined through
2846 EC_METHOD. For curves over prime fields, this returns the bit
2847 length of the modulus.
2848
2849 [Sheueling Chang Shantz and Douglas Stebila
2850 (Sun Microsystems Laboratories)]
2851
2852 *) New functions EC_GROUP_dup, EC_POINT_dup.
2853 (These simply call ..._new and ..._copy).
2854
2855 [Sheueling Chang Shantz and Douglas Stebila
2856 (Sun Microsystems Laboratories)]
2857
2858 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2859 Polynomials are represented as BIGNUMs (where the sign bit is not
2860 used) in the following functions [macros]:
2861
2862 BN_GF2m_add
2863 BN_GF2m_sub [= BN_GF2m_add]
2864 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2865 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2866 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2867 BN_GF2m_mod_inv
2868 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2869 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2870 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2871 BN_GF2m_cmp [= BN_ucmp]
2872
2873 (Note that only the 'mod' functions are actually for fields GF(2^m).
2874 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2875
2876 For some functions, an the irreducible polynomial defining a
2877 field can be given as an 'unsigned int[]' with strictly
2878 decreasing elements giving the indices of those bits that are set;
2879 i.e., p[] represents the polynomial
2880 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2881 where
2882 p[0] > p[1] > ... > p[k] = 0.
2883 This applies to the following functions:
2884
2885 BN_GF2m_mod_arr
2886 BN_GF2m_mod_mul_arr
2887 BN_GF2m_mod_sqr_arr
2888 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2889 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2890 BN_GF2m_mod_exp_arr
2891 BN_GF2m_mod_sqrt_arr
2892 BN_GF2m_mod_solve_quad_arr
2893 BN_GF2m_poly2arr
2894 BN_GF2m_arr2poly
2895
2896 Conversion can be performed by the following functions:
2897
2898 BN_GF2m_poly2arr
2899 BN_GF2m_arr2poly
2900
2901 bntest.c has additional tests for binary polynomial arithmetic.
2902
2903 Two implementations for BN_GF2m_mod_div() are available.
2904 The default algorithm simply uses BN_GF2m_mod_inv() and
2905 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2906 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2907 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2908
2909 [Sheueling Chang Shantz and Douglas Stebila
2910 (Sun Microsystems Laboratories)]
2911
2912 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2913 functionality is disabled at compile-time.
2914 [Douglas Stebila <douglas.stebila@sun.com>]
2915
2916 *) Change default behaviour of 'openssl asn1parse' so that more
2917 information is visible when viewing, e.g., a certificate:
2918
2919 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2920 mode the content of non-printable OCTET STRINGs is output in a
2921 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2922 avoid the appearance of a printable string.
2923 [Nils Larsch <nla@trustcenter.de>]
2924
2925 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2926 functions
2927 EC_GROUP_set_asn1_flag()
2928 EC_GROUP_get_asn1_flag()
2929 EC_GROUP_set_point_conversion_form()
2930 EC_GROUP_get_point_conversion_form()
2931 These control ASN1 encoding details:
2932 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2933 has been set to OPENSSL_EC_NAMED_CURVE.
2934 - Points are encoded in uncompressed form by default; options for
2935 asn1_for are as for point2oct, namely
2936 POINT_CONVERSION_COMPRESSED
2937 POINT_CONVERSION_UNCOMPRESSED
2938 POINT_CONVERSION_HYBRID
2939
2940 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2941 functions
2942 EC_GROUP_set_seed()
2943 EC_GROUP_get0_seed()
2944 EC_GROUP_get_seed_len()
2945 This is used only for ASN1 purposes (so far).
2946 [Nils Larsch <nla@trustcenter.de>]
2947
2948 *) Add 'field_type' member to EC_METHOD, which holds the NID
2949 of the appropriate field type OID. The new function
2950 EC_METHOD_get_field_type() returns this value.
2951 [Nils Larsch <nla@trustcenter.de>]
2952
2953 *) Add functions
2954 EC_POINT_point2bn()
2955 EC_POINT_bn2point()
2956 EC_POINT_point2hex()
2957 EC_POINT_hex2point()
2958 providing useful interfaces to EC_POINT_point2oct() and
2959 EC_POINT_oct2point().
2960 [Nils Larsch <nla@trustcenter.de>]
2961
2962 *) Change internals of the EC library so that the functions
2963 EC_GROUP_set_generator()
2964 EC_GROUP_get_generator()
2965 EC_GROUP_get_order()
2966 EC_GROUP_get_cofactor()
2967 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2968 to methods, which would lead to unnecessary code duplication when
2969 adding different types of curves.
2970 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2971
2972 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2973 arithmetic, and such that modified wNAFs are generated
2974 (which avoid length expansion in many cases).
2975 [Bodo Moeller]
2976
2977 *) Add a function EC_GROUP_check_discriminant() (defined via
2978 EC_METHOD) that verifies that the curve discriminant is non-zero.
2979
2980 Add a function EC_GROUP_check() that makes some sanity tests
2981 on a EC_GROUP, its generator and order. This includes
2982 EC_GROUP_check_discriminant().
2983 [Nils Larsch <nla@trustcenter.de>]
2984
2985 *) Add ECDSA in new directory crypto/ecdsa/.
2986
2987 Add applications 'openssl ecparam' and 'openssl ecdsa'
2988 (these are based on 'openssl dsaparam' and 'openssl dsa').
2989
2990 ECDSA support is also included in various other files across the
2991 library. Most notably,
2992 - 'openssl req' now has a '-newkey ecdsa:file' option;
2993 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2994 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2995 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2996 them suitable for ECDSA where domain parameters must be
2997 extracted before the specific public key;
2998 - ECDSA engine support has been added.
2999 [Nils Larsch <nla@trustcenter.de>]
3000
3001 *) Include some named elliptic curves, and add OIDs from X9.62,
3002 SECG, and WAP/WTLS. Each curve can be obtained from the new
3003 function
3004 EC_GROUP_new_by_curve_name(),
3005 and the list of available named curves can be obtained with
3006 EC_get_builtin_curves().
3007 Also add a 'curve_name' member to EC_GROUP objects, which can be
3008 accessed via
3009 EC_GROUP_set_curve_name()
3010 EC_GROUP_get_curve_name()
3011 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3012
3013 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3014 was actually never needed) and in BN_mul(). The removal in BN_mul()
3015 required a small change in bn_mul_part_recursive() and the addition
3016 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3017 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3018 bn_sub_words() and bn_add_words() except they take arrays with
3019 differing sizes.
3020 [Richard Levitte]
3021
3022 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3023
3024 *) Cleanse PEM buffers before freeing them since they may contain
3025 sensitive data.
3026 [Benjamin Bennett <ben@psc.edu>]
3027
3028 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3029 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3030 authentication-only ciphersuites.
3031 [Bodo Moeller]
3032
3033 *) Since AES128 and AES256 share a single mask bit in the logic of
3034 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3035 kludge to work properly if AES128 is available and AES256 isn't.
3036 [Victor Duchovni]
3037
3038 *) Expand security boundary to match 1.1.1 module.
3039 [Steve Henson]
3040
3041 *) Remove redundant features: hash file source, editing of test vectors
3042 modify fipsld to use external fips_premain.c signature.
3043 [Steve Henson]
3044
3045 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3046 run algorithm test programs.
3047 [Steve Henson]
3048
3049 *) Make algorithm test programs more tolerant of whitespace.
3050 [Steve Henson]
3051
3052 *) Have SSL/TLS server implementation tolerate "mismatched" record
3053 protocol version while receiving ClientHello even if the
3054 ClientHello is fragmented. (The server can't insist on the
3055 particular protocol version it has chosen before the ServerHello
3056 message has informed the client about his choice.)
3057 [Bodo Moeller]
3058
3059 *) Load error codes if they are not already present instead of using a
3060 static variable. This allows them to be cleanly unloaded and reloaded.
3061 [Steve Henson]
3062
3063 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3064
3065 *) Introduce limits to prevent malicious keys being able to
3066 cause a denial of service. (CVE-2006-2940)
3067 [Steve Henson, Bodo Moeller]
3068
3069 *) Fix ASN.1 parsing of certain invalid structures that can result
3070 in a denial of service. (CVE-2006-2937) [Steve Henson]
3071
3072 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3073 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3074
3075 *) Fix SSL client code which could crash if connecting to a
3076 malicious SSLv2 server. (CVE-2006-4343)
3077 [Tavis Ormandy and Will Drewry, Google Security Team]
3078
3079 *) Change ciphersuite string processing so that an explicit
3080 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3081 will no longer include "AES128-SHA"), and any other similar
3082 ciphersuite (same bitmap) from *other* protocol versions (so that
3083 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3084 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3085 changes from 0.9.8b and 0.9.8d.
3086 [Bodo Moeller]
3087
3088 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3089
3090 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3091 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3092
3093 *) Change the Unix randomness entropy gathering to use poll() when
3094 possible instead of select(), since the latter has some
3095 undesirable limitations.
3096 [Darryl Miles via Richard Levitte and Bodo Moeller]
3097
3098 *) Disable rogue ciphersuites:
3099
3100 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3101 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3102 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3103
3104 The latter two were purportedly from
3105 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3106 appear there.
3107
3108 Also deactive the remaining ciphersuites from
3109 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3110 unofficial, and the ID has long expired.
3111 [Bodo Moeller]
3112
3113 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3114 dual-core machines) and other potential thread-safety issues.
3115 [Bodo Moeller]
3116
3117 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3118
3119 *) Adapt fipsld and the build system to link against the validated FIPS
3120 module in FIPS mode.
3121 [Steve Henson]
3122
3123 *) Fixes for VC++ 2005 build under Windows.
3124 [Steve Henson]
3125
3126 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3127 from a Windows bash shell such as MSYS. It is autodetected from the
3128 "config" script when run from a VC++ environment. Modify standard VC++
3129 build to use fipscanister.o from the GNU make build.
3130 [Steve Henson]
3131
3132 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3133
3134 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3135 The value now differs depending on if you build for FIPS or not.
3136 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3137 safely run with a non-FIPSed libcrypto, as it may crash because of
3138 the difference induced by this change.
3139 [Andy Polyakov]
3140
3141 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3142
3143 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3144 (part of SSL_OP_ALL). This option used to disable the
3145 countermeasure against man-in-the-middle protocol-version
3146 rollback in the SSL 2.0 server implementation, which is a bad
3147 idea. (CVE-2005-2969)
3148
3149 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3150 for Information Security, National Institute of Advanced Industrial
3151 Science and Technology [AIST], Japan)]
3152
3153 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3154 mainly for FIPS compliance and not fully integrated at this stage.
3155 [Steve Henson]
3156
3157 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3158 the exponentiation using a fixed-length exponent. (Otherwise,
3159 the information leaked through timing could expose the secret key
3160 after many signatures; cf. Bleichenbacher's attack on DSA with
3161 biased k.)
3162 [Bodo Moeller]
3163
3164 *) Make a new fixed-window mod_exp implementation the default for
3165 RSA, DSA, and DH private-key operations so that the sequence of
3166 squares and multiplies and the memory access pattern are
3167 independent of the particular secret key. This will mitigate
3168 cache-timing and potential related attacks.
3169
3170 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3171 and this is automatically used by BN_mod_exp_mont() if the new flag
3172 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3173 will use this BN flag for private exponents unless the flag
3174 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3175 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3176
3177 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3178
3179 *) Change the client implementation for SSLv23_method() and
3180 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3181 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3182 (Previously, the SSL 2.0 backwards compatible Client Hello
3183 message format would be used even with SSL_OP_NO_SSLv2.)
3184 [Bodo Moeller]
3185
3186 *) Add support for smime-type MIME parameter in S/MIME messages which some
3187 clients need.
3188 [Steve Henson]
3189
3190 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3191 a threadsafe manner. Modify rsa code to use new function and add calls
3192 to dsa and dh code (which had race conditions before).
3193 [Steve Henson]
3194
3195 *) Include the fixed error library code in the C error file definitions
3196 instead of fixing them up at runtime. This keeps the error code
3197 structures constant.
3198 [Steve Henson]
3199
3200 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3201
3202 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3203 OpenSSL 0.9.8.]
3204
3205 *) Fixes for newer kerberos headers. NB: the casts are needed because
3206 the 'length' field is signed on one version and unsigned on another
3207 with no (?) obvious way to tell the difference, without these VC++
3208 complains. Also the "definition" of FAR (blank) is no longer included
3209 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3210 some needed definitions.
3211 [Steve Henson]
3212
3213 *) Undo Cygwin change.
3214 [Ulf Möller]
3215
3216 *) Added support for proxy certificates according to RFC 3820.
3217 Because they may be a security thread to unaware applications,
3218 they must be explicitely allowed in run-time. See
3219 docs/HOWTO/proxy_certificates.txt for further information.
3220 [Richard Levitte]
3221
3222 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3223
3224 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3225 server and client random values. Previously
3226 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3227 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3228
3229 This change has negligible security impact because:
3230
3231 1. Server and client random values still have 24 bytes of pseudo random
3232 data.
3233
3234 2. Server and client random values are sent in the clear in the initial
3235 handshake.
3236
3237 3. The master secret is derived using the premaster secret (48 bytes in
3238 size for static RSA ciphersuites) as well as client server and random
3239 values.
3240
3241 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3242 to our attention.
3243
3244 [Stephen Henson, reported by UK NISCC]
3245
3246 *) Use Windows randomness collection on Cygwin.
3247 [Ulf Möller]
3248
3249 *) Fix hang in EGD/PRNGD query when communication socket is closed
3250 prematurely by EGD/PRNGD.
3251 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3252
3253 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3254 [Steve Henson]
3255
3256 *) Back-port of selected performance improvements from development
3257 branch, as well as improved support for PowerPC platforms.
3258 [Andy Polyakov]
3259
3260 *) Add lots of checks for memory allocation failure, error codes to indicate
3261 failure and freeing up memory if a failure occurs.
3262 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3263
3264 *) Add new -passin argument to dgst.
3265 [Steve Henson]
3266
3267 *) Perform some character comparisons of different types in X509_NAME_cmp:
3268 this is needed for some certificates that reencode DNs into UTF8Strings
3269 (in violation of RFC3280) and can't or wont issue name rollover
3270 certificates.
3271 [Steve Henson]
3272
3273 *) Make an explicit check during certificate validation to see that
3274 the CA setting in each certificate on the chain is correct. As a
3275 side effect always do the following basic checks on extensions,
3276 not just when there's an associated purpose to the check:
3277
3278 - if there is an unhandled critical extension (unless the user
3279 has chosen to ignore this fault)
3280 - if the path length has been exceeded (if one is set at all)
3281 - that certain extensions fit the associated purpose (if one has
3282 been given)
3283 [Richard Levitte]
3284
3285 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3286
3287 *) Avoid a race condition when CRLs are checked in a multi threaded
3288 environment. This would happen due to the reordering of the revoked
3289 entries during signature checking and serial number lookup. Now the
3290 encoding is cached and the serial number sort performed under a lock.
3291 Add new STACK function sk_is_sorted().
3292 [Steve Henson]
3293
3294 *) Add Delta CRL to the extension code.
3295 [Steve Henson]
3296
3297 *) Various fixes to s3_pkt.c so alerts are sent properly.
3298 [David Holmes <d.holmes@f5.com>]
3299
3300 *) Reduce the chances of duplicate issuer name and serial numbers (in
3301 violation of RFC3280) using the OpenSSL certificate creation utilities.
3302 This is done by creating a random 64 bit value for the initial serial
3303 number when a serial number file is created or when a self signed
3304 certificate is created using 'openssl req -x509'. The initial serial
3305 number file is created using 'openssl x509 -next_serial' in CA.pl
3306 rather than being initialized to 1.
3307 [Steve Henson]
3308
3309 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3310
3311 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3312 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3313 [Joe Orton, Steve Henson]
3314
3315 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3316 (CVE-2004-0112)
3317 [Joe Orton, Steve Henson]
3318
3319 *) Make it possible to have multiple active certificates with the same
3320 subject in the CA index file. This is done only if the keyword
3321 'unique_subject' is set to 'no' in the main CA section (default
3322 if 'CA_default') of the configuration file. The value is saved
3323 with the database itself in a separate index attribute file,
3324 named like the index file with '.attr' appended to the name.
3325 [Richard Levitte]
3326
3327 *) X509 verify fixes. Disable broken certificate workarounds when
3328 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3329 keyUsage extension present. Don't accept CRLs with unhandled critical
3330 extensions: since verify currently doesn't process CRL extensions this
3331 rejects a CRL with *any* critical extensions. Add new verify error codes
3332 for these cases.
3333 [Steve Henson]
3334
3335 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3336 A clarification of RFC2560 will require the use of OCTET STRINGs and
3337 some implementations cannot handle the current raw format. Since OpenSSL
3338 copies and compares OCSP nonces as opaque blobs without any attempt at
3339 parsing them this should not create any compatibility issues.
3340 [Steve Henson]
3341
3342 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3343 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3344 this HMAC (and other) operations are several times slower than OpenSSL
3345 < 0.9.7.
3346 [Steve Henson]
3347
3348 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3349 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3350
3351 *) Use the correct content when signing type "other".
3352 [Steve Henson]
3353
3354 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3355
3356 *) Fix various bugs revealed by running the NISCC test suite:
3357
3358 Stop out of bounds reads in the ASN1 code when presented with
3359 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3360
3361 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3362
3363 If verify callback ignores invalid public key errors don't try to check
3364 certificate signature with the NULL public key.
3365
3366 [Steve Henson]
3367
3368 *) New -ignore_err option in ocsp application to stop the server
3369 exiting on the first error in a request.
3370 [Steve Henson]
3371
3372 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3373 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3374 specifications.
3375 [Steve Henson]
3376
3377 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3378 extra data after the compression methods not only for TLS 1.0
3379 but also for SSL 3.0 (as required by the specification).
3380 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3381
3382 *) Change X509_certificate_type() to mark the key as exported/exportable
3383 when it's 512 *bits* long, not 512 bytes.
3384 [Richard Levitte]
3385
3386 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3387 blocks during encryption.
3388 [Richard Levitte]
3389
3390 *) Various fixes to base64 BIO and non blocking I/O. On write
3391 flushes were not handled properly if the BIO retried. On read
3392 data was not being buffered properly and had various logic bugs.
3393 This also affects blocking I/O when the data being decoded is a
3394 certain size.
3395 [Steve Henson]
3396
3397 *) Various S/MIME bugfixes and compatibility changes:
3398 output correct application/pkcs7 MIME type if
3399 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3400 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3401 of files as .eml work). Correctly handle very long lines in MIME
3402 parser.
3403 [Steve Henson]
3404
3405 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3406
3407 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3408 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3409 a protocol version number mismatch like a decryption error
3410 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3411 [Bodo Moeller]
3412
3413 *) Turn on RSA blinding by default in the default implementation
3414 to avoid a timing attack. Applications that don't want it can call
3415 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3416 They would be ill-advised to do so in most cases.
3417 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3418
3419 *) Change RSA blinding code so that it works when the PRNG is not
3420 seeded (in this case, the secret RSA exponent is abused as
3421 an unpredictable seed -- if it is not unpredictable, there
3422 is no point in blinding anyway). Make RSA blinding thread-safe
3423 by remembering the creator's thread ID in rsa->blinding and
3424 having all other threads use local one-time blinding factors
3425 (this requires more computation than sharing rsa->blinding, but
3426 avoids excessive locking; and if an RSA object is not shared
3427 between threads, blinding will still be very fast).
3428 [Bodo Moeller]
3429
3430 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3431 ENGINE as defaults for all supported algorithms irrespective of
3432 the 'flags' parameter. 'flags' is now honoured, so applications
3433 should make sure they are passing it correctly.
3434 [Geoff Thorpe]
3435
3436 *) Target "mingw" now allows native Windows code to be generated in
3437 the Cygwin environment as well as with the MinGW compiler.
3438 [Ulf Moeller]
3439
3440 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3441
3442 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3443 via timing by performing a MAC computation even if incorrrect
3444 block cipher padding has been found. This is a countermeasure
3445 against active attacks where the attacker has to distinguish
3446 between bad padding and a MAC verification error. (CVE-2003-0078)
3447
3448 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3449 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3450 Martin Vuagnoux (EPFL, Ilion)]
3451
3452 *) Make the no-err option work as intended. The intention with no-err
3453 is not to have the whole error stack handling routines removed from
3454 libcrypto, it's only intended to remove all the function name and
3455 reason texts, thereby removing some of the footprint that may not
3456 be interesting if those errors aren't displayed anyway.
3457
3458 NOTE: it's still possible for any application or module to have it's
3459 own set of error texts inserted. The routines are there, just not
3460 used by default when no-err is given.
3461 [Richard Levitte]
3462
3463 *) Add support for FreeBSD on IA64.
3464 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3465
3466 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3467 Kerberos function mit_des_cbc_cksum(). Before this change,
3468 the value returned by DES_cbc_cksum() was like the one from
3469 mit_des_cbc_cksum(), except the bytes were swapped.
3470 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3471
3472 *) Allow an application to disable the automatic SSL chain building.
3473 Before this a rather primitive chain build was always performed in
3474 ssl3_output_cert_chain(): an application had no way to send the
3475 correct chain if the automatic operation produced an incorrect result.
3476
3477 Now the chain builder is disabled if either:
3478
3479 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3480
3481 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3482
3483 The reasoning behind this is that an application would not want the
3484 auto chain building to take place if extra chain certificates are
3485 present and it might also want a means of sending no additional
3486 certificates (for example the chain has two certificates and the
3487 root is omitted).
3488 [Steve Henson]
3489
3490 *) Add the possibility to build without the ENGINE framework.
3491 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3492
3493 *) Under Win32 gmtime() can return NULL: check return value in
3494 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3495 [Steve Henson]
3496
3497 *) DSA routines: under certain error conditions uninitialized BN objects
3498 could be freed. Solution: make sure initialization is performed early
3499 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3500 Nils Larsch <nla@trustcenter.de> via PR#459)
3501 [Lutz Jaenicke]
3502
3503 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3504 checked on reconnect on the client side, therefore session resumption
3505 could still fail with a "ssl session id is different" error. This
3506 behaviour is masked when SSL_OP_ALL is used due to
3507 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3508 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3509 followup to PR #377.
3510 [Lutz Jaenicke]
3511
3512 *) IA-32 assembler support enhancements: unified ELF targets, support
3513 for SCO/Caldera platforms, fix for Cygwin shared build.
3514 [Andy Polyakov]
3515
3516 *) Add support for FreeBSD on sparc64. As a consequence, support for
3517 FreeBSD on non-x86 processors is separate from x86 processors on
3518 the config script, much like the NetBSD support.
3519 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3520
3521 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3522
3523 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3524 OpenSSL 0.9.7.]
3525
3526 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3527 code (06) was taken as the first octet of the session ID and the last
3528 octet was ignored consequently. As a result SSLv2 client side session
3529 caching could not have worked due to the session ID mismatch between
3530 client and server.
3531 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3532 PR #377.
3533 [Lutz Jaenicke]
3534
3535 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3536 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3537 removed entirely.
3538 [Richard Levitte]
3539
3540 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3541 seems that in spite of existing for more than a year, many application
3542 author have done nothing to provide the necessary callbacks, which
3543 means that this particular engine will not work properly anywhere.
3544 This is a very unfortunate situation which forces us, in the name
3545 of usability, to give the hw_ncipher.c a static lock, which is part
3546 of libcrypto.
3547 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3548 appear in 0.9.8 or later. We EXPECT application authors to have
3549 dealt properly with this when 0.9.8 is released (unless we actually
3550 make such changes in the libcrypto locking code that changes will
3551 have to be made anyway).
3552 [Richard Levitte]
3553
3554 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3555 octets have been read, EOF or an error occurs. Without this change
3556 some truncated ASN1 structures will not produce an error.
3557 [Steve Henson]
3558
3559 *) Disable Heimdal support, since it hasn't been fully implemented.
3560 Still give the possibility to force the use of Heimdal, but with
3561 warnings and a request that patches get sent to openssl-dev.
3562 [Richard Levitte]
3563
3564 *) Add the VC-CE target, introduce the WINCE sysname, and add
3565 INSTALL.WCE and appropriate conditionals to make it build.
3566 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3567
3568 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3569 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3570 edit numbers of the version.
3571 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3572
3573 *) Introduce safe string copy and catenation functions
3574 (BUF_strlcpy() and BUF_strlcat()).
3575 [Ben Laurie (CHATS) and Richard Levitte]
3576
3577 *) Avoid using fixed-size buffers for one-line DNs.
3578 [Ben Laurie (CHATS)]
3579
3580 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3581 resizing buffers containing secrets, and use where appropriate.
3582 [Ben Laurie (CHATS)]
3583
3584 *) Avoid using fixed size buffers for configuration file location.
3585 [Ben Laurie (CHATS)]
3586
3587 *) Avoid filename truncation for various CA files.
3588 [Ben Laurie (CHATS)]
3589
3590 *) Use sizeof in preference to magic numbers.
3591 [Ben Laurie (CHATS)]
3592
3593 *) Avoid filename truncation in cert requests.
3594 [Ben Laurie (CHATS)]
3595
3596 *) Add assertions to check for (supposedly impossible) buffer
3597 overflows.
3598 [Ben Laurie (CHATS)]
3599
3600 *) Don't cache truncated DNS entries in the local cache (this could
3601 potentially lead to a spoofing attack).
3602 [Ben Laurie (CHATS)]
3603
3604 *) Fix various buffers to be large enough for hex/decimal
3605 representations in a platform independent manner.
3606 [Ben Laurie (CHATS)]
3607
3608 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3609 resizing buffers containing secrets, and use where appropriate.
3610 [Ben Laurie (CHATS)]
3611
3612 *) Add BIO_indent() to avoid much slightly worrying code to do
3613 indents.
3614 [Ben Laurie (CHATS)]
3615
3616 *) Convert sprintf()/BIO_puts() to BIO_printf().
3617 [Ben Laurie (CHATS)]
3618
3619 *) buffer_gets() could terminate with the buffer only half
3620 full. Fixed.
3621 [Ben Laurie (CHATS)]
3622
3623 *) Add assertions to prevent user-supplied crypto functions from
3624 overflowing internal buffers by having large block sizes, etc.
3625 [Ben Laurie (CHATS)]
3626
3627 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3628 unconditionally).
3629 [Ben Laurie (CHATS)]
3630
3631 *) Eliminate unused copy of key in RC4.
3632 [Ben Laurie (CHATS)]
3633
3634 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3635 [Ben Laurie (CHATS)]
3636
3637 *) Fix off-by-one error in EGD path.
3638 [Ben Laurie (CHATS)]
3639
3640 *) If RANDFILE path is too long, ignore instead of truncating.
3641 [Ben Laurie (CHATS)]
3642
3643 *) Eliminate unused and incorrectly sized X.509 structure
3644 CBCParameter.
3645 [Ben Laurie (CHATS)]
3646
3647 *) Eliminate unused and dangerous function knumber().
3648 [Ben Laurie (CHATS)]
3649
3650 *) Eliminate unused and dangerous structure, KSSL_ERR.
3651 [Ben Laurie (CHATS)]
3652
3653 *) Protect against overlong session ID context length in an encoded
3654 session object. Since these are local, this does not appear to be
3655 exploitable.
3656 [Ben Laurie (CHATS)]
3657
3658 *) Change from security patch (see 0.9.6e below) that did not affect
3659 the 0.9.6 release series:
3660
3661 Remote buffer overflow in SSL3 protocol - an attacker could
3662 supply an oversized master key in Kerberos-enabled versions.
3663 (CVE-2002-0657)
3664 [Ben Laurie (CHATS)]
3665
3666 *) Change the SSL kerb5 codes to match RFC 2712.
3667 [Richard Levitte]
3668
3669 *) Make -nameopt work fully for req and add -reqopt switch.
3670 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3671
3672 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3673 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3674
3675 *) Make sure tests can be performed even if the corresponding algorithms
3676 have been removed entirely. This was also the last step to make
3677 OpenSSL compilable with DJGPP under all reasonable conditions.
3678 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3679
3680 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3681 to allow version independent disabling of normally unselected ciphers,
3682 which may be activated as a side-effect of selecting a single cipher.
3683
3684 (E.g., cipher list string "RSA" enables ciphersuites that are left
3685 out of "ALL" because they do not provide symmetric encryption.
3686 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3687 [Lutz Jaenicke, Bodo Moeller]
3688
3689 *) Add appropriate support for separate platform-dependent build
3690 directories. The recommended way to make a platform-dependent
3691 build directory is the following (tested on Linux), maybe with
3692 some local tweaks:
3693
3694 # Place yourself outside of the OpenSSL source tree. In
3695 # this example, the environment variable OPENSSL_SOURCE
3696 # is assumed to contain the absolute OpenSSL source directory.
3697 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3698 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3699 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3700 mkdir -p `dirname $F`
3701 ln -s $OPENSSL_SOURCE/$F $F
3702 done
3703
3704 To be absolutely sure not to disturb the source tree, a "make clean"
3705 is a good thing. If it isn't successfull, don't worry about it,
3706 it probably means the source directory is very clean.
3707 [Richard Levitte]
3708
3709 *) Make sure any ENGINE control commands make local copies of string
3710 pointers passed to them whenever necessary. Otherwise it is possible
3711 the caller may have overwritten (or deallocated) the original string
3712 data when a later ENGINE operation tries to use the stored values.
3713 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3714
3715 *) Improve diagnostics in file reading and command-line digests.
3716 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3717
3718 *) Add AES modes CFB and OFB to the object database. Correct an
3719 error in AES-CFB decryption.
3720 [Richard Levitte]
3721
3722 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3723 allows existing EVP_CIPHER_CTX structures to be reused after
3724 calling EVP_*Final(). This behaviour is used by encryption
3725 BIOs and some applications. This has the side effect that
3726 applications must explicitly clean up cipher contexts with
3727 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3728 [Steve Henson]
3729
3730 *) Check the values of dna and dnb in bn_mul_recursive before calling
3731 bn_mul_comba (a non zero value means the a or b arrays do not contain
3732 n2 elements) and fallback to bn_mul_normal if either is not zero.
