]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
[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 *) Support for fixed DH ciphersuites: those requiring DH server
8 certificates.
9 [Steve Henson]
10
11 *) Transparently support X9.42 DH parameters when calling
12 PEM_read_bio_DHparameters. This means existing applications can handle
13 the new parameter format automatically.
14 [Steve Henson]
15
16 *) Initial experimental support for X9.42 DH parameter format: mainly
17 to support use of 'q' parameter for RFC5114 parameters.
18 [Steve Henson]
19
20 *) Add DH parameters from RFC5114 including test data to dhtest.
21 [Steve Henson]
22
23 *) Update fips_test_suite to support multiple command line options. New
24 test to induce all self test errors in sequence and check expected
25 failures.
26 [Steve Henson]
27
28 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
29 sign or verify all in one operation.
30 [Steve Henson]
31
32 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
33 test programs and fips_test_suite. Includes functionality to parse
34 the minimal script output of fipsalgest.pl directly.
35 [Steve Henson]
36
37 *) Add authorisation parameter to FIPS_module_mode_set().
38 [Steve Henson]
39
40 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
41 [Steve Henson]
42
43 *) Use separate DRBG fields for internal and external flags. New function
44 FIPS_drbg_health_check() to perform on demand health checking. Add
45 generation tests to fips_test_suite with reduced health check interval to
46 demonstrate periodic health checking. Add "nodh" option to
47 fips_test_suite to skip very slow DH test.
48 [Steve Henson]
49
50 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
51 based on NID.
52 [Steve Henson]
53
54 *) More extensive health check for DRBG checking many more failure modes.
55 New function FIPS_selftest_drbg_all() to handle every possible DRBG
56 combination: call this in fips_test_suite.
57 [Steve Henson]
58
59 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
60 and POST to handle Dual EC cases.
61 [Steve Henson]
62
63 *) Add support for canonical generation of DSA parameter 'g'. See
64 FIPS 186-3 A.2.3.
65
66 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
67 POST to handle HMAC cases.
68 [Steve Henson]
69
70 *) Add functions FIPS_module_version() and FIPS_module_version_text()
71 to return numberical and string versions of the FIPS module number.
72 [Steve Henson]
73
74 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
75 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
76 outside the validated module in the FIPS capable OpenSSL.
77 [Steve Henson]
78
79 *) Minor change to DRBG entropy callback semantics. In some cases
80 there is no mutiple of the block length between min_len and
81 max_len. Allow the callback to return more than max_len bytes
82 of entropy but discard any extra: it is the callback's responsibility
83 to ensure that the extra data discarded does not impact the
84 requested amount of entropy.
85 [Steve Henson]
86
87 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
88 information in FIPS186-3, SP800-57 and SP800-131A.
89 [Steve Henson]
90
91 *) CCM support via EVP. Interface is very similar to GCM case except we
92 must supply all data in one chunk (i.e. no update, final) and the
93 message length must be supplied if AAD is used. Add algorithm test
94 support.
95 [Steve Henson]
96
97 *) Initial version of POST overhaul. Add POST callback to allow the status
98 of POST to be monitored and/or failures induced. Modify fips_test_suite
99 to use callback. Always run all selftests even if one fails.
100 [Steve Henson]
101
102 *) XTS support including algorithm test driver in the fips_gcmtest program.
103 Note: this does increase the maximum key length from 32 to 64 bytes but
104 there should be no binary compatibility issues as existing applications
105 will never use XTS mode.
106 [Steve Henson]
107
108 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
109 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
110 performs algorithm blocking for unapproved PRNG types. Also do not
111 set PRNG type in FIPS_mode_set(): leave this to the application.
112 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
113 the standard OpenSSL PRNG: set additional data to a date time vector.
114 [Steve Henson]
115
116 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
117 This shouldn't present any incompatibility problems because applications
118 shouldn't be using these directly and any that are will need to rethink
119 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
120 [Steve Henson]
121
122 *) Extensive self tests and health checking required by SP800-90 DRBG.
123 Remove strength parameter from FIPS_drbg_instantiate and always
124 instantiate at maximum supported strength.
125 [Steve Henson]
126
127 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
128 [Steve Henson]
129
130 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
131 [Steve Henson]
132
133 *) New function DH_compute_key_padded() to compute a DH key and pad with
134 leading zeroes if needed: this complies with SP800-56A et al.
135 [Steve Henson]
136
137 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
138 anything, incomplete, subject to change and largely untested at present.
139 [Steve Henson]
140
141 *) Modify fipscanisteronly build option to only build the necessary object
142 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
143 [Steve Henson]
144
145 *) Add experimental option FIPSSYMS to give all symbols in
146 fipscanister.o and FIPS or fips prefix. This will avoid
147 conflicts with future versions of OpenSSL. Add perl script
148 util/fipsas.pl to preprocess assembly language source files
149 and rename any affected symbols.
150 [Steve Henson]
151
152 *) Add selftest checks and algorithm block of non-fips algorithms in
153 FIPS mode. Remove DES2 from selftests.
154 [Steve Henson]
155
156 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
157 return internal method without any ENGINE dependencies. Add new
158 tiny fips sign and verify functions.
159 [Steve Henson]
160
161 *) New build option no-ec2m to disable characteristic 2 code.
162 [Steve Henson]
163
164 *) New build option "fipscanisteronly". This only builds fipscanister.o
165 and (currently) associated fips utilities. Uses the file Makefile.fips
166 instead of Makefile.org as the prototype.
167 [Steve Henson]
168
169 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
170 Update fips_gcmtest to use IV generator.
171 [Steve Henson]
172
173 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
174 setting output buffer to NULL. The *Final function must be
175 called although it will not retrieve any additional data. The tag
176 can be set or retrieved with a ctrl. The IV length is by default 12
177 bytes (96 bits) but can be set to an alternative value. If the IV
178 length exceeds the maximum IV length (currently 16 bytes) it cannot be
179 set before the key.
180 [Steve Henson]
181
182 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
183 underlying do_cipher function handles all cipher semantics itself
184 including padding and finalisation. This is useful if (for example)
185 an ENGINE cipher handles block padding itself. The behaviour of
186 do_cipher is subtly changed if this flag is set: the return value
187 is the number of characters written to the output buffer (zero is
188 no longer an error code) or a negative error code. Also if the
189 input buffer is NULL and length 0 finalisation should be performed.
190 [Steve Henson]
191
192 *) If a candidate issuer certificate is already part of the constructed
193 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
194 [Steve Henson]
195
196 *) Improve forward-security support: add functions
197
198 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
199 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
200
201 for use by SSL/TLS servers; the callback function will be called whenever a
202 new session is created, and gets to decide whether the session may be
203 cached to make it resumable (return 0) or not (return 1). (As by the
204 SSL/TLS protocol specifications, the session_id sent by the server will be
205 empty to indicate that the session is not resumable; also, the server will
206 not generate RFC 4507 (RFC 5077) session tickets.)
207
208 A simple reasonable callback implementation is to return is_forward_secure.
209 This parameter will be set to 1 or 0 depending on the ciphersuite selected
210 by the SSL/TLS server library, indicating whether it can provide forward
211 security.
212 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
213
214 *) New function OPENSSL_gmtime_diff to find the difference in days
215 and seconds between two tm structures. This will be used to provide
216 additional functionality for ASN1_TIME.
217 [Steve Henson]
218
219 *) Add -trusted_first option which attempts to find certificates in the
220 trusted store even if an untrusted chain is also supplied.
221 [Steve Henson]
222
223 *) Initial experimental support for explicitly trusted non-root CAs.
224 OpenSSL still tries to build a complete chain to a root but if an
225 intermediate CA has a trust setting included that is used. The first
226 setting is used: whether to trust or reject.
227 [Steve Henson]
228
229 *) New -verify_name option in command line utilities to set verification
230 parameters by name.
231 [Steve Henson]
232
233 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
234 Add CMAC pkey methods.
235 [Steve Henson]
236
237 *) Experimental regnegotiation in s_server -www mode. If the client
238 browses /reneg connection is renegotiated. If /renegcert it is
239 renegotiated requesting a certificate.
240 [Steve Henson]
241
242 *) Add an "external" session cache for debugging purposes to s_server. This
243 should help trace issues which normally are only apparent in deployed
244 multi-process servers.
245 [Steve Henson]
246
247 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
248 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
249 BIO_set_cipher() and some obscure PEM functions were changed so they
250 can now return an error. The RAND changes required a change to the
251 RAND_METHOD structure.
252 [Steve Henson]
253
254 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
255 a gcc attribute to warn if the result of a function is ignored. This
256 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
257 whose return value is often ignored.
258 [Steve Henson]
259
260 Changes between 1.0.0f and 1.0.1 [xx XXX xxxx]
261
262 *) Add support for TLS/DTLS heartbeats.
263 [Robin Seggelmann <seggelmann@fh-muenster.de>]
264
265 *) Add support for SCTP.
266 [Robin Seggelmann <seggelmann@fh-muenster.de>]
267
268 *) Improved PRNG seeding for VOS.
269 [Paul Green <Paul.Green@stratus.com>]
270
271 *) Extensive assembler packs updates, most notably:
272
273 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
274 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
275 - x86_64: bit-sliced AES implementation;
276 - ARM: NEON support, contemporary platforms optimizations;
277 - s390x: z196 support;
278 - *: GHASH and GF(2^m) multiplication implementations;
279
280 [Andy Polyakov]
281
282 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
283 (removal of unnecessary code)
284 [Peter Sylvester <peter.sylvester@edelweb.fr>]
285
286 *) Add TLS key material exporter from RFC 5705.
287 [Eric Rescorla]
288
289 *) Add DTLS-SRTP negotiation from RFC 5764.
290 [Eric Rescorla]
291
292 *) Add Next Protocol Negotiation,
293 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
294 disabled with a no-npn flag to config or Configure. Code donated
295 by Google.
296 [Adam Langley <agl@google.com> and Ben Laurie]
297
298 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
299 NIST-P256, NIST-P521, with constant-time single point multiplication on
300 typical inputs. Compiler support for the nonstandard type __uint128_t is
301 required to use this (present in gcc 4.4 and later, for 64-bit builds).
302 Code made available under Apache License version 2.0.
303
304 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
305 line to include this in your build of OpenSSL, and run "make depend" (or
306 "make update"). This enables the following EC_METHODs:
307
308 EC_GFp_nistp224_method()
309 EC_GFp_nistp256_method()
310 EC_GFp_nistp521_method()
311
312 EC_GROUP_new_by_curve_name() will automatically use these (while
313 EC_GROUP_new_curve_GFp() currently prefers the more flexible
314 implementations).
315 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
316
317 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
318 all platforms. Move ssize_t definition from e_os.h to the public
319 header file e_os2.h as it now appears in public header file cms.h
320 [Steve Henson]
321
322 *) New -sigopt option to the ca, req and x509 utilities. Additional
323 signature parameters can be passed using this option and in
324 particular PSS.
325 [Steve Henson]
326
327 *) Add RSA PSS signing function. This will generate and set the
328 appropriate AlgorithmIdentifiers for PSS based on those in the
329 corresponding EVP_MD_CTX structure. No application support yet.
330 [Steve Henson]
331
332 *) Support for companion algorithm specific ASN1 signing routines.
333 New function ASN1_item_sign_ctx() signs a pre-initialised
334 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
335 the appropriate parameters.
336 [Steve Henson]
337
338 *) Add new algorithm specific ASN1 verification initialisation function
339 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
340 handling will be the same no matter what EVP_PKEY_METHOD is used.
341 Add a PSS handler to support verification of PSS signatures: checked
342 against a number of sample certificates.
343 [Steve Henson]
344
345 *) Add signature printing for PSS. Add PSS OIDs.
346 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
347
348 *) Add algorithm specific signature printing. An individual ASN1 method
349 can now print out signatures instead of the standard hex dump.
350
351 More complex signatures (e.g. PSS) can print out more meaningful
352 information. Include DSA version that prints out the signature
353 parameters r, s.
354 [Steve Henson]
355
356 *) Password based recipient info support for CMS library: implementing
357 RFC3211.
358 [Steve Henson]
359
360 *) Split password based encryption into PBES2 and PBKDF2 functions. This
361 neatly separates the code into cipher and PBE sections and is required
362 for some algorithms that split PBES2 into separate pieces (such as
363 password based CMS).
364 [Steve Henson]
365
366 *) Session-handling fixes:
367 - Fix handling of connections that are resuming with a session ID,
368 but also support Session Tickets.
369 - Fix a bug that suppressed issuing of a new ticket if the client
370 presented a ticket with an expired session.
371 - Try to set the ticket lifetime hint to something reasonable.
372 - Make tickets shorter by excluding irrelevant information.
373 - On the client side, don't ignore renewed tickets.
374 [Adam Langley, Bodo Moeller (Google)]
375
376 *) Fix PSK session representation.
377 [Bodo Moeller]
378
379 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
380
381 This work was sponsored by Intel.
382 [Andy Polyakov]
383
384 *) Add GCM support to TLS library. Some custom code is needed to split
385 the IV between the fixed (from PRF) and explicit (from TLS record)
386 portions. This adds all GCM ciphersuites supported by RFC5288 and
387 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
388 add a special AESGCM string for GCM only.
389 [Steve Henson]
390
391 *) Expand range of ctrls for AES GCM. Permit setting invocation
392 field on decrypt and retrieval of invocation field only on encrypt.
393 [Steve Henson]
394
395 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
396 As required by RFC5289 these ciphersuites cannot be used if for
397 versions of TLS earlier than 1.2.
398 [Steve Henson]
399
400 *) For FIPS capable OpenSSL interpret a NULL default public key method
401 as unset and return the appopriate default but do *not* set the default.
402 This means we can return the appopriate method in applications that
403 swicth between FIPS and non-FIPS modes.
404 [Steve Henson]
405
406 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
407 ENGINE is used then we cannot handle that in the FIPS module so we
408 keep original code iff non-FIPS operations are allowed.
409 [Steve Henson]
410
411 *) Add -attime option to openssl utilities.
412 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
413
414 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
415 [Steve Henson]
416
417 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
418 FIPS EC methods unconditionally for now.
419 [Steve Henson]
420
421 *) New build option no-ec2m to disable characteristic 2 code.
422 [Steve Henson]
423
424 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
425 all cases can be covered as some introduce binary incompatibilities.
426 [Steve Henson]
427
428 *) Redirect RSA operations to FIPS module including keygen,
429 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
430 [Steve Henson]
431
432 *) Add similar low level API blocking to ciphers.
433 [Steve Henson]
434
435 *) Low level digest APIs are not approved in FIPS mode: any attempt
436 to use these will cause a fatal error. Applications that *really* want
437 to use them can use the private_* version instead.
438 [Steve Henson]
439
440 *) Redirect cipher operations to FIPS module for FIPS builds.
441 [Steve Henson]
442
443 *) Redirect digest operations to FIPS module for FIPS builds.
444 [Steve Henson]
445
446 *) Update build system to add "fips" flag which will link in fipscanister.o
447 for static and shared library builds embedding a signature if needed.
448 [Steve Henson]
449
450 *) Output TLS supported curves in preference order instead of numerical
451 order. This is currently hardcoded for the highest order curves first.
452 This should be configurable so applications can judge speed vs strength.
453 [Steve Henson]
454
455 *) Add TLS v1.2 server support for client authentication.
456 [Steve Henson]
457
458 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
459 and enable MD5.
460 [Steve Henson]
461
462 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
463 FIPS modules versions.
464 [Steve Henson]
465
466 *) Add TLS v1.2 client side support for client authentication. Keep cache
467 of handshake records longer as we don't know the hash algorithm to use
468 until after the certificate request message is received.
469 [Steve Henson]
470
471 *) Initial TLS v1.2 client support. Add a default signature algorithms
472 extension including all the algorithms we support. Parse new signature
473 format in client key exchange. Relax some ECC signing restrictions for
474 TLS v1.2 as indicated in RFC5246.
475 [Steve Henson]
476
477 *) Add server support for TLS v1.2 signature algorithms extension. Switch
478 to new signature format when needed using client digest preference.
479 All server ciphersuites should now work correctly in TLS v1.2. No client
480 support yet and no support for client certificates.
481 [Steve Henson]
482
483 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
484 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
485 ciphersuites. At present only RSA key exchange ciphersuites work with
486 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
487 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
488 and version checking.
489 [Steve Henson]
490
491 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
492 with this defined it will not be affected by any changes to ssl internal
493 structures. Add several utility functions to allow openssl application
494 to work with OPENSSL_NO_SSL_INTERN defined.
495 [Steve Henson]
496
497 *) Add SRP support.
498 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
499
500 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
501 [Steve Henson]
502
503 *) Permit abbreviated handshakes when renegotiating using the function
504 SSL_renegotiate_abbreviated().
505 [Robin Seggelmann <seggelmann@fh-muenster.de>]
506
507 *) Add call to ENGINE_register_all_complete() to
508 ENGINE_load_builtin_engines(), so some implementations get used
509 automatically instead of needing explicit application support.
510 [Steve Henson]
511
512 *) Add support for TLS key exporter as described in RFC5705.
513 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
514
515 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
516 a few changes are required:
517
518 Add SSL_OP_NO_TLSv1_1 flag.
519 Add TLSv1_1 methods.
520 Update version checking logic to handle version 1.1.
521 Add explicit IV handling (ported from DTLS code).
522 Add command line options to s_client/s_server.
523 [Steve Henson]
524
525 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
526
527 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
528 Thanks to Antonio Martin, Enterprise Secure Access Research and
529 Development, Cisco Systems, Inc. for discovering this bug and
530 preparing a fix. (CVE-2012-0050)
531 [Antonio Martin]
532
533 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
534
535 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
536 of the Vaudenay padding oracle attack on CBC mode encryption
537 which enables an efficient plaintext recovery attack against
538 the OpenSSL implementation of DTLS. Their attack exploits timing
539 differences arising during decryption processing. A research
540 paper describing this attack can be found at:
541 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
542 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
543 Security Group at Royal Holloway, University of London
544 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
545 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
546 for preparing the fix. (CVE-2011-4108)
547 [Robin Seggelmann, Michael Tuexen]
548
549 *) Clear bytes used for block padding of SSL 3.0 records.
550 (CVE-2011-4576)
551 [Adam Langley (Google)]
552
553 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
554 Kadianakis <desnacked@gmail.com> for discovering this issue and
555 Adam Langley for preparing the fix. (CVE-2011-4619)
556 [Adam Langley (Google)]
557
558 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
559 [Andrey Kulikov <amdeich@gmail.com>]
560
561 *) Prevent malformed RFC3779 data triggering an assertion failure.
562 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
563 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
564 [Rob Austein <sra@hactrn.net>]
565
566 *) Improved PRNG seeding for VOS.
567 [Paul Green <Paul.Green@stratus.com>]
568
569 *) Fix ssl_ciph.c set-up race.
570 [Adam Langley (Google)]
571
572 *) Fix spurious failures in ecdsatest.c.
573 [Emilia Käsper (Google)]
574
575 *) Fix the BIO_f_buffer() implementation (which was mixing different
576 interpretations of the '..._len' fields).
577 [Adam Langley (Google)]
578
579 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
580 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
581 threads won't reuse the same blinding coefficients.
582
583 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
584 lock to call BN_BLINDING_invert_ex, and avoids one use of
585 BN_BLINDING_update for each BN_BLINDING structure (previously,
586 the last update always remained unused).
587 [Emilia Käsper (Google)]
588
589 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
590 [Bob Buckholz (Google)]
591
592 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
593
594 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
595 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
596 [Kaspar Brand <ossl@velox.ch>]
597
598 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
599 for multi-threaded use of ECDH. (CVE-2011-3210)
600 [Adam Langley (Google)]
601
602 *) Fix x509_name_ex_d2i memory leak on bad inputs.
603 [Bodo Moeller]
604
605 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
606 signature public key algorithm by using OID xref utilities instead.
607 Before this you could only use some ECC ciphersuites with SHA1 only.
608 [Steve Henson]
609
610 *) Add protection against ECDSA timing attacks as mentioned in the paper
611 by Billy Bob Brumley and Nicola Tuveri, see:
612
613 http://eprint.iacr.org/2011/232.pdf
614
615 [Billy Bob Brumley and Nicola Tuveri]
616
617 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
618
619 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
620 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
621
622 *) Fix bug in string printing code: if *any* escaping is enabled we must
623 escape the escape character (backslash) or the resulting string is
624 ambiguous.
625 [Steve Henson]
626
627 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
628
629 *) Disable code workaround for ancient and obsolete Netscape browsers
630 and servers: an attacker can use it in a ciphersuite downgrade attack.
631 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
632 [Steve Henson]
633
634 *) Fixed J-PAKE implementation error, originally discovered by
635 Sebastien Martini, further info and confirmation from Stefan
636 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
637 [Ben Laurie]
638
639 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
640
641 *) Fix extension code to avoid race conditions which can result in a buffer
642 overrun vulnerability: resumed sessions must not be modified as they can
643 be shared by multiple threads. CVE-2010-3864
644 [Steve Henson]
645
646 *) Fix WIN32 build system to correctly link an ENGINE directory into
647 a DLL.
648 [Steve Henson]
649
650 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
651
652 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
653 (CVE-2010-1633)
654 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
655
656 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
657
658 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
659 context. The operation can be customised via the ctrl mechanism in
660 case ENGINEs want to include additional functionality.
661 [Steve Henson]
662
663 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
664 [Steve Henson]
665
666 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
667 output hashes compatible with older versions of OpenSSL.
668 [Willy Weisz <weisz@vcpc.univie.ac.at>]
669
670 *) Fix compression algorithm handling: if resuming a session use the
671 compression algorithm of the resumed session instead of determining
672 it from client hello again. Don't allow server to change algorithm.
673 [Steve Henson]
674
675 *) Add load_crls() function to apps tidying load_certs() too. Add option
676 to verify utility to allow additional CRLs to be included.
677 [Steve Henson]
678
679 *) Update OCSP request code to permit adding custom headers to the request:
680 some responders need this.
681 [Steve Henson]
682
683 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
684 correctly.
685 [Julia Lawall <julia@diku.dk>]
686
687 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
688 needlessly dereferenced structures, used obsolete functions and
689 didn't handle all updated verify codes correctly.
690 [Steve Henson]
691
692 *) Disable MD2 in the default configuration.
693 [Steve Henson]
694
695 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
696 indicate the initial BIO being pushed or popped. This makes it possible
697 to determine whether the BIO is the one explicitly called or as a result
698 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
699 it handles reference counts correctly and doesn't zero out the I/O bio
700 when it is not being explicitly popped. WARNING: applications which
701 included workarounds for the old buggy behaviour will need to be modified
702 or they could free up already freed BIOs.
703 [Steve Henson]
704
705 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
706 renaming to all platforms (within the 0.9.8 branch, this was
707 done conditionally on Netware platforms to avoid a name clash).
708 [Guenter <lists@gknw.net>]
709
710 *) Add ECDHE and PSK support to DTLS.
711 [Michael Tuexen <tuexen@fh-muenster.de>]
712
713 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
714 be used on C++.
715 [Steve Henson]
716
717 *) Add "missing" function EVP_MD_flags() (without this the only way to
718 retrieve a digest flags is by accessing the structure directly. Update
719 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
720 or cipher is registered as in the "from" argument. Print out all
721 registered digests in the dgst usage message instead of manually
722 attempting to work them out.
723 [Steve Henson]
724
725 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
726 this allows the use of compression and extensions. Change default cipher
727 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
728 by default unless an application cipher string requests it.
729 [Steve Henson]
730
731 *) Alter match criteria in PKCS12_parse(). It used to try to use local
732 key ids to find matching certificates and keys but some PKCS#12 files
733 don't follow the (somewhat unwritten) rules and this strategy fails.
734 Now just gather all certificates together and the first private key
735 then look for the first certificate that matches the key.
736 [Steve Henson]
737
738 *) Support use of registered digest and cipher names for dgst and cipher
739 commands instead of having to add each one as a special case. So now
740 you can do:
741
742 openssl sha256 foo
743
744 as well as:
745
746 openssl dgst -sha256 foo
747
748 and this works for ENGINE based algorithms too.
749
750 [Steve Henson]
751
752 *) Update Gost ENGINE to support parameter files.
753 [Victor B. Wagner <vitus@cryptocom.ru>]
754
755 *) Support GeneralizedTime in ca utility.
756 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
757
758 *) Enhance the hash format used for certificate directory links. The new
759 form uses the canonical encoding (meaning equivalent names will work
760 even if they aren't identical) and uses SHA1 instead of MD5. This form
761 is incompatible with the older format and as a result c_rehash should
762 be used to rebuild symbolic links.
763 [Steve Henson]
764
765 *) Make PKCS#8 the default write format for private keys, replacing the
766 traditional format. This form is standardised, more secure and doesn't
767 include an implicit MD5 dependency.
768 [Steve Henson]
769
770 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
771 committed to OpenSSL should pass this lot as a minimum.
772 [Steve Henson]
773
774 *) Add session ticket override functionality for use by EAP-FAST.
775 [Jouni Malinen <j@w1.fi>]
776
777 *) Modify HMAC functions to return a value. Since these can be implemented
778 in an ENGINE errors can occur.
779 [Steve Henson]
780
781 *) Type-checked OBJ_bsearch_ex.
782 [Ben Laurie]
783
784 *) Type-checked OBJ_bsearch. Also some constification necessitated
785 by type-checking. Still to come: TXT_DB, bsearch(?),
786 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
787 CONF_VALUE.
788 [Ben Laurie]
789
790 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
791 seconds to a tm structure directly, instead of going through OS
792 specific date routines. This avoids any issues with OS routines such
793 as the year 2038 bug. New *_adj() functions for ASN1 time structures
794 and X509_time_adj_ex() to cover the extended range. The existing
795 X509_time_adj() is still usable and will no longer have any date issues.
796 [Steve Henson]
797
798 *) Delta CRL support. New use deltas option which will attempt to locate
799 and search any appropriate delta CRLs available.
800
801 This work was sponsored by Google.
802 [Steve Henson]
803
804 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
805 code and add additional score elements. Validate alternate CRL paths
806 as part of the CRL checking and indicate a new error "CRL path validation
807 error" in this case. Applications wanting additional details can use
808 the verify callback and check the new "parent" field. If this is not
809 NULL CRL path validation is taking place. Existing applications wont
810 see this because it requires extended CRL support which is off by
811 default.
812
813 This work was sponsored by Google.
814 [Steve Henson]
815
816 *) Support for freshest CRL extension.
817
818 This work was sponsored by Google.
819 [Steve Henson]
820
821 *) Initial indirect CRL support. Currently only supported in the CRLs
822 passed directly and not via lookup. Process certificate issuer
823 CRL entry extension and lookup CRL entries by bother issuer name
824 and serial number. Check and process CRL issuer entry in IDP extension.
825
826 This work was sponsored by Google.
827 [Steve Henson]
828
829 *) Add support for distinct certificate and CRL paths. The CRL issuer
830 certificate is validated separately in this case. Only enabled if
831 an extended CRL support flag is set: this flag will enable additional
832 CRL functionality in future.
833
834 This work was sponsored by Google.
835 [Steve Henson]
836
837 *) Add support for policy mappings extension.
838
839 This work was sponsored by Google.
840 [Steve Henson]
841
842 *) Fixes to pathlength constraint, self issued certificate handling,
843 policy processing to align with RFC3280 and PKITS tests.
844
845 This work was sponsored by Google.
846 [Steve Henson]
847
848 *) Support for name constraints certificate extension. DN, email, DNS
849 and URI types are currently supported.
850
851 This work was sponsored by Google.
852 [Steve Henson]
853
854 *) To cater for systems that provide a pointer-based thread ID rather
855 than numeric, deprecate the current numeric thread ID mechanism and
856 replace it with a structure and associated callback type. This
857 mechanism allows a numeric "hash" to be extracted from a thread ID in
858 either case, and on platforms where pointers are larger than 'long',
859 mixing is done to help ensure the numeric 'hash' is usable even if it
860 can't be guaranteed unique. The default mechanism is to use "&errno"
861 as a pointer-based thread ID to distinguish between threads.
862
863 Applications that want to provide their own thread IDs should now use
864 CRYPTO_THREADID_set_callback() to register a callback that will call
865 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
866
867 Note that ERR_remove_state() is now deprecated, because it is tied
868 to the assumption that thread IDs are numeric. ERR_remove_state(0)
869 to free the current thread's error state should be replaced by
870 ERR_remove_thread_state(NULL).
871
872 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
873 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
874 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
875 application was previously providing a numeric thread callback that
876 was inappropriate for distinguishing threads, then uniqueness might
877 have been obtained with &errno that happened immediately in the
878 intermediate development versions of OpenSSL; this is no longer the
879 case, the numeric thread callback will now override the automatic use
880 of &errno.)
881 [Geoff Thorpe, with help from Bodo Moeller]
882
883 *) Initial support for different CRL issuing certificates. This covers a
884 simple case where the self issued certificates in the chain exist and
885 the real CRL issuer is higher in the existing chain.
886
887 This work was sponsored by Google.
888 [Steve Henson]
889
890 *) Removed effectively defunct crypto/store from the build.
891 [Ben Laurie]
892
893 *) Revamp of STACK to provide stronger type-checking. Still to come:
894 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
895 ASN1_STRING, CONF_VALUE.
896 [Ben Laurie]
897
898 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
899 RAM on SSL connections. This option can save about 34k per idle SSL.
900 [Nick Mathewson]
901
902 *) Revamp of LHASH to provide stronger type-checking. Still to come:
903 STACK, TXT_DB, bsearch, qsort.
904 [Ben Laurie]
905
906 *) Initial support for Cryptographic Message Syntax (aka CMS) based
907 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
908 support for data, signedData, compressedData, digestedData and
909 encryptedData, envelopedData types included. Scripts to check against
910 RFC4134 examples draft and interop and consistency checks of many
911 content types and variants.
912 [Steve Henson]
913
914 *) Add options to enc utility to support use of zlib compression BIO.
915 [Steve Henson]
916
917 *) Extend mk1mf to support importing of options and assembly language
918 files from Configure script, currently only included in VC-WIN32.
919 The assembly language rules can now optionally generate the source
920 files from the associated perl scripts.
921 [Steve Henson]
922
923 *) Implement remaining functionality needed to support GOST ciphersuites.
924 Interop testing has been performed using CryptoPro implementations.
925 [Victor B. Wagner <vitus@cryptocom.ru>]
926
927 *) s390x assembler pack.
928 [Andy Polyakov]
929
930 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
931 "family."
932 [Andy Polyakov]
933
934 *) Implement Opaque PRF Input TLS extension as specified in
935 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
936 official specification yet and no extension type assignment by
937 IANA exists, this extension (for now) will have to be explicitly
938 enabled when building OpenSSL by providing the extension number
939 to use. For example, specify an option
940
941 -DTLSEXT_TYPE_opaque_prf_input=0x9527
942
943 to the "config" or "Configure" script to enable the extension,
944 assuming extension number 0x9527 (which is a completely arbitrary
945 and unofficial assignment based on the MD5 hash of the Internet
946 Draft). Note that by doing so, you potentially lose
947 interoperability with other TLS implementations since these might
948 be using the same extension number for other purposes.
949
950 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
951 opaque PRF input value to use in the handshake. This will create
952 an interal copy of the length-'len' string at 'src', and will
953 return non-zero for success.
954
955 To get more control and flexibility, provide a callback function
956 by using
957
958 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
959 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
960
961 where
962
963 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
964 void *arg;
965
966 Callback function 'cb' will be called in handshakes, and is
967 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
968 Argument 'arg' is for application purposes (the value as given to
969 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
970 be provided to the callback function). The callback function
971 has to return non-zero to report success: usually 1 to use opaque
972 PRF input just if possible, or 2 to enforce use of the opaque PRF
973 input. In the latter case, the library will abort the handshake
974 if opaque PRF input is not successfully negotiated.
975
976 Arguments 'peerinput' and 'len' given to the callback function
977 will always be NULL and 0 in the case of a client. A server will
978 see the client's opaque PRF input through these variables if
979 available (NULL and 0 otherwise). Note that if the server
980 provides an opaque PRF input, the length must be the same as the
981 length of the client's opaque PRF input.
982
983 Note that the callback function will only be called when creating
984 a new session (session resumption can resume whatever was
985 previously negotiated), and will not be called in SSL 2.0
986 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
987 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
988 for applications that need to enforce opaque PRF input.
989
990 [Bodo Moeller]
991
992 *) Update ssl code to support digests other than SHA1+MD5 for handshake
993 MAC.
994
995 [Victor B. Wagner <vitus@cryptocom.ru>]
996
997 *) Add RFC4507 support to OpenSSL. This includes the corrections in
998 RFC4507bis. The encrypted ticket format is an encrypted encoded
999 SSL_SESSION structure, that way new session features are automatically
1000 supported.
1001
1002 If a client application caches session in an SSL_SESSION structure
1003 support is transparent because tickets are now stored in the encoded
1004 SSL_SESSION.
1005
1006 The SSL_CTX structure automatically generates keys for ticket
1007 protection in servers so again support should be possible
1008 with no application modification.
1009
1010 If a client or server wishes to disable RFC4507 support then the option
1011 SSL_OP_NO_TICKET can be set.
1012
1013 Add a TLS extension debugging callback to allow the contents of any client
1014 or server extensions to be examined.
1015
1016 This work was sponsored by Google.
1017 [Steve Henson]
1018
1019 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1020 OpenSSL should now compile cleanly on gcc 4.2
1021 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1022
1023 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1024 support including streaming MAC support: this is required for GOST
1025 ciphersuite support.
1026 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1027
1028 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1029 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1030 to output in BER and PEM format.
1031 [Steve Henson]
1032
1033 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1034 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1035 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1036 ENGINE support for HMAC keys which are unextractable. New -mac and
1037 -macopt options to dgst utility.
1038 [Steve Henson]
1039
1040 *) New option -sigopt to dgst utility. Update dgst to use
1041 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1042 alternative signing paramaters such as X9.31 or PSS in the dgst
1043 utility.
1044 [Steve Henson]
1045
1046 *) Change ssl_cipher_apply_rule(), the internal function that does
1047 the work each time a ciphersuite string requests enabling
1048 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1049 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1050 the order of disabled ciphersuites such that those ciphersuites
1051 that most recently went from enabled to disabled not only stay
1052 in order with respect to each other, but also have higher priority
1053 than other disabled ciphersuites the next time ciphersuites are
1054 enabled again.
1055
1056 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1057 the same ciphersuites as with "HIGH" alone, but in a specific
1058 order where the PSK ciphersuites come first (since they are the
1059 most recently disabled ciphersuites when "HIGH" is parsed).
1060
1061 Also, change ssl_create_cipher_list() (using this new
1062 funcionality) such that between otherwise identical
1063 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1064 the default order.
1065 [Bodo Moeller]
1066
1067 *) Change ssl_create_cipher_list() so that it automatically
1068 arranges the ciphersuites in reasonable order before starting
1069 to process the rule string. Thus, the definition for "DEFAULT"
1070 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1071 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1072 This makes it much easier to arrive at a reasonable default order
1073 in applications for which anonymous ciphers are OK (meaning
1074 that you can't actually use DEFAULT).
1075 [Bodo Moeller; suggested by Victor Duchovni]
1076
1077 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1078 processing) into multiple integers instead of setting
1079 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1080 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1081 (These masks as well as the individual bit definitions are hidden
1082 away into the non-exported interface ssl/ssl_locl.h, so this
1083 change to the definition of the SSL_CIPHER structure shouldn't
1084 affect applications.) This give us more bits for each of these
1085 categories, so there is no longer a need to coagulate AES128 and
1086 AES256 into a single algorithm bit, and to coagulate Camellia128
1087 and Camellia256 into a single algorithm bit, which has led to all
1088 kinds of kludges.
1089
1090 Thus, among other things, the kludge introduced in 0.9.7m and
1091 0.9.8e for masking out AES256 independently of AES128 or masking
1092 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1093
1094 With the change, we also introduce new ciphersuite aliases that
1095 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1096 "CAMELLIA256".
1097 [Bodo Moeller]
1098
1099 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1100 Use the leftmost N bytes of the signature input if the input is
1101 larger than the prime q (with N being the size in bytes of q).
1102 [Nils Larsch]
1103
1104 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1105 it yet and it is largely untested.
1106 [Steve Henson]
1107
1108 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1109 [Nils Larsch]
1110
1111 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1112 some compilers (gcc 4.2 and later) reject their use. Safestack is
1113 reimplemented. Update ASN1 to avoid use of legacy functions.
1114 [Steve Henson]
1115
1116 *) Win32/64 targets are linked with Winsock2.
1117 [Andy Polyakov]
1118
1119 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1120 to external functions. This can be used to increase CRL handling
1121 efficiency especially when CRLs are very large by (for example) storing
1122 the CRL revoked certificates in a database.
1123 [Steve Henson]
1124
1125 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1126 new CRLs added to a directory can be used. New command line option
1127 -verify_return_error to s_client and s_server. This causes real errors
1128 to be returned by the verify callback instead of carrying on no matter
1129 what. This reflects the way a "real world" verify callback would behave.
1130 [Steve Henson]
1131
1132 *) GOST engine, supporting several GOST algorithms and public key formats.
1133 Kindly donated by Cryptocom.
1134 [Cryptocom]
1135
1136 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1137 partitioned by DP are handled but no indirect CRL or reason partitioning
1138 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1139 selected via a scoring technique which handles IDP and AKID in CRLs.
1140 [Steve Henson]
1141
1142 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1143 will ultimately be used for all verify operations: this will remove the
1144 X509_STORE dependency on certificate verification and allow alternative
1145 lookup methods. X509_STORE based implementations of these two callbacks.
1146 [Steve Henson]
1147
1148 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1149 Modify get_crl() to find a valid (unexpired) CRL if possible.
1150 [Steve Henson]
1151
1152 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1153 this would be called X509_CRL_cmp() but that name is already used by
1154 a function that just compares CRL issuer names. Cache several CRL
1155 extensions in X509_CRL structure and cache CRLDP in X509.
1156 [Steve Henson]
1157
1158 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1159 this maps equivalent X509_NAME structures into a consistent structure.
1160 Name comparison can then be performed rapidly using memcmp().
1161 [Steve Henson]
1162
1163 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1164 utility.
1165 [Steve Henson]
1166
1167 *) Allow digests to supply their own micalg string for S/MIME type using
1168 the ctrl EVP_MD_CTRL_MICALG.
1169 [Steve Henson]
1170
1171 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1172 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1173 ctrl. It can then customise the structure before and/or after signing
1174 if necessary.
1175 [Steve Henson]
1176
1177 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1178 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1179 to free up any added signature OIDs.
1180 [Steve Henson]
1181
1182 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1183 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1184 digest and cipher tables. New options added to openssl utility:
1185 list-message-digest-algorithms and list-cipher-algorithms.
1186 [Steve Henson]
1187
1188 *) Change the array representation of binary polynomials: the list
1189 of degrees of non-zero coefficients is now terminated with -1.
1190 Previously it was terminated with 0, which was also part of the
1191 value; thus, the array representation was not applicable to
1192 polynomials where t^0 has coefficient zero. This change makes
1193 the array representation useful in a more general context.
1194 [Douglas Stebila]
1195
1196 *) Various modifications and fixes to SSL/TLS cipher string
1197 handling. For ECC, the code now distinguishes between fixed ECDH
1198 with RSA certificates on the one hand and with ECDSA certificates
1199 on the other hand, since these are separate ciphersuites. The
1200 unused code for Fortezza ciphersuites has been removed.
1201
1202 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1203 (not "ECDHE"). For consistency with the code for DH
1204 certificates, use of ECDH certificates is now considered ECDH
1205 authentication, not RSA or ECDSA authentication (the latter is
1206 merely the CA's signing algorithm and not actively used in the
1207 protocol).
1208
1209 The temporary ciphersuite alias "ECCdraft" is no longer
1210 available, and ECC ciphersuites are no longer excluded from "ALL"
1211 and "DEFAULT". The following aliases now exist for RFC 4492
1212 ciphersuites, most of these by analogy with the DH case:
1213
1214 kECDHr - ECDH cert, signed with RSA
1215 kECDHe - ECDH cert, signed with ECDSA
1216 kECDH - ECDH cert (signed with either RSA or ECDSA)
1217 kEECDH - ephemeral ECDH
1218 ECDH - ECDH cert or ephemeral ECDH
1219
1220 aECDH - ECDH cert
1221 aECDSA - ECDSA cert
1222 ECDSA - ECDSA cert
1223
1224 AECDH - anonymous ECDH
1225 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1226
1227 [Bodo Moeller]
1228
1229 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1230 Use correct micalg parameters depending on digest(s) in signed message.
1231 [Steve Henson]
1232
1233 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1234 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1235 [Steve Henson]
1236
1237 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1238 an engine to register a method. Add ENGINE lookups for methods and
1239 functional reference processing.
1240 [Steve Henson]
1241
1242 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1243 EVP_{Sign,Verify}* which allow an application to customise the signature
1244 process.
1245 [Steve Henson]
1246
1247 *) New -resign option to smime utility. This adds one or more signers
1248 to an existing PKCS#7 signedData structure. Also -md option to use an
1249 alternative message digest algorithm for signing.
1250 [Steve Henson]
1251
1252 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1253 create PKCS7 structures containing multiple signers. Update smime
1254 application to support multiple signers.
1255 [Steve Henson]
1256
1257 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1258 digest MAC.
1259 [Steve Henson]
1260
1261 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1262 Reorganize PBE internals to lookup from a static table using NIDs,
1263 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1264 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1265 PRF which will be automatically used with PBES2.
1266 [Steve Henson]
1267
1268 *) Replace the algorithm specific calls to generate keys in "req" with the
1269 new API.
1270 [Steve Henson]
1271
1272 *) Update PKCS#7 enveloped data routines to use new API. This is now
1273 supported by any public key method supporting the encrypt operation. A
1274 ctrl is added to allow the public key algorithm to examine or modify
1275 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1276 a no op.
1277 [Steve Henson]
1278
1279 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1280 a default digest type to use. In most cases this will be SHA1 but some
1281 algorithms (such as GOST) need to specify an alternative digest. The
1282 return value indicates how strong the prefernce is 1 means optional and
1283 2 is mandatory (that is it is the only supported type). Modify
1284 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1285 use the default md. Update openssl utilities to use the default digest
1286 type for signing if it is not explicitly indicated.
1287 [Steve Henson]
1288
1289 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1290 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1291 signing method from the key type. This effectively removes the link
1292 between digests and public key types.
1293 [Steve Henson]
1294
1295 *) Add an OID cross reference table and utility functions. Its purpose is to
1296 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1297 rsaEncryption. This will allow some of the algorithm specific hackery
1298 needed to use the correct OID to be removed.
1299 [Steve Henson]
1300
1301 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1302 structures for PKCS7_sign(). They are now set up by the relevant public
1303 key ASN1 method.
1304 [Steve Henson]
1305
1306 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1307 [Steve Henson]
1308
1309 *) Add support for key derivation (agreement) in the API, DH method and
1310 pkeyutl.
