]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Add support for application defined signature algorithms for use with
[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 *) Add new functions to allow customised supported signature algorithms
8 for SSL and SSL_CTX structures. Add options to s_client and s_server
9 to support them.
10 [Steve Henson]
11
12 *) New function SSL_certs_clear() to delete all references to certificates
13 from an SSL structure. Before this once a certificate had been added
14 it couldn't be removed.
15 [Steve Henson]
16
17 *) Initial SSL tracing code. This parses out SSL/TLS records using the
18 message callback and prints the results. Needs compile time option
19 "enable-ssl-trace". New options to s_client and s_server to enable
20 tracing.
21 [Steve Henson]
22
23 *) New functions to retrieve certificate signature and signature
24 OID NID.
25 [Steve Henson]
26
27 *) Print out deprecated issuer and subject unique ID fields in
28 certificates.
29 [Steve Henson]
30
31 *) RFC 5878 support.
32 [Emilia Kasper, Adam Langley, Ben Laurie (Google)]
33
34 *) Support for automatic EC temporary key parameter selection. If enabled
35 the most preferred EC parameters are automatically used instead of
36 hardcoded fixed parameters. Now a server just has to call:
37 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
38 support ECDH and use the most appropriate parameters.
39 [Steve Henson]
40
41 *) Enhance and tidy EC curve and point format TLS extension code. Use
42 static structures instead of allocation if default values are used.
43 New ctrls to set curves we wish to support and to retrieve shared curves.
44 Print out shared curves in s_server. New options to s_server and s_client
45 to set list of supported curves.
46 [Steve Henson]
47
48 *) New ctrls to retrieve supported signature algorithms and
49 supported curve values as an array of NIDs. Extend openssl utility
50 to print out received values.
51 [Steve Henson]
52
53 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
54 between NIDs and the more common NIST names such as "P-256". Enhance
55 ecparam utility and ECC method to recognise the NIST names for curves.
56 [Steve Henson]
57
58 *) Enhance SSL/TLS certificate chain handling to support different
59 chains for each certificate instead of one chain in the parent SSL_CTX.
60 [Steve Henson]
61
62 *) Support for fixed DH ciphersuite client authentication: where both
63 server and client use DH certificates with common parameters.
64 [Steve Henson]
65
66 *) Support for fixed DH ciphersuites: those requiring DH server
67 certificates.
68 [Steve Henson]
69
70 *) Transparently support X9.42 DH parameters when calling
71 PEM_read_bio_DHparameters. This means existing applications can handle
72 the new parameter format automatically.
73 [Steve Henson]
74
75 *) Initial experimental support for X9.42 DH parameter format: mainly
76 to support use of 'q' parameter for RFC5114 parameters.
77 [Steve Henson]
78
79 *) Add DH parameters from RFC5114 including test data to dhtest.
80 [Steve Henson]
81
82 *) Update fips_test_suite to support multiple command line options. New
83 test to induce all self test errors in sequence and check expected
84 failures.
85 [Steve Henson]
86
87 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
88 sign or verify all in one operation.
89 [Steve Henson]
90
91 *) Add fips_algvs: a multicall fips utility incorporaing all the algorithm
92 test programs and fips_test_suite. Includes functionality to parse
93 the minimal script output of fipsalgest.pl directly.
94 [Steve Henson]
95
96 *) Add authorisation parameter to FIPS_module_mode_set().
97 [Steve Henson]
98
99 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
100 [Steve Henson]
101
102 *) Use separate DRBG fields for internal and external flags. New function
103 FIPS_drbg_health_check() to perform on demand health checking. Add
104 generation tests to fips_test_suite with reduced health check interval to
105 demonstrate periodic health checking. Add "nodh" option to
106 fips_test_suite to skip very slow DH test.
107 [Steve Henson]
108
109 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
110 based on NID.
111 [Steve Henson]
112
113 *) More extensive health check for DRBG checking many more failure modes.
114 New function FIPS_selftest_drbg_all() to handle every possible DRBG
115 combination: call this in fips_test_suite.
116 [Steve Henson]
117
118 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
119 and POST to handle Dual EC cases.
120 [Steve Henson]
121
122 *) Add support for canonical generation of DSA parameter 'g'. See
123 FIPS 186-3 A.2.3.
124
125 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
126 POST to handle HMAC cases.
127 [Steve Henson]
128
129 *) Add functions FIPS_module_version() and FIPS_module_version_text()
130 to return numberical and string versions of the FIPS module number.
131 [Steve Henson]
132
133 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
134 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
135 outside the validated module in the FIPS capable OpenSSL.
136 [Steve Henson]
137
138 *) Minor change to DRBG entropy callback semantics. In some cases
139 there is no mutiple of the block length between min_len and
140 max_len. Allow the callback to return more than max_len bytes
141 of entropy but discard any extra: it is the callback's responsibility
142 to ensure that the extra data discarded does not impact the
143 requested amount of entropy.
144 [Steve Henson]
145
146 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
147 information in FIPS186-3, SP800-57 and SP800-131A.
148 [Steve Henson]
149
150 *) CCM support via EVP. Interface is very similar to GCM case except we
151 must supply all data in one chunk (i.e. no update, final) and the
152 message length must be supplied if AAD is used. Add algorithm test
153 support.
154 [Steve Henson]
155
156 *) Initial version of POST overhaul. Add POST callback to allow the status
157 of POST to be monitored and/or failures induced. Modify fips_test_suite
158 to use callback. Always run all selftests even if one fails.
159 [Steve Henson]
160
161 *) XTS support including algorithm test driver in the fips_gcmtest program.
162 Note: this does increase the maximum key length from 32 to 64 bytes but
163 there should be no binary compatibility issues as existing applications
164 will never use XTS mode.
165 [Steve Henson]
166
167 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
168 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
169 performs algorithm blocking for unapproved PRNG types. Also do not
170 set PRNG type in FIPS_mode_set(): leave this to the application.
171 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
172 the standard OpenSSL PRNG: set additional data to a date time vector.
173 [Steve Henson]
174
175 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
176 This shouldn't present any incompatibility problems because applications
177 shouldn't be using these directly and any that are will need to rethink
178 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
179 [Steve Henson]
180
181 *) Extensive self tests and health checking required by SP800-90 DRBG.
182 Remove strength parameter from FIPS_drbg_instantiate and always
183 instantiate at maximum supported strength.
184 [Steve Henson]
185
186 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
187 [Steve Henson]
188
189 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
190 [Steve Henson]
191
192 *) New function DH_compute_key_padded() to compute a DH key and pad with
193 leading zeroes if needed: this complies with SP800-56A et al.
194 [Steve Henson]
195
196 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
197 anything, incomplete, subject to change and largely untested at present.
198 [Steve Henson]
199
200 *) Modify fipscanisteronly build option to only build the necessary object
201 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
202 [Steve Henson]
203
204 *) Add experimental option FIPSSYMS to give all symbols in
205 fipscanister.o and FIPS or fips prefix. This will avoid
206 conflicts with future versions of OpenSSL. Add perl script
207 util/fipsas.pl to preprocess assembly language source files
208 and rename any affected symbols.
209 [Steve Henson]
210
211 *) Add selftest checks and algorithm block of non-fips algorithms in
212 FIPS mode. Remove DES2 from selftests.
213 [Steve Henson]
214
215 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
216 return internal method without any ENGINE dependencies. Add new
217 tiny fips sign and verify functions.
218 [Steve Henson]
219
220 *) New build option no-ec2m to disable characteristic 2 code.
221 [Steve Henson]
222
223 *) New build option "fipscanisteronly". This only builds fipscanister.o
224 and (currently) associated fips utilities. Uses the file Makefile.fips
225 instead of Makefile.org as the prototype.
226 [Steve Henson]
227
228 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
229 Update fips_gcmtest to use IV generator.
230 [Steve Henson]
231
232 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
233 setting output buffer to NULL. The *Final function must be
234 called although it will not retrieve any additional data. The tag
235 can be set or retrieved with a ctrl. The IV length is by default 12
236 bytes (96 bits) but can be set to an alternative value. If the IV
237 length exceeds the maximum IV length (currently 16 bytes) it cannot be
238 set before the key.
239 [Steve Henson]
240
241 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
242 underlying do_cipher function handles all cipher semantics itself
243 including padding and finalisation. This is useful if (for example)
244 an ENGINE cipher handles block padding itself. The behaviour of
245 do_cipher is subtly changed if this flag is set: the return value
246 is the number of characters written to the output buffer (zero is
247 no longer an error code) or a negative error code. Also if the
248 input buffer is NULL and length 0 finalisation should be performed.
249 [Steve Henson]
250
251 *) If a candidate issuer certificate is already part of the constructed
252 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
253 [Steve Henson]
254
255 *) Improve forward-security support: add functions
256
257 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
258 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
259
260 for use by SSL/TLS servers; the callback function will be called whenever a
261 new session is created, and gets to decide whether the session may be
262 cached to make it resumable (return 0) or not (return 1). (As by the
263 SSL/TLS protocol specifications, the session_id sent by the server will be
264 empty to indicate that the session is not resumable; also, the server will
265 not generate RFC 4507 (RFC 5077) session tickets.)
266
267 A simple reasonable callback implementation is to return is_forward_secure.
268 This parameter will be set to 1 or 0 depending on the ciphersuite selected
269 by the SSL/TLS server library, indicating whether it can provide forward
270 security.
271 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
272
273 *) New function OPENSSL_gmtime_diff to find the difference in days
274 and seconds between two tm structures. This will be used to provide
275 additional functionality for ASN1_TIME.
276 [Steve Henson]
277
278 *) Add -trusted_first option which attempts to find certificates in the
279 trusted store even if an untrusted chain is also supplied.
280 [Steve Henson]
281
282 *) Initial experimental support for explicitly trusted non-root CAs.
283 OpenSSL still tries to build a complete chain to a root but if an
284 intermediate CA has a trust setting included that is used. The first
285 setting is used: whether to trust or reject.
286 [Steve Henson]
287
288 *) New -verify_name option in command line utilities to set verification
289 parameters by name.
290 [Steve Henson]
291
292 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
293 Add CMAC pkey methods.
294 [Steve Henson]
295
296 *) Experimental regnegotiation in s_server -www mode. If the client
297 browses /reneg connection is renegotiated. If /renegcert it is
298 renegotiated requesting a certificate.
299 [Steve Henson]
300
301 *) Add an "external" session cache for debugging purposes to s_server. This
302 should help trace issues which normally are only apparent in deployed
303 multi-process servers.
304 [Steve Henson]
305
306 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
307 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
308 BIO_set_cipher() and some obscure PEM functions were changed so they
309 can now return an error. The RAND changes required a change to the
310 RAND_METHOD structure.
311 [Steve Henson]
312
313 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
314 a gcc attribute to warn if the result of a function is ignored. This
315 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
316 whose return value is often ignored.
317 [Steve Henson]
318
319 Changes between 1.0.1c and 1.0.1d [xx XXX xxxx]
320
321 *) Fix possible deadlock when decoding public keys.
322 [Steve Henson]
323
324 *) Don't use TLS 1.0 record version number in initial client hello
325 if renegotiating.
326 [Steve Henson]
327
328 Changes between 1.0.1b and 1.0.1c [10 May 2012]
329
330 *) Sanity check record length before skipping explicit IV in TLS
331 1.2, 1.1 and DTLS to avoid DoS attack.
332
333 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
334 fuzzing as a service testing platform.
335 (CVE-2012-2333)
336 [Steve Henson]
337
338 *) Initialise tkeylen properly when encrypting CMS messages.
339 Thanks to Solar Designer of Openwall for reporting this issue.
340 [Steve Henson]
341
342 *) In FIPS mode don't try to use composite ciphers as they are not
343 approved.
344 [Steve Henson]
345
346 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
347
348 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
349 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
350 mean any application compiled against OpenSSL 1.0.0 headers setting
351 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
352 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
353 0x10000000L Any application which was previously compiled against
354 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
355 will need to be recompiled as a result. Letting be results in
356 inability to disable specifically TLS 1.1 and in client context,
357 in unlike event, limit maximum offered version to TLS 1.0 [see below].
358 [Steve Henson]
359
360 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
361 disable just protocol X, but all protocols above X *if* there are
362 protocols *below* X still enabled. In more practical terms it means
363 that if application wants to disable TLS1.0 in favor of TLS1.1 and
364 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
365 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
366 client side.
367 [Andy Polyakov]
368
369 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
370
371 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
372 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
373 in CRYPTO_realloc_clean.
374
375 Thanks to Tavis Ormandy, Google Security Team, for discovering this
376 issue and to Adam Langley <agl@chromium.org> for fixing it.
377 (CVE-2012-2110)
378 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
379
380 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
381 [Adam Langley]
382
383 *) Workarounds for some broken servers that "hang" if a client hello
384 record length exceeds 255 bytes:
385
386 1. Do not use record version number > TLS 1.0 in initial client
387 hello: some (but not all) hanging servers will now work.
388 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
389 the number of ciphers sent in the client hello. This should be
390 set to an even number, such as 50, for example by passing:
391 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
392 Most broken servers should now work.
393 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
394 TLS 1.2 client support entirely.
395 [Steve Henson]
396
397 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
398 [Andy Polyakov]
399
400 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
401
402 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
403 STRING form instead of a DigestInfo.
404 [Steve Henson]
405
406 *) The format used for MDC2 RSA signatures is inconsistent between EVP
407 and the RSA_sign/RSA_verify functions. This was made more apparent when
408 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
409 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
410 the correct format in RSA_verify so both forms transparently work.
411 [Steve Henson]
412
413 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
414 STRING form instead of a DigestInfo.
415 [Steve Henson]
416
417 *) Some servers which support TLS 1.0 can choke if we initially indicate
418 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
419 encrypted premaster secret. As a workaround use the maximum pemitted
420 client version in client hello, this should keep such servers happy
421 and still work with previous versions of OpenSSL.
422 [Steve Henson]
423
424 *) Add support for TLS/DTLS heartbeats.
425 [Robin Seggelmann <seggelmann@fh-muenster.de>]
426
427 *) Add support for SCTP.
428 [Robin Seggelmann <seggelmann@fh-muenster.de>]
429
430 *) Improved PRNG seeding for VOS.
431 [Paul Green <Paul.Green@stratus.com>]
432
433 *) Extensive assembler packs updates, most notably:
434
435 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
436 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
437 - x86_64: bit-sliced AES implementation;
438 - ARM: NEON support, contemporary platforms optimizations;
439 - s390x: z196 support;
440 - *: GHASH and GF(2^m) multiplication implementations;
441
442 [Andy Polyakov]
443
444 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
445 (removal of unnecessary code)
446 [Peter Sylvester <peter.sylvester@edelweb.fr>]
447
448 *) Add TLS key material exporter from RFC 5705.
449 [Eric Rescorla]
450
451 *) Add DTLS-SRTP negotiation from RFC 5764.
452 [Eric Rescorla]
453
454 *) Add Next Protocol Negotiation,
455 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
456 disabled with a no-npn flag to config or Configure. Code donated
457 by Google.
458 [Adam Langley <agl@google.com> and Ben Laurie]
459
460 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
461 NIST-P256, NIST-P521, with constant-time single point multiplication on
462 typical inputs. Compiler support for the nonstandard type __uint128_t is
463 required to use this (present in gcc 4.4 and later, for 64-bit builds).
464 Code made available under Apache License version 2.0.
465
466 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
467 line to include this in your build of OpenSSL, and run "make depend" (or
468 "make update"). This enables the following EC_METHODs:
469
470 EC_GFp_nistp224_method()
471 EC_GFp_nistp256_method()
472 EC_GFp_nistp521_method()
473
474 EC_GROUP_new_by_curve_name() will automatically use these (while
475 EC_GROUP_new_curve_GFp() currently prefers the more flexible
476 implementations).
477 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
478
479 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
480 all platforms. Move ssize_t definition from e_os.h to the public
481 header file e_os2.h as it now appears in public header file cms.h
482 [Steve Henson]
483
484 *) New -sigopt option to the ca, req and x509 utilities. Additional
485 signature parameters can be passed using this option and in
486 particular PSS.
487 [Steve Henson]
488
489 *) Add RSA PSS signing function. This will generate and set the
490 appropriate AlgorithmIdentifiers for PSS based on those in the
491 corresponding EVP_MD_CTX structure. No application support yet.
492 [Steve Henson]
493
494 *) Support for companion algorithm specific ASN1 signing routines.
495 New function ASN1_item_sign_ctx() signs a pre-initialised
496 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
497 the appropriate parameters.
498 [Steve Henson]
499
500 *) Add new algorithm specific ASN1 verification initialisation function
501 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
502 handling will be the same no matter what EVP_PKEY_METHOD is used.
503 Add a PSS handler to support verification of PSS signatures: checked
504 against a number of sample certificates.
505 [Steve Henson]
506
507 *) Add signature printing for PSS. Add PSS OIDs.
508 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
509
510 *) Add algorithm specific signature printing. An individual ASN1 method
511 can now print out signatures instead of the standard hex dump.
512
513 More complex signatures (e.g. PSS) can print out more meaningful
514 information. Include DSA version that prints out the signature
515 parameters r, s.
516 [Steve Henson]
517
518 *) Password based recipient info support for CMS library: implementing
519 RFC3211.
520 [Steve Henson]
521
522 *) Split password based encryption into PBES2 and PBKDF2 functions. This
523 neatly separates the code into cipher and PBE sections and is required
524 for some algorithms that split PBES2 into separate pieces (such as
525 password based CMS).
526 [Steve Henson]
527
528 *) Session-handling fixes:
529 - Fix handling of connections that are resuming with a session ID,
530 but also support Session Tickets.
531 - Fix a bug that suppressed issuing of a new ticket if the client
532 presented a ticket with an expired session.
533 - Try to set the ticket lifetime hint to something reasonable.
534 - Make tickets shorter by excluding irrelevant information.
535 - On the client side, don't ignore renewed tickets.
536 [Adam Langley, Bodo Moeller (Google)]
537
538 *) Fix PSK session representation.
539 [Bodo Moeller]
540
541 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
542
543 This work was sponsored by Intel.
544 [Andy Polyakov]
545
546 *) Add GCM support to TLS library. Some custom code is needed to split
547 the IV between the fixed (from PRF) and explicit (from TLS record)
548 portions. This adds all GCM ciphersuites supported by RFC5288 and
549 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
550 add a special AESGCM string for GCM only.
551 [Steve Henson]
552
553 *) Expand range of ctrls for AES GCM. Permit setting invocation
554 field on decrypt and retrieval of invocation field only on encrypt.