3733 [Steve Henson]
3734
3735 *) Fix escaping of non-ASCII characters when using the -subj option
3736 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3737 [Lutz Jaenicke]
3738
3739 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3740 form for "surname", serialNumber has no short form.
3741 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3742 therefore remove "mail" short name for "internet 7".
3743 The OID for unique identifiers in X509 certificates is
3744 x500UniqueIdentifier, not uniqueIdentifier.
3745 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3746 [Lutz Jaenicke]
3747
3748 *) Add an "init" command to the ENGINE config module and auto initialize
3749 ENGINEs. Without any "init" command the ENGINE will be initialized
3750 after all ctrl commands have been executed on it. If init=1 the
3751 ENGINE is initailized at that point (ctrls before that point are run
3752 on the uninitialized ENGINE and after on the initialized one). If
3753 init=0 then the ENGINE will not be iniatialized at all.
3754 [Steve Henson]
3755
3756 *) Fix the 'app_verify_callback' interface so that the user-defined
3757 argument is actually passed to the callback: In the
3758 SSL_CTX_set_cert_verify_callback() prototype, the callback
3759 declaration has been changed from
3760 int (*cb)()
3761 into
3762 int (*cb)(X509_STORE_CTX *,void *);
3763 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3764 i=s->ctx->app_verify_callback(&ctx)
3765 has been changed into
3766 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3767
3768 To update applications using SSL_CTX_set_cert_verify_callback(),
3769 a dummy argument can be added to their callback functions.
3770 [D. K. Smetters <smetters@parc.xerox.com>]
3771
3772 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3773 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3774
3775 *) Add and OPENSSL_LOAD_CONF define which will cause
3776 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3777 This allows older applications to transparently support certain
3778 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3779 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3780 load the config file and OPENSSL_add_all_algorithms_conf() which will
3781 always load it have also been added.
3782 [Steve Henson]
3783
3784 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3785 Adjust NIDs and EVP layer.
3786 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3787
3788 *) Config modules support in openssl utility.
3789
3790 Most commands now load modules from the config file,
3791 though in a few (such as version) this isn't done
3792 because it couldn't be used for anything.
3793
3794 In the case of ca and req the config file used is
3795 the same as the utility itself: that is the -config
3796 command line option can be used to specify an
3797 alternative file.
3798 [Steve Henson]
3799
3800 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3801 use "openssl_conf" if filename is NULL use default openssl config file.
3802 [Steve Henson]
3803
3804 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3805 config section name. Add a new flag to tolerate a missing config file
3806 and move code to CONF_modules_load_file().
3807 [Steve Henson]
3808
3809 *) Support for crypto accelerator cards from Accelerated Encryption
3810 Processing, www.aep.ie. (Use engine 'aep')
3811 The support was copied from 0.9.6c [engine] and adapted/corrected
3812 to work with the new engine framework.
3813 [AEP Inc. and Richard Levitte]
3814
3815 *) Support for SureWare crypto accelerator cards from Baltimore
3816 Technologies. (Use engine 'sureware')
3817 The support was copied from 0.9.6c [engine] and adapted
3818 to work with the new engine framework.
3819 [Richard Levitte]
3820
3821 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3822 make the newer ENGINE framework commands for the CHIL engine work.
3823 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3824
3825 *) Make it possible to produce shared libraries on ReliantUNIX.
3826 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3827
3828 *) Add the configuration target debug-linux-ppro.
3829 Make 'openssl rsa' use the general key loading routines
3830 implemented in apps.c, and make those routines able to
3831 handle the key format FORMAT_NETSCAPE and the variant
3832 FORMAT_IISSGC.
3833 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3834
3835 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3836 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3837
3838 *) Add -keyform to rsautl, and document -engine.
3839 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3840
3841 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3842 BIO_R_NO_SUCH_FILE error code rather than the generic
3843 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3844 [Ben Laurie]
3845
3846 *) Add new functions
3847 ERR_peek_last_error
3848 ERR_peek_last_error_line
3849 ERR_peek_last_error_line_data.
3850 These are similar to
3851 ERR_peek_error
3852 ERR_peek_error_line
3853 ERR_peek_error_line_data,
3854 but report on the latest error recorded rather than the first one
3855 still in the error queue.
3856 [Ben Laurie, Bodo Moeller]
3857
3858 *) default_algorithms option in ENGINE config module. This allows things
3859 like:
3860 default_algorithms = ALL
3861 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3862 [Steve Henson]
3863
3864 *) Prelminary ENGINE config module.
3865 [Steve Henson]
3866
3867 *) New experimental application configuration code.
3868 [Steve Henson]
3869
3870 *) Change the AES code to follow the same name structure as all other
3871 symmetric ciphers, and behave the same way. Move everything to
3872 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3873 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3874
3875 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3876 [Ben Laurie and Theo de Raadt]
3877
3878 *) Add option to output public keys in req command.
3879 [Massimiliano Pala madwolf@openca.org]
3880
3881 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3882 (up to about 10% better than before for P-192 and P-224).
3883 [Bodo Moeller]
3884
3885 *) New functions/macros
3886
3887 SSL_CTX_set_msg_callback(ctx, cb)
3888 SSL_CTX_set_msg_callback_arg(ctx, arg)
3889 SSL_set_msg_callback(ssl, cb)
3890 SSL_set_msg_callback_arg(ssl, arg)
3891
3892 to request calling a callback function
3893
3894 void cb(int write_p, int version, int content_type,
3895 const void *buf, size_t len, SSL *ssl, void *arg)
3896
3897 whenever a protocol message has been completely received
3898 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3899 protocol version according to which the SSL library interprets
3900 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3901 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3902 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3903 specification (change_cipher_spec(20), alert(21), handshake(22)).
3904 'buf' and 'len' point to the actual message, 'ssl' to the
3905 SSL object, and 'arg' is the application-defined value set by
3906 SSL[_CTX]_set_msg_callback_arg().
3907
3908 'openssl s_client' and 'openssl s_server' have new '-msg' options
3909 to enable a callback that displays all protocol messages.
3910 [Bodo Moeller]
3911
3912 *) Change the shared library support so shared libraries are built as
3913 soon as the corresponding static library is finished, and thereby get
3914 openssl and the test programs linked against the shared library.
3915 This still only happens when the keyword "shard" has been given to
3916 the configuration scripts.
3917
3918 NOTE: shared library support is still an experimental thing, and
3919 backward binary compatibility is still not guaranteed.
3920 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3921
3922 *) Add support for Subject Information Access extension.
3923 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3924
3925 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3926 additional bytes when new memory had to be allocated, not just
3927 when reusing an existing buffer.
3928 [Bodo Moeller]
3929
3930 *) New command line and configuration option 'utf8' for the req command.
3931 This allows field values to be specified as UTF8 strings.
3932 [Steve Henson]
3933
3934 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3935 runs for the former and machine-readable output for the latter.
3936 [Ben Laurie]
3937
3938 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3939 of the e-mail address in the DN (i.e., it will go into a certificate
3940 extension only). The new configuration file option 'email_in_dn = no'
3941 has the same effect.
3942 [Massimiliano Pala madwolf@openca.org]
3943
3944 *) Change all functions with names starting with des_ to be starting
3945 with DES_ instead. Add wrappers that are compatible with libdes,
3946 but are named _ossl_old_des_*. Finally, add macros that map the
3947 des_* symbols to the corresponding _ossl_old_des_* if libdes
3948 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3949 desired, the des_* symbols will be mapped to DES_*, with one
3950 exception.
3951
3952 Since we provide two compatibility mappings, the user needs to
3953 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3954 compatibility is desired. The default (i.e., when that macro
3955 isn't defined) is OpenSSL 0.9.6c compatibility.
3956
3957 There are also macros that enable and disable the support of old
3958 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3959 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3960 are defined, the default will apply: to support the old des routines.
3961
3962 In either case, one must include openssl/des.h to get the correct
3963 definitions. Do not try to just include openssl/des_old.h, that
3964 won't work.
3965
3966 NOTE: This is a major break of an old API into a new one. Software
3967 authors are encouraged to switch to the DES_ style functions. Some
3968 time in the future, des_old.h and the libdes compatibility functions
3969 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3970 default), and then completely removed.
3971 [Richard Levitte]
3972
3973 *) Test for certificates which contain unsupported critical extensions.
3974 If such a certificate is found during a verify operation it is
3975 rejected by default: this behaviour can be overridden by either
3976 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3977 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3978 X509_supported_extension() has also been added which returns 1 if a
3979 particular extension is supported.
3980 [Steve Henson]
3981
3982 *) Modify the behaviour of EVP cipher functions in similar way to digests
3983 to retain compatibility with existing code.
3984 [Steve Henson]
3985
3986 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3987 compatibility with existing code. In particular the 'ctx' parameter does
3988 not have to be to be initialized before the call to EVP_DigestInit() and
3989 it is tidied up after a call to EVP_DigestFinal(). New function
3990 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3991 EVP_MD_CTX_copy() changed to not require the destination to be
3992 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3993 requires the destination to be valid.
3994
3995 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3996 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3997 [Steve Henson]
3998
3999 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4000 so that complete 'Handshake' protocol structures are kept in memory
4001 instead of overwriting 'msg_type' and 'length' with 'body' data.
4002 [Bodo Moeller]
4003
4004 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4005 [Massimo Santin via Richard Levitte]
4006
4007 *) Major restructuring to the underlying ENGINE code. This includes
4008 reduction of linker bloat, separation of pure "ENGINE" manipulation
4009 (initialisation, etc) from functionality dealing with implementations
4010 of specific crypto iterfaces. This change also introduces integrated
4011 support for symmetric ciphers and digest implementations - so ENGINEs
4012 can now accelerate these by providing EVP_CIPHER and EVP_MD
4013 implementations of their own. This is detailed in crypto/engine/README
4014 as it couldn't be adequately described here. However, there are a few
4015 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4016 were changed in the original introduction of ENGINE code have now
4017 reverted back - the hooking from this code to ENGINE is now a good
4018 deal more passive and at run-time, operations deal directly with
4019 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4020 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4021 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4022 they were not being used by the framework as there is no concept of a
4023 BIGNUM_METHOD and they could not be generalised to the new
4024 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4025 ENGINE_cpy() has been removed as it cannot be consistently defined in
4026 the new code.
4027 [Geoff Thorpe]
4028
4029 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4030 [Steve Henson]
4031
4032 *) Change mkdef.pl to sort symbols that get the same entry number,
4033 and make sure the automatically generated functions ERR_load_*
4034 become part of libeay.num as well.
4035 [Richard Levitte]
4036
4037 *) New function SSL_renegotiate_pending(). This returns true once
4038 renegotiation has been requested (either SSL_renegotiate() call
4039 or HelloRequest/ClientHello receveived from the peer) and becomes
4040 false once a handshake has been completed.
4041 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4042 sends a HelloRequest, but does not ensure that a handshake takes
4043 place. SSL_renegotiate_pending() is useful for checking if the
4044 client has followed the request.)
4045 [Bodo Moeller]
4046
4047 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4048 By default, clients may request session resumption even during
4049 renegotiation (if session ID contexts permit); with this option,
4050 session resumption is possible only in the first handshake.
4051
4052 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4053 more bits available for options that should not be part of
4054 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4055 [Bodo Moeller]
4056
4057 *) Add some demos for certificate and certificate request creation.
4058 [Steve Henson]
4059
4060 *) Make maximum certificate chain size accepted from the peer application
4061 settable (SSL*_get/set_max_cert_list()), as proposed by
4062 "Douglas E. Engert" <deengert@anl.gov>.
4063 [Lutz Jaenicke]
4064
4065 *) Add support for shared libraries for Unixware-7
4066 (Boyd Lynn Gerber <gerberb@zenez.com>).
4067 [Lutz Jaenicke]
4068
4069 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4070 be done prior to destruction. Use this to unload error strings from
4071 ENGINEs that load their own error strings. NB: This adds two new API
4072 functions to "get" and "set" this destroy handler in an ENGINE.
4073 [Geoff Thorpe]
4074
4075 *) Alter all existing ENGINE implementations (except "openssl" and
4076 "openbsd") to dynamically instantiate their own error strings. This
4077 makes them more flexible to be built both as statically-linked ENGINEs
4078 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4079 Also, add stub code to each that makes building them as self-contained
4080 shared-libraries easier (see README.ENGINE).
4081 [Geoff Thorpe]
4082
4083 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4084 implementations into applications that are completely implemented in
4085 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4086 commands that can be used to configure what shared-library to load and
4087 to control aspects of the way it is handled. Also, made an update to
4088 the README.ENGINE file that brings its information up-to-date and
4089 provides some information and instructions on the "dynamic" ENGINE
4090 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4091 [Geoff Thorpe]
4092
4093 *) Make it possible to unload ranges of ERR strings with a new
4094 "ERR_unload_strings" function.
4095 [Geoff Thorpe]
4096
4097 *) Add a copy() function to EVP_MD.
4098 [Ben Laurie]
4099
4100 *) Make EVP_MD routines take a context pointer instead of just the
4101 md_data void pointer.
4102 [Ben Laurie]
4103
4104 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4105 that the digest can only process a single chunk of data
4106 (typically because it is provided by a piece of
4107 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4108 is only going to provide a single chunk of data, and hence the
4109 framework needn't accumulate the data for oneshot drivers.
4110 [Ben Laurie]
4111
4112 *) As with "ERR", make it possible to replace the underlying "ex_data"
4113 functions. This change also alters the storage and management of global
4114 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4115 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4116 index counters. The API functions that use this state have been changed
4117 to take a "class_index" rather than pointers to the class's local STACK
4118 and counter, and there is now an API function to dynamically create new
4119 classes. This centralisation allows us to (a) plug a lot of the
4120 thread-safety problems that existed, and (b) makes it possible to clean
4121 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4122 such data would previously have always leaked in application code and
4123 workarounds were in place to make the memory debugging turn a blind eye
4124 to it. Application code that doesn't use this new function will still
4125 leak as before, but their memory debugging output will announce it now
4126 rather than letting it slide.
4127
4128 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4129 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4130 has a return value to indicate success or failure.
4131 [Geoff Thorpe]
4132
4133 *) Make it possible to replace the underlying "ERR" functions such that the
4134 global state (2 LHASH tables and 2 locks) is only used by the "default"
4135 implementation. This change also adds two functions to "get" and "set"
4136 the implementation prior to it being automatically set the first time
4137 any other ERR function takes place. Ie. an application can call "get",
4138 pass the return value to a module it has just loaded, and that module
4139 can call its own "set" function using that value. This means the
4140 module's "ERR" operations will use (and modify) the error state in the
4141 application and not in its own statically linked copy of OpenSSL code.
4142 [Geoff Thorpe]
4143
4144 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4145 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4146 the operation, and provides a more encapsulated way for external code
4147 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4148 to use these functions rather than manually incrementing the counts.
4149
4150 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4151 [Geoff Thorpe]
4152
4153 *) Add EVP test program.
4154 [Ben Laurie]
4155
4156 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4157 [Ben Laurie]
4158
4159 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4160 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4161 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4162 These allow a CRL to be built without having to access X509_CRL fields
4163 directly. Modify 'ca' application to use new functions.
4164 [Steve Henson]
4165
4166 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4167 bug workarounds. Rollback attack detection is a security feature.
4168 The problem will only arise on OpenSSL servers when TLSv1 is not
4169 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4170 Software authors not wanting to support TLSv1 will have special reasons
4171 for their choice and can explicitly enable this option.
4172 [Bodo Moeller, Lutz Jaenicke]
4173
4174 *) Rationalise EVP so it can be extended: don't include a union of
4175 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4176 (similar to those existing for EVP_CIPHER_CTX).
4177 Usage example:
4178
4179 EVP_MD_CTX md;
4180
4181 EVP_MD_CTX_init(&md); /* new function call */
4182 EVP_DigestInit(&md, EVP_sha1());
4183 EVP_DigestUpdate(&md, in, len);
4184 EVP_DigestFinal(&md, out, NULL);
4185 EVP_MD_CTX_cleanup(&md); /* new function call */
4186
4187 [Ben Laurie]
4188
4189 *) Make DES key schedule conform to the usual scheme, as well as
4190 correcting its structure. This means that calls to DES functions
4191 now have to pass a pointer to a des_key_schedule instead of a
4192 plain des_key_schedule (which was actually always a pointer
4193 anyway): E.g.,
4194
4195 des_key_schedule ks;
4196
4197 des_set_key_checked(..., &ks);
4198 des_ncbc_encrypt(..., &ks, ...);
4199
4200 (Note that a later change renames 'des_...' into 'DES_...'.)
4201 [Ben Laurie]
4202
4203 *) Initial reduction of linker bloat: the use of some functions, such as
4204 PEM causes large amounts of unused functions to be linked in due to
4205 poor organisation. For example pem_all.c contains every PEM function
4206 which has a knock on effect of linking in large amounts of (unused)
4207 ASN1 code. Grouping together similar functions and splitting unrelated
4208 functions prevents this.
4209 [Steve Henson]
4210
4211 *) Cleanup of EVP macros.
4212 [Ben Laurie]
4213
4214 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4215 correct _ecb suffix.
4216 [Ben Laurie]
4217
4218 *) Add initial OCSP responder support to ocsp application. The
4219 revocation information is handled using the text based index
4220 use by the ca application. The responder can either handle
4221 requests generated internally, supplied in files (for example
4222 via a CGI script) or using an internal minimal server.
4223 [Steve Henson]
4224
4225 *) Add configuration choices to get zlib compression for TLS.
4226 [Richard Levitte]
4227
4228 *) Changes to Kerberos SSL for RFC 2712 compliance:
4229 1. Implemented real KerberosWrapper, instead of just using
4230 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4231 2. Implemented optional authenticator field of KerberosWrapper.
4232
4233 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4234 and authenticator structs; see crypto/krb5/.
4235
4236 Generalized Kerberos calls to support multiple Kerberos libraries.
4237 [Vern Staats <staatsvr@asc.hpc.mil>,
4238 Jeffrey Altman <jaltman@columbia.edu>
4239 via Richard Levitte]
4240
4241 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4242 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4243 values for each of the key sizes rather than having just
4244 parameters (and 'speed' generating keys each time).
4245 [Geoff Thorpe]
4246
4247 *) Speed up EVP routines.
4248 Before:
4249 encrypt
4250 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4251 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4252 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4253 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4254 decrypt
4255 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4256 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4257 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4258 After:
4259 encrypt
4260 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4261 decrypt
4262 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4263 [Ben Laurie]
4264
4265 *) Added the OS2-EMX target.
4266 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4267
4268 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4269 to support NCONF routines in extension code. New function CONF_set_nconf()
4270 to allow functions which take an NCONF to also handle the old LHASH
4271 structure: this means that the old CONF compatible routines can be
4272 retained (in particular wrt extensions) without having to duplicate the
4273 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4274 [Steve Henson]
4275
4276 *) Enhance the general user interface with mechanisms for inner control
4277 and with possibilities to have yes/no kind of prompts.
4278 [Richard Levitte]
4279
4280 *) Change all calls to low level digest routines in the library and
4281 applications to use EVP. Add missing calls to HMAC_cleanup() and
4282 don't assume HMAC_CTX can be copied using memcpy().
4283 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4284
4285 *) Add the possibility to control engines through control names but with
4286 arbitrary arguments instead of just a string.
4287 Change the key loaders to take a UI_METHOD instead of a callback
4288 function pointer. NOTE: this breaks binary compatibility with earlier
4289 versions of OpenSSL [engine].
4290 Adapt the nCipher code for these new conditions and add a card insertion
4291 callback.
4292 [Richard Levitte]
4293
4294 *) Enhance the general user interface with mechanisms to better support
4295 dialog box interfaces, application-defined prompts, the possibility
4296 to use defaults (for example default passwords from somewhere else)
4297 and interrupts/cancellations.
4298 [Richard Levitte]
4299
4300 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4301 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4302 [Steve Henson]
4303
4304 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4305 tidy up some unnecessarily weird code in 'sk_new()').
4306 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4307
4308 *) Change the key loading routines for ENGINEs to use the same kind
4309 callback (pem_password_cb) as all other routines that need this
4310 kind of callback.
4311 [Richard Levitte]
4312
4313 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4314 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4315 than this minimum value is recommended.
4316 [Lutz Jaenicke]
4317
4318 *) New random seeder for OpenVMS, using the system process statistics
4319 that are easily reachable.
4320 [Richard Levitte]
4321
4322 *) Windows apparently can't transparently handle global
4323 variables defined in DLLs. Initialisations such as:
4324
4325 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4326
4327 wont compile. This is used by the any applications that need to
4328 declare their own ASN1 modules. This was fixed by adding the option
4329 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4330 needed for static libraries under Win32.
4331 [Steve Henson]
4332
4333 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4334 setting of purpose and trust fields. New X509_STORE trust and
4335 purpose functions and tidy up setting in other SSL functions.
4336 [Steve Henson]
4337
4338 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4339 structure. These are inherited by X509_STORE_CTX when it is
4340 initialised. This allows various defaults to be set in the
4341 X509_STORE structure (such as flags for CRL checking and custom
4342 purpose or trust settings) for functions which only use X509_STORE_CTX
4343 internally such as S/MIME.
4344
4345 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4346 trust settings if they are not set in X509_STORE. This allows X509_STORE
4347 purposes and trust (in S/MIME for example) to override any set by default.
4348
4349 Add command line options for CRL checking to smime, s_client and s_server
4350 applications.
4351 [Steve Henson]
4352
4353 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4354 are set then the CRL is looked up in the X509_STORE structure and
4355 its validity and signature checked, then if the certificate is found
4356 in the CRL the verify fails with a revoked error.
4357
4358 Various new CRL related callbacks added to X509_STORE_CTX structure.
4359
4360 Command line options added to 'verify' application to support this.
4361
4362 This needs some additional work, such as being able to handle multiple
4363 CRLs with different times, extension based lookup (rather than just
4364 by subject name) and ultimately more complete V2 CRL extension
4365 handling.
4366 [Steve Henson]
4367
4368 *) Add a general user interface API (crypto/ui/). This is designed
4369 to replace things like des_read_password and friends (backward
4370 compatibility functions using this new API are provided).
4371 The purpose is to remove prompting functions from the DES code
4372 section as well as provide for prompting through dialog boxes in
4373 a window system and the like.
4374 [Richard Levitte]
4375
4376 *) Add "ex_data" support to ENGINE so implementations can add state at a
4377 per-structure level rather than having to store it globally.
4378 [Geoff]
4379
4380 *) Make it possible for ENGINE structures to be copied when retrieved by
4381 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4382 This causes the "original" ENGINE structure to act like a template,
4383 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4384 operational state can be localised to each ENGINE structure, despite the
4385 fact they all share the same "methods". New ENGINE structures returned in
4386 this case have no functional references and the return value is the single
4387 structural reference. This matches the single structural reference returned
4388 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4389 ENGINE structure.
4390 [Geoff]
4391
4392 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4393 needs to match any other type at all we need to manually clear the
4394 tag cache.
4395 [Steve Henson]
4396
4397 *) Changes to the "openssl engine" utility to include;
4398 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4399 about an ENGINE's available control commands.
4400 - executing control commands from command line arguments using the
4401 '-pre' and '-post' switches. '-post' is only used if '-t' is
4402 specified and the ENGINE is successfully initialised. The syntax for
4403 the individual commands are colon-separated, for example;
4404 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4405 [Geoff]
4406
4407 *) New dynamic control command support for ENGINEs. ENGINEs can now
4408 declare their own commands (numbers), names (strings), descriptions,
4409 and input types for run-time discovery by calling applications. A
4410 subset of these commands are implicitly classed as "executable"
4411 depending on their input type, and only these can be invoked through
4412 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4413 can be based on user input, config files, etc). The distinction is
4414 that "executable" commands cannot return anything other than a boolean
4415 result and can only support numeric or string input, whereas some
4416 discoverable commands may only be for direct use through
4417 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4418 pointers, or other custom uses. The "executable" commands are to
4419 support parameterisations of ENGINE behaviour that can be
4420 unambiguously defined by ENGINEs and used consistently across any
4421 OpenSSL-based application. Commands have been added to all the
4422 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4423 control over shared-library paths without source code alterations.
4424 [Geoff]
4425
4426 *) Changed all ENGINE implementations to dynamically allocate their
4427 ENGINEs rather than declaring them statically. Apart from this being
4428 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4429 this also allows the implementations to compile without using the
4430 internal engine_int.h header.
4431 [Geoff]
4432
4433 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4434 'const' value. Any code that should be able to modify a RAND_METHOD
4435 should already have non-const pointers to it (ie. they should only
4436 modify their own ones).
4437 [Geoff]
4438
4439 *) Made a variety of little tweaks to the ENGINE code.
4440 - "atalla" and "ubsec" string definitions were moved from header files
4441 to C code. "nuron" string definitions were placed in variables
4442 rather than hard-coded - allowing parameterisation of these values
4443 later on via ctrl() commands.
4444 - Removed unused "#if 0"'d code.
4445 - Fixed engine list iteration code so it uses ENGINE_free() to release
4446 structural references.
4447 - Constified the RAND_METHOD element of ENGINE structures.
4448 - Constified various get/set functions as appropriate and added
4449 missing functions (including a catch-all ENGINE_cpy that duplicates
4450 all ENGINE values onto a new ENGINE except reference counts/state).
4451 - Removed NULL parameter checks in get/set functions. Setting a method
4452 or function to NULL is a way of cancelling out a previously set
4453 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4454 and doesn't justify the extra error symbols and code.
4455 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4456 flags from engine_int.h to engine.h.
4457 - Changed prototypes for ENGINE handler functions (init(), finish(),
4458 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4459 [Geoff]
4460
4461 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4462 to the algorithm using long division. The binary algorithm can be
4463 used only if the modulus is odd. On 32-bit systems, it is faster
4464 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4465 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4466 up to 450 bits. In 64-bit environments, the binary algorithm
4467 appears to be advantageous for much longer moduli; here we use it
4468 for moduli up to 2048 bits.
4469 [Bodo Moeller]
4470
4471 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4472 could not support the combine flag in choice fields.
4473 [Steve Henson]
4474
4475 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4476 extensions from a certificate request to the certificate.
4477 [Steve Henson]
4478
4479 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4480 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4481 file: this allows the display of the certificate about to be
4482 signed to be customised, to allow certain fields to be included
4483 or excluded and extension details. The old system didn't display
4484 multicharacter strings properly, omitted fields not in the policy
4485 and couldn't display additional details such as extensions.
4486 [Steve Henson]
4487
4488 *) Function EC_POINTs_mul for multiple scalar multiplication
4489 of an arbitrary number of elliptic curve points
4490 \sum scalars[i]*points[i],
4491 optionally including the generator defined for the EC_GROUP:
4492 scalar*generator + \sum scalars[i]*points[i].
4493
4494 EC_POINT_mul is a simple wrapper function for the typical case
4495 that the point list has just one item (besides the optional
4496 generator).
4497 [Bodo Moeller]
4498
4499 *) First EC_METHODs for curves over GF(p):
4500
4501 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4502 operations and provides various method functions that can also
4503 operate with faster implementations of modular arithmetic.
4504
4505 EC_GFp_mont_method() reuses most functions that are part of
4506 EC_GFp_simple_method, but uses Montgomery arithmetic.
4507
4508 [Bodo Moeller; point addition and point doubling
4509 implementation directly derived from source code provided by
4510 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4511
4512 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4513 crypto/ec/ec_lib.c):
4514
4515 Curves are EC_GROUP objects (with an optional group generator)
4516 based on EC_METHODs that are built into the library.
4517
4518 Points are EC_POINT objects based on EC_GROUP objects.
4519
4520 Most of the framework would be able to handle curves over arbitrary
4521 finite fields, but as there are no obvious types for fields other
4522 than GF(p), some functions are limited to that for now.
4523 [Bodo Moeller]
4524
4525 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4526 that the file contains a complete HTTP response.
4527 [Richard Levitte]
4528
4529 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4530 change the def and num file printf format specifier from "%-40sXXX"
4531 to "%-39s XXX". The latter will always guarantee a space after the
4532 field while the former will cause them to run together if the field
4533 is 40 of more characters long.
4534 [Steve Henson]
4535
4536 *) Constify the cipher and digest 'method' functions and structures
4537 and modify related functions to take constant EVP_MD and EVP_CIPHER
4538 pointers.
4539 [Steve Henson]
4540
4541 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4542 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4543 [Bodo Moeller]
4544
4545 *) Modify EVP_Digest*() routines so they now return values. Although the
4546 internal software routines can never fail additional hardware versions
4547 might.
4548 [Steve Henson]
4549
4550 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4551
4552 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4553 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4554
4555 ASN1 error codes
4556 ERR_R_NESTED_ASN1_ERROR
4557 ...
4558 ERR_R_MISSING_ASN1_EOS
4559 were 4 .. 9, conflicting with
4560 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4561 ...
4562 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4563 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4564
4565 Add new error code 'ERR_R_INTERNAL_ERROR'.
4566 [Bodo Moeller]
4567
4568 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4569 suffices.
4570 [Bodo Moeller]
4571
4572 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4573 sets the subject name for a new request or supersedes the
4574 subject name in a given request. Formats that can be parsed are
4575 'CN=Some Name, OU=myOU, C=IT'
4576 and
4577 'CN=Some Name/OU=myOU/C=IT'.
4578
4579 Add options '-batch' and '-verbose' to 'openssl req'.
4580 [Massimiliano Pala <madwolf@hackmasters.net>]
4581
4582 *) Introduce the possibility to access global variables through
4583 functions on platform were that's the best way to handle exporting
4584 global variables in shared libraries. To enable this functionality,
4585 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4586 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4587 is normally done by Configure or something similar).