1311 [Steve Henson]
1312
1313 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1314 public and private key formats. As a side effect these add additional
1315 command line functionality not previously available: DSA signatures can be
1316 generated and verified using pkeyutl and DH key support and generation in
1317 pkey, genpkey.
1318 [Steve Henson]
1319
1320 *) BeOS support.
1321 [Oliver Tappe <zooey@hirschkaefer.de>]
1322
1323 *) New make target "install_html_docs" installs HTML renditions of the
1324 manual pages.
1325 [Oliver Tappe <zooey@hirschkaefer.de>]
1326
1327 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1328 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1329 support key and parameter generation and add initial key generation
1330 functionality for RSA.
1331 [Steve Henson]
1332
1333 *) Add functions for main EVP_PKEY_method operations. The undocumented
1334 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1335 EVP_PKEY_{encrypt,decrypt}_old.
1336 [Steve Henson]
1337
1338 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1339 key API, doesn't do much yet.
1340 [Steve Henson]
1341
1342 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1343 public key algorithms. New option to openssl utility:
1344 "list-public-key-algorithms" to print out info.
1345 [Steve Henson]
1346
1347 *) Implement the Supported Elliptic Curves Extension for
1348 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1349 [Douglas Stebila]
1350
1351 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1352 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1353 [Steve Henson]
1354
1355 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1356 utilities such as rsa, dsa, dsaparam etc except they process any key
1357 type.
1358 [Steve Henson]
1359
1360 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1361 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1362 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1363 structure.
1364 [Steve Henson]
1365
1366 *) Initial support for pluggable public key ASN1.
1367 De-spaghettify the public key ASN1 handling. Move public and private
1368 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1369 algorithm specific handling to a single module within the relevant
1370 algorithm directory. Add functions to allow (near) opaque processing
1371 of public and private key structures.
1372 [Steve Henson]
1373
1374 *) Implement the Supported Point Formats Extension for
1375 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1376 [Douglas Stebila]
1377
1378 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1379 for the psk identity [hint] and the psk callback functions to the
1380 SSL_SESSION, SSL and SSL_CTX structure.
1381
1382 New ciphersuites:
1383 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1384 PSK-AES256-CBC-SHA
1385
1386 New functions:
1387 SSL_CTX_use_psk_identity_hint
1388 SSL_get_psk_identity_hint
1389 SSL_get_psk_identity
1390 SSL_use_psk_identity_hint
1391
1392 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1393
1394 *) Add RFC 3161 compliant time stamp request creation, response generation
1395 and response verification functionality.
1396 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1397
1398 *) Add initial support for TLS extensions, specifically for the server_name
1399 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1400 have new members for a host name. The SSL data structure has an
1401 additional member SSL_CTX *initial_ctx so that new sessions can be
1402 stored in that context to allow for session resumption, even after the
1403 SSL has been switched to a new SSL_CTX in reaction to a client's
1404 server_name extension.
1405
1406 New functions (subject to change):
1407
1408 SSL_get_servername()
1409 SSL_get_servername_type()
1410 SSL_set_SSL_CTX()
1411
1412 New CTRL codes and macros (subject to change):
1413
1414 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1415 - SSL_CTX_set_tlsext_servername_callback()
1416 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1417 - SSL_CTX_set_tlsext_servername_arg()
1418 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1419
1420 openssl s_client has a new '-servername ...' option.
1421
1422 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1423 '-key2 ...', '-servername_fatal' (subject to change). This allows
1424 testing the HostName extension for a specific single host name ('-cert'
1425 and '-key' remain fallbacks for handshakes without HostName
1426 negotiation). If the unrecogninzed_name alert has to be sent, this by
1427 default is a warning; it becomes fatal with the '-servername_fatal'
1428 option.
1429
1430 [Peter Sylvester, Remy Allais, Christophe Renou]
1431
1432 *) Whirlpool hash implementation is added.
1433 [Andy Polyakov]
1434
1435 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1436 bn(64,32). Because of instruction set limitations it doesn't have
1437 any negative impact on performance. This was done mostly in order
1438 to make it possible to share assembler modules, such as bn_mul_mont
1439 implementations, between 32- and 64-bit builds without hassle.
1440 [Andy Polyakov]
1441
1442 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1443 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1444 macro.
1445 [Bodo Moeller]
1446
1447 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1448 dedicated Montgomery multiplication procedure, is introduced.
1449 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1450 "64-bit" performance on certain 32-bit targets.
1451 [Andy Polyakov]
1452
1453 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1454 in SSL structures. New SSL ctrl to set maximum send fragment size.
1455 Save memory by seeting the I/O buffer sizes dynamically instead of
1456 using the maximum available value.
1457 [Steve Henson]
1458
1459 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1460 in addition to the text details.
1461 [Bodo Moeller]
1462
1463 *) Very, very preliminary EXPERIMENTAL support for printing of general
1464 ASN1 structures. This currently produces rather ugly output and doesn't
1465 handle several customised structures at all.
1466 [Steve Henson]
1467
1468 *) Integrated support for PVK file format and some related formats such
1469 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1470 these in the 'rsa' and 'dsa' utilities.
1471 [Steve Henson]
1472
1473 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1474 [Steve Henson]
1475
1476 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1477 place for the (very old) "NETSCAPE" format certificates which are now
1478 handled using new ASN1 code equivalents.
1479 [Steve Henson]
1480
1481 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1482 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1483 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1484 [Nils Larsch]
1485
1486 *) Modify CRL distribution points extension code to print out previously
1487 unsupported fields. Enhance extension setting code to allow setting of
1488 all fields.
1489 [Steve Henson]
1490
1491 *) Add print and set support for Issuing Distribution Point CRL extension.
1492 [Steve Henson]
1493
1494 *) Change 'Configure' script to enable Camellia by default.
1495 [NTT]
1496
1497 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1498
1499 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1500 Thanks to Antonio Martin, Enterprise Secure Access Research and
1501 Development, Cisco Systems, Inc. for discovering this bug and
1502 preparing a fix. (CVE-2012-0050)
1503 [Antonio Martin]
1504
1505 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1506
1507 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1508 of the Vaudenay padding oracle attack on CBC mode encryption
1509 which enables an efficient plaintext recovery attack against
1510 the OpenSSL implementation of DTLS. Their attack exploits timing
1511 differences arising during decryption processing. A research
1512 paper describing this attack can be found at:
1513 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1514 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1515 Security Group at Royal Holloway, University of London
1516 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1517 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1518 for preparing the fix. (CVE-2011-4108)
1519 [Robin Seggelmann, Michael Tuexen]
1520
1521 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1522 [Ben Laurie, Kasper <ekasper@google.com>]
1523
1524 *) Clear bytes used for block padding of SSL 3.0 records.
1525 (CVE-2011-4576)
1526 [Adam Langley (Google)]
1527
1528 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1529 Kadianakis <desnacked@gmail.com> for discovering this issue and
1530 Adam Langley for preparing the fix. (CVE-2011-4619)
1531 [Adam Langley (Google)]
1532
1533 *) Prevent malformed RFC3779 data triggering an assertion failure.
1534 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1535 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1536 [Rob Austein <sra@hactrn.net>]
1537
1538 *) Fix ssl_ciph.c set-up race.
1539 [Adam Langley (Google)]
1540
1541 *) Fix spurious failures in ecdsatest.c.
1542 [Emilia Käsper (Google)]
1543
1544 *) Fix the BIO_f_buffer() implementation (which was mixing different
1545 interpretations of the '..._len' fields).
1546 [Adam Langley (Google)]
1547
1548 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1549 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1550 threads won't reuse the same blinding coefficients.
1551
1552 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1553 lock to call BN_BLINDING_invert_ex, and avoids one use of
1554 BN_BLINDING_update for each BN_BLINDING structure (previously,
1555 the last update always remained unused).
1556 [Emilia Käsper (Google)]
1557
1558 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1559 for multi-threaded use of ECDH.
1560 [Adam Langley (Google)]
1561
1562 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1563 [Bodo Moeller]
1564
1565 *) Add protection against ECDSA timing attacks as mentioned in the paper
1566 by Billy Bob Brumley and Nicola Tuveri, see:
1567
1568 http://eprint.iacr.org/2011/232.pdf
1569
1570 [Billy Bob Brumley and Nicola Tuveri]
1571
1572 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1573
1574 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1575 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1576
1577 *) Fix bug in string printing code: if *any* escaping is enabled we must
1578 escape the escape character (backslash) or the resulting string is
1579 ambiguous.
1580 [Steve Henson]
1581
1582 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1583
1584 *) Disable code workaround for ancient and obsolete Netscape browsers
1585 and servers: an attacker can use it in a ciphersuite downgrade attack.
1586 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1587 [Steve Henson]
1588
1589 *) Fixed J-PAKE implementation error, originally discovered by
1590 Sebastien Martini, further info and confirmation from Stefan
1591 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1592 [Ben Laurie]
1593
1594 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1595
1596 *) Fix extension code to avoid race conditions which can result in a buffer
1597 overrun vulnerability: resumed sessions must not be modified as they can
1598 be shared by multiple threads. CVE-2010-3864
1599 [Steve Henson]
1600
1601 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1602 [Steve Henson]
1603
1604 *) Don't reencode certificate when calculating signature: cache and use
1605 the original encoding instead. This makes signature verification of
1606 some broken encodings work correctly.
1607 [Steve Henson]
1608
1609 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1610 is also one of the inputs.
1611 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1612
1613 *) Don't repeatedly append PBE algorithms to table if they already exist.
1614 Sort table on each new add. This effectively makes the table read only
1615 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1616 etc are non-op.
1617 [Steve Henson]
1618
1619 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1620
1621 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1622 OpenSSL 1.0.0.]
1623
1624 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1625 access or freeing data twice (CVE-2010-0742)
1626 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1627
1628 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1629 common in certificates and some applications which only call
1630 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1631 [Steve Henson]
1632
1633 *) VMS fixes:
1634 Reduce copying into .apps and .test in makevms.com
1635 Don't try to use blank CA certificate in CA.com
1636 Allow use of C files from original directories in maketests.com
1637 [Steven M. Schweda" <sms@antinode.info>]
1638
1639 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1640
1641 *) When rejecting SSL/TLS records due to an incorrect version number, never
1642 update s->server with a new major version number. As of
1643 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1644 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1645 the previous behavior could result in a read attempt at NULL when
1646 receiving specific incorrect SSL/TLS records once record payload
1647 protection is active. (CVE-2010-0740)
1648 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1649
1650 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1651 could be crashed if the relevant tables were not present (e.g. chrooted).
1652 [Tomas Hoger <thoger@redhat.com>]
1653
1654 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1655
1656 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1657 [Martin Olsson, Neel Mehta]
1658
1659 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1660 accommodate for stack sorting, always a write lock!).
1661 [Bodo Moeller]
1662
1663 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1664 excessive delays in the RAND_poll(): over a minute. As a workaround
1665 include a time check in the inner Heap32Next loop too.
1666 [Steve Henson]
1667
1668 *) The code that handled flushing of data in SSL/TLS originally used the
1669 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1670 the problem outlined in PR#1949. The fix suggested there however can
1671 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1672 of Apache). So instead simplify the code to flush unconditionally.
1673 This should be fine since flushing with no data to flush is a no op.
1674 [Steve Henson]
1675
1676 *) Handle TLS versions 2.0 and later properly and correctly use the
1677 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1678 off ancient servers have a habit of sticking around for a while...
1679 [Steve Henson]
1680
1681 *) Modify compression code so it frees up structures without using the
1682 ex_data callbacks. This works around a problem where some applications
1683 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1684 restarting) then use compression (e.g. SSL with compression) later.
1685 This results in significant per-connection memory leaks and
1686 has caused some security issues including CVE-2008-1678 and
1687 CVE-2009-4355.
1688 [Steve Henson]
1689
1690 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1691 change when encrypting or decrypting.
1692 [Bodo Moeller]
1693
1694 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1695 connect and renegotiate with servers which do not support RI.
1696 Until RI is more widely deployed this option is enabled by default.
1697 [Steve Henson]
1698
1699 *) Add "missing" ssl ctrls to clear options and mode.
1700 [Steve Henson]
1701
1702 *) If client attempts to renegotiate and doesn't support RI respond with
1703 a no_renegotiation alert as required by RFC5746. Some renegotiating
1704 TLS clients will continue a connection gracefully when they receive
1705 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1706 waiting for a server hello which it will never receive. Now we treat a
1707 received no_renegotiation alert as a fatal error. This is because
1708 applications requesting a renegotiation might well expect it to succeed
1709 and would have no code in place to handle the server denying it so the
1710 only safe thing to do is to terminate the connection.
1711 [Steve Henson]
1712
1713 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1714 peer supports secure renegotiation and 0 otherwise. Print out peer
1715 renegotiation support in s_client/s_server.
1716 [Steve Henson]
1717
1718 *) Replace the highly broken and deprecated SPKAC certification method with
1719 the updated NID creation version. This should correctly handle UTF8.
1720 [Steve Henson]
1721
1722 *) Implement RFC5746. Re-enable renegotiation but require the extension
1723 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1724 turns out to be a bad idea. It has been replaced by
1725 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1726 SSL_CTX_set_options(). This is really not recommended unless you
1727 know what you are doing.
1728 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1729
1730 *) Fixes to stateless session resumption handling. Use initial_ctx when
1731 issuing and attempting to decrypt tickets in case it has changed during
1732 servername handling. Use a non-zero length session ID when attempting
1733 stateless session resumption: this makes it possible to determine if
1734 a resumption has occurred immediately after receiving server hello
1735 (several places in OpenSSL subtly assume this) instead of later in
1736 the handshake.
1737 [Steve Henson]
1738
1739 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1740 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1741 fixes for a few places where the return code is not checked
1742 correctly.
1743 [Julia Lawall <julia@diku.dk>]
1744
1745 *) Add --strict-warnings option to Configure script to include devteam
1746 warnings in other configurations.
1747 [Steve Henson]
1748
1749 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1750 makes it possible to install openssl libraries in locations which
1751 have names other than "lib", for example "/usr/lib64" which some
1752 systems need.
1753 [Steve Henson, based on patch from Jeremy Utley]
1754
1755 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1756 X690 8.9.12 and can produce some misleading textual output of OIDs.
1757 [Steve Henson, reported by Dan Kaminsky]
1758
1759 *) Delete MD2 from algorithm tables. This follows the recommendation in
1760 several standards that it is not used in new applications due to
1761 several cryptographic weaknesses. For binary compatibility reasons
1762 the MD2 API is still compiled in by default.
1763 [Steve Henson]
1764
1765 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1766 and restored.
1767 [Steve Henson]
1768
1769 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1770 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1771 clash.
1772 [Guenter <lists@gknw.net>]
1773
1774 *) Fix the server certificate chain building code to use X509_verify_cert(),
1775 it used to have an ad-hoc builder which was unable to cope with anything
1776 other than a simple chain.
1777 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1778
1779 *) Don't check self signed certificate signatures in X509_verify_cert()
1780 by default (a flag can override this): it just wastes time without
1781 adding any security. As a useful side effect self signed root CAs
1782 with non-FIPS digests are now usable in FIPS mode.
1783 [Steve Henson]
1784
1785 *) In dtls1_process_out_of_seq_message() the check if the current message
1786 is already buffered was missing. For every new message was memory
1787 allocated, allowing an attacker to perform an denial of service attack
1788 with sending out of seq handshake messages until there is no memory
1789 left. Additionally every future messege was buffered, even if the
1790 sequence number made no sense and would be part of another handshake.
1791 So only messages with sequence numbers less than 10 in advance will be
1792 buffered. (CVE-2009-1378)
1793 [Robin Seggelmann, discovered by Daniel Mentz]
1794
1795 *) Records are buffered if they arrive with a future epoch to be
1796 processed after finishing the corresponding handshake. There is
1797 currently no limitation to this buffer allowing an attacker to perform
1798 a DOS attack with sending records with future epochs until there is no
1799 memory left. This patch adds the pqueue_size() function to detemine
1800 the size of a buffer and limits the record buffer to 100 entries.
1801 (CVE-2009-1377)
1802 [Robin Seggelmann, discovered by Daniel Mentz]
1803
1804 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1805 parent structure is freed. (CVE-2009-1379)
1806 [Daniel Mentz]
1807
1808 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1809 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1810
1811 *) Add 2.5.4.* OIDs
1812 [Ilya O. <vrghost@gmail.com>]
1813
1814 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1815
1816 *) Disable renegotiation completely - this fixes a severe security
1817 problem (CVE-2009-3555) at the cost of breaking all
1818 renegotiation. Renegotiation can be re-enabled by setting
1819 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1820 run-time. This is really not recommended unless you know what
1821 you're doing.
1822 [Ben Laurie]
1823
1824 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1825
1826 *) Don't set val to NULL when freeing up structures, it is freed up by
1827 underlying code. If sizeof(void *) > sizeof(long) this can result in
1828 zeroing past the valid field. (CVE-2009-0789)
1829 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1830
1831 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1832 checked correctly. This would allow some invalid signed attributes to
1833 appear to verify correctly. (CVE-2009-0591)
1834 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1835
1836 *) Reject UniversalString and BMPString types with invalid lengths. This
1837 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1838 a legal length. (CVE-2009-0590)
1839 [Steve Henson]
1840
1841 *) Set S/MIME signing as the default purpose rather than setting it
1842 unconditionally. This allows applications to override it at the store
1843 level.
1844 [Steve Henson]
1845
1846 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1847 to handle some structures.
1848 [Steve Henson]
1849
1850 *) Improve efficiency of mem_gets: don't search whole buffer each time
1851 for a '\n'
1852 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1853
1854 *) New -hex option for openssl rand.
1855 [Matthieu Herrb]
1856
1857 *) Print out UTF8String and NumericString when parsing ASN1.
1858 [Steve Henson]
1859
1860 *) Support NumericString type for name components.
1861 [Steve Henson]
1862
1863 *) Allow CC in the environment to override the automatically chosen
1864 compiler. Note that nothing is done to ensure flags work with the
1865 chosen compiler.
1866 [Ben Laurie]
1867
1868 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1869
1870 *) Properly check EVP_VerifyFinal() and similar return values
1871 (CVE-2008-5077).
1872 [Ben Laurie, Bodo Moeller, Google Security Team]
1873
1874 *) Enable TLS extensions by default.
1875 [Ben Laurie]
1876
1877 *) Allow the CHIL engine to be loaded, whether the application is
1878 multithreaded or not. (This does not release the developer from the
1879 obligation to set up the dynamic locking callbacks.)
1880 [Sander Temme <sander@temme.net>]
1881
1882 *) Use correct exit code if there is an error in dgst command.
1883 [Steve Henson; problem pointed out by Roland Dirlewanger]
1884
1885 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1886 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1887 [Bodo Moeller]
1888
1889 *) Add experimental JPAKE support, including demo authentication in
1890 s_client and s_server.
1891 [Ben Laurie]
1892
1893 *) Set the comparison function in v3_addr_canonize().
1894 [Rob Austein <sra@hactrn.net>]
1895
1896 *) Add support for XMPP STARTTLS in s_client.
1897 [Philip Paeps <philip@freebsd.org>]
1898
1899 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1900 to ensure that even with this option, only ciphersuites in the
1901 server's preference list will be accepted. (Note that the option
1902 applies only when resuming a session, so the earlier behavior was
1903 just about the algorithm choice for symmetric cryptography.)
1904 [Bodo Moeller]
1905
1906 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1907
1908 *) Fix NULL pointer dereference if a DTLS server received
1909 ChangeCipherSpec as first record (CVE-2009-1386).
1910 [PR #1679]
1911
1912 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1913 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1914 [Nagendra Modadugu]
1915
1916 *) The fix in 0.9.8c that supposedly got rid of unsafe
1917 double-checked locking was incomplete for RSA blinding,
1918 addressing just one layer of what turns out to have been
1919 doubly unsafe triple-checked locking.
1920
1921 So now fix this for real by retiring the MONT_HELPER macro
1922 in crypto/rsa/rsa_eay.c.
1923
1924 [Bodo Moeller; problem pointed out by Marius Schilder]
1925
1926 *) Various precautionary measures:
1927
1928 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1929
1930 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1931 (NB: This would require knowledge of the secret session ticket key
1932 to exploit, in which case you'd be SOL either way.)
1933
1934 - Change bn_nist.c so that it will properly handle input BIGNUMs
1935 outside the expected range.
1936
1937 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1938 builds.
1939
1940 [Neel Mehta, Bodo Moeller]
1941
1942 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1943 the load fails. Useful for distros.
1944 [Ben Laurie and the FreeBSD team]
1945
1946 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1947 [Steve Henson]
1948
1949 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1950 [Huang Ying]
1951
1952 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1953
1954 This work was sponsored by Logica.
1955 [Steve Henson]
1956
1957 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1958 keystores. Support for SSL/TLS client authentication too.
1959 Not compiled unless enable-capieng specified to Configure.
1960
1961 This work was sponsored by Logica.
1962 [Steve Henson]
1963
1964 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1965 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1966 attribute creation routines such as certifcate requests and PKCS#12
1967 files.
1968 [Steve Henson]
1969
1970 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1971
1972 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1973 handshake which could lead to a cilent crash as found using the
1974 Codenomicon TLS test suite (CVE-2008-1672)
1975 [Steve Henson, Mark Cox]
1976
1977 *) Fix double free in TLS server name extensions which could lead to
1978 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1979 [Joe Orton]
1980
1981 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1982
1983 Clear the error queue to ensure that error entries left from
1984 older function calls do not interfere with the correct operation.
1985 [Lutz Jaenicke, Erik de Castro Lopo]
1986
1987 *) Remove root CA certificates of commercial CAs:
1988
1989 The OpenSSL project does not recommend any specific CA and does not
1990 have any policy with respect to including or excluding any CA.
1991 Therefore it does not make any sense to ship an arbitrary selection
1992 of root CA certificates with the OpenSSL software.
1993 [Lutz Jaenicke]
1994
1995 *) RSA OAEP patches to fix two separate invalid memory reads.
1996 The first one involves inputs when 'lzero' is greater than
1997 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1998 before the beginning of from). The second one involves inputs where
1999 the 'db' section contains nothing but zeroes (there is a one-byte
2000 invalid read after the end of 'db').
2001 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2002
2003 *) Partial backport from 0.9.9-dev:
2004
2005 Introduce bn_mul_mont (dedicated Montgomery multiplication
2006 procedure) as a candidate for BIGNUM assembler implementation.
2007 While 0.9.9-dev uses assembler for various architectures, only
2008 x86_64 is available by default here in the 0.9.8 branch, and
2009 32-bit x86 is available through a compile-time setting.
2010
2011 To try the 32-bit x86 assembler implementation, use Configure
2012 option "enable-montasm" (which exists only for this backport).
2013
2014 As "enable-montasm" for 32-bit x86 disclaims code stability
2015 anyway, in this constellation we activate additional code
2016 backported from 0.9.9-dev for further performance improvements,
2017 namely BN_from_montgomery_word. (To enable this otherwise,
2018 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2019
2020 [Andy Polyakov (backport partially by Bodo Moeller)]
2021
2022 *) Add TLS session ticket callback. This allows an application to set
2023 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2024 values. This is useful for key rollover for example where several key
2025 sets may exist with different names.
2026 [Steve Henson]
2027
2028 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2029 This was broken until now in 0.9.8 releases, such that the only way
2030 a registered ENGINE could be used (assuming it initialises
2031 successfully on the host) was to explicitly set it as the default
2032 for the relevant algorithms. This is in contradiction with 0.9.7
2033 behaviour and the documentation. With this fix, when an ENGINE is
2034 registered into a given algorithm's table of implementations, the
2035 'uptodate' flag is reset so that auto-discovery will be used next
2036 time a new context for that algorithm attempts to select an
2037 implementation.
2038 [Ian Lister (tweaked by Geoff Thorpe)]
2039
2040 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2041 implemention in the following ways:
2042
2043 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2044 hard coded.
2045
2046 Lack of BER streaming support means one pass streaming processing is
2047 only supported if data is detached: setting the streaming flag is
2048 ignored for embedded content.
2049
2050 CMS support is disabled by default and must be explicitly enabled
2051 with the enable-cms configuration option.
2052 [Steve Henson]
2053
2054 *) Update the GMP engine glue to do direct copies between BIGNUM and
2055 mpz_t when openssl and GMP use the same limb size. Otherwise the
2056 existing "conversion via a text string export" trick is still used.
2057 [Paul Sheer <paulsheer@gmail.com>]
2058
2059 *) Zlib compression BIO. This is a filter BIO which compressed and
2060 uncompresses any data passed through it.
2061 [Steve Henson]
2062
2063 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2064 RFC3394 compatible AES key wrapping.
2065 [Steve Henson]
2066
2067 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2068 sets string data without copying. X509_ALGOR_set0() and
2069 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2070 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2071 from an X509_ATTRIBUTE structure optionally checking it occurs only
2072 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2073 data.
2074 [Steve Henson]
2075
2076 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2077 to get the expected BN_FLG_CONSTTIME behavior.
2078 [Bodo Moeller (Google)]
2079
2080 *) Netware support:
2081
2082 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2083 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2084 - added some more tests to do_tests.pl
2085 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2086 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2087 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2088 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2089 - various changes to netware.pl to enable gcc-cross builds on Win32
2090 platform
2091 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2092 - various changes to fix missing prototype warnings
2093 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2094 - added AES, WHIRLPOOL and CPUID assembler code to build files
2095 - added missing AES assembler make rules to mk1mf.pl
2096 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2097 [Guenter Knauf <eflash@gmx.net>]
2098
2099 *) Implement certificate status request TLS extension defined in RFC3546.
2100 A client can set the appropriate parameters and receive the encoded
2101 OCSP response via a callback. A server can query the supplied parameters
2102 and set the encoded OCSP response in the callback. Add simplified examples
2103 to s_client and s_server.
2104 [Steve Henson]
2105
2106 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2107
2108 *) Fix various bugs:
2109 + Binary incompatibility of ssl_ctx_st structure
2110 + DTLS interoperation with non-compliant servers
2111 + Don't call get_session_cb() without proposed session
2112 + Fix ia64 assembler code
2113 [Andy Polyakov, Steve Henson]
2114
2115 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2116
2117 *) DTLS Handshake overhaul. There were longstanding issues with
2118 OpenSSL DTLS implementation, which were making it impossible for
2119 RFC 4347 compliant client to communicate with OpenSSL server.
2120 Unfortunately just fixing these incompatibilities would "cut off"
2121 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2122 server keeps tolerating non RFC compliant syntax. The opposite is
2123 not true, 0.9.8f client can not communicate with earlier server.
2124 This update even addresses CVE-2007-4995.
2125 [Andy Polyakov]
2126
2127 *) Changes to avoid need for function casts in OpenSSL: some compilers
2128 (gcc 4.2 and later) reject their use.
2129 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2130 Steve Henson]
2131
2132 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2133 RFC4507bis. The encrypted ticket format is an encrypted encoded
2134 SSL_SESSION structure, that way new session features are automatically
2135 supported.
2136
2137 If a client application caches session in an SSL_SESSION structure
2138 support is transparent because tickets are now stored in the encoded
2139 SSL_SESSION.
2140
2141 The SSL_CTX structure automatically generates keys for ticket
2142 protection in servers so again support should be possible
2143 with no application modification.
2144
2145 If a client or server wishes to disable RFC4507 support then the option
2146 SSL_OP_NO_TICKET can be set.
2147
2148 Add a TLS extension debugging callback to allow the contents of any client
2149 or server extensions to be examined.
2150
2151 This work was sponsored by Google.
2152 [Steve Henson]
2153
2154 *) Add initial support for TLS extensions, specifically for the server_name
2155 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2156 have new members for a host name. The SSL data structure has an
2157 additional member SSL_CTX *initial_ctx so that new sessions can be
2158 stored in that context to allow for session resumption, even after the
2159 SSL has been switched to a new SSL_CTX in reaction to a client's
2160 server_name extension.
2161
2162 New functions (subject to change):
2163
2164 SSL_get_servername()
2165 SSL_get_servername_type()
2166 SSL_set_SSL_CTX()
2167
2168 New CTRL codes and macros (subject to change):
2169
2170 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2171 - SSL_CTX_set_tlsext_servername_callback()
2172 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2173 - SSL_CTX_set_tlsext_servername_arg()
2174 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2175
2176 openssl s_client has a new '-servername ...' option.
2177
2178 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2179 '-key2 ...', '-servername_fatal' (subject to change). This allows
2180 testing the HostName extension for a specific single host name ('-cert'
2181 and '-key' remain fallbacks for handshakes without HostName
2182 negotiation). If the unrecogninzed_name alert has to be sent, this by
2183 default is a warning; it becomes fatal with the '-servername_fatal'
2184 option.
2185
2186 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2187
2188 *) Add AES and SSE2 assembly language support to VC++ build.
2189 [Steve Henson]
2190
2191 *) Mitigate attack on final subtraction in Montgomery reduction.
2192 [Andy Polyakov]
2193
2194 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2195 (which previously caused an internal error).
2196 [Bodo Moeller]
2197
2198 *) Squeeze another 10% out of IGE mode when in != out.
2199 [Ben Laurie]
2200
2201 *) AES IGE mode speedup.
2202 [Dean Gaudet (Google)]
2203
2204 *) Add the Korean symmetric 128-bit cipher SEED (see
2205 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2206 add SEED ciphersuites from RFC 4162:
2207
2208 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2209 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2210 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2211 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2212
2213 To minimize changes between patchlevels in the OpenSSL 0.9.8
2214 series, SEED remains excluded from compilation unless OpenSSL
2215 is configured with 'enable-seed'.
2216 [KISA, Bodo Moeller]
2217
2218 *) Mitigate branch prediction attacks, which can be practical if a
2219 single processor is shared, allowing a spy process to extract
2220 information. For detailed background information, see
2221 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2222 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2223 and Necessary Software Countermeasures"). The core of the change
2224 are new versions BN_div_no_branch() and
2225 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2226 respectively, which are slower, but avoid the security-relevant
2227 conditional branches. These are automatically called by BN_div()
2228 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2229 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2230 remove a conditional branch.
2231
2232 BN_FLG_CONSTTIME is the new name for the previous
2233 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2234 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2235 in the exponent causes BN_mod_exp_mont() to use the alternative
2236 implementation in BN_mod_exp_mont_consttime().) The old name
2237 remains as a deprecated alias.
2238
2239 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2240 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2241 constant-time implementations for more than just exponentiation.
2242 Here too the old name is kept as a deprecated alias.
2243
2244 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2245 the BN_BLINDING structure gets an independent copy of the
2246 modulus. This means that the previous "BIGNUM *m" argument to
2247 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2248 essentially becomes "const BIGNUM *m", although we can't actually
2249 change this in the header file before 0.9.9. It allows
2250 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2251 enable BN_FLG_CONSTTIME.
2252
2253 [Matthew D Wood (Intel Corp)]
2254
2255 *) In the SSL/TLS server implementation, be strict about session ID
2256 context matching (which matters if an application uses a single
2257 external cache for different purposes). Previously,
2258 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2259 set. This did ensure strict client verification, but meant that,
2260 with applications using a single external cache for quite
2261 different requirements, clients could circumvent ciphersuite
2262 restrictions for a given session ID context by starting a session
2263 in a different context.
2264 [Bodo Moeller]
2265
2266 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2267 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2268 authentication-only ciphersuites.
2269 [Bodo Moeller]
2270
2271 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2272 not complete and could lead to a possible single byte overflow
2273 (CVE-2007-5135) [Ben Laurie]
2274
2275 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2276
2277 *) Since AES128 and AES256 (and similarly Camellia128 and
2278 Camellia256) share a single mask bit in the logic of
2279 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2280 kludge to work properly if AES128 is available and AES256 isn't
2281 (or if Camellia128 is available and Camellia256 isn't).
2282 [Victor Duchovni]
2283
2284 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2285 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2286 When a point or a seed is encoded in a BIT STRING, we need to
2287 prevent the removal of trailing zero bits to get the proper DER
2288 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2289 of a NamedBitList, for which trailing 0 bits need to be removed.)
2290 [Bodo Moeller]
2291
2292 *) Have SSL/TLS server implementation tolerate "mismatched" record
2293 protocol version while receiving ClientHello even if the
2294 ClientHello is fragmented. (The server can't insist on the
2295 particular protocol version it has chosen before the ServerHello
2296 message has informed the client about his choice.)
2297 [Bodo Moeller]
2298
2299 *) Add RFC 3779 support.
2300 [Rob Austein for ARIN, Ben Laurie]
2301
2302 *) Load error codes if they are not already present instead of using a
2303 static variable. This allows them to be cleanly unloaded and reloaded.
2304 Improve header file function name parsing.
2305 [Steve Henson]
2306
2307 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2308 or CAPABILITY handshake as required by RFCs.
2309 [Goetz Babin-Ebell]
2310
2311 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2312
2313 *) Introduce limits to prevent malicious keys being able to
2314 cause a denial of service. (CVE-2006-2940)
2315 [Steve Henson, Bodo Moeller]
2316
2317 *) Fix ASN.1 parsing of certain invalid structures that can result
2318 in a denial of service. (CVE-2006-2937) [Steve Henson]
2319
2320 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2321 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2322
2323 *) Fix SSL client code which could crash if connecting to a
2324 malicious SSLv2 server. (CVE-2006-4343)
2325 [Tavis Ormandy and Will Drewry, Google Security Team]
2326
2327 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2328 match only those. Before that, "AES256-SHA" would be interpreted
2329 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2330 the same strength classification in 0.9.7h) as we currently only
2331 have a single AES bit in the ciphersuite description bitmap.
2332 That change, however, also applied to ciphersuite strings such as
2333 "RC4-MD5" that intentionally matched multiple ciphersuites --
2334 namely, SSL 2.0 ciphersuites in addition to the more common ones
2335 from SSL 3.0/TLS 1.0.
2336
2337 So we change the selection algorithm again: Naming an explicit
2338 ciphersuite selects this one ciphersuite, and any other similar
2339 ciphersuite (same bitmap) from *other* protocol versions.
2340 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2341 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2342
2343 Since SSL 2.0 does not have any ciphersuites for which the
2344 128/256 bit distinction would be relevant, this works for now.
2345 The proper fix will be to use different bits for AES128 and
2346 AES256, which would have avoided the problems from the beginning;
2347 however, bits are scarce, so we can only do this in a new release
2348 (not just a patchlevel) when we can change the SSL_CIPHER
2349 definition to split the single 'unsigned long mask' bitmap into
2350 multiple values to extend the available space.
2351
2352 [Bodo Moeller]
2353
2354 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2355
2356 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2357 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2358
2359 *) Add AES IGE and biIGE modes.
2360 [Ben Laurie]
2361
2362 *) Change the Unix randomness entropy gathering to use poll() when
2363 possible instead of select(), since the latter has some
2364 undesirable limitations.
2365 [Darryl Miles via Richard Levitte and Bodo Moeller]
2366
2367 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2368 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2369 cannot be implicitly activated as part of, e.g., the "AES" alias.
2370 However, please upgrade to OpenSSL 0.9.9[-dev] for
2371 non-experimental use of the ECC ciphersuites to get TLS extension
2372 support, which is required for curve and point format negotiation
2373 to avoid potential handshake problems.
2374 [Bodo Moeller]
2375
2376 *) Disable rogue ciphersuites:
2377
2378 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2379 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2380 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2381
2382 The latter two were purportedly from
2383 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2384 appear there.
2385
2386 Also deactivate the remaining ciphersuites from
2387 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2388 unofficial, and the ID has long expired.
2389 [Bodo Moeller]
2390
2391 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2392 dual-core machines) and other potential thread-safety issues.
2393 [Bodo Moeller]
2394
2395 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2396 versions), which is now available for royalty-free use
2397 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2398 Also, add Camellia TLS ciphersuites from RFC 4132.
2399
2400 To minimize changes between patchlevels in the OpenSSL 0.9.8
2401 series, Camellia remains excluded from compilation unless OpenSSL
2402 is configured with 'enable-camellia'.
2403 [NTT]
2404
2405 *) Disable the padding bug check when compression is in use. The padding
2406 bug check assumes the first packet is of even length, this is not
2407 necessarily true if compresssion is enabled and can result in false
2408 positives causing handshake failure. The actual bug test is ancient
2409 code so it is hoped that implementations will either have fixed it by
2410 now or any which still have the bug do not support compression.
2411 [Steve Henson]
2412
2413 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2414
2415 *) When applying a cipher rule check to see if string match is an explicit
2416 cipher suite and only match that one cipher suite if it is.
2417 [Steve Henson]
2418
2419 *) Link in manifests for VC++ if needed.
2420 [Austin Ziegler <halostatue@gmail.com>]
2421
2422 *) Update support for ECC-based TLS ciphersuites according to
2423 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2424 TLS extensions, which are supported starting with the 0.9.9
2425 branch, not in the OpenSSL 0.9.8 branch).
2426 [Douglas Stebila]
2427
2428 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2429 opaque EVP_CIPHER_CTX handling.
2430 [Steve Henson]
2431
2432 *) Fixes and enhancements to zlib compression code. We now only use
2433 "zlib1.dll" and use the default __cdecl calling convention on Win32
2434 to conform with the standards mentioned here:
2435 http://www.zlib.net/DLL_FAQ.txt
2436 Static zlib linking now works on Windows and the new --with-zlib-include
2437 --with-zlib-lib options to Configure can be used to supply the location
2438 of the headers and library. Gracefully handle case where zlib library
2439 can't be loaded.
2440 [Steve Henson]
2441
2442 *) Several fixes and enhancements to the OID generation code. The old code
2443 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2444 handle numbers larger than ULONG_MAX, truncated printing and had a
2445 non standard OBJ_obj2txt() behaviour.
2446 [Steve Henson]
2447
2448 *) Add support for building of engines under engine/ as shared libraries
2449 under VC++ build system.
2450 [Steve Henson]
2451
2452 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2453 Hopefully, we will not see any false combination of paths any more.
2454 [Richard Levitte]
2455
2456 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2457
2458 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2459 (part of SSL_OP_ALL). This option used to disable the
2460 countermeasure against man-in-the-middle protocol-version
2461 rollback in the SSL 2.0 server implementation, which is a bad
2462 idea. (CVE-2005-2969)
2463
2464 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2465 for Information Security, National Institute of Advanced Industrial
2466 Science and Technology [AIST], Japan)]
2467
2468 *) Add two function to clear and return the verify parameter flags.
2469 [Steve Henson]
2470
2471 *) Keep cipherlists sorted in the source instead of sorting them at
2472 runtime, thus removing the need for a lock.
2473 [Nils Larsch]
2474
2475 *) Avoid some small subgroup attacks in Diffie-Hellman.
2476 [Nick Mathewson and Ben Laurie]
2477
2478 *) Add functions for well-known primes.
2479 [Nick Mathewson]
2480
2481 *) Extended Windows CE support.
2482 [Satoshi Nakamura and Andy Polyakov]
2483
2484 *) Initialize SSL_METHOD structures at compile time instead of during
2485 runtime, thus removing the need for a lock.
2486 [Steve Henson]
2487
2488 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2489 attempting to decrypt each encrypted key in turn. Add support to
2490 smime utility.
2491 [Steve Henson]
2492
2493 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2494
2495 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2496 OpenSSL 0.9.8.]
2497
2498 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2499 [Richard Levitte]
2500
2501 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2502 key into the same file any more.
2503 [Richard Levitte]
2504
2505 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2506 [Andy Polyakov]
2507
2508 *) Add -utf8 command line and config file option to 'ca'.
2509 [Stefan <stf@udoma.org]
2510
2511 *) Removed the macro des_crypt(), as it seems to conflict with some
2512 libraries. Use DES_crypt().
2513 [Richard Levitte]
2514
2515 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2516 involves renaming the source and generated shared-libs for
2517 both. The engines will accept the corrected or legacy ids
2518 ('ncipher' and '4758_cca' respectively) when binding. NB,
2519 this only applies when building 'shared'.
2520 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2521
2522 *) Add attribute functions to EVP_PKEY structure. Modify
2523 PKCS12_create() to recognize a CSP name attribute and
2524 use it. Make -CSP option work again in pkcs12 utility.
2525 [Steve Henson]
2526
2527 *) Add new functionality to the bn blinding code:
2528 - automatic re-creation of the BN_BLINDING parameters after
2529 a fixed number of uses (currently 32)
2530 - add new function for parameter creation
2531 - introduce flags to control the update behaviour of the
2532 BN_BLINDING parameters
2533 - hide BN_BLINDING structure
2534 Add a second BN_BLINDING slot to the RSA structure to improve
2535 performance when a single RSA object is shared among several
2536 threads.
2537 [Nils Larsch]
2538
2539 *) Add support for DTLS.
2540 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2541
2542 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2543 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2544 [Walter Goulet]
2545
2546 *) Remove buggy and incompletet DH cert support from
2547 ssl/ssl_rsa.c and ssl/s3_both.c
2548 [Nils Larsch]
2549
2550 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2551 the apps/openssl applications.
2552 [Nils Larsch]
2553
2554 *) Compile clean with "-Wall -Wmissing-prototypes
2555 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2556 DEBUG_SAFESTACK must also be set.
2557 [Ben Laurie]
2558
2559 *) Change ./Configure so that certain algorithms can be disabled by default.
2560 The new counterpiece to "no-xxx" is "enable-xxx".
2561
2562 The patented RC5 and MDC2 algorithms will now be disabled unless
2563 "enable-rc5" and "enable-mdc2", respectively, are specified.
2564
2565 (IDEA remains enabled despite being patented. This is because IDEA
2566 is frequently required for interoperability, and there is no license
2567 fee for non-commercial use. As before, "no-idea" can be used to
2568 avoid this algorithm.)
2569
2570 [Bodo Moeller]
2571
2572 *) Add processing of proxy certificates (see RFC 3820). This work was
2573 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2574 EGEE (Enabling Grids for E-science in Europe).
2575 [Richard Levitte]
2576
2577 *) RC4 performance overhaul on modern architectures/implementations, such
2578 as Intel P4, IA-64 and AMD64.
2579 [Andy Polyakov]
2580
2581 *) New utility extract-section.pl. This can be used specify an alternative
2582 section number in a pod file instead of having to treat each file as
2583 a separate case in Makefile. This can be done by adding two lines to the
2584 pod file:
2585
2586 =for comment openssl_section:XXX
2587
2588 The blank line is mandatory.
2589
2590 [Steve Henson]
2591
2592 *) New arguments -certform, -keyform and -pass for s_client and s_server
2593 to allow alternative format key and certificate files and passphrase
2594 sources.
2595 [Steve Henson]
2596
2597 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2598 update associated structures and add various utility functions.
2599
2600 Add new policy related verify parameters, include policy checking in
2601 standard verify code. Enhance 'smime' application with extra parameters
2602 to support policy checking and print out.
2603 [Steve Henson]
2604
2605 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2606 Nehemiah processors. These extensions support AES encryption in hardware
2607 as well as RNG (though RNG support is currently disabled).