555 [Steve Henson]
556
557 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
558 As required by RFC5289 these ciphersuites cannot be used if for
559 versions of TLS earlier than 1.2.
560 [Steve Henson]
561
562 *) For FIPS capable OpenSSL interpret a NULL default public key method
563 as unset and return the appopriate default but do *not* set the default.
564 This means we can return the appopriate method in applications that
565 swicth between FIPS and non-FIPS modes.
566 [Steve Henson]
567
568 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
569 ENGINE is used then we cannot handle that in the FIPS module so we
570 keep original code iff non-FIPS operations are allowed.
571 [Steve Henson]
572
573 *) Add -attime option to openssl utilities.
574 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
575
576 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
577 [Steve Henson]
578
579 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
580 FIPS EC methods unconditionally for now.
581 [Steve Henson]
582
583 *) New build option no-ec2m to disable characteristic 2 code.
584 [Steve Henson]
585
586 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
587 all cases can be covered as some introduce binary incompatibilities.
588 [Steve Henson]
589
590 *) Redirect RSA operations to FIPS module including keygen,
591 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
592 [Steve Henson]
593
594 *) Add similar low level API blocking to ciphers.
595 [Steve Henson]
596
597 *) Low level digest APIs are not approved in FIPS mode: any attempt
598 to use these will cause a fatal error. Applications that *really* want
599 to use them can use the private_* version instead.
600 [Steve Henson]
601
602 *) Redirect cipher operations to FIPS module for FIPS builds.
603 [Steve Henson]
604
605 *) Redirect digest operations to FIPS module for FIPS builds.
606 [Steve Henson]
607
608 *) Update build system to add "fips" flag which will link in fipscanister.o
609 for static and shared library builds embedding a signature if needed.
610 [Steve Henson]
611
612 *) Output TLS supported curves in preference order instead of numerical
613 order. This is currently hardcoded for the highest order curves first.
614 This should be configurable so applications can judge speed vs strength.
615 [Steve Henson]
616
617 *) Add TLS v1.2 server support for client authentication.
618 [Steve Henson]
619
620 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
621 and enable MD5.
622 [Steve Henson]
623
624 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
625 FIPS modules versions.
626 [Steve Henson]
627
628 *) Add TLS v1.2 client side support for client authentication. Keep cache
629 of handshake records longer as we don't know the hash algorithm to use
630 until after the certificate request message is received.
631 [Steve Henson]
632
633 *) Initial TLS v1.2 client support. Add a default signature algorithms
634 extension including all the algorithms we support. Parse new signature
635 format in client key exchange. Relax some ECC signing restrictions for
636 TLS v1.2 as indicated in RFC5246.
637 [Steve Henson]
638
639 *) Add server support for TLS v1.2 signature algorithms extension. Switch
640 to new signature format when needed using client digest preference.
641 All server ciphersuites should now work correctly in TLS v1.2. No client
642 support yet and no support for client certificates.
643 [Steve Henson]
644
645 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
646 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
647 ciphersuites. At present only RSA key exchange ciphersuites work with
648 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
649 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
650 and version checking.
651 [Steve Henson]
652
653 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
654 with this defined it will not be affected by any changes to ssl internal
655 structures. Add several utility functions to allow openssl application
656 to work with OPENSSL_NO_SSL_INTERN defined.
657 [Steve Henson]
658
659 *) Add SRP support.
660 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
661
662 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
663 [Steve Henson]
664
665 *) Permit abbreviated handshakes when renegotiating using the function
666 SSL_renegotiate_abbreviated().
667 [Robin Seggelmann <seggelmann@fh-muenster.de>]
668
669 *) Add call to ENGINE_register_all_complete() to
670 ENGINE_load_builtin_engines(), so some implementations get used
671 automatically instead of needing explicit application support.
672 [Steve Henson]
673
674 *) Add support for TLS key exporter as described in RFC5705.
675 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
676
677 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
678 a few changes are required:
679
680 Add SSL_OP_NO_TLSv1_1 flag.
681 Add TLSv1_1 methods.
682 Update version checking logic to handle version 1.1.
683 Add explicit IV handling (ported from DTLS code).
684 Add command line options to s_client/s_server.
685 [Steve Henson]
686
687 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
688
689 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
690 in CMS and PKCS7 code. When RSA decryption fails use a random key for
691 content decryption and always return the same error. Note: this attack
692 needs on average 2^20 messages so it only affects automated senders. The
693 old behaviour can be reenabled in the CMS code by setting the
694 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
695 an MMA defence is not necessary.
696 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
697 this issue. (CVE-2012-0884)
698 [Steve Henson]
699
700 *) Fix CVE-2011-4619: make sure we really are receiving a
701 client hello before rejecting multiple SGC restarts. Thanks to
702 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
703 [Steve Henson]
704
705 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
706
707 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
708 Thanks to Antonio Martin, Enterprise Secure Access Research and
709 Development, Cisco Systems, Inc. for discovering this bug and
710 preparing a fix. (CVE-2012-0050)
711 [Antonio Martin]
712
713 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
714
715 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
716 of the Vaudenay padding oracle attack on CBC mode encryption
717 which enables an efficient plaintext recovery attack against
718 the OpenSSL implementation of DTLS. Their attack exploits timing
719 differences arising during decryption processing. A research
720 paper describing this attack can be found at:
721 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
722 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
723 Security Group at Royal Holloway, University of London
724 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
725 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
726 for preparing the fix. (CVE-2011-4108)
727 [Robin Seggelmann, Michael Tuexen]
728
729 *) Clear bytes used for block padding of SSL 3.0 records.
730 (CVE-2011-4576)
731 [Adam Langley (Google)]
732
733 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
734 Kadianakis <desnacked@gmail.com> for discovering this issue and
735 Adam Langley for preparing the fix. (CVE-2011-4619)
736 [Adam Langley (Google)]
737
738 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
739 [Andrey Kulikov <amdeich@gmail.com>]
740
741 *) Prevent malformed RFC3779 data triggering an assertion failure.
742 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
743 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
744 [Rob Austein <sra@hactrn.net>]
745
746 *) Improved PRNG seeding for VOS.
747 [Paul Green <Paul.Green@stratus.com>]
748
749 *) Fix ssl_ciph.c set-up race.
750 [Adam Langley (Google)]
751
752 *) Fix spurious failures in ecdsatest.c.
753 [Emilia Käsper (Google)]
754
755 *) Fix the BIO_f_buffer() implementation (which was mixing different
756 interpretations of the '..._len' fields).
757 [Adam Langley (Google)]
758
759 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
760 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
761 threads won't reuse the same blinding coefficients.
762
763 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
764 lock to call BN_BLINDING_invert_ex, and avoids one use of
765 BN_BLINDING_update for each BN_BLINDING structure (previously,
766 the last update always remained unused).
767 [Emilia Käsper (Google)]
768
769 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
770 [Bob Buckholz (Google)]
771
772 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
773
774 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
775 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
776 [Kaspar Brand <ossl@velox.ch>]
777
778 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
779 for multi-threaded use of ECDH. (CVE-2011-3210)
780 [Adam Langley (Google)]
781
782 *) Fix x509_name_ex_d2i memory leak on bad inputs.
783 [Bodo Moeller]
784
785 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
786 signature public key algorithm by using OID xref utilities instead.
787 Before this you could only use some ECC ciphersuites with SHA1 only.
788 [Steve Henson]
789
790 *) Add protection against ECDSA timing attacks as mentioned in the paper
791 by Billy Bob Brumley and Nicola Tuveri, see:
792
793 http://eprint.iacr.org/2011/232.pdf
794
795 [Billy Bob Brumley and Nicola Tuveri]
796
797 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
798
799 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
800 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
801
802 *) Fix bug in string printing code: if *any* escaping is enabled we must
803 escape the escape character (backslash) or the resulting string is
804 ambiguous.
805 [Steve Henson]
806
807 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
808
809 *) Disable code workaround for ancient and obsolete Netscape browsers
810 and servers: an attacker can use it in a ciphersuite downgrade attack.
811 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
812 [Steve Henson]
813
814 *) Fixed J-PAKE implementation error, originally discovered by
815 Sebastien Martini, further info and confirmation from Stefan
816 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
817 [Ben Laurie]
818
819 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
820
821 *) Fix extension code to avoid race conditions which can result in a buffer
822 overrun vulnerability: resumed sessions must not be modified as they can
823 be shared by multiple threads. CVE-2010-3864
824 [Steve Henson]
825
826 *) Fix WIN32 build system to correctly link an ENGINE directory into
827 a DLL.
828 [Steve Henson]
829
830 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
831
832 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
833 (CVE-2010-1633)
834 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
835
836 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
837
838 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
839 context. The operation can be customised via the ctrl mechanism in
840 case ENGINEs want to include additional functionality.
841 [Steve Henson]
842
843 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
844 [Steve Henson]
845
846 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
847 output hashes compatible with older versions of OpenSSL.
848 [Willy Weisz <weisz@vcpc.univie.ac.at>]
849
850 *) Fix compression algorithm handling: if resuming a session use the
851 compression algorithm of the resumed session instead of determining
852 it from client hello again. Don't allow server to change algorithm.
853 [Steve Henson]
854
855 *) Add load_crls() function to apps tidying load_certs() too. Add option
856 to verify utility to allow additional CRLs to be included.
857 [Steve Henson]
858
859 *) Update OCSP request code to permit adding custom headers to the request:
860 some responders need this.
861 [Steve Henson]
862
863 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
864 correctly.
865 [Julia Lawall <julia@diku.dk>]
866
867 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
868 needlessly dereferenced structures, used obsolete functions and
869 didn't handle all updated verify codes correctly.
870 [Steve Henson]
871
872 *) Disable MD2 in the default configuration.
873 [Steve Henson]
874
875 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
876 indicate the initial BIO being pushed or popped. This makes it possible
877 to determine whether the BIO is the one explicitly called or as a result
878 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
879 it handles reference counts correctly and doesn't zero out the I/O bio
880 when it is not being explicitly popped. WARNING: applications which
881 included workarounds for the old buggy behaviour will need to be modified
882 or they could free up already freed BIOs.
883 [Steve Henson]
884
885 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
886 renaming to all platforms (within the 0.9.8 branch, this was
887 done conditionally on Netware platforms to avoid a name clash).
888 [Guenter <lists@gknw.net>]
889
890 *) Add ECDHE and PSK support to DTLS.
891 [Michael Tuexen <tuexen@fh-muenster.de>]
892
893 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
894 be used on C++.
895 [Steve Henson]
896
897 *) Add "missing" function EVP_MD_flags() (without this the only way to
898 retrieve a digest flags is by accessing the structure directly. Update
899 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
900 or cipher is registered as in the "from" argument. Print out all
901 registered digests in the dgst usage message instead of manually
902 attempting to work them out.
903 [Steve Henson]
904
905 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
906 this allows the use of compression and extensions. Change default cipher
907 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
908 by default unless an application cipher string requests it.
909 [Steve Henson]
910
911 *) Alter match criteria in PKCS12_parse(). It used to try to use local
912 key ids to find matching certificates and keys but some PKCS#12 files
913 don't follow the (somewhat unwritten) rules and this strategy fails.
914 Now just gather all certificates together and the first private key
915 then look for the first certificate that matches the key.
916 [Steve Henson]
917
918 *) Support use of registered digest and cipher names for dgst and cipher
919 commands instead of having to add each one as a special case. So now
920 you can do:
921
922 openssl sha256 foo
923
924 as well as:
925
926 openssl dgst -sha256 foo
927
928 and this works for ENGINE based algorithms too.
929
930 [Steve Henson]
931
932 *) Update Gost ENGINE to support parameter files.
933 [Victor B. Wagner <vitus@cryptocom.ru>]
934
935 *) Support GeneralizedTime in ca utility.
936 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
937
938 *) Enhance the hash format used for certificate directory links. The new
939 form uses the canonical encoding (meaning equivalent names will work
940 even if they aren't identical) and uses SHA1 instead of MD5. This form
941 is incompatible with the older format and as a result c_rehash should
942 be used to rebuild symbolic links.
943 [Steve Henson]
944
945 *) Make PKCS#8 the default write format for private keys, replacing the
946 traditional format. This form is standardised, more secure and doesn't
947 include an implicit MD5 dependency.
948 [Steve Henson]
949
950 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
951 committed to OpenSSL should pass this lot as a minimum.
952 [Steve Henson]
953
954 *) Add session ticket override functionality for use by EAP-FAST.
955 [Jouni Malinen <j@w1.fi>]
956
957 *) Modify HMAC functions to return a value. Since these can be implemented
958 in an ENGINE errors can occur.
959 [Steve Henson]
960
961 *) Type-checked OBJ_bsearch_ex.
962 [Ben Laurie]
963
964 *) Type-checked OBJ_bsearch. Also some constification necessitated
965 by type-checking. Still to come: TXT_DB, bsearch(?),
966 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
967 CONF_VALUE.
968 [Ben Laurie]
969
970 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
971 seconds to a tm structure directly, instead of going through OS
972 specific date routines. This avoids any issues with OS routines such
973 as the year 2038 bug. New *_adj() functions for ASN1 time structures
974 and X509_time_adj_ex() to cover the extended range. The existing
975 X509_time_adj() is still usable and will no longer have any date issues.
976 [Steve Henson]
977
978 *) Delta CRL support. New use deltas option which will attempt to locate
979 and search any appropriate delta CRLs available.
980
981 This work was sponsored by Google.
982 [Steve Henson]
983
984 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
985 code and add additional score elements. Validate alternate CRL paths
986 as part of the CRL checking and indicate a new error "CRL path validation
987 error" in this case. Applications wanting additional details can use
988 the verify callback and check the new "parent" field. If this is not
989 NULL CRL path validation is taking place. Existing applications wont
990 see this because it requires extended CRL support which is off by
991 default.
992
993 This work was sponsored by Google.
994 [Steve Henson]
995
996 *) Support for freshest CRL extension.
997
998 This work was sponsored by Google.
999 [Steve Henson]
1000
1001 *) Initial indirect CRL support. Currently only supported in the CRLs
1002 passed directly and not via lookup. Process certificate issuer
1003 CRL entry extension and lookup CRL entries by bother issuer name
1004 and serial number. Check and process CRL issuer entry in IDP extension.
1005
1006 This work was sponsored by Google.
1007 [Steve Henson]
1008
1009 *) Add support for distinct certificate and CRL paths. The CRL issuer
1010 certificate is validated separately in this case. Only enabled if
1011 an extended CRL support flag is set: this flag will enable additional
1012 CRL functionality in future.
1013
1014 This work was sponsored by Google.
1015 [Steve Henson]
1016
1017 *) Add support for policy mappings extension.
1018
1019 This work was sponsored by Google.
1020 [Steve Henson]
1021
1022 *) Fixes to pathlength constraint, self issued certificate handling,
1023 policy processing to align with RFC3280 and PKITS tests.
1024
1025 This work was sponsored by Google.
1026 [Steve Henson]
1027
1028 *) Support for name constraints certificate extension. DN, email, DNS
1029 and URI types are currently supported.
1030
1031 This work was sponsored by Google.
1032 [Steve Henson]
1033
1034 *) To cater for systems that provide a pointer-based thread ID rather
1035 than numeric, deprecate the current numeric thread ID mechanism and
1036 replace it with a structure and associated callback type. This
1037 mechanism allows a numeric "hash" to be extracted from a thread ID in
1038 either case, and on platforms where pointers are larger than 'long',
1039 mixing is done to help ensure the numeric 'hash' is usable even if it
1040 can't be guaranteed unique. The default mechanism is to use "&errno"
1041 as a pointer-based thread ID to distinguish between threads.
1042
1043 Applications that want to provide their own thread IDs should now use
1044 CRYPTO_THREADID_set_callback() to register a callback that will call
1045 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1046
1047 Note that ERR_remove_state() is now deprecated, because it is tied
1048 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1049 to free the current thread's error state should be replaced by
1050 ERR_remove_thread_state(NULL).
1051
1052 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1053 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1054 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1055 application was previously providing a numeric thread callback that
1056 was inappropriate for distinguishing threads, then uniqueness might
1057 have been obtained with &errno that happened immediately in the
1058 intermediate development versions of OpenSSL; this is no longer the
1059 case, the numeric thread callback will now override the automatic use
1060 of &errno.)
1061 [Geoff Thorpe, with help from Bodo Moeller]
1062
1063 *) Initial support for different CRL issuing certificates. This covers a
1064 simple case where the self issued certificates in the chain exist and
1065 the real CRL issuer is higher in the existing chain.
1066
1067 This work was sponsored by Google.
1068 [Steve Henson]
1069
1070 *) Removed effectively defunct crypto/store from the build.
1071 [Ben Laurie]
1072
1073 *) Revamp of STACK to provide stronger type-checking. Still to come:
1074 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1075 ASN1_STRING, CONF_VALUE.
1076 [Ben Laurie]
1077
1078 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1079 RAM on SSL connections. This option can save about 34k per idle SSL.
1080 [Nick Mathewson]
1081
1082 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1083 STACK, TXT_DB, bsearch, qsort.
1084 [Ben Laurie]
1085
1086 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1087 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1088 support for data, signedData, compressedData, digestedData and
1089 encryptedData, envelopedData types included. Scripts to check against
1090 RFC4134 examples draft and interop and consistency checks of many
1091 content types and variants.
1092 [Steve Henson]
1093
1094 *) Add options to enc utility to support use of zlib compression BIO.
1095 [Steve Henson]
1096
1097 *) Extend mk1mf to support importing of options and assembly language
1098 files from Configure script, currently only included in VC-WIN32.
1099 The assembly language rules can now optionally generate the source
1100 files from the associated perl scripts.
1101 [Steve Henson]
1102
1103 *) Implement remaining functionality needed to support GOST ciphersuites.
1104 Interop testing has been performed using CryptoPro implementations.
1105 [Victor B. Wagner <vitus@cryptocom.ru>]
1106
1107 *) s390x assembler pack.
1108 [Andy Polyakov]
1109
1110 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1111 "family."
1112 [Andy Polyakov]
1113
1114 *) Implement Opaque PRF Input TLS extension as specified in
1115 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1116 official specification yet and no extension type assignment by
1117 IANA exists, this extension (for now) will have to be explicitly
1118 enabled when building OpenSSL by providing the extension number
1119 to use. For example, specify an option
1120
1121 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1122
1123 to the "config" or "Configure" script to enable the extension,
1124 assuming extension number 0x9527 (which is a completely arbitrary
1125 and unofficial assignment based on the MD5 hash of the Internet
1126 Draft). Note that by doing so, you potentially lose
1127 interoperability with other TLS implementations since these might
1128 be using the same extension number for other purposes.