4588
4589 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4590 in the source file (foo.c) like this:
4591
4592 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4593 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4594
4595 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4596 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4597
4598 OPENSSL_DECLARE_GLOBAL(int,foo);
4599 #define foo OPENSSL_GLOBAL_REF(foo)
4600 OPENSSL_DECLARE_GLOBAL(double,bar);
4601 #define bar OPENSSL_GLOBAL_REF(bar)
4602
4603 The #defines are very important, and therefore so is including the
4604 header file everywhere where the defined globals are used.
4605
4606 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4607 of ASN.1 items, but that structure is a bit different.
4608
4609 The largest change is in util/mkdef.pl which has been enhanced with
4610 better and easier to understand logic to choose which symbols should
4611 go into the Windows .def files as well as a number of fixes and code
4612 cleanup (among others, algorithm keywords are now sorted
4613 lexicographically to avoid constant rewrites).
4614 [Richard Levitte]
4615
4616 *) In BN_div() keep a copy of the sign of 'num' before writing the
4617 result to 'rm' because if rm==num the value will be overwritten
4618 and produce the wrong result if 'num' is negative: this caused
4619 problems with BN_mod() and BN_nnmod().
4620 [Steve Henson]
4621
4622 *) Function OCSP_request_verify(). This checks the signature on an
4623 OCSP request and verifies the signer certificate. The signer
4624 certificate is just checked for a generic purpose and OCSP request
4625 trust settings.
4626 [Steve Henson]
4627
4628 *) Add OCSP_check_validity() function to check the validity of OCSP
4629 responses. OCSP responses are prepared in real time and may only
4630 be a few seconds old. Simply checking that the current time lies
4631 between thisUpdate and nextUpdate max reject otherwise valid responses
4632 caused by either OCSP responder or client clock inaccuracy. Instead
4633 we allow thisUpdate and nextUpdate to fall within a certain period of
4634 the current time. The age of the response can also optionally be
4635 checked. Two new options -validity_period and -status_age added to
4636 ocsp utility.
4637 [Steve Henson]
4638
4639 *) If signature or public key algorithm is unrecognized print out its
4640 OID rather that just UNKNOWN.
4641 [Steve Henson]
4642
4643 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4644 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4645 ID to be generated from the issuer certificate alone which can then be
4646 passed to OCSP_id_issuer_cmp().
4647 [Steve Henson]
4648
4649 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4650 ASN1 modules to export functions returning ASN1_ITEM pointers
4651 instead of the ASN1_ITEM structures themselves. This adds several
4652 new macros which allow the underlying ASN1 function/structure to
4653 be accessed transparently. As a result code should not use ASN1_ITEM
4654 references directly (such as &X509_it) but instead use the relevant
4655 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4656 use of the new ASN1 code on platforms where exporting structures
4657 is problematical (for example in shared libraries) but exporting
4658 functions returning pointers to structures is not.
4659 [Steve Henson]
4660
4661 *) Add support for overriding the generation of SSL/TLS session IDs.
4662 These callbacks can be registered either in an SSL_CTX or per SSL.
4663 The purpose of this is to allow applications to control, if they wish,
4664 the arbitrary values chosen for use as session IDs, particularly as it
4665 can be useful for session caching in multiple-server environments. A
4666 command-line switch for testing this (and any client code that wishes
4667 to use such a feature) has been added to "s_server".
4668 [Geoff Thorpe, Lutz Jaenicke]
4669
4670 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4671 of the form '#if defined(...) || defined(...) || ...' and
4672 '#if !defined(...) && !defined(...) && ...'. This also avoids
4673 the growing number of special cases it was previously handling.
4674 [Richard Levitte]
4675
4676 *) Make all configuration macros available for application by making
4677 sure they are available in opensslconf.h, by giving them names starting
4678 with "OPENSSL_" to avoid conflicts with other packages and by making
4679 sure e_os2.h will cover all platform-specific cases together with
4680 opensslconf.h.
4681 Additionally, it is now possible to define configuration/platform-
4682 specific names (called "system identities"). In the C code, these
4683 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4684 macro with the name beginning with "OPENSSL_SYS_", which is determined
4685 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4686 what is available.
4687 [Richard Levitte]
4688
4689 *) New option -set_serial to 'req' and 'x509' this allows the serial
4690 number to use to be specified on the command line. Previously self
4691 signed certificates were hard coded with serial number 0 and the
4692 CA options of 'x509' had to use a serial number in a file which was
4693 auto incremented.
4694 [Steve Henson]
4695
4696 *) New options to 'ca' utility to support V2 CRL entry extensions.
4697 Currently CRL reason, invalidity date and hold instruction are
4698 supported. Add new CRL extensions to V3 code and some new objects.
4699 [Steve Henson]
4700
4701 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4702 disable standard block padding (aka PKCS#5 padding) in the EVP
4703 API, which was previously mandatory. This means that the data is
4704 not padded in any way and so the total length much be a multiple
4705 of the block size, otherwise an error occurs.
4706 [Steve Henson]
4707
4708 *) Initial (incomplete) OCSP SSL support.
4709 [Steve Henson]
4710
4711 *) New function OCSP_parse_url(). This splits up a URL into its host,
4712 port and path components: primarily to parse OCSP URLs. New -url
4713 option to ocsp utility.
4714 [Steve Henson]
4715
4716 *) New nonce behavior. The return value of OCSP_check_nonce() now
4717 reflects the various checks performed. Applications can decide
4718 whether to tolerate certain situations such as an absent nonce
4719 in a response when one was present in a request: the ocsp application
4720 just prints out a warning. New function OCSP_add1_basic_nonce()
4721 this is to allow responders to include a nonce in a response even if
4722 the request is nonce-less.
4723 [Steve Henson]
4724
4725 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4726 skipped when using openssl x509 multiple times on a single input file,
4727 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4728 [Bodo Moeller]
4729
4730 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4731 set string type: to handle setting ASN1_TIME structures. Fix ca
4732 utility to correctly initialize revocation date of CRLs.
4733 [Steve Henson]
4734
4735 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4736 the clients preferred ciphersuites and rather use its own preferences.
4737 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4738 Internet Explorer by ensuring unchanged hash method during stepup.
4739 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4740 [Lutz Jaenicke]
4741
4742 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4743 to aes and add a new 'exist' option to print out symbols that don't
4744 appear to exist.
4745 [Steve Henson]
4746
4747 *) Additional options to ocsp utility to allow flags to be set and
4748 additional certificates supplied.
4749 [Steve Henson]
4750
4751 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4752 OCSP client a number of certificate to only verify the response
4753 signature against.
4754 [Richard Levitte]
4755
4756 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4757 handle the new API. Currently only ECB, CBC modes supported. Add new
4758 AES OIDs.
4759
4760 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4761 Encryption Standard (AES) Ciphersuites for Transport Layer
4762 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4763 not enabled by default and were not part of the "ALL" ciphersuite
4764 alias because they were not yet official; they could be
4765 explicitly requested by specifying the "AESdraft" ciphersuite
4766 group alias. In the final release of OpenSSL 0.9.7, the group
4767 alias is called "AES" and is part of "ALL".)
4768 [Ben Laurie, Steve Henson, Bodo Moeller]
4769
4770 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4771 request to response.
4772 [Steve Henson]
4773
4774 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4775 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4776 extract information from a certificate request. OCSP_response_create()
4777 creates a response and optionally adds a basic response structure.
4778 OCSP_basic_add1_status() adds a complete single response to a basic
4779 response and returns the OCSP_SINGLERESP structure just added (to allow
4780 extensions to be included for example). OCSP_basic_add1_cert() adds a
4781 certificate to a basic response and OCSP_basic_sign() signs a basic
4782 response with various flags. New helper functions ASN1_TIME_check()
4783 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4784 (converts ASN1_TIME to GeneralizedTime).
4785 [Steve Henson]
4786
4787 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4788 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4789 structure from a certificate. X509_pubkey_digest() digests the public_key
4790 contents: this is used in various key identifiers.
4791 [Steve Henson]
4792
4793 *) Make sk_sort() tolerate a NULL argument.
4794 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4795
4796 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4797 passed by the function are trusted implicitly. If any of them signed the
4798 response then it is assumed to be valid and is not verified.
4799 [Steve Henson]
4800
4801 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4802 to data. This was previously part of the PKCS7 ASN1 code. This
4803 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4804 [Steve Henson, reported by Kenneth R. Robinette
4805 <support@securenetterm.com>]
4806
4807 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4808 routines: without these tracing memory leaks is very painful.
4809 Fix leaks in PKCS12 and PKCS7 routines.
4810 [Steve Henson]
4811
4812 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4813 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4814 effectively meant GeneralizedTime would never be used. Now it
4815 is initialised to -1 but X509_time_adj() now has to check the value
4816 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4817 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4818 [Steve Henson, reported by Kenneth R. Robinette
4819 <support@securenetterm.com>]
4820
4821 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4822 result in a zero length in the ASN1_INTEGER structure which was
4823 not consistent with the structure when d2i_ASN1_INTEGER() was used
4824 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4825 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4826 where it did not print out a minus for negative ASN1_INTEGER.
4827 [Steve Henson]
4828
4829 *) Add summary printout to ocsp utility. The various functions which
4830 convert status values to strings have been renamed to:
4831 OCSP_response_status_str(), OCSP_cert_status_str() and
4832 OCSP_crl_reason_str() and are no longer static. New options
4833 to verify nonce values and to disable verification. OCSP response
4834 printout format cleaned up.
4835 [Steve Henson]
4836
4837 *) Add additional OCSP certificate checks. These are those specified
4838 in RFC2560. This consists of two separate checks: the CA of the
4839 certificate being checked must either be the OCSP signer certificate
4840 or the issuer of the OCSP signer certificate. In the latter case the
4841 OCSP signer certificate must contain the OCSP signing extended key
4842 usage. This check is performed by attempting to match the OCSP
4843 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4844 in the OCSP_CERTID structures of the response.
4845 [Steve Henson]
4846
4847 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4848 and related routines. This uses the standard OpenSSL certificate
4849 verify routines to perform initial checks (just CA validity) and
4850 to obtain the certificate chain. Then additional checks will be
4851 performed on the chain. Currently the root CA is checked to see
4852 if it is explicitly trusted for OCSP signing. This is used to set
4853 a root CA as a global signing root: that is any certificate that
4854 chains to that CA is an acceptable OCSP signing certificate.
4855 [Steve Henson]
4856
4857 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4858 extensions from a separate configuration file.
4859 As when reading extensions from the main configuration file,
4860 the '-extensions ...' option may be used for specifying the
4861 section to use.
4862 [Massimiliano Pala <madwolf@comune.modena.it>]
4863
4864 *) New OCSP utility. Allows OCSP requests to be generated or
4865 read. The request can be sent to a responder and the output
4866 parsed, outputed or printed in text form. Not complete yet:
4867 still needs to check the OCSP response validity.
4868 [Steve Henson]
4869
4870 *) New subcommands for 'openssl ca':
4871 'openssl ca -status <serial>' prints the status of the cert with
4872 the given serial number (according to the index file).
4873 'openssl ca -updatedb' updates the expiry status of certificates
4874 in the index file.
4875 [Massimiliano Pala <madwolf@comune.modena.it>]
4876
4877 *) New '-newreq-nodes' command option to CA.pl. This is like
4878 '-newreq', but calls 'openssl req' with the '-nodes' option
4879 so that the resulting key is not encrypted.
4880 [Damien Miller <djm@mindrot.org>]
4881
4882 *) New configuration for the GNU Hurd.
4883 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4884
4885 *) Initial code to implement OCSP basic response verify. This
4886 is currently incomplete. Currently just finds the signer's
4887 certificate and verifies the signature on the response.
4888 [Steve Henson]
4889
4890 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4891 value of OPENSSLDIR. This is available via the new '-d' option
4892 to 'openssl version', and is also included in 'openssl version -a'.
4893 [Bodo Moeller]
4894
4895 *) Allowing defining memory allocation callbacks that will be given
4896 file name and line number information in additional arguments
4897 (a const char* and an int). The basic functionality remains, as
4898 well as the original possibility to just replace malloc(),
4899 realloc() and free() by functions that do not know about these
4900 additional arguments. To register and find out the current
4901 settings for extended allocation functions, the following
4902 functions are provided:
4903
4904 CRYPTO_set_mem_ex_functions
4905 CRYPTO_set_locked_mem_ex_functions
4906 CRYPTO_get_mem_ex_functions
4907 CRYPTO_get_locked_mem_ex_functions
4908
4909 These work the same way as CRYPTO_set_mem_functions and friends.
4910 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4911 extended allocation function is enabled.
4912 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4913 a conventional allocation function is enabled.
4914 [Richard Levitte, Bodo Moeller]
4915
4916 *) Finish off removing the remaining LHASH function pointer casts.
4917 There should no longer be any prototype-casting required when using
4918 the LHASH abstraction, and any casts that remain are "bugs". See
4919 the callback types and macros at the head of lhash.h for details
4920 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4921 [Geoff Thorpe]
4922
4923 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4924 If /dev/[u]random devices are not available or do not return enough
4925 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4926 be queried.
4927 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4928 /etc/entropy will be queried once each in this sequence, quering stops
4929 when enough entropy was collected without querying more sockets.
4930 [Lutz Jaenicke]
4931
4932 *) Change the Unix RAND_poll() variant to be able to poll several
4933 random devices, as specified by DEVRANDOM, until a sufficient amount
4934 of data has been collected. We spend at most 10 ms on each file
4935 (select timeout) and read in non-blocking mode. DEVRANDOM now
4936 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4937 (previously it was just the string "/dev/urandom"), so on typical
4938 platforms the 10 ms delay will never occur.
4939 Also separate out the Unix variant to its own file, rand_unix.c.
4940 For VMS, there's a currently-empty rand_vms.c.
4941 [Richard Levitte]
4942
4943 *) Move OCSP client related routines to ocsp_cl.c. These
4944 provide utility functions which an application needing
4945 to issue a request to an OCSP responder and analyse the
4946 response will typically need: as opposed to those which an
4947 OCSP responder itself would need which will be added later.
4948
4949 OCSP_request_sign() signs an OCSP request with an API similar
4950 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4951 response. OCSP_response_get1_basic() extracts basic response
4952 from response. OCSP_resp_find_status(): finds and extracts status
4953 information from an OCSP_CERTID structure (which will be created
4954 when the request structure is built). These are built from lower
4955 level functions which work on OCSP_SINGLERESP structures but
4956 wont normally be used unless the application wishes to examine
4957 extensions in the OCSP response for example.
4958
4959 Replace nonce routines with a pair of functions.
4960 OCSP_request_add1_nonce() adds a nonce value and optionally
4961 generates a random value. OCSP_check_nonce() checks the
4962 validity of the nonce in an OCSP response.
4963 [Steve Henson]
4964
4965 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4966 This doesn't copy the supplied OCSP_CERTID and avoids the
4967 need to free up the newly created id. Change return type
4968 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4969 This can then be used to add extensions to the request.
4970 Deleted OCSP_request_new(), since most of its functionality
4971 is now in OCSP_REQUEST_new() (and the case insensitive name
4972 clash) apart from the ability to set the request name which
4973 will be added elsewhere.
4974 [Steve Henson]
4975
4976 *) Update OCSP API. Remove obsolete extensions argument from
4977 various functions. Extensions are now handled using the new
4978 OCSP extension code. New simple OCSP HTTP function which
4979 can be used to send requests and parse the response.
4980 [Steve Henson]
4981
4982 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4983 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4984 uses the special reorder version of SET OF to sort the attributes
4985 and reorder them to match the encoded order. This resolves a long
4986 standing problem: a verify on a PKCS7 structure just after signing
4987 it used to fail because the attribute order did not match the
4988 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4989 it uses the received order. This is necessary to tolerate some broken
4990 software that does not order SET OF. This is handled by encoding
4991 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4992 to produce the required SET OF.
4993 [Steve Henson]
4994
4995 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4996 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4997 files to get correct declarations of the ASN.1 item variables.
4998 [Richard Levitte]
4999
5000 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5001 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5002 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5003 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5004 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5005 ASN1_ITEM and no wrapper functions.
5006 [Steve Henson]
5007
5008 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5009 replace the old function pointer based I/O routines. Change most of
5010 the *_d2i_bio() and *_d2i_fp() functions to use these.
5011 [Steve Henson]
5012
5013 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5014 lines, recognice more "algorithms" that can be deselected, and make
5015 it complain about algorithm deselection that isn't recognised.
5016 [Richard Levitte]
5017
5018 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5019 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5020 to use new functions. Add NO_ASN1_OLD which can be set to remove
5021 some old style ASN1 functions: this can be used to determine if old
5022 code will still work when these eventually go away.
5023 [Steve Henson]
5024
5025 *) New extension functions for OCSP structures, these follow the
5026 same conventions as certificates and CRLs.
5027 [Steve Henson]
5028
5029 *) New function X509V3_add1_i2d(). This automatically encodes and
5030 adds an extension. Its behaviour can be customised with various
5031 flags to append, replace or delete. Various wrappers added for
5032 certifcates and CRLs.
5033 [Steve Henson]
5034
5035 *) Fix to avoid calling the underlying ASN1 print routine when
5036 an extension cannot be parsed. Correct a typo in the
5037 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5038 [Steve Henson]
5039
5040 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5041 entries for variables.
5042 [Steve Henson]
5043
5044 *) Add functionality to apps/openssl.c for detecting locking
5045 problems: As the program is single-threaded, all we have
5046 to do is register a locking callback using an array for
5047 storing which locks are currently held by the program.
5048 [Bodo Moeller]
5049
5050 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5051 SSL_get_ex_data_X509_STORE_idx(), which is used in
5052 ssl_verify_cert_chain() and thus can be called at any time
5053 during TLS/SSL handshakes so that thread-safety is essential.
5054 Unfortunately, the ex_data design is not at all suited
5055 for multi-threaded use, so it probably should be abolished.
5056 [Bodo Moeller]
5057
5058 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5059 [Broadcom, tweaked and integrated by Geoff Thorpe]
5060
5061 *) Move common extension printing code to new function
5062 X509V3_print_extensions(). Reorganise OCSP print routines and
5063 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5064 [Steve Henson]
5065
5066 *) New function X509_signature_print() to remove duplication in some
5067 print routines.
5068 [Steve Henson]
5069
5070 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5071 set (this was treated exactly the same as SET OF previously). This
5072 is used to reorder the STACK representing the structure to match the
5073 encoding. This will be used to get round a problem where a PKCS7
5074 structure which was signed could not be verified because the STACK
5075 order did not reflect the encoded order.
5076 [Steve Henson]
5077
5078 *) Reimplement the OCSP ASN1 module using the new code.
5079 [Steve Henson]
5080
5081 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5082 for its ASN1 operations. The old style function pointers still exist
5083 for now but they will eventually go away.
5084 [Steve Henson]
5085
5086 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5087 completely replaces the old ASN1 functionality with a table driven
5088 encoder and decoder which interprets an ASN1_ITEM structure describing
5089 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5090 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5091 has also been converted to the new form.
5092 [Steve Henson]
5093
5094 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5095 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5096 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5097 for negative moduli.
5098 [Bodo Moeller]
5099
5100 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5101 of not touching the result's sign bit.
5102 [Bodo Moeller]
5103
5104 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5105 set.
5106 [Bodo Moeller]
5107
5108 *) Changed the LHASH code to use prototypes for callbacks, and created
5109 macros to declare and implement thin (optionally static) functions
5110 that provide type-safety and avoid function pointer casting for the
5111 type-specific callbacks.
5112 [Geoff Thorpe]
5113
5114 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5115 RFC 2712.
5116 [Veers Staats <staatsvr@asc.hpc.mil>,
5117 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5118
5119 *) Reformat the FAQ so the different questions and answers can be divided
5120 in sections depending on the subject.
5121 [Richard Levitte]
5122
5123 *) Have the zlib compression code load ZLIB.DLL dynamically under
5124 Windows.
5125 [Richard Levitte]
5126
5127 *) New function BN_mod_sqrt for computing square roots modulo a prime
5128 (using the probabilistic Tonelli-Shanks algorithm unless
5129 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5130 be handled deterministically).
5131 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5132
5133 *) Make BN_mod_inverse faster by explicitly handling small quotients
5134 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5135 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5136 [Bodo Moeller]
5137
5138 *) New function BN_kronecker.
5139 [Bodo Moeller]
5140
5141 *) Fix BN_gcd so that it works on negative inputs; the result is
5142 positive unless both parameters are zero.
5143 Previously something reasonably close to an infinite loop was
5144 possible because numbers could be growing instead of shrinking
5145 in the implementation of Euclid's algorithm.
5146 [Bodo Moeller]
5147
5148 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5149 sign of the number in question.
5150
5151 Fix BN_is_word(a,w) to work correctly for w == 0.
5152
5153 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5154 because its test if the absolute value of 'a' equals 'w'.
5155 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5156 it exists mostly for use in the implementations of BN_is_zero(),
5157 BN_is_one(), and BN_is_word().
5158 [Bodo Moeller]
5159
5160 *) New function BN_swap.
5161 [Bodo Moeller]
5162
5163 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5164 the exponentiation functions are more likely to produce reasonable
5165 results on negative inputs.
5166 [Bodo Moeller]
5167
5168 *) Change BN_mod_mul so that the result is always non-negative.
5169 Previously, it could be negative if one of the factors was negative;
5170 I don't think anyone really wanted that behaviour.
5171 [Bodo Moeller]
5172
5173 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5174 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5175 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5176 and add new functions:
5177
5178 BN_nnmod
5179 BN_mod_sqr
5180 BN_mod_add
5181 BN_mod_add_quick
5182 BN_mod_sub
5183 BN_mod_sub_quick
5184 BN_mod_lshift1
5185 BN_mod_lshift1_quick
5186 BN_mod_lshift
5187 BN_mod_lshift_quick
5188
5189 These functions always generate non-negative results.
5190
5191 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5192 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5193
5194 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5195 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5196 be reduced modulo m.
5197 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5198
5199 #if 0
5200 The following entry accidentily appeared in the CHANGES file
5201 distributed with OpenSSL 0.9.7. The modifications described in
5202 it do *not* apply to OpenSSL 0.9.7.
5203
5204 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5205 was actually never needed) and in BN_mul(). The removal in BN_mul()
5206 required a small change in bn_mul_part_recursive() and the addition
5207 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5208 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5209 bn_sub_words() and bn_add_words() except they take arrays with
5210 differing sizes.
5211 [Richard Levitte]
5212 #endif
5213
5214 *) In 'openssl passwd', verify passwords read from the terminal
5215 unless the '-salt' option is used (which usually means that
5216 verification would just waste user's time since the resulting
5217 hash is going to be compared with some given password hash)
5218 or the new '-noverify' option is used.
5219
5220 This is an incompatible change, but it does not affect
5221 non-interactive use of 'openssl passwd' (passwords on the command
5222 line, '-stdin' option, '-in ...' option) and thus should not
5223 cause any problems.
5224 [Bodo Moeller]
5225
5226 *) Remove all references to RSAref, since there's no more need for it.
5227 [Richard Levitte]
5228
5229 *) Make DSO load along a path given through an environment variable
5230 (SHLIB_PATH) with shl_load().
5231 [Richard Levitte]
5232
5233 *) Constify the ENGINE code as a result of BIGNUM constification.
5234 Also constify the RSA code and most things related to it. In a
5235 few places, most notable in the depth of the ASN.1 code, ugly
5236 casts back to non-const were required (to be solved at a later
5237 time)
5238 [Richard Levitte]
5239
5240 *) Make it so the openssl application has all engines loaded by default.
5241 [Richard Levitte]
5242
5243 *) Constify the BIGNUM routines a little more.
5244 [Richard Levitte]
5245
5246 *) Add the following functions:
5247
5248 ENGINE_load_cswift()
5249 ENGINE_load_chil()
5250 ENGINE_load_atalla()
5251 ENGINE_load_nuron()
5252 ENGINE_load_builtin_engines()
5253
5254 That way, an application can itself choose if external engines that
5255 are built-in in OpenSSL shall ever be used or not. The benefit is
5256 that applications won't have to be linked with libdl or other dso
5257 libraries unless it's really needed.
5258
5259 Changed 'openssl engine' to load all engines on demand.
5260 Changed the engine header files to avoid the duplication of some
5261 declarations (they differed!).
5262 [Richard Levitte]
5263
5264 *) 'openssl engine' can now list capabilities.
5265 [Richard Levitte]
5266
5267 *) Better error reporting in 'openssl engine'.
5268 [Richard Levitte]
5269
5270 *) Never call load_dh_param(NULL) in s_server.
5271 [Bodo Moeller]
5272
5273 *) Add engine application. It can currently list engines by name and
5274 identity, and test if they are actually available.
5275 [Richard Levitte]
5276
5277 *) Improve RPM specification file by forcing symbolic linking and making
5278 sure the installed documentation is also owned by root.root.
5279 [Damien Miller <djm@mindrot.org>]
5280
5281 *) Give the OpenSSL applications more possibilities to make use of
5282 keys (public as well as private) handled by engines.
5283 [Richard Levitte]
5284
5285 *) Add OCSP code that comes from CertCo.
5286 [Richard Levitte]
5287
5288 *) Add VMS support for the Rijndael code.
5289 [Richard Levitte]
5290
5291 *) Added untested support for Nuron crypto accelerator.
5292 [Ben Laurie]
5293
5294 *) Add support for external cryptographic devices. This code was
5295 previously distributed separately as the "engine" branch.
5296 [Geoff Thorpe, Richard Levitte]
5297
5298 *) Rework the filename-translation in the DSO code. It is now possible to
5299 have far greater control over how a "name" is turned into a filename
5300 depending on the operating environment and any oddities about the
5301 different shared library filenames on each system.
5302 [Geoff Thorpe]
5303
5304 *) Support threads on FreeBSD-elf in Configure.
5305 [Richard Levitte]
5306
5307 *) Fix for SHA1 assembly problem with MASM: it produces
5308 warnings about corrupt line number information when assembling
5309 with debugging information. This is caused by the overlapping
5310 of two sections.
5311 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5312
5313 *) NCONF changes.
5314 NCONF_get_number() has no error checking at all. As a replacement,
5315 NCONF_get_number_e() is defined (_e for "error checking") and is
5316 promoted strongly. The old NCONF_get_number is kept around for
5317 binary backward compatibility.
5318 Make it possible for methods to load from something other than a BIO,
5319 by providing a function pointer that is given a name instead of a BIO.
5320 For example, this could be used to load configuration data from an
5321 LDAP server.
5322 [Richard Levitte]
5323
5324 *) Fix for non blocking accept BIOs. Added new I/O special reason
5325 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5326 with non blocking I/O was not possible because no retry code was
5327 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5328 this case.
5329 [Steve Henson]
5330
5331 *) Added the beginnings of Rijndael support.
5332 [Ben Laurie]
5333
5334 *) Fix for bug in DirectoryString mask setting. Add support for
5335 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5336 to allow certificate printing to more controllable, additional
5337 'certopt' option to 'x509' to allow new printing options to be
5338 set.
5339 [Steve Henson]
5340
5341 *) Clean old EAY MD5 hack from e_os.h.
5342 [Richard Levitte]
5343
5344 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5345
5346 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5347 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5348 [Joe Orton, Steve Henson]
5349
5350 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5351
5352 *) Fix additional bug revealed by the NISCC test suite:
5353
5354 Stop bug triggering large recursion when presented with
5355 certain ASN.1 tags (CVE-2003-0851)
5356 [Steve Henson]
5357
5358 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5359
5360 *) Fix various bugs revealed by running the NISCC test suite:
5361
5362 Stop out of bounds reads in the ASN1 code when presented with
5363 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5364
5365 If verify callback ignores invalid public key errors don't try to check
5366 certificate signature with the NULL public key.
5367
5368 [Steve Henson]
5369
5370 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5371 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5372 specifications.
5373 [Steve Henson]
5374
5375 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5376 extra data after the compression methods not only for TLS 1.0
5377 but also for SSL 3.0 (as required by the specification).
5378 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5379
5380 *) Change X509_certificate_type() to mark the key as exported/exportable
5381 when it's 512 *bits* long, not 512 bytes.
5382 [Richard Levitte]
5383
5384 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5385
5386 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5387 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5388 a protocol version number mismatch like a decryption error
5389 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5390 [Bodo Moeller]
5391
5392 *) Turn on RSA blinding by default in the default implementation
5393 to avoid a timing attack. Applications that don't want it can call
5394 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5395 They would be ill-advised to do so in most cases.
5396 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5397
5398 *) Change RSA blinding code so that it works when the PRNG is not
5399 seeded (in this case, the secret RSA exponent is abused as
5400 an unpredictable seed -- if it is not unpredictable, there
5401 is no point in blinding anyway). Make RSA blinding thread-safe
5402 by remembering the creator's thread ID in rsa->blinding and
5403 having all other threads use local one-time blinding factors
5404 (this requires more computation than sharing rsa->blinding, but
5405 avoids excessive locking; and if an RSA object is not shared
5406 between threads, blinding will still be very fast).
5407 [Bodo Moeller]
5408
5409 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5410
5411 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5412 via timing by performing a MAC computation even if incorrrect
5413 block cipher padding has been found. This is a countermeasure
5414 against active attacks where the attacker has to distinguish
5415 between bad padding and a MAC verification error. (CVE-2003-0078)
5416
5417 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5418 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5419 Martin Vuagnoux (EPFL, Ilion)]
5420
5421 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5422
5423 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5424 memory from it's contents. This is done with a counter that will
5425 place alternating values in each byte. This can be used to solve
5426 two issues: 1) the removal of calls to memset() by highly optimizing
5427 compilers, and 2) cleansing with other values than 0, since those can
5428 be read through on certain media, for example a swap space on disk.