2608 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2609
2610 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2611 [Geoff Thorpe]
2612
2613 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2614 [Andy Polyakov and a number of other people]
2615
2616 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2617 implementation contributed by IBM.
2618 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2619
2620 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2621 exponent rather than 'unsigned long'. There is a corresponding change to
2622 the new 'rsa_keygen' element of the RSA_METHOD structure.
2623 [Jelte Jansen, Geoff Thorpe]
2624
2625 *) Functionality for creating the initial serial number file is now
2626 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2627
2628 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2629 number file to 1, which is bound to cause problems. To avoid
2630 the problems while respecting compatibility between different 0.9.7
2631 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2632 CA.pl for serial number initialization. With the new release 0.9.8,
2633 we can fix the problem directly in the 'ca' utility.)
2634 [Steve Henson]
2635
2636 *) Reduced header interdepencies by declaring more opaque objects in
2637 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2638 give fewer recursive includes, which could break lazy source code - so
2639 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2640 developers should define this symbol when building and using openssl to
2641 ensure they track the recommended behaviour, interfaces, [etc], but
2642 backwards-compatible behaviour prevails when this isn't defined.
2643 [Geoff Thorpe]
2644
2645 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2646 [Steve Henson]
2647
2648 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2649 This will generate a random key of the appropriate length based on the
2650 cipher context. The EVP_CIPHER can provide its own random key generation
2651 routine to support keys of a specific form. This is used in the des and
2652 3des routines to generate a key of the correct parity. Update S/MIME
2653 code to use new functions and hence generate correct parity DES keys.
2654 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2655 valid (weak or incorrect parity).
2656 [Steve Henson]
2657
2658 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2659 as looking them up. This is useful when the verified structure may contain
2660 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2661 present unless the new PKCS7_NO_CRL flag is asserted.
2662 [Steve Henson]
2663
2664 *) Extend ASN1 oid configuration module. It now additionally accepts the
2665 syntax:
2666
2667 shortName = some long name, 1.2.3.4
2668 [Steve Henson]
2669
2670 *) Reimplemented the BN_CTX implementation. There is now no more static
2671 limitation on the number of variables it can handle nor the depth of the
2672 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2673 information can now expand as required, and rather than having a single
2674 static array of bignums, BN_CTX now uses a linked-list of such arrays
2675 allowing it to expand on demand whilst maintaining the usefulness of
2676 BN_CTX's "bundling".
2677 [Geoff Thorpe]
2678
2679 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2680 to allow all RSA operations to function using a single BN_CTX.
2681 [Geoff Thorpe]
2682
2683 *) Preliminary support for certificate policy evaluation and checking. This
2684 is initially intended to pass the tests outlined in "Conformance Testing
2685 of Relying Party Client Certificate Path Processing Logic" v1.07.
2686 [Steve Henson]
2687
2688 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2689 remained unused and not that useful. A variety of other little bignum
2690 tweaks and fixes have also been made continuing on from the audit (see
2691 below).
2692 [Geoff Thorpe]
2693
2694 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2695 associated ASN1, EVP and SSL functions and old ASN1 macros.
2696 [Richard Levitte]
2697
2698 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2699 and this should never fail. So the return value from the use of
2700 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2701 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2702 [Geoff Thorpe]
2703
2704 *) BN_CTX_get() should return zero-valued bignums, providing the same
2705 initialised value as BN_new().
2706 [Geoff Thorpe, suggested by Ulf Möller]
2707
2708 *) Support for inhibitAnyPolicy certificate extension.
2709 [Steve Henson]
2710
2711 *) An audit of the BIGNUM code is underway, for which debugging code is
2712 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2713 is considered valid when processing BIGNUMs, and causes execution to
2714 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2715 further steps are taken to deliberately pollute unused data in BIGNUM
2716 structures to try and expose faulty code further on. For now, openssl will
2717 (in its default mode of operation) continue to tolerate the inconsistent
2718 forms that it has tolerated in the past, but authors and packagers should
2719 consider trying openssl and their own applications when compiled with
2720 these debugging symbols defined. It will help highlight potential bugs in
2721 their own code, and will improve the test coverage for OpenSSL itself. At
2722 some point, these tighter rules will become openssl's default to improve
2723 maintainability, though the assert()s and other overheads will remain only
2724 in debugging configurations. See bn.h for more details.
2725 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2726
2727 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2728 that can only be obtained through BN_CTX_new() (which implicitly
2729 initialises it). The presence of this function only made it possible
2730 to overwrite an existing structure (and cause memory leaks).
2731 [Geoff Thorpe]
2732
2733 *) Because of the callback-based approach for implementing LHASH as a
2734 template type, lh_insert() adds opaque objects to hash-tables and
2735 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2736 to clean up those corresponding objects before destroying the hash table
2737 (and losing the object pointers). So some over-zealous constifications in
2738 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2739 objects as "const" and the lh_doall[_arg] callback wrappers are not
2740 prototyped to have "const" restrictions on the object pointers they are
2741 given (and so aren't required to cast them away any more).
2742 [Geoff Thorpe]
2743
2744 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2745 (speed) prefers to use its own implementation. The two implementations
2746 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2747 its object type properly exposed (MS_TM) instead of casting to/from "char
2748 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2749 aren't necessarily the greatest nomenclatures - but this is what was used
2750 internally to the implementation so I've used that for now.
2751 [Geoff Thorpe]
2752
2753 *) Ensure that deprecated functions do not get compiled when
2754 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2755 the self-tests were still using deprecated key-generation functions so
2756 these have been updated also.
2757 [Geoff Thorpe]
2758
2759 *) Reorganise PKCS#7 code to separate the digest location functionality
2760 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2761 New function PKCS7_set_digest() to set the digest type for PKCS#7
2762 digestedData type. Add additional code to correctly generate the
2763 digestedData type and add support for this type in PKCS7 initialization
2764 functions.
2765 [Steve Henson]
2766
2767 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2768 structure of type "other".
2769 [Steve Henson]
2770
2771 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2772 sure the loop does correctly stop and breaking ("division by zero")
2773 modulus operations are not performed. The (pre-generated) prime
2774 table crypto/bn/bn_prime.h was already correct, but it could not be
2775 re-generated on some platforms because of the "division by zero"
2776 situation in the script.
2777 [Ralf S. Engelschall]
2778
2779 *) Update support for ECC-based TLS ciphersuites according to
2780 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2781 SHA-1 now is only used for "small" curves (where the
2782 representation of a field element takes up to 24 bytes); for
2783 larger curves, the field element resulting from ECDH is directly
2784 used as premaster secret.
2785 [Douglas Stebila (Sun Microsystems Laboratories)]
2786
2787 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2788 curve secp160r1 to the tests.
2789 [Douglas Stebila (Sun Microsystems Laboratories)]
2790
2791 *) Add the possibility to load symbols globally with DSO.
2792 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2793
2794 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2795 control of the error stack.
2796 [Richard Levitte]
2797
2798 *) Add support for STORE in ENGINE.
2799 [Richard Levitte]
2800
2801 *) Add the STORE type. The intention is to provide a common interface
2802 to certificate and key stores, be they simple file-based stores, or
2803 HSM-type store, or LDAP stores, or...
2804 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2805 [Richard Levitte]
2806
2807 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2808 pass a list of arguments to any function as well as provide a way
2809 for a function to pass data back to the caller.
2810 [Richard Levitte]
2811
2812 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2813 works like BUF_strdup() but can be used to duplicate a portion of
2814 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2815 a memory area.
2816 [Richard Levitte]
2817
2818 *) Add the function sk_find_ex() which works like sk_find(), but will
2819 return an index to an element even if an exact match couldn't be
2820 found. The index is guaranteed to point at the element where the
2821 searched-for key would be inserted to preserve sorting order.
2822 [Richard Levitte]
2823
2824 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2825 takes an extra flags argument for optional functionality. Currently,
2826 the following flags are defined:
2827
2828 OBJ_BSEARCH_VALUE_ON_NOMATCH
2829 This one gets OBJ_bsearch_ex() to return a pointer to the first
2830 element where the comparing function returns a negative or zero
2831 number.
2832
2833 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2834 This one gets OBJ_bsearch_ex() to return a pointer to the first
2835 element where the comparing function returns zero. This is useful
2836 if there are more than one element where the comparing function
2837 returns zero.
2838 [Richard Levitte]
2839
2840 *) Make it possible to create self-signed certificates with 'openssl ca'
2841 in such a way that the self-signed certificate becomes part of the
2842 CA database and uses the same mechanisms for serial number generation
2843 as all other certificate signing. The new flag '-selfsign' enables
2844 this functionality. Adapt CA.sh and CA.pl.in.
2845 [Richard Levitte]
2846
2847 *) Add functionality to check the public key of a certificate request
2848 against a given private. This is useful to check that a certificate
2849 request can be signed by that key (self-signing).
2850 [Richard Levitte]
2851
2852 *) Make it possible to have multiple active certificates with the same
2853 subject in the CA index file. This is done only if the keyword
2854 'unique_subject' is set to 'no' in the main CA section (default
2855 if 'CA_default') of the configuration file. The value is saved
2856 with the database itself in a separate index attribute file,
2857 named like the index file with '.attr' appended to the name.
2858 [Richard Levitte]
2859
2860 *) Generate muti valued AVAs using '+' notation in config files for
2861 req and dirName.
2862 [Steve Henson]
2863
2864 *) Support for nameConstraints certificate extension.
2865 [Steve Henson]
2866
2867 *) Support for policyConstraints certificate extension.
2868 [Steve Henson]
2869
2870 *) Support for policyMappings certificate extension.
2871 [Steve Henson]
2872
2873 *) Make sure the default DSA_METHOD implementation only uses its
2874 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2875 and change its own handlers to be NULL so as to remove unnecessary
2876 indirection. This lets alternative implementations fallback to the
2877 default implementation more easily.
2878 [Geoff Thorpe]
2879
2880 *) Support for directoryName in GeneralName related extensions
2881 in config files.
2882 [Steve Henson]
2883
2884 *) Make it possible to link applications using Makefile.shared.
2885 Make that possible even when linking against static libraries!
2886 [Richard Levitte]
2887
2888 *) Support for single pass processing for S/MIME signing. This now
2889 means that S/MIME signing can be done from a pipe, in addition
2890 cleartext signing (multipart/signed type) is effectively streaming
2891 and the signed data does not need to be all held in memory.
2892
2893 This is done with a new flag PKCS7_STREAM. When this flag is set
2894 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2895 is done after the data is output (and digests calculated) in
2896 SMIME_write_PKCS7().
2897 [Steve Henson]
2898
2899 *) Add full support for -rpath/-R, both in shared libraries and
2900 applications, at least on the platforms where it's known how
2901 to do it.
2902 [Richard Levitte]
2903
2904 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2905 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2906 will now compute a table of multiples of the generator that
2907 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2908 faster (notably in the case of a single point multiplication,
2909 scalar * generator).
2910 [Nils Larsch, Bodo Moeller]
2911
2912 *) IPv6 support for certificate extensions. The various extensions
2913 which use the IP:a.b.c.d can now take IPv6 addresses using the
2914 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2915 correctly.
2916 [Steve Henson]
2917
2918 *) Added an ENGINE that implements RSA by performing private key
2919 exponentiations with the GMP library. The conversions to and from
2920 GMP's mpz_t format aren't optimised nor are any montgomery forms
2921 cached, and on x86 it appears OpenSSL's own performance has caught up.
2922 However there are likely to be other architectures where GMP could
2923 provide a boost. This ENGINE is not built in by default, but it can be
2924 specified at Configure time and should be accompanied by the necessary
2925 linker additions, eg;
2926 ./config -DOPENSSL_USE_GMP -lgmp
2927 [Geoff Thorpe]
2928
2929 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2930 testing availability of engines with "-t" - the old behaviour is
2931 produced by increasing the feature's verbosity with "-tt".
2932 [Geoff Thorpe]
2933
2934 *) ECDSA routines: under certain error conditions uninitialized BN objects
2935 could be freed. Solution: make sure initialization is performed early
2936 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2937 via PR#459)
2938 [Lutz Jaenicke]
2939
2940 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2941 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2942 software implementations. For DSA and DH, parameter generation can
2943 also be overriden by providing the appropriate method callbacks.
2944 [Geoff Thorpe]
2945
2946 *) Change the "progress" mechanism used in key-generation and
2947 primality testing to functions that take a new BN_GENCB pointer in
2948 place of callback/argument pairs. The new API functions have "_ex"
2949 postfixes and the older functions are reimplemented as wrappers for
2950 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2951 declarations of the old functions to help (graceful) attempts to
2952 migrate to the new functions. Also, the new key-generation API
2953 functions operate on a caller-supplied key-structure and return
2954 success/failure rather than returning a key or NULL - this is to
2955 help make "keygen" another member function of RSA_METHOD etc.
2956
2957 Example for using the new callback interface:
2958
2959 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2960 void *my_arg = ...;
2961 BN_GENCB my_cb;
2962
2963 BN_GENCB_set(&my_cb, my_callback, my_arg);
2964
2965 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2966 /* For the meaning of a, b in calls to my_callback(), see the
2967 * documentation of the function that calls the callback.
2968 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2969 * my_callback should return 1 if it wants BN_is_prime_ex()
2970 * to continue, or 0 to stop.
2971 */
2972
2973 [Geoff Thorpe]
2974
2975 *) Change the ZLIB compression method to be stateful, and make it
2976 available to TLS with the number defined in
2977 draft-ietf-tls-compression-04.txt.
2978 [Richard Levitte]
2979
2980 *) Add the ASN.1 structures and functions for CertificatePair, which
2981 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2982
2983 CertificatePair ::= SEQUENCE {
2984 forward [0] Certificate OPTIONAL,
2985 reverse [1] Certificate OPTIONAL,
2986 -- at least one of the pair shall be present -- }
2987
2988 Also implement the PEM functions to read and write certificate
2989 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2990
2991 This needed to be defined, mostly for the sake of the LDAP
2992 attribute crossCertificatePair, but may prove useful elsewhere as
2993 well.
2994 [Richard Levitte]
2995
2996 *) Make it possible to inhibit symlinking of shared libraries in
2997 Makefile.shared, for Cygwin's sake.
2998 [Richard Levitte]
2999
3000 *) Extend the BIGNUM API by creating a function
3001 void BN_set_negative(BIGNUM *a, int neg);
3002 and a macro that behave like
3003 int BN_is_negative(const BIGNUM *a);
3004
3005 to avoid the need to access 'a->neg' directly in applications.
3006 [Nils Larsch]
3007
3008 *) Implement fast modular reduction for pseudo-Mersenne primes
3009 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3010 EC_GROUP_new_curve_GFp() will now automatically use this
3011 if applicable.
3012 [Nils Larsch <nla@trustcenter.de>]
3013
3014 *) Add new lock type (CRYPTO_LOCK_BN).
3015 [Bodo Moeller]
3016
3017 *) Change the ENGINE framework to automatically load engines
3018 dynamically from specific directories unless they could be
3019 found to already be built in or loaded. Move all the
3020 current engines except for the cryptodev one to a new
3021 directory engines/.
3022 The engines in engines/ are built as shared libraries if
3023 the "shared" options was given to ./Configure or ./config.
3024 Otherwise, they are inserted in libcrypto.a.
3025 /usr/local/ssl/engines is the default directory for dynamic
3026 engines, but that can be overriden at configure time through
3027 the usual use of --prefix and/or --openssldir, and at run
3028 time with the environment variable OPENSSL_ENGINES.
3029 [Geoff Thorpe and Richard Levitte]
3030
3031 *) Add Makefile.shared, a helper makefile to build shared
3032 libraries. Addapt Makefile.org.
3033 [Richard Levitte]
3034
3035 *) Add version info to Win32 DLLs.
3036 [Peter 'Luna' Runestig" <peter@runestig.com>]
3037
3038 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3039 can be added using this API to created arbitrary PKCS#12
3040 files while avoiding the low level API.
3041
3042 New options to PKCS12_create(), key or cert can be NULL and
3043 will then be omitted from the output file. The encryption
3044 algorithm NIDs can be set to -1 for no encryption, the mac
3045 iteration count can be set to 0 to omit the mac.
3046
3047 Enhance pkcs12 utility by making the -nokeys and -nocerts
3048 options work when creating a PKCS#12 file. New option -nomac
3049 to omit the mac, NONE can be set for an encryption algorithm.
3050 New code is modified to use the enhanced PKCS12_create()
3051 instead of the low level API.
3052 [Steve Henson]
3053
3054 *) Extend ASN1 encoder to support indefinite length constructed
3055 encoding. This can output sequences tags and octet strings in
3056 this form. Modify pk7_asn1.c to support indefinite length
3057 encoding. This is experimental and needs additional code to
3058 be useful, such as an ASN1 bio and some enhanced streaming
3059 PKCS#7 code.
3060
3061 Extend template encode functionality so that tagging is passed
3062 down to the template encoder.
3063 [Steve Henson]
3064
3065 *) Let 'openssl req' fail if an argument to '-newkey' is not
3066 recognized instead of using RSA as a default.
3067 [Bodo Moeller]
3068
3069 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3070 As these are not official, they are not included in "ALL";
3071 the "ECCdraft" ciphersuite group alias can be used to select them.
3072 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3073
3074 *) Add ECDH engine support.
3075 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3076
3077 *) Add ECDH in new directory crypto/ecdh/.
3078 [Douglas Stebila (Sun Microsystems Laboratories)]
3079
3080 *) Let BN_rand_range() abort with an error after 100 iterations
3081 without success (which indicates a broken PRNG).
3082 [Bodo Moeller]
3083
3084 *) Change BN_mod_sqrt() so that it verifies that the input value
3085 is really the square of the return value. (Previously,
3086 BN_mod_sqrt would show GIGO behaviour.)
3087 [Bodo Moeller]
3088
3089 *) Add named elliptic curves over binary fields from X9.62, SECG,
3090 and WAP/WTLS; add OIDs that were still missing.
3091
3092 [Sheueling Chang Shantz and Douglas Stebila
3093 (Sun Microsystems Laboratories)]
3094
3095 *) Extend the EC library for elliptic curves over binary fields
3096 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3097 New EC_METHOD:
3098
3099 EC_GF2m_simple_method
3100
3101 New API functions:
3102
3103 EC_GROUP_new_curve_GF2m
3104 EC_GROUP_set_curve_GF2m
3105 EC_GROUP_get_curve_GF2m
3106 EC_POINT_set_affine_coordinates_GF2m
3107 EC_POINT_get_affine_coordinates_GF2m
3108 EC_POINT_set_compressed_coordinates_GF2m
3109
3110 Point compression for binary fields is disabled by default for
3111 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3112 enable it).
3113
3114 As binary polynomials are represented as BIGNUMs, various members
3115 of the EC_GROUP and EC_POINT data structures can be shared
3116 between the implementations for prime fields and binary fields;
3117 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3118 are essentially identical to their ..._GFp counterparts.
3119 (For simplicity, the '..._GFp' prefix has been dropped from
3120 various internal method names.)
3121
3122 An internal 'field_div' method (similar to 'field_mul' and
3123 'field_sqr') has been added; this is used only for binary fields.
3124
3125 [Sheueling Chang Shantz and Douglas Stebila
3126 (Sun Microsystems Laboratories)]
3127
3128 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3129 through methods ('mul', 'precompute_mult').
3130
3131 The generic implementations (now internally called 'ec_wNAF_mul'
3132 and 'ec_wNAF_precomputed_mult') remain the default if these
3133 methods are undefined.
3134
3135 [Sheueling Chang Shantz and Douglas Stebila
3136 (Sun Microsystems Laboratories)]
3137
3138 *) New function EC_GROUP_get_degree, which is defined through
3139 EC_METHOD. For curves over prime fields, this returns the bit
3140 length of the modulus.
3141
3142 [Sheueling Chang Shantz and Douglas Stebila
3143 (Sun Microsystems Laboratories)]
3144
3145 *) New functions EC_GROUP_dup, EC_POINT_dup.
3146 (These simply call ..._new and ..._copy).
3147
3148 [Sheueling Chang Shantz and Douglas Stebila
3149 (Sun Microsystems Laboratories)]
3150
3151 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3152 Polynomials are represented as BIGNUMs (where the sign bit is not
3153 used) in the following functions [macros]:
3154
3155 BN_GF2m_add
3156 BN_GF2m_sub [= BN_GF2m_add]
3157 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3158 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3159 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3160 BN_GF2m_mod_inv
3161 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3162 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3163 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3164 BN_GF2m_cmp [= BN_ucmp]
3165
3166 (Note that only the 'mod' functions are actually for fields GF(2^m).
3167 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3168
3169 For some functions, an the irreducible polynomial defining a
3170 field can be given as an 'unsigned int[]' with strictly
3171 decreasing elements giving the indices of those bits that are set;
3172 i.e., p[] represents the polynomial
3173 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3174 where
3175 p[0] > p[1] > ... > p[k] = 0.
3176 This applies to the following functions:
3177
3178 BN_GF2m_mod_arr
3179 BN_GF2m_mod_mul_arr
3180 BN_GF2m_mod_sqr_arr
3181 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3182 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3183 BN_GF2m_mod_exp_arr
3184 BN_GF2m_mod_sqrt_arr
3185 BN_GF2m_mod_solve_quad_arr
3186 BN_GF2m_poly2arr
3187 BN_GF2m_arr2poly
3188
3189 Conversion can be performed by the following functions:
3190
3191 BN_GF2m_poly2arr
3192 BN_GF2m_arr2poly
3193
3194 bntest.c has additional tests for binary polynomial arithmetic.
3195
3196 Two implementations for BN_GF2m_mod_div() are available.
3197 The default algorithm simply uses BN_GF2m_mod_inv() and
3198 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3199 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3200 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3201
3202 [Sheueling Chang Shantz and Douglas Stebila
3203 (Sun Microsystems Laboratories)]
3204
3205 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3206 functionality is disabled at compile-time.
3207 [Douglas Stebila <douglas.stebila@sun.com>]
3208
3209 *) Change default behaviour of 'openssl asn1parse' so that more
3210 information is visible when viewing, e.g., a certificate:
3211
3212 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3213 mode the content of non-printable OCTET STRINGs is output in a
3214 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3215 avoid the appearance of a printable string.
3216 [Nils Larsch <nla@trustcenter.de>]
3217
3218 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3219 functions
3220 EC_GROUP_set_asn1_flag()
3221 EC_GROUP_get_asn1_flag()
3222 EC_GROUP_set_point_conversion_form()
3223 EC_GROUP_get_point_conversion_form()
3224 These control ASN1 encoding details:
3225 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3226 has been set to OPENSSL_EC_NAMED_CURVE.
3227 - Points are encoded in uncompressed form by default; options for
3228 asn1_for are as for point2oct, namely
3229 POINT_CONVERSION_COMPRESSED
3230 POINT_CONVERSION_UNCOMPRESSED
3231 POINT_CONVERSION_HYBRID
3232
3233 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3234 functions
3235 EC_GROUP_set_seed()
3236 EC_GROUP_get0_seed()
3237 EC_GROUP_get_seed_len()
3238 This is used only for ASN1 purposes (so far).
3239 [Nils Larsch <nla@trustcenter.de>]
3240
3241 *) Add 'field_type' member to EC_METHOD, which holds the NID
3242 of the appropriate field type OID. The new function
3243 EC_METHOD_get_field_type() returns this value.
3244 [Nils Larsch <nla@trustcenter.de>]
3245
3246 *) Add functions
3247 EC_POINT_point2bn()
3248 EC_POINT_bn2point()
3249 EC_POINT_point2hex()
3250 EC_POINT_hex2point()
3251 providing useful interfaces to EC_POINT_point2oct() and
3252 EC_POINT_oct2point().
3253 [Nils Larsch <nla@trustcenter.de>]
3254
3255 *) Change internals of the EC library so that the functions
3256 EC_GROUP_set_generator()
3257 EC_GROUP_get_generator()
3258 EC_GROUP_get_order()
3259 EC_GROUP_get_cofactor()
3260 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3261 to methods, which would lead to unnecessary code duplication when
3262 adding different types of curves.
3263 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3264
3265 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3266 arithmetic, and such that modified wNAFs are generated
3267 (which avoid length expansion in many cases).
3268 [Bodo Moeller]
3269
3270 *) Add a function EC_GROUP_check_discriminant() (defined via
3271 EC_METHOD) that verifies that the curve discriminant is non-zero.
3272
3273 Add a function EC_GROUP_check() that makes some sanity tests
3274 on a EC_GROUP, its generator and order. This includes
3275 EC_GROUP_check_discriminant().
3276 [Nils Larsch <nla@trustcenter.de>]
3277
3278 *) Add ECDSA in new directory crypto/ecdsa/.
3279
3280 Add applications 'openssl ecparam' and 'openssl ecdsa'
3281 (these are based on 'openssl dsaparam' and 'openssl dsa').
3282
3283 ECDSA support is also included in various other files across the
3284 library. Most notably,
3285 - 'openssl req' now has a '-newkey ecdsa:file' option;
3286 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3287 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3288 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3289 them suitable for ECDSA where domain parameters must be
3290 extracted before the specific public key;
3291 - ECDSA engine support has been added.
3292 [Nils Larsch <nla@trustcenter.de>]
3293
3294 *) Include some named elliptic curves, and add OIDs from X9.62,
3295 SECG, and WAP/WTLS. Each curve can be obtained from the new
3296 function
3297 EC_GROUP_new_by_curve_name(),
3298 and the list of available named curves can be obtained with
3299 EC_get_builtin_curves().
3300 Also add a 'curve_name' member to EC_GROUP objects, which can be
3301 accessed via
3302 EC_GROUP_set_curve_name()
3303 EC_GROUP_get_curve_name()
3304 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3305
3306 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3307 was actually never needed) and in BN_mul(). The removal in BN_mul()
3308 required a small change in bn_mul_part_recursive() and the addition
3309 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3310 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3311 bn_sub_words() and bn_add_words() except they take arrays with
3312 differing sizes.
3313 [Richard Levitte]
3314
3315 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3316
3317 *) Cleanse PEM buffers before freeing them since they may contain
3318 sensitive data.
3319 [Benjamin Bennett <ben@psc.edu>]
3320
3321 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3322 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3323 authentication-only ciphersuites.
3324 [Bodo Moeller]
3325
3326 *) Since AES128 and AES256 share a single mask bit in the logic of
3327 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3328 kludge to work properly if AES128 is available and AES256 isn't.
3329 [Victor Duchovni]
3330
3331 *) Expand security boundary to match 1.1.1 module.
3332 [Steve Henson]
3333
3334 *) Remove redundant features: hash file source, editing of test vectors
3335 modify fipsld to use external fips_premain.c signature.
3336 [Steve Henson]
3337
3338 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3339 run algorithm test programs.
3340 [Steve Henson]
3341
3342 *) Make algorithm test programs more tolerant of whitespace.
3343 [Steve Henson]
3344
3345 *) Have SSL/TLS server implementation tolerate "mismatched" record
3346 protocol version while receiving ClientHello even if the
3347 ClientHello is fragmented. (The server can't insist on the
3348 particular protocol version it has chosen before the ServerHello
3349 message has informed the client about his choice.)
3350 [Bodo Moeller]
3351
3352 *) Load error codes if they are not already present instead of using a
3353 static variable. This allows them to be cleanly unloaded and reloaded.
3354 [Steve Henson]
3355
3356 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3357
3358 *) Introduce limits to prevent malicious keys being able to
3359 cause a denial of service. (CVE-2006-2940)
3360 [Steve Henson, Bodo Moeller]
3361
3362 *) Fix ASN.1 parsing of certain invalid structures that can result
3363 in a denial of service. (CVE-2006-2937) [Steve Henson]
3364
3365 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3366 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3367
3368 *) Fix SSL client code which could crash if connecting to a
3369 malicious SSLv2 server. (CVE-2006-4343)
3370 [Tavis Ormandy and Will Drewry, Google Security Team]
3371
3372 *) Change ciphersuite string processing so that an explicit
3373 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3374 will no longer include "AES128-SHA"), and any other similar
3375 ciphersuite (same bitmap) from *other* protocol versions (so that
3376 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3377 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3378 changes from 0.9.8b and 0.9.8d.
3379 [Bodo Moeller]
3380
3381 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3382
3383 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3384 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3385
3386 *) Change the Unix randomness entropy gathering to use poll() when
3387 possible instead of select(), since the latter has some
3388 undesirable limitations.
3389 [Darryl Miles via Richard Levitte and Bodo Moeller]
3390
3391 *) Disable rogue ciphersuites:
3392
3393 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3394 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3395 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3396
3397 The latter two were purportedly from
3398 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3399 appear there.
3400
3401 Also deactive the remaining ciphersuites from
3402 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3403 unofficial, and the ID has long expired.
3404 [Bodo Moeller]
3405
3406 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3407 dual-core machines) and other potential thread-safety issues.
3408 [Bodo Moeller]
3409
3410 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3411
3412 *) Adapt fipsld and the build system to link against the validated FIPS
3413 module in FIPS mode.
3414 [Steve Henson]
3415
3416 *) Fixes for VC++ 2005 build under Windows.
3417 [Steve Henson]
3418
3419 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3420 from a Windows bash shell such as MSYS. It is autodetected from the
3421 "config" script when run from a VC++ environment. Modify standard VC++
3422 build to use fipscanister.o from the GNU make build.
3423 [Steve Henson]
3424
3425 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3426
3427 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3428 The value now differs depending on if you build for FIPS or not.
3429 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3430 safely run with a non-FIPSed libcrypto, as it may crash because of
3431 the difference induced by this change.
3432 [Andy Polyakov]
3433
3434 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3435
3436 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3437 (part of SSL_OP_ALL). This option used to disable the
3438 countermeasure against man-in-the-middle protocol-version
3439 rollback in the SSL 2.0 server implementation, which is a bad
3440 idea. (CVE-2005-2969)
3441
3442 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3443 for Information Security, National Institute of Advanced Industrial
3444 Science and Technology [AIST], Japan)]
3445
3446 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3447 mainly for FIPS compliance and not fully integrated at this stage.
3448 [Steve Henson]
3449
3450 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3451 the exponentiation using a fixed-length exponent. (Otherwise,
3452 the information leaked through timing could expose the secret key
3453 after many signatures; cf. Bleichenbacher's attack on DSA with
3454 biased k.)
3455 [Bodo Moeller]
3456
3457 *) Make a new fixed-window mod_exp implementation the default for
3458 RSA, DSA, and DH private-key operations so that the sequence of
3459 squares and multiplies and the memory access pattern are
3460 independent of the particular secret key. This will mitigate
3461 cache-timing and potential related attacks.
3462
3463 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3464 and this is automatically used by BN_mod_exp_mont() if the new flag
3465 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3466 will use this BN flag for private exponents unless the flag
3467 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3468 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3469
3470 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3471
3472 *) Change the client implementation for SSLv23_method() and
3473 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3474 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3475 (Previously, the SSL 2.0 backwards compatible Client Hello
3476 message format would be used even with SSL_OP_NO_SSLv2.)
3477 [Bodo Moeller]
3478
3479 *) Add support for smime-type MIME parameter in S/MIME messages which some
3480 clients need.
3481 [Steve Henson]
3482
3483 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3484 a threadsafe manner. Modify rsa code to use new function and add calls
3485 to dsa and dh code (which had race conditions before).
3486 [Steve Henson]
3487
3488 *) Include the fixed error library code in the C error file definitions
3489 instead of fixing them up at runtime. This keeps the error code
3490 structures constant.
3491 [Steve Henson]
3492
3493 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3494
3495 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3496 OpenSSL 0.9.8.]
3497
3498 *) Fixes for newer kerberos headers. NB: the casts are needed because
3499 the 'length' field is signed on one version and unsigned on another
3500 with no (?) obvious way to tell the difference, without these VC++
3501 complains. Also the "definition" of FAR (blank) is no longer included
3502 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3503 some needed definitions.
3504 [Steve Henson]
3505
3506 *) Undo Cygwin change.
3507 [Ulf Möller]
3508
3509 *) Added support for proxy certificates according to RFC 3820.
3510 Because they may be a security thread to unaware applications,
3511 they must be explicitely allowed in run-time. See
3512 docs/HOWTO/proxy_certificates.txt for further information.
3513 [Richard Levitte]
3514
3515 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3516
3517 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3518 server and client random values. Previously
3519 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3520 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3521
3522 This change has negligible security impact because:
3523
3524 1. Server and client random values still have 24 bytes of pseudo random
3525 data.
3526
3527 2. Server and client random values are sent in the clear in the initial
3528 handshake.
3529
3530 3. The master secret is derived using the premaster secret (48 bytes in
3531 size for static RSA ciphersuites) as well as client server and random
3532 values.
3533
3534 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3535 to our attention.
3536
3537 [Stephen Henson, reported by UK NISCC]
3538
3539 *) Use Windows randomness collection on Cygwin.
3540 [Ulf Möller]
3541
3542 *) Fix hang in EGD/PRNGD query when communication socket is closed
3543 prematurely by EGD/PRNGD.
3544 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3545
3546 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3547 [Steve Henson]
3548
3549 *) Back-port of selected performance improvements from development
3550 branch, as well as improved support for PowerPC platforms.
3551 [Andy Polyakov]
3552
3553 *) Add lots of checks for memory allocation failure, error codes to indicate
3554 failure and freeing up memory if a failure occurs.
3555 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3556
3557 *) Add new -passin argument to dgst.
3558 [Steve Henson]
3559
3560 *) Perform some character comparisons of different types in X509_NAME_cmp:
3561 this is needed for some certificates that reencode DNs into UTF8Strings
3562 (in violation of RFC3280) and can't or wont issue name rollover
3563 certificates.
3564 [Steve Henson]
3565
3566 *) Make an explicit check during certificate validation to see that
3567 the CA setting in each certificate on the chain is correct. As a
3568 side effect always do the following basic checks on extensions,
3569 not just when there's an associated purpose to the check:
3570
3571 - if there is an unhandled critical extension (unless the user
3572 has chosen to ignore this fault)
3573 - if the path length has been exceeded (if one is set at all)
3574 - that certain extensions fit the associated purpose (if one has
3575 been given)
3576 [Richard Levitte]
3577
3578 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3579
3580 *) Avoid a race condition when CRLs are checked in a multi threaded
3581 environment. This would happen due to the reordering of the revoked
3582 entries during signature checking and serial number lookup. Now the
3583 encoding is cached and the serial number sort performed under a lock.
3584 Add new STACK function sk_is_sorted().
3585 [Steve Henson]
3586
3587 *) Add Delta CRL to the extension code.
3588 [Steve Henson]
3589
3590 *) Various fixes to s3_pkt.c so alerts are sent properly.
3591 [David Holmes <d.holmes@f5.com>]
3592
3593 *) Reduce the chances of duplicate issuer name and serial numbers (in
3594 violation of RFC3280) using the OpenSSL certificate creation utilities.
3595 This is done by creating a random 64 bit value for the initial serial
3596 number when a serial number file is created or when a self signed
3597 certificate is created using 'openssl req -x509'. The initial serial
3598 number file is created using 'openssl x509 -next_serial' in CA.pl
3599 rather than being initialized to 1.
3600 [Steve Henson]
3601
3602 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3603
3604 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3605 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3606 [Joe Orton, Steve Henson]
3607
3608 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3609 (CVE-2004-0112)
3610 [Joe Orton, Steve Henson]
3611
3612 *) Make it possible to have multiple active certificates with the same
3613 subject in the CA index file. This is done only if the keyword
3614 'unique_subject' is set to 'no' in the main CA section (default
3615 if 'CA_default') of the configuration file. The value is saved
3616 with the database itself in a separate index attribute file,
3617 named like the index file with '.attr' appended to the name.
3618 [Richard Levitte]
3619
3620 *) X509 verify fixes. Disable broken certificate workarounds when
3621 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3622 keyUsage extension present. Don't accept CRLs with unhandled critical
3623 extensions: since verify currently doesn't process CRL extensions this
3624 rejects a CRL with *any* critical extensions. Add new verify error codes
3625 for these cases.
3626 [Steve Henson]
3627
3628 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3629 A clarification of RFC2560 will require the use of OCTET STRINGs and
3630 some implementations cannot handle the current raw format. Since OpenSSL
3631 copies and compares OCSP nonces as opaque blobs without any attempt at
3632 parsing them this should not create any compatibility issues.
3633 [Steve Henson]
3634
3635 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3636 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3637 this HMAC (and other) operations are several times slower than OpenSSL
3638 < 0.9.7.
3639 [Steve Henson]
3640
3641 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3642 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3643
3644 *) Use the correct content when signing type "other".
3645 [Steve Henson]
3646
3647 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3648
3649 *) Fix various bugs revealed by running the NISCC test suite:
3650
3651 Stop out of bounds reads in the ASN1 code when presented with
3652 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3653
3654 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3655
3656 If verify callback ignores invalid public key errors don't try to check
3657 certificate signature with the NULL public key.
3658
3659 [Steve Henson]
3660
3661 *) New -ignore_err option in ocsp application to stop the server
3662 exiting on the first error in a request.
3663 [Steve Henson]
3664
3665 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3666 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3667 specifications.
3668 [Steve Henson]
3669
3670 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3671 extra data after the compression methods not only for TLS 1.0
3672 but also for SSL 3.0 (as required by the specification).
3673 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3674
3675 *) Change X509_certificate_type() to mark the key as exported/exportable
3676 when it's 512 *bits* long, not 512 bytes.
3677 [Richard Levitte]
3678
3679 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3680 blocks during encryption.
3681 [Richard Levitte]
3682
3683 *) Various fixes to base64 BIO and non blocking I/O. On write
3684 flushes were not handled properly if the BIO retried. On read
3685 data was not being buffered properly and had various logic bugs.
3686 This also affects blocking I/O when the data being decoded is a
3687 certain size.
3688 [Steve Henson]
3689
3690 *) Various S/MIME bugfixes and compatibility changes:
3691 output correct application/pkcs7 MIME type if
3692 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3693 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3694 of files as .eml work). Correctly handle very long lines in MIME
3695 parser.
3696 [Steve Henson]
3697
3698 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3699
3700 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3701 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3702 a protocol version number mismatch like a decryption error
3703 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3704 [Bodo Moeller]
3705
3706 *) Turn on RSA blinding by default in the default implementation
3707 to avoid a timing attack. Applications that don't want it can call
3708 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3709 They would be ill-advised to do so in most cases.
3710 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3711
3712 *) Change RSA blinding code so that it works when the PRNG is not
3713 seeded (in this case, the secret RSA exponent is abused as
3714 an unpredictable seed -- if it is not unpredictable, there
3715 is no point in blinding anyway). Make RSA blinding thread-safe
3716 by remembering the creator's thread ID in rsa->blinding and
3717 having all other threads use local one-time blinding factors
3718 (this requires more computation than sharing rsa->blinding, but
3719 avoids excessive locking; and if an RSA object is not shared
3720 between threads, blinding will still be very fast).
3721 [Bodo Moeller]
3722
3723 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3724 ENGINE as defaults for all supported algorithms irrespective of
3725 the 'flags' parameter. 'flags' is now honoured, so applications
3726 should make sure they are passing it correctly.
3727 [Geoff Thorpe]
3728
3729 *) Target "mingw" now allows native Windows code to be generated in
3730 the Cygwin environment as well as with the MinGW compiler.
3731 [Ulf Moeller]
3732
3733 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3734
3735 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3736 via timing by performing a MAC computation even if incorrrect
3737 block cipher padding has been found. This is a countermeasure
3738 against active attacks where the attacker has to distinguish
3739 between bad padding and a MAC verification error. (CVE-2003-0078)
3740
3741 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3742 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3743 Martin Vuagnoux (EPFL, Ilion)]
3744
3745 *) Make the no-err option work as intended. The intention with no-err
3746 is not to have the whole error stack handling routines removed from
3747 libcrypto, it's only intended to remove all the function name and
3748 reason texts, thereby removing some of the footprint that may not
3749 be interesting if those errors aren't displayed anyway.
3750
3751 NOTE: it's still possible for any application or module to have it's
3752 own set of error texts inserted. The routines are there, just not
3753 used by default when no-err is given.
3754 [Richard Levitte]
3755
3756 *) Add support for FreeBSD on IA64.
3757 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3758
3759 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3760 Kerberos function mit_des_cbc_cksum(). Before this change,
3761 the value returned by DES_cbc_cksum() was like the one from
3762 mit_des_cbc_cksum(), except the bytes were swapped.
3763 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3764
3765 *) Allow an application to disable the automatic SSL chain building.
3766 Before this a rather primitive chain build was always performed in
3767 ssl3_output_cert_chain(): an application had no way to send the
3768 correct chain if the automatic operation produced an incorrect result.
3769
3770 Now the chain builder is disabled if either:
3771
3772 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3773
3774 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3775
3776 The reasoning behind this is that an application would not want the
3777 auto chain building to take place if extra chain certificates are
3778 present and it might also want a means of sending no additional
3779 certificates (for example the chain has two certificates and the
3780 root is omitted).
3781 [Steve Henson]
3782
3783 *) Add the possibility to build without the ENGINE framework.
3784 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3785
3786 *) Under Win32 gmtime() can return NULL: check return value in
3787 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3788 [Steve Henson]
3789
3790 *) DSA routines: under certain error conditions uninitialized BN objects
3791 could be freed. Solution: make sure initialization is performed early
3792 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3793 Nils Larsch <nla@trustcenter.de> via PR#459)
3794 [Lutz Jaenicke]
3795
3796 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3797 checked on reconnect on the client side, therefore session resumption
3798 could still fail with a "ssl session id is different" error. This
3799 behaviour is masked when SSL_OP_ALL is used due to
3800 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3801 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3802 followup to PR #377.
3803 [Lutz Jaenicke]
3804
3805 *) IA-32 assembler support enhancements: unified ELF targets, support
3806 for SCO/Caldera platforms, fix for Cygwin shared build.
3807 [Andy Polyakov]
3808
3809 *) Add support for FreeBSD on sparc64. As a consequence, support for
3810 FreeBSD on non-x86 processors is separate from x86 processors on
3811 the config script, much like the NetBSD support.
3812 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3813
3814 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3815
3816 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3817 OpenSSL 0.9.7.]
3818
3819 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3820 code (06) was taken as the first octet of the session ID and the last
3821 octet was ignored consequently. As a result SSLv2 client side session
3822 caching could not have worked due to the session ID mismatch between
3823 client and server.
3824 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3825 PR #377.
3826 [Lutz Jaenicke]
3827
3828 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3829 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3830 removed entirely.
3831 [Richard Levitte]
3832
3833 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3834 seems that in spite of existing for more than a year, many application
3835 author have done nothing to provide the necessary callbacks, which
3836 means that this particular engine will not work properly anywhere.
3837 This is a very unfortunate situation which forces us, in the name
3838 of usability, to give the hw_ncipher.c a static lock, which is part
3839 of libcrypto.
3840 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3841 appear in 0.9.8 or later. We EXPECT application authors to have
3842 dealt properly with this when 0.9.8 is released (unless we actually
3843 make such changes in the libcrypto locking code that changes will
3844 have to be made anyway).