1129
1130 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1131 opaque PRF input value to use in the handshake. This will create
1132 an interal copy of the length-'len' string at 'src', and will
1133 return non-zero for success.
1134
1135 To get more control and flexibility, provide a callback function
1136 by using
1137
1138 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1139 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1140
1141 where
1142
1143 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1144 void *arg;
1145
1146 Callback function 'cb' will be called in handshakes, and is
1147 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1148 Argument 'arg' is for application purposes (the value as given to
1149 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1150 be provided to the callback function). The callback function
1151 has to return non-zero to report success: usually 1 to use opaque
1152 PRF input just if possible, or 2 to enforce use of the opaque PRF
1153 input. In the latter case, the library will abort the handshake
1154 if opaque PRF input is not successfully negotiated.
1155
1156 Arguments 'peerinput' and 'len' given to the callback function
1157 will always be NULL and 0 in the case of a client. A server will
1158 see the client's opaque PRF input through these variables if
1159 available (NULL and 0 otherwise). Note that if the server
1160 provides an opaque PRF input, the length must be the same as the
1161 length of the client's opaque PRF input.
1162
1163 Note that the callback function will only be called when creating
1164 a new session (session resumption can resume whatever was
1165 previously negotiated), and will not be called in SSL 2.0
1166 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1167 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1168 for applications that need to enforce opaque PRF input.
1169
1170 [Bodo Moeller]
1171
1172 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1173 MAC.
1174
1175 [Victor B. Wagner <vitus@cryptocom.ru>]
1176
1177 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1178 RFC4507bis. The encrypted ticket format is an encrypted encoded
1179 SSL_SESSION structure, that way new session features are automatically
1180 supported.
1181
1182 If a client application caches session in an SSL_SESSION structure
1183 support is transparent because tickets are now stored in the encoded
1184 SSL_SESSION.
1185
1186 The SSL_CTX structure automatically generates keys for ticket
1187 protection in servers so again support should be possible
1188 with no application modification.
1189
1190 If a client or server wishes to disable RFC4507 support then the option
1191 SSL_OP_NO_TICKET can be set.
1192
1193 Add a TLS extension debugging callback to allow the contents of any client
1194 or server extensions to be examined.
1195
1196 This work was sponsored by Google.
1197 [Steve Henson]
1198
1199 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1200 OpenSSL should now compile cleanly on gcc 4.2
1201 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1202
1203 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1204 support including streaming MAC support: this is required for GOST
1205 ciphersuite support.
1206 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1207
1208 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1209 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1210 to output in BER and PEM format.
1211 [Steve Henson]
1212
1213 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1214 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1215 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1216 ENGINE support for HMAC keys which are unextractable. New -mac and
1217 -macopt options to dgst utility.
1218 [Steve Henson]
1219
1220 *) New option -sigopt to dgst utility. Update dgst to use
1221 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1222 alternative signing paramaters such as X9.31 or PSS in the dgst
1223 utility.
1224 [Steve Henson]
1225
1226 *) Change ssl_cipher_apply_rule(), the internal function that does
1227 the work each time a ciphersuite string requests enabling
1228 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1229 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1230 the order of disabled ciphersuites such that those ciphersuites
1231 that most recently went from enabled to disabled not only stay
1232 in order with respect to each other, but also have higher priority
1233 than other disabled ciphersuites the next time ciphersuites are
1234 enabled again.
1235
1236 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1237 the same ciphersuites as with "HIGH" alone, but in a specific
1238 order where the PSK ciphersuites come first (since they are the
1239 most recently disabled ciphersuites when "HIGH" is parsed).
1240
1241 Also, change ssl_create_cipher_list() (using this new
1242 funcionality) such that between otherwise identical
1243 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1244 the default order.
1245 [Bodo Moeller]
1246
1247 *) Change ssl_create_cipher_list() so that it automatically
1248 arranges the ciphersuites in reasonable order before starting
1249 to process the rule string. Thus, the definition for "DEFAULT"
1250 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1251 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1252 This makes it much easier to arrive at a reasonable default order
1253 in applications for which anonymous ciphers are OK (meaning
1254 that you can't actually use DEFAULT).
1255 [Bodo Moeller; suggested by Victor Duchovni]
1256
1257 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1258 processing) into multiple integers instead of setting
1259 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1260 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1261 (These masks as well as the individual bit definitions are hidden
1262 away into the non-exported interface ssl/ssl_locl.h, so this
1263 change to the definition of the SSL_CIPHER structure shouldn't
1264 affect applications.) This give us more bits for each of these
1265 categories, so there is no longer a need to coagulate AES128 and
1266 AES256 into a single algorithm bit, and to coagulate Camellia128
1267 and Camellia256 into a single algorithm bit, which has led to all
1268 kinds of kludges.
1269
1270 Thus, among other things, the kludge introduced in 0.9.7m and
1271 0.9.8e for masking out AES256 independently of AES128 or masking
1272 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1273
1274 With the change, we also introduce new ciphersuite aliases that
1275 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1276 "CAMELLIA256".
1277 [Bodo Moeller]
1278
1279 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1280 Use the leftmost N bytes of the signature input if the input is
1281 larger than the prime q (with N being the size in bytes of q).
1282 [Nils Larsch]
1283
1284 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1285 it yet and it is largely untested.
1286 [Steve Henson]
1287
1288 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1289 [Nils Larsch]
1290
1291 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1292 some compilers (gcc 4.2 and later) reject their use. Safestack is
1293 reimplemented. Update ASN1 to avoid use of legacy functions.
1294 [Steve Henson]
1295
1296 *) Win32/64 targets are linked with Winsock2.
1297 [Andy Polyakov]
1298
1299 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1300 to external functions. This can be used to increase CRL handling
1301 efficiency especially when CRLs are very large by (for example) storing
1302 the CRL revoked certificates in a database.
1303 [Steve Henson]
1304
1305 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1306 new CRLs added to a directory can be used. New command line option
1307 -verify_return_error to s_client and s_server. This causes real errors
1308 to be returned by the verify callback instead of carrying on no matter
1309 what. This reflects the way a "real world" verify callback would behave.
1310 [Steve Henson]
1311
1312 *) GOST engine, supporting several GOST algorithms and public key formats.
1313 Kindly donated by Cryptocom.
1314 [Cryptocom]
1315
1316 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1317 partitioned by DP are handled but no indirect CRL or reason partitioning
1318 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1319 selected via a scoring technique which handles IDP and AKID in CRLs.
1320 [Steve Henson]
1321
1322 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1323 will ultimately be used for all verify operations: this will remove the
1324 X509_STORE dependency on certificate verification and allow alternative
1325 lookup methods. X509_STORE based implementations of these two callbacks.
1326 [Steve Henson]
1327
1328 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1329 Modify get_crl() to find a valid (unexpired) CRL if possible.
1330 [Steve Henson]
1331
1332 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1333 this would be called X509_CRL_cmp() but that name is already used by
1334 a function that just compares CRL issuer names. Cache several CRL
1335 extensions in X509_CRL structure and cache CRLDP in X509.
1336 [Steve Henson]
1337
1338 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1339 this maps equivalent X509_NAME structures into a consistent structure.
1340 Name comparison can then be performed rapidly using memcmp().
1341 [Steve Henson]
1342
1343 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
1344 utility.
1345 [Steve Henson]
1346
1347 *) Allow digests to supply their own micalg string for S/MIME type using
1348 the ctrl EVP_MD_CTRL_MICALG.
1349 [Steve Henson]
1350
1351 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
1352 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
1353 ctrl. It can then customise the structure before and/or after signing
1354 if necessary.
1355 [Steve Henson]
1356
1357 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1358 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1359 to free up any added signature OIDs.
1360 [Steve Henson]
1361
1362 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1363 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1364 digest and cipher tables. New options added to openssl utility:
1365 list-message-digest-algorithms and list-cipher-algorithms.
1366 [Steve Henson]
1367
1368 *) Change the array representation of binary polynomials: the list
1369 of degrees of non-zero coefficients is now terminated with -1.
1370 Previously it was terminated with 0, which was also part of the
1371 value; thus, the array representation was not applicable to
1372 polynomials where t^0 has coefficient zero. This change makes
1373 the array representation useful in a more general context.
1374 [Douglas Stebila]
1375
1376 *) Various modifications and fixes to SSL/TLS cipher string
1377 handling. For ECC, the code now distinguishes between fixed ECDH
1378 with RSA certificates on the one hand and with ECDSA certificates
1379 on the other hand, since these are separate ciphersuites. The
1380 unused code for Fortezza ciphersuites has been removed.
1381
1382 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1383 (not "ECDHE"). For consistency with the code for DH
1384 certificates, use of ECDH certificates is now considered ECDH
1385 authentication, not RSA or ECDSA authentication (the latter is
1386 merely the CA's signing algorithm and not actively used in the
1387 protocol).
1388
1389 The temporary ciphersuite alias "ECCdraft" is no longer
1390 available, and ECC ciphersuites are no longer excluded from "ALL"
1391 and "DEFAULT". The following aliases now exist for RFC 4492
1392 ciphersuites, most of these by analogy with the DH case:
1393
1394 kECDHr - ECDH cert, signed with RSA
1395 kECDHe - ECDH cert, signed with ECDSA
1396 kECDH - ECDH cert (signed with either RSA or ECDSA)
1397 kEECDH - ephemeral ECDH
1398 ECDH - ECDH cert or ephemeral ECDH
1399
1400 aECDH - ECDH cert
1401 aECDSA - ECDSA cert
1402 ECDSA - ECDSA cert
1403
1404 AECDH - anonymous ECDH
1405 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1406
1407 [Bodo Moeller]
1408
1409 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1410 Use correct micalg parameters depending on digest(s) in signed message.
1411 [Steve Henson]
1412
1413 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1414 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1415 [Steve Henson]
1416
1417 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1418 an engine to register a method. Add ENGINE lookups for methods and
1419 functional reference processing.
1420 [Steve Henson]
1421
1422 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1423 EVP_{Sign,Verify}* which allow an application to customise the signature
1424 process.
1425 [Steve Henson]
1426
1427 *) New -resign option to smime utility. This adds one or more signers
1428 to an existing PKCS#7 signedData structure. Also -md option to use an
1429 alternative message digest algorithm for signing.
1430 [Steve Henson]
1431
1432 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1433 create PKCS7 structures containing multiple signers. Update smime
1434 application to support multiple signers.
1435 [Steve Henson]
1436
1437 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1438 digest MAC.
1439 [Steve Henson]
1440
1441 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1442 Reorganize PBE internals to lookup from a static table using NIDs,
1443 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1444 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1445 PRF which will be automatically used with PBES2.
1446 [Steve Henson]
1447
1448 *) Replace the algorithm specific calls to generate keys in "req" with the
1449 new API.
1450 [Steve Henson]
1451
1452 *) Update PKCS#7 enveloped data routines to use new API. This is now
1453 supported by any public key method supporting the encrypt operation. A
1454 ctrl is added to allow the public key algorithm to examine or modify
1455 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1456 a no op.
1457 [Steve Henson]
1458
1459 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1460 a default digest type to use. In most cases this will be SHA1 but some
1461 algorithms (such as GOST) need to specify an alternative digest. The
1462 return value indicates how strong the prefernce is 1 means optional and
1463 2 is mandatory (that is it is the only supported type). Modify
1464 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1465 use the default md. Update openssl utilities to use the default digest
1466 type for signing if it is not explicitly indicated.
1467 [Steve Henson]
1468
1469 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1470 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1471 signing method from the key type. This effectively removes the link
1472 between digests and public key types.
1473 [Steve Henson]
1474
1475 *) Add an OID cross reference table and utility functions. Its purpose is to
1476 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1477 rsaEncryption. This will allow some of the algorithm specific hackery
1478 needed to use the correct OID to be removed.
1479 [Steve Henson]
1480
1481 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1482 structures for PKCS7_sign(). They are now set up by the relevant public
1483 key ASN1 method.
1484 [Steve Henson]
1485
1486 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1487 [Steve Henson]
1488
1489 *) Add support for key derivation (agreement) in the API, DH method and
1490 pkeyutl.
1491 [Steve Henson]
1492
1493 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1494 public and private key formats. As a side effect these add additional
1495 command line functionality not previously available: DSA signatures can be
1496 generated and verified using pkeyutl and DH key support and generation in
1497 pkey, genpkey.
1498 [Steve Henson]
1499
1500 *) BeOS support.
1501 [Oliver Tappe <zooey@hirschkaefer.de>]
1502
1503 *) New make target "install_html_docs" installs HTML renditions of the
1504 manual pages.
1505 [Oliver Tappe <zooey@hirschkaefer.de>]
1506
1507 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1508 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1509 support key and parameter generation and add initial key generation
1510 functionality for RSA.
1511 [Steve Henson]
1512
1513 *) Add functions for main EVP_PKEY_method operations. The undocumented
1514 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1515 EVP_PKEY_{encrypt,decrypt}_old.
1516 [Steve Henson]
1517
1518 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1519 key API, doesn't do much yet.
1520 [Steve Henson]
1521
1522 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1523 public key algorithms. New option to openssl utility:
1524 "list-public-key-algorithms" to print out info.
1525 [Steve Henson]
1526
1527 *) Implement the Supported Elliptic Curves Extension for
1528 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1529 [Douglas Stebila]
1530
1531 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1532 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1533 [Steve Henson]
1534
1535 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1536 utilities such as rsa, dsa, dsaparam etc except they process any key
1537 type.
1538 [Steve Henson]
1539
1540 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1541 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1542 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1543 structure.
1544 [Steve Henson]
1545
1546 *) Initial support for pluggable public key ASN1.
1547 De-spaghettify the public key ASN1 handling. Move public and private
1548 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1549 algorithm specific handling to a single module within the relevant
1550 algorithm directory. Add functions to allow (near) opaque processing
1551 of public and private key structures.
1552 [Steve Henson]
1553
1554 *) Implement the Supported Point Formats Extension for
1555 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1556 [Douglas Stebila]
1557
1558 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1559 for the psk identity [hint] and the psk callback functions to the
1560 SSL_SESSION, SSL and SSL_CTX structure.
1561
1562 New ciphersuites:
1563 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1564 PSK-AES256-CBC-SHA
1565
1566 New functions:
1567 SSL_CTX_use_psk_identity_hint
1568 SSL_get_psk_identity_hint
1569 SSL_get_psk_identity
1570 SSL_use_psk_identity_hint
1571
1572 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1573
1574 *) Add RFC 3161 compliant time stamp request creation, response generation
1575 and response verification functionality.
1576 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1577
1578 *) Add initial support for TLS extensions, specifically for the server_name
1579 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1580 have new members for a host name. The SSL data structure has an
1581 additional member SSL_CTX *initial_ctx so that new sessions can be
1582 stored in that context to allow for session resumption, even after the
1583 SSL has been switched to a new SSL_CTX in reaction to a client's
1584 server_name extension.
1585
1586 New functions (subject to change):
1587
1588 SSL_get_servername()
1589 SSL_get_servername_type()
1590 SSL_set_SSL_CTX()
1591
1592 New CTRL codes and macros (subject to change):
1593
1594 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1595 - SSL_CTX_set_tlsext_servername_callback()
1596 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1597 - SSL_CTX_set_tlsext_servername_arg()
1598 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1599
1600 openssl s_client has a new '-servername ...' option.
1601
1602 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1603 '-key2 ...', '-servername_fatal' (subject to change). This allows
1604 testing the HostName extension for a specific single host name ('-cert'
1605 and '-key' remain fallbacks for handshakes without HostName
1606 negotiation). If the unrecogninzed_name alert has to be sent, this by
1607 default is a warning; it becomes fatal with the '-servername_fatal'
1608 option.
1609
1610 [Peter Sylvester, Remy Allais, Christophe Renou]
1611
1612 *) Whirlpool hash implementation is added.
1613 [Andy Polyakov]
1614
1615 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1616 bn(64,32). Because of instruction set limitations it doesn't have
1617 any negative impact on performance. This was done mostly in order
1618 to make it possible to share assembler modules, such as bn_mul_mont
1619 implementations, between 32- and 64-bit builds without hassle.
1620 [Andy Polyakov]
1621
1622 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1623 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1624 macro.
1625 [Bodo Moeller]
1626
1627 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1628 dedicated Montgomery multiplication procedure, is introduced.
1629 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1630 "64-bit" performance on certain 32-bit targets.
1631 [Andy Polyakov]
1632
1633 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1634 in SSL structures. New SSL ctrl to set maximum send fragment size.
1635 Save memory by seeting the I/O buffer sizes dynamically instead of
1636 using the maximum available value.
1637 [Steve Henson]
1638
1639 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1640 in addition to the text details.
1641 [Bodo Moeller]
1642
1643 *) Very, very preliminary EXPERIMENTAL support for printing of general
1644 ASN1 structures. This currently produces rather ugly output and doesn't
1645 handle several customised structures at all.
1646 [Steve Henson]
1647
1648 *) Integrated support for PVK file format and some related formats such
1649 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1650 these in the 'rsa' and 'dsa' utilities.
1651 [Steve Henson]
1652
1653 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1654 [Steve Henson]
1655
1656 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1657 place for the (very old) "NETSCAPE" format certificates which are now
1658 handled using new ASN1 code equivalents.
1659 [Steve Henson]
1660
1661 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1662 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1663 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1664 [Nils Larsch]
1665
1666 *) Modify CRL distribution points extension code to print out previously
1667 unsupported fields. Enhance extension setting code to allow setting of
1668 all fields.
1669 [Steve Henson]
1670
1671 *) Add print and set support for Issuing Distribution Point CRL extension.
1672 [Steve Henson]
1673
1674 *) Change 'Configure' script to enable Camellia by default.
1675 [NTT]
1676
1677 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
1678
1679 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1680 Thanks to Antonio Martin, Enterprise Secure Access Research and
1681 Development, Cisco Systems, Inc. for discovering this bug and
1682 preparing a fix. (CVE-2012-0050)
1683 [Antonio Martin]
1684
1685 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
1686
1687 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1688 of the Vaudenay padding oracle attack on CBC mode encryption
1689 which enables an efficient plaintext recovery attack against
1690 the OpenSSL implementation of DTLS. Their attack exploits timing
1691 differences arising during decryption processing. A research
1692 paper describing this attack can be found at:
1693 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1694 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1695 Security Group at Royal Holloway, University of London
1696 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1697 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1698 for preparing the fix. (CVE-2011-4108)
1699 [Robin Seggelmann, Michael Tuexen]
1700
1701 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
1702 [Ben Laurie, Kasper <ekasper@google.com>]
1703
1704 *) Clear bytes used for block padding of SSL 3.0 records.