5429 [Geoff Thorpe]
5430
5431 *) Bugfix: client side session caching did not work with external caching,
5432 because the session->cipher setting was not restored when reloading
5433 from the external cache. This problem was masked, when
5434 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5435 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5436 [Lutz Jaenicke]
5437
5438 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5439 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5440 [Zeev Lieber <zeev-l@yahoo.com>]
5441
5442 *) Undo an undocumented change introduced in 0.9.6e which caused
5443 repeated calls to OpenSSL_add_all_ciphers() and
5444 OpenSSL_add_all_digests() to be ignored, even after calling
5445 EVP_cleanup().
5446 [Richard Levitte]
5447
5448 *) Change the default configuration reader to deal with last line not
5449 being properly terminated.
5450 [Richard Levitte]
5451
5452 *) Change X509_NAME_cmp() so it applies the special rules on handling
5453 DN values that are of type PrintableString, as well as RDNs of type
5454 emailAddress where the value has the type ia5String.
5455 [stefank@valicert.com via Richard Levitte]
5456
5457 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5458 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5459 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5460 the bitwise-OR of the two for use by the majority of applications
5461 wanting this behaviour, and update the docs. The documented
5462 behaviour and actual behaviour were inconsistent and had been
5463 changing anyway, so this is more a bug-fix than a behavioural
5464 change.
5465 [Geoff Thorpe, diagnosed by Nadav Har'El]
5466
5467 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5468 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5469 [Bodo Moeller]
5470
5471 *) Fix initialization code race conditions in
5472 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5473 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5474 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5475 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5476 ssl2_get_cipher_by_char(),
5477 ssl3_get_cipher_by_char().
5478 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5479
5480 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5481 the cached sessions are flushed, as the remove_cb() might use ex_data
5482 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5483 (see [openssl.org #212]).
5484 [Geoff Thorpe, Lutz Jaenicke]
5485
5486 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5487 length, instead of the encoding length to d2i_ASN1_OBJECT.
5488 [Steve Henson]
5489
5490 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5491
5492 *) [In 0.9.6g-engine release:]
5493 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5494 [Lynn Gazis <lgazis@rainbow.com>]
5495
5496 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5497
5498 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5499 and get fix the header length calculation.
5500 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5501 Alon Kantor <alonk@checkpoint.com> (and others),
5502 Steve Henson]
5503
5504 *) Use proper error handling instead of 'assertions' in buffer
5505 overflow checks added in 0.9.6e. This prevents DoS (the
5506 assertions could call abort()).
5507 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5508
5509 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5510
5511 *) Add various sanity checks to asn1_get_length() to reject
5512 the ASN1 length bytes if they exceed sizeof(long), will appear
5513 negative or the content length exceeds the length of the
5514 supplied buffer.
5515 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5516
5517 *) Fix cipher selection routines: ciphers without encryption had no flags
5518 for the cipher strength set and where therefore not handled correctly
5519 by the selection routines (PR #130).
5520 [Lutz Jaenicke]
5521
5522 *) Fix EVP_dsa_sha macro.
5523 [Nils Larsch]
5524
5525 *) New option
5526 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5527 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5528 that was added in OpenSSL 0.9.6d.
5529
5530 As the countermeasure turned out to be incompatible with some
5531 broken SSL implementations, the new option is part of SSL_OP_ALL.
5532 SSL_OP_ALL is usually employed when compatibility with weird SSL
5533 implementations is desired (e.g. '-bugs' option to 's_client' and
5534 's_server'), so the new option is automatically set in many
5535 applications.
5536 [Bodo Moeller]
5537
5538 *) Changes in security patch:
5539
5540 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5541 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5542 Air Force Materiel Command, USAF, under agreement number
5543 F30602-01-2-0537.
5544
5545 *) Add various sanity checks to asn1_get_length() to reject
5546 the ASN1 length bytes if they exceed sizeof(long), will appear
5547 negative or the content length exceeds the length of the
5548 supplied buffer. (CVE-2002-0659)
5549 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5550
5551 *) Assertions for various potential buffer overflows, not known to
5552 happen in practice.
5553 [Ben Laurie (CHATS)]
5554
5555 *) Various temporary buffers to hold ASCII versions of integers were
5556 too small for 64 bit platforms. (CVE-2002-0655)
5557 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5558
5559 *) Remote buffer overflow in SSL3 protocol - an attacker could
5560 supply an oversized session ID to a client. (CVE-2002-0656)
5561 [Ben Laurie (CHATS)]
5562
5563 *) Remote buffer overflow in SSL2 protocol - an attacker could
5564 supply an oversized client master key. (CVE-2002-0656)
5565 [Ben Laurie (CHATS)]
5566
5567 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5568
5569 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5570 encoded as NULL) with id-dsa-with-sha1.
5571 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5572
5573 *) Check various X509_...() return values in apps/req.c.
5574 [Nils Larsch <nla@trustcenter.de>]
5575
5576 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5577 an end-of-file condition would erronously be flagged, when the CRLF
5578 was just at the end of a processed block. The bug was discovered when
5579 processing data through a buffering memory BIO handing the data to a
5580 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5581 <ptsekov@syntrex.com> and Nedelcho Stanev.
5582 [Lutz Jaenicke]
5583
5584 *) Implement a countermeasure against a vulnerability recently found
5585 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5586 before application data chunks to avoid the use of known IVs
5587 with data potentially chosen by the attacker.
5588 [Bodo Moeller]
5589
5590 *) Fix length checks in ssl3_get_client_hello().
5591 [Bodo Moeller]
5592
5593 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5594 to prevent ssl3_read_internal() from incorrectly assuming that
5595 ssl3_read_bytes() found application data while handshake
5596 processing was enabled when in fact s->s3->in_read_app_data was
5597 merely automatically cleared during the initial handshake.
5598 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5599
5600 *) Fix object definitions for Private and Enterprise: they were not
5601 recognized in their shortname (=lowercase) representation. Extend
5602 obj_dat.pl to issue an error when using undefined keywords instead
5603 of silently ignoring the problem (Svenning Sorensen
5604 <sss@sss.dnsalias.net>).
5605 [Lutz Jaenicke]
5606
5607 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5608 generators, i.e. generators other than 2 and 5. (Previously, the
5609 code did not properly initialise the 'add' and 'rem' values to
5610 BN_generate_prime().)
5611
5612 In the new general case, we do not insist that 'generator' is
5613 actually a primitive root: This requirement is rather pointless;
5614 a generator of the order-q subgroup is just as good, if not
5615 better.
5616 [Bodo Moeller]
5617
5618 *) Map new X509 verification errors to alerts. Discovered and submitted by
5619 Tom Wu <tom@arcot.com>.
5620 [Lutz Jaenicke]
5621
5622 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5623 returning non-zero before the data has been completely received
5624 when using non-blocking I/O.
5625 [Bodo Moeller; problem pointed out by John Hughes]
5626
5627 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5628 [Ben Laurie, Lutz Jaenicke]
5629
5630 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5631 Yoram Zahavi <YoramZ@gilian.com>).
5632 [Lutz Jaenicke]
5633
5634 *) Add information about CygWin 1.3 and on, and preserve proper
5635 configuration for the versions before that.
5636 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5637
5638 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5639 check whether we deal with a copy of a session and do not delete from
5640 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5641 <izhar@checkpoint.com>.
5642 [Lutz Jaenicke]
5643
5644 *) Do not store session data into the internal session cache, if it
5645 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5646 flag is set). Proposed by Aslam <aslam@funk.com>.
5647 [Lutz Jaenicke]
5648
5649 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5650 value is 0.
5651 [Richard Levitte]
5652
5653 *) [In 0.9.6d-engine release:]
5654 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5655 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5656
5657 *) Add the configuration target linux-s390x.
5658 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5659
5660 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5661 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5662 variable as an indication that a ClientHello message has been
5663 received. As the flag value will be lost between multiple
5664 invocations of ssl3_accept when using non-blocking I/O, the
5665 function may not be aware that a handshake has actually taken
5666 place, thus preventing a new session from being added to the
5667 session cache.
5668
5669 To avoid this problem, we now set s->new_session to 2 instead of
5670 using a local variable.
5671 [Lutz Jaenicke, Bodo Moeller]
5672
5673 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5674 if the SSL_R_LENGTH_MISMATCH error is detected.
5675 [Geoff Thorpe, Bodo Moeller]
5676
5677 *) New 'shared_ldflag' column in Configure platform table.
5678 [Richard Levitte]
5679
5680 *) Fix EVP_CIPHER_mode macro.
5681 ["Dan S. Camper" <dan@bti.net>]
5682
5683 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5684 type, we must throw them away by setting rr->length to 0.
5685 [D P Chang <dpc@qualys.com>]
5686
5687 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5688
5689 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5690 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5691 worked incorrectly for those cases where range = 10..._2 and
5692 3*range is two bits longer than range.)
5693 [Bodo Moeller]
5694
5695 *) Only add signing time to PKCS7 structures if it is not already
5696 present.
5697 [Steve Henson]
5698
5699 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5700 OBJ_ld_ce should be OBJ_id_ce.
5701 Also some ip-pda OIDs in crypto/objects/objects.txt were
5702 incorrect (cf. RFC 3039).
5703 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5704
5705 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5706 returns early because it has nothing to do.
5707 [Andy Schneider <andy.schneider@bjss.co.uk>]
5708
5709 *) [In 0.9.6c-engine release:]
5710 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5711 [Andy Schneider <andy.schneider@bjss.co.uk>]
5712
5713 *) [In 0.9.6c-engine release:]
5714 Add support for Cryptographic Appliance's keyserver technology.
5715 (Use engine 'keyclient')
5716 [Cryptographic Appliances and Geoff Thorpe]
5717
5718 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5719 is called via tools/c89.sh because arguments have to be
5720 rearranged (all '-L' options must appear before the first object
5721 modules).
5722 [Richard Shapiro <rshapiro@abinitio.com>]
5723
5724 *) [In 0.9.6c-engine release:]
5725 Add support for Broadcom crypto accelerator cards, backported
5726 from 0.9.7.
5727 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5728
5729 *) [In 0.9.6c-engine release:]
5730 Add support for SureWare crypto accelerator cards from
5731 Baltimore Technologies. (Use engine 'sureware')
5732 [Baltimore Technologies and Mark Cox]
5733
5734 *) [In 0.9.6c-engine release:]
5735 Add support for crypto accelerator cards from Accelerated
5736 Encryption Processing, www.aep.ie. (Use engine 'aep')
5737 [AEP Inc. and Mark Cox]
5738
5739 *) Add a configuration entry for gcc on UnixWare.
5740 [Gary Benson <gbenson@redhat.com>]
5741
5742 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5743 messages are stored in a single piece (fixed-length part and
5744 variable-length part combined) and fix various bugs found on the way.
5745 [Bodo Moeller]
5746
5747 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5748 instead. BIO_gethostbyname() does not know what timeouts are
5749 appropriate, so entries would stay in cache even when they have
5750 become invalid.
5751 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5752
5753 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5754 faced with a pathologically small ClientHello fragment that does
5755 not contain client_version: Instead of aborting with an error,
5756 simply choose the highest available protocol version (i.e.,
5757 TLS 1.0 unless it is disabled). In practice, ClientHello
5758 messages are never sent like this, but this change gives us
5759 strictly correct behaviour at least for TLS.
5760 [Bodo Moeller]
5761
5762 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5763 never resets s->method to s->ctx->method when called from within
5764 one of the SSL handshake functions.
5765 [Bodo Moeller; problem pointed out by Niko Baric]
5766
5767 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5768 (sent using the client's version number) if client_version is
5769 smaller than the protocol version in use. Also change
5770 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5771 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5772 the client will at least see that alert.
5773 [Bodo Moeller]
5774
5775 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5776 correctly.
5777 [Bodo Moeller]
5778
5779 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5780 client receives HelloRequest while in a handshake.
5781 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5782
5783 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5784 should end in 'break', not 'goto end' which circuments various
5785 cleanups done in state SSL_ST_OK. But session related stuff
5786 must be disabled for SSL_ST_OK in the case that we just sent a
5787 HelloRequest.
5788
5789 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5790 before just sending a HelloRequest.
5791 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5792
5793 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5794 reveal whether illegal block cipher padding was found or a MAC
5795 verification error occured. (Neither SSLerr() codes nor alerts
5796 are directly visible to potential attackers, but the information
5797 may leak via logfiles.)
5798
5799 Similar changes are not required for the SSL 2.0 implementation
5800 because the number of padding bytes is sent in clear for SSL 2.0,
5801 and the extra bytes are just ignored. However ssl/s2_pkt.c
5802 failed to verify that the purported number of padding bytes is in
5803 the legal range.
5804 [Bodo Moeller]
5805
5806 *) Add OpenUNIX-8 support including shared libraries
5807 (Boyd Lynn Gerber <gerberb@zenez.com>).
5808 [Lutz Jaenicke]
5809
5810 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5811 'wristwatch attack' using huge encoding parameters (cf.
5812 James H. Manger's CRYPTO 2001 paper). Note that the
5813 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5814 encoding parameters and hence was not vulnerable.
5815 [Bodo Moeller]
5816
5817 *) BN_sqr() bug fix.
5818 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5819
5820 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5821 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5822 followed by modular reduction.
5823 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5824
5825 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5826 equivalent based on BN_pseudo_rand() instead of BN_rand().
5827 [Bodo Moeller]
5828
5829 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5830 This function was broken, as the check for a new client hello message
5831 to handle SGC did not allow these large messages.
5832 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5833 [Lutz Jaenicke]
5834
5835 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5836 [Lutz Jaenicke]
5837
5838 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5839 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5840 [Lutz Jaenicke]
5841
5842 *) Rework the configuration and shared library support for Tru64 Unix.
5843 The configuration part makes use of modern compiler features and
5844 still retains old compiler behavior for those that run older versions
5845 of the OS. The shared library support part includes a variant that
5846 uses the RPATH feature, and is available through the special
5847 configuration target "alpha-cc-rpath", which will never be selected
5848 automatically.
5849 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5850
5851 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5852 with the same message size as in ssl3_get_certificate_request().
5853 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5854 messages might inadvertently be reject as too long.
5855 [Petr Lampa <lampa@fee.vutbr.cz>]
5856
5857 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5858 [Andy Polyakov]
5859
5860 *) Modified SSL library such that the verify_callback that has been set
5861 specificly for an SSL object with SSL_set_verify() is actually being
5862 used. Before the change, a verify_callback set with this function was
5863 ignored and the verify_callback() set in the SSL_CTX at the time of
5864 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5865 to allow the necessary settings.
5866 [Lutz Jaenicke]
5867
5868 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5869 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5870 done automatically (in contradiction to the requirements of the C
5871 standard). This made problems when used from OpenSSH.
5872 [Lutz Jaenicke]
5873
5874 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5875 dh->length and always used
5876
5877 BN_rand_range(priv_key, dh->p).
5878
5879 BN_rand_range() is not necessary for Diffie-Hellman, and this
5880 specific range makes Diffie-Hellman unnecessarily inefficient if
5881 dh->length (recommended exponent length) is much smaller than the
5882 length of dh->p. We could use BN_rand_range() if the order of
5883 the subgroup was stored in the DH structure, but we only have
5884 dh->length.
5885
5886 So switch back to
5887
5888 BN_rand(priv_key, l, ...)
5889
5890 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5891 otherwise.
5892 [Bodo Moeller]
5893
5894 *) In
5895
5896 RSA_eay_public_encrypt
5897 RSA_eay_private_decrypt
5898 RSA_eay_private_encrypt (signing)
5899 RSA_eay_public_decrypt (signature verification)
5900
5901 (default implementations for RSA_public_encrypt,
5902 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5903 always reject numbers >= n.
5904 [Bodo Moeller]
5905
5906 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5907 to synchronize access to 'locking_thread'. This is necessary on
5908 systems where access to 'locking_thread' (an 'unsigned long'
5909 variable) is not atomic.
5910 [Bodo Moeller]
5911
5912 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5913 *before* setting the 'crypto_lock_rand' flag. The previous code had
5914 a race condition if 0 is a valid thread ID.
5915 [Travis Vitek <vitek@roguewave.com>]
5916
5917 *) Add support for shared libraries under Irix.
5918 [Albert Chin-A-Young <china@thewrittenword.com>]
5919
5920 *) Add configuration option to build on Linux on both big-endian and
5921 little-endian MIPS.
5922 [Ralf Baechle <ralf@uni-koblenz.de>]
5923
5924 *) Add the possibility to create shared libraries on HP-UX.
5925 [Richard Levitte]
5926
5927 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5928
5929 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5930 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5931 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5932 PRNG state recovery was possible based on the output of
5933 one PRNG request appropriately sized to gain knowledge on
5934 'md' followed by enough consecutive 1-byte PRNG requests
5935 to traverse all of 'state'.
5936
5937 1. When updating 'md_local' (the current thread's copy of 'md')
5938 during PRNG output generation, hash all of the previous
5939 'md_local' value, not just the half used for PRNG output.
5940
5941 2. Make the number of bytes from 'state' included into the hash
5942 independent from the number of PRNG bytes requested.
5943
5944 The first measure alone would be sufficient to avoid
5945 Markku-Juhani's attack. (Actually it had never occurred
5946 to me that the half of 'md_local' used for chaining was the
5947 half from which PRNG output bytes were taken -- I had always
5948 assumed that the secret half would be used.) The second
5949 measure makes sure that additional data from 'state' is never
5950 mixed into 'md_local' in small portions; this heuristically
5951 further strengthens the PRNG.
5952 [Bodo Moeller]
5953
5954 *) Fix crypto/bn/asm/mips3.s.
5955 [Andy Polyakov]
5956
5957 *) When only the key is given to "enc", the IV is undefined. Print out
5958 an error message in this case.
5959 [Lutz Jaenicke]
5960
5961 *) Handle special case when X509_NAME is empty in X509 printing routines.
5962 [Steve Henson]
5963
5964 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5965 positive and less than q.
5966 [Bodo Moeller]
5967
5968 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5969 used: it isn't thread safe and the add_lock_callback should handle
5970 that itself.
5971 [Paul Rose <Paul.Rose@bridge.com>]
5972
5973 *) Verify that incoming data obeys the block size in
5974 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5975 [Bodo Moeller]
5976
5977 *) Fix OAEP check.
5978 [Ulf Möller, Bodo Möller]
5979
5980 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5981 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5982 when fixing the server behaviour for backwards-compatible 'client
5983 hello' messages. (Note that the attack is impractical against
5984 SSL 3.0 and TLS 1.0 anyway because length and version checking
5985 means that the probability of guessing a valid ciphertext is
5986 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5987 paper.)
5988
5989 Before 0.9.5, the countermeasure (hide the error by generating a
5990 random 'decryption result') did not work properly because
5991 ERR_clear_error() was missing, meaning that SSL_get_error() would
5992 detect the supposedly ignored error.
5993
5994 Both problems are now fixed.
5995 [Bodo Moeller]
5996
5997 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5998 (previously it was 1024).
5999 [Bodo Moeller]
6000
6001 *) Fix for compatibility mode trust settings: ignore trust settings
6002 unless some valid trust or reject settings are present.
6003 [Steve Henson]
6004
6005 *) Fix for blowfish EVP: its a variable length cipher.
6006 [Steve Henson]
6007
6008 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6009 parameters in DSA public key structures and return an error in the
6010 DSA routines if parameters are absent.
6011 [Steve Henson]
6012
6013 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6014 in the current directory if neither $RANDFILE nor $HOME was set.
6015 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6016 caused some confusion to Windows users who haven't defined $HOME.
6017 Thus RAND_file_name() is changed again: e_os.h can define a
6018 DEFAULT_HOME, which will be used if $HOME is not set.
6019 For Windows, we use "C:"; on other platforms, we still require
6020 environment variables.
6021
6022 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6023 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6024 having multiple threads call RAND_poll() concurrently.
6025 [Bodo Moeller]
6026
6027 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6028 combination of a flag and a thread ID variable.
6029 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6030 flag), *other* threads can enter ssleay_add_bytes without obeying
6031 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6032 that they do not hold after the first thread unsets add_do_not_lock).
6033 [Bodo Moeller]
6034
6035 *) Change bctest again: '-x' expressions are not available in all
6036 versions of 'test'.
6037 [Bodo Moeller]
6038
6039 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6040
6041 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6042 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6043
6044 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6045 the default extension for executables, if any. Also, make the perl
6046 scripts that use symlink() to test if it really exists and use "cp"
6047 if it doesn't. All this made OpenSSL compilable and installable in
6048 CygWin.
6049 [Richard Levitte]
6050
6051 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6052 If SEQUENCE is length is indefinite just set c->slen to the total
6053 amount of data available.
6054 [Steve Henson, reported by shige@FreeBSD.org]
6055 [This change does not apply to 0.9.7.]
6056
6057 *) Change bctest to avoid here-documents inside command substitution
6058 (workaround for FreeBSD /bin/sh bug).
6059 For compatibility with Ultrix, avoid shell functions (introduced
6060 in the bctest version that searches along $PATH).
6061 [Bodo Moeller]
6062
6063 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6064 with des_encrypt() defined on some operating systems, like Solaris
6065 and UnixWare.
6066 [Richard Levitte]
6067
6068 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6069 On the Importance of Eliminating Errors in Cryptographic
6070 Computations, J. Cryptology 14 (2001) 2, 101-119,
6071 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6072 [Ulf Moeller]
6073
6074 *) MIPS assembler BIGNUM division bug fix.
6075 [Andy Polyakov]
6076
6077 *) Disabled incorrect Alpha assembler code.
6078 [Richard Levitte]
6079
6080 *) Fix PKCS#7 decode routines so they correctly update the length
6081 after reading an EOC for the EXPLICIT tag.
6082 [Steve Henson]
6083 [This change does not apply to 0.9.7.]
6084
6085 *) Fix bug in PKCS#12 key generation routines. This was triggered
6086 if a 3DES key was generated with a 0 initial byte. Include
6087 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6088 (but broken) behaviour.
6089 [Steve Henson]
6090
6091 *) Enhance bctest to search for a working bc along $PATH and print
6092 it when found.
6093 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6094
6095 *) Fix memory leaks in err.c: free err_data string if necessary;
6096 don't write to the wrong index in ERR_set_error_data.
6097 [Bodo Moeller]
6098
6099 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6100 did not exist.
6101 [Bodo Moeller]
6102
6103 *) Replace rdtsc with _emit statements for VC++ version 5.
6104 [Jeremy Cooper <jeremy@baymoo.org>]
6105
6106 *) Make it possible to reuse SSLv2 sessions.
6107 [Richard Levitte]
6108
6109 *) In copy_email() check for >= 0 as a return value for
6110 X509_NAME_get_index_by_NID() since 0 is a valid index.
6111 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6112
6113 *) Avoid coredump with unsupported or invalid public keys by checking if
6114 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6115 PKCS7_verify() fails with non detached data.
6116 [Steve Henson]
6117
6118 *) Don't use getenv in library functions when run as setuid/setgid.
6119 New function OPENSSL_issetugid().
6120 [Ulf Moeller]
6121
6122 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6123 due to incorrect handling of multi-threading:
6124
6125 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6126
6127 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6128
6129 3. Count how many times MemCheck_off() has been called so that
6130 nested use can be treated correctly. This also avoids
6131 inband-signalling in the previous code (which relied on the
6132 assumption that thread ID 0 is impossible).
6133 [Bodo Moeller]
6134
6135 *) Add "-rand" option also to s_client and s_server.
6136 [Lutz Jaenicke]
6137
6138 *) Fix CPU detection on Irix 6.x.
6139 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6140 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6141
6142 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6143 was empty.
6144 [Steve Henson]
6145 [This change does not apply to 0.9.7.]
6146
6147 *) Use the cached encoding of an X509_NAME structure rather than
6148 copying it. This is apparently the reason for the libsafe "errors"
6149 but the code is actually correct.
6150 [Steve Henson]
6151
6152 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6153 Bleichenbacher's DSA attack.
6154 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6155 to be set and top=0 forces the highest bit to be set; top=-1 is new
6156 and leaves the highest bit random.
6157 [Ulf Moeller, Bodo Moeller]
6158
6159 *) In the NCONF_...-based implementations for CONF_... queries
6160 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6161 a temporary CONF structure with the data component set to NULL
6162 (which gives segmentation faults in lh_retrieve).
6163 Instead, use NULL for the CONF pointer in CONF_get_string and
6164 CONF_get_number (which may use environment variables) and directly
6165 return NULL from CONF_get_section.
6166 [Bodo Moeller]
6167
6168 *) Fix potential buffer overrun for EBCDIC.
6169 [Ulf Moeller]
6170
6171 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6172 keyUsage if basicConstraints absent for a CA.
6173 [Steve Henson]
6174
6175 *) Make SMIME_write_PKCS7() write mail header values with a format that
6176 is more generally accepted (no spaces before the semicolon), since
6177 some programs can't parse those values properly otherwise. Also make
6178 sure BIO's that break lines after each write do not create invalid
6179 headers.
6180 [Richard Levitte]
6181
6182 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6183 macros previously used would not encode an empty SEQUENCE OF
6184 and break the signature.
6185 [Steve Henson]
6186 [This change does not apply to 0.9.7.]
6187
6188 *) Zero the premaster secret after deriving the master secret in
6189 DH ciphersuites.
6190 [Steve Henson]
6191
6192 *) Add some EVP_add_digest_alias registrations (as found in
6193 OpenSSL_add_all_digests()) to SSL_library_init()
6194 aka OpenSSL_add_ssl_algorithms(). This provides improved
6195 compatibility with peers using X.509 certificates
6196 with unconventional AlgorithmIdentifier OIDs.
6197 [Bodo Moeller]
6198
6199 *) Fix for Irix with NO_ASM.
6200 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6201
6202 *) ./config script fixes.
6203 [Ulf Moeller, Richard Levitte]
6204
6205 *) Fix 'openssl passwd -1'.
6206 [Bodo Moeller]
6207
6208 *) Change PKCS12_key_gen_asc() so it can cope with non null
6209 terminated strings whose length is passed in the passlen
6210 parameter, for example from PEM callbacks. This was done
6211 by adding an extra length parameter to asc2uni().
6212 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6213
6214 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6215 call failed, free the DSA structure.
6216 [Bodo Moeller]
6217
6218 *) Fix to uni2asc() to cope with zero length Unicode strings.
6219 These are present in some PKCS#12 files.
6220 [Steve Henson]
6221
6222 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6223 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6224 when writing a 32767 byte record.
6225 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6226
6227 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6228 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6229
6230 (RSA objects have a reference count access to which is protected
6231 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6232 so they are meant to be shared between threads.)
6233 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6234 "Reddie, Steven" <Steven.Reddie@ca.com>]
6235
6236 *) Fix a deadlock in CRYPTO_mem_leaks().
6237 [Bodo Moeller]
6238
6239 *) Use better test patterns in bntest.
6240 [Ulf Möller]
6241
6242 *) rand_win.c fix for Borland C.
6243 [Ulf Möller]
6244
6245 *) BN_rshift bugfix for n == 0.
6246 [Bodo Moeller]
6247
6248 *) Add a 'bctest' script that checks for some known 'bc' bugs
6249 so that 'make test' does not abort just because 'bc' is broken.
6250 [Bodo Moeller]
6251
6252 *) Store verify_result within SSL_SESSION also for client side to
6253 avoid potential security hole. (Re-used sessions on the client side
6254 always resulted in verify_result==X509_V_OK, not using the original
6255 result of the server certificate verification.)
6256 [Lutz Jaenicke]
6257
6258 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6259 SSL3_RT_APPLICATION_DATA, return 0.
6260 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6261 [Bodo Moeller]
6262
6263 *) Fix SSL_peek:
6264 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6265 releases, have been re-implemented by renaming the previous
6266 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6267 and ssl3_read_internal, respectively, and adding 'peek' parameters
6268 to them. The new ssl[23]_{read,peek} functions are calls to
6269 ssl[23]_read_internal with the 'peek' flag set appropriately.
6270 A 'peek' parameter has also been added to ssl3_read_bytes, which
6271 does the actual work for ssl3_read_internal.
6272 [Bodo Moeller]
6273
6274 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6275 the method-specific "init()" handler. Also clean up ex_data after
6276 calling the method-specific "finish()" handler. Previously, this was
6277 happening the other way round.
6278 [Geoff Thorpe]
6279
6280 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6281 The previous value, 12, was not always sufficient for BN_mod_exp().
6282 [Bodo Moeller]
6283
6284 *) Make sure that shared libraries get the internal name engine with
6285 the full version number and not just 0. This should mark the
6286 shared libraries as not backward compatible. Of course, this should
6287 be changed again when we can guarantee backward binary compatibility.
6288 [Richard Levitte]
6289
6290 *) Fix typo in get_cert_by_subject() in by_dir.c
6291 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6292
6293 *) Rework the system to generate shared libraries:
6294
6295 - Make note of the expected extension for the shared libraries and
6296 if there is a need for symbolic links from for example libcrypto.so.0
6297 to libcrypto.so.0.9.7. There is extended info in Configure for
6298 that.
6299
6300 - Make as few rebuilds of the shared libraries as possible.
6301
6302 - Still avoid linking the OpenSSL programs with the shared libraries.
6303
6304 - When installing, install the shared libraries separately from the
6305 static ones.
6306 [Richard Levitte]
6307
6308 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6309
6310 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6311 and not in SSL_clear because the latter is also used by the
6312 accept/connect functions; previously, the settings made by
6313 SSL_set_read_ahead would be lost during the handshake.
6314 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6315
6316 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6317 Previously, it would create entries for disableed algorithms no
6318 matter what.
6319 [Richard Levitte]
6320
6321 *) Added several new manual pages for SSL_* function.
6322 [Lutz Jaenicke]
6323
6324 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6325
6326 *) In ssl23_get_client_hello, generate an error message when faced
6327 with an initial SSL 3.0/TLS record that is too small to contain the
6328 first two bytes of the ClientHello message, i.e. client_version.