3845 [Richard Levitte]
3846
3847 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3848 octets have been read, EOF or an error occurs. Without this change
3849 some truncated ASN1 structures will not produce an error.
3850 [Steve Henson]
3851
3852 *) Disable Heimdal support, since it hasn't been fully implemented.
3853 Still give the possibility to force the use of Heimdal, but with
3854 warnings and a request that patches get sent to openssl-dev.
3855 [Richard Levitte]
3856
3857 *) Add the VC-CE target, introduce the WINCE sysname, and add
3858 INSTALL.WCE and appropriate conditionals to make it build.
3859 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3860
3861 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3862 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3863 edit numbers of the version.
3864 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3865
3866 *) Introduce safe string copy and catenation functions
3867 (BUF_strlcpy() and BUF_strlcat()).
3868 [Ben Laurie (CHATS) and Richard Levitte]
3869
3870 *) Avoid using fixed-size buffers for one-line DNs.
3871 [Ben Laurie (CHATS)]
3872
3873 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3874 resizing buffers containing secrets, and use where appropriate.
3875 [Ben Laurie (CHATS)]
3876
3877 *) Avoid using fixed size buffers for configuration file location.
3878 [Ben Laurie (CHATS)]
3879
3880 *) Avoid filename truncation for various CA files.
3881 [Ben Laurie (CHATS)]
3882
3883 *) Use sizeof in preference to magic numbers.
3884 [Ben Laurie (CHATS)]
3885
3886 *) Avoid filename truncation in cert requests.
3887 [Ben Laurie (CHATS)]
3888
3889 *) Add assertions to check for (supposedly impossible) buffer
3890 overflows.
3891 [Ben Laurie (CHATS)]
3892
3893 *) Don't cache truncated DNS entries in the local cache (this could
3894 potentially lead to a spoofing attack).
3895 [Ben Laurie (CHATS)]
3896
3897 *) Fix various buffers to be large enough for hex/decimal
3898 representations in a platform independent manner.
3899 [Ben Laurie (CHATS)]
3900
3901 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3902 resizing buffers containing secrets, and use where appropriate.
3903 [Ben Laurie (CHATS)]
3904
3905 *) Add BIO_indent() to avoid much slightly worrying code to do
3906 indents.
3907 [Ben Laurie (CHATS)]
3908
3909 *) Convert sprintf()/BIO_puts() to BIO_printf().
3910 [Ben Laurie (CHATS)]
3911
3912 *) buffer_gets() could terminate with the buffer only half
3913 full. Fixed.
3914 [Ben Laurie (CHATS)]
3915
3916 *) Add assertions to prevent user-supplied crypto functions from
3917 overflowing internal buffers by having large block sizes, etc.
3918 [Ben Laurie (CHATS)]
3919
3920 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3921 unconditionally).
3922 [Ben Laurie (CHATS)]
3923
3924 *) Eliminate unused copy of key in RC4.
3925 [Ben Laurie (CHATS)]
3926
3927 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3928 [Ben Laurie (CHATS)]
3929
3930 *) Fix off-by-one error in EGD path.
3931 [Ben Laurie (CHATS)]
3932
3933 *) If RANDFILE path is too long, ignore instead of truncating.
3934 [Ben Laurie (CHATS)]
3935
3936 *) Eliminate unused and incorrectly sized X.509 structure
3937 CBCParameter.
3938 [Ben Laurie (CHATS)]
3939
3940 *) Eliminate unused and dangerous function knumber().
3941 [Ben Laurie (CHATS)]
3942
3943 *) Eliminate unused and dangerous structure, KSSL_ERR.
3944 [Ben Laurie (CHATS)]
3945
3946 *) Protect against overlong session ID context length in an encoded
3947 session object. Since these are local, this does not appear to be
3948 exploitable.
3949 [Ben Laurie (CHATS)]
3950
3951 *) Change from security patch (see 0.9.6e below) that did not affect
3952 the 0.9.6 release series:
3953
3954 Remote buffer overflow in SSL3 protocol - an attacker could
3955 supply an oversized master key in Kerberos-enabled versions.
3956 (CVE-2002-0657)
3957 [Ben Laurie (CHATS)]
3958
3959 *) Change the SSL kerb5 codes to match RFC 2712.
3960 [Richard Levitte]
3961
3962 *) Make -nameopt work fully for req and add -reqopt switch.
3963 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3964
3965 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3966 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3967
3968 *) Make sure tests can be performed even if the corresponding algorithms
3969 have been removed entirely. This was also the last step to make
3970 OpenSSL compilable with DJGPP under all reasonable conditions.
3971 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3972
3973 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3974 to allow version independent disabling of normally unselected ciphers,
3975 which may be activated as a side-effect of selecting a single cipher.
3976
3977 (E.g., cipher list string "RSA" enables ciphersuites that are left
3978 out of "ALL" because they do not provide symmetric encryption.
3979 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3980 [Lutz Jaenicke, Bodo Moeller]
3981
3982 *) Add appropriate support for separate platform-dependent build
3983 directories. The recommended way to make a platform-dependent
3984 build directory is the following (tested on Linux), maybe with
3985 some local tweaks:
3986
3987 # Place yourself outside of the OpenSSL source tree. In
3988 # this example, the environment variable OPENSSL_SOURCE
3989 # is assumed to contain the absolute OpenSSL source directory.
3990 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3991 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3992 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3993 mkdir -p `dirname $F`
3994 ln -s $OPENSSL_SOURCE/$F $F
3995 done
3996
3997 To be absolutely sure not to disturb the source tree, a "make clean"
3998 is a good thing. If it isn't successfull, don't worry about it,
3999 it probably means the source directory is very clean.
4000 [Richard Levitte]
4001
4002 *) Make sure any ENGINE control commands make local copies of string
4003 pointers passed to them whenever necessary. Otherwise it is possible
4004 the caller may have overwritten (or deallocated) the original string
4005 data when a later ENGINE operation tries to use the stored values.
4006 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4007
4008 *) Improve diagnostics in file reading and command-line digests.
4009 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4010
4011 *) Add AES modes CFB and OFB to the object database. Correct an
4012 error in AES-CFB decryption.
4013 [Richard Levitte]
4014
4015 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4016 allows existing EVP_CIPHER_CTX structures to be reused after
4017 calling EVP_*Final(). This behaviour is used by encryption
4018 BIOs and some applications. This has the side effect that
4019 applications must explicitly clean up cipher contexts with
4020 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4021 [Steve Henson]
4022
4023 *) Check the values of dna and dnb in bn_mul_recursive before calling
4024 bn_mul_comba (a non zero value means the a or b arrays do not contain
4025 n2 elements) and fallback to bn_mul_normal if either is not zero.
4026 [Steve Henson]
4027
4028 *) Fix escaping of non-ASCII characters when using the -subj option
4029 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4030 [Lutz Jaenicke]
4031
4032 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4033 form for "surname", serialNumber has no short form.
4034 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4035 therefore remove "mail" short name for "internet 7".
4036 The OID for unique identifiers in X509 certificates is
4037 x500UniqueIdentifier, not uniqueIdentifier.
4038 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4039 [Lutz Jaenicke]
4040
4041 *) Add an "init" command to the ENGINE config module and auto initialize
4042 ENGINEs. Without any "init" command the ENGINE will be initialized
4043 after all ctrl commands have been executed on it. If init=1 the
4044 ENGINE is initailized at that point (ctrls before that point are run
4045 on the uninitialized ENGINE and after on the initialized one). If
4046 init=0 then the ENGINE will not be iniatialized at all.
4047 [Steve Henson]
4048
4049 *) Fix the 'app_verify_callback' interface so that the user-defined
4050 argument is actually passed to the callback: In the
4051 SSL_CTX_set_cert_verify_callback() prototype, the callback
4052 declaration has been changed from
4053 int (*cb)()
4054 into
4055 int (*cb)(X509_STORE_CTX *,void *);
4056 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4057 i=s->ctx->app_verify_callback(&ctx)
4058 has been changed into
4059 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4060
4061 To update applications using SSL_CTX_set_cert_verify_callback(),
4062 a dummy argument can be added to their callback functions.
4063 [D. K. Smetters <smetters@parc.xerox.com>]
4064
4065 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4066 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4067
4068 *) Add and OPENSSL_LOAD_CONF define which will cause
4069 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4070 This allows older applications to transparently support certain
4071 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4072 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4073 load the config file and OPENSSL_add_all_algorithms_conf() which will
4074 always load it have also been added.
4075 [Steve Henson]
4076
4077 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4078 Adjust NIDs and EVP layer.
4079 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4080
4081 *) Config modules support in openssl utility.
4082
4083 Most commands now load modules from the config file,
4084 though in a few (such as version) this isn't done
4085 because it couldn't be used for anything.
4086
4087 In the case of ca and req the config file used is
4088 the same as the utility itself: that is the -config
4089 command line option can be used to specify an
4090 alternative file.
4091 [Steve Henson]
4092
4093 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4094 use "openssl_conf" if filename is NULL use default openssl config file.
4095 [Steve Henson]
4096
4097 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4098 config section name. Add a new flag to tolerate a missing config file
4099 and move code to CONF_modules_load_file().
4100 [Steve Henson]
4101
4102 *) Support for crypto accelerator cards from Accelerated Encryption
4103 Processing, www.aep.ie. (Use engine 'aep')
4104 The support was copied from 0.9.6c [engine] and adapted/corrected
4105 to work with the new engine framework.
4106 [AEP Inc. and Richard Levitte]
4107
4108 *) Support for SureWare crypto accelerator cards from Baltimore
4109 Technologies. (Use engine 'sureware')
4110 The support was copied from 0.9.6c [engine] and adapted
4111 to work with the new engine framework.
4112 [Richard Levitte]
4113
4114 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4115 make the newer ENGINE framework commands for the CHIL engine work.
4116 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4117
4118 *) Make it possible to produce shared libraries on ReliantUNIX.
4119 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4120
4121 *) Add the configuration target debug-linux-ppro.
4122 Make 'openssl rsa' use the general key loading routines
4123 implemented in apps.c, and make those routines able to
4124 handle the key format FORMAT_NETSCAPE and the variant
4125 FORMAT_IISSGC.
4126 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4127
4128 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4129 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4130
4131 *) Add -keyform to rsautl, and document -engine.
4132 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4133
4134 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4135 BIO_R_NO_SUCH_FILE error code rather than the generic
4136 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4137 [Ben Laurie]
4138
4139 *) Add new functions
4140 ERR_peek_last_error
4141 ERR_peek_last_error_line
4142 ERR_peek_last_error_line_data.
4143 These are similar to
4144 ERR_peek_error
4145 ERR_peek_error_line
4146 ERR_peek_error_line_data,
4147 but report on the latest error recorded rather than the first one
4148 still in the error queue.
4149 [Ben Laurie, Bodo Moeller]
4150
4151 *) default_algorithms option in ENGINE config module. This allows things
4152 like:
4153 default_algorithms = ALL
4154 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4155 [Steve Henson]
4156
4157 *) Prelminary ENGINE config module.
4158 [Steve Henson]
4159
4160 *) New experimental application configuration code.
4161 [Steve Henson]
4162
4163 *) Change the AES code to follow the same name structure as all other
4164 symmetric ciphers, and behave the same way. Move everything to
4165 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4166 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4167
4168 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4169 [Ben Laurie and Theo de Raadt]
4170
4171 *) Add option to output public keys in req command.
4172 [Massimiliano Pala madwolf@openca.org]
4173
4174 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4175 (up to about 10% better than before for P-192 and P-224).
4176 [Bodo Moeller]
4177
4178 *) New functions/macros
4179
4180 SSL_CTX_set_msg_callback(ctx, cb)
4181 SSL_CTX_set_msg_callback_arg(ctx, arg)
4182 SSL_set_msg_callback(ssl, cb)
4183 SSL_set_msg_callback_arg(ssl, arg)
4184
4185 to request calling a callback function
4186
4187 void cb(int write_p, int version, int content_type,
4188 const void *buf, size_t len, SSL *ssl, void *arg)
4189
4190 whenever a protocol message has been completely received
4191 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4192 protocol version according to which the SSL library interprets
4193 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4194 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4195 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4196 specification (change_cipher_spec(20), alert(21), handshake(22)).
4197 'buf' and 'len' point to the actual message, 'ssl' to the
4198 SSL object, and 'arg' is the application-defined value set by
4199 SSL[_CTX]_set_msg_callback_arg().
4200
4201 'openssl s_client' and 'openssl s_server' have new '-msg' options
4202 to enable a callback that displays all protocol messages.
4203 [Bodo Moeller]
4204
4205 *) Change the shared library support so shared libraries are built as
4206 soon as the corresponding static library is finished, and thereby get
4207 openssl and the test programs linked against the shared library.
4208 This still only happens when the keyword "shard" has been given to
4209 the configuration scripts.
4210
4211 NOTE: shared library support is still an experimental thing, and
4212 backward binary compatibility is still not guaranteed.
4213 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4214
4215 *) Add support for Subject Information Access extension.
4216 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4217
4218 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4219 additional bytes when new memory had to be allocated, not just
4220 when reusing an existing buffer.
4221 [Bodo Moeller]
4222
4223 *) New command line and configuration option 'utf8' for the req command.
4224 This allows field values to be specified as UTF8 strings.
4225 [Steve Henson]
4226
4227 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4228 runs for the former and machine-readable output for the latter.
4229 [Ben Laurie]
4230
4231 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4232 of the e-mail address in the DN (i.e., it will go into a certificate
4233 extension only). The new configuration file option 'email_in_dn = no'
4234 has the same effect.
4235 [Massimiliano Pala madwolf@openca.org]
4236
4237 *) Change all functions with names starting with des_ to be starting
4238 with DES_ instead. Add wrappers that are compatible with libdes,
4239 but are named _ossl_old_des_*. Finally, add macros that map the
4240 des_* symbols to the corresponding _ossl_old_des_* if libdes
4241 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4242 desired, the des_* symbols will be mapped to DES_*, with one
4243 exception.
4244
4245 Since we provide two compatibility mappings, the user needs to
4246 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4247 compatibility is desired. The default (i.e., when that macro
4248 isn't defined) is OpenSSL 0.9.6c compatibility.
4249
4250 There are also macros that enable and disable the support of old
4251 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4252 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4253 are defined, the default will apply: to support the old des routines.
4254
4255 In either case, one must include openssl/des.h to get the correct
4256 definitions. Do not try to just include openssl/des_old.h, that
4257 won't work.
4258
4259 NOTE: This is a major break of an old API into a new one. Software
4260 authors are encouraged to switch to the DES_ style functions. Some
4261 time in the future, des_old.h and the libdes compatibility functions
4262 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4263 default), and then completely removed.
4264 [Richard Levitte]
4265
4266 *) Test for certificates which contain unsupported critical extensions.
4267 If such a certificate is found during a verify operation it is
4268 rejected by default: this behaviour can be overridden by either
4269 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4270 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4271 X509_supported_extension() has also been added which returns 1 if a
4272 particular extension is supported.
4273 [Steve Henson]
4274
4275 *) Modify the behaviour of EVP cipher functions in similar way to digests
4276 to retain compatibility with existing code.
4277 [Steve Henson]
4278
4279 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4280 compatibility with existing code. In particular the 'ctx' parameter does
4281 not have to be to be initialized before the call to EVP_DigestInit() and
4282 it is tidied up after a call to EVP_DigestFinal(). New function
4283 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4284 EVP_MD_CTX_copy() changed to not require the destination to be
4285 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4286 requires the destination to be valid.
4287
4288 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4289 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4290 [Steve Henson]
4291
4292 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4293 so that complete 'Handshake' protocol structures are kept in memory
4294 instead of overwriting 'msg_type' and 'length' with 'body' data.
4295 [Bodo Moeller]
4296
4297 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4298 [Massimo Santin via Richard Levitte]
4299
4300 *) Major restructuring to the underlying ENGINE code. This includes
4301 reduction of linker bloat, separation of pure "ENGINE" manipulation
4302 (initialisation, etc) from functionality dealing with implementations
4303 of specific crypto iterfaces. This change also introduces integrated
4304 support for symmetric ciphers and digest implementations - so ENGINEs
4305 can now accelerate these by providing EVP_CIPHER and EVP_MD
4306 implementations of their own. This is detailed in crypto/engine/README
4307 as it couldn't be adequately described here. However, there are a few
4308 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4309 were changed in the original introduction of ENGINE code have now
4310 reverted back - the hooking from this code to ENGINE is now a good
4311 deal more passive and at run-time, operations deal directly with
4312 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4313 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4314 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4315 they were not being used by the framework as there is no concept of a
4316 BIGNUM_METHOD and they could not be generalised to the new
4317 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4318 ENGINE_cpy() has been removed as it cannot be consistently defined in
4319 the new code.
4320 [Geoff Thorpe]
4321
4322 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4323 [Steve Henson]
4324
4325 *) Change mkdef.pl to sort symbols that get the same entry number,
4326 and make sure the automatically generated functions ERR_load_*
4327 become part of libeay.num as well.
4328 [Richard Levitte]
4329
4330 *) New function SSL_renegotiate_pending(). This returns true once
4331 renegotiation has been requested (either SSL_renegotiate() call
4332 or HelloRequest/ClientHello receveived from the peer) and becomes
4333 false once a handshake has been completed.
4334 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4335 sends a HelloRequest, but does not ensure that a handshake takes
4336 place. SSL_renegotiate_pending() is useful for checking if the
4337 client has followed the request.)
4338 [Bodo Moeller]
4339
4340 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4341 By default, clients may request session resumption even during
4342 renegotiation (if session ID contexts permit); with this option,
4343 session resumption is possible only in the first handshake.
4344
4345 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4346 more bits available for options that should not be part of
4347 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4348 [Bodo Moeller]
4349
4350 *) Add some demos for certificate and certificate request creation.
4351 [Steve Henson]
4352
4353 *) Make maximum certificate chain size accepted from the peer application
4354 settable (SSL*_get/set_max_cert_list()), as proposed by
4355 "Douglas E. Engert" <deengert@anl.gov>.
4356 [Lutz Jaenicke]
4357
4358 *) Add support for shared libraries for Unixware-7
4359 (Boyd Lynn Gerber <gerberb@zenez.com>).
4360 [Lutz Jaenicke]
4361
4362 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4363 be done prior to destruction. Use this to unload error strings from
4364 ENGINEs that load their own error strings. NB: This adds two new API
4365 functions to "get" and "set" this destroy handler in an ENGINE.
4366 [Geoff Thorpe]
4367
4368 *) Alter all existing ENGINE implementations (except "openssl" and
4369 "openbsd") to dynamically instantiate their own error strings. This
4370 makes them more flexible to be built both as statically-linked ENGINEs
4371 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4372 Also, add stub code to each that makes building them as self-contained
4373 shared-libraries easier (see README.ENGINE).
4374 [Geoff Thorpe]
4375
4376 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4377 implementations into applications that are completely implemented in
4378 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4379 commands that can be used to configure what shared-library to load and
4380 to control aspects of the way it is handled. Also, made an update to
4381 the README.ENGINE file that brings its information up-to-date and
4382 provides some information and instructions on the "dynamic" ENGINE
4383 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4384 [Geoff Thorpe]
4385
4386 *) Make it possible to unload ranges of ERR strings with a new
4387 "ERR_unload_strings" function.
4388 [Geoff Thorpe]
4389
4390 *) Add a copy() function to EVP_MD.
4391 [Ben Laurie]
4392
4393 *) Make EVP_MD routines take a context pointer instead of just the
4394 md_data void pointer.
4395 [Ben Laurie]
4396
4397 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4398 that the digest can only process a single chunk of data
4399 (typically because it is provided by a piece of
4400 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4401 is only going to provide a single chunk of data, and hence the
4402 framework needn't accumulate the data for oneshot drivers.
4403 [Ben Laurie]
4404
4405 *) As with "ERR", make it possible to replace the underlying "ex_data"
4406 functions. This change also alters the storage and management of global
4407 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4408 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4409 index counters. The API functions that use this state have been changed
4410 to take a "class_index" rather than pointers to the class's local STACK
4411 and counter, and there is now an API function to dynamically create new
4412 classes. This centralisation allows us to (a) plug a lot of the
4413 thread-safety problems that existed, and (b) makes it possible to clean
4414 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4415 such data would previously have always leaked in application code and
4416 workarounds were in place to make the memory debugging turn a blind eye
4417 to it. Application code that doesn't use this new function will still
4418 leak as before, but their memory debugging output will announce it now
4419 rather than letting it slide.
4420
4421 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4422 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4423 has a return value to indicate success or failure.
4424 [Geoff Thorpe]
4425
4426 *) Make it possible to replace the underlying "ERR" functions such that the
4427 global state (2 LHASH tables and 2 locks) is only used by the "default"
4428 implementation. This change also adds two functions to "get" and "set"
4429 the implementation prior to it being automatically set the first time
4430 any other ERR function takes place. Ie. an application can call "get",
4431 pass the return value to a module it has just loaded, and that module
4432 can call its own "set" function using that value. This means the
4433 module's "ERR" operations will use (and modify) the error state in the
4434 application and not in its own statically linked copy of OpenSSL code.
4435 [Geoff Thorpe]
4436
4437 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4438 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4439 the operation, and provides a more encapsulated way for external code
4440 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4441 to use these functions rather than manually incrementing the counts.
4442
4443 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4444 [Geoff Thorpe]
4445
4446 *) Add EVP test program.
4447 [Ben Laurie]
4448
4449 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4450 [Ben Laurie]
4451
4452 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4453 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4454 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4455 These allow a CRL to be built without having to access X509_CRL fields
4456 directly. Modify 'ca' application to use new functions.
4457 [Steve Henson]
4458
4459 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4460 bug workarounds. Rollback attack detection is a security feature.
4461 The problem will only arise on OpenSSL servers when TLSv1 is not
4462 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4463 Software authors not wanting to support TLSv1 will have special reasons
4464 for their choice and can explicitly enable this option.
4465 [Bodo Moeller, Lutz Jaenicke]
4466
4467 *) Rationalise EVP so it can be extended: don't include a union of
4468 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4469 (similar to those existing for EVP_CIPHER_CTX).
4470 Usage example:
4471
4472 EVP_MD_CTX md;
4473
4474 EVP_MD_CTX_init(&md); /* new function call */
4475 EVP_DigestInit(&md, EVP_sha1());
4476 EVP_DigestUpdate(&md, in, len);
4477 EVP_DigestFinal(&md, out, NULL);
4478 EVP_MD_CTX_cleanup(&md); /* new function call */
4479
4480 [Ben Laurie]
4481
4482 *) Make DES key schedule conform to the usual scheme, as well as
4483 correcting its structure. This means that calls to DES functions
4484 now have to pass a pointer to a des_key_schedule instead of a
4485 plain des_key_schedule (which was actually always a pointer
4486 anyway): E.g.,
4487
4488 des_key_schedule ks;
4489
4490 des_set_key_checked(..., &ks);
4491 des_ncbc_encrypt(..., &ks, ...);
4492
4493 (Note that a later change renames 'des_...' into 'DES_...'.)
4494 [Ben Laurie]
4495
4496 *) Initial reduction of linker bloat: the use of some functions, such as
4497 PEM causes large amounts of unused functions to be linked in due to
4498 poor organisation. For example pem_all.c contains every PEM function
4499 which has a knock on effect of linking in large amounts of (unused)
4500 ASN1 code. Grouping together similar functions and splitting unrelated
4501 functions prevents this.
4502 [Steve Henson]
4503
4504 *) Cleanup of EVP macros.
4505 [Ben Laurie]
4506
4507 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4508 correct _ecb suffix.
4509 [Ben Laurie]
4510
4511 *) Add initial OCSP responder support to ocsp application. The
4512 revocation information is handled using the text based index
4513 use by the ca application. The responder can either handle
4514 requests generated internally, supplied in files (for example
4515 via a CGI script) or using an internal minimal server.
4516 [Steve Henson]
4517
4518 *) Add configuration choices to get zlib compression for TLS.
4519 [Richard Levitte]
4520
4521 *) Changes to Kerberos SSL for RFC 2712 compliance:
4522 1. Implemented real KerberosWrapper, instead of just using
4523 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4524 2. Implemented optional authenticator field of KerberosWrapper.
4525
4526 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4527 and authenticator structs; see crypto/krb5/.
4528
4529 Generalized Kerberos calls to support multiple Kerberos libraries.
4530 [Vern Staats <staatsvr@asc.hpc.mil>,
4531 Jeffrey Altman <jaltman@columbia.edu>
4532 via Richard Levitte]
4533
4534 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4535 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4536 values for each of the key sizes rather than having just
4537 parameters (and 'speed' generating keys each time).
4538 [Geoff Thorpe]
4539
4540 *) Speed up EVP routines.
4541 Before:
4542 encrypt
4543 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4544 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4545 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4546 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4547 decrypt
4548 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4549 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4550 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4551 After:
4552 encrypt
4553 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4554 decrypt
4555 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4556 [Ben Laurie]
4557
4558 *) Added the OS2-EMX target.
4559 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4560
4561 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4562 to support NCONF routines in extension code. New function CONF_set_nconf()
4563 to allow functions which take an NCONF to also handle the old LHASH
4564 structure: this means that the old CONF compatible routines can be
4565 retained (in particular wrt extensions) without having to duplicate the
4566 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4567 [Steve Henson]
4568
4569 *) Enhance the general user interface with mechanisms for inner control
4570 and with possibilities to have yes/no kind of prompts.
4571 [Richard Levitte]
4572
4573 *) Change all calls to low level digest routines in the library and
4574 applications to use EVP. Add missing calls to HMAC_cleanup() and
4575 don't assume HMAC_CTX can be copied using memcpy().
4576 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4577
4578 *) Add the possibility to control engines through control names but with
4579 arbitrary arguments instead of just a string.
4580 Change the key loaders to take a UI_METHOD instead of a callback
4581 function pointer. NOTE: this breaks binary compatibility with earlier
4582 versions of OpenSSL [engine].
4583 Adapt the nCipher code for these new conditions and add a card insertion
4584 callback.
4585 [Richard Levitte]
4586
4587 *) Enhance the general user interface with mechanisms to better support
4588 dialog box interfaces, application-defined prompts, the possibility
4589 to use defaults (for example default passwords from somewhere else)
4590 and interrupts/cancellations.
4591 [Richard Levitte]
4592
4593 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4594 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4595 [Steve Henson]
4596
4597 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4598 tidy up some unnecessarily weird code in 'sk_new()').
4599 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4600
4601 *) Change the key loading routines for ENGINEs to use the same kind
4602 callback (pem_password_cb) as all other routines that need this
4603 kind of callback.
4604 [Richard Levitte]
4605
4606 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4607 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4608 than this minimum value is recommended.
4609 [Lutz Jaenicke]
4610
4611 *) New random seeder for OpenVMS, using the system process statistics
4612 that are easily reachable.
4613 [Richard Levitte]
4614
4615 *) Windows apparently can't transparently handle global
4616 variables defined in DLLs. Initialisations such as:
4617
4618 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4619
4620 wont compile. This is used by the any applications that need to
4621 declare their own ASN1 modules. This was fixed by adding the option
4622 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4623 needed for static libraries under Win32.
4624 [Steve Henson]
4625
4626 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4627 setting of purpose and trust fields. New X509_STORE trust and
4628 purpose functions and tidy up setting in other SSL functions.
4629 [Steve Henson]
4630
4631 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4632 structure. These are inherited by X509_STORE_CTX when it is
4633 initialised. This allows various defaults to be set in the
4634 X509_STORE structure (such as flags for CRL checking and custom
4635 purpose or trust settings) for functions which only use X509_STORE_CTX
4636 internally such as S/MIME.
4637
4638 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4639 trust settings if they are not set in X509_STORE. This allows X509_STORE
4640 purposes and trust (in S/MIME for example) to override any set by default.
4641
4642 Add command line options for CRL checking to smime, s_client and s_server
4643 applications.
4644 [Steve Henson]
4645
4646 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4647 are set then the CRL is looked up in the X509_STORE structure and
4648 its validity and signature checked, then if the certificate is found
4649 in the CRL the verify fails with a revoked error.
4650
4651 Various new CRL related callbacks added to X509_STORE_CTX structure.
4652
4653 Command line options added to 'verify' application to support this.
4654
4655 This needs some additional work, such as being able to handle multiple
4656 CRLs with different times, extension based lookup (rather than just
4657 by subject name) and ultimately more complete V2 CRL extension
4658 handling.
4659 [Steve Henson]
4660
4661 *) Add a general user interface API (crypto/ui/). This is designed
4662 to replace things like des_read_password and friends (backward
4663 compatibility functions using this new API are provided).
4664 The purpose is to remove prompting functions from the DES code
4665 section as well as provide for prompting through dialog boxes in
4666 a window system and the like.
4667 [Richard Levitte]
4668
4669 *) Add "ex_data" support to ENGINE so implementations can add state at a
4670 per-structure level rather than having to store it globally.
4671 [Geoff]
4672
4673 *) Make it possible for ENGINE structures to be copied when retrieved by
4674 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4675 This causes the "original" ENGINE structure to act like a template,
4676 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4677 operational state can be localised to each ENGINE structure, despite the
4678 fact they all share the same "methods". New ENGINE structures returned in
4679 this case have no functional references and the return value is the single
4680 structural reference. This matches the single structural reference returned
4681 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4682 ENGINE structure.
4683 [Geoff]
4684
4685 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4686 needs to match any other type at all we need to manually clear the
4687 tag cache.
4688 [Steve Henson]
4689
4690 *) Changes to the "openssl engine" utility to include;
4691 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4692 about an ENGINE's available control commands.
4693 - executing control commands from command line arguments using the
4694 '-pre' and '-post' switches. '-post' is only used if '-t' is
4695 specified and the ENGINE is successfully initialised. The syntax for
4696 the individual commands are colon-separated, for example;
4697 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4698 [Geoff]
4699
4700 *) New dynamic control command support for ENGINEs. ENGINEs can now
4701 declare their own commands (numbers), names (strings), descriptions,
4702 and input types for run-time discovery by calling applications. A
4703 subset of these commands are implicitly classed as "executable"
4704 depending on their input type, and only these can be invoked through
4705 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4706 can be based on user input, config files, etc). The distinction is
4707 that "executable" commands cannot return anything other than a boolean
4708 result and can only support numeric or string input, whereas some
4709 discoverable commands may only be for direct use through
4710 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4711 pointers, or other custom uses. The "executable" commands are to
4712 support parameterisations of ENGINE behaviour that can be
4713 unambiguously defined by ENGINEs and used consistently across any
4714 OpenSSL-based application. Commands have been added to all the
4715 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4716 control over shared-library paths without source code alterations.
4717 [Geoff]
4718
4719 *) Changed all ENGINE implementations to dynamically allocate their
4720 ENGINEs rather than declaring them statically. Apart from this being
4721 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4722 this also allows the implementations to compile without using the
4723 internal engine_int.h header.
4724 [Geoff]
4725
4726 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4727 'const' value. Any code that should be able to modify a RAND_METHOD
4728 should already have non-const pointers to it (ie. they should only
4729 modify their own ones).
4730 [Geoff]
4731
4732 *) Made a variety of little tweaks to the ENGINE code.
4733 - "atalla" and "ubsec" string definitions were moved from header files
4734 to C code. "nuron" string definitions were placed in variables
4735 rather than hard-coded - allowing parameterisation of these values
4736 later on via ctrl() commands.
4737 - Removed unused "#if 0"'d code.
4738 - Fixed engine list iteration code so it uses ENGINE_free() to release
4739 structural references.
4740 - Constified the RAND_METHOD element of ENGINE structures.
4741 - Constified various get/set functions as appropriate and added
4742 missing functions (including a catch-all ENGINE_cpy that duplicates
4743 all ENGINE values onto a new ENGINE except reference counts/state).
4744 - Removed NULL parameter checks in get/set functions. Setting a method
4745 or function to NULL is a way of cancelling out a previously set
4746 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4747 and doesn't justify the extra error symbols and code.
4748 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4749 flags from engine_int.h to engine.h.
4750 - Changed prototypes for ENGINE handler functions (init(), finish(),
4751 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4752 [Geoff]
4753
4754 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4755 to the algorithm using long division. The binary algorithm can be
4756 used only if the modulus is odd. On 32-bit systems, it is faster
4757 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4758 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4759 up to 450 bits. In 64-bit environments, the binary algorithm
4760 appears to be advantageous for much longer moduli; here we use it
4761 for moduli up to 2048 bits.
4762 [Bodo Moeller]
4763
4764 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4765 could not support the combine flag in choice fields.
4766 [Steve Henson]
4767
4768 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4769 extensions from a certificate request to the certificate.
4770 [Steve Henson]
4771
4772 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4773 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4774 file: this allows the display of the certificate about to be
4775 signed to be customised, to allow certain fields to be included
4776 or excluded and extension details. The old system didn't display
4777 multicharacter strings properly, omitted fields not in the policy
4778 and couldn't display additional details such as extensions.
4779 [Steve Henson]
4780
4781 *) Function EC_POINTs_mul for multiple scalar multiplication
4782 of an arbitrary number of elliptic curve points
4783 \sum scalars[i]*points[i],
4784 optionally including the generator defined for the EC_GROUP:
4785 scalar*generator + \sum scalars[i]*points[i].
4786
4787 EC_POINT_mul is a simple wrapper function for the typical case
4788 that the point list has just one item (besides the optional
4789 generator).
4790 [Bodo Moeller]
4791
4792 *) First EC_METHODs for curves over GF(p):
4793
4794 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4795 operations and provides various method functions that can also
4796 operate with faster implementations of modular arithmetic.
4797
4798 EC_GFp_mont_method() reuses most functions that are part of
4799 EC_GFp_simple_method, but uses Montgomery arithmetic.
4800
4801 [Bodo Moeller; point addition and point doubling
4802 implementation directly derived from source code provided by
4803 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4804
4805 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4806 crypto/ec/ec_lib.c):
4807
4808 Curves are EC_GROUP objects (with an optional group generator)
4809 based on EC_METHODs that are built into the library.
4810
4811 Points are EC_POINT objects based on EC_GROUP objects.
4812
4813 Most of the framework would be able to handle curves over arbitrary
4814 finite fields, but as there are no obvious types for fields other
4815 than GF(p), some functions are limited to that for now.
4816 [Bodo Moeller]
4817
4818 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4819 that the file contains a complete HTTP response.
4820 [Richard Levitte]
4821
4822 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4823 change the def and num file printf format specifier from "%-40sXXX"
4824 to "%-39s XXX". The latter will always guarantee a space after the
4825 field while the former will cause them to run together if the field
4826 is 40 of more characters long.
4827 [Steve Henson]
4828
4829 *) Constify the cipher and digest 'method' functions and structures
4830 and modify related functions to take constant EVP_MD and EVP_CIPHER
4831 pointers.
4832 [Steve Henson]
4833
4834 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4835 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4836 [Bodo Moeller]
4837
4838 *) Modify EVP_Digest*() routines so they now return values. Although the
4839 internal software routines can never fail additional hardware versions
4840 might.
4841 [Steve Henson]
4842
4843 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4844
4845 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4846 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4847
4848 ASN1 error codes
4849 ERR_R_NESTED_ASN1_ERROR
4850 ...
4851 ERR_R_MISSING_ASN1_EOS
4852 were 4 .. 9, conflicting with
4853 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4854 ...
4855 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4856 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4857
4858 Add new error code 'ERR_R_INTERNAL_ERROR'.
4859 [Bodo Moeller]
4860
4861 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4862 suffices.
4863 [Bodo Moeller]
4864
4865 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4866 sets the subject name for a new request or supersedes the
4867 subject name in a given request. Formats that can be parsed are
4868 'CN=Some Name, OU=myOU, C=IT'
4869 and
4870 'CN=Some Name/OU=myOU/C=IT'.
4871
4872 Add options '-batch' and '-verbose' to 'openssl req'.
4873 [Massimiliano Pala <madwolf@hackmasters.net>]
4874
4875 *) Introduce the possibility to access global variables through
4876 functions on platform were that's the best way to handle exporting
4877 global variables in shared libraries. To enable this functionality,
4878 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4879 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4880 is normally done by Configure or something similar).
4881
4882 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4883 in the source file (foo.c) like this:
4884
4885 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4886 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4887
4888 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4889 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4890
4891 OPENSSL_DECLARE_GLOBAL(int,foo);
4892 #define foo OPENSSL_GLOBAL_REF(foo)
4893 OPENSSL_DECLARE_GLOBAL(double,bar);
4894 #define bar OPENSSL_GLOBAL_REF(bar)
4895
4896 The #defines are very important, and therefore so is including the
4897 header file everywhere where the defined globals are used.
4898
4899 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4900 of ASN.1 items, but that structure is a bit different.
4901
4902 The largest change is in util/mkdef.pl which has been enhanced with
4903 better and easier to understand logic to choose which symbols should
4904 go into the Windows .def files as well as a number of fixes and code
4905 cleanup (among others, algorithm keywords are now sorted
4906 lexicographically to avoid constant rewrites).
4907 [Richard Levitte]
4908
4909 *) In BN_div() keep a copy of the sign of 'num' before writing the
4910 result to 'rm' because if rm==num the value will be overwritten
4911 and produce the wrong result if 'num' is negative: this caused
4912 problems with BN_mod() and BN_nnmod().
4913 [Steve Henson]
4914
4915 *) Function OCSP_request_verify(). This checks the signature on an
4916 OCSP request and verifies the signer certificate. The signer
4917 certificate is just checked for a generic purpose and OCSP request
4918 trust settings.
4919 [Steve Henson]
4920
4921 *) Add OCSP_check_validity() function to check the validity of OCSP
4922 responses. OCSP responses are prepared in real time and may only
4923 be a few seconds old. Simply checking that the current time lies
4924 between thisUpdate and nextUpdate max reject otherwise valid responses
4925 caused by either OCSP responder or client clock inaccuracy. Instead
4926 we allow thisUpdate and nextUpdate to fall within a certain period of
4927 the current time. The age of the response can also optionally be
4928 checked. Two new options -validity_period and -status_age added to
4929 ocsp utility.
4930 [Steve Henson]
4931
4932 *) If signature or public key algorithm is unrecognized print out its
4933 OID rather that just UNKNOWN.
4934 [Steve Henson]
4935
4936 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4937 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4938 ID to be generated from the issuer certificate alone which can then be
4939 passed to OCSP_id_issuer_cmp().
4940 [Steve Henson]
4941
4942 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4943 ASN1 modules to export functions returning ASN1_ITEM pointers
4944 instead of the ASN1_ITEM structures themselves. This adds several
4945 new macros which allow the underlying ASN1 function/structure to
4946 be accessed transparently. As a result code should not use ASN1_ITEM
4947 references directly (such as &X509_it) but instead use the relevant
4948 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4949 use of the new ASN1 code on platforms where exporting structures
4950 is problematical (for example in shared libraries) but exporting
4951 functions returning pointers to structures is not.
4952 [Steve Henson]
4953
4954 *) Add support for overriding the generation of SSL/TLS session IDs.
4955 These callbacks can be registered either in an SSL_CTX or per SSL.
4956 The purpose of this is to allow applications to control, if they wish,
4957 the arbitrary values chosen for use as session IDs, particularly as it
4958 can be useful for session caching in multiple-server environments. A
4959 command-line switch for testing this (and any client code that wishes
4960 to use such a feature) has been added to "s_server".
4961 [Geoff Thorpe, Lutz Jaenicke]
4962
4963 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4964 of the form '#if defined(...) || defined(...) || ...' and
4965 '#if !defined(...) && !defined(...) && ...'. This also avoids
4966 the growing number of special cases it was previously handling.
4967 [Richard Levitte]
4968
4969 *) Make all configuration macros available for application by making
4970 sure they are available in opensslconf.h, by giving them names starting
4971 with "OPENSSL_" to avoid conflicts with other packages and by making
4972 sure e_os2.h will cover all platform-specific cases together with
4973 opensslconf.h.
4974 Additionally, it is now possible to define configuration/platform-
4975 specific names (called "system identities"). In the C code, these
4976 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4977 macro with the name beginning with "OPENSSL_SYS_", which is determined
4978 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4979 what is available.
4980 [Richard Levitte]
4981
4982 *) New option -set_serial to 'req' and 'x509' this allows the serial
4983 number to use to be specified on the command line. Previously self
4984 signed certificates were hard coded with serial number 0 and the
4985 CA options of 'x509' had to use a serial number in a file which was
4986 auto incremented.
4987 [Steve Henson]
4988
4989 *) New options to 'ca' utility to support V2 CRL entry extensions.
4990 Currently CRL reason, invalidity date and hold instruction are
4991 supported. Add new CRL extensions to V3 code and some new objects.
4992 [Steve Henson]
4993
4994 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4995 disable standard block padding (aka PKCS#5 padding) in the EVP
4996 API, which was previously mandatory. This means that the data is
4997 not padded in any way and so the total length much be a multiple
4998 of the block size, otherwise an error occurs.
4999 [Steve Henson]
5000
5001 *) Initial (incomplete) OCSP SSL support.
5002 [Steve Henson]
5003
5004 *) New function OCSP_parse_url(). This splits up a URL into its host,
5005 port and path components: primarily to parse OCSP URLs. New -url
5006 option to ocsp utility.
5007 [Steve Henson]
5008
5009 *) New nonce behavior. The return value of OCSP_check_nonce() now
5010 reflects the various checks performed. Applications can decide
5011 whether to tolerate certain situations such as an absent nonce
5012 in a response when one was present in a request: the ocsp application
5013 just prints out a warning. New function OCSP_add1_basic_nonce()
5014 this is to allow responders to include a nonce in a response even if
5015 the request is nonce-less.
5016 [Steve Henson]
5017
5018 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5019 skipped when using openssl x509 multiple times on a single input file,
5020 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5021 [Bodo Moeller]
5022
5023 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5024 set string type: to handle setting ASN1_TIME structures. Fix ca
5025 utility to correctly initialize revocation date of CRLs.
5026 [Steve Henson]
5027
5028 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5029 the clients preferred ciphersuites and rather use its own preferences.
5030 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5031 Internet Explorer by ensuring unchanged hash method during stepup.
5032 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5033 [Lutz Jaenicke]
5034
5035 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5036 to aes and add a new 'exist' option to print out symbols that don't
5037 appear to exist.
5038 [Steve Henson]
5039
5040 *) Additional options to ocsp utility to allow flags to be set and
5041 additional certificates supplied.
5042 [Steve Henson]
5043
5044 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5045 OCSP client a number of certificate to only verify the response
5046 signature against.
5047 [Richard Levitte]
5048
5049 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5050 handle the new API. Currently only ECB, CBC modes supported. Add new
5051 AES OIDs.
5052
5053 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5054 Encryption Standard (AES) Ciphersuites for Transport Layer
5055 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5056 not enabled by default and were not part of the "ALL" ciphersuite
5057 alias because they were not yet official; they could be
5058 explicitly requested by specifying the "AESdraft" ciphersuite
5059 group alias. In the final release of OpenSSL 0.9.7, the group
5060 alias is called "AES" and is part of "ALL".)