1705 (CVE-2011-4576)
1706 [Adam Langley (Google)]
1707
1708 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1709 Kadianakis <desnacked@gmail.com> for discovering this issue and
1710 Adam Langley for preparing the fix. (CVE-2011-4619)
1711 [Adam Langley (Google)]
1712
1713 *) Prevent malformed RFC3779 data triggering an assertion failure.
1714 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1715 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1716 [Rob Austein <sra@hactrn.net>]
1717
1718 *) Fix ssl_ciph.c set-up race.
1719 [Adam Langley (Google)]
1720
1721 *) Fix spurious failures in ecdsatest.c.
1722 [Emilia Käsper (Google)]
1723
1724 *) Fix the BIO_f_buffer() implementation (which was mixing different
1725 interpretations of the '..._len' fields).
1726 [Adam Langley (Google)]
1727
1728 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1729 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1730 threads won't reuse the same blinding coefficients.
1731
1732 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1733 lock to call BN_BLINDING_invert_ex, and avoids one use of
1734 BN_BLINDING_update for each BN_BLINDING structure (previously,
1735 the last update always remained unused).
1736 [Emilia Käsper (Google)]
1737
1738 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1739 for multi-threaded use of ECDH.
1740 [Adam Langley (Google)]
1741
1742 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1743 [Bodo Moeller]
1744
1745 *) Add protection against ECDSA timing attacks as mentioned in the paper
1746 by Billy Bob Brumley and Nicola Tuveri, see:
1747
1748 http://eprint.iacr.org/2011/232.pdf
1749
1750 [Billy Bob Brumley and Nicola Tuveri]
1751
1752 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1753
1754 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1755 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1756
1757 *) Fix bug in string printing code: if *any* escaping is enabled we must
1758 escape the escape character (backslash) or the resulting string is
1759 ambiguous.
1760 [Steve Henson]
1761
1762 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1763
1764 *) Disable code workaround for ancient and obsolete Netscape browsers
1765 and servers: an attacker can use it in a ciphersuite downgrade attack.
1766 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1767 [Steve Henson]
1768
1769 *) Fixed J-PAKE implementation error, originally discovered by
1770 Sebastien Martini, further info and confirmation from Stefan
1771 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1772 [Ben Laurie]
1773
1774 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1775
1776 *) Fix extension code to avoid race conditions which can result in a buffer
1777 overrun vulnerability: resumed sessions must not be modified as they can
1778 be shared by multiple threads. CVE-2010-3864
1779 [Steve Henson]
1780
1781 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1782 [Steve Henson]
1783
1784 *) Don't reencode certificate when calculating signature: cache and use
1785 the original encoding instead. This makes signature verification of
1786 some broken encodings work correctly.
1787 [Steve Henson]
1788
1789 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1790 is also one of the inputs.
1791 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1792
1793 *) Don't repeatedly append PBE algorithms to table if they already exist.
1794 Sort table on each new add. This effectively makes the table read only
1795 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1796 etc are non-op.
1797 [Steve Henson]
1798
1799 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1800
1801 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1802 OpenSSL 1.0.0.]
1803
1804 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1805 access or freeing data twice (CVE-2010-0742)
1806 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1807
1808 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1809 common in certificates and some applications which only call
1810 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1811 [Steve Henson]
1812
1813 *) VMS fixes:
1814 Reduce copying into .apps and .test in makevms.com
1815 Don't try to use blank CA certificate in CA.com
1816 Allow use of C files from original directories in maketests.com
1817 [Steven M. Schweda" <sms@antinode.info>]
1818
1819 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1820
1821 *) When rejecting SSL/TLS records due to an incorrect version number, never
1822 update s->server with a new major version number. As of
1823 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1824 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1825 the previous behavior could result in a read attempt at NULL when
1826 receiving specific incorrect SSL/TLS records once record payload
1827 protection is active. (CVE-2010-0740)
1828 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1829
1830 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1831 could be crashed if the relevant tables were not present (e.g. chrooted).
1832 [Tomas Hoger <thoger@redhat.com>]
1833
1834 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1835
1836 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1837 [Martin Olsson, Neel Mehta]
1838
1839 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1840 accommodate for stack sorting, always a write lock!).
1841 [Bodo Moeller]
1842
1843 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1844 excessive delays in the RAND_poll(): over a minute. As a workaround
1845 include a time check in the inner Heap32Next loop too.
1846 [Steve Henson]
1847
1848 *) The code that handled flushing of data in SSL/TLS originally used the
1849 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1850 the problem outlined in PR#1949. The fix suggested there however can
1851 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1852 of Apache). So instead simplify the code to flush unconditionally.
1853 This should be fine since flushing with no data to flush is a no op.
1854 [Steve Henson]
1855
1856 *) Handle TLS versions 2.0 and later properly and correctly use the
1857 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1858 off ancient servers have a habit of sticking around for a while...
1859 [Steve Henson]
1860
1861 *) Modify compression code so it frees up structures without using the
1862 ex_data callbacks. This works around a problem where some applications
1863 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1864 restarting) then use compression (e.g. SSL with compression) later.
1865 This results in significant per-connection memory leaks and
1866 has caused some security issues including CVE-2008-1678 and
1867 CVE-2009-4355.
1868 [Steve Henson]
1869
1870 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1871 change when encrypting or decrypting.
1872 [Bodo Moeller]
1873
1874 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1875 connect and renegotiate with servers which do not support RI.
1876 Until RI is more widely deployed this option is enabled by default.
1877 [Steve Henson]
1878
1879 *) Add "missing" ssl ctrls to clear options and mode.
1880 [Steve Henson]
1881
1882 *) If client attempts to renegotiate and doesn't support RI respond with
1883 a no_renegotiation alert as required by RFC5746. Some renegotiating
1884 TLS clients will continue a connection gracefully when they receive
1885 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1886 waiting for a server hello which it will never receive. Now we treat a
1887 received no_renegotiation alert as a fatal error. This is because
1888 applications requesting a renegotiation might well expect it to succeed
1889 and would have no code in place to handle the server denying it so the
1890 only safe thing to do is to terminate the connection.
1891 [Steve Henson]
1892
1893 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1894 peer supports secure renegotiation and 0 otherwise. Print out peer
1895 renegotiation support in s_client/s_server.
1896 [Steve Henson]
1897
1898 *) Replace the highly broken and deprecated SPKAC certification method with
1899 the updated NID creation version. This should correctly handle UTF8.
1900 [Steve Henson]
1901
1902 *) Implement RFC5746. Re-enable renegotiation but require the extension
1903 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1904 turns out to be a bad idea. It has been replaced by
1905 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1906 SSL_CTX_set_options(). This is really not recommended unless you
1907 know what you are doing.
1908 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1909
1910 *) Fixes to stateless session resumption handling. Use initial_ctx when
1911 issuing and attempting to decrypt tickets in case it has changed during
1912 servername handling. Use a non-zero length session ID when attempting
1913 stateless session resumption: this makes it possible to determine if
1914 a resumption has occurred immediately after receiving server hello
1915 (several places in OpenSSL subtly assume this) instead of later in
1916 the handshake.
1917 [Steve Henson]
1918
1919 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1920 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1921 fixes for a few places where the return code is not checked
1922 correctly.
1923 [Julia Lawall <julia@diku.dk>]
1924
1925 *) Add --strict-warnings option to Configure script to include devteam
1926 warnings in other configurations.
1927 [Steve Henson]
1928
1929 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1930 makes it possible to install openssl libraries in locations which
1931 have names other than "lib", for example "/usr/lib64" which some
1932 systems need.
1933 [Steve Henson, based on patch from Jeremy Utley]
1934
1935 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1936 X690 8.9.12 and can produce some misleading textual output of OIDs.
1937 [Steve Henson, reported by Dan Kaminsky]
1938
1939 *) Delete MD2 from algorithm tables. This follows the recommendation in
1940 several standards that it is not used in new applications due to
1941 several cryptographic weaknesses. For binary compatibility reasons
1942 the MD2 API is still compiled in by default.
1943 [Steve Henson]
1944
1945 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1946 and restored.
1947 [Steve Henson]
1948
1949 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1950 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1951 clash.
1952 [Guenter <lists@gknw.net>]
1953
1954 *) Fix the server certificate chain building code to use X509_verify_cert(),
1955 it used to have an ad-hoc builder which was unable to cope with anything
1956 other than a simple chain.
1957 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1958
1959 *) Don't check self signed certificate signatures in X509_verify_cert()
1960 by default (a flag can override this): it just wastes time without
1961 adding any security. As a useful side effect self signed root CAs
1962 with non-FIPS digests are now usable in FIPS mode.
1963 [Steve Henson]
1964
1965 *) In dtls1_process_out_of_seq_message() the check if the current message
1966 is already buffered was missing. For every new message was memory
1967 allocated, allowing an attacker to perform an denial of service attack
1968 with sending out of seq handshake messages until there is no memory
1969 left. Additionally every future messege was buffered, even if the
1970 sequence number made no sense and would be part of another handshake.
1971 So only messages with sequence numbers less than 10 in advance will be
1972 buffered. (CVE-2009-1378)
1973 [Robin Seggelmann, discovered by Daniel Mentz]
1974
1975 *) Records are buffered if they arrive with a future epoch to be
1976 processed after finishing the corresponding handshake. There is
1977 currently no limitation to this buffer allowing an attacker to perform
1978 a DOS attack with sending records with future epochs until there is no
1979 memory left. This patch adds the pqueue_size() function to detemine
1980 the size of a buffer and limits the record buffer to 100 entries.
1981 (CVE-2009-1377)
1982 [Robin Seggelmann, discovered by Daniel Mentz]
1983
1984 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1985 parent structure is freed. (CVE-2009-1379)
1986 [Daniel Mentz]
1987
1988 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1989 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1990
1991 *) Add 2.5.4.* OIDs
1992 [Ilya O. <vrghost@gmail.com>]
1993
1994 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1995
1996 *) Disable renegotiation completely - this fixes a severe security
1997 problem (CVE-2009-3555) at the cost of breaking all
1998 renegotiation. Renegotiation can be re-enabled by setting
1999 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2000 run-time. This is really not recommended unless you know what
2001 you're doing.
2002 [Ben Laurie]
2003
2004 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2005
2006 *) Don't set val to NULL when freeing up structures, it is freed up by
2007 underlying code. If sizeof(void *) > sizeof(long) this can result in
2008 zeroing past the valid field. (CVE-2009-0789)
2009 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2010
2011 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2012 checked correctly. This would allow some invalid signed attributes to
2013 appear to verify correctly. (CVE-2009-0591)
2014 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2015
2016 *) Reject UniversalString and BMPString types with invalid lengths. This
2017 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2018 a legal length. (CVE-2009-0590)
2019 [Steve Henson]
2020
2021 *) Set S/MIME signing as the default purpose rather than setting it
2022 unconditionally. This allows applications to override it at the store
2023 level.
2024 [Steve Henson]
2025
2026 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2027 to handle some structures.
2028 [Steve Henson]
2029
2030 *) Improve efficiency of mem_gets: don't search whole buffer each time
2031 for a '\n'
2032 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2033
2034 *) New -hex option for openssl rand.
2035 [Matthieu Herrb]
2036
2037 *) Print out UTF8String and NumericString when parsing ASN1.
2038 [Steve Henson]
2039
2040 *) Support NumericString type for name components.
2041 [Steve Henson]
2042
2043 *) Allow CC in the environment to override the automatically chosen
2044 compiler. Note that nothing is done to ensure flags work with the
2045 chosen compiler.
2046 [Ben Laurie]
2047
2048 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2049
2050 *) Properly check EVP_VerifyFinal() and similar return values
2051 (CVE-2008-5077).
2052 [Ben Laurie, Bodo Moeller, Google Security Team]
2053
2054 *) Enable TLS extensions by default.
2055 [Ben Laurie]
2056
2057 *) Allow the CHIL engine to be loaded, whether the application is
2058 multithreaded or not. (This does not release the developer from the
2059 obligation to set up the dynamic locking callbacks.)
2060 [Sander Temme <sander@temme.net>]
2061
2062 *) Use correct exit code if there is an error in dgst command.
2063 [Steve Henson; problem pointed out by Roland Dirlewanger]
2064
2065 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2066 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2067 [Bodo Moeller]
2068
2069 *) Add experimental JPAKE support, including demo authentication in
2070 s_client and s_server.
2071 [Ben Laurie]
2072
2073 *) Set the comparison function in v3_addr_canonize().
2074 [Rob Austein <sra@hactrn.net>]
2075
2076 *) Add support for XMPP STARTTLS in s_client.
2077 [Philip Paeps <philip@freebsd.org>]
2078
2079 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2080 to ensure that even with this option, only ciphersuites in the
2081 server's preference list will be accepted. (Note that the option
2082 applies only when resuming a session, so the earlier behavior was
2083 just about the algorithm choice for symmetric cryptography.)
2084 [Bodo Moeller]
2085
2086 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2087
2088 *) Fix NULL pointer dereference if a DTLS server received
2089 ChangeCipherSpec as first record (CVE-2009-1386).
2090 [PR #1679]
2091
2092 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
2093 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2094 [Nagendra Modadugu]
2095
2096 *) The fix in 0.9.8c that supposedly got rid of unsafe
2097 double-checked locking was incomplete for RSA blinding,
2098 addressing just one layer of what turns out to have been
2099 doubly unsafe triple-checked locking.
2100
2101 So now fix this for real by retiring the MONT_HELPER macro
2102 in crypto/rsa/rsa_eay.c.
2103
2104 [Bodo Moeller; problem pointed out by Marius Schilder]
2105
2106 *) Various precautionary measures:
2107
2108 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2109
2110 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2111 (NB: This would require knowledge of the secret session ticket key
2112 to exploit, in which case you'd be SOL either way.)
2113
2114 - Change bn_nist.c so that it will properly handle input BIGNUMs
2115 outside the expected range.
2116
2117 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2118 builds.
2119
2120 [Neel Mehta, Bodo Moeller]
2121
2122 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2123 the load fails. Useful for distros.
2124 [Ben Laurie and the FreeBSD team]
2125
2126 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2127 [Steve Henson]
2128
2129 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2130 [Huang Ying]
2131
2132 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2133
2134 This work was sponsored by Logica.
2135 [Steve Henson]
2136
2137 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2138 keystores. Support for SSL/TLS client authentication too.
2139 Not compiled unless enable-capieng specified to Configure.
2140
2141 This work was sponsored by Logica.
2142 [Steve Henson]
2143
2144 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2145 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2146 attribute creation routines such as certifcate requests and PKCS#12
2147 files.
2148 [Steve Henson]
2149
2150 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2151
2152 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2153 handshake which could lead to a cilent crash as found using the
2154 Codenomicon TLS test suite (CVE-2008-1672)
2155 [Steve Henson, Mark Cox]
2156
2157 *) Fix double free in TLS server name extensions which could lead to
2158 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2159 [Joe Orton]
2160
2161 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2162
2163 Clear the error queue to ensure that error entries left from
2164 older function calls do not interfere with the correct operation.
2165 [Lutz Jaenicke, Erik de Castro Lopo]
2166
2167 *) Remove root CA certificates of commercial CAs:
2168
2169 The OpenSSL project does not recommend any specific CA and does not
2170 have any policy with respect to including or excluding any CA.
2171 Therefore it does not make any sense to ship an arbitrary selection
2172 of root CA certificates with the OpenSSL software.
2173 [Lutz Jaenicke]
2174
2175 *) RSA OAEP patches to fix two separate invalid memory reads.
2176 The first one involves inputs when 'lzero' is greater than
2177 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2178 before the beginning of from). The second one involves inputs where
2179 the 'db' section contains nothing but zeroes (there is a one-byte
2180 invalid read after the end of 'db').
2181 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2182
2183 *) Partial backport from 0.9.9-dev:
2184
2185 Introduce bn_mul_mont (dedicated Montgomery multiplication
2186 procedure) as a candidate for BIGNUM assembler implementation.
2187 While 0.9.9-dev uses assembler for various architectures, only
2188 x86_64 is available by default here in the 0.9.8 branch, and
2189 32-bit x86 is available through a compile-time setting.
2190
2191 To try the 32-bit x86 assembler implementation, use Configure
2192 option "enable-montasm" (which exists only for this backport).
2193
2194 As "enable-montasm" for 32-bit x86 disclaims code stability
2195 anyway, in this constellation we activate additional code
2196 backported from 0.9.9-dev for further performance improvements,
2197 namely BN_from_montgomery_word. (To enable this otherwise,
2198 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2199
2200 [Andy Polyakov (backport partially by Bodo Moeller)]
2201
2202 *) Add TLS session ticket callback. This allows an application to set
2203 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2204 values. This is useful for key rollover for example where several key
2205 sets may exist with different names.
2206 [Steve Henson]
2207
2208 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2209 This was broken until now in 0.9.8 releases, such that the only way
2210 a registered ENGINE could be used (assuming it initialises
2211 successfully on the host) was to explicitly set it as the default
2212 for the relevant algorithms. This is in contradiction with 0.9.7
2213 behaviour and the documentation. With this fix, when an ENGINE is
2214 registered into a given algorithm's table of implementations, the
2215 'uptodate' flag is reset so that auto-discovery will be used next
2216 time a new context for that algorithm attempts to select an
2217 implementation.
2218 [Ian Lister (tweaked by Geoff Thorpe)]
2219
2220 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2221 implemention in the following ways:
2222
2223 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2224 hard coded.
2225
2226 Lack of BER streaming support means one pass streaming processing is
2227 only supported if data is detached: setting the streaming flag is
2228 ignored for embedded content.
2229
2230 CMS support is disabled by default and must be explicitly enabled
2231 with the enable-cms configuration option.
2232 [Steve Henson]
2233
2234 *) Update the GMP engine glue to do direct copies between BIGNUM and
2235 mpz_t when openssl and GMP use the same limb size. Otherwise the
2236 existing "conversion via a text string export" trick is still used.
2237 [Paul Sheer <paulsheer@gmail.com>]
2238
2239 *) Zlib compression BIO. This is a filter BIO which compressed and
2240 uncompresses any data passed through it.
2241 [Steve Henson]
2242
2243 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2244 RFC3394 compatible AES key wrapping.
2245 [Steve Henson]
2246
2247 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2248 sets string data without copying. X509_ALGOR_set0() and
2249 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2250 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2251 from an X509_ATTRIBUTE structure optionally checking it occurs only
2252 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2253 data.