6329 (Note that this is a pathologic case that probably has never happened
6330 in real life.) The previous approach was to use the version number
6331 from the record header as a substitute; but our protocol choice
6332 should not depend on that one because it is not authenticated
6333 by the Finished messages.
6334 [Bodo Moeller]
6335
6336 *) More robust randomness gathering functions for Windows.
6337 [Jeffrey Altman <jaltman@columbia.edu>]
6338
6339 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6340 not set then we don't setup the error code for issuer check errors
6341 to avoid possibly overwriting other errors which the callback does
6342 handle. If an application does set the flag then we assume it knows
6343 what it is doing and can handle the new informational codes
6344 appropriately.
6345 [Steve Henson]
6346
6347 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6348 a general "ANY" type, as such it should be able to decode anything
6349 including tagged types. However it didn't check the class so it would
6350 wrongly interpret tagged types in the same way as their universal
6351 counterpart and unknown types were just rejected. Changed so that the
6352 tagged and unknown types are handled in the same way as a SEQUENCE:
6353 that is the encoding is stored intact. There is also a new type
6354 "V_ASN1_OTHER" which is used when the class is not universal, in this
6355 case we have no idea what the actual type is so we just lump them all
6356 together.
6357 [Steve Henson]
6358
6359 *) On VMS, stdout may very well lead to a file that is written to
6360 in a record-oriented fashion. That means that every write() will
6361 write a separate record, which will be read separately by the
6362 programs trying to read from it. This can be very confusing.
6363
6364 The solution is to put a BIO filter in the way that will buffer
6365 text until a linefeed is reached, and then write everything a
6366 line at a time, so every record written will be an actual line,
6367 not chunks of lines and not (usually doesn't happen, but I've
6368 seen it once) several lines in one record. BIO_f_linebuffer() is
6369 the answer.
6370
6371 Currently, it's a VMS-only method, because that's where it has
6372 been tested well enough.
6373 [Richard Levitte]
6374
6375 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6376 it can return incorrect results.
6377 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6378 but it was in 0.9.6-beta[12].)
6379 [Bodo Moeller]
6380
6381 *) Disable the check for content being present when verifying detached
6382 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6383 include zero length content when signing messages.
6384 [Steve Henson]
6385
6386 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6387 BIO_ctrl (for BIO pairs).
6388 [Bodo Möller]
6389
6390 *) Add DSO method for VMS.
6391 [Richard Levitte]
6392
6393 *) Bug fix: Montgomery multiplication could produce results with the
6394 wrong sign.
6395 [Ulf Möller]
6396
6397 *) Add RPM specification openssl.spec and modify it to build three
6398 packages. The default package contains applications, application
6399 documentation and run-time libraries. The devel package contains
6400 include files, static libraries and function documentation. The
6401 doc package contains the contents of the doc directory. The original
6402 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6403 [Richard Levitte]
6404
6405 *) Add a large number of documentation files for many SSL routines.
6406 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6407
6408 *) Add a configuration entry for Sony News 4.
6409 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6410
6411 *) Don't set the two most significant bits to one when generating a
6412 random number < q in the DSA library.
6413 [Ulf Möller]
6414
6415 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6416 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6417 the underlying transport is blocking) if a handshake took place.
6418 (The default behaviour is needed by applications such as s_client
6419 and s_server that use select() to determine when to use SSL_read;
6420 but for applications that know in advance when to expect data, it
6421 just makes things more complicated.)
6422 [Bodo Moeller]
6423
6424 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6425 from EGD.
6426 [Ben Laurie]
6427
6428 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6429 work better on such systems.
6430 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6431
6432 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6433 Update PKCS12_parse() so it copies the friendlyName and the
6434 keyid to the certificates aux info.
6435 [Steve Henson]
6436
6437 *) Fix bug in PKCS7_verify() which caused an infinite loop
6438 if there was more than one signature.
6439 [Sven Uszpelkat <su@celocom.de>]
6440
6441 *) Major change in util/mkdef.pl to include extra information
6442 about each symbol, as well as presentig variables as well
6443 as functions. This change means that there's n more need
6444 to rebuild the .num files when some algorithms are excluded.
6445 [Richard Levitte]
6446
6447 *) Allow the verify time to be set by an application,
6448 rather than always using the current time.
6449 [Steve Henson]
6450
6451 *) Phase 2 verify code reorganisation. The certificate
6452 verify code now looks up an issuer certificate by a
6453 number of criteria: subject name, authority key id
6454 and key usage. It also verifies self signed certificates
6455 by the same criteria. The main comparison function is
6456 X509_check_issued() which performs these checks.
6457
6458 Lot of changes were necessary in order to support this
6459 without completely rewriting the lookup code.
6460
6461 Authority and subject key identifier are now cached.
6462
6463 The LHASH 'certs' is X509_STORE has now been replaced
6464 by a STACK_OF(X509_OBJECT). This is mainly because an
6465 LHASH can't store or retrieve multiple objects with
6466 the same hash value.
6467
6468 As a result various functions (which were all internal
6469 use only) have changed to handle the new X509_STORE
6470 structure. This will break anything that messed round
6471 with X509_STORE internally.
6472
6473 The functions X509_STORE_add_cert() now checks for an
6474 exact match, rather than just subject name.
6475
6476 The X509_STORE API doesn't directly support the retrieval
6477 of multiple certificates matching a given criteria, however
6478 this can be worked round by performing a lookup first
6479 (which will fill the cache with candidate certificates)
6480 and then examining the cache for matches. This is probably
6481 the best we can do without throwing out X509_LOOKUP
6482 entirely (maybe later...).
6483
6484 The X509_VERIFY_CTX structure has been enhanced considerably.
6485
6486 All certificate lookup operations now go via a get_issuer()
6487 callback. Although this currently uses an X509_STORE it
6488 can be replaced by custom lookups. This is a simple way
6489 to bypass the X509_STORE hackery necessary to make this
6490 work and makes it possible to use more efficient techniques
6491 in future. A very simple version which uses a simple
6492 STACK for its trusted certificate store is also provided
6493 using X509_STORE_CTX_trusted_stack().
6494
6495 The verify_cb() and verify() callbacks now have equivalents
6496 in the X509_STORE_CTX structure.
6497
6498 X509_STORE_CTX also has a 'flags' field which can be used
6499 to customise the verify behaviour.
6500 [Steve Henson]
6501
6502 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6503 excludes S/MIME capabilities.
6504 [Steve Henson]
6505
6506 *) When a certificate request is read in keep a copy of the
6507 original encoding of the signed data and use it when outputing
6508 again. Signatures then use the original encoding rather than
6509 a decoded, encoded version which may cause problems if the
6510 request is improperly encoded.
6511 [Steve Henson]
6512
6513 *) For consistency with other BIO_puts implementations, call
6514 buffer_write(b, ...) directly in buffer_puts instead of calling
6515 BIO_write(b, ...).
6516
6517 In BIO_puts, increment b->num_write as in BIO_write.
6518 [Peter.Sylvester@EdelWeb.fr]
6519
6520 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6521 BN_zero, we may not return a BIGNUM with an array consisting of
6522 words set to zero.)
6523 [Bodo Moeller]
6524
6525 *) Avoid calling abort() from within the library when problems are
6526 detected, except if preprocessor symbols have been defined
6527 (such as REF_CHECK, BN_DEBUG etc.).
6528 [Bodo Moeller]
6529
6530 *) New openssl application 'rsautl'. This utility can be
6531 used for low level RSA operations. DER public key
6532 BIO/fp routines also added.
6533 [Steve Henson]
6534
6535 *) New Configure entry and patches for compiling on QNX 4.
6536 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6537
6538 *) A demo state-machine implementation was sponsored by
6539 Nuron (http://www.nuron.com/) and is now available in
6540 demos/state_machine.
6541 [Ben Laurie]
6542
6543 *) New options added to the 'dgst' utility for signature
6544 generation and verification.
6545 [Steve Henson]
6546
6547 *) Unrecognized PKCS#7 content types are now handled via a
6548 catch all ASN1_TYPE structure. This allows unsupported
6549 types to be stored as a "blob" and an application can
6550 encode and decode it manually.
6551 [Steve Henson]
6552
6553 *) Fix various signed/unsigned issues to make a_strex.c
6554 compile under VC++.
6555 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6556
6557 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6558 length if passed a buffer. ASN1_INTEGER_to_BN failed
6559 if passed a NULL BN and its argument was negative.
6560 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6561
6562 *) Modification to PKCS#7 encoding routines to output definite
6563 length encoding. Since currently the whole structures are in
6564 memory there's not real point in using indefinite length
6565 constructed encoding. However if OpenSSL is compiled with
6566 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6567 [Steve Henson]
6568
6569 *) Added BIO_vprintf() and BIO_vsnprintf().
6570 [Richard Levitte]
6571
6572 *) Added more prefixes to parse for in the the strings written
6573 through a logging bio, to cover all the levels that are available
6574 through syslog. The prefixes are now:
6575
6576 PANIC, EMERG, EMR => LOG_EMERG
6577 ALERT, ALR => LOG_ALERT
6578 CRIT, CRI => LOG_CRIT
6579 ERROR, ERR => LOG_ERR
6580 WARNING, WARN, WAR => LOG_WARNING
6581 NOTICE, NOTE, NOT => LOG_NOTICE
6582 INFO, INF => LOG_INFO
6583 DEBUG, DBG => LOG_DEBUG
6584
6585 and as before, if none of those prefixes are present at the
6586 beginning of the string, LOG_ERR is chosen.
6587
6588 On Win32, the LOG_* levels are mapped according to this:
6589
6590 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6591 LOG_WARNING => EVENTLOG_WARNING_TYPE
6592 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6593
6594 [Richard Levitte]
6595
6596 *) Made it possible to reconfigure with just the configuration
6597 argument "reconf" or "reconfigure". The command line arguments
6598 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6599 and are retrieved from there when reconfiguring.
6600 [Richard Levitte]
6601
6602 *) MD4 implemented.
6603 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6604
6605 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6606 [Richard Levitte]
6607
6608 *) The obj_dat.pl script was messing up the sorting of object
6609 names. The reason was that it compared the quoted version
6610 of strings as a result "OCSP" > "OCSP Signing" because
6611 " > SPACE. Changed script to store unquoted versions of
6612 names and add quotes on output. It was also omitting some
6613 names from the lookup table if they were given a default
6614 value (that is if SN is missing it is given the same
6615 value as LN and vice versa), these are now added on the
6616 grounds that if an object has a name we should be able to
6617 look it up. Finally added warning output when duplicate
6618 short or long names are found.
6619 [Steve Henson]
6620
6621 *) Changes needed for Tandem NSK.
6622 [Scott Uroff <scott@xypro.com>]
6623
6624 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6625 RSA_padding_check_SSLv23(), special padding was never detected
6626 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6627 version rollback attacks was not effective.
6628
6629 In s23_clnt.c, don't use special rollback-attack detection padding
6630 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6631 client; similarly, in s23_srvr.c, don't do the rollback check if
6632 SSL 2.0 is the only protocol enabled in the server.
6633 [Bodo Moeller]
6634
6635 *) Make it possible to get hexdumps of unprintable data with 'openssl
6636 asn1parse'. By implication, the functions ASN1_parse_dump() and
6637 BIO_dump_indent() are added.
6638 [Richard Levitte]
6639
6640 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6641 these print out strings and name structures based on various
6642 flags including RFC2253 support and proper handling of
6643 multibyte characters. Added options to the 'x509' utility
6644 to allow the various flags to be set.
6645 [Steve Henson]
6646
6647 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6648 Also change the functions X509_cmp_current_time() and
6649 X509_gmtime_adj() work with an ASN1_TIME structure,
6650 this will enable certificates using GeneralizedTime in validity
6651 dates to be checked.
6652 [Steve Henson]
6653
6654 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6655 negative public key encodings) on by default,
6656 NO_NEG_PUBKEY_BUG can be set to disable it.
6657 [Steve Henson]
6658
6659 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6660 content octets. An i2c_ASN1_OBJECT is unnecessary because
6661 the encoding can be trivially obtained from the structure.
6662 [Steve Henson]
6663
6664 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6665 not read locks (CRYPTO_r_[un]lock).
6666 [Bodo Moeller]
6667
6668 *) A first attempt at creating official support for shared
6669 libraries through configuration. I've kept it so the
6670 default is static libraries only, and the OpenSSL programs
6671 are always statically linked for now, but there are
6672 preparations for dynamic linking in place.
6673 This has been tested on Linux and Tru64.
6674 [Richard Levitte]
6675
6676 *) Randomness polling function for Win9x, as described in:
6677 Peter Gutmann, Software Generation of Practically Strong
6678 Random Numbers.
6679 [Ulf Möller]
6680
6681 *) Fix so PRNG is seeded in req if using an already existing
6682 DSA key.
6683 [Steve Henson]
6684
6685 *) New options to smime application. -inform and -outform
6686 allow alternative formats for the S/MIME message including
6687 PEM and DER. The -content option allows the content to be
6688 specified separately. This should allow things like Netscape
6689 form signing output easier to verify.
6690 [Steve Henson]
6691
6692 *) Fix the ASN1 encoding of tags using the 'long form'.
6693 [Steve Henson]
6694
6695 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6696 STRING types. These convert content octets to and from the
6697 underlying type. The actual tag and length octets are
6698 already assumed to have been read in and checked. These
6699 are needed because all other string types have virtually
6700 identical handling apart from the tag. By having versions
6701 of the ASN1 functions that just operate on content octets
6702 IMPLICIT tagging can be handled properly. It also allows
6703 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6704 and ASN1_INTEGER are identical apart from the tag.
6705 [Steve Henson]
6706
6707 *) Change the handling of OID objects as follows:
6708
6709 - New object identifiers are inserted in objects.txt, following
6710 the syntax given in objects.README.
6711 - objects.pl is used to process obj_mac.num and create a new
6712 obj_mac.h.
6713 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6714 obj_mac.h.
6715
6716 This is currently kind of a hack, and the perl code in objects.pl
6717 isn't very elegant, but it works as I intended. The simplest way
6718 to check that it worked correctly is to look in obj_dat.h and
6719 check the array nid_objs and make sure the objects haven't moved
6720 around (this is important!). Additions are OK, as well as
6721 consistent name changes.
6722 [Richard Levitte]
6723
6724 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6725 [Bodo Moeller]
6726
6727 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6728 The given file adds to whatever has already been seeded into the
6729 random pool through the RANDFILE configuration file option or
6730 environment variable, or the default random state file.
6731 [Richard Levitte]
6732
6733 *) mkstack.pl now sorts each macro group into lexical order.
6734 Previously the output order depended on the order the files
6735 appeared in the directory, resulting in needless rewriting
6736 of safestack.h .
6737 [Steve Henson]
6738
6739 *) Patches to make OpenSSL compile under Win32 again. Mostly
6740 work arounds for the VC++ problem that it treats func() as
6741 func(void). Also stripped out the parts of mkdef.pl that
6742 added extra typesafe functions: these no longer exist.
6743 [Steve Henson]
6744
6745 *) Reorganisation of the stack code. The macros are now all
6746 collected in safestack.h . Each macro is defined in terms of
6747 a "stack macro" of the form SKM_<name>(type, a, b). The
6748 DEBUG_SAFESTACK is now handled in terms of function casts,
6749 this has the advantage of retaining type safety without the
6750 use of additional functions. If DEBUG_SAFESTACK is not defined
6751 then the non typesafe macros are used instead. Also modified the
6752 mkstack.pl script to handle the new form. Needs testing to see
6753 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6754 the default if no major problems. Similar behaviour for ASN1_SET_OF
6755 and PKCS12_STACK_OF.
6756 [Steve Henson]
6757
6758 *) When some versions of IIS use the 'NET' form of private key the
6759 key derivation algorithm is different. Normally MD5(password) is
6760 used as a 128 bit RC4 key. In the modified case
6761 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6762 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6763 as the old Netscape_RSA functions except they have an additional
6764 'sgckey' parameter which uses the modified algorithm. Also added
6765 an -sgckey command line option to the rsa utility. Thanks to
6766 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6767 algorithm to openssl-dev.
6768 [Steve Henson]
6769
6770 *) The evp_local.h macros were using 'c.##kname' which resulted in
6771 invalid expansion on some systems (SCO 5.0.5 for example).
6772 Corrected to 'c.kname'.
6773 [Phillip Porch <root@theporch.com>]
6774
6775 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6776 a STACK of email addresses from a certificate or request, these look
6777 in the subject name and the subject alternative name extensions and
6778 omit any duplicate addresses.
6779 [Steve Henson]
6780
6781 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6782 This makes DSA verification about 2 % faster.
6783 [Bodo Moeller]
6784
6785 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6786 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6787 plus overhead for 1024 bit moduli).
6788 This makes exponentiations about 0.5 % faster for 1024 bit
6789 exponents (as measured by "openssl speed rsa2048").
6790 [Bodo Moeller]
6791
6792 *) Rename memory handling macros to avoid conflicts with other
6793 software:
6794 Malloc => OPENSSL_malloc
6795 Malloc_locked => OPENSSL_malloc_locked
6796 Realloc => OPENSSL_realloc
6797 Free => OPENSSL_free
6798 [Richard Levitte]
6799
6800 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6801 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6802 [Bodo Moeller]
6803
6804 *) CygWin32 support.
6805 [John Jarvie <jjarvie@newsguy.com>]
6806
6807 *) The type-safe stack code has been rejigged. It is now only compiled
6808 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6809 by default all type-specific stack functions are "#define"d back to
6810 standard stack functions. This results in more streamlined output
6811 but retains the type-safety checking possibilities of the original
6812 approach.
6813 [Geoff Thorpe]
6814
6815 *) The STACK code has been cleaned up, and certain type declarations
6816 that didn't make a lot of sense have been brought in line. This has
6817 also involved a cleanup of sorts in safestack.h to more correctly
6818 map type-safe stack functions onto their plain stack counterparts.
6819 This work has also resulted in a variety of "const"ifications of
6820 lots of the code, especially "_cmp" operations which should normally
6821 be prototyped with "const" parameters anyway.
6822 [Geoff Thorpe]
6823
6824 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6825 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6826 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6827 where all of 'md' is used each time the PRNG is used, but 'state'
6828 is used only indexed by a cyclic counter. As entropy may not be
6829 well distributed from the beginning, 'md' is important as a
6830 chaining variable. However, the output function chains only half
6831 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6832 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6833 in all of 'state' being rewritten, with the new values depending
6834 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6835 [Bodo Moeller]
6836
6837 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6838 the handshake is continued after ssl_verify_cert_chain();
6839 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6840 can lead to 'unexplainable' connection aborts later.
6841 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6842
6843 *) Major EVP API cipher revision.
6844 Add hooks for extra EVP features. This allows various cipher
6845 parameters to be set in the EVP interface. Support added for variable
6846 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6847 setting of RC2 and RC5 parameters.
6848
6849 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6850 ciphers.
6851
6852 Remove lots of duplicated code from the EVP library. For example *every*
6853 cipher init() function handles the 'iv' in the same way according to the
6854 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6855 for CFB and OFB modes they zero ctx->num.
6856
6857 New functionality allows removal of S/MIME code RC2 hack.
6858
6859 Most of the routines have the same form and so can be declared in terms
6860 of macros.
6861
6862 By shifting this to the top level EVP_CipherInit() it can be removed from
6863 all individual ciphers. If the cipher wants to handle IVs or keys
6864 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6865 flags.
6866
6867 Change lots of functions like EVP_EncryptUpdate() to now return a
6868 value: although software versions of the algorithms cannot fail
6869 any installed hardware versions can.
6870 [Steve Henson]
6871
6872 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6873 this option is set, tolerate broken clients that send the negotiated
6874 protocol version number instead of the requested protocol version
6875 number.
6876 [Bodo Moeller]
6877
6878 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6879 i.e. non-zero for export ciphersuites, zero otherwise.
6880 Previous versions had this flag inverted, inconsistent with
6881 rsa_tmp_cb (..._TMP_RSA_CB).
6882 [Bodo Moeller; problem reported by Amit Chopra]
6883
6884 *) Add missing DSA library text string. Work around for some IIS
6885 key files with invalid SEQUENCE encoding.
6886 [Steve Henson]
6887
6888 *) Add a document (doc/standards.txt) that list all kinds of standards
6889 and so on that are implemented in OpenSSL.
6890 [Richard Levitte]
6891
6892 *) Enhance c_rehash script. Old version would mishandle certificates
6893 with the same subject name hash and wouldn't handle CRLs at all.
6894 Added -fingerprint option to crl utility, to support new c_rehash
6895 features.
6896 [Steve Henson]
6897
6898 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6899 [Ulf Möller]
6900
6901 *) Fix for SSL server purpose checking. Server checking was
6902 rejecting certificates which had extended key usage present
6903 but no ssl client purpose.
6904 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6905
6906 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6907 is a little unclear about how a blank password is handled.
6908 Since the password in encoded as a BMPString with terminating
6909 double NULL a zero length password would end up as just the
6910 double NULL. However no password at all is different and is
6911 handled differently in the PKCS#12 key generation code. NS
6912 treats a blank password as zero length. MSIE treats it as no
6913 password on export: but it will try both on import. We now do
6914 the same: PKCS12_parse() tries zero length and no password if
6915 the password is set to "" or NULL (NULL is now a valid password:
6916 it wasn't before) as does the pkcs12 application.
6917 [Steve Henson]
6918
6919 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6920 perror when PEM_read_bio_X509_REQ fails, the error message must
6921 be obtained from the error queue.
6922 [Bodo Moeller]
6923
6924 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6925 it in ERR_remove_state if appropriate, and change ERR_get_state
6926 accordingly to avoid race conditions (this is necessary because
6927 thread_hash is no longer constant once set).
6928 [Bodo Moeller]
6929
6930 *) Bugfix for linux-elf makefile.one.
6931 [Ulf Möller]
6932
6933 *) RSA_get_default_method() will now cause a default
6934 RSA_METHOD to be chosen if one doesn't exist already.
6935 Previously this was only set during a call to RSA_new()
6936 or RSA_new_method(NULL) meaning it was possible for
6937 RSA_get_default_method() to return NULL.
6938 [Geoff Thorpe]
6939
6940 *) Added native name translation to the existing DSO code
6941 that will convert (if the flag to do so is set) filenames
6942 that are sufficiently small and have no path information
6943 into a canonical native form. Eg. "blah" converted to
6944 "libblah.so" or "blah.dll" etc.
6945 [Geoff Thorpe]
6946
6947 *) New function ERR_error_string_n(e, buf, len) which is like
6948 ERR_error_string(e, buf), but writes at most 'len' bytes
6949 including the 0 terminator. For ERR_error_string_n, 'buf'
6950 may not be NULL.
6951 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6952
6953 *) CONF library reworked to become more general. A new CONF
6954 configuration file reader "class" is implemented as well as a
6955 new functions (NCONF_*, for "New CONF") to handle it. The now
6956 old CONF_* functions are still there, but are reimplemented to
6957 work in terms of the new functions. Also, a set of functions
6958 to handle the internal storage of the configuration data is
6959 provided to make it easier to write new configuration file
6960 reader "classes" (I can definitely see something reading a
6961 configuration file in XML format, for example), called _CONF_*,
6962 or "the configuration storage API"...
6963
6964 The new configuration file reading functions are:
6965
6966 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6967 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6968
6969 NCONF_default, NCONF_WIN32
6970
6971 NCONF_dump_fp, NCONF_dump_bio
6972
6973 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6974 NCONF_new creates a new CONF object. This works in the same way
6975 as other interfaces in OpenSSL, like the BIO interface.
6976 NCONF_dump_* dump the internal storage of the configuration file,
6977 which is useful for debugging. All other functions take the same
6978 arguments as the old CONF_* functions wth the exception of the
6979 first that must be a `CONF *' instead of a `LHASH *'.
6980
6981 To make it easer to use the new classes with the old CONF_* functions,
6982 the function CONF_set_default_method is provided.
6983 [Richard Levitte]
6984
6985 *) Add '-tls1' option to 'openssl ciphers', which was already
6986 mentioned in the documentation but had not been implemented.
6987 (This option is not yet really useful because even the additional
6988 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6989 [Bodo Moeller]
6990
6991 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6992 OpenSSL-based applications) load shared libraries and bind to
6993 them in a portable way.
6994 [Geoff Thorpe, with contributions from Richard Levitte]
6995
6996 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6997
6998 *) Make sure _lrotl and _lrotr are only used with MSVC.
6999
7000 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7001 (the default implementation of RAND_status).
7002
7003 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7004 to '-clrext' (= clear extensions), as intended and documented.
7005 [Bodo Moeller; inconsistency pointed out by Michael Attili
7006 <attili@amaxo.com>]
7007
7008 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7009 was larger than the MD block size.
7010 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7011
7012 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7013 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7014 using the passed key: if the passed key was a private key the result
7015 of X509_print(), for example, would be to print out all the private key
7016 components.
7017 [Steve Henson]
7018
7019 *) des_quad_cksum() byte order bug fix.
7020 [Ulf Möller, using the problem description in krb4-0.9.7, where
7021 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7022
7023 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7024 discouraged.
7025 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7026
7027 *) For easily testing in shell scripts whether some command
7028 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7029 returns with exit code 0 iff no command of the given name is available.
7030 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7031 the output goes to stdout and nothing is printed to stderr.
7032 Additional arguments are always ignored.
7033
7034 Since for each cipher there is a command of the same name,
7035 the 'no-cipher' compilation switches can be tested this way.
7036
7037 ('openssl no-XXX' is not able to detect pseudo-commands such
7038 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7039 [Bodo Moeller]
7040
7041 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7042 [Bodo Moeller]
7043
7044 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7045 is set; it will be thrown away anyway because each handshake creates
7046 its own key.
7047 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7048 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7049 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7050 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7051 [Bodo Moeller]
7052
7053 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7054 'Q' and 'R' lose their special meanings (quit/renegotiate).
7055 This is part of what -quiet does; unlike -quiet, -ign_eof
7056 does not suppress any output.
7057 [Richard Levitte]
7058
7059 *) Add compatibility options to the purpose and trust code. The
7060 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7061 accepts a certificate or CA, this was the previous behaviour,
7062 with all the associated security issues.
7063
7064 X509_TRUST_COMPAT is the old trust behaviour: only and
7065 automatically trust self signed roots in certificate store. A
7066 new trust setting X509_TRUST_DEFAULT is used to specify that
7067 a purpose has no associated trust setting and it should instead
7068 use the value in the default purpose.
7069 [Steve Henson]
7070
7071 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7072 and fix a memory leak.
7073 [Steve Henson]
7074
7075 *) In util/mkerr.pl (which implements 'make errors'), preserve
7076 reason strings from the previous version of the .c file, as
7077 the default to have only downcase letters (and digits) in
7078 automatically generated reasons codes is not always appropriate.
7079 [Bodo Moeller]
7080
7081 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7082 using strerror. Previously, ERR_reason_error_string() returned
7083 library names as reason strings for SYSerr; but SYSerr is a special
7084 case where small numbers are errno values, not library numbers.
7085 [Bodo Moeller]
7086
7087 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7088 converts DSA parameters into DH parameters. (When creating parameters,
7089 DSA_generate_parameters is used.)
7090 [Bodo Moeller]
7091
7092 *) Include 'length' (recommended exponent length) in C code generated
7093 by 'openssl dhparam -C'.
7094 [Bodo Moeller]
7095
7096 *) The second argument to set_label in perlasm was already being used
7097 so couldn't be used as a "file scope" flag. Moved to third argument
7098 which was free.
7099 [Steve Henson]
7100
7101 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7102 instead of RAND_bytes for encryption IVs and salts.
7103 [Bodo Moeller]
7104
7105 *) Include RAND_status() into RAND_METHOD instead of implementing
7106 it only for md_rand.c Otherwise replacing the PRNG by calling
7107 RAND_set_rand_method would be impossible.
7108 [Bodo Moeller]
7109
7110 *) Don't let DSA_generate_key() enter an infinite loop if the random
7111 number generation fails.
7112 [Bodo Moeller]
7113
7114 *) New 'rand' application for creating pseudo-random output.
7115 [Bodo Moeller]
7116
7117 *) Added configuration support for Linux/IA64
7118 [Rolf Haberrecker <rolf@suse.de>]
7119
7120 *) Assembler module support for Mingw32.
7121 [Ulf Möller]
7122
7123 *) Shared library support for HPUX (in shlib/).
7124 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7125
7126 *) Shared library support for Solaris gcc.
7127 [Lutz Behnke <behnke@trustcenter.de>]
7128
7129 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7130
7131 *) PKCS7_encrypt() was adding text MIME headers twice because they
7132 were added manually and by SMIME_crlf_copy().
7133 [Steve Henson]
7134
7135 *) In bntest.c don't call BN_rand with zero bits argument.
7136 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7137
7138 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7139 case was implemented. This caused BN_div_recp() to fail occasionally.
7140 [Ulf Möller]
7141
7142 *) Add an optional second argument to the set_label() in the perl
7143 assembly language builder. If this argument exists and is set
7144 to 1 it signals that the assembler should use a symbol whose
7145 scope is the entire file, not just the current function. This
7146 is needed with MASM which uses the format label:: for this scope.
7147 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7148
7149 *) Change the ASN1 types so they are typedefs by default. Before
7150 almost all types were #define'd to ASN1_STRING which was causing
7151 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7152 for example.
7153 [Steve Henson]
7154
7155 *) Change names of new functions to the new get1/get0 naming
7156 convention: After 'get1', the caller owns a reference count
7157 and has to call ..._free; 'get0' returns a pointer to some
7158 data structure without incrementing reference counters.
7159 (Some of the existing 'get' functions increment a reference
7160 counter, some don't.)
7161 Similarly, 'set1' and 'add1' functions increase reference
7162 counters or duplicate objects.
7163 [Steve Henson]
7164
7165 *) Allow for the possibility of temp RSA key generation failure:
7166 the code used to assume it always worked and crashed on failure.