5061 [Ben Laurie, Steve Henson, Bodo Moeller]
5062
5063 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5064 request to response.
5065 [Steve Henson]
5066
5067 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5068 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5069 extract information from a certificate request. OCSP_response_create()
5070 creates a response and optionally adds a basic response structure.
5071 OCSP_basic_add1_status() adds a complete single response to a basic
5072 response and returns the OCSP_SINGLERESP structure just added (to allow
5073 extensions to be included for example). OCSP_basic_add1_cert() adds a
5074 certificate to a basic response and OCSP_basic_sign() signs a basic
5075 response with various flags. New helper functions ASN1_TIME_check()
5076 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5077 (converts ASN1_TIME to GeneralizedTime).
5078 [Steve Henson]
5079
5080 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5081 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5082 structure from a certificate. X509_pubkey_digest() digests the public_key
5083 contents: this is used in various key identifiers.
5084 [Steve Henson]
5085
5086 *) Make sk_sort() tolerate a NULL argument.
5087 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5088
5089 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5090 passed by the function are trusted implicitly. If any of them signed the
5091 response then it is assumed to be valid and is not verified.
5092 [Steve Henson]
5093
5094 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5095 to data. This was previously part of the PKCS7 ASN1 code. This
5096 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5097 [Steve Henson, reported by Kenneth R. Robinette
5098 <support@securenetterm.com>]
5099
5100 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5101 routines: without these tracing memory leaks is very painful.
5102 Fix leaks in PKCS12 and PKCS7 routines.
5103 [Steve Henson]
5104
5105 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5106 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5107 effectively meant GeneralizedTime would never be used. Now it
5108 is initialised to -1 but X509_time_adj() now has to check the value
5109 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5110 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5111 [Steve Henson, reported by Kenneth R. Robinette
5112 <support@securenetterm.com>]
5113
5114 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5115 result in a zero length in the ASN1_INTEGER structure which was
5116 not consistent with the structure when d2i_ASN1_INTEGER() was used
5117 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5118 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5119 where it did not print out a minus for negative ASN1_INTEGER.
5120 [Steve Henson]
5121
5122 *) Add summary printout to ocsp utility. The various functions which
5123 convert status values to strings have been renamed to:
5124 OCSP_response_status_str(), OCSP_cert_status_str() and
5125 OCSP_crl_reason_str() and are no longer static. New options
5126 to verify nonce values and to disable verification. OCSP response
5127 printout format cleaned up.
5128 [Steve Henson]
5129
5130 *) Add additional OCSP certificate checks. These are those specified
5131 in RFC2560. This consists of two separate checks: the CA of the
5132 certificate being checked must either be the OCSP signer certificate
5133 or the issuer of the OCSP signer certificate. In the latter case the
5134 OCSP signer certificate must contain the OCSP signing extended key
5135 usage. This check is performed by attempting to match the OCSP
5136 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5137 in the OCSP_CERTID structures of the response.
5138 [Steve Henson]
5139
5140 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5141 and related routines. This uses the standard OpenSSL certificate
5142 verify routines to perform initial checks (just CA validity) and
5143 to obtain the certificate chain. Then additional checks will be
5144 performed on the chain. Currently the root CA is checked to see
5145 if it is explicitly trusted for OCSP signing. This is used to set
5146 a root CA as a global signing root: that is any certificate that
5147 chains to that CA is an acceptable OCSP signing certificate.
5148 [Steve Henson]
5149
5150 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5151 extensions from a separate configuration file.
5152 As when reading extensions from the main configuration file,
5153 the '-extensions ...' option may be used for specifying the
5154 section to use.
5155 [Massimiliano Pala <madwolf@comune.modena.it>]
5156
5157 *) New OCSP utility. Allows OCSP requests to be generated or
5158 read. The request can be sent to a responder and the output
5159 parsed, outputed or printed in text form. Not complete yet:
5160 still needs to check the OCSP response validity.
5161 [Steve Henson]
5162
5163 *) New subcommands for 'openssl ca':
5164 'openssl ca -status <serial>' prints the status of the cert with
5165 the given serial number (according to the index file).
5166 'openssl ca -updatedb' updates the expiry status of certificates
5167 in the index file.
5168 [Massimiliano Pala <madwolf@comune.modena.it>]
5169
5170 *) New '-newreq-nodes' command option to CA.pl. This is like
5171 '-newreq', but calls 'openssl req' with the '-nodes' option
5172 so that the resulting key is not encrypted.
5173 [Damien Miller <djm@mindrot.org>]
5174
5175 *) New configuration for the GNU Hurd.
5176 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5177
5178 *) Initial code to implement OCSP basic response verify. This
5179 is currently incomplete. Currently just finds the signer's
5180 certificate and verifies the signature on the response.
5181 [Steve Henson]
5182
5183 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5184 value of OPENSSLDIR. This is available via the new '-d' option
5185 to 'openssl version', and is also included in 'openssl version -a'.
5186 [Bodo Moeller]
5187
5188 *) Allowing defining memory allocation callbacks that will be given
5189 file name and line number information in additional arguments
5190 (a const char* and an int). The basic functionality remains, as
5191 well as the original possibility to just replace malloc(),
5192 realloc() and free() by functions that do not know about these
5193 additional arguments. To register and find out the current
5194 settings for extended allocation functions, the following
5195 functions are provided:
5196
5197 CRYPTO_set_mem_ex_functions
5198 CRYPTO_set_locked_mem_ex_functions
5199 CRYPTO_get_mem_ex_functions
5200 CRYPTO_get_locked_mem_ex_functions
5201
5202 These work the same way as CRYPTO_set_mem_functions and friends.
5203 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5204 extended allocation function is enabled.
5205 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5206 a conventional allocation function is enabled.
5207 [Richard Levitte, Bodo Moeller]
5208
5209 *) Finish off removing the remaining LHASH function pointer casts.
5210 There should no longer be any prototype-casting required when using
5211 the LHASH abstraction, and any casts that remain are "bugs". See
5212 the callback types and macros at the head of lhash.h for details
5213 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5214 [Geoff Thorpe]
5215
5216 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5217 If /dev/[u]random devices are not available or do not return enough
5218 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5219 be queried.
5220 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5221 /etc/entropy will be queried once each in this sequence, quering stops
5222 when enough entropy was collected without querying more sockets.
5223 [Lutz Jaenicke]
5224
5225 *) Change the Unix RAND_poll() variant to be able to poll several
5226 random devices, as specified by DEVRANDOM, until a sufficient amount
5227 of data has been collected. We spend at most 10 ms on each file
5228 (select timeout) and read in non-blocking mode. DEVRANDOM now
5229 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5230 (previously it was just the string "/dev/urandom"), so on typical
5231 platforms the 10 ms delay will never occur.
5232 Also separate out the Unix variant to its own file, rand_unix.c.
5233 For VMS, there's a currently-empty rand_vms.c.
5234 [Richard Levitte]
5235
5236 *) Move OCSP client related routines to ocsp_cl.c. These
5237 provide utility functions which an application needing
5238 to issue a request to an OCSP responder and analyse the
5239 response will typically need: as opposed to those which an
5240 OCSP responder itself would need which will be added later.
5241
5242 OCSP_request_sign() signs an OCSP request with an API similar
5243 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5244 response. OCSP_response_get1_basic() extracts basic response
5245 from response. OCSP_resp_find_status(): finds and extracts status
5246 information from an OCSP_CERTID structure (which will be created
5247 when the request structure is built). These are built from lower
5248 level functions which work on OCSP_SINGLERESP structures but
5249 wont normally be used unless the application wishes to examine
5250 extensions in the OCSP response for example.
5251
5252 Replace nonce routines with a pair of functions.
5253 OCSP_request_add1_nonce() adds a nonce value and optionally
5254 generates a random value. OCSP_check_nonce() checks the
5255 validity of the nonce in an OCSP response.
5256 [Steve Henson]
5257
5258 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5259 This doesn't copy the supplied OCSP_CERTID and avoids the
5260 need to free up the newly created id. Change return type
5261 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5262 This can then be used to add extensions to the request.
5263 Deleted OCSP_request_new(), since most of its functionality
5264 is now in OCSP_REQUEST_new() (and the case insensitive name
5265 clash) apart from the ability to set the request name which
5266 will be added elsewhere.
5267 [Steve Henson]
5268
5269 *) Update OCSP API. Remove obsolete extensions argument from
5270 various functions. Extensions are now handled using the new
5271 OCSP extension code. New simple OCSP HTTP function which
5272 can be used to send requests and parse the response.
5273 [Steve Henson]
5274
5275 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5276 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5277 uses the special reorder version of SET OF to sort the attributes
5278 and reorder them to match the encoded order. This resolves a long
5279 standing problem: a verify on a PKCS7 structure just after signing
5280 it used to fail because the attribute order did not match the
5281 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5282 it uses the received order. This is necessary to tolerate some broken
5283 software that does not order SET OF. This is handled by encoding
5284 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5285 to produce the required SET OF.
5286 [Steve Henson]
5287
5288 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5289 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5290 files to get correct declarations of the ASN.1 item variables.
5291 [Richard Levitte]
5292
5293 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5294 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5295 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5296 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5297 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5298 ASN1_ITEM and no wrapper functions.
5299 [Steve Henson]
5300
5301 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5302 replace the old function pointer based I/O routines. Change most of
5303 the *_d2i_bio() and *_d2i_fp() functions to use these.
5304 [Steve Henson]
5305
5306 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5307 lines, recognice more "algorithms" that can be deselected, and make
5308 it complain about algorithm deselection that isn't recognised.
5309 [Richard Levitte]
5310
5311 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5312 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5313 to use new functions. Add NO_ASN1_OLD which can be set to remove
5314 some old style ASN1 functions: this can be used to determine if old
5315 code will still work when these eventually go away.
5316 [Steve Henson]
5317
5318 *) New extension functions for OCSP structures, these follow the
5319 same conventions as certificates and CRLs.
5320 [Steve Henson]
5321
5322 *) New function X509V3_add1_i2d(). This automatically encodes and
5323 adds an extension. Its behaviour can be customised with various
5324 flags to append, replace or delete. Various wrappers added for
5325 certifcates and CRLs.
5326 [Steve Henson]
5327
5328 *) Fix to avoid calling the underlying ASN1 print routine when
5329 an extension cannot be parsed. Correct a typo in the
5330 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5331 [Steve Henson]
5332
5333 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5334 entries for variables.
5335 [Steve Henson]
5336
5337 *) Add functionality to apps/openssl.c for detecting locking
5338 problems: As the program is single-threaded, all we have
5339 to do is register a locking callback using an array for
5340 storing which locks are currently held by the program.
5341 [Bodo Moeller]
5342
5343 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5344 SSL_get_ex_data_X509_STORE_idx(), which is used in
5345 ssl_verify_cert_chain() and thus can be called at any time
5346 during TLS/SSL handshakes so that thread-safety is essential.
5347 Unfortunately, the ex_data design is not at all suited
5348 for multi-threaded use, so it probably should be abolished.
5349 [Bodo Moeller]
5350
5351 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5352 [Broadcom, tweaked and integrated by Geoff Thorpe]
5353
5354 *) Move common extension printing code to new function
5355 X509V3_print_extensions(). Reorganise OCSP print routines and
5356 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5357 [Steve Henson]
5358
5359 *) New function X509_signature_print() to remove duplication in some
5360 print routines.
5361 [Steve Henson]
5362
5363 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5364 set (this was treated exactly the same as SET OF previously). This
5365 is used to reorder the STACK representing the structure to match the
5366 encoding. This will be used to get round a problem where a PKCS7
5367 structure which was signed could not be verified because the STACK
5368 order did not reflect the encoded order.
5369 [Steve Henson]
5370
5371 *) Reimplement the OCSP ASN1 module using the new code.
5372 [Steve Henson]
5373
5374 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5375 for its ASN1 operations. The old style function pointers still exist
5376 for now but they will eventually go away.
5377 [Steve Henson]
5378
5379 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5380 completely replaces the old ASN1 functionality with a table driven
5381 encoder and decoder which interprets an ASN1_ITEM structure describing
5382 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5383 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5384 has also been converted to the new form.
5385 [Steve Henson]
5386
5387 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5388 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5389 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5390 for negative moduli.
5391 [Bodo Moeller]
5392
5393 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5394 of not touching the result's sign bit.
5395 [Bodo Moeller]
5396
5397 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5398 set.
5399 [Bodo Moeller]
5400
5401 *) Changed the LHASH code to use prototypes for callbacks, and created
5402 macros to declare and implement thin (optionally static) functions
5403 that provide type-safety and avoid function pointer casting for the
5404 type-specific callbacks.
5405 [Geoff Thorpe]
5406
5407 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5408 RFC 2712.
5409 [Veers Staats <staatsvr@asc.hpc.mil>,
5410 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5411
5412 *) Reformat the FAQ so the different questions and answers can be divided
5413 in sections depending on the subject.
5414 [Richard Levitte]
5415
5416 *) Have the zlib compression code load ZLIB.DLL dynamically under
5417 Windows.
5418 [Richard Levitte]
5419
5420 *) New function BN_mod_sqrt for computing square roots modulo a prime
5421 (using the probabilistic Tonelli-Shanks algorithm unless
5422 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5423 be handled deterministically).
5424 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5425
5426 *) Make BN_mod_inverse faster by explicitly handling small quotients
5427 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5428 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5429 [Bodo Moeller]
5430
5431 *) New function BN_kronecker.
5432 [Bodo Moeller]
5433
5434 *) Fix BN_gcd so that it works on negative inputs; the result is
5435 positive unless both parameters are zero.
5436 Previously something reasonably close to an infinite loop was
5437 possible because numbers could be growing instead of shrinking
5438 in the implementation of Euclid's algorithm.
5439 [Bodo Moeller]
5440
5441 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5442 sign of the number in question.
5443
5444 Fix BN_is_word(a,w) to work correctly for w == 0.
5445
5446 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5447 because its test if the absolute value of 'a' equals 'w'.
5448 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5449 it exists mostly for use in the implementations of BN_is_zero(),
5450 BN_is_one(), and BN_is_word().
5451 [Bodo Moeller]
5452
5453 *) New function BN_swap.
5454 [Bodo Moeller]
5455
5456 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5457 the exponentiation functions are more likely to produce reasonable
5458 results on negative inputs.
5459 [Bodo Moeller]
5460
5461 *) Change BN_mod_mul so that the result is always non-negative.
5462 Previously, it could be negative if one of the factors was negative;
5463 I don't think anyone really wanted that behaviour.
5464 [Bodo Moeller]
5465
5466 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5467 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5468 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5469 and add new functions:
5470
5471 BN_nnmod
5472 BN_mod_sqr
5473 BN_mod_add
5474 BN_mod_add_quick
5475 BN_mod_sub
5476 BN_mod_sub_quick
5477 BN_mod_lshift1
5478 BN_mod_lshift1_quick
5479 BN_mod_lshift
5480 BN_mod_lshift_quick
5481
5482 These functions always generate non-negative results.
5483
5484 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5485 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5486
5487 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5488 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5489 be reduced modulo m.
5490 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5491
5492 #if 0
5493 The following entry accidentily appeared in the CHANGES file
5494 distributed with OpenSSL 0.9.7. The modifications described in
5495 it do *not* apply to OpenSSL 0.9.7.
5496
5497 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5498 was actually never needed) and in BN_mul(). The removal in BN_mul()
5499 required a small change in bn_mul_part_recursive() and the addition
5500 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5501 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5502 bn_sub_words() and bn_add_words() except they take arrays with
5503 differing sizes.
5504 [Richard Levitte]
5505 #endif
5506
5507 *) In 'openssl passwd', verify passwords read from the terminal
5508 unless the '-salt' option is used (which usually means that
5509 verification would just waste user's time since the resulting
5510 hash is going to be compared with some given password hash)
5511 or the new '-noverify' option is used.
5512
5513 This is an incompatible change, but it does not affect
5514 non-interactive use of 'openssl passwd' (passwords on the command
5515 line, '-stdin' option, '-in ...' option) and thus should not
5516 cause any problems.
5517 [Bodo Moeller]
5518
5519 *) Remove all references to RSAref, since there's no more need for it.
5520 [Richard Levitte]
5521
5522 *) Make DSO load along a path given through an environment variable
5523 (SHLIB_PATH) with shl_load().
5524 [Richard Levitte]
5525
5526 *) Constify the ENGINE code as a result of BIGNUM constification.
5527 Also constify the RSA code and most things related to it. In a
5528 few places, most notable in the depth of the ASN.1 code, ugly
5529 casts back to non-const were required (to be solved at a later
5530 time)
5531 [Richard Levitte]
5532
5533 *) Make it so the openssl application has all engines loaded by default.
5534 [Richard Levitte]
5535
5536 *) Constify the BIGNUM routines a little more.
5537 [Richard Levitte]
5538
5539 *) Add the following functions:
5540
5541 ENGINE_load_cswift()
5542 ENGINE_load_chil()
5543 ENGINE_load_atalla()
5544 ENGINE_load_nuron()
5545 ENGINE_load_builtin_engines()
5546
5547 That way, an application can itself choose if external engines that
5548 are built-in in OpenSSL shall ever be used or not. The benefit is
5549 that applications won't have to be linked with libdl or other dso
5550 libraries unless it's really needed.
5551
5552 Changed 'openssl engine' to load all engines on demand.
5553 Changed the engine header files to avoid the duplication of some
5554 declarations (they differed!).
5555 [Richard Levitte]
5556
5557 *) 'openssl engine' can now list capabilities.
5558 [Richard Levitte]
5559
5560 *) Better error reporting in 'openssl engine'.
5561 [Richard Levitte]
5562
5563 *) Never call load_dh_param(NULL) in s_server.
5564 [Bodo Moeller]
5565
5566 *) Add engine application. It can currently list engines by name and
5567 identity, and test if they are actually available.
5568 [Richard Levitte]
5569
5570 *) Improve RPM specification file by forcing symbolic linking and making
5571 sure the installed documentation is also owned by root.root.
5572 [Damien Miller <djm@mindrot.org>]
5573
5574 *) Give the OpenSSL applications more possibilities to make use of
5575 keys (public as well as private) handled by engines.
5576 [Richard Levitte]
5577
5578 *) Add OCSP code that comes from CertCo.
5579 [Richard Levitte]
5580
5581 *) Add VMS support for the Rijndael code.
5582 [Richard Levitte]
5583
5584 *) Added untested support for Nuron crypto accelerator.
5585 [Ben Laurie]
5586
5587 *) Add support for external cryptographic devices. This code was
5588 previously distributed separately as the "engine" branch.
5589 [Geoff Thorpe, Richard Levitte]
5590
5591 *) Rework the filename-translation in the DSO code. It is now possible to
5592 have far greater control over how a "name" is turned into a filename
5593 depending on the operating environment and any oddities about the
5594 different shared library filenames on each system.
5595 [Geoff Thorpe]
5596
5597 *) Support threads on FreeBSD-elf in Configure.
5598 [Richard Levitte]
5599
5600 *) Fix for SHA1 assembly problem with MASM: it produces
5601 warnings about corrupt line number information when assembling
5602 with debugging information. This is caused by the overlapping
5603 of two sections.
5604 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5605
5606 *) NCONF changes.
5607 NCONF_get_number() has no error checking at all. As a replacement,
5608 NCONF_get_number_e() is defined (_e for "error checking") and is
5609 promoted strongly. The old NCONF_get_number is kept around for
5610 binary backward compatibility.
5611 Make it possible for methods to load from something other than a BIO,
5612 by providing a function pointer that is given a name instead of a BIO.
5613 For example, this could be used to load configuration data from an
5614 LDAP server.
5615 [Richard Levitte]
5616
5617 *) Fix for non blocking accept BIOs. Added new I/O special reason
5618 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5619 with non blocking I/O was not possible because no retry code was
5620 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5621 this case.
5622 [Steve Henson]
5623
5624 *) Added the beginnings of Rijndael support.
5625 [Ben Laurie]
5626
5627 *) Fix for bug in DirectoryString mask setting. Add support for
5628 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5629 to allow certificate printing to more controllable, additional
5630 'certopt' option to 'x509' to allow new printing options to be
5631 set.
5632 [Steve Henson]
5633
5634 *) Clean old EAY MD5 hack from e_os.h.
5635 [Richard Levitte]
5636
5637 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5638
5639 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5640 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5641 [Joe Orton, Steve Henson]
5642
5643 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5644
5645 *) Fix additional bug revealed by the NISCC test suite:
5646
5647 Stop bug triggering large recursion when presented with
5648 certain ASN.1 tags (CVE-2003-0851)
5649 [Steve Henson]
5650
5651 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5652
5653 *) Fix various bugs revealed by running the NISCC test suite:
5654
5655 Stop out of bounds reads in the ASN1 code when presented with
5656 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5657
5658 If verify callback ignores invalid public key errors don't try to check
5659 certificate signature with the NULL public key.
5660
5661 [Steve Henson]
5662
5663 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5664 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5665 specifications.
5666 [Steve Henson]
5667
5668 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5669 extra data after the compression methods not only for TLS 1.0
5670 but also for SSL 3.0 (as required by the specification).
5671 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5672
5673 *) Change X509_certificate_type() to mark the key as exported/exportable
5674 when it's 512 *bits* long, not 512 bytes.
5675 [Richard Levitte]
5676
5677 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5678
5679 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5680 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5681 a protocol version number mismatch like a decryption error
5682 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5683 [Bodo Moeller]
5684
5685 *) Turn on RSA blinding by default in the default implementation
5686 to avoid a timing attack. Applications that don't want it can call
5687 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5688 They would be ill-advised to do so in most cases.
5689 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5690
5691 *) Change RSA blinding code so that it works when the PRNG is not
5692 seeded (in this case, the secret RSA exponent is abused as
5693 an unpredictable seed -- if it is not unpredictable, there
5694 is no point in blinding anyway). Make RSA blinding thread-safe
5695 by remembering the creator's thread ID in rsa->blinding and
5696 having all other threads use local one-time blinding factors
5697 (this requires more computation than sharing rsa->blinding, but
5698 avoids excessive locking; and if an RSA object is not shared
5699 between threads, blinding will still be very fast).
5700 [Bodo Moeller]
5701
5702 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5703
5704 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5705 via timing by performing a MAC computation even if incorrrect
5706 block cipher padding has been found. This is a countermeasure
5707 against active attacks where the attacker has to distinguish
5708 between bad padding and a MAC verification error. (CVE-2003-0078)
5709
5710 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5711 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5712 Martin Vuagnoux (EPFL, Ilion)]
5713
5714 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5715
5716 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5717 memory from it's contents. This is done with a counter that will
5718 place alternating values in each byte. This can be used to solve
5719 two issues: 1) the removal of calls to memset() by highly optimizing
5720 compilers, and 2) cleansing with other values than 0, since those can
5721 be read through on certain media, for example a swap space on disk.
5722 [Geoff Thorpe]
5723
5724 *) Bugfix: client side session caching did not work with external caching,
5725 because the session->cipher setting was not restored when reloading
5726 from the external cache. This problem was masked, when
5727 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5728 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5729 [Lutz Jaenicke]
5730
5731 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5732 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5733 [Zeev Lieber <zeev-l@yahoo.com>]
5734
5735 *) Undo an undocumented change introduced in 0.9.6e which caused
5736 repeated calls to OpenSSL_add_all_ciphers() and
5737 OpenSSL_add_all_digests() to be ignored, even after calling
5738 EVP_cleanup().
5739 [Richard Levitte]
5740
5741 *) Change the default configuration reader to deal with last line not
5742 being properly terminated.
5743 [Richard Levitte]
5744
5745 *) Change X509_NAME_cmp() so it applies the special rules on handling
5746 DN values that are of type PrintableString, as well as RDNs of type
5747 emailAddress where the value has the type ia5String.
5748 [stefank@valicert.com via Richard Levitte]
5749
5750 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5751 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5752 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5753 the bitwise-OR of the two for use by the majority of applications
5754 wanting this behaviour, and update the docs. The documented
5755 behaviour and actual behaviour were inconsistent and had been
5756 changing anyway, so this is more a bug-fix than a behavioural
5757 change.
5758 [Geoff Thorpe, diagnosed by Nadav Har'El]
5759
5760 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5761 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5762 [Bodo Moeller]
5763
5764 *) Fix initialization code race conditions in
5765 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5766 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5767 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5768 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5769 ssl2_get_cipher_by_char(),
5770 ssl3_get_cipher_by_char().
5771 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5772
5773 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5774 the cached sessions are flushed, as the remove_cb() might use ex_data
5775 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5776 (see [openssl.org #212]).
5777 [Geoff Thorpe, Lutz Jaenicke]
5778
5779 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5780 length, instead of the encoding length to d2i_ASN1_OBJECT.
5781 [Steve Henson]
5782
5783 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5784
5785 *) [In 0.9.6g-engine release:]
5786 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5787 [Lynn Gazis <lgazis@rainbow.com>]
5788
5789 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5790
5791 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5792 and get fix the header length calculation.
5793 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5794 Alon Kantor <alonk@checkpoint.com> (and others),
5795 Steve Henson]
5796
5797 *) Use proper error handling instead of 'assertions' in buffer
5798 overflow checks added in 0.9.6e. This prevents DoS (the
5799 assertions could call abort()).
5800 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5801
5802 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5803
5804 *) Add various sanity checks to asn1_get_length() to reject
5805 the ASN1 length bytes if they exceed sizeof(long), will appear
5806 negative or the content length exceeds the length of the
5807 supplied buffer.
5808 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5809
5810 *) Fix cipher selection routines: ciphers without encryption had no flags
5811 for the cipher strength set and where therefore not handled correctly
5812 by the selection routines (PR #130).
5813 [Lutz Jaenicke]
5814
5815 *) Fix EVP_dsa_sha macro.
5816 [Nils Larsch]
5817
5818 *) New option
5819 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5820 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5821 that was added in OpenSSL 0.9.6d.
5822
5823 As the countermeasure turned out to be incompatible with some
5824 broken SSL implementations, the new option is part of SSL_OP_ALL.
5825 SSL_OP_ALL is usually employed when compatibility with weird SSL
5826 implementations is desired (e.g. '-bugs' option to 's_client' and
5827 's_server'), so the new option is automatically set in many
5828 applications.
5829 [Bodo Moeller]
5830
5831 *) Changes in security patch:
5832
5833 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5834 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5835 Air Force Materiel Command, USAF, under agreement number
5836 F30602-01-2-0537.
5837
5838 *) Add various sanity checks to asn1_get_length() to reject
5839 the ASN1 length bytes if they exceed sizeof(long), will appear
5840 negative or the content length exceeds the length of the
5841 supplied buffer. (CVE-2002-0659)
5842 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5843
5844 *) Assertions for various potential buffer overflows, not known to
5845 happen in practice.
5846 [Ben Laurie (CHATS)]
5847
5848 *) Various temporary buffers to hold ASCII versions of integers were
5849 too small for 64 bit platforms. (CVE-2002-0655)
5850 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5851
5852 *) Remote buffer overflow in SSL3 protocol - an attacker could
5853 supply an oversized session ID to a client. (CVE-2002-0656)
5854 [Ben Laurie (CHATS)]
5855
5856 *) Remote buffer overflow in SSL2 protocol - an attacker could
5857 supply an oversized client master key. (CVE-2002-0656)
5858 [Ben Laurie (CHATS)]
5859
5860 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5861
5862 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5863 encoded as NULL) with id-dsa-with-sha1.
5864 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5865
5866 *) Check various X509_...() return values in apps/req.c.
5867 [Nils Larsch <nla@trustcenter.de>]
5868
5869 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5870 an end-of-file condition would erronously be flagged, when the CRLF
5871 was just at the end of a processed block. The bug was discovered when
5872 processing data through a buffering memory BIO handing the data to a
5873 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5874 <ptsekov@syntrex.com> and Nedelcho Stanev.
5875 [Lutz Jaenicke]
5876
5877 *) Implement a countermeasure against a vulnerability recently found
5878 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5879 before application data chunks to avoid the use of known IVs
5880 with data potentially chosen by the attacker.
5881 [Bodo Moeller]
5882
5883 *) Fix length checks in ssl3_get_client_hello().
5884 [Bodo Moeller]
5885
5886 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5887 to prevent ssl3_read_internal() from incorrectly assuming that
5888 ssl3_read_bytes() found application data while handshake
5889 processing was enabled when in fact s->s3->in_read_app_data was
5890 merely automatically cleared during the initial handshake.
5891 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5892
5893 *) Fix object definitions for Private and Enterprise: they were not
5894 recognized in their shortname (=lowercase) representation. Extend
5895 obj_dat.pl to issue an error when using undefined keywords instead
5896 of silently ignoring the problem (Svenning Sorensen
5897 <sss@sss.dnsalias.net>).
5898 [Lutz Jaenicke]
5899
5900 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5901 generators, i.e. generators other than 2 and 5. (Previously, the
5902 code did not properly initialise the 'add' and 'rem' values to
5903 BN_generate_prime().)
5904
5905 In the new general case, we do not insist that 'generator' is
5906 actually a primitive root: This requirement is rather pointless;
5907 a generator of the order-q subgroup is just as good, if not
5908 better.
5909 [Bodo Moeller]
5910
5911 *) Map new X509 verification errors to alerts. Discovered and submitted by
5912 Tom Wu <tom@arcot.com>.
5913 [Lutz Jaenicke]
5914
5915 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5916 returning non-zero before the data has been completely received
5917 when using non-blocking I/O.
5918 [Bodo Moeller; problem pointed out by John Hughes]
5919
5920 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5921 [Ben Laurie, Lutz Jaenicke]
5922
5923 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5924 Yoram Zahavi <YoramZ@gilian.com>).
5925 [Lutz Jaenicke]
5926
5927 *) Add information about CygWin 1.3 and on, and preserve proper
5928 configuration for the versions before that.
5929 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5930
5931 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5932 check whether we deal with a copy of a session and do not delete from
5933 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5934 <izhar@checkpoint.com>.
5935 [Lutz Jaenicke]
5936
5937 *) Do not store session data into the internal session cache, if it
5938 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5939 flag is set). Proposed by Aslam <aslam@funk.com>.
5940 [Lutz Jaenicke]
5941
5942 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5943 value is 0.
5944 [Richard Levitte]
5945
5946 *) [In 0.9.6d-engine release:]
5947 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5948 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5949
5950 *) Add the configuration target linux-s390x.
5951 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5952
5953 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5954 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5955 variable as an indication that a ClientHello message has been
5956 received. As the flag value will be lost between multiple
5957 invocations of ssl3_accept when using non-blocking I/O, the
5958 function may not be aware that a handshake has actually taken
5959 place, thus preventing a new session from being added to the
5960 session cache.
5961
5962 To avoid this problem, we now set s->new_session to 2 instead of
5963 using a local variable.
5964 [Lutz Jaenicke, Bodo Moeller]
5965
5966 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5967 if the SSL_R_LENGTH_MISMATCH error is detected.
5968 [Geoff Thorpe, Bodo Moeller]
5969
5970 *) New 'shared_ldflag' column in Configure platform table.
5971 [Richard Levitte]
5972
5973 *) Fix EVP_CIPHER_mode macro.
5974 ["Dan S. Camper" <dan@bti.net>]
5975
5976 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5977 type, we must throw them away by setting rr->length to 0.
5978 [D P Chang <dpc@qualys.com>]
5979
5980 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5981
5982 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5983 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5984 worked incorrectly for those cases where range = 10..._2 and
5985 3*range is two bits longer than range.)
5986 [Bodo Moeller]
5987
5988 *) Only add signing time to PKCS7 structures if it is not already
5989 present.
5990 [Steve Henson]
5991
5992 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5993 OBJ_ld_ce should be OBJ_id_ce.
5994 Also some ip-pda OIDs in crypto/objects/objects.txt were
5995 incorrect (cf. RFC 3039).
5996 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5997
5998 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5999 returns early because it has nothing to do.
6000 [Andy Schneider <andy.schneider@bjss.co.uk>]
6001
6002 *) [In 0.9.6c-engine release:]
6003 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6004 [Andy Schneider <andy.schneider@bjss.co.uk>]
6005
6006 *) [In 0.9.6c-engine release:]
6007 Add support for Cryptographic Appliance's keyserver technology.
6008 (Use engine 'keyclient')
6009 [Cryptographic Appliances and Geoff Thorpe]
6010
6011 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6012 is called via tools/c89.sh because arguments have to be
6013 rearranged (all '-L' options must appear before the first object
6014 modules).
6015 [Richard Shapiro <rshapiro@abinitio.com>]
6016
6017 *) [In 0.9.6c-engine release:]
6018 Add support for Broadcom crypto accelerator cards, backported
6019 from 0.9.7.
6020 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6021
6022 *) [In 0.9.6c-engine release:]
6023 Add support for SureWare crypto accelerator cards from
6024 Baltimore Technologies. (Use engine 'sureware')
6025 [Baltimore Technologies and Mark Cox]
6026
6027 *) [In 0.9.6c-engine release:]
6028 Add support for crypto accelerator cards from Accelerated
6029 Encryption Processing, www.aep.ie. (Use engine 'aep')
6030 [AEP Inc. and Mark Cox]
6031
6032 *) Add a configuration entry for gcc on UnixWare.
6033 [Gary Benson <gbenson@redhat.com>]
6034
6035 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6036 messages are stored in a single piece (fixed-length part and
6037 variable-length part combined) and fix various bugs found on the way.
6038 [Bodo Moeller]
6039
6040 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6041 instead. BIO_gethostbyname() does not know what timeouts are
6042 appropriate, so entries would stay in cache even when they have
6043 become invalid.
6044 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6045
6046 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6047 faced with a pathologically small ClientHello fragment that does
6048 not contain client_version: Instead of aborting with an error,
6049 simply choose the highest available protocol version (i.e.,
6050 TLS 1.0 unless it is disabled). In practice, ClientHello
6051 messages are never sent like this, but this change gives us
6052 strictly correct behaviour at least for TLS.
6053 [Bodo Moeller]
6054
6055 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6056 never resets s->method to s->ctx->method when called from within
6057 one of the SSL handshake functions.
6058 [Bodo Moeller; problem pointed out by Niko Baric]
6059
6060 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6061 (sent using the client's version number) if client_version is
6062 smaller than the protocol version in use. Also change
6063 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6064 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6065 the client will at least see that alert.
6066 [Bodo Moeller]
6067
6068 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6069 correctly.
6070 [Bodo Moeller]
6071
6072 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6073 client receives HelloRequest while in a handshake.
6074 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6075
6076 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6077 should end in 'break', not 'goto end' which circuments various
6078 cleanups done in state SSL_ST_OK. But session related stuff
6079 must be disabled for SSL_ST_OK in the case that we just sent a
6080 HelloRequest.
6081
6082 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6083 before just sending a HelloRequest.
6084 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6085
6086 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6087 reveal whether illegal block cipher padding was found or a MAC
6088 verification error occured. (Neither SSLerr() codes nor alerts
6089 are directly visible to potential attackers, but the information
6090 may leak via logfiles.)
6091
6092 Similar changes are not required for the SSL 2.0 implementation
6093 because the number of padding bytes is sent in clear for SSL 2.0,
6094 and the extra bytes are just ignored. However ssl/s2_pkt.c
6095 failed to verify that the purported number of padding bytes is in
6096 the legal range.
6097 [Bodo Moeller]
6098
6099 *) Add OpenUNIX-8 support including shared libraries
6100 (Boyd Lynn Gerber <gerberb@zenez.com>).
6101 [Lutz Jaenicke]
6102
6103 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6104 'wristwatch attack' using huge encoding parameters (cf.
6105 James H. Manger's CRYPTO 2001 paper). Note that the
6106 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6107 encoding parameters and hence was not vulnerable.
6108 [Bodo Moeller]
6109
6110 *) BN_sqr() bug fix.
6111 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6112
6113 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6114 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6115 followed by modular reduction.
6116 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6117
6118 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6119 equivalent based on BN_pseudo_rand() instead of BN_rand().
6120 [Bodo Moeller]
6121
6122 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6123 This function was broken, as the check for a new client hello message
6124 to handle SGC did not allow these large messages.
6125 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6126 [Lutz Jaenicke]
6127
6128 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6129 [Lutz Jaenicke]
6130
6131 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6132 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6133 [Lutz Jaenicke]
6134
6135 *) Rework the configuration and shared library support for Tru64 Unix.
6136 The configuration part makes use of modern compiler features and
6137 still retains old compiler behavior for those that run older versions
6138 of the OS. The shared library support part includes a variant that
6139 uses the RPATH feature, and is available through the special
6140 configuration target "alpha-cc-rpath", which will never be selected
6141 automatically.
6142 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6143
6144 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6145 with the same message size as in ssl3_get_certificate_request().
6146 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6147 messages might inadvertently be reject as too long.
6148 [Petr Lampa <lampa@fee.vutbr.cz>]
6149
6150 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6151 [Andy Polyakov]
6152
6153 *) Modified SSL library such that the verify_callback that has been set
6154 specificly for an SSL object with SSL_set_verify() is actually being
6155 used. Before the change, a verify_callback set with this function was
6156 ignored and the verify_callback() set in the SSL_CTX at the time of
6157 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6158 to allow the necessary settings.
6159 [Lutz Jaenicke]
6160
6161 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6162 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6163 done automatically (in contradiction to the requirements of the C
6164 standard). This made problems when used from OpenSSH.
6165 [Lutz Jaenicke]
6166
6167 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6168 dh->length and always used
6169
6170 BN_rand_range(priv_key, dh->p).
6171
6172 BN_rand_range() is not necessary for Diffie-Hellman, and this
6173 specific range makes Diffie-Hellman unnecessarily inefficient if
6174 dh->length (recommended exponent length) is much smaller than the
6175 length of dh->p. We could use BN_rand_range() if the order of
6176 the subgroup was stored in the DH structure, but we only have
6177 dh->length.
6178
6179 So switch back to
6180
6181 BN_rand(priv_key, l, ...)
6182
6183 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6184 otherwise.
6185 [Bodo Moeller]
6186
6187 *) In
6188
6189 RSA_eay_public_encrypt
6190 RSA_eay_private_decrypt
6191 RSA_eay_private_encrypt (signing)
6192 RSA_eay_public_decrypt (signature verification)
6193
6194 (default implementations for RSA_public_encrypt,
6195 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6196 always reject numbers >= n.
6197 [Bodo Moeller]
6198
6199 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6200 to synchronize access to 'locking_thread'. This is necessary on
6201 systems where access to 'locking_thread' (an 'unsigned long'
6202 variable) is not atomic.
6203 [Bodo Moeller]
6204
6205 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6206 *before* setting the 'crypto_lock_rand' flag. The previous code had
6207 a race condition if 0 is a valid thread ID.
6208 [Travis Vitek <vitek@roguewave.com>]
6209
6210 *) Add support for shared libraries under Irix.
6211 [Albert Chin-A-Young <china@thewrittenword.com>]
6212
6213 *) Add configuration option to build on Linux on both big-endian and
6214 little-endian MIPS.
6215 [Ralf Baechle <ralf@uni-koblenz.de>]
6216
6217 *) Add the possibility to create shared libraries on HP-UX.
6218 [Richard Levitte]
6219
6220 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6221
6222 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6223 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6224 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6225 PRNG state recovery was possible based on the output of
6226 one PRNG request appropriately sized to gain knowledge on
6227 'md' followed by enough consecutive 1-byte PRNG requests
6228 to traverse all of 'state'.
6229
6230 1. When updating 'md_local' (the current thread's copy of 'md')
6231 during PRNG output generation, hash all of the previous
6232 'md_local' value, not just the half used for PRNG output.
6233
6234 2. Make the number of bytes from 'state' included into the hash
6235 independent from the number of PRNG bytes requested.
6236
6237 The first measure alone would be sufficient to avoid
6238 Markku-Juhani's attack. (Actually it had never occurred
6239 to me that the half of 'md_local' used for chaining was the
6240 half from which PRNG output bytes were taken -- I had always
6241 assumed that the secret half would be used.) The second
6242 measure makes sure that additional data from 'state' is never
6243 mixed into 'md_local' in small portions; this heuristically
6244 further strengthens the PRNG.
6245 [Bodo Moeller]
6246
6247 *) Fix crypto/bn/asm/mips3.s.
6248 [Andy Polyakov]
6249
6250 *) When only the key is given to "enc", the IV is undefined. Print out
6251 an error message in this case.
6252 [Lutz Jaenicke]
6253
6254 *) Handle special case when X509_NAME is empty in X509 printing routines.
6255 [Steve Henson]
6256
6257 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6258 positive and less than q.
6259 [Bodo Moeller]
6260
6261 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6262 used: it isn't thread safe and the add_lock_callback should handle
6263 that itself.
6264 [Paul Rose <Paul.Rose@bridge.com>]
6265
6266 *) Verify that incoming data obeys the block size in
6267 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6268 [Bodo Moeller]
6269
6270 *) Fix OAEP check.
6271 [Ulf Möller, Bodo Möller]
6272
6273 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6274 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6275 when fixing the server behaviour for backwards-compatible 'client
6276 hello' messages. (Note that the attack is impractical against
6277 SSL 3.0 and TLS 1.0 anyway because length and version checking
6278 means that the probability of guessing a valid ciphertext is
6279 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6280 paper.)
6281
6282 Before 0.9.5, the countermeasure (hide the error by generating a
6283 random 'decryption result') did not work properly because
6284 ERR_clear_error() was missing, meaning that SSL_get_error() would
6285 detect the supposedly ignored error.
6286
6287 Both problems are now fixed.
6288 [Bodo Moeller]
6289
6290 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6291 (previously it was 1024).
6292 [Bodo Moeller]
6293
6294 *) Fix for compatibility mode trust settings: ignore trust settings
6295 unless some valid trust or reject settings are present.
6296 [Steve Henson]
6297
6298 *) Fix for blowfish EVP: its a variable length cipher.
6299 [Steve Henson]
6300
6301 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6302 parameters in DSA public key structures and return an error in the
6303 DSA routines if parameters are absent.
6304 [Steve Henson]
6305
6306 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6307 in the current directory if neither $RANDFILE nor $HOME was set.
6308 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6309 caused some confusion to Windows users who haven't defined $HOME.
6310 Thus RAND_file_name() is changed again: e_os.h can define a
6311 DEFAULT_HOME, which will be used if $HOME is not set.
6312 For Windows, we use "C:"; on other platforms, we still require
6313 environment variables.
6314
6315 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6316 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6317 having multiple threads call RAND_poll() concurrently.
6318 [Bodo Moeller]
6319
6320 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6321 combination of a flag and a thread ID variable.
6322 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6323 flag), *other* threads can enter ssleay_add_bytes without obeying
6324 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6325 that they do not hold after the first thread unsets add_do_not_lock).
6326 [Bodo Moeller]
6327
6328 *) Change bctest again: '-x' expressions are not available in all
6329 versions of 'test'.