2254 [Steve Henson]
2255
2256 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2257 to get the expected BN_FLG_CONSTTIME behavior.
2258 [Bodo Moeller (Google)]
2259
2260 *) Netware support:
2261
2262 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
2263 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
2264 - added some more tests to do_tests.pl
2265 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
2266 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
2267 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
2268 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
2269 - various changes to netware.pl to enable gcc-cross builds on Win32
2270 platform
2271 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
2272 - various changes to fix missing prototype warnings
2273 - fixed x86nasm.pl to create correct asm files for NASM COFF output
2274 - added AES, WHIRLPOOL and CPUID assembler code to build files
2275 - added missing AES assembler make rules to mk1mf.pl
2276 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
2277 [Guenter Knauf <eflash@gmx.net>]
2278
2279 *) Implement certificate status request TLS extension defined in RFC3546.
2280 A client can set the appropriate parameters and receive the encoded
2281 OCSP response via a callback. A server can query the supplied parameters
2282 and set the encoded OCSP response in the callback. Add simplified examples
2283 to s_client and s_server.
2284 [Steve Henson]
2285
2286 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
2287
2288 *) Fix various bugs:
2289 + Binary incompatibility of ssl_ctx_st structure
2290 + DTLS interoperation with non-compliant servers
2291 + Don't call get_session_cb() without proposed session
2292 + Fix ia64 assembler code
2293 [Andy Polyakov, Steve Henson]
2294
2295 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
2296
2297 *) DTLS Handshake overhaul. There were longstanding issues with
2298 OpenSSL DTLS implementation, which were making it impossible for
2299 RFC 4347 compliant client to communicate with OpenSSL server.
2300 Unfortunately just fixing these incompatibilities would "cut off"
2301 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
2302 server keeps tolerating non RFC compliant syntax. The opposite is
2303 not true, 0.9.8f client can not communicate with earlier server.
2304 This update even addresses CVE-2007-4995.
2305 [Andy Polyakov]
2306
2307 *) Changes to avoid need for function casts in OpenSSL: some compilers
2308 (gcc 4.2 and later) reject their use.
2309 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
2310 Steve Henson]
2311
2312 *) Add RFC4507 support to OpenSSL. This includes the corrections in
2313 RFC4507bis. The encrypted ticket format is an encrypted encoded
2314 SSL_SESSION structure, that way new session features are automatically
2315 supported.
2316
2317 If a client application caches session in an SSL_SESSION structure
2318 support is transparent because tickets are now stored in the encoded
2319 SSL_SESSION.
2320
2321 The SSL_CTX structure automatically generates keys for ticket
2322 protection in servers so again support should be possible
2323 with no application modification.
2324
2325 If a client or server wishes to disable RFC4507 support then the option
2326 SSL_OP_NO_TICKET can be set.
2327
2328 Add a TLS extension debugging callback to allow the contents of any client
2329 or server extensions to be examined.
2330
2331 This work was sponsored by Google.
2332 [Steve Henson]
2333
2334 *) Add initial support for TLS extensions, specifically for the server_name
2335 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2336 have new members for a host name. The SSL data structure has an
2337 additional member SSL_CTX *initial_ctx so that new sessions can be
2338 stored in that context to allow for session resumption, even after the
2339 SSL has been switched to a new SSL_CTX in reaction to a client's
2340 server_name extension.
2341
2342 New functions (subject to change):
2343
2344 SSL_get_servername()
2345 SSL_get_servername_type()
2346 SSL_set_SSL_CTX()
2347
2348 New CTRL codes and macros (subject to change):
2349
2350 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2351 - SSL_CTX_set_tlsext_servername_callback()
2352 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2353 - SSL_CTX_set_tlsext_servername_arg()
2354 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2355
2356 openssl s_client has a new '-servername ...' option.
2357
2358 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2359 '-key2 ...', '-servername_fatal' (subject to change). This allows
2360 testing the HostName extension for a specific single host name ('-cert'
2361 and '-key' remain fallbacks for handshakes without HostName
2362 negotiation). If the unrecogninzed_name alert has to be sent, this by
2363 default is a warning; it becomes fatal with the '-servername_fatal'
2364 option.
2365
2366 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
2367
2368 *) Add AES and SSE2 assembly language support to VC++ build.
2369 [Steve Henson]
2370
2371 *) Mitigate attack on final subtraction in Montgomery reduction.
2372 [Andy Polyakov]
2373
2374 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
2375 (which previously caused an internal error).
2376 [Bodo Moeller]
2377
2378 *) Squeeze another 10% out of IGE mode when in != out.
2379 [Ben Laurie]
2380
2381 *) AES IGE mode speedup.
2382 [Dean Gaudet (Google)]
2383
2384 *) Add the Korean symmetric 128-bit cipher SEED (see
2385 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
2386 add SEED ciphersuites from RFC 4162:
2387
2388 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
2389 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
2390 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
2391 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
2392
2393 To minimize changes between patchlevels in the OpenSSL 0.9.8
2394 series, SEED remains excluded from compilation unless OpenSSL
2395 is configured with 'enable-seed'.
2396 [KISA, Bodo Moeller]
2397
2398 *) Mitigate branch prediction attacks, which can be practical if a
2399 single processor is shared, allowing a spy process to extract
2400 information. For detailed background information, see
2401 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
2402 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
2403 and Necessary Software Countermeasures"). The core of the change
2404 are new versions BN_div_no_branch() and
2405 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
2406 respectively, which are slower, but avoid the security-relevant
2407 conditional branches. These are automatically called by BN_div()
2408 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
2409 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
2410 remove a conditional branch.
2411
2412 BN_FLG_CONSTTIME is the new name for the previous
2413 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2414 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2415 in the exponent causes BN_mod_exp_mont() to use the alternative
2416 implementation in BN_mod_exp_mont_consttime().) The old name
2417 remains as a deprecated alias.
2418
2419 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2420 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2421 constant-time implementations for more than just exponentiation.
2422 Here too the old name is kept as a deprecated alias.
2423
2424 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2425 the BN_BLINDING structure gets an independent copy of the
2426 modulus. This means that the previous "BIGNUM *m" argument to
2427 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2428 essentially becomes "const BIGNUM *m", although we can't actually
2429 change this in the header file before 0.9.9. It allows
2430 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2431 enable BN_FLG_CONSTTIME.
2432
2433 [Matthew D Wood (Intel Corp)]
2434
2435 *) In the SSL/TLS server implementation, be strict about session ID
2436 context matching (which matters if an application uses a single
2437 external cache for different purposes). Previously,
2438 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2439 set. This did ensure strict client verification, but meant that,
2440 with applications using a single external cache for quite
2441 different requirements, clients could circumvent ciphersuite
2442 restrictions for a given session ID context by starting a session
2443 in a different context.
2444 [Bodo Moeller]
2445
2446 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2447 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2448 authentication-only ciphersuites.
2449 [Bodo Moeller]
2450
2451 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2452 not complete and could lead to a possible single byte overflow
2453 (CVE-2007-5135) [Ben Laurie]
2454
2455 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2456
2457 *) Since AES128 and AES256 (and similarly Camellia128 and
2458 Camellia256) share a single mask bit in the logic of
2459 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2460 kludge to work properly if AES128 is available and AES256 isn't
2461 (or if Camellia128 is available and Camellia256 isn't).
2462 [Victor Duchovni]
2463
2464 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2465 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2466 When a point or a seed is encoded in a BIT STRING, we need to
2467 prevent the removal of trailing zero bits to get the proper DER
2468 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2469 of a NamedBitList, for which trailing 0 bits need to be removed.)
2470 [Bodo Moeller]
2471
2472 *) Have SSL/TLS server implementation tolerate "mismatched" record
2473 protocol version while receiving ClientHello even if the
2474 ClientHello is fragmented. (The server can't insist on the
2475 particular protocol version it has chosen before the ServerHello
2476 message has informed the client about his choice.)
2477 [Bodo Moeller]
2478
2479 *) Add RFC 3779 support.
2480 [Rob Austein for ARIN, Ben Laurie]
2481
2482 *) Load error codes if they are not already present instead of using a
2483 static variable. This allows them to be cleanly unloaded and reloaded.
2484 Improve header file function name parsing.
2485 [Steve Henson]
2486
2487 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2488 or CAPABILITY handshake as required by RFCs.
2489 [Goetz Babin-Ebell]
2490
2491 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2492
2493 *) Introduce limits to prevent malicious keys being able to
2494 cause a denial of service. (CVE-2006-2940)
2495 [Steve Henson, Bodo Moeller]
2496
2497 *) Fix ASN.1 parsing of certain invalid structures that can result
2498 in a denial of service. (CVE-2006-2937) [Steve Henson]
2499
2500 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2501 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2502
2503 *) Fix SSL client code which could crash if connecting to a
2504 malicious SSLv2 server. (CVE-2006-4343)
2505 [Tavis Ormandy and Will Drewry, Google Security Team]
2506
2507 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2508 match only those. Before that, "AES256-SHA" would be interpreted
2509 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2510 the same strength classification in 0.9.7h) as we currently only
2511 have a single AES bit in the ciphersuite description bitmap.
2512 That change, however, also applied to ciphersuite strings such as
2513 "RC4-MD5" that intentionally matched multiple ciphersuites --
2514 namely, SSL 2.0 ciphersuites in addition to the more common ones
2515 from SSL 3.0/TLS 1.0.
2516
2517 So we change the selection algorithm again: Naming an explicit
2518 ciphersuite selects this one ciphersuite, and any other similar
2519 ciphersuite (same bitmap) from *other* protocol versions.
2520 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2521 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2522
2523 Since SSL 2.0 does not have any ciphersuites for which the
2524 128/256 bit distinction would be relevant, this works for now.
2525 The proper fix will be to use different bits for AES128 and
2526 AES256, which would have avoided the problems from the beginning;
2527 however, bits are scarce, so we can only do this in a new release
2528 (not just a patchlevel) when we can change the SSL_CIPHER
2529 definition to split the single 'unsigned long mask' bitmap into
2530 multiple values to extend the available space.
2531
2532 [Bodo Moeller]
2533
2534 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2535
2536 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2537 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2538
2539 *) Add AES IGE and biIGE modes.
2540 [Ben Laurie]
2541
2542 *) Change the Unix randomness entropy gathering to use poll() when
2543 possible instead of select(), since the latter has some
2544 undesirable limitations.
2545 [Darryl Miles via Richard Levitte and Bodo Moeller]
2546
2547 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2548 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2549 cannot be implicitly activated as part of, e.g., the "AES" alias.
2550 However, please upgrade to OpenSSL 0.9.9[-dev] for
2551 non-experimental use of the ECC ciphersuites to get TLS extension
2552 support, which is required for curve and point format negotiation
2553 to avoid potential handshake problems.
2554 [Bodo Moeller]
2555
2556 *) Disable rogue ciphersuites:
2557
2558 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2559 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2560 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2561
2562 The latter two were purportedly from
2563 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2564 appear there.
2565
2566 Also deactivate the remaining ciphersuites from
2567 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2568 unofficial, and the ID has long expired.
2569 [Bodo Moeller]
2570
2571 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2572 dual-core machines) and other potential thread-safety issues.
2573 [Bodo Moeller]
2574
2575 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2576 versions), which is now available for royalty-free use
2577 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2578 Also, add Camellia TLS ciphersuites from RFC 4132.
2579
2580 To minimize changes between patchlevels in the OpenSSL 0.9.8
2581 series, Camellia remains excluded from compilation unless OpenSSL
2582 is configured with 'enable-camellia'.
2583 [NTT]
2584
2585 *) Disable the padding bug check when compression is in use. The padding
2586 bug check assumes the first packet is of even length, this is not
2587 necessarily true if compresssion is enabled and can result in false
2588 positives causing handshake failure. The actual bug test is ancient
2589 code so it is hoped that implementations will either have fixed it by
2590 now or any which still have the bug do not support compression.
2591 [Steve Henson]
2592
2593 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2594
2595 *) When applying a cipher rule check to see if string match is an explicit
2596 cipher suite and only match that one cipher suite if it is.
2597 [Steve Henson]
2598
2599 *) Link in manifests for VC++ if needed.
2600 [Austin Ziegler <halostatue@gmail.com>]
2601
2602 *) Update support for ECC-based TLS ciphersuites according to
2603 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2604 TLS extensions, which are supported starting with the 0.9.9
2605 branch, not in the OpenSSL 0.9.8 branch).
2606 [Douglas Stebila]
2607
2608 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2609 opaque EVP_CIPHER_CTX handling.
2610 [Steve Henson]
2611
2612 *) Fixes and enhancements to zlib compression code. We now only use
2613 "zlib1.dll" and use the default __cdecl calling convention on Win32
2614 to conform with the standards mentioned here:
2615 http://www.zlib.net/DLL_FAQ.txt
2616 Static zlib linking now works on Windows and the new --with-zlib-include
2617 --with-zlib-lib options to Configure can be used to supply the location
2618 of the headers and library. Gracefully handle case where zlib library
2619 can't be loaded.
2620 [Steve Henson]
2621
2622 *) Several fixes and enhancements to the OID generation code. The old code
2623 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2624 handle numbers larger than ULONG_MAX, truncated printing and had a
2625 non standard OBJ_obj2txt() behaviour.
2626 [Steve Henson]
2627
2628 *) Add support for building of engines under engine/ as shared libraries
2629 under VC++ build system.
2630 [Steve Henson]
2631
2632 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2633 Hopefully, we will not see any false combination of paths any more.
2634 [Richard Levitte]
2635
2636 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2637
2638 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2639 (part of SSL_OP_ALL). This option used to disable the
2640 countermeasure against man-in-the-middle protocol-version
2641 rollback in the SSL 2.0 server implementation, which is a bad
2642 idea. (CVE-2005-2969)
2643
2644 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2645 for Information Security, National Institute of Advanced Industrial
2646 Science and Technology [AIST], Japan)]
2647
2648 *) Add two function to clear and return the verify parameter flags.
2649 [Steve Henson]
2650
2651 *) Keep cipherlists sorted in the source instead of sorting them at
2652 runtime, thus removing the need for a lock.
2653 [Nils Larsch]
2654
2655 *) Avoid some small subgroup attacks in Diffie-Hellman.
2656 [Nick Mathewson and Ben Laurie]
2657
2658 *) Add functions for well-known primes.
2659 [Nick Mathewson]
2660
2661 *) Extended Windows CE support.
2662 [Satoshi Nakamura and Andy Polyakov]
2663
2664 *) Initialize SSL_METHOD structures at compile time instead of during
2665 runtime, thus removing the need for a lock.
2666 [Steve Henson]
2667
2668 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2669 attempting to decrypt each encrypted key in turn. Add support to
2670 smime utility.
2671 [Steve Henson]
2672
2673 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2674
2675 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2676 OpenSSL 0.9.8.]
2677
2678 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2679 [Richard Levitte]
2680
2681 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2682 key into the same file any more.
2683 [Richard Levitte]
2684
2685 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2686 [Andy Polyakov]
2687
2688 *) Add -utf8 command line and config file option to 'ca'.
2689 [Stefan <stf@udoma.org]
2690
2691 *) Removed the macro des_crypt(), as it seems to conflict with some
2692 libraries. Use DES_crypt().
2693 [Richard Levitte]
2694
2695 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2696 involves renaming the source and generated shared-libs for
2697 both. The engines will accept the corrected or legacy ids
2698 ('ncipher' and '4758_cca' respectively) when binding. NB,
2699 this only applies when building 'shared'.
2700 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2701
2702 *) Add attribute functions to EVP_PKEY structure. Modify
2703 PKCS12_create() to recognize a CSP name attribute and
2704 use it. Make -CSP option work again in pkcs12 utility.
2705 [Steve Henson]
2706
2707 *) Add new functionality to the bn blinding code:
2708 - automatic re-creation of the BN_BLINDING parameters after
2709 a fixed number of uses (currently 32)
2710 - add new function for parameter creation
2711 - introduce flags to control the update behaviour of the
2712 BN_BLINDING parameters
2713 - hide BN_BLINDING structure
2714 Add a second BN_BLINDING slot to the RSA structure to improve
2715 performance when a single RSA object is shared among several
2716 threads.
2717 [Nils Larsch]
2718
2719 *) Add support for DTLS.
2720 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2721
2722 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2723 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2724 [Walter Goulet]
2725
2726 *) Remove buggy and incompletet DH cert support from
2727 ssl/ssl_rsa.c and ssl/s3_both.c
2728 [Nils Larsch]
2729
2730 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2731 the apps/openssl applications.
2732 [Nils Larsch]
2733
2734 *) Compile clean with "-Wall -Wmissing-prototypes
2735 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2736 DEBUG_SAFESTACK must also be set.
2737 [Ben Laurie]
2738
2739 *) Change ./Configure so that certain algorithms can be disabled by default.
2740 The new counterpiece to "no-xxx" is "enable-xxx".
2741
2742 The patented RC5 and MDC2 algorithms will now be disabled unless
2743 "enable-rc5" and "enable-mdc2", respectively, are specified.
2744
2745 (IDEA remains enabled despite being patented. This is because IDEA
2746 is frequently required for interoperability, and there is no license
2747 fee for non-commercial use. As before, "no-idea" can be used to
2748 avoid this algorithm.)
2749
2750 [Bodo Moeller]
2751
2752 *) Add processing of proxy certificates (see RFC 3820). This work was
2753 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2754 EGEE (Enabling Grids for E-science in Europe).
2755 [Richard Levitte]
2756
2757 *) RC4 performance overhaul on modern architectures/implementations, such
2758 as Intel P4, IA-64 and AMD64.
2759 [Andy Polyakov]
2760
2761 *) New utility extract-section.pl. This can be used specify an alternative
2762 section number in a pod file instead of having to treat each file as
2763 a separate case in Makefile. This can be done by adding two lines to the
2764 pod file:
2765
2766 =for comment openssl_section:XXX
2767
2768 The blank line is mandatory.
2769
2770 [Steve Henson]
2771
2772 *) New arguments -certform, -keyform and -pass for s_client and s_server
2773 to allow alternative format key and certificate files and passphrase
2774 sources.
2775 [Steve Henson]
2776
2777 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2778 update associated structures and add various utility functions.
2779
2780 Add new policy related verify parameters, include policy checking in
2781 standard verify code. Enhance 'smime' application with extra parameters
2782 to support policy checking and print out.
2783 [Steve Henson]
2784
2785 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2786 Nehemiah processors. These extensions support AES encryption in hardware
2787 as well as RNG (though RNG support is currently disabled).