7167 [Steve Henson]
7168
7169 *) Fix potential buffer overrun problem in BIO_printf().
7170 [Ulf Möller, using public domain code by Patrick Powell; problem
7171 pointed out by David Sacerdote <das33@cornell.edu>]
7172
7173 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7174 RAND_egd() and RAND_status(). In the command line application,
7175 the EGD socket can be specified like a seed file using RANDFILE
7176 or -rand.
7177 [Ulf Möller]
7178
7179 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7180 Some CAs (e.g. Verisign) distribute certificates in this form.
7181 [Steve Henson]
7182
7183 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7184 list to exclude them. This means that no special compilation option
7185 is needed to use anonymous DH: it just needs to be included in the
7186 cipher list.
7187 [Steve Henson]
7188
7189 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7190 EVP_MD_type. The old functionality is available in a new macro called
7191 EVP_MD_md(). Change code that uses it and update docs.
7192 [Steve Henson]
7193
7194 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7195 where the 'void *' argument is replaced by a function pointer argument.
7196 Previously 'void *' was abused to point to functions, which works on
7197 many platforms, but is not correct. As these functions are usually
7198 called by macros defined in OpenSSL header files, most source code
7199 should work without changes.
7200 [Richard Levitte]
7201
7202 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7203 sections with information on -D... compiler switches used for
7204 compiling the library so that applications can see them. To enable
7205 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7206 must be defined. E.g.,
7207 #define OPENSSL_ALGORITHM_DEFINES
7208 #include <openssl/opensslconf.h>
7209 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7210 [Richard Levitte, Ulf and Bodo Möller]
7211
7212 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7213 record layer.
7214 [Bodo Moeller]
7215
7216 *) Change the 'other' type in certificate aux info to a STACK_OF
7217 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7218 the required ASN1 format: arbitrary types determined by an OID.
7219 [Steve Henson]
7220
7221 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7222 argument to 'req'. This is not because the function is newer or
7223 better than others it just uses the work 'NEW' in the certificate
7224 request header lines. Some software needs this.
7225 [Steve Henson]
7226
7227 *) Reorganise password command line arguments: now passwords can be
7228 obtained from various sources. Delete the PEM_cb function and make
7229 it the default behaviour: i.e. if the callback is NULL and the
7230 usrdata argument is not NULL interpret it as a null terminated pass
7231 phrase. If usrdata and the callback are NULL then the pass phrase
7232 is prompted for as usual.
7233 [Steve Henson]
7234
7235 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7236 the support is automatically enabled. The resulting binaries will
7237 autodetect the card and use it if present.
7238 [Ben Laurie and Compaq Inc.]
7239
7240 *) Work around for Netscape hang bug. This sends certificate request
7241 and server done in one record. Since this is perfectly legal in the
7242 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7243 the bugs/SSLv3 entry for more info.
7244 [Steve Henson]
7245
7246 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7247 [Andy Polyakov]
7248
7249 *) Add -rand argument to smime and pkcs12 applications and read/write
7250 of seed file.
7251 [Steve Henson]
7252
7253 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7254 [Bodo Moeller]
7255
7256 *) Add command line password options to the remaining applications.
7257 [Steve Henson]
7258
7259 *) Bug fix for BN_div_recp() for numerators with an even number of
7260 bits.
7261 [Ulf Möller]
7262
7263 *) More tests in bntest.c, and changed test_bn output.
7264 [Ulf Möller]
7265
7266 *) ./config recognizes MacOS X now.
7267 [Andy Polyakov]
7268
7269 *) Bug fix for BN_div() when the first words of num and divsor are
7270 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7271 [Ulf Möller]
7272
7273 *) Add support for various broken PKCS#8 formats, and command line
7274 options to produce them.
7275 [Steve Henson]
7276
7277 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7278 get temporary BIGNUMs from a BN_CTX.
7279 [Ulf Möller]
7280
7281 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7282 for p == 0.
7283 [Ulf Möller]
7284
7285 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7286 include a #define from the old name to the new. The original intent
7287 was that statically linked binaries could for example just call
7288 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7289 link with digests. This never worked becayse SSLeay_add_all_digests()
7290 and SSLeay_add_all_ciphers() were in the same source file so calling
7291 one would link with the other. They are now in separate source files.
7292 [Steve Henson]
7293
7294 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7295 [Steve Henson]
7296
7297 *) Use a less unusual form of the Miller-Rabin primality test (it used
7298 a binary algorithm for exponentiation integrated into the Miller-Rabin
7299 loop, our standard modexp algorithms are faster).
7300 [Bodo Moeller]
7301
7302 *) Support for the EBCDIC character set completed.
7303 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7304
7305 *) Source code cleanups: use const where appropriate, eliminate casts,
7306 use void * instead of char * in lhash.
7307 [Ulf Möller]
7308
7309 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7310 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7311 this the server could overwrite ephemeral keys that the client
7312 has already seen).
7313 [Bodo Moeller]
7314
7315 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7316 using 50 iterations of the Rabin-Miller test.
7317
7318 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7319 iterations of the Rabin-Miller test as required by the appendix
7320 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7321 As BN_is_prime_fasttest includes trial division, DSA parameter
7322 generation becomes much faster.
7323
7324 This implies a change for the callback functions in DSA_is_prime
7325 and DSA_generate_parameters: The callback function is called once
7326 for each positive witness in the Rabin-Miller test, not just
7327 occasionally in the inner loop; and the parameters to the
7328 callback function now provide an iteration count for the outer
7329 loop rather than for the current invocation of the inner loop.
7330 DSA_generate_parameters additionally can call the callback
7331 function with an 'iteration count' of -1, meaning that a
7332 candidate has passed the trial division test (when q is generated
7333 from an application-provided seed, trial division is skipped).
7334 [Bodo Moeller]
7335
7336 *) New function BN_is_prime_fasttest that optionally does trial
7337 division before starting the Rabin-Miller test and has
7338 an additional BN_CTX * argument (whereas BN_is_prime always
7339 has to allocate at least one BN_CTX).
7340 'callback(1, -1, cb_arg)' is called when a number has passed the
7341 trial division stage.
7342 [Bodo Moeller]
7343
7344 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7345 as ASN1_TIME.
7346 [Steve Henson]
7347
7348 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7349 [Steve Henson]
7350
7351 *) New function BN_pseudo_rand().
7352 [Ulf Möller]
7353
7354 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7355 bignum version of BN_from_montgomery() with the working code from
7356 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7357 the comments.
7358 [Ulf Möller]
7359
7360 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7361 made it impossible to use the same SSL_SESSION data structure in
7362 SSL2 clients in multiple threads.
7363 [Bodo Moeller]
7364
7365 *) The return value of RAND_load_file() no longer counts bytes obtained
7366 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7367 to seed the PRNG (previously an explicit byte count was required).
7368 [Ulf Möller, Bodo Möller]
7369
7370 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7371 used (char *) instead of (void *) and had casts all over the place.
7372 [Steve Henson]
7373
7374 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7375 [Ulf Möller]
7376
7377 *) Retain source code compatibility for BN_prime_checks macro:
7378 BN_is_prime(..., BN_prime_checks, ...) now uses
7379 BN_prime_checks_for_size to determine the appropriate number of
7380 Rabin-Miller iterations.
7381 [Ulf Möller]
7382
7383 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7384 DH_CHECK_P_NOT_SAFE_PRIME.
7385 (Check if this is true? OpenPGP calls them "strong".)
7386 [Ulf Möller]
7387
7388 *) Merge the functionality of "dh" and "gendh" programs into a new program
7389 "dhparam". The old programs are retained for now but will handle DH keys
7390 (instead of parameters) in future.
7391 [Steve Henson]
7392
7393 *) Make the ciphers, s_server and s_client programs check the return values
7394 when a new cipher list is set.
7395 [Steve Henson]
7396
7397 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7398 ciphers. Before when the 56bit ciphers were enabled the sorting was
7399 wrong.
7400
7401 The syntax for the cipher sorting has been extended to support sorting by
7402 cipher-strength (using the strength_bits hard coded in the tables).
7403 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7404
7405 Fix a bug in the cipher-command parser: when supplying a cipher command
7406 string with an "undefined" symbol (neither command nor alphanumeric
7407 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7408 an error is flagged.
7409
7410 Due to the strength-sorting extension, the code of the
7411 ssl_create_cipher_list() function was completely rearranged. I hope that
7412 the readability was also increased :-)
7413 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7414
7415 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7416 for the first serial number and places 2 in the serial number file. This
7417 avoids problems when the root CA is created with serial number zero and
7418 the first user certificate has the same issuer name and serial number
7419 as the root CA.
7420 [Steve Henson]
7421
7422 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7423 the new code. Add documentation for this stuff.
7424 [Steve Henson]
7425
7426 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7427 X509_*() to X509at_*() on the grounds that they don't handle X509
7428 structures and behave in an analagous way to the X509v3 functions:
7429 they shouldn't be called directly but wrapper functions should be used
7430 instead.
7431
7432 So we also now have some wrapper functions that call the X509at functions
7433 when passed certificate requests. (TO DO: similar things can be done with
7434 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7435 things. Some of these need some d2i or i2d and print functionality
7436 because they handle more complex structures.)
7437 [Steve Henson]
7438
7439 *) Add missing #ifndefs that caused missing symbols when building libssl
7440 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7441 NO_RSA in ssl/s2*.c.
7442 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7443
7444 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7445 has a return value which indicates the quality of the random data
7446 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7447 error queue. New function RAND_pseudo_bytes() generates output that is
7448 guaranteed to be unique but not unpredictable. RAND_add is like
7449 RAND_seed, but takes an extra argument for an entropy estimate
7450 (RAND_seed always assumes full entropy).
7451 [Ulf Möller]
7452
7453 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7454 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7455 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7456 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7457 false-positive rate of at most 2^-80 for random input.
7458 [Bodo Moeller]
7459
7460 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7461 [Bodo Moeller]
7462
7463 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7464 in the 0.9.5 release), this returns the chain
7465 from an X509_CTX structure with a dup of the stack and all
7466 the X509 reference counts upped: so the stack will exist
7467 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7468 to use this.
7469
7470 Also make SSL_SESSION_print() print out the verify return
7471 code.
7472 [Steve Henson]
7473
7474 *) Add manpage for the pkcs12 command. Also change the default
7475 behaviour so MAC iteration counts are used unless the new
7476 -nomaciter option is used. This improves file security and
7477 only older versions of MSIE (4.0 for example) need it.
7478 [Steve Henson]
7479
7480 *) Honor the no-xxx Configure options when creating .DEF files.
7481 [Ulf Möller]
7482
7483 *) Add PKCS#10 attributes to field table: challengePassword,
7484 unstructuredName and unstructuredAddress. These are taken from
7485 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7486 international characters are used.
7487
7488 More changes to X509_ATTRIBUTE code: allow the setting of types
7489 based on strings. Remove the 'loc' parameter when adding
7490 attributes because these will be a SET OF encoding which is sorted
7491 in ASN1 order.
7492 [Steve Henson]
7493
7494 *) Initial changes to the 'req' utility to allow request generation
7495 automation. This will allow an application to just generate a template
7496 file containing all the field values and have req construct the
7497 request.
7498
7499 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7500 used all over the place including certificate requests and PKCS#7
7501 structures. They are currently handled manually where necessary with
7502 some primitive wrappers for PKCS#7. The new functions behave in a
7503 manner analogous to the X509 extension functions: they allow
7504 attributes to be looked up by NID and added.
7505
7506 Later something similar to the X509V3 code would be desirable to
7507 automatically handle the encoding, decoding and printing of the
7508 more complex types. The string types like challengePassword can
7509 be handled by the string table functions.
7510
7511 Also modified the multi byte string table handling. Now there is
7512 a 'global mask' which masks out certain types. The table itself
7513 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7514 is useful when for example there is only one permissible type
7515 (as in countryName) and using the mask might result in no valid
7516 types at all.
7517 [Steve Henson]
7518
7519 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7520 SSL_get_peer_finished to allow applications to obtain the latest
7521 Finished messages sent to the peer or expected from the peer,
7522 respectively. (SSL_get_peer_finished is usually the Finished message
7523 actually received from the peer, otherwise the protocol will be aborted.)
7524
7525 As the Finished message are message digests of the complete handshake
7526 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7527 be used for external authentication procedures when the authentication
7528 provided by SSL/TLS is not desired or is not enough.
7529 [Bodo Moeller]
7530
7531 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7532 the host supports BWX extension and if Compaq C is present on the
7533 $PATH. Just exploiting of the BWX extension results in 20-30%
7534 performance kick for some algorithms, e.g. DES and RC4 to mention
7535 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7536 SHA1.
7537 [Andy Polyakov]
7538
7539 *) Add support for MS "fast SGC". This is arguably a violation of the
7540 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7541 weak crypto and after checking the certificate is SGC a second one
7542 with strong crypto. MS SGC stops the first handshake after receiving
7543 the server certificate message and sends a second client hello. Since
7544 a server will typically do all the time consuming operations before
7545 expecting any further messages from the client (server key exchange
7546 is the most expensive) there is little difference between the two.
7547
7548 To get OpenSSL to support MS SGC we have to permit a second client
7549 hello message after we have sent server done. In addition we have to
7550 reset the MAC if we do get this second client hello.
7551 [Steve Henson]
7552
7553 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7554 if a DER encoded private key is RSA or DSA traditional format. Changed
7555 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7556 format DER encoded private key. Newer code should use PKCS#8 format which
7557 has the key type encoded in the ASN1 structure. Added DER private key
7558 support to pkcs8 application.
7559 [Steve Henson]
7560
7561 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7562 ciphersuites has been selected (as required by the SSL 3/TLS 1
7563 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7564 is set, we interpret this as a request to violate the specification
7565 (the worst that can happen is a handshake failure, and 'correct'
7566 behaviour would result in a handshake failure anyway).
7567 [Bodo Moeller]
7568
7569 *) In SSL_CTX_add_session, take into account that there might be multiple
7570 SSL_SESSION structures with the same session ID (e.g. when two threads
7571 concurrently obtain them from an external cache).
7572 The internal cache can handle only one SSL_SESSION with a given ID,
7573 so if there's a conflict, we now throw out the old one to achieve
7574 consistency.
7575 [Bodo Moeller]
7576
7577 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7578 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7579 some routines that use cipher OIDs: some ciphers do not have OIDs
7580 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7581 example.
7582 [Steve Henson]
7583
7584 *) Simplify the trust setting structure and code. Now we just have
7585 two sequences of OIDs for trusted and rejected settings. These will
7586 typically have values the same as the extended key usage extension
7587 and any application specific purposes.
7588
7589 The trust checking code now has a default behaviour: it will just
7590 check for an object with the same NID as the passed id. Functions can
7591 be provided to override either the default behaviour or the behaviour
7592 for a given id. SSL client, server and email already have functions
7593 in place for compatibility: they check the NID and also return "trusted"
7594 if the certificate is self signed.
7595 [Steve Henson]
7596
7597 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7598 traditional format into an EVP_PKEY structure.
7599 [Steve Henson]
7600
7601 *) Add a password callback function PEM_cb() which either prompts for
7602 a password if usr_data is NULL or otherwise assumes it is a null
7603 terminated password. Allow passwords to be passed on command line
7604 environment or config files in a few more utilities.
7605 [Steve Henson]
7606
7607 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7608 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7609 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7610 Update documentation.
7611 [Steve Henson]
7612
7613 *) Support for ASN1 "NULL" type. This could be handled before by using
7614 ASN1_TYPE but there wasn't any function that would try to read a NULL
7615 and produce an error if it couldn't. For compatibility we also have
7616 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7617 don't allocate anything because they don't need to.
7618 [Steve Henson]
7619
7620 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7621 for details.
7622 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7623
7624 *) Rebuild of the memory allocation routines used by OpenSSL code and
7625 possibly others as well. The purpose is to make an interface that
7626 provide hooks so anyone can build a separate set of allocation and
7627 deallocation routines to be used by OpenSSL, for example memory
7628 pool implementations, or something else, which was previously hard
7629 since Malloc(), Realloc() and Free() were defined as macros having
7630 the values malloc, realloc and free, respectively (except for Win32
7631 compilations). The same is provided for memory debugging code.
7632 OpenSSL already comes with functionality to find memory leaks, but
7633 this gives people a chance to debug other memory problems.
7634
7635 With these changes, a new set of functions and macros have appeared:
7636
7637 CRYPTO_set_mem_debug_functions() [F]
7638 CRYPTO_get_mem_debug_functions() [F]
7639 CRYPTO_dbg_set_options() [F]
7640 CRYPTO_dbg_get_options() [F]
7641 CRYPTO_malloc_debug_init() [M]
7642
7643 The memory debug functions are NULL by default, unless the library
7644 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7645 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7646 gives the standard debugging functions that come with OpenSSL) or
7647 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7648 provided by the library user) must be used. When the standard
7649 debugging functions are used, CRYPTO_dbg_set_options can be used to
7650 request additional information:
7651 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7652 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7653
7654 Also, things like CRYPTO_set_mem_functions will always give the
7655 expected result (the new set of functions is used for allocation
7656 and deallocation) at all times, regardless of platform and compiler
7657 options.
7658
7659 To finish it up, some functions that were never use in any other
7660 way than through macros have a new API and new semantic:
7661
7662 CRYPTO_dbg_malloc()
7663 CRYPTO_dbg_realloc()
7664 CRYPTO_dbg_free()
7665
7666 All macros of value have retained their old syntax.
7667 [Richard Levitte and Bodo Moeller]
7668
7669 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7670 ordering of SMIMECapabilities wasn't in "strength order" and there
7671 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7672 algorithm.
7673 [Steve Henson]
7674
7675 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7676 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7677 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7678
7679 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7680 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7681 functionality to handle multipart/signed properly) and a utility
7682 called 'smime' to call all this stuff. This is based on code I
7683 originally wrote for Celo who have kindly allowed it to be
7684 included in OpenSSL.
7685 [Steve Henson]
7686
7687 *) Add variants des_set_key_checked and des_set_key_unchecked of
7688 des_set_key (aka des_key_sched). Global variable des_check_key
7689 decides which of these is called by des_set_key; this way
7690 des_check_key behaves as it always did, but applications and
7691 the library itself, which was buggy for des_check_key == 1,
7692 have a cleaner way to pick the version they need.
7693 [Bodo Moeller]
7694
7695 *) New function PKCS12_newpass() which changes the password of a
7696 PKCS12 structure.
7697 [Steve Henson]
7698
7699 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7700 dynamic mix. In both cases the ids can be used as an index into the
7701 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7702 functions so they accept a list of the field values and the
7703 application doesn't need to directly manipulate the X509_TRUST
7704 structure.
7705 [Steve Henson]
7706
7707 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7708 need initialising.
7709 [Steve Henson]
7710
7711 *) Modify the way the V3 extension code looks up extensions. This now
7712 works in a similar way to the object code: we have some "standard"
7713 extensions in a static table which is searched with OBJ_bsearch()
7714 and the application can add dynamic ones if needed. The file
7715 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7716 updated whenever a new extension is added to the core code and kept
7717 in ext_nid order. There is a simple program 'tabtest.c' which checks
7718 this. New extensions are not added too often so this file can readily
7719 be maintained manually.
7720
7721 There are two big advantages in doing things this way. The extensions
7722 can be looked up immediately and no longer need to be "added" using
7723 X509V3_add_standard_extensions(): this function now does nothing.
7724 [Side note: I get *lots* of email saying the extension code doesn't
7725 work because people forget to call this function]
7726 Also no dynamic allocation is done unless new extensions are added:
7727 so if we don't add custom extensions there is no need to call
7728 X509V3_EXT_cleanup().
7729 [Steve Henson]
7730
7731 *) Modify enc utility's salting as follows: make salting the default. Add a
7732 magic header, so unsalted files fail gracefully instead of just decrypting
7733 to garbage. This is because not salting is a big security hole, so people
7734 should be discouraged from doing it.
7735 [Ben Laurie]
7736
7737 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7738 digest to be passed on the command line but it only used this
7739 parameter when signing a certificate. Modified so all relevant
7740 operations are affected by the digest parameter including the
7741 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7742 DSA key was used because it didn't fix the digest.
7743 [Steve Henson]
7744
7745 *) Initial certificate chain verify code. Currently tests the untrusted
7746 certificates for consistency with the verify purpose (which is set
7747 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7748
7749 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7750 this is because it will reject chains with invalid extensions whereas
7751 every previous version of OpenSSL and SSLeay made no checks at all.
7752
7753 Trust code: checks the root CA for the relevant trust settings. Trust
7754 settings have an initial value consistent with the verify purpose: e.g.
7755 if the verify purpose is for SSL client use it expects the CA to be
7756 trusted for SSL client use. However the default value can be changed to
7757 permit custom trust settings: one example of this would be to only trust
7758 certificates from a specific "secure" set of CAs.
7759
7760 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7761 which should be used for version portability: especially since the
7762 verify structure is likely to change more often now.
7763
7764 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7765 to set them. If not set then assume SSL clients will verify SSL servers
7766 and vice versa.
7767
7768 Two new options to the verify program: -untrusted allows a set of
7769 untrusted certificates to be passed in and -purpose which sets the
7770 intended purpose of the certificate. If a purpose is set then the
7771 new chain verify code is used to check extension consistency.
7772 [Steve Henson]
7773
7774 *) Support for the authority information access extension.
7775 [Steve Henson]
7776
7777 *) Modify RSA and DSA PEM read routines to transparently handle
7778 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7779 public keys in a format compatible with certificate
7780 SubjectPublicKeyInfo structures. Unfortunately there were already
7781 functions called *_PublicKey_* which used various odd formats so
7782 these are retained for compatibility: however the DSA variants were
7783 never in a public release so they have been deleted. Changed dsa/rsa
7784 utilities to handle the new format: note no releases ever handled public
7785 keys so we should be OK.
7786
7787 The primary motivation for this change is to avoid the same fiasco
7788 that dogs private keys: there are several incompatible private key
7789 formats some of which are standard and some OpenSSL specific and
7790 require various evil hacks to allow partial transparent handling and
7791 even then it doesn't work with DER formats. Given the option anything
7792 other than PKCS#8 should be dumped: but the other formats have to
7793 stay in the name of compatibility.
7794
7795 With public keys and the benefit of hindsight one standard format
7796 is used which works with EVP_PKEY, RSA or DSA structures: though
7797 it clearly returns an error if you try to read the wrong kind of key.
7798
7799 Added a -pubkey option to the 'x509' utility to output the public key.
7800 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7801 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7802 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7803 that do the same as the EVP_PKEY_assign_*() except they up the
7804 reference count of the added key (they don't "swallow" the
7805 supplied key).
7806 [Steve Henson]
7807
7808 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7809 CRLs would fail if the file contained no certificates or no CRLs:
7810 added a new function to read in both types and return the number
7811 read: this means that if none are read it will be an error. The
7812 DER versions of the certificate and CRL reader would always fail
7813 because it isn't possible to mix certificates and CRLs in DER format
7814 without choking one or the other routine. Changed this to just read
7815 a certificate: this is the best we can do. Also modified the code
7816 in apps/verify.c to take notice of return codes: it was previously
7817 attempting to read in certificates from NULL pointers and ignoring
7818 any errors: this is one reason why the cert and CRL reader seemed
7819 to work. It doesn't check return codes from the default certificate
7820 routines: these may well fail if the certificates aren't installed.
7821 [Steve Henson]
7822
7823 *) Code to support otherName option in GeneralName.
7824 [Steve Henson]
7825
7826 *) First update to verify code. Change the verify utility
7827 so it warns if it is passed a self signed certificate:
7828 for consistency with the normal behaviour. X509_verify
7829 has been modified to it will now verify a self signed
7830 certificate if *exactly* the same certificate appears
7831 in the store: it was previously impossible to trust a
7832 single self signed certificate. This means that:
7833 openssl verify ss.pem
7834 now gives a warning about a self signed certificate but
7835 openssl verify -CAfile ss.pem ss.pem
7836 is OK.
7837 [Steve Henson]
7838
7839 *) For servers, store verify_result in SSL_SESSION data structure
7840 (and add it to external session representation).
7841 This is needed when client certificate verifications fails,
7842 but an application-provided verification callback (set by
7843 SSL_CTX_set_cert_verify_callback) allows accepting the session
7844 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7845 but returns 1): When the session is reused, we have to set
7846 ssl->verify_result to the appropriate error code to avoid
7847 security holes.
7848 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7849
7850 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7851 case in PKCS7_dataInit() where the signed PKCS7 structure
7852 didn't contain any existing data because it was being created.
7853 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7854
7855 *) Add a salt to the key derivation routines in enc.c. This
7856 forms the first 8 bytes of the encrypted file. Also add a
7857 -S option to allow a salt to be input on the command line.
7858 [Steve Henson]
7859
7860 *) New function X509_cmp(). Oddly enough there wasn't a function
7861 to compare two certificates. We do this by working out the SHA1
7862 hash and comparing that. X509_cmp() will be needed by the trust
7863 code.
7864 [Steve Henson]
7865
7866 *) SSL_get1_session() is like SSL_get_session(), but increments
7867 the reference count in the SSL_SESSION returned.
7868 [Geoff Thorpe <geoff@eu.c2.net>]
7869
7870 *) Fix for 'req': it was adding a null to request attributes.
7871 Also change the X509_LOOKUP and X509_INFO code to handle
7872 certificate auxiliary information.
7873 [Steve Henson]
7874
7875 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7876 the 'enc' command.
7877 [Steve Henson]
7878
7879 *) Add the possibility to add extra information to the memory leak
7880 detecting output, to form tracebacks, showing from where each
7881 allocation was originated: CRYPTO_push_info("constant string") adds
7882 the string plus current file name and line number to a per-thread
7883 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7884 is like calling CYRPTO_pop_info() until the stack is empty.
7885 Also updated memory leak detection code to be multi-thread-safe.
7886 [Richard Levitte]
7887
7888 *) Add options -text and -noout to pkcs7 utility and delete the
7889 encryption options which never did anything. Update docs.
7890 [Steve Henson]
7891
7892 *) Add options to some of the utilities to allow the pass phrase
7893 to be included on either the command line (not recommended on
7894 OSes like Unix) or read from the environment. Update the
7895 manpages and fix a few bugs.
7896 [Steve Henson]
7897
7898 *) Add a few manpages for some of the openssl commands.
7899 [Steve Henson]
7900
7901 *) Fix the -revoke option in ca. It was freeing up memory twice,
7902 leaking and not finding already revoked certificates.
7903 [Steve Henson]
7904
7905 *) Extensive changes to support certificate auxiliary information.
7906 This involves the use of X509_CERT_AUX structure and X509_AUX
7907 functions. An X509_AUX function such as PEM_read_X509_AUX()
7908 can still read in a certificate file in the usual way but it
7909 will also read in any additional "auxiliary information". By
7910 doing things this way a fair degree of compatibility can be
7911 retained: existing certificates can have this information added
7912 using the new 'x509' options.
7913
7914 Current auxiliary information includes an "alias" and some trust
7915 settings. The trust settings will ultimately be used in enhanced
7916 certificate chain verification routines: currently a certificate
7917 can only be trusted if it is self signed and then it is trusted
7918 for all purposes.
7919 [Steve Henson]
7920
7921 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7922 The problem was that one of the replacement routines had not been working
7923 since SSLeay releases. For now the offending routine has been replaced
7924 with non-optimised assembler. Even so, this now gives around 95%
7925 performance improvement for 1024 bit RSA signs.
7926 [Mark Cox]
7927
7928 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7929 handling. Most clients have the effective key size in bits equal to
7930 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7931 A few however don't do this and instead use the size of the decrypted key
7932 to determine the RC2 key length and the AlgorithmIdentifier to determine
7933 the effective key length. In this case the effective key length can still
7934 be 40 bits but the key length can be 168 bits for example. This is fixed
7935 by manually forcing an RC2 key into the EVP_PKEY structure because the
7936 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7937 the key length and effective key length are equal.
7938 [Steve Henson]
7939
7940 *) Add a bunch of functions that should simplify the creation of
7941 X509_NAME structures. Now you should be able to do:
7942 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7943 and have it automatically work out the correct field type and fill in
7944 the structures. The more adventurous can try:
7945 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7946 and it will (hopefully) work out the correct multibyte encoding.
7947 [Steve Henson]
7948
7949 *) Change the 'req' utility to use the new field handling and multibyte
7950 copy routines. Before the DN field creation was handled in an ad hoc
7951 way in req, ca, and x509 which was rather broken and didn't support
7952 BMPStrings or UTF8Strings. Since some software doesn't implement
7953 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7954 using the dirstring_type option. See the new comment in the default
7955 openssl.cnf for more info.
7956 [Steve Henson]
7957
7958 *) Make crypto/rand/md_rand.c more robust:
7959 - Assure unique random numbers after fork().
7960 - Make sure that concurrent threads access the global counter and
7961 md serializably so that we never lose entropy in them
7962 or use exactly the same state in multiple threads.
7963 Access to the large state is not always serializable because
7964 the additional locking could be a performance killer, and
7965 md should be large enough anyway.
7966 [Bodo Moeller]
7967
7968 *) New file apps/app_rand.c with commonly needed functionality
7969 for handling the random seed file.
7970
7971 Use the random seed file in some applications that previously did not:
7972 ca,
7973 dsaparam -genkey (which also ignored its '-rand' option),
7974 s_client,
7975 s_server,
7976 x509 (when signing).
7977 Except on systems with /dev/urandom, it is crucial to have a random
7978 seed file at least for key creation, DSA signing, and for DH exchanges;
7979 for RSA signatures we could do without one.
7980
7981 gendh and gendsa (unlike genrsa) used to read only the first byte
7982 of each file listed in the '-rand' option. The function as previously
7983 found in genrsa is now in app_rand.c and is used by all programs
7984 that support '-rand'.
7985 [Bodo Moeller]
7986
7987 *) In RAND_write_file, use mode 0600 for creating files;
7988 don't just chmod when it may be too late.