6330 [Bodo Moeller]
6331
6332 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6333
6334 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6335 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6336
6337 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6338 the default extension for executables, if any. Also, make the perl
6339 scripts that use symlink() to test if it really exists and use "cp"
6340 if it doesn't. All this made OpenSSL compilable and installable in
6341 CygWin.
6342 [Richard Levitte]
6343
6344 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6345 If SEQUENCE is length is indefinite just set c->slen to the total
6346 amount of data available.
6347 [Steve Henson, reported by shige@FreeBSD.org]
6348 [This change does not apply to 0.9.7.]
6349
6350 *) Change bctest to avoid here-documents inside command substitution
6351 (workaround for FreeBSD /bin/sh bug).
6352 For compatibility with Ultrix, avoid shell functions (introduced
6353 in the bctest version that searches along $PATH).
6354 [Bodo Moeller]
6355
6356 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6357 with des_encrypt() defined on some operating systems, like Solaris
6358 and UnixWare.
6359 [Richard Levitte]
6360
6361 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6362 On the Importance of Eliminating Errors in Cryptographic
6363 Computations, J. Cryptology 14 (2001) 2, 101-119,
6364 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6365 [Ulf Moeller]
6366
6367 *) MIPS assembler BIGNUM division bug fix.
6368 [Andy Polyakov]
6369
6370 *) Disabled incorrect Alpha assembler code.
6371 [Richard Levitte]
6372
6373 *) Fix PKCS#7 decode routines so they correctly update the length
6374 after reading an EOC for the EXPLICIT tag.
6375 [Steve Henson]
6376 [This change does not apply to 0.9.7.]
6377
6378 *) Fix bug in PKCS#12 key generation routines. This was triggered
6379 if a 3DES key was generated with a 0 initial byte. Include
6380 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6381 (but broken) behaviour.
6382 [Steve Henson]
6383
6384 *) Enhance bctest to search for a working bc along $PATH and print
6385 it when found.
6386 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6387
6388 *) Fix memory leaks in err.c: free err_data string if necessary;
6389 don't write to the wrong index in ERR_set_error_data.
6390 [Bodo Moeller]
6391
6392 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6393 did not exist.
6394 [Bodo Moeller]
6395
6396 *) Replace rdtsc with _emit statements for VC++ version 5.
6397 [Jeremy Cooper <jeremy@baymoo.org>]
6398
6399 *) Make it possible to reuse SSLv2 sessions.
6400 [Richard Levitte]
6401
6402 *) In copy_email() check for >= 0 as a return value for
6403 X509_NAME_get_index_by_NID() since 0 is a valid index.
6404 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6405
6406 *) Avoid coredump with unsupported or invalid public keys by checking if
6407 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6408 PKCS7_verify() fails with non detached data.
6409 [Steve Henson]
6410
6411 *) Don't use getenv in library functions when run as setuid/setgid.
6412 New function OPENSSL_issetugid().
6413 [Ulf Moeller]
6414
6415 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6416 due to incorrect handling of multi-threading:
6417
6418 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6419
6420 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6421
6422 3. Count how many times MemCheck_off() has been called so that
6423 nested use can be treated correctly. This also avoids
6424 inband-signalling in the previous code (which relied on the
6425 assumption that thread ID 0 is impossible).
6426 [Bodo Moeller]
6427
6428 *) Add "-rand" option also to s_client and s_server.
6429 [Lutz Jaenicke]
6430
6431 *) Fix CPU detection on Irix 6.x.
6432 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6433 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6434
6435 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6436 was empty.
6437 [Steve Henson]
6438 [This change does not apply to 0.9.7.]
6439
6440 *) Use the cached encoding of an X509_NAME structure rather than
6441 copying it. This is apparently the reason for the libsafe "errors"
6442 but the code is actually correct.
6443 [Steve Henson]
6444
6445 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6446 Bleichenbacher's DSA attack.
6447 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6448 to be set and top=0 forces the highest bit to be set; top=-1 is new
6449 and leaves the highest bit random.
6450 [Ulf Moeller, Bodo Moeller]
6451
6452 *) In the NCONF_...-based implementations for CONF_... queries
6453 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6454 a temporary CONF structure with the data component set to NULL
6455 (which gives segmentation faults in lh_retrieve).
6456 Instead, use NULL for the CONF pointer in CONF_get_string and
6457 CONF_get_number (which may use environment variables) and directly
6458 return NULL from CONF_get_section.
6459 [Bodo Moeller]
6460
6461 *) Fix potential buffer overrun for EBCDIC.
6462 [Ulf Moeller]
6463
6464 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6465 keyUsage if basicConstraints absent for a CA.
6466 [Steve Henson]
6467
6468 *) Make SMIME_write_PKCS7() write mail header values with a format that
6469 is more generally accepted (no spaces before the semicolon), since
6470 some programs can't parse those values properly otherwise. Also make
6471 sure BIO's that break lines after each write do not create invalid
6472 headers.
6473 [Richard Levitte]
6474
6475 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6476 macros previously used would not encode an empty SEQUENCE OF
6477 and break the signature.
6478 [Steve Henson]
6479 [This change does not apply to 0.9.7.]
6480
6481 *) Zero the premaster secret after deriving the master secret in
6482 DH ciphersuites.
6483 [Steve Henson]
6484
6485 *) Add some EVP_add_digest_alias registrations (as found in
6486 OpenSSL_add_all_digests()) to SSL_library_init()
6487 aka OpenSSL_add_ssl_algorithms(). This provides improved
6488 compatibility with peers using X.509 certificates
6489 with unconventional AlgorithmIdentifier OIDs.
6490 [Bodo Moeller]
6491
6492 *) Fix for Irix with NO_ASM.
6493 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6494
6495 *) ./config script fixes.
6496 [Ulf Moeller, Richard Levitte]
6497
6498 *) Fix 'openssl passwd -1'.
6499 [Bodo Moeller]
6500
6501 *) Change PKCS12_key_gen_asc() so it can cope with non null
6502 terminated strings whose length is passed in the passlen
6503 parameter, for example from PEM callbacks. This was done
6504 by adding an extra length parameter to asc2uni().
6505 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6506
6507 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6508 call failed, free the DSA structure.
6509 [Bodo Moeller]
6510
6511 *) Fix to uni2asc() to cope with zero length Unicode strings.
6512 These are present in some PKCS#12 files.
6513 [Steve Henson]
6514
6515 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6516 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6517 when writing a 32767 byte record.
6518 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6519
6520 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6521 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6522
6523 (RSA objects have a reference count access to which is protected
6524 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6525 so they are meant to be shared between threads.)
6526 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6527 "Reddie, Steven" <Steven.Reddie@ca.com>]
6528
6529 *) Fix a deadlock in CRYPTO_mem_leaks().
6530 [Bodo Moeller]
6531
6532 *) Use better test patterns in bntest.
6533 [Ulf Möller]
6534
6535 *) rand_win.c fix for Borland C.
6536 [Ulf Möller]
6537
6538 *) BN_rshift bugfix for n == 0.
6539 [Bodo Moeller]
6540
6541 *) Add a 'bctest' script that checks for some known 'bc' bugs
6542 so that 'make test' does not abort just because 'bc' is broken.
6543 [Bodo Moeller]
6544
6545 *) Store verify_result within SSL_SESSION also for client side to
6546 avoid potential security hole. (Re-used sessions on the client side
6547 always resulted in verify_result==X509_V_OK, not using the original
6548 result of the server certificate verification.)
6549 [Lutz Jaenicke]
6550
6551 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6552 SSL3_RT_APPLICATION_DATA, return 0.
6553 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6554 [Bodo Moeller]
6555
6556 *) Fix SSL_peek:
6557 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6558 releases, have been re-implemented by renaming the previous
6559 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6560 and ssl3_read_internal, respectively, and adding 'peek' parameters
6561 to them. The new ssl[23]_{read,peek} functions are calls to
6562 ssl[23]_read_internal with the 'peek' flag set appropriately.
6563 A 'peek' parameter has also been added to ssl3_read_bytes, which
6564 does the actual work for ssl3_read_internal.
6565 [Bodo Moeller]
6566
6567 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6568 the method-specific "init()" handler. Also clean up ex_data after
6569 calling the method-specific "finish()" handler. Previously, this was
6570 happening the other way round.
6571 [Geoff Thorpe]
6572
6573 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6574 The previous value, 12, was not always sufficient for BN_mod_exp().
6575 [Bodo Moeller]
6576
6577 *) Make sure that shared libraries get the internal name engine with
6578 the full version number and not just 0. This should mark the
6579 shared libraries as not backward compatible. Of course, this should
6580 be changed again when we can guarantee backward binary compatibility.
6581 [Richard Levitte]
6582
6583 *) Fix typo in get_cert_by_subject() in by_dir.c
6584 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6585
6586 *) Rework the system to generate shared libraries:
6587
6588 - Make note of the expected extension for the shared libraries and
6589 if there is a need for symbolic links from for example libcrypto.so.0
6590 to libcrypto.so.0.9.7. There is extended info in Configure for
6591 that.
6592
6593 - Make as few rebuilds of the shared libraries as possible.
6594
6595 - Still avoid linking the OpenSSL programs with the shared libraries.
6596
6597 - When installing, install the shared libraries separately from the
6598 static ones.
6599 [Richard Levitte]
6600
6601 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6602
6603 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6604 and not in SSL_clear because the latter is also used by the
6605 accept/connect functions; previously, the settings made by
6606 SSL_set_read_ahead would be lost during the handshake.
6607 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6608
6609 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6610 Previously, it would create entries for disableed algorithms no
6611 matter what.
6612 [Richard Levitte]
6613
6614 *) Added several new manual pages for SSL_* function.
6615 [Lutz Jaenicke]
6616
6617 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6618
6619 *) In ssl23_get_client_hello, generate an error message when faced
6620 with an initial SSL 3.0/TLS record that is too small to contain the
6621 first two bytes of the ClientHello message, i.e. client_version.
6622 (Note that this is a pathologic case that probably has never happened
6623 in real life.) The previous approach was to use the version number
6624 from the record header as a substitute; but our protocol choice
6625 should not depend on that one because it is not authenticated
6626 by the Finished messages.
6627 [Bodo Moeller]
6628
6629 *) More robust randomness gathering functions for Windows.
6630 [Jeffrey Altman <jaltman@columbia.edu>]
6631
6632 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6633 not set then we don't setup the error code for issuer check errors
6634 to avoid possibly overwriting other errors which the callback does
6635 handle. If an application does set the flag then we assume it knows
6636 what it is doing and can handle the new informational codes
6637 appropriately.
6638 [Steve Henson]
6639
6640 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6641 a general "ANY" type, as such it should be able to decode anything
6642 including tagged types. However it didn't check the class so it would
6643 wrongly interpret tagged types in the same way as their universal
6644 counterpart and unknown types were just rejected. Changed so that the
6645 tagged and unknown types are handled in the same way as a SEQUENCE:
6646 that is the encoding is stored intact. There is also a new type
6647 "V_ASN1_OTHER" which is used when the class is not universal, in this
6648 case we have no idea what the actual type is so we just lump them all
6649 together.
6650 [Steve Henson]
6651
6652 *) On VMS, stdout may very well lead to a file that is written to
6653 in a record-oriented fashion. That means that every write() will
6654 write a separate record, which will be read separately by the
6655 programs trying to read from it. This can be very confusing.
6656
6657 The solution is to put a BIO filter in the way that will buffer
6658 text until a linefeed is reached, and then write everything a
6659 line at a time, so every record written will be an actual line,
6660 not chunks of lines and not (usually doesn't happen, but I've
6661 seen it once) several lines in one record. BIO_f_linebuffer() is
6662 the answer.
6663
6664 Currently, it's a VMS-only method, because that's where it has
6665 been tested well enough.
6666 [Richard Levitte]
6667
6668 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6669 it can return incorrect results.
6670 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6671 but it was in 0.9.6-beta[12].)
6672 [Bodo Moeller]
6673
6674 *) Disable the check for content being present when verifying detached
6675 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6676 include zero length content when signing messages.
6677 [Steve Henson]
6678
6679 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6680 BIO_ctrl (for BIO pairs).
6681 [Bodo Möller]
6682
6683 *) Add DSO method for VMS.
6684 [Richard Levitte]
6685
6686 *) Bug fix: Montgomery multiplication could produce results with the
6687 wrong sign.
6688 [Ulf Möller]
6689
6690 *) Add RPM specification openssl.spec and modify it to build three
6691 packages. The default package contains applications, application
6692 documentation and run-time libraries. The devel package contains
6693 include files, static libraries and function documentation. The
6694 doc package contains the contents of the doc directory. The original
6695 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6696 [Richard Levitte]
6697
6698 *) Add a large number of documentation files for many SSL routines.
6699 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6700
6701 *) Add a configuration entry for Sony News 4.
6702 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6703
6704 *) Don't set the two most significant bits to one when generating a
6705 random number < q in the DSA library.
6706 [Ulf Möller]
6707
6708 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6709 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6710 the underlying transport is blocking) if a handshake took place.
6711 (The default behaviour is needed by applications such as s_client
6712 and s_server that use select() to determine when to use SSL_read;
6713 but for applications that know in advance when to expect data, it
6714 just makes things more complicated.)
6715 [Bodo Moeller]
6716
6717 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6718 from EGD.
6719 [Ben Laurie]
6720
6721 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6722 work better on such systems.
6723 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6724
6725 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6726 Update PKCS12_parse() so it copies the friendlyName and the
6727 keyid to the certificates aux info.
6728 [Steve Henson]
6729
6730 *) Fix bug in PKCS7_verify() which caused an infinite loop
6731 if there was more than one signature.
6732 [Sven Uszpelkat <su@celocom.de>]
6733
6734 *) Major change in util/mkdef.pl to include extra information
6735 about each symbol, as well as presentig variables as well
6736 as functions. This change means that there's n more need
6737 to rebuild the .num files when some algorithms are excluded.
6738 [Richard Levitte]
6739
6740 *) Allow the verify time to be set by an application,
6741 rather than always using the current time.
6742 [Steve Henson]
6743
6744 *) Phase 2 verify code reorganisation. The certificate
6745 verify code now looks up an issuer certificate by a
6746 number of criteria: subject name, authority key id
6747 and key usage. It also verifies self signed certificates
6748 by the same criteria. The main comparison function is
6749 X509_check_issued() which performs these checks.
6750
6751 Lot of changes were necessary in order to support this
6752 without completely rewriting the lookup code.
6753
6754 Authority and subject key identifier are now cached.
6755
6756 The LHASH 'certs' is X509_STORE has now been replaced
6757 by a STACK_OF(X509_OBJECT). This is mainly because an
6758 LHASH can't store or retrieve multiple objects with
6759 the same hash value.
6760
6761 As a result various functions (which were all internal
6762 use only) have changed to handle the new X509_STORE
6763 structure. This will break anything that messed round
6764 with X509_STORE internally.
6765
6766 The functions X509_STORE_add_cert() now checks for an
6767 exact match, rather than just subject name.
6768
6769 The X509_STORE API doesn't directly support the retrieval
6770 of multiple certificates matching a given criteria, however
6771 this can be worked round by performing a lookup first
6772 (which will fill the cache with candidate certificates)
6773 and then examining the cache for matches. This is probably
6774 the best we can do without throwing out X509_LOOKUP
6775 entirely (maybe later...).
6776
6777 The X509_VERIFY_CTX structure has been enhanced considerably.
6778
6779 All certificate lookup operations now go via a get_issuer()
6780 callback. Although this currently uses an X509_STORE it
6781 can be replaced by custom lookups. This is a simple way
6782 to bypass the X509_STORE hackery necessary to make this
6783 work and makes it possible to use more efficient techniques
6784 in future. A very simple version which uses a simple
6785 STACK for its trusted certificate store is also provided
6786 using X509_STORE_CTX_trusted_stack().
6787
6788 The verify_cb() and verify() callbacks now have equivalents
6789 in the X509_STORE_CTX structure.
6790
6791 X509_STORE_CTX also has a 'flags' field which can be used
6792 to customise the verify behaviour.
6793 [Steve Henson]
6794
6795 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6796 excludes S/MIME capabilities.
6797 [Steve Henson]
6798
6799 *) When a certificate request is read in keep a copy of the
6800 original encoding of the signed data and use it when outputing
6801 again. Signatures then use the original encoding rather than
6802 a decoded, encoded version which may cause problems if the
6803 request is improperly encoded.
6804 [Steve Henson]
6805
6806 *) For consistency with other BIO_puts implementations, call
6807 buffer_write(b, ...) directly in buffer_puts instead of calling
6808 BIO_write(b, ...).
6809
6810 In BIO_puts, increment b->num_write as in BIO_write.
6811 [Peter.Sylvester@EdelWeb.fr]
6812
6813 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6814 BN_zero, we may not return a BIGNUM with an array consisting of
6815 words set to zero.)
6816 [Bodo Moeller]
6817
6818 *) Avoid calling abort() from within the library when problems are
6819 detected, except if preprocessor symbols have been defined
6820 (such as REF_CHECK, BN_DEBUG etc.).
6821 [Bodo Moeller]
6822
6823 *) New openssl application 'rsautl'. This utility can be
6824 used for low level RSA operations. DER public key
6825 BIO/fp routines also added.
6826 [Steve Henson]
6827
6828 *) New Configure entry and patches for compiling on QNX 4.
6829 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6830
6831 *) A demo state-machine implementation was sponsored by
6832 Nuron (http://www.nuron.com/) and is now available in
6833 demos/state_machine.
6834 [Ben Laurie]
6835
6836 *) New options added to the 'dgst' utility for signature
6837 generation and verification.
6838 [Steve Henson]
6839
6840 *) Unrecognized PKCS#7 content types are now handled via a
6841 catch all ASN1_TYPE structure. This allows unsupported
6842 types to be stored as a "blob" and an application can
6843 encode and decode it manually.
6844 [Steve Henson]
6845
6846 *) Fix various signed/unsigned issues to make a_strex.c
6847 compile under VC++.
6848 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6849
6850 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6851 length if passed a buffer. ASN1_INTEGER_to_BN failed
6852 if passed a NULL BN and its argument was negative.
6853 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6854
6855 *) Modification to PKCS#7 encoding routines to output definite
6856 length encoding. Since currently the whole structures are in
6857 memory there's not real point in using indefinite length
6858 constructed encoding. However if OpenSSL is compiled with
6859 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6860 [Steve Henson]
6861
6862 *) Added BIO_vprintf() and BIO_vsnprintf().
6863 [Richard Levitte]
6864
6865 *) Added more prefixes to parse for in the the strings written
6866 through a logging bio, to cover all the levels that are available
6867 through syslog. The prefixes are now:
6868
6869 PANIC, EMERG, EMR => LOG_EMERG
6870 ALERT, ALR => LOG_ALERT
6871 CRIT, CRI => LOG_CRIT
6872 ERROR, ERR => LOG_ERR
6873 WARNING, WARN, WAR => LOG_WARNING
6874 NOTICE, NOTE, NOT => LOG_NOTICE
6875 INFO, INF => LOG_INFO
6876 DEBUG, DBG => LOG_DEBUG
6877
6878 and as before, if none of those prefixes are present at the
6879 beginning of the string, LOG_ERR is chosen.
6880
6881 On Win32, the LOG_* levels are mapped according to this:
6882
6883 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6884 LOG_WARNING => EVENTLOG_WARNING_TYPE
6885 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6886
6887 [Richard Levitte]
6888
6889 *) Made it possible to reconfigure with just the configuration
6890 argument "reconf" or "reconfigure". The command line arguments
6891 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6892 and are retrieved from there when reconfiguring.
6893 [Richard Levitte]
6894
6895 *) MD4 implemented.
6896 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6897
6898 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6899 [Richard Levitte]
6900
6901 *) The obj_dat.pl script was messing up the sorting of object
6902 names. The reason was that it compared the quoted version
6903 of strings as a result "OCSP" > "OCSP Signing" because
6904 " > SPACE. Changed script to store unquoted versions of
6905 names and add quotes on output. It was also omitting some
6906 names from the lookup table if they were given a default
6907 value (that is if SN is missing it is given the same
6908 value as LN and vice versa), these are now added on the
6909 grounds that if an object has a name we should be able to
6910 look it up. Finally added warning output when duplicate
6911 short or long names are found.
6912 [Steve Henson]
6913
6914 *) Changes needed for Tandem NSK.
6915 [Scott Uroff <scott@xypro.com>]
6916
6917 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6918 RSA_padding_check_SSLv23(), special padding was never detected
6919 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6920 version rollback attacks was not effective.
6921
6922 In s23_clnt.c, don't use special rollback-attack detection padding
6923 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6924 client; similarly, in s23_srvr.c, don't do the rollback check if
6925 SSL 2.0 is the only protocol enabled in the server.
6926 [Bodo Moeller]
6927
6928 *) Make it possible to get hexdumps of unprintable data with 'openssl
6929 asn1parse'. By implication, the functions ASN1_parse_dump() and
6930 BIO_dump_indent() are added.
6931 [Richard Levitte]
6932
6933 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6934 these print out strings and name structures based on various
6935 flags including RFC2253 support and proper handling of
6936 multibyte characters. Added options to the 'x509' utility
6937 to allow the various flags to be set.
6938 [Steve Henson]
6939
6940 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6941 Also change the functions X509_cmp_current_time() and
6942 X509_gmtime_adj() work with an ASN1_TIME structure,
6943 this will enable certificates using GeneralizedTime in validity
6944 dates to be checked.
6945 [Steve Henson]
6946
6947 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6948 negative public key encodings) on by default,
6949 NO_NEG_PUBKEY_BUG can be set to disable it.
6950 [Steve Henson]
6951
6952 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6953 content octets. An i2c_ASN1_OBJECT is unnecessary because
6954 the encoding can be trivially obtained from the structure.
6955 [Steve Henson]
6956
6957 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6958 not read locks (CRYPTO_r_[un]lock).
6959 [Bodo Moeller]
6960
6961 *) A first attempt at creating official support for shared
6962 libraries through configuration. I've kept it so the
6963 default is static libraries only, and the OpenSSL programs
6964 are always statically linked for now, but there are
6965 preparations for dynamic linking in place.
6966 This has been tested on Linux and Tru64.
6967 [Richard Levitte]
6968
6969 *) Randomness polling function for Win9x, as described in:
6970 Peter Gutmann, Software Generation of Practically Strong
6971 Random Numbers.
6972 [Ulf Möller]
6973
6974 *) Fix so PRNG is seeded in req if using an already existing
6975 DSA key.
6976 [Steve Henson]
6977
6978 *) New options to smime application. -inform and -outform
6979 allow alternative formats for the S/MIME message including
6980 PEM and DER. The -content option allows the content to be
6981 specified separately. This should allow things like Netscape
6982 form signing output easier to verify.
6983 [Steve Henson]
6984
6985 *) Fix the ASN1 encoding of tags using the 'long form'.
6986 [Steve Henson]
6987
6988 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6989 STRING types. These convert content octets to and from the
6990 underlying type. The actual tag and length octets are
6991 already assumed to have been read in and checked. These
6992 are needed because all other string types have virtually
6993 identical handling apart from the tag. By having versions
6994 of the ASN1 functions that just operate on content octets
6995 IMPLICIT tagging can be handled properly. It also allows
6996 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6997 and ASN1_INTEGER are identical apart from the tag.
6998 [Steve Henson]
6999
7000 *) Change the handling of OID objects as follows:
7001
7002 - New object identifiers are inserted in objects.txt, following
7003 the syntax given in objects.README.
7004 - objects.pl is used to process obj_mac.num and create a new
7005 obj_mac.h.
7006 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7007 obj_mac.h.
7008
7009 This is currently kind of a hack, and the perl code in objects.pl
7010 isn't very elegant, but it works as I intended. The simplest way
7011 to check that it worked correctly is to look in obj_dat.h and
7012 check the array nid_objs and make sure the objects haven't moved
7013 around (this is important!). Additions are OK, as well as
7014 consistent name changes.
7015 [Richard Levitte]
7016
7017 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7018 [Bodo Moeller]
7019
7020 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7021 The given file adds to whatever has already been seeded into the
7022 random pool through the RANDFILE configuration file option or
7023 environment variable, or the default random state file.
7024 [Richard Levitte]
7025
7026 *) mkstack.pl now sorts each macro group into lexical order.
7027 Previously the output order depended on the order the files
7028 appeared in the directory, resulting in needless rewriting
7029 of safestack.h .
7030 [Steve Henson]
7031
7032 *) Patches to make OpenSSL compile under Win32 again. Mostly
7033 work arounds for the VC++ problem that it treats func() as
7034 func(void). Also stripped out the parts of mkdef.pl that
7035 added extra typesafe functions: these no longer exist.
7036 [Steve Henson]
7037
7038 *) Reorganisation of the stack code. The macros are now all
7039 collected in safestack.h . Each macro is defined in terms of
7040 a "stack macro" of the form SKM_<name>(type, a, b). The
7041 DEBUG_SAFESTACK is now handled in terms of function casts,
7042 this has the advantage of retaining type safety without the
7043 use of additional functions. If DEBUG_SAFESTACK is not defined
7044 then the non typesafe macros are used instead. Also modified the
7045 mkstack.pl script to handle the new form. Needs testing to see
7046 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7047 the default if no major problems. Similar behaviour for ASN1_SET_OF
7048 and PKCS12_STACK_OF.
7049 [Steve Henson]
7050
7051 *) When some versions of IIS use the 'NET' form of private key the
7052 key derivation algorithm is different. Normally MD5(password) is
7053 used as a 128 bit RC4 key. In the modified case
7054 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7055 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7056 as the old Netscape_RSA functions except they have an additional
7057 'sgckey' parameter which uses the modified algorithm. Also added
7058 an -sgckey command line option to the rsa utility. Thanks to
7059 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7060 algorithm to openssl-dev.
7061 [Steve Henson]
7062
7063 *) The evp_local.h macros were using 'c.##kname' which resulted in
7064 invalid expansion on some systems (SCO 5.0.5 for example).
7065 Corrected to 'c.kname'.
7066 [Phillip Porch <root@theporch.com>]
7067
7068 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7069 a STACK of email addresses from a certificate or request, these look
7070 in the subject name and the subject alternative name extensions and
7071 omit any duplicate addresses.
7072 [Steve Henson]
7073
7074 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7075 This makes DSA verification about 2 % faster.
7076 [Bodo Moeller]
7077
7078 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7079 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7080 plus overhead for 1024 bit moduli).
7081 This makes exponentiations about 0.5 % faster for 1024 bit
7082 exponents (as measured by "openssl speed rsa2048").
7083 [Bodo Moeller]
7084
7085 *) Rename memory handling macros to avoid conflicts with other
7086 software:
7087 Malloc => OPENSSL_malloc
7088 Malloc_locked => OPENSSL_malloc_locked
7089 Realloc => OPENSSL_realloc
7090 Free => OPENSSL_free
7091 [Richard Levitte]
7092
7093 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7094 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7095 [Bodo Moeller]
7096
7097 *) CygWin32 support.
7098 [John Jarvie <jjarvie@newsguy.com>]
7099
7100 *) The type-safe stack code has been rejigged. It is now only compiled
7101 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7102 by default all type-specific stack functions are "#define"d back to
7103 standard stack functions. This results in more streamlined output
7104 but retains the type-safety checking possibilities of the original
7105 approach.
7106 [Geoff Thorpe]
7107
7108 *) The STACK code has been cleaned up, and certain type declarations
7109 that didn't make a lot of sense have been brought in line. This has
7110 also involved a cleanup of sorts in safestack.h to more correctly
7111 map type-safe stack functions onto their plain stack counterparts.
7112 This work has also resulted in a variety of "const"ifications of
7113 lots of the code, especially "_cmp" operations which should normally
7114 be prototyped with "const" parameters anyway.
7115 [Geoff Thorpe]
7116
7117 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7118 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7119 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7120 where all of 'md' is used each time the PRNG is used, but 'state'
7121 is used only indexed by a cyclic counter. As entropy may not be
7122 well distributed from the beginning, 'md' is important as a
7123 chaining variable. However, the output function chains only half
7124 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7125 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7126 in all of 'state' being rewritten, with the new values depending
7127 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7128 [Bodo Moeller]
7129
7130 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7131 the handshake is continued after ssl_verify_cert_chain();
7132 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7133 can lead to 'unexplainable' connection aborts later.
7134 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7135
7136 *) Major EVP API cipher revision.
7137 Add hooks for extra EVP features. This allows various cipher
7138 parameters to be set in the EVP interface. Support added for variable
7139 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7140 setting of RC2 and RC5 parameters.
7141
7142 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7143 ciphers.
7144
7145 Remove lots of duplicated code from the EVP library. For example *every*
7146 cipher init() function handles the 'iv' in the same way according to the
7147 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7148 for CFB and OFB modes they zero ctx->num.
7149
7150 New functionality allows removal of S/MIME code RC2 hack.
7151
7152 Most of the routines have the same form and so can be declared in terms
7153 of macros.
7154
7155 By shifting this to the top level EVP_CipherInit() it can be removed from
7156 all individual ciphers. If the cipher wants to handle IVs or keys
7157 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7158 flags.
7159
7160 Change lots of functions like EVP_EncryptUpdate() to now return a
7161 value: although software versions of the algorithms cannot fail
7162 any installed hardware versions can.
7163 [Steve Henson]
7164
7165 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7166 this option is set, tolerate broken clients that send the negotiated
7167 protocol version number instead of the requested protocol version
7168 number.
7169 [Bodo Moeller]
7170
7171 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7172 i.e. non-zero for export ciphersuites, zero otherwise.
7173 Previous versions had this flag inverted, inconsistent with
7174 rsa_tmp_cb (..._TMP_RSA_CB).
7175 [Bodo Moeller; problem reported by Amit Chopra]
7176
7177 *) Add missing DSA library text string. Work around for some IIS
7178 key files with invalid SEQUENCE encoding.
7179 [Steve Henson]
7180
7181 *) Add a document (doc/standards.txt) that list all kinds of standards
7182 and so on that are implemented in OpenSSL.
7183 [Richard Levitte]
7184
7185 *) Enhance c_rehash script. Old version would mishandle certificates
7186 with the same subject name hash and wouldn't handle CRLs at all.
7187 Added -fingerprint option to crl utility, to support new c_rehash
7188 features.
7189 [Steve Henson]
7190
7191 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7192 [Ulf Möller]
7193
7194 *) Fix for SSL server purpose checking. Server checking was
7195 rejecting certificates which had extended key usage present
7196 but no ssl client purpose.
7197 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7198
7199 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7200 is a little unclear about how a blank password is handled.
7201 Since the password in encoded as a BMPString with terminating
7202 double NULL a zero length password would end up as just the
7203 double NULL. However no password at all is different and is
7204 handled differently in the PKCS#12 key generation code. NS
7205 treats a blank password as zero length. MSIE treats it as no
7206 password on export: but it will try both on import. We now do
7207 the same: PKCS12_parse() tries zero length and no password if
7208 the password is set to "" or NULL (NULL is now a valid password:
7209 it wasn't before) as does the pkcs12 application.
7210 [Steve Henson]
7211
7212 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7213 perror when PEM_read_bio_X509_REQ fails, the error message must
7214 be obtained from the error queue.
7215 [Bodo Moeller]
7216
7217 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7218 it in ERR_remove_state if appropriate, and change ERR_get_state
7219 accordingly to avoid race conditions (this is necessary because
7220 thread_hash is no longer constant once set).
7221 [Bodo Moeller]
7222
7223 *) Bugfix for linux-elf makefile.one.
7224 [Ulf Möller]
7225
7226 *) RSA_get_default_method() will now cause a default
7227 RSA_METHOD to be chosen if one doesn't exist already.
7228 Previously this was only set during a call to RSA_new()
7229 or RSA_new_method(NULL) meaning it was possible for
7230 RSA_get_default_method() to return NULL.
7231 [Geoff Thorpe]
7232
7233 *) Added native name translation to the existing DSO code
7234 that will convert (if the flag to do so is set) filenames
7235 that are sufficiently small and have no path information
7236 into a canonical native form. Eg. "blah" converted to
7237 "libblah.so" or "blah.dll" etc.
7238 [Geoff Thorpe]
7239
7240 *) New function ERR_error_string_n(e, buf, len) which is like
7241 ERR_error_string(e, buf), but writes at most 'len' bytes
7242 including the 0 terminator. For ERR_error_string_n, 'buf'
7243 may not be NULL.
7244 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7245
7246 *) CONF library reworked to become more general. A new CONF
7247 configuration file reader "class" is implemented as well as a
7248 new functions (NCONF_*, for "New CONF") to handle it. The now
7249 old CONF_* functions are still there, but are reimplemented to
7250 work in terms of the new functions. Also, a set of functions
7251 to handle the internal storage of the configuration data is
7252 provided to make it easier to write new configuration file
7253 reader "classes" (I can definitely see something reading a
7254 configuration file in XML format, for example), called _CONF_*,
7255 or "the configuration storage API"...
7256
7257 The new configuration file reading functions are:
7258
7259 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7260 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7261
7262 NCONF_default, NCONF_WIN32
7263
7264 NCONF_dump_fp, NCONF_dump_bio
7265
7266 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7267 NCONF_new creates a new CONF object. This works in the same way
7268 as other interfaces in OpenSSL, like the BIO interface.
7269 NCONF_dump_* dump the internal storage of the configuration file,
7270 which is useful for debugging. All other functions take the same
7271 arguments as the old CONF_* functions wth the exception of the
7272 first that must be a `CONF *' instead of a `LHASH *'.
7273
7274 To make it easer to use the new classes with the old CONF_* functions,
7275 the function CONF_set_default_method is provided.
7276 [Richard Levitte]
7277
7278 *) Add '-tls1' option to 'openssl ciphers', which was already
7279 mentioned in the documentation but had not been implemented.
7280 (This option is not yet really useful because even the additional
7281 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7282 [Bodo Moeller]
7283
7284 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7285 OpenSSL-based applications) load shared libraries and bind to
7286 them in a portable way.
7287 [Geoff Thorpe, with contributions from Richard Levitte]
7288
7289 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7290
7291 *) Make sure _lrotl and _lrotr are only used with MSVC.
7292
7293 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7294 (the default implementation of RAND_status).
7295
7296 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7297 to '-clrext' (= clear extensions), as intended and documented.
7298 [Bodo Moeller; inconsistency pointed out by Michael Attili
7299 <attili@amaxo.com>]
7300
7301 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7302 was larger than the MD block size.
7303 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7304
7305 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7306 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7307 using the passed key: if the passed key was a private key the result
7308 of X509_print(), for example, would be to print out all the private key
7309 components.
7310 [Steve Henson]
7311
7312 *) des_quad_cksum() byte order bug fix.
7313 [Ulf Möller, using the problem description in krb4-0.9.7, where
7314 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7315
7316 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7317 discouraged.
7318 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7319
7320 *) For easily testing in shell scripts whether some command
7321 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7322 returns with exit code 0 iff no command of the given name is available.
7323 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7324 the output goes to stdout and nothing is printed to stderr.
7325 Additional arguments are always ignored.
7326
7327 Since for each cipher there is a command of the same name,
7328 the 'no-cipher' compilation switches can be tested this way.
7329
7330 ('openssl no-XXX' is not able to detect pseudo-commands such
7331 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7332 [Bodo Moeller]
7333
7334 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7335 [Bodo Moeller]
7336
7337 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7338 is set; it will be thrown away anyway because each handshake creates
7339 its own key.
7340 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7341 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7342 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7343 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7344 [Bodo Moeller]
7345
7346 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7347 'Q' and 'R' lose their special meanings (quit/renegotiate).
7348 This is part of what -quiet does; unlike -quiet, -ign_eof
7349 does not suppress any output.
7350 [Richard Levitte]
7351
7352 *) Add compatibility options to the purpose and trust code. The
7353 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7354 accepts a certificate or CA, this was the previous behaviour,
7355 with all the associated security issues.
7356
7357 X509_TRUST_COMPAT is the old trust behaviour: only and
7358 automatically trust self signed roots in certificate store. A
7359 new trust setting X509_TRUST_DEFAULT is used to specify that
7360 a purpose has no associated trust setting and it should instead
7361 use the value in the default purpose.
7362 [Steve Henson]
7363
7364 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7365 and fix a memory leak.
7366 [Steve Henson]
7367
7368 *) In util/mkerr.pl (which implements 'make errors'), preserve
7369 reason strings from the previous version of the .c file, as
7370 the default to have only downcase letters (and digits) in
7371 automatically generated reasons codes is not always appropriate.
7372 [Bodo Moeller]
7373
7374 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7375 using strerror. Previously, ERR_reason_error_string() returned
7376 library names as reason strings for SYSerr; but SYSerr is a special
7377 case where small numbers are errno values, not library numbers.
7378 [Bodo Moeller]
7379
7380 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7381 converts DSA parameters into DH parameters. (When creating parameters,
7382 DSA_generate_parameters is used.)
7383 [Bodo Moeller]
7384
7385 *) Include 'length' (recommended exponent length) in C code generated
7386 by 'openssl dhparam -C'.
7387 [Bodo Moeller]
7388
7389 *) The second argument to set_label in perlasm was already being used
7390 so couldn't be used as a "file scope" flag. Moved to third argument
7391 which was free.
7392 [Steve Henson]
7393
7394 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7395 instead of RAND_bytes for encryption IVs and salts.
7396 [Bodo Moeller]
7397
7398 *) Include RAND_status() into RAND_METHOD instead of implementing
7399 it only for md_rand.c Otherwise replacing the PRNG by calling
7400 RAND_set_rand_method would be impossible.
7401 [Bodo Moeller]
7402
7403 *) Don't let DSA_generate_key() enter an infinite loop if the random
7404 number generation fails.
7405 [Bodo Moeller]
7406
7407 *) New 'rand' application for creating pseudo-random output.
7408 [Bodo Moeller]
7409
7410 *) Added configuration support for Linux/IA64
7411 [Rolf Haberrecker <rolf@suse.de>]
7412
7413 *) Assembler module support for Mingw32.
7414 [Ulf Möller]
7415
7416 *) Shared library support for HPUX (in shlib/).
7417 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7418
7419 *) Shared library support for Solaris gcc.
7420 [Lutz Behnke <behnke@trustcenter.de>]
7421
7422 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7423
7424 *) PKCS7_encrypt() was adding text MIME headers twice because they
7425 were added manually and by SMIME_crlf_copy().
7426 [Steve Henson]
7427
7428 *) In bntest.c don't call BN_rand with zero bits argument.
7429 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7430
7431 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7432 case was implemented. This caused BN_div_recp() to fail occasionally.
7433 [Ulf Möller]
7434
7435 *) Add an optional second argument to the set_label() in the perl
7436 assembly language builder. If this argument exists and is set
7437 to 1 it signals that the assembler should use a symbol whose
7438 scope is the entire file, not just the current function. This
7439 is needed with MASM which uses the format label:: for this scope.
7440 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7441
7442 *) Change the ASN1 types so they are typedefs by default. Before
7443 almost all types were #define'd to ASN1_STRING which was causing
7444 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7445 for example.
7446 [Steve Henson]
7447
7448 *) Change names of new functions to the new get1/get0 naming
7449 convention: After 'get1', the caller owns a reference count
7450 and has to call ..._free; 'get0' returns a pointer to some
7451 data structure without incrementing reference counters.
7452 (Some of the existing 'get' functions increment a reference
7453 counter, some don't.)
7454 Similarly, 'set1' and 'add1' functions increase reference
7455 counters or duplicate objects.
7456 [Steve Henson]
7457
7458 *) Allow for the possibility of temp RSA key generation failure:
7459 the code used to assume it always worked and crashed on failure.
7460 [Steve Henson]
7461
7462 *) Fix potential buffer overrun problem in BIO_printf().
7463 [Ulf Möller, using public domain code by Patrick Powell; problem
7464 pointed out by David Sacerdote <das33@cornell.edu>]
7465
7466 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7467 RAND_egd() and RAND_status(). In the command line application,
7468 the EGD socket can be specified like a seed file using RANDFILE
7469 or -rand.
7470 [Ulf Möller]
7471
7472 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7473 Some CAs (e.g. Verisign) distribute certificates in this form.
7474 [Steve Henson]
7475
7476 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7477 list to exclude them. This means that no special compilation option
7478 is needed to use anonymous DH: it just needs to be included in the
7479 cipher list.
7480 [Steve Henson]
7481
7482 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7483 EVP_MD_type. The old functionality is available in a new macro called
7484 EVP_MD_md(). Change code that uses it and update docs.
7485 [Steve Henson]
7486
7487 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7488 where the 'void *' argument is replaced by a function pointer argument.
7489 Previously 'void *' was abused to point to functions, which works on
7490 many platforms, but is not correct. As these functions are usually
7491 called by macros defined in OpenSSL header files, most source code
7492 should work without changes.
7493 [Richard Levitte]
7494
7495 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7496 sections with information on -D... compiler switches used for
7497 compiling the library so that applications can see them. To enable
7498 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7499 must be defined. E.g.,
7500 #define OPENSSL_ALGORITHM_DEFINES
7501 #include <openssl/opensslconf.h>
7502 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7503 [Richard Levitte, Ulf and Bodo Möller]
7504
7505 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7506 record layer.
7507 [Bodo Moeller]
7508
7509 *) Change the 'other' type in certificate aux info to a STACK_OF
7510 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7511 the required ASN1 format: arbitrary types determined by an OID.
7512 [Steve Henson]
7513
7514 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7515 argument to 'req'. This is not because the function is newer or
7516 better than others it just uses the work 'NEW' in the certificate
7517 request header lines. Some software needs this.
7518 [Steve Henson]
7519
7520 *) Reorganise password command line arguments: now passwords can be
7521 obtained from various sources. Delete the PEM_cb function and make
7522 it the default behaviour: i.e. if the callback is NULL and the
7523 usrdata argument is not NULL interpret it as a null terminated pass
7524 phrase. If usrdata and the callback are NULL then the pass phrase
7525 is prompted for as usual.
7526 [Steve Henson]
7527
7528 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7529 the support is automatically enabled. The resulting binaries will
7530 autodetect the card and use it if present.
7531 [Ben Laurie and Compaq Inc.]
7532
7533 *) Work around for Netscape hang bug. This sends certificate request
7534 and server done in one record. Since this is perfectly legal in the
7535 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7536 the bugs/SSLv3 entry for more info.
7537 [Steve Henson]
7538
7539 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7540 [Andy Polyakov]
7541
7542 *) Add -rand argument to smime and pkcs12 applications and read/write
7543 of seed file.
7544 [Steve Henson]
7545
7546 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7547 [Bodo Moeller]
7548
7549 *) Add command line password options to the remaining applications.
7550 [Steve Henson]
7551
7552 *) Bug fix for BN_div_recp() for numerators with an even number of
7553 bits.
7554 [Ulf Möller]
7555
7556 *) More tests in bntest.c, and changed test_bn output.
7557 [Ulf Möller]
7558
7559 *) ./config recognizes MacOS X now.