2788 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2789
2790 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2791 [Geoff Thorpe]
2792
2793 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2794 [Andy Polyakov and a number of other people]
2795
2796 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2797 implementation contributed by IBM.
2798 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2799
2800 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2801 exponent rather than 'unsigned long'. There is a corresponding change to
2802 the new 'rsa_keygen' element of the RSA_METHOD structure.
2803 [Jelte Jansen, Geoff Thorpe]
2804
2805 *) Functionality for creating the initial serial number file is now
2806 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2807
2808 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2809 number file to 1, which is bound to cause problems. To avoid
2810 the problems while respecting compatibility between different 0.9.7
2811 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2812 CA.pl for serial number initialization. With the new release 0.9.8,
2813 we can fix the problem directly in the 'ca' utility.)
2814 [Steve Henson]
2815
2816 *) Reduced header interdepencies by declaring more opaque objects in
2817 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2818 give fewer recursive includes, which could break lazy source code - so
2819 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2820 developers should define this symbol when building and using openssl to
2821 ensure they track the recommended behaviour, interfaces, [etc], but
2822 backwards-compatible behaviour prevails when this isn't defined.
2823 [Geoff Thorpe]
2824
2825 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2826 [Steve Henson]
2827
2828 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2829 This will generate a random key of the appropriate length based on the
2830 cipher context. The EVP_CIPHER can provide its own random key generation
2831 routine to support keys of a specific form. This is used in the des and
2832 3des routines to generate a key of the correct parity. Update S/MIME
2833 code to use new functions and hence generate correct parity DES keys.
2834 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2835 valid (weak or incorrect parity).
2836 [Steve Henson]
2837
2838 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2839 as looking them up. This is useful when the verified structure may contain
2840 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2841 present unless the new PKCS7_NO_CRL flag is asserted.
2842 [Steve Henson]
2843
2844 *) Extend ASN1 oid configuration module. It now additionally accepts the
2845 syntax:
2846
2847 shortName = some long name, 1.2.3.4
2848 [Steve Henson]
2849
2850 *) Reimplemented the BN_CTX implementation. There is now no more static
2851 limitation on the number of variables it can handle nor the depth of the
2852 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2853 information can now expand as required, and rather than having a single
2854 static array of bignums, BN_CTX now uses a linked-list of such arrays
2855 allowing it to expand on demand whilst maintaining the usefulness of
2856 BN_CTX's "bundling".
2857 [Geoff Thorpe]
2858
2859 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2860 to allow all RSA operations to function using a single BN_CTX.
2861 [Geoff Thorpe]
2862
2863 *) Preliminary support for certificate policy evaluation and checking. This
2864 is initially intended to pass the tests outlined in "Conformance Testing
2865 of Relying Party Client Certificate Path Processing Logic" v1.07.
2866 [Steve Henson]
2867
2868 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2869 remained unused and not that useful. A variety of other little bignum
2870 tweaks and fixes have also been made continuing on from the audit (see
2871 below).
2872 [Geoff Thorpe]
2873
2874 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2875 associated ASN1, EVP and SSL functions and old ASN1 macros.
2876 [Richard Levitte]
2877
2878 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2879 and this should never fail. So the return value from the use of
2880 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2881 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2882 [Geoff Thorpe]
2883
2884 *) BN_CTX_get() should return zero-valued bignums, providing the same
2885 initialised value as BN_new().
2886 [Geoff Thorpe, suggested by Ulf Möller]
2887
2888 *) Support for inhibitAnyPolicy certificate extension.
2889 [Steve Henson]
2890
2891 *) An audit of the BIGNUM code is underway, for which debugging code is
2892 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2893 is considered valid when processing BIGNUMs, and causes execution to
2894 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2895 further steps are taken to deliberately pollute unused data in BIGNUM
2896 structures to try and expose faulty code further on. For now, openssl will
2897 (in its default mode of operation) continue to tolerate the inconsistent
2898 forms that it has tolerated in the past, but authors and packagers should
2899 consider trying openssl and their own applications when compiled with
2900 these debugging symbols defined. It will help highlight potential bugs in
2901 their own code, and will improve the test coverage for OpenSSL itself. At
2902 some point, these tighter rules will become openssl's default to improve
2903 maintainability, though the assert()s and other overheads will remain only
2904 in debugging configurations. See bn.h for more details.
2905 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2906
2907 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2908 that can only be obtained through BN_CTX_new() (which implicitly
2909 initialises it). The presence of this function only made it possible
2910 to overwrite an existing structure (and cause memory leaks).
2911 [Geoff Thorpe]
2912
2913 *) Because of the callback-based approach for implementing LHASH as a
2914 template type, lh_insert() adds opaque objects to hash-tables and
2915 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2916 to clean up those corresponding objects before destroying the hash table
2917 (and losing the object pointers). So some over-zealous constifications in
2918 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2919 objects as "const" and the lh_doall[_arg] callback wrappers are not
2920 prototyped to have "const" restrictions on the object pointers they are
2921 given (and so aren't required to cast them away any more).
2922 [Geoff Thorpe]
2923
2924 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2925 (speed) prefers to use its own implementation. The two implementations
2926 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2927 its object type properly exposed (MS_TM) instead of casting to/from "char
2928 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2929 aren't necessarily the greatest nomenclatures - but this is what was used
2930 internally to the implementation so I've used that for now.
2931 [Geoff Thorpe]
2932
2933 *) Ensure that deprecated functions do not get compiled when
2934 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2935 the self-tests were still using deprecated key-generation functions so
2936 these have been updated also.
2937 [Geoff Thorpe]
2938
2939 *) Reorganise PKCS#7 code to separate the digest location functionality
2940 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2941 New function PKCS7_set_digest() to set the digest type for PKCS#7
2942 digestedData type. Add additional code to correctly generate the
2943 digestedData type and add support for this type in PKCS7 initialization
2944 functions.
2945 [Steve Henson]
2946
2947 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2948 structure of type "other".
2949 [Steve Henson]
2950
2951 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2952 sure the loop does correctly stop and breaking ("division by zero")
2953 modulus operations are not performed. The (pre-generated) prime
2954 table crypto/bn/bn_prime.h was already correct, but it could not be
2955 re-generated on some platforms because of the "division by zero"
2956 situation in the script.
2957 [Ralf S. Engelschall]
2958
2959 *) Update support for ECC-based TLS ciphersuites according to
2960 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2961 SHA-1 now is only used for "small" curves (where the
2962 representation of a field element takes up to 24 bytes); for
2963 larger curves, the field element resulting from ECDH is directly
2964 used as premaster secret.
2965 [Douglas Stebila (Sun Microsystems Laboratories)]
2966
2967 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2968 curve secp160r1 to the tests.
2969 [Douglas Stebila (Sun Microsystems Laboratories)]
2970
2971 *) Add the possibility to load symbols globally with DSO.
2972 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2973
2974 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2975 control of the error stack.
2976 [Richard Levitte]
2977
2978 *) Add support for STORE in ENGINE.
2979 [Richard Levitte]
2980
2981 *) Add the STORE type. The intention is to provide a common interface
2982 to certificate and key stores, be they simple file-based stores, or
2983 HSM-type store, or LDAP stores, or...
2984 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2985 [Richard Levitte]
2986
2987 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2988 pass a list of arguments to any function as well as provide a way
2989 for a function to pass data back to the caller.
2990 [Richard Levitte]
2991
2992 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2993 works like BUF_strdup() but can be used to duplicate a portion of
2994 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2995 a memory area.
2996 [Richard Levitte]
2997
2998 *) Add the function sk_find_ex() which works like sk_find(), but will
2999 return an index to an element even if an exact match couldn't be
3000 found. The index is guaranteed to point at the element where the
3001 searched-for key would be inserted to preserve sorting order.
3002 [Richard Levitte]
3003
3004 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3005 takes an extra flags argument for optional functionality. Currently,
3006 the following flags are defined:
3007
3008 OBJ_BSEARCH_VALUE_ON_NOMATCH
3009 This one gets OBJ_bsearch_ex() to return a pointer to the first
3010 element where the comparing function returns a negative or zero
3011 number.
3012
3013 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3014 This one gets OBJ_bsearch_ex() to return a pointer to the first
3015 element where the comparing function returns zero. This is useful
3016 if there are more than one element where the comparing function
3017 returns zero.
3018 [Richard Levitte]
3019
3020 *) Make it possible to create self-signed certificates with 'openssl ca'
3021 in such a way that the self-signed certificate becomes part of the
3022 CA database and uses the same mechanisms for serial number generation
3023 as all other certificate signing. The new flag '-selfsign' enables
3024 this functionality. Adapt CA.sh and CA.pl.in.
3025 [Richard Levitte]
3026
3027 *) Add functionality to check the public key of a certificate request
3028 against a given private. This is useful to check that a certificate
3029 request can be signed by that key (self-signing).
3030 [Richard Levitte]
3031
3032 *) Make it possible to have multiple active certificates with the same
3033 subject in the CA index file. This is done only if the keyword
3034 'unique_subject' is set to 'no' in the main CA section (default
3035 if 'CA_default') of the configuration file. The value is saved
3036 with the database itself in a separate index attribute file,
3037 named like the index file with '.attr' appended to the name.
3038 [Richard Levitte]
3039
3040 *) Generate muti valued AVAs using '+' notation in config files for
3041 req and dirName.
3042 [Steve Henson]
3043
3044 *) Support for nameConstraints certificate extension.
3045 [Steve Henson]
3046
3047 *) Support for policyConstraints certificate extension.
3048 [Steve Henson]
3049
3050 *) Support for policyMappings certificate extension.
3051 [Steve Henson]
3052
3053 *) Make sure the default DSA_METHOD implementation only uses its
3054 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3055 and change its own handlers to be NULL so as to remove unnecessary
3056 indirection. This lets alternative implementations fallback to the
3057 default implementation more easily.
3058 [Geoff Thorpe]
3059
3060 *) Support for directoryName in GeneralName related extensions
3061 in config files.
3062 [Steve Henson]
3063
3064 *) Make it possible to link applications using Makefile.shared.
3065 Make that possible even when linking against static libraries!
3066 [Richard Levitte]
3067
3068 *) Support for single pass processing for S/MIME signing. This now
3069 means that S/MIME signing can be done from a pipe, in addition
3070 cleartext signing (multipart/signed type) is effectively streaming
3071 and the signed data does not need to be all held in memory.
3072
3073 This is done with a new flag PKCS7_STREAM. When this flag is set
3074 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3075 is done after the data is output (and digests calculated) in
3076 SMIME_write_PKCS7().
3077 [Steve Henson]
3078
3079 *) Add full support for -rpath/-R, both in shared libraries and
3080 applications, at least on the platforms where it's known how
3081 to do it.
3082 [Richard Levitte]
3083
3084 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3085 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3086 will now compute a table of multiples of the generator that
3087 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3088 faster (notably in the case of a single point multiplication,
3089 scalar * generator).
3090 [Nils Larsch, Bodo Moeller]
3091
3092 *) IPv6 support for certificate extensions. The various extensions
3093 which use the IP:a.b.c.d can now take IPv6 addresses using the
3094 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3095 correctly.
3096 [Steve Henson]
3097
3098 *) Added an ENGINE that implements RSA by performing private key
3099 exponentiations with the GMP library. The conversions to and from
3100 GMP's mpz_t format aren't optimised nor are any montgomery forms
3101 cached, and on x86 it appears OpenSSL's own performance has caught up.
3102 However there are likely to be other architectures where GMP could
3103 provide a boost. This ENGINE is not built in by default, but it can be
3104 specified at Configure time and should be accompanied by the necessary
3105 linker additions, eg;
3106 ./config -DOPENSSL_USE_GMP -lgmp
3107 [Geoff Thorpe]
3108
3109 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3110 testing availability of engines with "-t" - the old behaviour is
3111 produced by increasing the feature's verbosity with "-tt".
3112 [Geoff Thorpe]
3113
3114 *) ECDSA routines: under certain error conditions uninitialized BN objects
3115 could be freed. Solution: make sure initialization is performed early
3116 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3117 via PR#459)
3118 [Lutz Jaenicke]
3119
3120 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3121 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3122 software implementations. For DSA and DH, parameter generation can
3123 also be overriden by providing the appropriate method callbacks.
3124 [Geoff Thorpe]
3125
3126 *) Change the "progress" mechanism used in key-generation and
3127 primality testing to functions that take a new BN_GENCB pointer in
3128 place of callback/argument pairs. The new API functions have "_ex"
3129 postfixes and the older functions are reimplemented as wrappers for
3130 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3131 declarations of the old functions to help (graceful) attempts to
3132 migrate to the new functions. Also, the new key-generation API
3133 functions operate on a caller-supplied key-structure and return
3134 success/failure rather than returning a key or NULL - this is to
3135 help make "keygen" another member function of RSA_METHOD etc.
3136
3137 Example for using the new callback interface:
3138
3139 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3140 void *my_arg = ...;
3141 BN_GENCB my_cb;
3142
3143 BN_GENCB_set(&my_cb, my_callback, my_arg);
3144
3145 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3146 /* For the meaning of a, b in calls to my_callback(), see the
3147 * documentation of the function that calls the callback.
3148 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3149 * my_callback should return 1 if it wants BN_is_prime_ex()
3150 * to continue, or 0 to stop.
3151 */
3152
3153 [Geoff Thorpe]
3154
3155 *) Change the ZLIB compression method to be stateful, and make it
3156 available to TLS with the number defined in
3157 draft-ietf-tls-compression-04.txt.
3158 [Richard Levitte]
3159
3160 *) Add the ASN.1 structures and functions for CertificatePair, which
3161 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3162
3163 CertificatePair ::= SEQUENCE {
3164 forward [0] Certificate OPTIONAL,
3165 reverse [1] Certificate OPTIONAL,
3166 -- at least one of the pair shall be present -- }
3167
3168 Also implement the PEM functions to read and write certificate
3169 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3170
3171 This needed to be defined, mostly for the sake of the LDAP
3172 attribute crossCertificatePair, but may prove useful elsewhere as
3173 well.
3174 [Richard Levitte]
3175
3176 *) Make it possible to inhibit symlinking of shared libraries in
3177 Makefile.shared, for Cygwin's sake.
3178 [Richard Levitte]
3179
3180 *) Extend the BIGNUM API by creating a function
3181 void BN_set_negative(BIGNUM *a, int neg);
3182 and a macro that behave like
3183 int BN_is_negative(const BIGNUM *a);
3184
3185 to avoid the need to access 'a->neg' directly in applications.
3186 [Nils Larsch]
3187
3188 *) Implement fast modular reduction for pseudo-Mersenne primes
3189 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3190 EC_GROUP_new_curve_GFp() will now automatically use this
3191 if applicable.
3192 [Nils Larsch <nla@trustcenter.de>]
3193
3194 *) Add new lock type (CRYPTO_LOCK_BN).
3195 [Bodo Moeller]
3196
3197 *) Change the ENGINE framework to automatically load engines
3198 dynamically from specific directories unless they could be
3199 found to already be built in or loaded. Move all the
3200 current engines except for the cryptodev one to a new
3201 directory engines/.
3202 The engines in engines/ are built as shared libraries if
3203 the "shared" options was given to ./Configure or ./config.
3204 Otherwise, they are inserted in libcrypto.a.
3205 /usr/local/ssl/engines is the default directory for dynamic
3206 engines, but that can be overriden at configure time through
3207 the usual use of --prefix and/or --openssldir, and at run
3208 time with the environment variable OPENSSL_ENGINES.
3209 [Geoff Thorpe and Richard Levitte]
3210
3211 *) Add Makefile.shared, a helper makefile to build shared
3212 libraries. Addapt Makefile.org.
3213 [Richard Levitte]
3214
3215 *) Add version info to Win32 DLLs.
3216 [Peter 'Luna' Runestig" <peter@runestig.com>]
3217
3218 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3219 can be added using this API to created arbitrary PKCS#12
3220 files while avoiding the low level API.
3221
3222 New options to PKCS12_create(), key or cert can be NULL and
3223 will then be omitted from the output file. The encryption
3224 algorithm NIDs can be set to -1 for no encryption, the mac
3225 iteration count can be set to 0 to omit the mac.
3226
3227 Enhance pkcs12 utility by making the -nokeys and -nocerts
3228 options work when creating a PKCS#12 file. New option -nomac
3229 to omit the mac, NONE can be set for an encryption algorithm.
3230 New code is modified to use the enhanced PKCS12_create()
3231 instead of the low level API.
3232 [Steve Henson]
3233
3234 *) Extend ASN1 encoder to support indefinite length constructed
3235 encoding. This can output sequences tags and octet strings in
3236 this form. Modify pk7_asn1.c to support indefinite length
3237 encoding. This is experimental and needs additional code to
3238 be useful, such as an ASN1 bio and some enhanced streaming
3239 PKCS#7 code.
3240
3241 Extend template encode functionality so that tagging is passed
3242 down to the template encoder.
3243 [Steve Henson]
3244
3245 *) Let 'openssl req' fail if an argument to '-newkey' is not
3246 recognized instead of using RSA as a default.
3247 [Bodo Moeller]
3248
3249 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3250 As these are not official, they are not included in "ALL";
3251 the "ECCdraft" ciphersuite group alias can be used to select them.
3252 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3253
3254 *) Add ECDH engine support.
3255 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3256
3257 *) Add ECDH in new directory crypto/ecdh/.
3258 [Douglas Stebila (Sun Microsystems Laboratories)]
3259
3260 *) Let BN_rand_range() abort with an error after 100 iterations
3261 without success (which indicates a broken PRNG).
3262 [Bodo Moeller]
3263
3264 *) Change BN_mod_sqrt() so that it verifies that the input value
3265 is really the square of the return value. (Previously,
3266 BN_mod_sqrt would show GIGO behaviour.)
3267 [Bodo Moeller]
3268
3269 *) Add named elliptic curves over binary fields from X9.62, SECG,
3270 and WAP/WTLS; add OIDs that were still missing.
3271
3272 [Sheueling Chang Shantz and Douglas Stebila
3273 (Sun Microsystems Laboratories)]
3274
3275 *) Extend the EC library for elliptic curves over binary fields
3276 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
3277 New EC_METHOD:
3278
3279 EC_GF2m_simple_method
3280
3281 New API functions:
3282
3283 EC_GROUP_new_curve_GF2m
3284 EC_GROUP_set_curve_GF2m
3285 EC_GROUP_get_curve_GF2m
3286 EC_POINT_set_affine_coordinates_GF2m
3287 EC_POINT_get_affine_coordinates_GF2m
3288 EC_POINT_set_compressed_coordinates_GF2m
3289
3290 Point compression for binary fields is disabled by default for
3291 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
3292 enable it).