7989 [Bodo Moeller]
7990
7991 *) Report an error from X509_STORE_load_locations
7992 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7993 [Bill Perry]
7994
7995 *) New function ASN1_mbstring_copy() this copies a string in either
7996 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7997 into an ASN1_STRING type. A mask of permissible types is passed
7998 and it chooses the "minimal" type to use or an error if not type
7999 is suitable.
8000 [Steve Henson]
8001
8002 *) Add function equivalents to the various macros in asn1.h. The old
8003 macros are retained with an M_ prefix. Code inside the library can
8004 use the M_ macros. External code (including the openssl utility)
8005 should *NOT* in order to be "shared library friendly".
8006 [Steve Henson]
8007
8008 *) Add various functions that can check a certificate's extensions
8009 to see if it usable for various purposes such as SSL client,
8010 server or S/MIME and CAs of these types. This is currently
8011 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8012 verification. Also added a -purpose flag to x509 utility to
8013 print out all the purposes.
8014 [Steve Henson]
8015
8016 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8017 functions.
8018 [Steve Henson]
8019
8020 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8021 for, obtain and decode and extension and obtain its critical flag.
8022 This allows all the necessary extension code to be handled in a
8023 single function call.
8024 [Steve Henson]
8025
8026 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8027 platforms. See crypto/rc4/rc4_enc.c for further details.
8028 [Andy Polyakov]
8029
8030 *) New -noout option to asn1parse. This causes no output to be produced
8031 its main use is when combined with -strparse and -out to extract data
8032 from a file (which may not be in ASN.1 format).
8033 [Steve Henson]
8034
8035 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8036 when producing the local key id.
8037 [Richard Levitte <levitte@stacken.kth.se>]
8038
8039 *) New option -dhparam in s_server. This allows a DH parameter file to be
8040 stated explicitly. If it is not stated then it tries the first server
8041 certificate file. The previous behaviour hard coded the filename
8042 "server.pem".
8043 [Steve Henson]
8044
8045 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8046 a public key to be input or output. For example:
8047 openssl rsa -in key.pem -pubout -out pubkey.pem
8048 Also added necessary DSA public key functions to handle this.
8049 [Steve Henson]
8050
8051 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8052 in the message. This was handled by allowing
8053 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8054 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8055
8056 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8057 to the end of the strings whereas this didn't. This would cause problems
8058 if strings read with d2i_ASN1_bytes() were later modified.
8059 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8060
8061 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8062 data and it contains EOF it will end up returning an error. This is
8063 caused by input 46 bytes long. The cause is due to the way base64
8064 BIOs find the start of base64 encoded data. They do this by trying a
8065 trial decode on each line until they find one that works. When they
8066 do a flag is set and it starts again knowing it can pass all the
8067 data directly through the decoder. Unfortunately it doesn't reset
8068 the context it uses. This means that if EOF is reached an attempt
8069 is made to pass two EOFs through the context and this causes the
8070 resulting error. This can also cause other problems as well. As is
8071 usual with these problems it takes *ages* to find and the fix is
8072 trivial: move one line.
8073 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8074
8075 *) Ugly workaround to get s_client and s_server working under Windows. The
8076 old code wouldn't work because it needed to select() on sockets and the
8077 tty (for keypresses and to see if data could be written). Win32 only
8078 supports select() on sockets so we select() with a 1s timeout on the
8079 sockets and then see if any characters are waiting to be read, if none
8080 are present then we retry, we also assume we can always write data to
8081 the tty. This isn't nice because the code then blocks until we've
8082 received a complete line of data and it is effectively polling the
8083 keyboard at 1s intervals: however it's quite a bit better than not
8084 working at all :-) A dedicated Windows application might handle this
8085 with an event loop for example.
8086 [Steve Henson]
8087
8088 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8089 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8090 will be called when RSA_sign() and RSA_verify() are used. This is useful
8091 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8092 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8093 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8094 This necessitated the support of an extra signature type NID_md5_sha1
8095 for SSL signatures and modifications to the SSL library to use it instead
8096 of calling RSA_public_decrypt() and RSA_private_encrypt().
8097 [Steve Henson]
8098
8099 *) Add new -verify -CAfile and -CApath options to the crl program, these
8100 will lookup a CRL issuers certificate and verify the signature in a
8101 similar way to the verify program. Tidy up the crl program so it
8102 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8103 less strict. It will now permit CRL extensions even if it is not
8104 a V2 CRL: this will allow it to tolerate some broken CRLs.
8105 [Steve Henson]
8106
8107 *) Initialize all non-automatic variables each time one of the openssl
8108 sub-programs is started (this is necessary as they may be started
8109 multiple times from the "OpenSSL>" prompt).
8110 [Lennart Bang, Bodo Moeller]
8111
8112 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8113 removing all other RSA functionality (this is what NO_RSA does). This
8114 is so (for example) those in the US can disable those operations covered
8115 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8116 key generation.
8117 [Steve Henson]
8118
8119 *) Non-copying interface to BIO pairs.
8120 (still largely untested)
8121 [Bodo Moeller]
8122
8123 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8124 ASCII string. This was handled independently in various places before.
8125 [Steve Henson]
8126
8127 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8128 UTF8 strings a character at a time.
8129 [Steve Henson]
8130
8131 *) Use client_version from client hello to select the protocol
8132 (s23_srvr.c) and for RSA client key exchange verification
8133 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8134 [Bodo Moeller]
8135
8136 *) Add various utility functions to handle SPKACs, these were previously
8137 handled by poking round in the structure internals. Added new function
8138 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8139 print, verify and generate SPKACs. Based on an original idea from
8140 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8141 [Steve Henson]
8142
8143 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8144 [Andy Polyakov]
8145
8146 *) Allow the config file extension section to be overwritten on the
8147 command line. Based on an original idea from Massimiliano Pala
8148 <madwolf@comune.modena.it>. The new option is called -extensions
8149 and can be applied to ca, req and x509. Also -reqexts to override
8150 the request extensions in req and -crlexts to override the crl extensions
8151 in ca.
8152 [Steve Henson]
8153
8154 *) Add new feature to the SPKAC handling in ca. Now you can include
8155 the same field multiple times by preceding it by "XXXX." for example:
8156 1.OU="Unit name 1"
8157 2.OU="Unit name 2"
8158 this is the same syntax as used in the req config file.
8159 [Steve Henson]
8160
8161 *) Allow certificate extensions to be added to certificate requests. These
8162 are specified in a 'req_extensions' option of the req section of the
8163 config file. They can be printed out with the -text option to req but
8164 are otherwise ignored at present.
8165 [Steve Henson]
8166
8167 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8168 data read consists of only the final block it would not decrypted because
8169 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8170 A misplaced 'break' also meant the decrypted final block might not be
8171 copied until the next read.
8172 [Steve Henson]
8173
8174 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8175 a few extra parameters to the DH structure: these will be useful if
8176 for example we want the value of 'q' or implement X9.42 DH.
8177 [Steve Henson]
8178
8179 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8180 provides hooks that allow the default DSA functions or functions on a
8181 "per key" basis to be replaced. This allows hardware acceleration and
8182 hardware key storage to be handled without major modification to the
8183 library. Also added low level modexp hooks and CRYPTO_EX structure and
8184 associated functions.
8185 [Steve Henson]
8186
8187 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8188 as "read only": it can't be written to and the buffer it points to will
8189 not be freed. Reading from a read only BIO is much more efficient than
8190 a normal memory BIO. This was added because there are several times when
8191 an area of memory needs to be read from a BIO. The previous method was
8192 to create a memory BIO and write the data to it, this results in two
8193 copies of the data and an O(n^2) reading algorithm. There is a new
8194 function BIO_new_mem_buf() which creates a read only memory BIO from
8195 an area of memory. Also modified the PKCS#7 routines to use read only
8196 memory BIOs.
8197 [Steve Henson]
8198
8199 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8200 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8201 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8202 but a retry condition occured while trying to read the rest.
8203 [Bodo Moeller]
8204
8205 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8206 NID_pkcs7_encrypted by default: this was wrong since this should almost
8207 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8208 the encrypted data type: this is a more sensible place to put it and it
8209 allows the PKCS#12 code to be tidied up that duplicated this
8210 functionality.
8211 [Steve Henson]
8212
8213 *) Changed obj_dat.pl script so it takes its input and output files on
8214 the command line. This should avoid shell escape redirection problems
8215 under Win32.
8216 [Steve Henson]
8217
8218 *) Initial support for certificate extension requests, these are included
8219 in things like Xenroll certificate requests. Included functions to allow
8220 extensions to be obtained and added.
8221 [Steve Henson]
8222
8223 *) -crlf option to s_client and s_server for sending newlines as
8224 CRLF (as required by many protocols).
8225 [Bodo Moeller]
8226
8227 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8228
8229 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8230 [Ralf S. Engelschall]
8231
8232 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8233 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8234
8235 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8236 program.
8237 [Steve Henson]
8238
8239 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8240 DH parameters/keys (q is lost during that conversion, but the resulting
8241 DH parameters contain its length).
8242
8243 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8244 much faster than DH_generate_parameters (which creates parameters
8245 where p = 2*q + 1), and also the smaller q makes DH computations
8246 much more efficient (160-bit exponentiation instead of 1024-bit
8247 exponentiation); so this provides a convenient way to support DHE
8248 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8249 utter importance to use
8250 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8251 or
8252 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8253 when such DH parameters are used, because otherwise small subgroup
8254 attacks may become possible!
8255 [Bodo Moeller]
8256
8257 *) Avoid memory leak in i2d_DHparams.
8258 [Bodo Moeller]
8259
8260 *) Allow the -k option to be used more than once in the enc program:
8261 this allows the same encrypted message to be read by multiple recipients.
8262 [Steve Henson]
8263
8264 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8265 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8266 it will always use the numerical form of the OID, even if it has a short
8267 or long name.
8268 [Steve Henson]
8269
8270 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8271 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8272 otherwise bn_mod_exp was called. In the case of hardware keys for example
8273 no private key components need be present and it might store extra data
8274 in the RSA structure, which cannot be accessed from bn_mod_exp.
8275 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8276 private key operations.
8277 [Steve Henson]
8278
8279 *) Added support for SPARC Linux.
8280 [Andy Polyakov]
8281
8282 *) pem_password_cb function type incompatibly changed from
8283 typedef int pem_password_cb(char *buf, int size, int rwflag);
8284 to
8285 ....(char *buf, int size, int rwflag, void *userdata);
8286 so that applications can pass data to their callbacks:
8287 The PEM[_ASN1]_{read,write}... functions and macros now take an
8288 additional void * argument, which is just handed through whenever
8289 the password callback is called.
8290 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8291
8292 New function SSL_CTX_set_default_passwd_cb_userdata.
8293
8294 Compatibility note: As many C implementations push function arguments
8295 onto the stack in reverse order, the new library version is likely to
8296 interoperate with programs that have been compiled with the old
8297 pem_password_cb definition (PEM_whatever takes some data that
8298 happens to be on the stack as its last argument, and the callback
8299 just ignores this garbage); but there is no guarantee whatsoever that
8300 this will work.
8301
8302 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8303 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8304 problems not only on Windows, but also on some Unix platforms.
8305 To avoid problematic command lines, these definitions are now in an
8306 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8307 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8308 [Bodo Moeller]
8309
8310 *) MIPS III/IV assembler module is reimplemented.
8311 [Andy Polyakov]
8312
8313 *) More DES library cleanups: remove references to srand/rand and
8314 delete an unused file.
8315 [Ulf Möller]
8316
8317 *) Add support for the the free Netwide assembler (NASM) under Win32,
8318 since not many people have MASM (ml) and it can be hard to obtain.
8319 This is currently experimental but it seems to work OK and pass all
8320 the tests. Check out INSTALL.W32 for info.
8321 [Steve Henson]
8322
8323 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8324 without temporary keys kept an extra copy of the server key,
8325 and connections with temporary keys did not free everything in case
8326 of an error.
8327 [Bodo Moeller]
8328
8329 *) New function RSA_check_key and new openssl rsa option -check
8330 for verifying the consistency of RSA keys.
8331 [Ulf Moeller, Bodo Moeller]
8332
8333 *) Various changes to make Win32 compile work:
8334 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8335 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8336 comparison" warnings.
8337 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8338 [Steve Henson]
8339
8340 *) Add a debugging option to PKCS#5 v2 key generation function: when
8341 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8342 derived keys are printed to stderr.
8343 [Steve Henson]
8344
8345 *) Copy the flags in ASN1_STRING_dup().
8346 [Roman E. Pavlov <pre@mo.msk.ru>]
8347
8348 *) The x509 application mishandled signing requests containing DSA
8349 keys when the signing key was also DSA and the parameters didn't match.
8350
8351 It was supposed to omit the parameters when they matched the signing key:
8352 the verifying software was then supposed to automatically use the CA's
8353 parameters if they were absent from the end user certificate.
8354
8355 Omitting parameters is no longer recommended. The test was also
8356 the wrong way round! This was probably due to unusual behaviour in
8357 EVP_cmp_parameters() which returns 1 if the parameters match.
8358 This meant that parameters were omitted when they *didn't* match and
8359 the certificate was useless. Certificates signed with 'ca' didn't have
8360 this bug.
8361 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8362
8363 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8364 The interface is as follows:
8365 Applications can use
8366 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8367 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8368 "off" is now the default.
8369 The library internally uses
8370 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8371 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8372 to disable memory-checking temporarily.
8373
8374 Some inconsistent states that previously were possible (and were
8375 even the default) are now avoided.
8376
8377 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8378 with each memory chunk allocated; this is occasionally more helpful
8379 than just having a counter.
8380
8381 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8382
8383 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8384 extensions.
8385 [Bodo Moeller]
8386
8387 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8388 which largely parallels "options", but is for changing API behaviour,
8389 whereas "options" are about protocol behaviour.
8390 Initial "mode" flags are:
8391
8392 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8393 a single record has been written.
8394 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8395 retries use the same buffer location.
8396 (But all of the contents must be
8397 copied!)
8398 [Bodo Moeller]
8399
8400 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8401 worked.
8402
8403 *) Fix problems with no-hmac etc.
8404 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8405
8406 *) New functions RSA_get_default_method(), RSA_set_method() and
8407 RSA_get_method(). These allows replacement of RSA_METHODs without having
8408 to mess around with the internals of an RSA structure.
8409 [Steve Henson]
8410
8411 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8412 Also really enable memory leak checks in openssl.c and in some
8413 test programs.
8414 [Chad C. Mulligan, Bodo Moeller]
8415
8416 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8417 up the length of negative integers. This has now been simplified to just
8418 store the length when it is first determined and use it later, rather
8419 than trying to keep track of where data is copied and updating it to
8420 point to the end.
8421 [Steve Henson, reported by Brien Wheeler
8422 <bwheeler@authentica-security.com>]
8423
8424 *) Add a new function PKCS7_signatureVerify. This allows the verification
8425 of a PKCS#7 signature but with the signing certificate passed to the
8426 function itself. This contrasts with PKCS7_dataVerify which assumes the
8427 certificate is present in the PKCS#7 structure. This isn't always the
8428 case: certificates can be omitted from a PKCS#7 structure and be
8429 distributed by "out of band" means (such as a certificate database).
8430 [Steve Henson]
8431
8432 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8433 function prototypes in pem.h, also change util/mkdef.pl to add the
8434 necessary function names.
8435 [Steve Henson]
8436
8437 *) mk1mf.pl (used by Windows builds) did not properly read the
8438 options set by Configure in the top level Makefile, and Configure
8439 was not even able to write more than one option correctly.
8440 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8441 [Bodo Moeller]
8442
8443 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8444 file to be loaded from a BIO or FILE pointer. The BIO version will
8445 for example allow memory BIOs to contain config info.
8446 [Steve Henson]
8447
8448 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8449 Whoever hopes to achieve shared-library compatibility across versions
8450 must use this, not the compile-time macro.
8451 (Exercise 0.9.4: Which is the minimum library version required by
8452 such programs?)
8453 Note: All this applies only to multi-threaded programs, others don't
8454 need locks.
8455 [Bodo Moeller]
8456
8457 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8458 through a BIO pair triggered the default case, i.e.
8459 SSLerr(...,SSL_R_UNKNOWN_STATE).
8460 [Bodo Moeller]
8461
8462 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8463 can use the SSL library even if none of the specific BIOs is
8464 appropriate.
8465 [Bodo Moeller]
8466
8467 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8468 for the encoded length.
8469 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8470
8471 *) Add initial documentation of the X509V3 functions.
8472 [Steve Henson]
8473
8474 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8475 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8476 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8477 secure PKCS#8 private key format with a high iteration count.
8478 [Steve Henson]
8479
8480 *) Fix determination of Perl interpreter: A perl or perl5
8481 _directory_ in $PATH was also accepted as the interpreter.
8482 [Ralf S. Engelschall]
8483
8484 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8485 wrong with it but it was very old and did things like calling
8486 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8487 unusual formatting.
8488 [Steve Henson]
8489
8490 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8491 to use the new extension code.
8492 [Steve Henson]
8493
8494 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8495 with macros. This should make it easier to change their form, add extra
8496 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8497 constant.
8498 [Steve Henson]
8499
8500 *) Add to configuration table a new entry that can specify an alternative
8501 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8502 according to Mark Crispin <MRC@Panda.COM>.
8503 [Bodo Moeller]
8504
8505 #if 0
8506 *) DES CBC did not update the IV. Weird.
8507 [Ben Laurie]
8508 #else
8509 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8510 Changing the behaviour of the former might break existing programs --
8511 where IV updating is needed, des_ncbc_encrypt can be used.
8512 #endif
8513
8514 *) When bntest is run from "make test" it drives bc to check its
8515 calculations, as well as internally checking them. If an internal check
8516 fails, it needs to cause bc to give a non-zero result or make test carries
8517 on without noticing the failure. Fixed.
8518 [Ben Laurie]
8519
8520 *) DES library cleanups.
8521 [Ulf Möller]
8522
8523 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8524 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8525 ciphers. NOTE: although the key derivation function has been verified
8526 against some published test vectors it has not been extensively tested
8527 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8528 of v2.0.
8529 [Steve Henson]
8530
8531 *) Instead of "mkdir -p", which is not fully portable, use new
8532 Perl script "util/mkdir-p.pl".
8533 [Bodo Moeller]
8534
8535 *) Rewrite the way password based encryption (PBE) is handled. It used to
8536 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8537 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8538 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8539 the 'parameter' field of the AlgorithmIdentifier is passed to the
8540 underlying key generation function so it must do its own ASN1 parsing.
8541 This has also changed the EVP_PBE_CipherInit() function which now has a
8542 'parameter' argument instead of literal salt and iteration count values
8543 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8544 [Steve Henson]
8545
8546 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8547 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8548 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8549 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8550 value was just used as a "magic string" and not used directly its
8551 value doesn't matter.
8552 [Steve Henson]
8553
8554 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8555 support mutable.
8556 [Ben Laurie]
8557
8558 *) "linux-sparc64" configuration (ultrapenguin).
8559 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8560 "linux-sparc" configuration.
8561 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8562
8563 *) config now generates no-xxx options for missing ciphers.
8564 [Ulf Möller]
8565
8566 *) Support the EBCDIC character set (work in progress).
8567 File ebcdic.c not yet included because it has a different license.
8568 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8569
8570 *) Support BS2000/OSD-POSIX.
8571 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8572
8573 *) Make callbacks for key generation use void * instead of char *.
8574 [Ben Laurie]
8575
8576 *) Make S/MIME samples compile (not yet tested).
8577 [Ben Laurie]
8578
8579 *) Additional typesafe stacks.
8580 [Ben Laurie]
8581
8582 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8583 [Bodo Moeller]
8584
8585
8586 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8587
8588 *) New configuration variant "sco5-gcc".
8589
8590 *) Updated some demos.
8591 [Sean O Riordain, Wade Scholine]
8592
8593 *) Add missing BIO_free at exit of pkcs12 application.
8594 [Wu Zhigang]
8595
8596 *) Fix memory leak in conf.c.
8597 [Steve Henson]
8598
8599 *) Updates for Win32 to assembler version of MD5.
8600 [Steve Henson]
8601
8602 *) Set #! path to perl in apps/der_chop to where we found it
8603 instead of using a fixed path.
8604 [Bodo Moeller]
8605
8606 *) SHA library changes for irix64-mips4-cc.
8607 [Andy Polyakov]
8608
8609 *) Improvements for VMS support.
8610 [Richard Levitte]
8611
8612
8613 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8614
8615 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8616 This also avoids the problems with SC4.2 and unpatched SC5.
8617 [Andy Polyakov <appro@fy.chalmers.se>]
8618
8619 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8620 These are required because of the typesafe stack would otherwise break
8621 existing code. If old code used a structure member which used to be STACK
8622 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8623 sk_num or sk_value it would produce an error because the num, data members
8624 are not present in STACK_OF. Now it just produces a warning. sk_set
8625 replaces the old method of assigning a value to sk_value
8626 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8627 that does this will no longer work (and should use sk_set instead) but
8628 this could be regarded as a "questionable" behaviour anyway.
8629 [Steve Henson]
8630
8631 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8632 correctly handle encrypted S/MIME data.
8633 [Steve Henson]
8634
8635 *) Change type of various DES function arguments from des_cblock
8636 (which means, in function argument declarations, pointer to char)
8637 to des_cblock * (meaning pointer to array with 8 char elements),
8638 which allows the compiler to do more typechecking; it was like
8639 that back in SSLeay, but with lots of ugly casts.
8640
8641 Introduce new type const_des_cblock.
8642 [Bodo Moeller]
8643
8644 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8645 problems: find RecipientInfo structure that matches recipient certificate
8646 and initialise the ASN1 structures properly based on passed cipher.
8647 [Steve Henson]
8648
8649 *) Belatedly make the BN tests actually check the results.
8650 [Ben Laurie]
8651
8652 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8653 to and from BNs: it was completely broken. New compilation option
8654 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8655 key elements as negative integers.
8656 [Steve Henson]
8657
8658 *) Reorganize and speed up MD5.
8659 [Andy Polyakov <appro@fy.chalmers.se>]
8660
8661 *) VMS support.
8662 [Richard Levitte <richard@levitte.org>]
8663
8664 *) New option -out to asn1parse to allow the parsed structure to be
8665 output to a file. This is most useful when combined with the -strparse
8666 option to examine the output of things like OCTET STRINGS.
8667 [Steve Henson]
8668
8669 *) Make SSL library a little more fool-proof by not requiring any longer
8670 that SSL_set_{accept,connect}_state be called before
8671 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8672 in many applications because usually everything *appeared* to work as
8673 intended anyway -- now it really works as intended).
8674 [Bodo Moeller]
8675
8676 *) Move openssl.cnf out of lib/.
8677 [Ulf Möller]
8678
8679 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8680 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8681 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8682 [Ralf S. Engelschall]
8683
8684 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8685 handle PKCS#7 enveloped data properly.
8686 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8687
8688 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8689 copying pointers. The cert_st handling is changed by this in
8690 various ways (and thus what used to be known as ctx->default_cert
8691 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8692 any longer when s->cert does not give us what we need).
8693 ssl_cert_instantiate becomes obsolete by this change.
8694 As soon as we've got the new code right (possibly it already is?),
8695 we have solved a couple of bugs of the earlier code where s->cert
8696 was used as if it could not have been shared with other SSL structures.
8697
8698 Note that using the SSL API in certain dirty ways now will result
8699 in different behaviour than observed with earlier library versions:
8700 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8701 does not influence s as it used to.
8702
8703 In order to clean up things more thoroughly, inside SSL_SESSION
8704 we don't use CERT any longer, but a new structure SESS_CERT
8705 that holds per-session data (if available); currently, this is
8706 the peer's certificate chain and, for clients, the server's certificate
8707 and temporary key. CERT holds only those values that can have
8708 meaningful defaults in an SSL_CTX.
8709 [Bodo Moeller]
8710
8711 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8712 from the internal representation. Various PKCS#7 fixes: remove some
8713 evil casts and set the enc_dig_alg field properly based on the signing
8714 key type.
8715 [Steve Henson]
8716
8717 *) Allow PKCS#12 password to be set from the command line or the
8718 environment. Let 'ca' get its config file name from the environment
8719 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8720 and 'x509').
8721 [Steve Henson]
8722
8723 *) Allow certificate policies extension to use an IA5STRING for the
8724 organization field. This is contrary to the PKIX definition but
8725 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8726 extension option.
8727 [Steve Henson]
8728
8729 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8730 without disallowing inline assembler and the like for non-pedantic builds.
8731 [Ben Laurie]
8732
8733 *) Support Borland C++ builder.
8734 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8735
8736 *) Support Mingw32.
8737 [Ulf Möller]
8738
8739 *) SHA-1 cleanups and performance enhancements.
8740 [Andy Polyakov <appro@fy.chalmers.se>]
8741
8742 *) Sparc v8plus assembler for the bignum library.
8743 [Andy Polyakov <appro@fy.chalmers.se>]
8744
8745 *) Accept any -xxx and +xxx compiler options in Configure.
8746 [Ulf Möller]
8747
8748 *) Update HPUX configuration.
8749 [Anonymous]
8750
8751 *) Add missing sk_<type>_unshift() function to safestack.h
8752 [Ralf S. Engelschall]
8753
8754 *) New function SSL_CTX_use_certificate_chain_file that sets the
8755 "extra_cert"s in addition to the certificate. (This makes sense
8756 only for "PEM" format files, as chains as a whole are not
8757 DER-encoded.)
8758 [Bodo Moeller]
8759
8760 *) Support verify_depth from the SSL API.
8761 x509_vfy.c had what can be considered an off-by-one-error:
8762 Its depth (which was not part of the external interface)
8763 was actually counting the number of certificates in a chain;
8764 now it really counts the depth.
8765 [Bodo Moeller]
8766
8767 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8768 instead of X509err, which often resulted in confusing error
8769 messages since the error codes are not globally unique
8770 (e.g. an alleged error in ssl3_accept when a certificate
8771 didn't match the private key).
8772
8773 *) New function SSL_CTX_set_session_id_context that allows to set a default
8774 value (so that you don't need SSL_set_session_id_context for each
8775 connection using the SSL_CTX).
8776 [Bodo Moeller]
8777
8778 *) OAEP decoding bug fix.
8779 [Ulf Möller]
8780
8781 *) Support INSTALL_PREFIX for package builders, as proposed by
8782 David Harris.
8783 [Bodo Moeller]
8784
8785 *) New Configure options "threads" and "no-threads". For systems
8786 where the proper compiler options are known (currently Solaris
8787 and Linux), "threads" is the default.
8788 [Bodo Moeller]
8789
8790 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8791 [Bodo Moeller]
8792
8793 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8794 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8795 such as /usr/local/bin.
8796 [Bodo Moeller]
8797
8798 *) "make linux-shared" to build shared libraries.
8799 [Niels Poppe <niels@netbox.org>]
8800
8801 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8802 [Ulf Möller]
8803
8804 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8805 extension adding in x509 utility.
8806 [Steve Henson]
8807
8808 *) Remove NOPROTO sections and error code comments.
8809 [Ulf Möller]
8810
8811 *) Partial rewrite of the DEF file generator to now parse the ANSI
8812 prototypes.
8813 [Steve Henson]
8814
8815 *) New Configure options --prefix=DIR and --openssldir=DIR.
8816 [Ulf Möller]
8817
8818 *) Complete rewrite of the error code script(s). It is all now handled
8819 by one script at the top level which handles error code gathering,
8820 header rewriting and C source file generation. It should be much better
8821 than the old method: it now uses a modified version of Ulf's parser to
8822 read the ANSI prototypes in all header files (thus the old K&R definitions
8823 aren't needed for error creation any more) and do a better job of
8824 translating function codes into names. The old 'ASN1 error code imbedded
8825 in a comment' is no longer necessary and it doesn't use .err files which
8826 have now been deleted. Also the error code call doesn't have to appear all
8827 on one line (which resulted in some large lines...).
8828 [Steve Henson]
8829
8830 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8831 [Bodo Moeller]
8832
8833 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8834 0 (which usually indicates a closed connection), but continue reading.
8835 [Bodo Moeller]
8836
8837 *) Fix some race conditions.
8838 [Bodo Moeller]
8839
8840 *) Add support for CRL distribution points extension. Add Certificate
8841 Policies and CRL distribution points documentation.
8842 [Steve Henson]
8843
8844 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8845 [Ulf Möller]
8846
8847 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8848 8 of keying material. Merlin has also confirmed interop with this fix
8849 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8850 [Merlin Hughes <merlin@baltimore.ie>]
8851
8852 *) Fix lots of warnings.
8853 [Richard Levitte <levitte@stacken.kth.se>]
8854
8855 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8856 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8857 [Richard Levitte <levitte@stacken.kth.se>]
8858
8859 *) Fix problems with sizeof(long) == 8.
8860 [Andy Polyakov <appro@fy.chalmers.se>]
8861
8862 *) Change functions to ANSI C.
8863 [Ulf Möller]
8864
8865 *) Fix typos in error codes.
8866 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8867
8868 *) Remove defunct assembler files from Configure.
8869 [Ulf Möller]
8870
8871 *) SPARC v8 assembler BIGNUM implementation.
8872 [Andy Polyakov <appro@fy.chalmers.se>]
8873
8874 *) Support for Certificate Policies extension: both print and set.
8875 Various additions to support the r2i method this uses.
8876 [Steve Henson]
8877
8878 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8879 return a const string when you are expecting an allocated buffer.
8880 [Ben Laurie]
8881
8882 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8883 types DirectoryString and DisplayText.
8884 [Steve Henson]
8885
8886 *) Add code to allow r2i extensions to access the configuration database,
8887 add an LHASH database driver and add several ctx helper functions.
8888 [Steve Henson]
8889
8890 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8891 fail when they extended the size of a BIGNUM.