7560 [Andy Polyakov]
7561
7562 *) Bug fix for BN_div() when the first words of num and divsor are
7563 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7564 [Ulf Möller]
7565
7566 *) Add support for various broken PKCS#8 formats, and command line
7567 options to produce them.
7568 [Steve Henson]
7569
7570 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7571 get temporary BIGNUMs from a BN_CTX.
7572 [Ulf Möller]
7573
7574 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7575 for p == 0.
7576 [Ulf Möller]
7577
7578 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7579 include a #define from the old name to the new. The original intent
7580 was that statically linked binaries could for example just call
7581 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7582 link with digests. This never worked becayse SSLeay_add_all_digests()
7583 and SSLeay_add_all_ciphers() were in the same source file so calling
7584 one would link with the other. They are now in separate source files.
7585 [Steve Henson]
7586
7587 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7588 [Steve Henson]
7589
7590 *) Use a less unusual form of the Miller-Rabin primality test (it used
7591 a binary algorithm for exponentiation integrated into the Miller-Rabin
7592 loop, our standard modexp algorithms are faster).
7593 [Bodo Moeller]
7594
7595 *) Support for the EBCDIC character set completed.
7596 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7597
7598 *) Source code cleanups: use const where appropriate, eliminate casts,
7599 use void * instead of char * in lhash.
7600 [Ulf Möller]
7601
7602 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7603 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7604 this the server could overwrite ephemeral keys that the client
7605 has already seen).
7606 [Bodo Moeller]
7607
7608 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7609 using 50 iterations of the Rabin-Miller test.
7610
7611 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7612 iterations of the Rabin-Miller test as required by the appendix
7613 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7614 As BN_is_prime_fasttest includes trial division, DSA parameter
7615 generation becomes much faster.
7616
7617 This implies a change for the callback functions in DSA_is_prime
7618 and DSA_generate_parameters: The callback function is called once
7619 for each positive witness in the Rabin-Miller test, not just
7620 occasionally in the inner loop; and the parameters to the
7621 callback function now provide an iteration count for the outer
7622 loop rather than for the current invocation of the inner loop.
7623 DSA_generate_parameters additionally can call the callback
7624 function with an 'iteration count' of -1, meaning that a
7625 candidate has passed the trial division test (when q is generated
7626 from an application-provided seed, trial division is skipped).
7627 [Bodo Moeller]
7628
7629 *) New function BN_is_prime_fasttest that optionally does trial
7630 division before starting the Rabin-Miller test and has
7631 an additional BN_CTX * argument (whereas BN_is_prime always
7632 has to allocate at least one BN_CTX).
7633 'callback(1, -1, cb_arg)' is called when a number has passed the
7634 trial division stage.
7635 [Bodo Moeller]
7636
7637 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7638 as ASN1_TIME.
7639 [Steve Henson]
7640
7641 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7642 [Steve Henson]
7643
7644 *) New function BN_pseudo_rand().
7645 [Ulf Möller]
7646
7647 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7648 bignum version of BN_from_montgomery() with the working code from
7649 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7650 the comments.
7651 [Ulf Möller]
7652
7653 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7654 made it impossible to use the same SSL_SESSION data structure in
7655 SSL2 clients in multiple threads.
7656 [Bodo Moeller]
7657
7658 *) The return value of RAND_load_file() no longer counts bytes obtained
7659 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7660 to seed the PRNG (previously an explicit byte count was required).
7661 [Ulf Möller, Bodo Möller]
7662
7663 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7664 used (char *) instead of (void *) and had casts all over the place.
7665 [Steve Henson]
7666
7667 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7668 [Ulf Möller]
7669
7670 *) Retain source code compatibility for BN_prime_checks macro:
7671 BN_is_prime(..., BN_prime_checks, ...) now uses
7672 BN_prime_checks_for_size to determine the appropriate number of
7673 Rabin-Miller iterations.
7674 [Ulf Möller]
7675
7676 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7677 DH_CHECK_P_NOT_SAFE_PRIME.
7678 (Check if this is true? OpenPGP calls them "strong".)
7679 [Ulf Möller]
7680
7681 *) Merge the functionality of "dh" and "gendh" programs into a new program
7682 "dhparam". The old programs are retained for now but will handle DH keys
7683 (instead of parameters) in future.
7684 [Steve Henson]
7685
7686 *) Make the ciphers, s_server and s_client programs check the return values
7687 when a new cipher list is set.
7688 [Steve Henson]
7689
7690 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7691 ciphers. Before when the 56bit ciphers were enabled the sorting was
7692 wrong.
7693
7694 The syntax for the cipher sorting has been extended to support sorting by
7695 cipher-strength (using the strength_bits hard coded in the tables).
7696 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7697
7698 Fix a bug in the cipher-command parser: when supplying a cipher command
7699 string with an "undefined" symbol (neither command nor alphanumeric
7700 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7701 an error is flagged.
7702
7703 Due to the strength-sorting extension, the code of the
7704 ssl_create_cipher_list() function was completely rearranged. I hope that
7705 the readability was also increased :-)
7706 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7707
7708 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7709 for the first serial number and places 2 in the serial number file. This
7710 avoids problems when the root CA is created with serial number zero and
7711 the first user certificate has the same issuer name and serial number
7712 as the root CA.
7713 [Steve Henson]
7714
7715 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7716 the new code. Add documentation for this stuff.
7717 [Steve Henson]
7718
7719 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7720 X509_*() to X509at_*() on the grounds that they don't handle X509
7721 structures and behave in an analagous way to the X509v3 functions:
7722 they shouldn't be called directly but wrapper functions should be used
7723 instead.
7724
7725 So we also now have some wrapper functions that call the X509at functions
7726 when passed certificate requests. (TO DO: similar things can be done with
7727 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7728 things. Some of these need some d2i or i2d and print functionality
7729 because they handle more complex structures.)
7730 [Steve Henson]
7731
7732 *) Add missing #ifndefs that caused missing symbols when building libssl
7733 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7734 NO_RSA in ssl/s2*.c.
7735 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7736
7737 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7738 has a return value which indicates the quality of the random data
7739 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7740 error queue. New function RAND_pseudo_bytes() generates output that is
7741 guaranteed to be unique but not unpredictable. RAND_add is like
7742 RAND_seed, but takes an extra argument for an entropy estimate
7743 (RAND_seed always assumes full entropy).
7744 [Ulf Möller]
7745
7746 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7747 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7748 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7749 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7750 false-positive rate of at most 2^-80 for random input.
7751 [Bodo Moeller]
7752
7753 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7754 [Bodo Moeller]
7755
7756 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7757 in the 0.9.5 release), this returns the chain
7758 from an X509_CTX structure with a dup of the stack and all
7759 the X509 reference counts upped: so the stack will exist
7760 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7761 to use this.
7762
7763 Also make SSL_SESSION_print() print out the verify return
7764 code.
7765 [Steve Henson]
7766
7767 *) Add manpage for the pkcs12 command. Also change the default
7768 behaviour so MAC iteration counts are used unless the new
7769 -nomaciter option is used. This improves file security and
7770 only older versions of MSIE (4.0 for example) need it.
7771 [Steve Henson]
7772
7773 *) Honor the no-xxx Configure options when creating .DEF files.
7774 [Ulf Möller]
7775
7776 *) Add PKCS#10 attributes to field table: challengePassword,
7777 unstructuredName and unstructuredAddress. These are taken from
7778 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7779 international characters are used.
7780
7781 More changes to X509_ATTRIBUTE code: allow the setting of types
7782 based on strings. Remove the 'loc' parameter when adding
7783 attributes because these will be a SET OF encoding which is sorted
7784 in ASN1 order.
7785 [Steve Henson]
7786
7787 *) Initial changes to the 'req' utility to allow request generation
7788 automation. This will allow an application to just generate a template
7789 file containing all the field values and have req construct the
7790 request.
7791
7792 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7793 used all over the place including certificate requests and PKCS#7
7794 structures. They are currently handled manually where necessary with
7795 some primitive wrappers for PKCS#7. The new functions behave in a
7796 manner analogous to the X509 extension functions: they allow
7797 attributes to be looked up by NID and added.
7798
7799 Later something similar to the X509V3 code would be desirable to
7800 automatically handle the encoding, decoding and printing of the
7801 more complex types. The string types like challengePassword can
7802 be handled by the string table functions.
7803
7804 Also modified the multi byte string table handling. Now there is
7805 a 'global mask' which masks out certain types. The table itself
7806 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7807 is useful when for example there is only one permissible type
7808 (as in countryName) and using the mask might result in no valid
7809 types at all.
7810 [Steve Henson]
7811
7812 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7813 SSL_get_peer_finished to allow applications to obtain the latest
7814 Finished messages sent to the peer or expected from the peer,
7815 respectively. (SSL_get_peer_finished is usually the Finished message
7816 actually received from the peer, otherwise the protocol will be aborted.)
7817
7818 As the Finished message are message digests of the complete handshake
7819 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7820 be used for external authentication procedures when the authentication
7821 provided by SSL/TLS is not desired or is not enough.
7822 [Bodo Moeller]
7823
7824 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7825 the host supports BWX extension and if Compaq C is present on the
7826 $PATH. Just exploiting of the BWX extension results in 20-30%
7827 performance kick for some algorithms, e.g. DES and RC4 to mention
7828 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7829 SHA1.
7830 [Andy Polyakov]
7831
7832 *) Add support for MS "fast SGC". This is arguably a violation of the
7833 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7834 weak crypto and after checking the certificate is SGC a second one
7835 with strong crypto. MS SGC stops the first handshake after receiving
7836 the server certificate message and sends a second client hello. Since
7837 a server will typically do all the time consuming operations before
7838 expecting any further messages from the client (server key exchange
7839 is the most expensive) there is little difference between the two.
7840
7841 To get OpenSSL to support MS SGC we have to permit a second client
7842 hello message after we have sent server done. In addition we have to
7843 reset the MAC if we do get this second client hello.
7844 [Steve Henson]
7845
7846 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7847 if a DER encoded private key is RSA or DSA traditional format. Changed
7848 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7849 format DER encoded private key. Newer code should use PKCS#8 format which
7850 has the key type encoded in the ASN1 structure. Added DER private key
7851 support to pkcs8 application.
7852 [Steve Henson]
7853
7854 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7855 ciphersuites has been selected (as required by the SSL 3/TLS 1
7856 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7857 is set, we interpret this as a request to violate the specification
7858 (the worst that can happen is a handshake failure, and 'correct'
7859 behaviour would result in a handshake failure anyway).
7860 [Bodo Moeller]
7861
7862 *) In SSL_CTX_add_session, take into account that there might be multiple
7863 SSL_SESSION structures with the same session ID (e.g. when two threads
7864 concurrently obtain them from an external cache).
7865 The internal cache can handle only one SSL_SESSION with a given ID,
7866 so if there's a conflict, we now throw out the old one to achieve
7867 consistency.
7868 [Bodo Moeller]
7869
7870 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7871 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7872 some routines that use cipher OIDs: some ciphers do not have OIDs
7873 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7874 example.
7875 [Steve Henson]
7876
7877 *) Simplify the trust setting structure and code. Now we just have
7878 two sequences of OIDs for trusted and rejected settings. These will
7879 typically have values the same as the extended key usage extension
7880 and any application specific purposes.
7881
7882 The trust checking code now has a default behaviour: it will just
7883 check for an object with the same NID as the passed id. Functions can
7884 be provided to override either the default behaviour or the behaviour
7885 for a given id. SSL client, server and email already have functions
7886 in place for compatibility: they check the NID and also return "trusted"
7887 if the certificate is self signed.
7888 [Steve Henson]
7889
7890 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7891 traditional format into an EVP_PKEY structure.
7892 [Steve Henson]
7893
7894 *) Add a password callback function PEM_cb() which either prompts for
7895 a password if usr_data is NULL or otherwise assumes it is a null
7896 terminated password. Allow passwords to be passed on command line
7897 environment or config files in a few more utilities.
7898 [Steve Henson]
7899
7900 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7901 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7902 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7903 Update documentation.
7904 [Steve Henson]
7905
7906 *) Support for ASN1 "NULL" type. This could be handled before by using
7907 ASN1_TYPE but there wasn't any function that would try to read a NULL
7908 and produce an error if it couldn't. For compatibility we also have
7909 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7910 don't allocate anything because they don't need to.
7911 [Steve Henson]
7912
7913 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7914 for details.
7915 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7916
7917 *) Rebuild of the memory allocation routines used by OpenSSL code and
7918 possibly others as well. The purpose is to make an interface that
7919 provide hooks so anyone can build a separate set of allocation and
7920 deallocation routines to be used by OpenSSL, for example memory
7921 pool implementations, or something else, which was previously hard
7922 since Malloc(), Realloc() and Free() were defined as macros having
7923 the values malloc, realloc and free, respectively (except for Win32
7924 compilations). The same is provided for memory debugging code.
7925 OpenSSL already comes with functionality to find memory leaks, but
7926 this gives people a chance to debug other memory problems.
7927
7928 With these changes, a new set of functions and macros have appeared:
7929
7930 CRYPTO_set_mem_debug_functions() [F]
7931 CRYPTO_get_mem_debug_functions() [F]
7932 CRYPTO_dbg_set_options() [F]
7933 CRYPTO_dbg_get_options() [F]
7934 CRYPTO_malloc_debug_init() [M]
7935
7936 The memory debug functions are NULL by default, unless the library
7937 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7938 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7939 gives the standard debugging functions that come with OpenSSL) or
7940 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7941 provided by the library user) must be used. When the standard
7942 debugging functions are used, CRYPTO_dbg_set_options can be used to
7943 request additional information:
7944 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7945 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7946
7947 Also, things like CRYPTO_set_mem_functions will always give the
7948 expected result (the new set of functions is used for allocation
7949 and deallocation) at all times, regardless of platform and compiler
7950 options.
7951
7952 To finish it up, some functions that were never use in any other
7953 way than through macros have a new API and new semantic:
7954
7955 CRYPTO_dbg_malloc()
7956 CRYPTO_dbg_realloc()
7957 CRYPTO_dbg_free()
7958
7959 All macros of value have retained their old syntax.
7960 [Richard Levitte and Bodo Moeller]
7961
7962 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7963 ordering of SMIMECapabilities wasn't in "strength order" and there
7964 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7965 algorithm.
7966 [Steve Henson]
7967
7968 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7969 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7970 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7971
7972 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7973 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7974 functionality to handle multipart/signed properly) and a utility
7975 called 'smime' to call all this stuff. This is based on code I
7976 originally wrote for Celo who have kindly allowed it to be
7977 included in OpenSSL.
7978 [Steve Henson]
7979
7980 *) Add variants des_set_key_checked and des_set_key_unchecked of
7981 des_set_key (aka des_key_sched). Global variable des_check_key
7982 decides which of these is called by des_set_key; this way
7983 des_check_key behaves as it always did, but applications and
7984 the library itself, which was buggy for des_check_key == 1,
7985 have a cleaner way to pick the version they need.
7986 [Bodo Moeller]
7987
7988 *) New function PKCS12_newpass() which changes the password of a
7989 PKCS12 structure.
7990 [Steve Henson]
7991
7992 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7993 dynamic mix. In both cases the ids can be used as an index into the
7994 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7995 functions so they accept a list of the field values and the
7996 application doesn't need to directly manipulate the X509_TRUST
7997 structure.
7998 [Steve Henson]
7999
8000 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8001 need initialising.
8002 [Steve Henson]
8003
8004 *) Modify the way the V3 extension code looks up extensions. This now
8005 works in a similar way to the object code: we have some "standard"
8006 extensions in a static table which is searched with OBJ_bsearch()
8007 and the application can add dynamic ones if needed. The file
8008 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8009 updated whenever a new extension is added to the core code and kept
8010 in ext_nid order. There is a simple program 'tabtest.c' which checks
8011 this. New extensions are not added too often so this file can readily
8012 be maintained manually.
8013
8014 There are two big advantages in doing things this way. The extensions
8015 can be looked up immediately and no longer need to be "added" using
8016 X509V3_add_standard_extensions(): this function now does nothing.
8017 [Side note: I get *lots* of email saying the extension code doesn't
8018 work because people forget to call this function]
8019 Also no dynamic allocation is done unless new extensions are added:
8020 so if we don't add custom extensions there is no need to call
8021 X509V3_EXT_cleanup().
8022 [Steve Henson]
8023
8024 *) Modify enc utility's salting as follows: make salting the default. Add a
8025 magic header, so unsalted files fail gracefully instead of just decrypting
8026 to garbage. This is because not salting is a big security hole, so people
8027 should be discouraged from doing it.
8028 [Ben Laurie]
8029
8030 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8031 digest to be passed on the command line but it only used this
8032 parameter when signing a certificate. Modified so all relevant
8033 operations are affected by the digest parameter including the
8034 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8035 DSA key was used because it didn't fix the digest.
8036 [Steve Henson]
8037
8038 *) Initial certificate chain verify code. Currently tests the untrusted
8039 certificates for consistency with the verify purpose (which is set
8040 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8041
8042 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8043 this is because it will reject chains with invalid extensions whereas
8044 every previous version of OpenSSL and SSLeay made no checks at all.
8045
8046 Trust code: checks the root CA for the relevant trust settings. Trust
8047 settings have an initial value consistent with the verify purpose: e.g.
8048 if the verify purpose is for SSL client use it expects the CA to be
8049 trusted for SSL client use. However the default value can be changed to
8050 permit custom trust settings: one example of this would be to only trust
8051 certificates from a specific "secure" set of CAs.
8052
8053 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8054 which should be used for version portability: especially since the
8055 verify structure is likely to change more often now.
8056
8057 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8058 to set them. If not set then assume SSL clients will verify SSL servers
8059 and vice versa.
8060
8061 Two new options to the verify program: -untrusted allows a set of
8062 untrusted certificates to be passed in and -purpose which sets the
8063 intended purpose of the certificate. If a purpose is set then the
8064 new chain verify code is used to check extension consistency.
8065 [Steve Henson]
8066
8067 *) Support for the authority information access extension.
8068 [Steve Henson]
8069
8070 *) Modify RSA and DSA PEM read routines to transparently handle
8071 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8072 public keys in a format compatible with certificate
8073 SubjectPublicKeyInfo structures. Unfortunately there were already
8074 functions called *_PublicKey_* which used various odd formats so
8075 these are retained for compatibility: however the DSA variants were
8076 never in a public release so they have been deleted. Changed dsa/rsa
8077 utilities to handle the new format: note no releases ever handled public
8078 keys so we should be OK.
8079
8080 The primary motivation for this change is to avoid the same fiasco
8081 that dogs private keys: there are several incompatible private key
8082 formats some of which are standard and some OpenSSL specific and
8083 require various evil hacks to allow partial transparent handling and
8084 even then it doesn't work with DER formats. Given the option anything
8085 other than PKCS#8 should be dumped: but the other formats have to
8086 stay in the name of compatibility.
8087
8088 With public keys and the benefit of hindsight one standard format
8089 is used which works with EVP_PKEY, RSA or DSA structures: though
8090 it clearly returns an error if you try to read the wrong kind of key.
8091
8092 Added a -pubkey option to the 'x509' utility to output the public key.
8093 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8094 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8095 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8096 that do the same as the EVP_PKEY_assign_*() except they up the
8097 reference count of the added key (they don't "swallow" the
8098 supplied key).
8099 [Steve Henson]
8100
8101 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8102 CRLs would fail if the file contained no certificates or no CRLs:
8103 added a new function to read in both types and return the number
8104 read: this means that if none are read it will be an error. The
8105 DER versions of the certificate and CRL reader would always fail
8106 because it isn't possible to mix certificates and CRLs in DER format
8107 without choking one or the other routine. Changed this to just read
8108 a certificate: this is the best we can do. Also modified the code
8109 in apps/verify.c to take notice of return codes: it was previously
8110 attempting to read in certificates from NULL pointers and ignoring
8111 any errors: this is one reason why the cert and CRL reader seemed
8112 to work. It doesn't check return codes from the default certificate
8113 routines: these may well fail if the certificates aren't installed.
8114 [Steve Henson]
8115
8116 *) Code to support otherName option in GeneralName.
8117 [Steve Henson]
8118
8119 *) First update to verify code. Change the verify utility
8120 so it warns if it is passed a self signed certificate:
8121 for consistency with the normal behaviour. X509_verify
8122 has been modified to it will now verify a self signed
8123 certificate if *exactly* the same certificate appears
8124 in the store: it was previously impossible to trust a
8125 single self signed certificate. This means that:
8126 openssl verify ss.pem
8127 now gives a warning about a self signed certificate but
8128 openssl verify -CAfile ss.pem ss.pem
8129 is OK.
8130 [Steve Henson]
8131
8132 *) For servers, store verify_result in SSL_SESSION data structure
8133 (and add it to external session representation).
8134 This is needed when client certificate verifications fails,
8135 but an application-provided verification callback (set by
8136 SSL_CTX_set_cert_verify_callback) allows accepting the session
8137 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8138 but returns 1): When the session is reused, we have to set
8139 ssl->verify_result to the appropriate error code to avoid
8140 security holes.
8141 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8142
8143 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8144 case in PKCS7_dataInit() where the signed PKCS7 structure
8145 didn't contain any existing data because it was being created.
8146 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8147
8148 *) Add a salt to the key derivation routines in enc.c. This
8149 forms the first 8 bytes of the encrypted file. Also add a
8150 -S option to allow a salt to be input on the command line.
8151 [Steve Henson]
8152
8153 *) New function X509_cmp(). Oddly enough there wasn't a function
8154 to compare two certificates. We do this by working out the SHA1
8155 hash and comparing that. X509_cmp() will be needed by the trust
8156 code.
8157 [Steve Henson]
8158
8159 *) SSL_get1_session() is like SSL_get_session(), but increments
8160 the reference count in the SSL_SESSION returned.
8161 [Geoff Thorpe <geoff@eu.c2.net>]
8162
8163 *) Fix for 'req': it was adding a null to request attributes.
8164 Also change the X509_LOOKUP and X509_INFO code to handle
8165 certificate auxiliary information.
8166 [Steve Henson]
8167
8168 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8169 the 'enc' command.
8170 [Steve Henson]
8171
8172 *) Add the possibility to add extra information to the memory leak
8173 detecting output, to form tracebacks, showing from where each
8174 allocation was originated: CRYPTO_push_info("constant string") adds
8175 the string plus current file name and line number to a per-thread
8176 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8177 is like calling CYRPTO_pop_info() until the stack is empty.
8178 Also updated memory leak detection code to be multi-thread-safe.
8179 [Richard Levitte]
8180
8181 *) Add options -text and -noout to pkcs7 utility and delete the
8182 encryption options which never did anything. Update docs.
8183 [Steve Henson]
8184
8185 *) Add options to some of the utilities to allow the pass phrase
8186 to be included on either the command line (not recommended on
8187 OSes like Unix) or read from the environment. Update the
8188 manpages and fix a few bugs.
8189 [Steve Henson]
8190
8191 *) Add a few manpages for some of the openssl commands.
8192 [Steve Henson]
8193
8194 *) Fix the -revoke option in ca. It was freeing up memory twice,
8195 leaking and not finding already revoked certificates.
8196 [Steve Henson]
8197
8198 *) Extensive changes to support certificate auxiliary information.
8199 This involves the use of X509_CERT_AUX structure and X509_AUX
8200 functions. An X509_AUX function such as PEM_read_X509_AUX()
8201 can still read in a certificate file in the usual way but it
8202 will also read in any additional "auxiliary information". By
8203 doing things this way a fair degree of compatibility can be
8204 retained: existing certificates can have this information added
8205 using the new 'x509' options.
8206
8207 Current auxiliary information includes an "alias" and some trust
8208 settings. The trust settings will ultimately be used in enhanced
8209 certificate chain verification routines: currently a certificate
8210 can only be trusted if it is self signed and then it is trusted
8211 for all purposes.
8212 [Steve Henson]
8213
8214 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8215 The problem was that one of the replacement routines had not been working
8216 since SSLeay releases. For now the offending routine has been replaced
8217 with non-optimised assembler. Even so, this now gives around 95%
8218 performance improvement for 1024 bit RSA signs.
8219 [Mark Cox]
8220
8221 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8222 handling. Most clients have the effective key size in bits equal to
8223 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8224 A few however don't do this and instead use the size of the decrypted key
8225 to determine the RC2 key length and the AlgorithmIdentifier to determine
8226 the effective key length. In this case the effective key length can still
8227 be 40 bits but the key length can be 168 bits for example. This is fixed
8228 by manually forcing an RC2 key into the EVP_PKEY structure because the
8229 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8230 the key length and effective key length are equal.
8231 [Steve Henson]
8232
8233 *) Add a bunch of functions that should simplify the creation of
8234 X509_NAME structures. Now you should be able to do:
8235 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8236 and have it automatically work out the correct field type and fill in
8237 the structures. The more adventurous can try:
8238 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8239 and it will (hopefully) work out the correct multibyte encoding.
8240 [Steve Henson]
8241
8242 *) Change the 'req' utility to use the new field handling and multibyte
8243 copy routines. Before the DN field creation was handled in an ad hoc
8244 way in req, ca, and x509 which was rather broken and didn't support
8245 BMPStrings or UTF8Strings. Since some software doesn't implement
8246 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8247 using the dirstring_type option. See the new comment in the default
8248 openssl.cnf for more info.
8249 [Steve Henson]
8250
8251 *) Make crypto/rand/md_rand.c more robust:
8252 - Assure unique random numbers after fork().
8253 - Make sure that concurrent threads access the global counter and
8254 md serializably so that we never lose entropy in them
8255 or use exactly the same state in multiple threads.
8256 Access to the large state is not always serializable because
8257 the additional locking could be a performance killer, and
8258 md should be large enough anyway.
8259 [Bodo Moeller]
8260
8261 *) New file apps/app_rand.c with commonly needed functionality
8262 for handling the random seed file.
8263
8264 Use the random seed file in some applications that previously did not:
8265 ca,
8266 dsaparam -genkey (which also ignored its '-rand' option),
8267 s_client,
8268 s_server,
8269 x509 (when signing).
8270 Except on systems with /dev/urandom, it is crucial to have a random
8271 seed file at least for key creation, DSA signing, and for DH exchanges;
8272 for RSA signatures we could do without one.
8273
8274 gendh and gendsa (unlike genrsa) used to read only the first byte
8275 of each file listed in the '-rand' option. The function as previously
8276 found in genrsa is now in app_rand.c and is used by all programs
8277 that support '-rand'.
8278 [Bodo Moeller]
8279
8280 *) In RAND_write_file, use mode 0600 for creating files;
8281 don't just chmod when it may be too late.
8282 [Bodo Moeller]
8283
8284 *) Report an error from X509_STORE_load_locations
8285 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8286 [Bill Perry]
8287
8288 *) New function ASN1_mbstring_copy() this copies a string in either
8289 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8290 into an ASN1_STRING type. A mask of permissible types is passed
8291 and it chooses the "minimal" type to use or an error if not type
8292 is suitable.
8293 [Steve Henson]
8294
8295 *) Add function equivalents to the various macros in asn1.h. The old
8296 macros are retained with an M_ prefix. Code inside the library can
8297 use the M_ macros. External code (including the openssl utility)
8298 should *NOT* in order to be "shared library friendly".
8299 [Steve Henson]
8300
8301 *) Add various functions that can check a certificate's extensions
8302 to see if it usable for various purposes such as SSL client,
8303 server or S/MIME and CAs of these types. This is currently
8304 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8305 verification. Also added a -purpose flag to x509 utility to
8306 print out all the purposes.
8307 [Steve Henson]
8308
8309 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8310 functions.
8311 [Steve Henson]
8312
8313 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8314 for, obtain and decode and extension and obtain its critical flag.
8315 This allows all the necessary extension code to be handled in a
8316 single function call.
8317 [Steve Henson]
8318
8319 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8320 platforms. See crypto/rc4/rc4_enc.c for further details.
8321 [Andy Polyakov]
8322
8323 *) New -noout option to asn1parse. This causes no output to be produced
8324 its main use is when combined with -strparse and -out to extract data
8325 from a file (which may not be in ASN.1 format).
8326 [Steve Henson]
8327
8328 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8329 when producing the local key id.
8330 [Richard Levitte <levitte@stacken.kth.se>]
8331
8332 *) New option -dhparam in s_server. This allows a DH parameter file to be
8333 stated explicitly. If it is not stated then it tries the first server
8334 certificate file. The previous behaviour hard coded the filename
8335 "server.pem".
8336 [Steve Henson]
8337
8338 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8339 a public key to be input or output. For example:
8340 openssl rsa -in key.pem -pubout -out pubkey.pem
8341 Also added necessary DSA public key functions to handle this.
8342 [Steve Henson]
8343
8344 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8345 in the message. This was handled by allowing
8346 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8347 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8348
8349 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8350 to the end of the strings whereas this didn't. This would cause problems
8351 if strings read with d2i_ASN1_bytes() were later modified.
8352 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8353
8354 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8355 data and it contains EOF it will end up returning an error. This is
8356 caused by input 46 bytes long. The cause is due to the way base64
8357 BIOs find the start of base64 encoded data. They do this by trying a
8358 trial decode on each line until they find one that works. When they
8359 do a flag is set and it starts again knowing it can pass all the
8360 data directly through the decoder. Unfortunately it doesn't reset
8361 the context it uses. This means that if EOF is reached an attempt
8362 is made to pass two EOFs through the context and this causes the
8363 resulting error. This can also cause other problems as well. As is
8364 usual with these problems it takes *ages* to find and the fix is
8365 trivial: move one line.
8366 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8367
8368 *) Ugly workaround to get s_client and s_server working under Windows. The
8369 old code wouldn't work because it needed to select() on sockets and the
8370 tty (for keypresses and to see if data could be written). Win32 only
8371 supports select() on sockets so we select() with a 1s timeout on the
8372 sockets and then see if any characters are waiting to be read, if none
8373 are present then we retry, we also assume we can always write data to
8374 the tty. This isn't nice because the code then blocks until we've
8375 received a complete line of data and it is effectively polling the
8376 keyboard at 1s intervals: however it's quite a bit better than not
8377 working at all :-) A dedicated Windows application might handle this
8378 with an event loop for example.
8379 [Steve Henson]
8380
8381 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8382 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8383 will be called when RSA_sign() and RSA_verify() are used. This is useful
8384 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8385 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8386 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8387 This necessitated the support of an extra signature type NID_md5_sha1
8388 for SSL signatures and modifications to the SSL library to use it instead
8389 of calling RSA_public_decrypt() and RSA_private_encrypt().
8390 [Steve Henson]
8391
8392 *) Add new -verify -CAfile and -CApath options to the crl program, these
8393 will lookup a CRL issuers certificate and verify the signature in a
8394 similar way to the verify program. Tidy up the crl program so it
8395 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8396 less strict. It will now permit CRL extensions even if it is not
8397 a V2 CRL: this will allow it to tolerate some broken CRLs.
8398 [Steve Henson]
8399
8400 *) Initialize all non-automatic variables each time one of the openssl
8401 sub-programs is started (this is necessary as they may be started
8402 multiple times from the "OpenSSL>" prompt).
8403 [Lennart Bang, Bodo Moeller]
8404
8405 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8406 removing all other RSA functionality (this is what NO_RSA does). This
8407 is so (for example) those in the US can disable those operations covered
8408 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8409 key generation.
8410 [Steve Henson]
8411
8412 *) Non-copying interface to BIO pairs.
8413 (still largely untested)
8414 [Bodo Moeller]
8415
8416 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8417 ASCII string. This was handled independently in various places before.
8418 [Steve Henson]
8419
8420 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8421 UTF8 strings a character at a time.
8422 [Steve Henson]
8423
8424 *) Use client_version from client hello to select the protocol
8425 (s23_srvr.c) and for RSA client key exchange verification
8426 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8427 [Bodo Moeller]
8428
8429 *) Add various utility functions to handle SPKACs, these were previously
8430 handled by poking round in the structure internals. Added new function
8431 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8432 print, verify and generate SPKACs. Based on an original idea from
8433 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8434 [Steve Henson]
8435
8436 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8437 [Andy Polyakov]
8438
8439 *) Allow the config file extension section to be overwritten on the
8440 command line. Based on an original idea from Massimiliano Pala
8441 <madwolf@comune.modena.it>. The new option is called -extensions
8442 and can be applied to ca, req and x509. Also -reqexts to override
8443 the request extensions in req and -crlexts to override the crl extensions
8444 in ca.
8445 [Steve Henson]
8446
8447 *) Add new feature to the SPKAC handling in ca. Now you can include
8448 the same field multiple times by preceding it by "XXXX." for example:
8449 1.OU="Unit name 1"
8450 2.OU="Unit name 2"
8451 this is the same syntax as used in the req config file.
8452 [Steve Henson]
8453
8454 *) Allow certificate extensions to be added to certificate requests. These
8455 are specified in a 'req_extensions' option of the req section of the
8456 config file. They can be printed out with the -text option to req but
8457 are otherwise ignored at present.
8458 [Steve Henson]
8459
8460 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8461 data read consists of only the final block it would not decrypted because
8462 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8463 A misplaced 'break' also meant the decrypted final block might not be
8464 copied until the next read.
8465 [Steve Henson]
8466
8467 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8468 a few extra parameters to the DH structure: these will be useful if
8469 for example we want the value of 'q' or implement X9.42 DH.
8470 [Steve Henson]
8471
8472 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8473 provides hooks that allow the default DSA functions or functions on a
8474 "per key" basis to be replaced. This allows hardware acceleration and
8475 hardware key storage to be handled without major modification to the
8476 library. Also added low level modexp hooks and CRYPTO_EX structure and
8477 associated functions.
8478 [Steve Henson]
8479
8480 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8481 as "read only": it can't be written to and the buffer it points to will
8482 not be freed. Reading from a read only BIO is much more efficient than
8483 a normal memory BIO. This was added because there are several times when
8484 an area of memory needs to be read from a BIO. The previous method was
8485 to create a memory BIO and write the data to it, this results in two
8486 copies of the data and an O(n^2) reading algorithm. There is a new
8487 function BIO_new_mem_buf() which creates a read only memory BIO from
8488 an area of memory. Also modified the PKCS#7 routines to use read only
8489 memory BIOs.
8490 [Steve Henson]
8491
8492 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8493 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8494 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8495 but a retry condition occured while trying to read the rest.
8496 [Bodo Moeller]
8497
8498 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8499 NID_pkcs7_encrypted by default: this was wrong since this should almost
8500 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8501 the encrypted data type: this is a more sensible place to put it and it
8502 allows the PKCS#12 code to be tidied up that duplicated this
8503 functionality.
8504 [Steve Henson]
8505
8506 *) Changed obj_dat.pl script so it takes its input and output files on
8507 the command line. This should avoid shell escape redirection problems
8508 under Win32.
8509 [Steve Henson]
8510
8511 *) Initial support for certificate extension requests, these are included
8512 in things like Xenroll certificate requests. Included functions to allow
8513 extensions to be obtained and added.
8514 [Steve Henson]
8515
8516 *) -crlf option to s_client and s_server for sending newlines as
8517 CRLF (as required by many protocols).
8518 [Bodo Moeller]
8519
8520 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8521
8522 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8523 [Ralf S. Engelschall]
8524
8525 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8526 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8527
8528 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8529 program.
8530 [Steve Henson]
8531
8532 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8533 DH parameters/keys (q is lost during that conversion, but the resulting
8534 DH parameters contain its length).
8535
8536 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8537 much faster than DH_generate_parameters (which creates parameters
8538 where p = 2*q + 1), and also the smaller q makes DH computations
8539 much more efficient (160-bit exponentiation instead of 1024-bit
8540 exponentiation); so this provides a convenient way to support DHE
8541 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8542 utter importance to use
8543 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8544 or
8545 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8546 when such DH parameters are used, because otherwise small subgroup
8547 attacks may become possible!
8548 [Bodo Moeller]
8549
8550 *) Avoid memory leak in i2d_DHparams.
8551 [Bodo Moeller]
8552
8553 *) Allow the -k option to be used more than once in the enc program:
8554 this allows the same encrypted message to be read by multiple recipients.
8555 [Steve Henson]
8556
8557 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8558 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8559 it will always use the numerical form of the OID, even if it has a short
8560 or long name.
8561 [Steve Henson]
8562
8563 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8564 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8565 otherwise bn_mod_exp was called. In the case of hardware keys for example
8566 no private key components need be present and it might store extra data
8567 in the RSA structure, which cannot be accessed from bn_mod_exp.
8568 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8569 private key operations.
8570 [Steve Henson]
8571
8572 *) Added support for SPARC Linux.
8573 [Andy Polyakov]
8574
8575 *) pem_password_cb function type incompatibly changed from
8576 typedef int pem_password_cb(char *buf, int size, int rwflag);
8577 to
8578 ....(char *buf, int size, int rwflag, void *userdata);
8579 so that applications can pass data to their callbacks:
8580 The PEM[_ASN1]_{read,write}... functions and macros now take an
8581 additional void * argument, which is just handed through whenever
8582 the password callback is called.
8583 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8584
8585 New function SSL_CTX_set_default_passwd_cb_userdata.
8586
8587 Compatibility note: As many C implementations push function arguments
8588 onto the stack in reverse order, the new library version is likely to
8589 interoperate with programs that have been compiled with the old
8590 pem_password_cb definition (PEM_whatever takes some data that
8591 happens to be on the stack as its last argument, and the callback
8592 just ignores this garbage); but there is no guarantee whatsoever that
8593 this will work.
8594
8595 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8596 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8597 problems not only on Windows, but also on some Unix platforms.
8598 To avoid problematic command lines, these definitions are now in an
8599 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8600 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8601 [Bodo Moeller]
8602
8603 *) MIPS III/IV assembler module is reimplemented.
8604 [Andy Polyakov]
8605
8606 *) More DES library cleanups: remove references to srand/rand and
8607 delete an unused file.
8608 [Ulf Möller]
8609
8610 *) Add support for the the free Netwide assembler (NASM) under Win32,
8611 since not many people have MASM (ml) and it can be hard to obtain.
8612 This is currently experimental but it seems to work OK and pass all
8613 the tests. Check out INSTALL.W32 for info.
8614 [Steve Henson]
8615
8616 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8617 without temporary keys kept an extra copy of the server key,
8618 and connections with temporary keys did not free everything in case
8619 of an error.
8620 [Bodo Moeller]
8621
8622 *) New function RSA_check_key and new openssl rsa option -check
8623 for verifying the consistency of RSA keys.
8624 [Ulf Moeller, Bodo Moeller]
8625
8626 *) Various changes to make Win32 compile work:
8627 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8628 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8629 comparison" warnings.
8630 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8631 [Steve Henson]
8632
8633 *) Add a debugging option to PKCS#5 v2 key generation function: when
8634 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8635 derived keys are printed to stderr.
8636 [Steve Henson]
8637
8638 *) Copy the flags in ASN1_STRING_dup().
8639 [Roman E. Pavlov <pre@mo.msk.ru>]
8640
8641 *) The x509 application mishandled signing requests containing DSA
8642 keys when the signing key was also DSA and the parameters didn't match.
8643
8644 It was supposed to omit the parameters when they matched the signing key:
8645 the verifying software was then supposed to automatically use the CA's
8646 parameters if they were absent from the end user certificate.
8647
8648 Omitting parameters is no longer recommended. The test was also
8649 the wrong way round! This was probably due to unusual behaviour in
8650 EVP_cmp_parameters() which returns 1 if the parameters match.
8651 This meant that parameters were omitted when they *didn't* match and
8652 the certificate was useless. Certificates signed with 'ca' didn't have
8653 this bug.
8654 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8655
8656 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8657 The interface is as follows:
8658 Applications can use
8659 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8660 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8661 "off" is now the default.
8662 The library internally uses
8663 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8664 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8665 to disable memory-checking temporarily.
8666
8667 Some inconsistent states that previously were possible (and were
8668 even the default) are now avoided.
8669
8670 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8671 with each memory chunk allocated; this is occasionally more helpful
8672 than just having a counter.
8673
8674 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8675
8676 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8677 extensions.
8678 [Bodo Moeller]
8679
8680 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8681 which largely parallels "options", but is for changing API behaviour,
8682 whereas "options" are about protocol behaviour.
8683 Initial "mode" flags are:
8684
8685 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8686 a single record has been written.
8687 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8688 retries use the same buffer location.
8689 (But all of the contents must be
8690 copied!)
8691 [Bodo Moeller]
8692
8693 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8694 worked.
8695
8696 *) Fix problems with no-hmac etc.
8697 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8698
8699 *) New functions RSA_get_default_method(), RSA_set_method() and
8700 RSA_get_method(). These allows replacement of RSA_METHODs without having
8701 to mess around with the internals of an RSA structure.
8702 [Steve Henson]
8703
8704 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8705 Also really enable memory leak checks in openssl.c and in some
8706 test programs.
8707 [Chad C. Mulligan, Bodo Moeller]
8708
8709 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8710 up the length of negative integers. This has now been simplified to just
8711 store the length when it is first determined and use it later, rather
8712 than trying to keep track of where data is copied and updating it to
8713 point to the end.
8714 [Steve Henson, reported by Brien Wheeler
8715 <bwheeler@authentica-security.com>]
8716
8717 *) Add a new function PKCS7_signatureVerify. This allows the verification
8718 of a PKCS#7 signature but with the signing certificate passed to the
8719 function itself. This contrasts with PKCS7_dataVerify which assumes the
8720 certificate is present in the PKCS#7 structure. This isn't always the
8721 case: certificates can be omitted from a PKCS#7 structure and be
8722 distributed by "out of band" means (such as a certificate database).
8723 [Steve Henson]
8724
8725 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8726 function prototypes in pem.h, also change util/mkdef.pl to add the
8727 necessary function names.
8728 [Steve Henson]
8729
8730 *) mk1mf.pl (used by Windows builds) did not properly read the
8731 options set by Configure in the top level Makefile, and Configure
8732 was not even able to write more than one option correctly.
8733 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8734 [Bodo Moeller]
8735
8736 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8737 file to be loaded from a BIO or FILE pointer. The BIO version will
8738 for example allow memory BIOs to contain config info.
8739 [Steve Henson]
8740
8741 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8742 Whoever hopes to achieve shared-library compatibility across versions
8743 must use this, not the compile-time macro.
8744 (Exercise 0.9.4: Which is the minimum library version required by
8745 such programs?)
8746 Note: All this applies only to multi-threaded programs, others don't
8747 need locks.
8748 [Bodo Moeller]
8749
8750 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8751 through a BIO pair triggered the default case, i.e.
8752 SSLerr(...,SSL_R_UNKNOWN_STATE).
8753 [Bodo Moeller]
8754
8755 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8756 can use the SSL library even if none of the specific BIOs is
8757 appropriate.
8758 [Bodo Moeller]
8759
8760 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8761 for the encoded length.
8762 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8763
8764 *) Add initial documentation of the X509V3 functions.
8765 [Steve Henson]
8766
8767 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8768 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8769 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8770 secure PKCS#8 private key format with a high iteration count.