3293
3294 As binary polynomials are represented as BIGNUMs, various members
3295 of the EC_GROUP and EC_POINT data structures can be shared
3296 between the implementations for prime fields and binary fields;
3297 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
3298 are essentially identical to their ..._GFp counterparts.
3299 (For simplicity, the '..._GFp' prefix has been dropped from
3300 various internal method names.)
3301
3302 An internal 'field_div' method (similar to 'field_mul' and
3303 'field_sqr') has been added; this is used only for binary fields.
3304
3305 [Sheueling Chang Shantz and Douglas Stebila
3306 (Sun Microsystems Laboratories)]
3307
3308 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
3309 through methods ('mul', 'precompute_mult').
3310
3311 The generic implementations (now internally called 'ec_wNAF_mul'
3312 and 'ec_wNAF_precomputed_mult') remain the default if these
3313 methods are undefined.
3314
3315 [Sheueling Chang Shantz and Douglas Stebila
3316 (Sun Microsystems Laboratories)]
3317
3318 *) New function EC_GROUP_get_degree, which is defined through
3319 EC_METHOD. For curves over prime fields, this returns the bit
3320 length of the modulus.
3321
3322 [Sheueling Chang Shantz and Douglas Stebila
3323 (Sun Microsystems Laboratories)]
3324
3325 *) New functions EC_GROUP_dup, EC_POINT_dup.
3326 (These simply call ..._new and ..._copy).
3327
3328 [Sheueling Chang Shantz and Douglas Stebila
3329 (Sun Microsystems Laboratories)]
3330
3331 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
3332 Polynomials are represented as BIGNUMs (where the sign bit is not
3333 used) in the following functions [macros]:
3334
3335 BN_GF2m_add
3336 BN_GF2m_sub [= BN_GF2m_add]
3337 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
3338 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
3339 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
3340 BN_GF2m_mod_inv
3341 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
3342 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
3343 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
3344 BN_GF2m_cmp [= BN_ucmp]
3345
3346 (Note that only the 'mod' functions are actually for fields GF(2^m).
3347 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
3348
3349 For some functions, an the irreducible polynomial defining a
3350 field can be given as an 'unsigned int[]' with strictly
3351 decreasing elements giving the indices of those bits that are set;
3352 i.e., p[] represents the polynomial
3353 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
3354 where
3355 p[0] > p[1] > ... > p[k] = 0.
3356 This applies to the following functions:
3357
3358 BN_GF2m_mod_arr
3359 BN_GF2m_mod_mul_arr
3360 BN_GF2m_mod_sqr_arr
3361 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
3362 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
3363 BN_GF2m_mod_exp_arr
3364 BN_GF2m_mod_sqrt_arr
3365 BN_GF2m_mod_solve_quad_arr
3366 BN_GF2m_poly2arr
3367 BN_GF2m_arr2poly
3368
3369 Conversion can be performed by the following functions:
3370
3371 BN_GF2m_poly2arr
3372 BN_GF2m_arr2poly
3373
3374 bntest.c has additional tests for binary polynomial arithmetic.
3375
3376 Two implementations for BN_GF2m_mod_div() are available.
3377 The default algorithm simply uses BN_GF2m_mod_inv() and
3378 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
3379 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
3380 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
3381
3382 [Sheueling Chang Shantz and Douglas Stebila
3383 (Sun Microsystems Laboratories)]
3384
3385 *) Add new error code 'ERR_R_DISABLED' that can be used when some
3386 functionality is disabled at compile-time.
3387 [Douglas Stebila <douglas.stebila@sun.com>]
3388
3389 *) Change default behaviour of 'openssl asn1parse' so that more
3390 information is visible when viewing, e.g., a certificate:
3391
3392 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
3393 mode the content of non-printable OCTET STRINGs is output in a
3394 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
3395 avoid the appearance of a printable string.
3396 [Nils Larsch <nla@trustcenter.de>]
3397
3398 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
3399 functions
3400 EC_GROUP_set_asn1_flag()
3401 EC_GROUP_get_asn1_flag()
3402 EC_GROUP_set_point_conversion_form()
3403 EC_GROUP_get_point_conversion_form()
3404 These control ASN1 encoding details:
3405 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
3406 has been set to OPENSSL_EC_NAMED_CURVE.
3407 - Points are encoded in uncompressed form by default; options for
3408 asn1_for are as for point2oct, namely
3409 POINT_CONVERSION_COMPRESSED
3410 POINT_CONVERSION_UNCOMPRESSED
3411 POINT_CONVERSION_HYBRID
3412
3413 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3414 functions
3415 EC_GROUP_set_seed()
3416 EC_GROUP_get0_seed()
3417 EC_GROUP_get_seed_len()
3418 This is used only for ASN1 purposes (so far).
3419 [Nils Larsch <nla@trustcenter.de>]
3420
3421 *) Add 'field_type' member to EC_METHOD, which holds the NID
3422 of the appropriate field type OID. The new function
3423 EC_METHOD_get_field_type() returns this value.
3424 [Nils Larsch <nla@trustcenter.de>]
3425
3426 *) Add functions
3427 EC_POINT_point2bn()
3428 EC_POINT_bn2point()
3429 EC_POINT_point2hex()
3430 EC_POINT_hex2point()
3431 providing useful interfaces to EC_POINT_point2oct() and
3432 EC_POINT_oct2point().
3433 [Nils Larsch <nla@trustcenter.de>]
3434
3435 *) Change internals of the EC library so that the functions
3436 EC_GROUP_set_generator()
3437 EC_GROUP_get_generator()
3438 EC_GROUP_get_order()
3439 EC_GROUP_get_cofactor()
3440 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3441 to methods, which would lead to unnecessary code duplication when
3442 adding different types of curves.
3443 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3444
3445 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3446 arithmetic, and such that modified wNAFs are generated
3447 (which avoid length expansion in many cases).
3448 [Bodo Moeller]
3449
3450 *) Add a function EC_GROUP_check_discriminant() (defined via
3451 EC_METHOD) that verifies that the curve discriminant is non-zero.
3452
3453 Add a function EC_GROUP_check() that makes some sanity tests
3454 on a EC_GROUP, its generator and order. This includes
3455 EC_GROUP_check_discriminant().
3456 [Nils Larsch <nla@trustcenter.de>]
3457
3458 *) Add ECDSA in new directory crypto/ecdsa/.
3459
3460 Add applications 'openssl ecparam' and 'openssl ecdsa'
3461 (these are based on 'openssl dsaparam' and 'openssl dsa').
3462
3463 ECDSA support is also included in various other files across the
3464 library. Most notably,
3465 - 'openssl req' now has a '-newkey ecdsa:file' option;
3466 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3467 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3468 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3469 them suitable for ECDSA where domain parameters must be
3470 extracted before the specific public key;
3471 - ECDSA engine support has been added.
3472 [Nils Larsch <nla@trustcenter.de>]
3473
3474 *) Include some named elliptic curves, and add OIDs from X9.62,
3475 SECG, and WAP/WTLS. Each curve can be obtained from the new
3476 function
3477 EC_GROUP_new_by_curve_name(),
3478 and the list of available named curves can be obtained with
3479 EC_get_builtin_curves().
3480 Also add a 'curve_name' member to EC_GROUP objects, which can be
3481 accessed via
3482 EC_GROUP_set_curve_name()
3483 EC_GROUP_get_curve_name()
3484 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3485
3486 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3487 was actually never needed) and in BN_mul(). The removal in BN_mul()
3488 required a small change in bn_mul_part_recursive() and the addition
3489 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3490 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3491 bn_sub_words() and bn_add_words() except they take arrays with
3492 differing sizes.
3493 [Richard Levitte]
3494
3495 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3496
3497 *) Cleanse PEM buffers before freeing them since they may contain
3498 sensitive data.
3499 [Benjamin Bennett <ben@psc.edu>]
3500
3501 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3502 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3503 authentication-only ciphersuites.
3504 [Bodo Moeller]
3505
3506 *) Since AES128 and AES256 share a single mask bit in the logic of
3507 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3508 kludge to work properly if AES128 is available and AES256 isn't.
3509 [Victor Duchovni]
3510
3511 *) Expand security boundary to match 1.1.1 module.
3512 [Steve Henson]
3513
3514 *) Remove redundant features: hash file source, editing of test vectors
3515 modify fipsld to use external fips_premain.c signature.
3516 [Steve Henson]
3517
3518 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3519 run algorithm test programs.
3520 [Steve Henson]
3521
3522 *) Make algorithm test programs more tolerant of whitespace.
3523 [Steve Henson]
3524
3525 *) Have SSL/TLS server implementation tolerate "mismatched" record
3526 protocol version while receiving ClientHello even if the
3527 ClientHello is fragmented. (The server can't insist on the
3528 particular protocol version it has chosen before the ServerHello
3529 message has informed the client about his choice.)
3530 [Bodo Moeller]
3531
3532 *) Load error codes if they are not already present instead of using a
3533 static variable. This allows them to be cleanly unloaded and reloaded.
3534 [Steve Henson]
3535
3536 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3537
3538 *) Introduce limits to prevent malicious keys being able to
3539 cause a denial of service. (CVE-2006-2940)
3540 [Steve Henson, Bodo Moeller]
3541
3542 *) Fix ASN.1 parsing of certain invalid structures that can result
3543 in a denial of service. (CVE-2006-2937) [Steve Henson]
3544
3545 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3546 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3547
3548 *) Fix SSL client code which could crash if connecting to a
3549 malicious SSLv2 server. (CVE-2006-4343)
3550 [Tavis Ormandy and Will Drewry, Google Security Team]
3551
3552 *) Change ciphersuite string processing so that an explicit
3553 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3554 will no longer include "AES128-SHA"), and any other similar
3555 ciphersuite (same bitmap) from *other* protocol versions (so that
3556 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3557 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3558 changes from 0.9.8b and 0.9.8d.
3559 [Bodo Moeller]
3560
3561 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3562
3563 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3564 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3565
3566 *) Change the Unix randomness entropy gathering to use poll() when
3567 possible instead of select(), since the latter has some
3568 undesirable limitations.
3569 [Darryl Miles via Richard Levitte and Bodo Moeller]
3570
3571 *) Disable rogue ciphersuites:
3572
3573 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3574 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3575 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3576
3577 The latter two were purportedly from
3578 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3579 appear there.
3580
3581 Also deactive the remaining ciphersuites from
3582 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3583 unofficial, and the ID has long expired.
3584 [Bodo Moeller]
3585
3586 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3587 dual-core machines) and other potential thread-safety issues.
3588 [Bodo Moeller]
3589
3590 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3591
3592 *) Adapt fipsld and the build system to link against the validated FIPS
3593 module in FIPS mode.
3594 [Steve Henson]
3595
3596 *) Fixes for VC++ 2005 build under Windows.
3597 [Steve Henson]
3598
3599 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3600 from a Windows bash shell such as MSYS. It is autodetected from the
3601 "config" script when run from a VC++ environment. Modify standard VC++
3602 build to use fipscanister.o from the GNU make build.
3603 [Steve Henson]
3604
3605 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3606
3607 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3608 The value now differs depending on if you build for FIPS or not.
3609 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3610 safely run with a non-FIPSed libcrypto, as it may crash because of
3611 the difference induced by this change.
3612 [Andy Polyakov]
3613
3614 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3615
3616 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3617 (part of SSL_OP_ALL). This option used to disable the
3618 countermeasure against man-in-the-middle protocol-version
3619 rollback in the SSL 2.0 server implementation, which is a bad
3620 idea. (CVE-2005-2969)
3621
3622 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3623 for Information Security, National Institute of Advanced Industrial
3624 Science and Technology [AIST], Japan)]
3625
3626 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3627 mainly for FIPS compliance and not fully integrated at this stage.
3628 [Steve Henson]
3629
3630 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3631 the exponentiation using a fixed-length exponent. (Otherwise,
3632 the information leaked through timing could expose the secret key
3633 after many signatures; cf. Bleichenbacher's attack on DSA with
3634 biased k.)
3635 [Bodo Moeller]
3636
3637 *) Make a new fixed-window mod_exp implementation the default for
3638 RSA, DSA, and DH private-key operations so that the sequence of
3639 squares and multiplies and the memory access pattern are
3640 independent of the particular secret key. This will mitigate
3641 cache-timing and potential related attacks.
3642
3643 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3644 and this is automatically used by BN_mod_exp_mont() if the new flag
3645 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3646 will use this BN flag for private exponents unless the flag
3647 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3648 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3649
3650 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3651
3652 *) Change the client implementation for SSLv23_method() and
3653 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3654 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3655 (Previously, the SSL 2.0 backwards compatible Client Hello
3656 message format would be used even with SSL_OP_NO_SSLv2.)
3657 [Bodo Moeller]
3658
3659 *) Add support for smime-type MIME parameter in S/MIME messages which some
3660 clients need.
3661 [Steve Henson]
3662
3663 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3664 a threadsafe manner. Modify rsa code to use new function and add calls
3665 to dsa and dh code (which had race conditions before).
3666 [Steve Henson]
3667
3668 *) Include the fixed error library code in the C error file definitions
3669 instead of fixing them up at runtime. This keeps the error code
3670 structures constant.
3671 [Steve Henson]
3672
3673 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3674
3675 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3676 OpenSSL 0.9.8.]
3677
3678 *) Fixes for newer kerberos headers. NB: the casts are needed because
3679 the 'length' field is signed on one version and unsigned on another
3680 with no (?) obvious way to tell the difference, without these VC++
3681 complains. Also the "definition" of FAR (blank) is no longer included
3682 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3683 some needed definitions.
3684 [Steve Henson]
3685
3686 *) Undo Cygwin change.
3687 [Ulf Möller]
3688
3689 *) Added support for proxy certificates according to RFC 3820.
3690 Because they may be a security thread to unaware applications,
3691 they must be explicitely allowed in run-time. See
3692 docs/HOWTO/proxy_certificates.txt for further information.
3693 [Richard Levitte]
3694
3695 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3696
3697 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3698 server and client random values. Previously
3699 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3700 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3701
3702 This change has negligible security impact because:
3703
3704 1. Server and client random values still have 24 bytes of pseudo random
3705 data.
3706
3707 2. Server and client random values are sent in the clear in the initial
3708 handshake.
3709
3710 3. The master secret is derived using the premaster secret (48 bytes in
3711 size for static RSA ciphersuites) as well as client server and random
3712 values.
3713
3714 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3715 to our attention.
3716
3717 [Stephen Henson, reported by UK NISCC]
3718
3719 *) Use Windows randomness collection on Cygwin.
3720 [Ulf Möller]
3721
3722 *) Fix hang in EGD/PRNGD query when communication socket is closed
3723 prematurely by EGD/PRNGD.
3724 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3725
3726 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3727 [Steve Henson]
3728
3729 *) Back-port of selected performance improvements from development
3730 branch, as well as improved support for PowerPC platforms.
3731 [Andy Polyakov]
3732
3733 *) Add lots of checks for memory allocation failure, error codes to indicate
3734 failure and freeing up memory if a failure occurs.
3735 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3736
3737 *) Add new -passin argument to dgst.
3738 [Steve Henson]
3739
3740 *) Perform some character comparisons of different types in X509_NAME_cmp:
3741 this is needed for some certificates that reencode DNs into UTF8Strings
3742 (in violation of RFC3280) and can't or wont issue name rollover
3743 certificates.
3744 [Steve Henson]
3745
3746 *) Make an explicit check during certificate validation to see that
3747 the CA setting in each certificate on the chain is correct. As a
3748 side effect always do the following basic checks on extensions,
3749 not just when there's an associated purpose to the check:
3750
3751 - if there is an unhandled critical extension (unless the user
3752 has chosen to ignore this fault)
3753 - if the path length has been exceeded (if one is set at all)
3754 - that certain extensions fit the associated purpose (if one has
3755 been given)
3756 [Richard Levitte]
3757
3758 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3759
3760 *) Avoid a race condition when CRLs are checked in a multi threaded
3761 environment. This would happen due to the reordering of the revoked
3762 entries during signature checking and serial number lookup. Now the
3763 encoding is cached and the serial number sort performed under a lock.
3764 Add new STACK function sk_is_sorted().
3765 [Steve Henson]
3766
3767 *) Add Delta CRL to the extension code.
3768 [Steve Henson]
3769
3770 *) Various fixes to s3_pkt.c so alerts are sent properly.
3771 [David Holmes <d.holmes@f5.com>]
3772
3773 *) Reduce the chances of duplicate issuer name and serial numbers (in
3774 violation of RFC3280) using the OpenSSL certificate creation utilities.
3775 This is done by creating a random 64 bit value for the initial serial
3776 number when a serial number file is created or when a self signed
3777 certificate is created using 'openssl req -x509'. The initial serial
3778 number file is created using 'openssl x509 -next_serial' in CA.pl
3779 rather than being initialized to 1.
3780 [Steve Henson]
3781
3782 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3783
3784 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3785 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3786 [Joe Orton, Steve Henson]
3787
3788 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3789 (CVE-2004-0112)
3790 [Joe Orton, Steve Henson]
3791
3792 *) Make it possible to have multiple active certificates with the same
3793 subject in the CA index file. This is done only if the keyword
3794 'unique_subject' is set to 'no' in the main CA section (default
3795 if 'CA_default') of the configuration file. The value is saved
3796 with the database itself in a separate index attribute file,
3797 named like the index file with '.attr' appended to the name.
3798 [Richard Levitte]
3799
3800 *) X509 verify fixes. Disable broken certificate workarounds when
3801 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3802 keyUsage extension present. Don't accept CRLs with unhandled critical
3803 extensions: since verify currently doesn't process CRL extensions this
3804 rejects a CRL with *any* critical extensions. Add new verify error codes
3805 for these cases.
3806 [Steve Henson]
3807
3808 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3809 A clarification of RFC2560 will require the use of OCTET STRINGs and
3810 some implementations cannot handle the current raw format. Since OpenSSL
3811 copies and compares OCSP nonces as opaque blobs without any attempt at
3812 parsing them this should not create any compatibility issues.
3813 [Steve Henson]
3814
3815 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3816 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3817 this HMAC (and other) operations are several times slower than OpenSSL
3818 < 0.9.7.
3819 [Steve Henson]
3820
3821 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3822 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3823
3824 *) Use the correct content when signing type "other".