8892 [Steve Henson]
8893
8894 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8895 support typesafe stack.
8896 [Steve Henson]
8897
8898 *) Fix typo in SSL_[gs]et_options().
8899 [Nils Frostberg <nils@medcom.se>]
8900
8901 *) Delete various functions and files that belonged to the (now obsolete)
8902 old X509V3 handling code.
8903 [Steve Henson]
8904
8905 *) New Configure option "rsaref".
8906 [Ulf Möller]
8907
8908 *) Don't auto-generate pem.h.
8909 [Bodo Moeller]
8910
8911 *) Introduce type-safe ASN.1 SETs.
8912 [Ben Laurie]
8913
8914 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8915 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8916
8917 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8918 that links with OpenSSL (well at least cause lots of warnings), but fear
8919 not: the conversion is trivial, and it eliminates loads of evil casts. A
8920 few STACKed things have been converted already. Feel free to convert more.
8921 In the fullness of time, I'll do away with the STACK type altogether.
8922 [Ben Laurie]
8923
8924 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8925 specified in <certfile> by updating the entry in the index.txt file.
8926 This way one no longer has to edit the index.txt file manually for
8927 revoking a certificate. The -revoke option does the gory details now.
8928 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8929
8930 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8931 `-text' option at all and this way the `-noout -text' combination was
8932 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8933 [Ralf S. Engelschall]
8934
8935 *) Make sure a corresponding plain text error message exists for the
8936 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8937 verify callback function determined that a certificate was revoked.
8938 [Ralf S. Engelschall]
8939
8940 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8941 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8942 all available cipers including rc5, which was forgotten until now.
8943 In order to let the testing shell script know which algorithms
8944 are available, a new (up to now undocumented) command
8945 "openssl list-cipher-commands" is used.
8946 [Bodo Moeller]
8947
8948 *) Bugfix: s_client occasionally would sleep in select() when
8949 it should have checked SSL_pending() first.
8950 [Bodo Moeller]
8951
8952 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8953 the raw DSA values prior to ASN.1 encoding.
8954 [Ulf Möller]
8955
8956 *) Tweaks to Configure
8957 [Niels Poppe <niels@netbox.org>]
8958
8959 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8960 yet...
8961 [Steve Henson]
8962
8963 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8964 [Ulf Möller]
8965
8966 *) New config option to avoid instructions that are illegal on the 80386.
8967 The default code is faster, but requires at least a 486.
8968 [Ulf Möller]
8969
8970 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8971 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8972 same as SSL2_VERSION anyway.
8973 [Bodo Moeller]
8974
8975 *) New "-showcerts" option for s_client.
8976 [Bodo Moeller]
8977
8978 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8979 application. Various cleanups and fixes.
8980 [Steve Henson]
8981
8982 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8983 modify error routines to work internally. Add error codes and PBE init
8984 to library startup routines.
8985 [Steve Henson]
8986
8987 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8988 packing functions to asn1 and evp. Changed function names and error
8989 codes along the way.
8990 [Steve Henson]
8991
8992 *) PKCS12 integration: and so it begins... First of several patches to
8993 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8994 objects to objects.h
8995 [Steve Henson]
8996
8997 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8998 and display support for Thawte strong extranet extension.
8999 [Steve Henson]
9000
9001 *) Add LinuxPPC support.
9002 [Jeff Dubrule <igor@pobox.org>]
9003
9004 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9005 bn_div_words in alpha.s.
9006 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9007
9008 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9009 OAEP isn't supported when OpenSSL is built with RSAref.
9010 [Ulf Moeller <ulf@fitug.de>]
9011
9012 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9013 so they no longer are missing under -DNOPROTO.
9014 [Soren S. Jorvang <soren@t.dk>]
9015
9016
9017 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9018
9019 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9020 doesn't work when the session is reused. Coming soon!
9021 [Ben Laurie]
9022
9023 *) Fix a security hole, that allows sessions to be reused in the wrong
9024 context thus bypassing client cert protection! All software that uses
9025 client certs and session caches in multiple contexts NEEDS PATCHING to
9026 allow session reuse! A fuller solution is in the works.
9027 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9028
9029 *) Some more source tree cleanups (removed obsolete files
9030 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9031 permission on "config" script to be executable) and a fix for the INSTALL
9032 document.
9033 [Ulf Moeller <ulf@fitug.de>]
9034
9035 *) Remove some legacy and erroneous uses of malloc, free instead of
9036 Malloc, Free.
9037 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9038
9039 *) Make rsa_oaep_test return non-zero on error.
9040 [Ulf Moeller <ulf@fitug.de>]
9041
9042 *) Add support for native Solaris shared libraries. Configure
9043 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9044 if someone would make that last step automatic.
9045 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9046
9047 *) ctx_size was not built with the right compiler during "make links". Fixed.
9048 [Ben Laurie]
9049
9050 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9051 except NULL ciphers". This means the default cipher list will no longer
9052 enable NULL ciphers. They need to be specifically enabled e.g. with
9053 the string "DEFAULT:eNULL".
9054 [Steve Henson]
9055
9056 *) Fix to RSA private encryption routines: if p < q then it would
9057 occasionally produce an invalid result. This will only happen with
9058 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9059 [Steve Henson]
9060
9061 *) Be less restrictive and allow also `perl util/perlpath.pl
9062 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9063 because this way one can also use an interpreter named `perl5' (which is
9064 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9065 installed as `perl').
9066 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9067
9068 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9069 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9070
9071 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9072 advapi32.lib to Win32 build and change the pem test comparision
9073 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9074 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9075 and crypto/des/ede_cbcm_enc.c.
9076 [Steve Henson]
9077
9078 *) DES quad checksum was broken on big-endian architectures. Fixed.
9079 [Ben Laurie]
9080
9081 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9082 Win32 test batch file so it (might) work again. The Win32 test batch file
9083 is horrible: I feel ill....
9084 [Steve Henson]
9085
9086 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9087 in e_os.h. Audit of header files to check ANSI and non ANSI
9088 sections: 10 functions were absent from non ANSI section and not exported
9089 from Windows DLLs. Fixed up libeay.num for new functions.
9090 [Steve Henson]
9091
9092 *) Make `openssl version' output lines consistent.
9093 [Ralf S. Engelschall]
9094
9095 *) Fix Win32 symbol export lists for BIO functions: Added
9096 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9097 to ms/libeay{16,32}.def.
9098 [Ralf S. Engelschall]
9099
9100 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9101 fine under Unix and passes some trivial tests I've now added. But the
9102 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9103 added to make sure no one expects that this stuff really works in the
9104 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9105 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9106 openssl_bio.xs.
9107 [Ralf S. Engelschall]
9108
9109 *) Fix the generation of two part addresses in perl.
9110 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9111
9112 *) Add config entry for Linux on MIPS.
9113 [John Tobey <jtobey@channel1.com>]
9114
9115 *) Make links whenever Configure is run, unless we are on Windoze.
9116 [Ben Laurie]
9117
9118 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9119 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9120 in CRLs.
9121 [Steve Henson]
9122
9123 *) Add a useful kludge to allow package maintainers to specify compiler and
9124 other platforms details on the command line without having to patch the
9125 Configure script everytime: One now can use ``perl Configure
9126 <id>:<details>'', i.e. platform ids are allowed to have details appended
9127 to them (seperated by colons). This is treated as there would be a static
9128 pre-configured entry in Configure's %table under key <id> with value
9129 <details> and ``perl Configure <id>'' is called. So, when you want to
9130 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9131 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9132 now, which overrides the FreeBSD-elf entry on-the-fly.
9133 [Ralf S. Engelschall]
9134
9135 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9136 [Ben Laurie]
9137
9138 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9139 on the `perl Configure ...' command line. This way one can compile
9140 OpenSSL libraries with Position Independent Code (PIC) which is needed
9141 for linking it into DSOs.
9142 [Ralf S. Engelschall]
9143
9144 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9145 Fixed.
9146 [Ben Laurie]
9147
9148 *) Cleaned up the LICENSE document: The official contact for any license
9149 questions now is the OpenSSL core team under openssl-core@openssl.org.
9150 And add a paragraph about the dual-license situation to make sure people
9151 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9152 to the OpenSSL toolkit.
9153 [Ralf S. Engelschall]
9154
9155 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9156 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9157 Additonally cleaned up the `make links' target: Remove unnecessary
9158 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9159 to speed processing and no longer clutter the display with confusing
9160 stuff. Instead only the actually done links are displayed.
9161 [Ralf S. Engelschall]
9162
9163 *) Permit null encryption ciphersuites, used for authentication only. It used
9164 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9165 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9166 encryption.
9167 [Ben Laurie]
9168
9169 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9170 signed attributes when verifying signatures (this would break them),
9171 the detached data encoding was wrong and public keys obtained using
9172 X509_get_pubkey() weren't freed.
9173 [Steve Henson]
9174
9175 *) Add text documentation for the BUFFER functions. Also added a work around
9176 to a Win95 console bug. This was triggered by the password read stuff: the
9177 last character typed gets carried over to the next fread(). If you were
9178 generating a new cert request using 'req' for example then the last
9179 character of the passphrase would be CR which would then enter the first
9180 field as blank.
9181 [Steve Henson]
9182
9183 *) Added the new `Includes OpenSSL Cryptography Software' button as
9184 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9185 button and can be used by applications based on OpenSSL to show the
9186 relationship to the OpenSSL project.
9187 [Ralf S. Engelschall]
9188
9189 *) Remove confusing variables in function signatures in files
9190 ssl/ssl_lib.c and ssl/ssl.h.
9191 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9192
9193 *) Don't install bss_file.c under PREFIX/include/
9194 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9195
9196 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9197 functions that return function pointers and has support for NT specific
9198 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9199 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9200 unsigned to signed types: this was killing the Win32 compile.
9201 [Steve Henson]
9202
9203 *) Add new certificate file to stack functions,
9204 SSL_add_dir_cert_subjects_to_stack() and
9205 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9206 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9207 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9208 This means that Apache-SSL and similar packages don't have to mess around
9209 to add as many CAs as they want to the preferred list.
9210 [Ben Laurie]
9211
9212 *) Experiment with doxygen documentation. Currently only partially applied to
9213 ssl/ssl_lib.c.
9214 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9215 openssl.doxy as the configuration file.
9216 [Ben Laurie]
9217
9218 *) Get rid of remaining C++-style comments which strict C compilers hate.
9219 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9220
9221 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9222 compiled in by default: it has problems with large keys.
9223 [Steve Henson]
9224
9225 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9226 DH private keys and/or callback functions which directly correspond to
9227 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9228 is needed for applications which have to configure certificates on a
9229 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9230 (e.g. s_server).
9231 For the RSA certificate situation is makes no difference, but
9232 for the DSA certificate situation this fixes the "no shared cipher"
9233 problem where the OpenSSL cipher selection procedure failed because the
9234 temporary keys were not overtaken from the context and the API provided
9235 no way to reconfigure them.
9236 The new functions now let applications reconfigure the stuff and they
9237 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9238 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9239 non-public-API function ssl_cert_instantiate() is used as a helper
9240 function and also to reduce code redundancy inside ssl_rsa.c.
9241 [Ralf S. Engelschall]
9242
9243 *) Move s_server -dcert and -dkey options out of the undocumented feature
9244 area because they are useful for the DSA situation and should be
9245 recognized by the users.
9246 [Ralf S. Engelschall]
9247
9248 *) Fix the cipher decision scheme for export ciphers: the export bits are
9249 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9250 SSL_EXP_MASK. So, the original variable has to be used instead of the
9251 already masked variable.
9252 [Richard Levitte <levitte@stacken.kth.se>]
9253
9254 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9255 [Richard Levitte <levitte@stacken.kth.se>]
9256
9257 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9258 from `int' to `unsigned int' because it's a length and initialized by
9259 EVP_DigestFinal() which expects an `unsigned int *'.
9260 [Richard Levitte <levitte@stacken.kth.se>]
9261
9262 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9263 script. Instead use the usual Shell->Perl transition trick.
9264 [Ralf S. Engelschall]
9265
9266 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9267 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9268 -noout -modulus' as it's already the case for `openssl rsa -noout
9269 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9270 currently the public key is printed (a decision which was already done by
9271 `openssl dsa -modulus' in the past) which serves a similar purpose.
9272 Additionally the NO_RSA no longer completely removes the whole -modulus
9273 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9274 now, too.
9275 [Ralf S. Engelschall]
9276
9277 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9278 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9279 [Arne Ansper <arne@ats.cyber.ee>]
9280
9281 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9282 to be added. Now both 'req' and 'ca' can use new objects defined in the
9283 config file.
9284 [Steve Henson]
9285
9286 *) Add cool BIO that does syslog (or event log on NT).
9287 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9288
9289 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9290 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9291 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9292 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9293 [Ben Laurie]
9294
9295 *) Add preliminary config info for new extension code.
9296 [Steve Henson]
9297
9298 *) Make RSA_NO_PADDING really use no padding.
9299 [Ulf Moeller <ulf@fitug.de>]
9300
9301 *) Generate errors when private/public key check is done.
9302 [Ben Laurie]
9303
9304 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9305 for some CRL extensions and new objects added.
9306 [Steve Henson]
9307
9308 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9309 key usage extension and fuller support for authority key id.
9310 [Steve Henson]
9311
9312 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9313 padding method for RSA, which is recommended for new applications in PKCS
9314 #1 v2.0 (RFC 2437, October 1998).
9315 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9316 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9317 against Bleichbacher's attack on RSA.
9318 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9319 Ben Laurie]
9320
9321 *) Updates to the new SSL compression code
9322 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9323
9324 *) Fix so that the version number in the master secret, when passed
9325 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9326 (because the server will not accept higher), that the version number
9327 is 0x03,0x01, not 0x03,0x00
9328 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9329
9330 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9331 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9332 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9333 [Steve Henson]
9334
9335 *) Support for RAW extensions where an arbitrary extension can be
9336 created by including its DER encoding. See apps/openssl.cnf for
9337 an example.
9338 [Steve Henson]
9339
9340 *) Make sure latest Perl versions don't interpret some generated C array
9341 code as Perl array code in the crypto/err/err_genc.pl script.
9342 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9343
9344 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9345 not many people have the assembler. Various Win32 compilation fixes and
9346 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9347 build instructions.
9348 [Steve Henson]
9349
9350 *) Modify configure script 'Configure' to automatically create crypto/date.h
9351 file under Win32 and also build pem.h from pem.org. New script
9352 util/mkfiles.pl to create the MINFO file on environments that can't do a
9353 'make files': perl util/mkfiles.pl >MINFO should work.
9354 [Steve Henson]
9355
9356 *) Major rework of DES function declarations, in the pursuit of correctness
9357 and purity. As a result, many evil casts evaporated, and some weirdness,
9358 too. You may find this causes warnings in your code. Zapping your evil
9359 casts will probably fix them. Mostly.
9360 [Ben Laurie]
9361
9362 *) Fix for a typo in asn1.h. Bug fix to object creation script
9363 obj_dat.pl. It considered a zero in an object definition to mean
9364 "end of object": none of the objects in objects.h have any zeros
9365 so it wasn't spotted.
9366 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9367
9368 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9369 Masking (CBCM). In the absence of test vectors, the best I have been able
9370 to do is check that the decrypt undoes the encrypt, so far. Send me test
9371 vectors if you have them.
9372 [Ben Laurie]
9373
9374 *) Correct calculation of key length for export ciphers (too much space was
9375 allocated for null ciphers). This has not been tested!
9376 [Ben Laurie]
9377
9378 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9379 message is now correct (it understands "crypto" and "ssl" on its
9380 command line). There is also now an "update" option. This will update
9381 the util/ssleay.num and util/libeay.num files with any new functions.
9382 If you do a:
9383 perl util/mkdef.pl crypto ssl update
9384 it will update them.
9385 [Steve Henson]
9386
9387 *) Overhauled the Perl interface (perl/*):
9388 - ported BN stuff to OpenSSL's different BN library
9389 - made the perl/ source tree CVS-aware
9390 - renamed the package from SSLeay to OpenSSL (the files still contain
9391 their history because I've copied them in the repository)
9392 - removed obsolete files (the test scripts will be replaced
9393 by better Test::Harness variants in the future)
9394 [Ralf S. Engelschall]
9395
9396 *) First cut for a very conservative source tree cleanup:
9397 1. merge various obsolete readme texts into doc/ssleay.txt
9398 where we collect the old documents and readme texts.
9399 2. remove the first part of files where I'm already sure that we no
9400 longer need them because of three reasons: either they are just temporary
9401 files which were left by Eric or they are preserved original files where
9402 I've verified that the diff is also available in the CVS via "cvs diff
9403 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9404 the crypto/md/ stuff).
9405 [Ralf S. Engelschall]
9406
9407 *) More extension code. Incomplete support for subject and issuer alt
9408 name, issuer and authority key id. Change the i2v function parameters
9409 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9410 what that's for :-) Fix to ASN1 macro which messed up
9411 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9412 [Steve Henson]
9413
9414 *) Preliminary support for ENUMERATED type. This is largely copied from the
9415 INTEGER code.
9416 [Steve Henson]
9417
9418 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9419 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9420
9421 *) Make sure `make rehash' target really finds the `openssl' program.
9422 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9423
9424 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9425 like to hear about it if this slows down other processors.
9426 [Ben Laurie]
9427
9428 *) Add CygWin32 platform information to Configure script.
9429 [Alan Batie <batie@aahz.jf.intel.com>]
9430
9431 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9432 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9433
9434 *) New program nseq to manipulate netscape certificate sequences
9435 [Steve Henson]
9436
9437 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9438 few typos.
9439 [Steve Henson]
9440
9441 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9442 but the BN code had some problems that would cause failures when
9443 doing certificate verification and some other functions.
9444 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9445
9446 *) Add ASN1 and PEM code to support netscape certificate sequences.
9447 [Steve Henson]
9448
9449 *) Add ASN1 and PEM code to support netscape certificate sequences.
9450 [Steve Henson]
9451
9452 *) Add several PKIX and private extended key usage OIDs.
9453 [Steve Henson]
9454
9455 *) Modify the 'ca' program to handle the new extension code. Modify
9456 openssl.cnf for new extension format, add comments.
9457 [Steve Henson]
9458
9459 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9460 and add a sample to openssl.cnf so req -x509 now adds appropriate
9461 CA extensions.
9462 [Steve Henson]
9463
9464 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9465 error code, add initial support to X509_print() and x509 application.
9466 [Steve Henson]
9467
9468 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9469 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9470 stuff is currently isolated and isn't even compiled yet.
9471 [Steve Henson]
9472
9473 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9474 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9475 Removed the versions check from X509 routines when loading extensions:
9476 this allows certain broken certificates that don't set the version
9477 properly to be processed.
9478 [Steve Henson]
9479
9480 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9481 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9482 can still be regenerated with "make depend".
9483 [Ben Laurie]
9484
9485 *) Spelling mistake in C version of CAST-128.
9486 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9487
9488 *) Changes to the error generation code. The perl script err-code.pl
9489 now reads in the old error codes and retains the old numbers, only
9490 adding new ones if necessary. It also only changes the .err files if new
9491 codes are added. The makefiles have been modified to only insert errors
9492 when needed (to avoid needlessly modifying header files). This is done
9493 by only inserting errors if the .err file is newer than the auto generated
9494 C file. To rebuild all the error codes from scratch (the old behaviour)
9495 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9496 or delete all the .err files.
9497 [Steve Henson]
9498
9499 *) CAST-128 was incorrectly implemented for short keys. The C version has
9500 been fixed, but is untested. The assembler versions are also fixed, but
9501 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9502 to regenerate it if needed.
9503 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9504 Hagino <itojun@kame.net>]
9505
9506 *) File was opened incorrectly in randfile.c.
9507 [Ulf Möller <ulf@fitug.de>]
9508
9509 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9510 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9511 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9512 al: it's just almost always a UTCTime. Note this patch adds new error
9513 codes so do a "make errors" if there are problems.
9514 [Steve Henson]
9515
9516 *) Correct Linux 1 recognition in config.
9517 [Ulf Möller <ulf@fitug.de>]
9518
9519 *) Remove pointless MD5 hash when using DSA keys in ca.
9520 [Anonymous <nobody@replay.com>]
9521
9522 *) Generate an error if given an empty string as a cert directory. Also
9523 generate an error if handed NULL (previously returned 0 to indicate an
9524 error, but didn't set one).
9525 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9526
9527 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9528 [Ben Laurie]
9529
9530 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9531 parameters. This was causing a warning which killed off the Win32 compile.
9532 [Steve Henson]
9533
9534 *) Remove C++ style comments from crypto/bn/bn_local.h.
9535 [Neil Costigan <neil.costigan@celocom.com>]
9536
9537 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9538 based on a text string, looking up short and long names and finally
9539 "dot" format. The "dot" format stuff didn't work. Added new function
9540 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9541 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9542 OID is not part of the table.
9543 [Steve Henson]
9544
9545 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9546 X509_LOOKUP_by_alias().
9547 [Ben Laurie]
9548
9549 *) Sort openssl functions by name.
9550 [Ben Laurie]
9551
9552 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9553 encryption from sample DSA keys (in case anyone is interested the password
9554 was "1234").
9555 [Steve Henson]
9556
9557 *) Make _all_ *_free functions accept a NULL pointer.
9558 [Frans Heymans <fheymans@isaserver.be>]
9559
9560 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9561 NULL pointers.
9562 [Anonymous <nobody@replay.com>]
9563
9564 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9565 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9566
9567 *) Don't blow it for numeric -newkey arguments to apps/req.
9568 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9569
9570 *) Temp key "for export" tests were wrong in s3_srvr.c.
9571 [Anonymous <nobody@replay.com>]
9572
9573 *) Add prototype for temp key callback functions
9574 SSL_CTX_set_tmp_{rsa,dh}_callback().
9575 [Ben Laurie]
9576
9577 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9578 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9579 [Steve Henson]
9580
9581 *) X509_name_add_entry() freed the wrong thing after an error.
9582 [Arne Ansper <arne@ats.cyber.ee>]
9583
9584 *) rsa_eay.c would attempt to free a NULL context.
9585 [Arne Ansper <arne@ats.cyber.ee>]
9586
9587 *) BIO_s_socket() had a broken should_retry() on Windoze.
9588 [Arne Ansper <arne@ats.cyber.ee>]
9589
9590 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9591 [Arne Ansper <arne@ats.cyber.ee>]
9592
9593 *) Make sure the already existing X509_STORE->depth variable is initialized
9594 in X509_STORE_new(), but document the fact that this variable is still
9595 unused in the certificate verification process.
9596 [Ralf S. Engelschall]
9597
9598 *) Fix the various library and apps files to free up pkeys obtained from
9599 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9600 [Steve Henson]
9601
9602 *) Fix reference counting in X509_PUBKEY_get(). This makes
9603 demos/maurice/example2.c work, amongst others, probably.
9604 [Steve Henson and Ben Laurie]
9605
9606 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9607 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9608 are no longer created. This way we have a single and consistent command
9609 line interface `openssl <command>', similar to `cvs <command>'.
9610 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9611
9612 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9613 BIT STRING wrapper always have zero unused bits.
9614 [Steve Henson]
9615
9616 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9617 [Steve Henson]
9618
9619 *) Make the top-level INSTALL documentation easier to understand.
9620 [Paul Sutton]
9621
9622 *) Makefiles updated to exit if an error occurs in a sub-directory
9623 make (including if user presses ^C) [Paul Sutton]
9624
9625 *) Make Montgomery context stuff explicit in RSA data structure.
9626 [Ben Laurie]
9627
9628 *) Fix build order of pem and err to allow for generated pem.h.
9629 [Ben Laurie]
9630
9631 *) Fix renumbering bug in X509_NAME_delete_entry().
9632 [Ben Laurie]
9633
9634 *) Enhanced the err-ins.pl script so it makes the error library number
9635 global and can add a library name. This is needed for external ASN1 and
9636 other error libraries.
9637 [Steve Henson]
9638
9639 *) Fixed sk_insert which never worked properly.
9640 [Steve Henson]
9641
9642 *) Fix ASN1 macros so they can handle indefinite length construted
9643 EXPLICIT tags. Some non standard certificates use these: they can now
9644 be read in.
9645 [Steve Henson]
9646
9647 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9648 into a single doc/ssleay.txt bundle. This way the information is still
9649 preserved but no longer messes up this directory. Now it's new room for
9650 the new set of documenation files.
9651 [Ralf S. Engelschall]
9652
9653 *) SETs were incorrectly DER encoded. This was a major pain, because they
9654 shared code with SEQUENCEs, which aren't coded the same. This means that
9655 almost everything to do with SETs or SEQUENCEs has either changed name or
9656 number of arguments.
9657 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9658
9659 *) Fix test data to work with the above.
9660 [Ben Laurie]
9661
9662 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9663 was already fixed by Eric for 0.9.1 it seems.
9664 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9665
9666 *) Autodetect FreeBSD3.
9667 [Ben Laurie]
9668
9669 *) Fix various bugs in Configure. This affects the following platforms:
9670 nextstep
9671 ncr-scde
9672 unixware-2.0
9673 unixware-2.0-pentium
9674 sco5-cc.
9675 [Ben Laurie]
9676
9677 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9678 before they are needed.
9679 [Ben Laurie]
9680
9681 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9682 [Ben Laurie]
9683
9684
9685 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9686
9687 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9688 changed SSLeay to OpenSSL in version strings.
9689 [Ralf S. Engelschall]
9690
9691 *) Some fixups to the top-level documents.
9692 [Paul Sutton]
9693
9694 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9695 because the symlink to include/ was missing.
9696 [Ralf S. Engelschall]
9697
9698 *) Incorporated the popular no-RSA/DSA-only patches
9699 which allow to compile a RSA-free SSLeay.
9700 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9701
9702 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9703 when "ssleay" is still not found.
9704 [Ralf S. Engelschall]
9705
9706 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9707 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9708
9709 *) Updated the README file.
9710 [Ralf S. Engelschall]
9711
9712 *) Added various .cvsignore files in the CVS repository subdirs
9713 to make a "cvs update" really silent.
9714 [Ralf S. Engelschall]
9715
9716 *) Recompiled the error-definition header files and added
9717 missing symbols to the Win32 linker tables.
9718 [Ralf S. Engelschall]
9719
9720 *) Cleaned up the top-level documents;
9721 o new files: CHANGES and LICENSE
9722 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9723 o merged COPYRIGHT into LICENSE
9724 o removed obsolete TODO file
9725 o renamed MICROSOFT to INSTALL.W32
9726 [Ralf S. Engelschall]
9727
9728 *) Removed dummy files from the 0.9.1b source tree:
9729 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9730 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9731 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9732 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9733 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9734 [Ralf S. Engelschall]
9735
9736 *) Added various platform portability fixes.
9737 [Mark J. Cox]
9738
9739 *) The Genesis of the OpenSSL rpject:
9740 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9741 Young and Tim J. Hudson created while they were working for C2Net until
9742 summer 1998.
9743 [The OpenSSL Project]
9744
9745
9746 Changes between 0.9.0b and 0.9.1b [not released]
9747
9748 *) Updated a few CA certificates under certs/
9749 [Eric A. Young]
9750
9751 *) Changed some BIGNUM api stuff.
9752 [Eric A. Young]
9753
9754 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9755 DGUX x86, Linux Alpha, etc.
9756 [Eric A. Young]
9757
9758 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9759 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9760 available).
9761 [Eric A. Young]
9762
9763 *) Add -strparse option to asn1pars program which parses nested
9764 binary structures
9765 [Dr Stephen Henson <shenson@bigfoot.com>]
9766
9767 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9768 [Eric A. Young]
9769
9770 *) DSA fix for "ca" program.
9771 [Eric A. Young]
9772
9773 *) Added "-genkey" option to "dsaparam" program.
9774 [Eric A. Young]
9775
9776 *) Added RIPE MD160 (rmd160) message digest.
9777 [Eric A. Young]
9778
9779 *) Added -a (all) option to "ssleay version" command.
9780 [Eric A. Young]
9781
9782 *) Added PLATFORM define which is the id given to Configure.
9783 [Eric A. Young]
9784
9785 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9786 [Eric A. Young]
9787
9788 *) Extended the ASN.1 parser routines.
9789 [Eric A. Young]
9790
9791 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9792 [Eric A. Young]
9793
9794 *) Added a BN_CTX to the BN library.
9795 [Eric A. Young]
9796
9797 *) Fixed the weak key values in DES library
9798 [Eric A. Young]
9799
9800 *) Changed API in EVP library for cipher aliases.
9801 [Eric A. Young]
9802
9803 *) Added support for RC2/64bit cipher.
9804 [Eric A. Young]
9805
9806 *) Converted the lhash library to the crypto/mem.c functions.
9807 [Eric A. Young]
9808
9809 *) Added more recognized ASN.1 object ids.
9810 [Eric A. Young]
9811
9812 *) Added more RSA padding checks for SSL/TLS.
9813 [Eric A. Young]
9814
9815 *) Added BIO proxy/filter functionality.
9816 [Eric A. Young]
9817
9818 *) Added extra_certs to SSL_CTX which can be used
9819 send extra CA certificates to the client in the CA cert chain sending
9820 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9821 [Eric A. Young]
9822
9823 *) Now Fortezza is denied in the authentication phase because
9824 this is key exchange mechanism is not supported by SSLeay at all.
9825 [Eric A. Young]
9826
9827 *) Additional PKCS1 checks.
9828 [Eric A. Young]
9829
9830 *) Support the string "TLSv1" for all TLS v1 ciphers.
9831 [Eric A. Young]
9832
9833 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9834 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9835 [Eric A. Young]
9836
9837 *) Fixed a few memory leaks.
9838 [Eric A. Young]
9839
9840 *) Fixed various code and comment typos.
9841 [Eric A. Young]
9842
9843 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9844 bytes sent in the client random.
9845 [Edward Bishop <ebishop@spyglass.com>]
9846