8771 [Steve Henson]
8772
8773 *) Fix determination of Perl interpreter: A perl or perl5
8774 _directory_ in $PATH was also accepted as the interpreter.
8775 [Ralf S. Engelschall]
8776
8777 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8778 wrong with it but it was very old and did things like calling
8779 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8780 unusual formatting.
8781 [Steve Henson]
8782
8783 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8784 to use the new extension code.
8785 [Steve Henson]
8786
8787 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8788 with macros. This should make it easier to change their form, add extra
8789 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8790 constant.
8791 [Steve Henson]
8792
8793 *) Add to configuration table a new entry that can specify an alternative
8794 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8795 according to Mark Crispin <MRC@Panda.COM>.
8796 [Bodo Moeller]
8797
8798 #if 0
8799 *) DES CBC did not update the IV. Weird.
8800 [Ben Laurie]
8801 #else
8802 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8803 Changing the behaviour of the former might break existing programs --
8804 where IV updating is needed, des_ncbc_encrypt can be used.
8805 #endif
8806
8807 *) When bntest is run from "make test" it drives bc to check its
8808 calculations, as well as internally checking them. If an internal check
8809 fails, it needs to cause bc to give a non-zero result or make test carries
8810 on without noticing the failure. Fixed.
8811 [Ben Laurie]
8812
8813 *) DES library cleanups.
8814 [Ulf Möller]
8815
8816 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8817 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8818 ciphers. NOTE: although the key derivation function has been verified
8819 against some published test vectors it has not been extensively tested
8820 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8821 of v2.0.
8822 [Steve Henson]
8823
8824 *) Instead of "mkdir -p", which is not fully portable, use new
8825 Perl script "util/mkdir-p.pl".
8826 [Bodo Moeller]
8827
8828 *) Rewrite the way password based encryption (PBE) is handled. It used to
8829 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8830 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8831 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8832 the 'parameter' field of the AlgorithmIdentifier is passed to the
8833 underlying key generation function so it must do its own ASN1 parsing.
8834 This has also changed the EVP_PBE_CipherInit() function which now has a
8835 'parameter' argument instead of literal salt and iteration count values
8836 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8837 [Steve Henson]
8838
8839 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8840 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8841 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8842 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8843 value was just used as a "magic string" and not used directly its
8844 value doesn't matter.
8845 [Steve Henson]
8846
8847 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8848 support mutable.
8849 [Ben Laurie]
8850
8851 *) "linux-sparc64" configuration (ultrapenguin).
8852 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8853 "linux-sparc" configuration.
8854 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8855
8856 *) config now generates no-xxx options for missing ciphers.
8857 [Ulf Möller]
8858
8859 *) Support the EBCDIC character set (work in progress).
8860 File ebcdic.c not yet included because it has a different license.
8861 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8862
8863 *) Support BS2000/OSD-POSIX.
8864 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8865
8866 *) Make callbacks for key generation use void * instead of char *.
8867 [Ben Laurie]
8868
8869 *) Make S/MIME samples compile (not yet tested).
8870 [Ben Laurie]
8871
8872 *) Additional typesafe stacks.
8873 [Ben Laurie]
8874
8875 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8876 [Bodo Moeller]
8877
8878
8879 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8880
8881 *) New configuration variant "sco5-gcc".
8882
8883 *) Updated some demos.
8884 [Sean O Riordain, Wade Scholine]
8885
8886 *) Add missing BIO_free at exit of pkcs12 application.
8887 [Wu Zhigang]
8888
8889 *) Fix memory leak in conf.c.
8890 [Steve Henson]
8891
8892 *) Updates for Win32 to assembler version of MD5.
8893 [Steve Henson]
8894
8895 *) Set #! path to perl in apps/der_chop to where we found it
8896 instead of using a fixed path.
8897 [Bodo Moeller]
8898
8899 *) SHA library changes for irix64-mips4-cc.
8900 [Andy Polyakov]
8901
8902 *) Improvements for VMS support.
8903 [Richard Levitte]
8904
8905
8906 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8907
8908 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8909 This also avoids the problems with SC4.2 and unpatched SC5.
8910 [Andy Polyakov <appro@fy.chalmers.se>]
8911
8912 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8913 These are required because of the typesafe stack would otherwise break
8914 existing code. If old code used a structure member which used to be STACK
8915 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8916 sk_num or sk_value it would produce an error because the num, data members
8917 are not present in STACK_OF. Now it just produces a warning. sk_set
8918 replaces the old method of assigning a value to sk_value
8919 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8920 that does this will no longer work (and should use sk_set instead) but
8921 this could be regarded as a "questionable" behaviour anyway.
8922 [Steve Henson]
8923
8924 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8925 correctly handle encrypted S/MIME data.
8926 [Steve Henson]
8927
8928 *) Change type of various DES function arguments from des_cblock
8929 (which means, in function argument declarations, pointer to char)
8930 to des_cblock * (meaning pointer to array with 8 char elements),
8931 which allows the compiler to do more typechecking; it was like
8932 that back in SSLeay, but with lots of ugly casts.
8933
8934 Introduce new type const_des_cblock.
8935 [Bodo Moeller]
8936
8937 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8938 problems: find RecipientInfo structure that matches recipient certificate
8939 and initialise the ASN1 structures properly based on passed cipher.
8940 [Steve Henson]
8941
8942 *) Belatedly make the BN tests actually check the results.
8943 [Ben Laurie]
8944
8945 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8946 to and from BNs: it was completely broken. New compilation option
8947 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8948 key elements as negative integers.
8949 [Steve Henson]
8950
8951 *) Reorganize and speed up MD5.
8952 [Andy Polyakov <appro@fy.chalmers.se>]
8953
8954 *) VMS support.
8955 [Richard Levitte <richard@levitte.org>]
8956
8957 *) New option -out to asn1parse to allow the parsed structure to be
8958 output to a file. This is most useful when combined with the -strparse
8959 option to examine the output of things like OCTET STRINGS.
8960 [Steve Henson]
8961
8962 *) Make SSL library a little more fool-proof by not requiring any longer
8963 that SSL_set_{accept,connect}_state be called before
8964 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8965 in many applications because usually everything *appeared* to work as
8966 intended anyway -- now it really works as intended).
8967 [Bodo Moeller]
8968
8969 *) Move openssl.cnf out of lib/.
8970 [Ulf Möller]
8971
8972 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8973 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8974 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8975 [Ralf S. Engelschall]
8976
8977 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8978 handle PKCS#7 enveloped data properly.
8979 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8980
8981 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8982 copying pointers. The cert_st handling is changed by this in
8983 various ways (and thus what used to be known as ctx->default_cert
8984 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8985 any longer when s->cert does not give us what we need).
8986 ssl_cert_instantiate becomes obsolete by this change.
8987 As soon as we've got the new code right (possibly it already is?),
8988 we have solved a couple of bugs of the earlier code where s->cert
8989 was used as if it could not have been shared with other SSL structures.
8990
8991 Note that using the SSL API in certain dirty ways now will result
8992 in different behaviour than observed with earlier library versions:
8993 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8994 does not influence s as it used to.
8995
8996 In order to clean up things more thoroughly, inside SSL_SESSION
8997 we don't use CERT any longer, but a new structure SESS_CERT
8998 that holds per-session data (if available); currently, this is
8999 the peer's certificate chain and, for clients, the server's certificate
9000 and temporary key. CERT holds only those values that can have
9001 meaningful defaults in an SSL_CTX.
9002 [Bodo Moeller]
9003
9004 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9005 from the internal representation. Various PKCS#7 fixes: remove some
9006 evil casts and set the enc_dig_alg field properly based on the signing
9007 key type.
9008 [Steve Henson]
9009
9010 *) Allow PKCS#12 password to be set from the command line or the
9011 environment. Let 'ca' get its config file name from the environment
9012 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9013 and 'x509').
9014 [Steve Henson]
9015
9016 *) Allow certificate policies extension to use an IA5STRING for the
9017 organization field. This is contrary to the PKIX definition but
9018 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9019 extension option.
9020 [Steve Henson]
9021
9022 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9023 without disallowing inline assembler and the like for non-pedantic builds.
9024 [Ben Laurie]
9025
9026 *) Support Borland C++ builder.
9027 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9028
9029 *) Support Mingw32.
9030 [Ulf Möller]
9031
9032 *) SHA-1 cleanups and performance enhancements.
9033 [Andy Polyakov <appro@fy.chalmers.se>]
9034
9035 *) Sparc v8plus assembler for the bignum library.
9036 [Andy Polyakov <appro@fy.chalmers.se>]
9037
9038 *) Accept any -xxx and +xxx compiler options in Configure.
9039 [Ulf Möller]
9040
9041 *) Update HPUX configuration.
9042 [Anonymous]
9043
9044 *) Add missing sk_<type>_unshift() function to safestack.h
9045 [Ralf S. Engelschall]
9046
9047 *) New function SSL_CTX_use_certificate_chain_file that sets the
9048 "extra_cert"s in addition to the certificate. (This makes sense
9049 only for "PEM" format files, as chains as a whole are not
9050 DER-encoded.)
9051 [Bodo Moeller]
9052
9053 *) Support verify_depth from the SSL API.
9054 x509_vfy.c had what can be considered an off-by-one-error:
9055 Its depth (which was not part of the external interface)
9056 was actually counting the number of certificates in a chain;
9057 now it really counts the depth.
9058 [Bodo Moeller]
9059
9060 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9061 instead of X509err, which often resulted in confusing error
9062 messages since the error codes are not globally unique
9063 (e.g. an alleged error in ssl3_accept when a certificate
9064 didn't match the private key).
9065
9066 *) New function SSL_CTX_set_session_id_context that allows to set a default
9067 value (so that you don't need SSL_set_session_id_context for each
9068 connection using the SSL_CTX).
9069 [Bodo Moeller]
9070
9071 *) OAEP decoding bug fix.
9072 [Ulf Möller]
9073
9074 *) Support INSTALL_PREFIX for package builders, as proposed by
9075 David Harris.
9076 [Bodo Moeller]
9077
9078 *) New Configure options "threads" and "no-threads". For systems
9079 where the proper compiler options are known (currently Solaris
9080 and Linux), "threads" is the default.
9081 [Bodo Moeller]
9082
9083 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9084 [Bodo Moeller]
9085
9086 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9087 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9088 such as /usr/local/bin.
9089 [Bodo Moeller]
9090
9091 *) "make linux-shared" to build shared libraries.
9092 [Niels Poppe <niels@netbox.org>]
9093
9094 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9095 [Ulf Möller]
9096
9097 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9098 extension adding in x509 utility.
9099 [Steve Henson]
9100
9101 *) Remove NOPROTO sections and error code comments.
9102 [Ulf Möller]
9103
9104 *) Partial rewrite of the DEF file generator to now parse the ANSI
9105 prototypes.
9106 [Steve Henson]
9107
9108 *) New Configure options --prefix=DIR and --openssldir=DIR.
9109 [Ulf Möller]
9110
9111 *) Complete rewrite of the error code script(s). It is all now handled
9112 by one script at the top level which handles error code gathering,
9113 header rewriting and C source file generation. It should be much better
9114 than the old method: it now uses a modified version of Ulf's parser to
9115 read the ANSI prototypes in all header files (thus the old K&R definitions
9116 aren't needed for error creation any more) and do a better job of
9117 translating function codes into names. The old 'ASN1 error code imbedded
9118 in a comment' is no longer necessary and it doesn't use .err files which
9119 have now been deleted. Also the error code call doesn't have to appear all
9120 on one line (which resulted in some large lines...).
9121 [Steve Henson]
9122
9123 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9124 [Bodo Moeller]
9125
9126 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9127 0 (which usually indicates a closed connection), but continue reading.
9128 [Bodo Moeller]
9129
9130 *) Fix some race conditions.
9131 [Bodo Moeller]
9132
9133 *) Add support for CRL distribution points extension. Add Certificate
9134 Policies and CRL distribution points documentation.
9135 [Steve Henson]
9136
9137 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9138 [Ulf Möller]
9139
9140 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9141 8 of keying material. Merlin has also confirmed interop with this fix
9142 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9143 [Merlin Hughes <merlin@baltimore.ie>]
9144
9145 *) Fix lots of warnings.
9146 [Richard Levitte <levitte@stacken.kth.se>]
9147
9148 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9149 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9150 [Richard Levitte <levitte@stacken.kth.se>]
9151
9152 *) Fix problems with sizeof(long) == 8.
9153 [Andy Polyakov <appro@fy.chalmers.se>]
9154
9155 *) Change functions to ANSI C.
9156 [Ulf Möller]
9157
9158 *) Fix typos in error codes.
9159 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9160
9161 *) Remove defunct assembler files from Configure.
9162 [Ulf Möller]
9163
9164 *) SPARC v8 assembler BIGNUM implementation.
9165 [Andy Polyakov <appro@fy.chalmers.se>]
9166
9167 *) Support for Certificate Policies extension: both print and set.
9168 Various additions to support the r2i method this uses.
9169 [Steve Henson]
9170
9171 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9172 return a const string when you are expecting an allocated buffer.
9173 [Ben Laurie]
9174
9175 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9176 types DirectoryString and DisplayText.
9177 [Steve Henson]
9178
9179 *) Add code to allow r2i extensions to access the configuration database,
9180 add an LHASH database driver and add several ctx helper functions.
9181 [Steve Henson]
9182
9183 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9184 fail when they extended the size of a BIGNUM.
9185 [Steve Henson]
9186
9187 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9188 support typesafe stack.
9189 [Steve Henson]
9190
9191 *) Fix typo in SSL_[gs]et_options().
9192 [Nils Frostberg <nils@medcom.se>]
9193
9194 *) Delete various functions and files that belonged to the (now obsolete)
9195 old X509V3 handling code.
9196 [Steve Henson]
9197
9198 *) New Configure option "rsaref".
9199 [Ulf Möller]
9200
9201 *) Don't auto-generate pem.h.
9202 [Bodo Moeller]
9203
9204 *) Introduce type-safe ASN.1 SETs.
9205 [Ben Laurie]
9206
9207 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9208 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9209
9210 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9211 that links with OpenSSL (well at least cause lots of warnings), but fear
9212 not: the conversion is trivial, and it eliminates loads of evil casts. A
9213 few STACKed things have been converted already. Feel free to convert more.
9214 In the fullness of time, I'll do away with the STACK type altogether.
9215 [Ben Laurie]
9216
9217 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9218 specified in <certfile> by updating the entry in the index.txt file.
9219 This way one no longer has to edit the index.txt file manually for
9220 revoking a certificate. The -revoke option does the gory details now.
9221 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9222
9223 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9224 `-text' option at all and this way the `-noout -text' combination was
9225 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9226 [Ralf S. Engelschall]
9227
9228 *) Make sure a corresponding plain text error message exists for the
9229 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9230 verify callback function determined that a certificate was revoked.
9231 [Ralf S. Engelschall]
9232
9233 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9234 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9235 all available cipers including rc5, which was forgotten until now.
9236 In order to let the testing shell script know which algorithms
9237 are available, a new (up to now undocumented) command
9238 "openssl list-cipher-commands" is used.
9239 [Bodo Moeller]
9240
9241 *) Bugfix: s_client occasionally would sleep in select() when
9242 it should have checked SSL_pending() first.
9243 [Bodo Moeller]
9244
9245 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9246 the raw DSA values prior to ASN.1 encoding.
9247 [Ulf Möller]
9248
9249 *) Tweaks to Configure
9250 [Niels Poppe <niels@netbox.org>]
9251
9252 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9253 yet...
9254 [Steve Henson]
9255
9256 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9257 [Ulf Möller]
9258
9259 *) New config option to avoid instructions that are illegal on the 80386.
9260 The default code is faster, but requires at least a 486.
9261 [Ulf Möller]
9262
9263 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9264 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9265 same as SSL2_VERSION anyway.
9266 [Bodo Moeller]
9267
9268 *) New "-showcerts" option for s_client.
9269 [Bodo Moeller]
9270
9271 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9272 application. Various cleanups and fixes.
9273 [Steve Henson]
9274
9275 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9276 modify error routines to work internally. Add error codes and PBE init
9277 to library startup routines.
9278 [Steve Henson]
9279
9280 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9281 packing functions to asn1 and evp. Changed function names and error
9282 codes along the way.
9283 [Steve Henson]
9284
9285 *) PKCS12 integration: and so it begins... First of several patches to
9286 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9287 objects to objects.h
9288 [Steve Henson]
9289
9290 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9291 and display support for Thawte strong extranet extension.
9292 [Steve Henson]
9293
9294 *) Add LinuxPPC support.
9295 [Jeff Dubrule <igor@pobox.org>]
9296
9297 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9298 bn_div_words in alpha.s.
9299 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9300
9301 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9302 OAEP isn't supported when OpenSSL is built with RSAref.
9303 [Ulf Moeller <ulf@fitug.de>]
9304
9305 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9306 so they no longer are missing under -DNOPROTO.
9307 [Soren S. Jorvang <soren@t.dk>]
9308
9309
9310 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9311
9312 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9313 doesn't work when the session is reused. Coming soon!
9314 [Ben Laurie]
9315
9316 *) Fix a security hole, that allows sessions to be reused in the wrong
9317 context thus bypassing client cert protection! All software that uses
9318 client certs and session caches in multiple contexts NEEDS PATCHING to
9319 allow session reuse! A fuller solution is in the works.
9320 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9321
9322 *) Some more source tree cleanups (removed obsolete files
9323 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9324 permission on "config" script to be executable) and a fix for the INSTALL
9325 document.
9326 [Ulf Moeller <ulf@fitug.de>]
9327
9328 *) Remove some legacy and erroneous uses of malloc, free instead of
9329 Malloc, Free.
9330 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9331
9332 *) Make rsa_oaep_test return non-zero on error.
9333 [Ulf Moeller <ulf@fitug.de>]
9334
9335 *) Add support for native Solaris shared libraries. Configure
9336 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9337 if someone would make that last step automatic.
9338 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9339
9340 *) ctx_size was not built with the right compiler during "make links". Fixed.
9341 [Ben Laurie]
9342
9343 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9344 except NULL ciphers". This means the default cipher list will no longer
9345 enable NULL ciphers. They need to be specifically enabled e.g. with
9346 the string "DEFAULT:eNULL".
9347 [Steve Henson]
9348
9349 *) Fix to RSA private encryption routines: if p < q then it would
9350 occasionally produce an invalid result. This will only happen with
9351 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9352 [Steve Henson]
9353
9354 *) Be less restrictive and allow also `perl util/perlpath.pl
9355 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9356 because this way one can also use an interpreter named `perl5' (which is
9357 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9358 installed as `perl').
9359 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9360
9361 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9362 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9363
9364 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9365 advapi32.lib to Win32 build and change the pem test comparision
9366 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9367 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9368 and crypto/des/ede_cbcm_enc.c.
9369 [Steve Henson]
9370
9371 *) DES quad checksum was broken on big-endian architectures. Fixed.
9372 [Ben Laurie]
9373
9374 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9375 Win32 test batch file so it (might) work again. The Win32 test batch file
9376 is horrible: I feel ill....
9377 [Steve Henson]
9378
9379 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9380 in e_os.h. Audit of header files to check ANSI and non ANSI
9381 sections: 10 functions were absent from non ANSI section and not exported
9382 from Windows DLLs. Fixed up libeay.num for new functions.
9383 [Steve Henson]
9384
9385 *) Make `openssl version' output lines consistent.
9386 [Ralf S. Engelschall]
9387
9388 *) Fix Win32 symbol export lists for BIO functions: Added
9389 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9390 to ms/libeay{16,32}.def.
9391 [Ralf S. Engelschall]
9392
9393 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9394 fine under Unix and passes some trivial tests I've now added. But the
9395 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9396 added to make sure no one expects that this stuff really works in the
9397 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9398 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9399 openssl_bio.xs.
9400 [Ralf S. Engelschall]
9401
9402 *) Fix the generation of two part addresses in perl.
9403 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9404
9405 *) Add config entry for Linux on MIPS.
9406 [John Tobey <jtobey@channel1.com>]
9407
9408 *) Make links whenever Configure is run, unless we are on Windoze.
9409 [Ben Laurie]
9410
9411 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9412 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9413 in CRLs.
9414 [Steve Henson]
9415
9416 *) Add a useful kludge to allow package maintainers to specify compiler and
9417 other platforms details on the command line without having to patch the
9418 Configure script everytime: One now can use ``perl Configure
9419 <id>:<details>'', i.e. platform ids are allowed to have details appended
9420 to them (seperated by colons). This is treated as there would be a static
9421 pre-configured entry in Configure's %table under key <id> with value
9422 <details> and ``perl Configure <id>'' is called. So, when you want to
9423 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9424 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9425 now, which overrides the FreeBSD-elf entry on-the-fly.
9426 [Ralf S. Engelschall]
9427
9428 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9429 [Ben Laurie]
9430
9431 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9432 on the `perl Configure ...' command line. This way one can compile
9433 OpenSSL libraries with Position Independent Code (PIC) which is needed
9434 for linking it into DSOs.
9435 [Ralf S. Engelschall]
9436
9437 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9438 Fixed.
9439 [Ben Laurie]
9440
9441 *) Cleaned up the LICENSE document: The official contact for any license
9442 questions now is the OpenSSL core team under openssl-core@openssl.org.
9443 And add a paragraph about the dual-license situation to make sure people
9444 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9445 to the OpenSSL toolkit.
9446 [Ralf S. Engelschall]
9447
9448 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9449 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9450 Additonally cleaned up the `make links' target: Remove unnecessary
9451 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9452 to speed processing and no longer clutter the display with confusing
9453 stuff. Instead only the actually done links are displayed.
9454 [Ralf S. Engelschall]
9455
9456 *) Permit null encryption ciphersuites, used for authentication only. It used
9457 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9458 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9459 encryption.
9460 [Ben Laurie]
9461
9462 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9463 signed attributes when verifying signatures (this would break them),
9464 the detached data encoding was wrong and public keys obtained using
9465 X509_get_pubkey() weren't freed.
9466 [Steve Henson]
9467
9468 *) Add text documentation for the BUFFER functions. Also added a work around
9469 to a Win95 console bug. This was triggered by the password read stuff: the
9470 last character typed gets carried over to the next fread(). If you were
9471 generating a new cert request using 'req' for example then the last
9472 character of the passphrase would be CR which would then enter the first
9473 field as blank.
9474 [Steve Henson]
9475
9476 *) Added the new `Includes OpenSSL Cryptography Software' button as
9477 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9478 button and can be used by applications based on OpenSSL to show the
9479 relationship to the OpenSSL project.
9480 [Ralf S. Engelschall]
9481
9482 *) Remove confusing variables in function signatures in files
9483 ssl/ssl_lib.c and ssl/ssl.h.
9484 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9485
9486 *) Don't install bss_file.c under PREFIX/include/
9487 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9488
9489 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9490 functions that return function pointers and has support for NT specific
9491 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9492 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9493 unsigned to signed types: this was killing the Win32 compile.
9494 [Steve Henson]
9495
9496 *) Add new certificate file to stack functions,
9497 SSL_add_dir_cert_subjects_to_stack() and
9498 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9499 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9500 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9501 This means that Apache-SSL and similar packages don't have to mess around
9502 to add as many CAs as they want to the preferred list.
9503 [Ben Laurie]
9504
9505 *) Experiment with doxygen documentation. Currently only partially applied to
9506 ssl/ssl_lib.c.
9507 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9508 openssl.doxy as the configuration file.
9509 [Ben Laurie]
9510
9511 *) Get rid of remaining C++-style comments which strict C compilers hate.
9512 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9513
9514 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9515 compiled in by default: it has problems with large keys.
9516 [Steve Henson]
9517
9518 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9519 DH private keys and/or callback functions which directly correspond to
9520 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9521 is needed for applications which have to configure certificates on a
9522 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9523 (e.g. s_server).
9524 For the RSA certificate situation is makes no difference, but
9525 for the DSA certificate situation this fixes the "no shared cipher"
9526 problem where the OpenSSL cipher selection procedure failed because the
9527 temporary keys were not overtaken from the context and the API provided
9528 no way to reconfigure them.
9529 The new functions now let applications reconfigure the stuff and they
9530 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9531 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9532 non-public-API function ssl_cert_instantiate() is used as a helper
9533 function and also to reduce code redundancy inside ssl_rsa.c.
9534 [Ralf S. Engelschall]
9535
9536 *) Move s_server -dcert and -dkey options out of the undocumented feature
9537 area because they are useful for the DSA situation and should be
9538 recognized by the users.
9539 [Ralf S. Engelschall]
9540
9541 *) Fix the cipher decision scheme for export ciphers: the export bits are
9542 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9543 SSL_EXP_MASK. So, the original variable has to be used instead of the
9544 already masked variable.
9545 [Richard Levitte <levitte@stacken.kth.se>]
9546
9547 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9548 [Richard Levitte <levitte@stacken.kth.se>]
9549
9550 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9551 from `int' to `unsigned int' because it's a length and initialized by
9552 EVP_DigestFinal() which expects an `unsigned int *'.
9553 [Richard Levitte <levitte@stacken.kth.se>]
9554
9555 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9556 script. Instead use the usual Shell->Perl transition trick.
9557 [Ralf S. Engelschall]
9558
9559 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9560 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9561 -noout -modulus' as it's already the case for `openssl rsa -noout
9562 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9563 currently the public key is printed (a decision which was already done by
9564 `openssl dsa -modulus' in the past) which serves a similar purpose.
9565 Additionally the NO_RSA no longer completely removes the whole -modulus
9566 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9567 now, too.
9568 [Ralf S. Engelschall]
9569
9570 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9571 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9572 [Arne Ansper <arne@ats.cyber.ee>]
9573
9574 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9575 to be added. Now both 'req' and 'ca' can use new objects defined in the
9576 config file.
9577 [Steve Henson]
9578
9579 *) Add cool BIO that does syslog (or event log on NT).
9580 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9581
9582 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9583 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9584 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9585 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9586 [Ben Laurie]
9587
9588 *) Add preliminary config info for new extension code.
9589 [Steve Henson]
9590
9591 *) Make RSA_NO_PADDING really use no padding.
9592 [Ulf Moeller <ulf@fitug.de>]
9593
9594 *) Generate errors when private/public key check is done.
9595 [Ben Laurie]
9596
9597 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9598 for some CRL extensions and new objects added.
9599 [Steve Henson]
9600
9601 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9602 key usage extension and fuller support for authority key id.
9603 [Steve Henson]
9604
9605 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9606 padding method for RSA, which is recommended for new applications in PKCS
9607 #1 v2.0 (RFC 2437, October 1998).
9608 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9609 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9610 against Bleichbacher's attack on RSA.
9611 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9612 Ben Laurie]
9613
9614 *) Updates to the new SSL compression code
9615 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9616
9617 *) Fix so that the version number in the master secret, when passed
9618 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9619 (because the server will not accept higher), that the version number
9620 is 0x03,0x01, not 0x03,0x00
9621 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9622
9623 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9624 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9625 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9626 [Steve Henson]
9627
9628 *) Support for RAW extensions where an arbitrary extension can be
9629 created by including its DER encoding. See apps/openssl.cnf for
9630 an example.
9631 [Steve Henson]
9632
9633 *) Make sure latest Perl versions don't interpret some generated C array
9634 code as Perl array code in the crypto/err/err_genc.pl script.
9635 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9636
9637 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9638 not many people have the assembler. Various Win32 compilation fixes and
9639 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9640 build instructions.
9641 [Steve Henson]
9642
9643 *) Modify configure script 'Configure' to automatically create crypto/date.h
9644 file under Win32 and also build pem.h from pem.org. New script
9645 util/mkfiles.pl to create the MINFO file on environments that can't do a
9646 'make files': perl util/mkfiles.pl >MINFO should work.
9647 [Steve Henson]
9648
9649 *) Major rework of DES function declarations, in the pursuit of correctness
9650 and purity. As a result, many evil casts evaporated, and some weirdness,
9651 too. You may find this causes warnings in your code. Zapping your evil
9652 casts will probably fix them. Mostly.
9653 [Ben Laurie]
9654
9655 *) Fix for a typo in asn1.h. Bug fix to object creation script
9656 obj_dat.pl. It considered a zero in an object definition to mean
9657 "end of object": none of the objects in objects.h have any zeros
9658 so it wasn't spotted.
9659 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9660
9661 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9662 Masking (CBCM). In the absence of test vectors, the best I have been able
9663 to do is check that the decrypt undoes the encrypt, so far. Send me test
9664 vectors if you have them.
9665 [Ben Laurie]
9666
9667 *) Correct calculation of key length for export ciphers (too much space was
9668 allocated for null ciphers). This has not been tested!
9669 [Ben Laurie]
9670
9671 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9672 message is now correct (it understands "crypto" and "ssl" on its
9673 command line). There is also now an "update" option. This will update
9674 the util/ssleay.num and util/libeay.num files with any new functions.
9675 If you do a:
9676 perl util/mkdef.pl crypto ssl update
9677 it will update them.
9678 [Steve Henson]
9679
9680 *) Overhauled the Perl interface (perl/*):
9681 - ported BN stuff to OpenSSL's different BN library
9682 - made the perl/ source tree CVS-aware
9683 - renamed the package from SSLeay to OpenSSL (the files still contain
9684 their history because I've copied them in the repository)
9685 - removed obsolete files (the test scripts will be replaced
9686 by better Test::Harness variants in the future)
9687 [Ralf S. Engelschall]
9688
9689 *) First cut for a very conservative source tree cleanup:
9690 1. merge various obsolete readme texts into doc/ssleay.txt
9691 where we collect the old documents and readme texts.
9692 2. remove the first part of files where I'm already sure that we no
9693 longer need them because of three reasons: either they are just temporary
9694 files which were left by Eric or they are preserved original files where
9695 I've verified that the diff is also available in the CVS via "cvs diff
9696 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9697 the crypto/md/ stuff).
9698 [Ralf S. Engelschall]
9699
9700 *) More extension code. Incomplete support for subject and issuer alt
9701 name, issuer and authority key id. Change the i2v function parameters
9702 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9703 what that's for :-) Fix to ASN1 macro which messed up
9704 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9705 [Steve Henson]
9706
9707 *) Preliminary support for ENUMERATED type. This is largely copied from the
9708 INTEGER code.
9709 [Steve Henson]
9710
9711 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9712 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9713
9714 *) Make sure `make rehash' target really finds the `openssl' program.
9715 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9716
9717 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9718 like to hear about it if this slows down other processors.
9719 [Ben Laurie]
9720
9721 *) Add CygWin32 platform information to Configure script.
9722 [Alan Batie <batie@aahz.jf.intel.com>]
9723
9724 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9725 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9726
9727 *) New program nseq to manipulate netscape certificate sequences
9728 [Steve Henson]
9729
9730 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9731 few typos.
9732 [Steve Henson]
9733
9734 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9735 but the BN code had some problems that would cause failures when
9736 doing certificate verification and some other functions.
9737 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9738
9739 *) Add ASN1 and PEM code to support netscape certificate sequences.
9740 [Steve Henson]
9741
9742 *) Add ASN1 and PEM code to support netscape certificate sequences.
9743 [Steve Henson]
9744
9745 *) Add several PKIX and private extended key usage OIDs.
9746 [Steve Henson]
9747
9748 *) Modify the 'ca' program to handle the new extension code. Modify
9749 openssl.cnf for new extension format, add comments.
9750 [Steve Henson]
9751
9752 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9753 and add a sample to openssl.cnf so req -x509 now adds appropriate
9754 CA extensions.
9755 [Steve Henson]
9756
9757 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9758 error code, add initial support to X509_print() and x509 application.
9759 [Steve Henson]
9760
9761 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9762 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9763 stuff is currently isolated and isn't even compiled yet.
9764 [Steve Henson]
9765
9766 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9767 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9768 Removed the versions check from X509 routines when loading extensions:
9769 this allows certain broken certificates that don't set the version
9770 properly to be processed.
9771 [Steve Henson]
9772
9773 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9774 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9775 can still be regenerated with "make depend".
9776 [Ben Laurie]
9777
9778 *) Spelling mistake in C version of CAST-128.
9779 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9780
9781 *) Changes to the error generation code. The perl script err-code.pl
9782 now reads in the old error codes and retains the old numbers, only
9783 adding new ones if necessary. It also only changes the .err files if new
9784 codes are added. The makefiles have been modified to only insert errors
9785 when needed (to avoid needlessly modifying header files). This is done
9786 by only inserting errors if the .err file is newer than the auto generated
9787 C file. To rebuild all the error codes from scratch (the old behaviour)
9788 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9789 or delete all the .err files.
9790 [Steve Henson]
9791
9792 *) CAST-128 was incorrectly implemented for short keys. The C version has
9793 been fixed, but is untested. The assembler versions are also fixed, but
9794 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9795 to regenerate it if needed.
9796 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9797 Hagino <itojun@kame.net>]
9798
9799 *) File was opened incorrectly in randfile.c.
9800 [Ulf Möller <ulf@fitug.de>]
9801
9802 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9803 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9804 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9805 al: it's just almost always a UTCTime. Note this patch adds new error
9806 codes so do a "make errors" if there are problems.
9807 [Steve Henson]
9808
9809 *) Correct Linux 1 recognition in config.
9810 [Ulf Möller <ulf@fitug.de>]
9811
9812 *) Remove pointless MD5 hash when using DSA keys in ca.
9813 [Anonymous <nobody@replay.com>]
9814
9815 *) Generate an error if given an empty string as a cert directory. Also
9816 generate an error if handed NULL (previously returned 0 to indicate an
9817 error, but didn't set one).
9818 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9819
9820 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9821 [Ben Laurie]
9822
9823 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9824 parameters. This was causing a warning which killed off the Win32 compile.
9825 [Steve Henson]
9826
9827 *) Remove C++ style comments from crypto/bn/bn_local.h.
9828 [Neil Costigan <neil.costigan@celocom.com>]
9829
9830 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9831 based on a text string, looking up short and long names and finally
9832 "dot" format. The "dot" format stuff didn't work. Added new function
9833 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9834 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9835 OID is not part of the table.
9836 [Steve Henson]
9837
9838 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9839 X509_LOOKUP_by_alias().
9840 [Ben Laurie]
9841
9842 *) Sort openssl functions by name.
9843 [Ben Laurie]
9844
9845 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9846 encryption from sample DSA keys (in case anyone is interested the password
9847 was "1234").
9848 [Steve Henson]
9849
9850 *) Make _all_ *_free functions accept a NULL pointer.
9851 [Frans Heymans <fheymans@isaserver.be>]
9852
9853 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9854 NULL pointers.
9855 [Anonymous <nobody@replay.com>]
9856
9857 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9858 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9859
9860 *) Don't blow it for numeric -newkey arguments to apps/req.
9861 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9862
9863 *) Temp key "for export" tests were wrong in s3_srvr.c.
9864 [Anonymous <nobody@replay.com>]
9865
9866 *) Add prototype for temp key callback functions
9867 SSL_CTX_set_tmp_{rsa,dh}_callback().
9868 [Ben Laurie]
9869
9870 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9871 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9872 [Steve Henson]
9873
9874 *) X509_name_add_entry() freed the wrong thing after an error.
9875 [Arne Ansper <arne@ats.cyber.ee>]
9876
9877 *) rsa_eay.c would attempt to free a NULL context.
9878 [Arne Ansper <arne@ats.cyber.ee>]
9879
9880 *) BIO_s_socket() had a broken should_retry() on Windoze.
9881 [Arne Ansper <arne@ats.cyber.ee>]
9882
9883 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9884 [Arne Ansper <arne@ats.cyber.ee>]
9885
9886 *) Make sure the already existing X509_STORE->depth variable is initialized
9887 in X509_STORE_new(), but document the fact that this variable is still
9888 unused in the certificate verification process.
9889 [Ralf S. Engelschall]
9890
9891 *) Fix the various library and apps files to free up pkeys obtained from
9892 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9893 [Steve Henson]
9894
9895 *) Fix reference counting in X509_PUBKEY_get(). This makes
9896 demos/maurice/example2.c work, amongst others, probably.
9897 [Steve Henson and Ben Laurie]
9898
9899 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9900 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9901 are no longer created. This way we have a single and consistent command
9902 line interface `openssl <command>', similar to `cvs <command>'.
9903 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9904
9905 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9906 BIT STRING wrapper always have zero unused bits.
9907 [Steve Henson]
9908
9909 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9910 [Steve Henson]
9911
9912 *) Make the top-level INSTALL documentation easier to understand.
9913 [Paul Sutton]
9914
9915 *) Makefiles updated to exit if an error occurs in a sub-directory
9916 make (including if user presses ^C) [Paul Sutton]
9917
9918 *) Make Montgomery context stuff explicit in RSA data structure.
9919 [Ben Laurie]
9920
9921 *) Fix build order of pem and err to allow for generated pem.h.
9922 [Ben Laurie]
9923
9924 *) Fix renumbering bug in X509_NAME_delete_entry().
9925 [Ben Laurie]
9926
9927 *) Enhanced the err-ins.pl script so it makes the error library number
9928 global and can add a library name. This is needed for external ASN1 and
9929 other error libraries.
9930 [Steve Henson]
9931
9932 *) Fixed sk_insert which never worked properly.
9933 [Steve Henson]
9934
9935 *) Fix ASN1 macros so they can handle indefinite length construted
9936 EXPLICIT tags. Some non standard certificates use these: they can now
9937 be read in.
9938 [Steve Henson]
9939
9940 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9941 into a single doc/ssleay.txt bundle. This way the information is still
9942 preserved but no longer messes up this directory. Now it's new room for
9943 the new set of documenation files.
9944 [Ralf S. Engelschall]
9945
9946 *) SETs were incorrectly DER encoded. This was a major pain, because they
9947 shared code with SEQUENCEs, which aren't coded the same. This means that
9948 almost everything to do with SETs or SEQUENCEs has either changed name or
9949 number of arguments.
9950 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9951
9952 *) Fix test data to work with the above.
9953 [Ben Laurie]
9954
9955 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9956 was already fixed by Eric for 0.9.1 it seems.
9957 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9958
9959 *) Autodetect FreeBSD3.
9960 [Ben Laurie]
9961
9962 *) Fix various bugs in Configure. This affects the following platforms:
9963 nextstep
9964 ncr-scde
9965 unixware-2.0
9966 unixware-2.0-pentium
9967 sco5-cc.
9968 [Ben Laurie]
9969
9970 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9971 before they are needed.
9972 [Ben Laurie]
9973
9974 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9975 [Ben Laurie]
9976
9977
9978 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9979
9980 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9981 changed SSLeay to OpenSSL in version strings.
9982 [Ralf S. Engelschall]
9983
9984 *) Some fixups to the top-level documents.
9985 [Paul Sutton]
9986
9987 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9988 because the symlink to include/ was missing.
9989 [Ralf S. Engelschall]
9990
9991 *) Incorporated the popular no-RSA/DSA-only patches
9992 which allow to compile a RSA-free SSLeay.
9993 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9994
9995 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9996 when "ssleay" is still not found.
9997 [Ralf S. Engelschall]
9998
9999 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10000 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10001
10002 *) Updated the README file.
10003 [Ralf S. Engelschall]
10004
10005 *) Added various .cvsignore files in the CVS repository subdirs
10006 to make a "cvs update" really silent.
10007 [Ralf S. Engelschall]
10008
10009 *) Recompiled the error-definition header files and added
10010 missing symbols to the Win32 linker tables.
10011 [Ralf S. Engelschall]
10012
10013 *) Cleaned up the top-level documents;
10014 o new files: CHANGES and LICENSE
10015 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10016 o merged COPYRIGHT into LICENSE
10017 o removed obsolete TODO file
10018 o renamed MICROSOFT to INSTALL.W32
10019 [Ralf S. Engelschall]
10020
10021 *) Removed dummy files from the 0.9.1b source tree:
10022 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10023 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10024 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10025 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10026 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10027 [Ralf S. Engelschall]
10028
10029 *) Added various platform portability fixes.
10030 [Mark J. Cox]
10031
10032 *) The Genesis of the OpenSSL rpject:
10033 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10034 Young and Tim J. Hudson created while they were working for C2Net until
10035 summer 1998.
10036 [The OpenSSL Project]
10037
10038
10039 Changes between 0.9.0b and 0.9.1b [not released]
10040
10041 *) Updated a few CA certificates under certs/
10042 [Eric A. Young]
10043
10044 *) Changed some BIGNUM api stuff.
10045 [Eric A. Young]
10046
10047 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10048 DGUX x86, Linux Alpha, etc.
10049 [Eric A. Young]
10050
10051 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10052 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10053 available).
10054 [Eric A. Young]
10055
10056 *) Add -strparse option to asn1pars program which parses nested
10057 binary structures
10058 [Dr Stephen Henson <shenson@bigfoot.com>]
10059
10060 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10061 [Eric A. Young]
10062
10063 *) DSA fix for "ca" program.
10064 [Eric A. Young]
10065
10066 *) Added "-genkey" option to "dsaparam" program.
10067 [Eric A. Young]
10068
10069 *) Added RIPE MD160 (rmd160) message digest.
10070 [Eric A. Young]
10071
10072 *) Added -a (all) option to "ssleay version" command.
10073 [Eric A. Young]
10074
10075 *) Added PLATFORM define which is the id given to Configure.
10076 [Eric A. Young]
10077
10078 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10079 [Eric A. Young]
10080
10081 *) Extended the ASN.1 parser routines.
10082 [Eric A. Young]
10083
10084 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10085 [Eric A. Young]
10086
10087 *) Added a BN_CTX to the BN library.
10088 [Eric A. Young]
10089
10090 *) Fixed the weak key values in DES library
10091 [Eric A. Young]
10092
10093 *) Changed API in EVP library for cipher aliases.
10094 [Eric A. Young]
10095
10096 *) Added support for RC2/64bit cipher.
10097 [Eric A. Young]
10098
10099 *) Converted the lhash library to the crypto/mem.c functions.
10100 [Eric A. Young]
10101
10102 *) Added more recognized ASN.1 object ids.
10103 [Eric A. Young]
10104
10105 *) Added more RSA padding checks for SSL/TLS.
10106 [Eric A. Young]
10107
10108 *) Added BIO proxy/filter functionality.
10109 [Eric A. Young]
10110
10111 *) Added extra_certs to SSL_CTX which can be used
10112 send extra CA certificates to the client in the CA cert chain sending
10113 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10114 [Eric A. Young]
10115
10116 *) Now Fortezza is denied in the authentication phase because
10117 this is key exchange mechanism is not supported by SSLeay at all.
10118 [Eric A. Young]
10119
10120 *) Additional PKCS1 checks.
10121 [Eric A. Young]
10122
10123 *) Support the string "TLSv1" for all TLS v1 ciphers.
10124 [Eric A. Young]
10125
10126 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10127 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10128 [Eric A. Young]
10129
10130 *) Fixed a few memory leaks.
10131 [Eric A. Young]
10132
10133 *) Fixed various code and comment typos.
10134 [Eric A. Young]
10135
10136 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10137 bytes sent in the client random.
10138 [Edward Bishop <ebishop@spyglass.com>]
10139