3825 [Steve Henson]
3826
3827 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3828
3829 *) Fix various bugs revealed by running the NISCC test suite:
3830
3831 Stop out of bounds reads in the ASN1 code when presented with
3832 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3833
3834 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3835
3836 If verify callback ignores invalid public key errors don't try to check
3837 certificate signature with the NULL public key.
3838
3839 [Steve Henson]
3840
3841 *) New -ignore_err option in ocsp application to stop the server
3842 exiting on the first error in a request.
3843 [Steve Henson]
3844
3845 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3846 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3847 specifications.
3848 [Steve Henson]
3849
3850 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3851 extra data after the compression methods not only for TLS 1.0
3852 but also for SSL 3.0 (as required by the specification).
3853 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3854
3855 *) Change X509_certificate_type() to mark the key as exported/exportable
3856 when it's 512 *bits* long, not 512 bytes.
3857 [Richard Levitte]
3858
3859 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3860 blocks during encryption.
3861 [Richard Levitte]
3862
3863 *) Various fixes to base64 BIO and non blocking I/O. On write
3864 flushes were not handled properly if the BIO retried. On read
3865 data was not being buffered properly and had various logic bugs.
3866 This also affects blocking I/O when the data being decoded is a
3867 certain size.
3868 [Steve Henson]
3869
3870 *) Various S/MIME bugfixes and compatibility changes:
3871 output correct application/pkcs7 MIME type if
3872 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3873 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3874 of files as .eml work). Correctly handle very long lines in MIME
3875 parser.
3876 [Steve Henson]
3877
3878 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3879
3880 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3881 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3882 a protocol version number mismatch like a decryption error
3883 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3884 [Bodo Moeller]
3885
3886 *) Turn on RSA blinding by default in the default implementation
3887 to avoid a timing attack. Applications that don't want it can call
3888 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3889 They would be ill-advised to do so in most cases.
3890 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3891
3892 *) Change RSA blinding code so that it works when the PRNG is not
3893 seeded (in this case, the secret RSA exponent is abused as
3894 an unpredictable seed -- if it is not unpredictable, there
3895 is no point in blinding anyway). Make RSA blinding thread-safe
3896 by remembering the creator's thread ID in rsa->blinding and
3897 having all other threads use local one-time blinding factors
3898 (this requires more computation than sharing rsa->blinding, but
3899 avoids excessive locking; and if an RSA object is not shared
3900 between threads, blinding will still be very fast).
3901 [Bodo Moeller]
3902
3903 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3904 ENGINE as defaults for all supported algorithms irrespective of
3905 the 'flags' parameter. 'flags' is now honoured, so applications
3906 should make sure they are passing it correctly.
3907 [Geoff Thorpe]
3908
3909 *) Target "mingw" now allows native Windows code to be generated in
3910 the Cygwin environment as well as with the MinGW compiler.
3911 [Ulf Moeller]
3912
3913 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3914
3915 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3916 via timing by performing a MAC computation even if incorrrect
3917 block cipher padding has been found. This is a countermeasure
3918 against active attacks where the attacker has to distinguish
3919 between bad padding and a MAC verification error. (CVE-2003-0078)
3920
3921 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3922 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3923 Martin Vuagnoux (EPFL, Ilion)]
3924
3925 *) Make the no-err option work as intended. The intention with no-err
3926 is not to have the whole error stack handling routines removed from
3927 libcrypto, it's only intended to remove all the function name and
3928 reason texts, thereby removing some of the footprint that may not
3929 be interesting if those errors aren't displayed anyway.
3930
3931 NOTE: it's still possible for any application or module to have it's
3932 own set of error texts inserted. The routines are there, just not
3933 used by default when no-err is given.
3934 [Richard Levitte]
3935
3936 *) Add support for FreeBSD on IA64.
3937 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3938
3939 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3940 Kerberos function mit_des_cbc_cksum(). Before this change,
3941 the value returned by DES_cbc_cksum() was like the one from
3942 mit_des_cbc_cksum(), except the bytes were swapped.
3943 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3944
3945 *) Allow an application to disable the automatic SSL chain building.
3946 Before this a rather primitive chain build was always performed in
3947 ssl3_output_cert_chain(): an application had no way to send the
3948 correct chain if the automatic operation produced an incorrect result.
3949
3950 Now the chain builder is disabled if either:
3951
3952 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3953
3954 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3955
3956 The reasoning behind this is that an application would not want the
3957 auto chain building to take place if extra chain certificates are
3958 present and it might also want a means of sending no additional
3959 certificates (for example the chain has two certificates and the
3960 root is omitted).
3961 [Steve Henson]
3962
3963 *) Add the possibility to build without the ENGINE framework.
3964 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3965
3966 *) Under Win32 gmtime() can return NULL: check return value in
3967 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3968 [Steve Henson]
3969
3970 *) DSA routines: under certain error conditions uninitialized BN objects
3971 could be freed. Solution: make sure initialization is performed early
3972 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3973 Nils Larsch <nla@trustcenter.de> via PR#459)
3974 [Lutz Jaenicke]
3975
3976 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3977 checked on reconnect on the client side, therefore session resumption
3978 could still fail with a "ssl session id is different" error. This
3979 behaviour is masked when SSL_OP_ALL is used due to
3980 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3981 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3982 followup to PR #377.
3983 [Lutz Jaenicke]
3984
3985 *) IA-32 assembler support enhancements: unified ELF targets, support
3986 for SCO/Caldera platforms, fix for Cygwin shared build.
3987 [Andy Polyakov]
3988
3989 *) Add support for FreeBSD on sparc64. As a consequence, support for
3990 FreeBSD on non-x86 processors is separate from x86 processors on
3991 the config script, much like the NetBSD support.
3992 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3993
3994 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3995
3996 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3997 OpenSSL 0.9.7.]
3998
3999 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4000 code (06) was taken as the first octet of the session ID and the last
4001 octet was ignored consequently. As a result SSLv2 client side session
4002 caching could not have worked due to the session ID mismatch between
4003 client and server.
4004 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4005 PR #377.
4006 [Lutz Jaenicke]
4007
4008 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4009 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4010 removed entirely.
4011 [Richard Levitte]
4012
4013 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4014 seems that in spite of existing for more than a year, many application
4015 author have done nothing to provide the necessary callbacks, which
4016 means that this particular engine will not work properly anywhere.
4017 This is a very unfortunate situation which forces us, in the name
4018 of usability, to give the hw_ncipher.c a static lock, which is part
4019 of libcrypto.
4020 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4021 appear in 0.9.8 or later. We EXPECT application authors to have
4022 dealt properly with this when 0.9.8 is released (unless we actually
4023 make such changes in the libcrypto locking code that changes will
4024 have to be made anyway).
4025 [Richard Levitte]
4026
4027 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4028 octets have been read, EOF or an error occurs. Without this change
4029 some truncated ASN1 structures will not produce an error.
4030 [Steve Henson]
4031
4032 *) Disable Heimdal support, since it hasn't been fully implemented.
4033 Still give the possibility to force the use of Heimdal, but with
4034 warnings and a request that patches get sent to openssl-dev.
4035 [Richard Levitte]
4036
4037 *) Add the VC-CE target, introduce the WINCE sysname, and add
4038 INSTALL.WCE and appropriate conditionals to make it build.
4039 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4040
4041 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4042 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4043 edit numbers of the version.
4044 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4045
4046 *) Introduce safe string copy and catenation functions
4047 (BUF_strlcpy() and BUF_strlcat()).
4048 [Ben Laurie (CHATS) and Richard Levitte]
4049
4050 *) Avoid using fixed-size buffers for one-line DNs.
4051 [Ben Laurie (CHATS)]
4052
4053 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4054 resizing buffers containing secrets, and use where appropriate.
4055 [Ben Laurie (CHATS)]
4056
4057 *) Avoid using fixed size buffers for configuration file location.
4058 [Ben Laurie (CHATS)]
4059
4060 *) Avoid filename truncation for various CA files.
4061 [Ben Laurie (CHATS)]
4062
4063 *) Use sizeof in preference to magic numbers.
4064 [Ben Laurie (CHATS)]
4065
4066 *) Avoid filename truncation in cert requests.
4067 [Ben Laurie (CHATS)]
4068
4069 *) Add assertions to check for (supposedly impossible) buffer
4070 overflows.
4071 [Ben Laurie (CHATS)]
4072
4073 *) Don't cache truncated DNS entries in the local cache (this could
4074 potentially lead to a spoofing attack).
4075 [Ben Laurie (CHATS)]
4076
4077 *) Fix various buffers to be large enough for hex/decimal
4078 representations in a platform independent manner.
4079 [Ben Laurie (CHATS)]
4080
4081 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4082 resizing buffers containing secrets, and use where appropriate.
4083 [Ben Laurie (CHATS)]
4084
4085 *) Add BIO_indent() to avoid much slightly worrying code to do
4086 indents.
4087 [Ben Laurie (CHATS)]
4088
4089 *) Convert sprintf()/BIO_puts() to BIO_printf().
4090 [Ben Laurie (CHATS)]
4091
4092 *) buffer_gets() could terminate with the buffer only half
4093 full. Fixed.
4094 [Ben Laurie (CHATS)]
4095
4096 *) Add assertions to prevent user-supplied crypto functions from
4097 overflowing internal buffers by having large block sizes, etc.
4098 [Ben Laurie (CHATS)]
4099
4100 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4101 unconditionally).
4102 [Ben Laurie (CHATS)]
4103
4104 *) Eliminate unused copy of key in RC4.
4105 [Ben Laurie (CHATS)]
4106
4107 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4108 [Ben Laurie (CHATS)]
4109
4110 *) Fix off-by-one error in EGD path.
4111 [Ben Laurie (CHATS)]
4112
4113 *) If RANDFILE path is too long, ignore instead of truncating.
4114 [Ben Laurie (CHATS)]
4115
4116 *) Eliminate unused and incorrectly sized X.509 structure
4117 CBCParameter.
4118 [Ben Laurie (CHATS)]
4119
4120 *) Eliminate unused and dangerous function knumber().
4121 [Ben Laurie (CHATS)]
4122
4123 *) Eliminate unused and dangerous structure, KSSL_ERR.
4124 [Ben Laurie (CHATS)]
4125
4126 *) Protect against overlong session ID context length in an encoded
4127 session object. Since these are local, this does not appear to be
4128 exploitable.
4129 [Ben Laurie (CHATS)]
4130
4131 *) Change from security patch (see 0.9.6e below) that did not affect
4132 the 0.9.6 release series:
4133
4134 Remote buffer overflow in SSL3 protocol - an attacker could
4135 supply an oversized master key in Kerberos-enabled versions.
4136 (CVE-2002-0657)
4137 [Ben Laurie (CHATS)]
4138
4139 *) Change the SSL kerb5 codes to match RFC 2712.
4140 [Richard Levitte]
4141
4142 *) Make -nameopt work fully for req and add -reqopt switch.
4143 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4144
4145 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4146 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4147
4148 *) Make sure tests can be performed even if the corresponding algorithms
4149 have been removed entirely. This was also the last step to make
4150 OpenSSL compilable with DJGPP under all reasonable conditions.
4151 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4152
4153 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4154 to allow version independent disabling of normally unselected ciphers,
4155 which may be activated as a side-effect of selecting a single cipher.
4156
4157 (E.g., cipher list string "RSA" enables ciphersuites that are left
4158 out of "ALL" because they do not provide symmetric encryption.
4159 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4160 [Lutz Jaenicke, Bodo Moeller]
4161
4162 *) Add appropriate support for separate platform-dependent build
4163 directories. The recommended way to make a platform-dependent
4164 build directory is the following (tested on Linux), maybe with
4165 some local tweaks:
4166
4167 # Place yourself outside of the OpenSSL source tree. In
4168 # this example, the environment variable OPENSSL_SOURCE
4169 # is assumed to contain the absolute OpenSSL source directory.
4170 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4171 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4172 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4173 mkdir -p `dirname $F`
4174 ln -s $OPENSSL_SOURCE/$F $F
4175 done
4176
4177 To be absolutely sure not to disturb the source tree, a "make clean"
4178 is a good thing. If it isn't successfull, don't worry about it,
4179 it probably means the source directory is very clean.
4180 [Richard Levitte]
4181
4182 *) Make sure any ENGINE control commands make local copies of string
4183 pointers passed to them whenever necessary. Otherwise it is possible
4184 the caller may have overwritten (or deallocated) the original string
4185 data when a later ENGINE operation tries to use the stored values.
4186 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4187
4188 *) Improve diagnostics in file reading and command-line digests.
4189 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4190
4191 *) Add AES modes CFB and OFB to the object database. Correct an
4192 error in AES-CFB decryption.
4193 [Richard Levitte]
4194
4195 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4196 allows existing EVP_CIPHER_CTX structures to be reused after
4197 calling EVP_*Final(). This behaviour is used by encryption
4198 BIOs and some applications. This has the side effect that
4199 applications must explicitly clean up cipher contexts with
4200 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4201 [Steve Henson]
4202
4203 *) Check the values of dna and dnb in bn_mul_recursive before calling
4204 bn_mul_comba (a non zero value means the a or b arrays do not contain
4205 n2 elements) and fallback to bn_mul_normal if either is not zero.
4206 [Steve Henson]
4207
4208 *) Fix escaping of non-ASCII characters when using the -subj option
4209 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4210 [Lutz Jaenicke]
4211
4212 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4213 form for "surname", serialNumber has no short form.
4214 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4215 therefore remove "mail" short name for "internet 7".
4216 The OID for unique identifiers in X509 certificates is
4217 x500UniqueIdentifier, not uniqueIdentifier.
4218 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4219 [Lutz Jaenicke]
4220
4221 *) Add an "init" command to the ENGINE config module and auto initialize
4222 ENGINEs. Without any "init" command the ENGINE will be initialized
4223 after all ctrl commands have been executed on it. If init=1 the
4224 ENGINE is initailized at that point (ctrls before that point are run
4225 on the uninitialized ENGINE and after on the initialized one). If
4226 init=0 then the ENGINE will not be iniatialized at all.
4227 [Steve Henson]
4228
4229 *) Fix the 'app_verify_callback' interface so that the user-defined
4230 argument is actually passed to the callback: In the
4231 SSL_CTX_set_cert_verify_callback() prototype, the callback
4232 declaration has been changed from
4233 int (*cb)()
4234 into
4235 int (*cb)(X509_STORE_CTX *,void *);
4236 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4237 i=s->ctx->app_verify_callback(&ctx)
4238 has been changed into
4239 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4240
4241 To update applications using SSL_CTX_set_cert_verify_callback(),
4242 a dummy argument can be added to their callback functions.
4243 [D. K. Smetters <smetters@parc.xerox.com>]
4244
4245 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4246 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4247
4248 *) Add and OPENSSL_LOAD_CONF define which will cause
4249 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4250 This allows older applications to transparently support certain
4251 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4252 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4253 load the config file and OPENSSL_add_all_algorithms_conf() which will
4254 always load it have also been added.
4255 [Steve Henson]
4256
4257 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4258 Adjust NIDs and EVP layer.
4259 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4260
4261 *) Config modules support in openssl utility.
4262
4263 Most commands now load modules from the config file,
4264 though in a few (such as version) this isn't done
4265 because it couldn't be used for anything.
4266
4267 In the case of ca and req the config file used is
4268 the same as the utility itself: that is the -config
4269 command line option can be used to specify an
4270 alternative file.
4271 [Steve Henson]
4272
4273 *) Move default behaviour from OPENSSL_config(). If appname is NULL
4274 use "openssl_conf" if filename is NULL use default openssl config file.
4275 [Steve Henson]
4276
4277 *) Add an argument to OPENSSL_config() to allow the use of an alternative
4278 config section name. Add a new flag to tolerate a missing config file
4279 and move code to CONF_modules_load_file().
4280 [Steve Henson]
4281
4282 *) Support for crypto accelerator cards from Accelerated Encryption
4283 Processing, www.aep.ie. (Use engine 'aep')
4284 The support was copied from 0.9.6c [engine] and adapted/corrected
4285 to work with the new engine framework.
4286 [AEP Inc. and Richard Levitte]
4287
4288 *) Support for SureWare crypto accelerator cards from Baltimore
4289 Technologies. (Use engine 'sureware')
4290 The support was copied from 0.9.6c [engine] and adapted
4291 to work with the new engine framework.
4292 [Richard Levitte]
4293
4294 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
4295 make the newer ENGINE framework commands for the CHIL engine work.
4296 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
4297
4298 *) Make it possible to produce shared libraries on ReliantUNIX.
4299 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
4300
4301 *) Add the configuration target debug-linux-ppro.
4302 Make 'openssl rsa' use the general key loading routines
4303 implemented in apps.c, and make those routines able to
4304 handle the key format FORMAT_NETSCAPE and the variant
4305 FORMAT_IISSGC.
4306 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4307
4308 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4309 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4310
4311 *) Add -keyform to rsautl, and document -engine.
4312 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
4313
4314 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
4315 BIO_R_NO_SUCH_FILE error code rather than the generic
4316 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
4317 [Ben Laurie]
4318
4319 *) Add new functions
4320 ERR_peek_last_error
4321 ERR_peek_last_error_line
4322 ERR_peek_last_error_line_data.
4323 These are similar to
4324 ERR_peek_error
4325 ERR_peek_error_line
4326 ERR_peek_error_line_data,
4327 but report on the latest error recorded rather than the first one
4328 still in the error queue.
4329 [Ben Laurie, Bodo Moeller]
4330
4331 *) default_algorithms option in ENGINE config module. This allows things
4332 like:
4333 default_algorithms = ALL
4334 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
4335 [Steve Henson]
4336
4337 *) Prelminary ENGINE config module.
4338 [Steve Henson]
4339
4340 *) New experimental application configuration code.
4341 [Steve Henson]
4342
4343 *) Change the AES code to follow the same name structure as all other
4344 symmetric ciphers, and behave the same way. Move everything to
4345 the directory crypto/aes, thereby obsoleting crypto/rijndael.
4346 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
4347
4348 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
4349 [Ben Laurie and Theo de Raadt]
4350
4351 *) Add option to output public keys in req command.
4352 [Massimiliano Pala madwolf@openca.org]
4353
4354 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
4355 (up to about 10% better than before for P-192 and P-224).
4356 [Bodo Moeller]
4357
4358 *) New functions/macros
4359
4360 SSL_CTX_set_msg_callback(ctx, cb)
4361 SSL_CTX_set_msg_callback_arg(ctx, arg)
4362 SSL_set_msg_callback(ssl, cb)
4363 SSL_set_msg_callback_arg(ssl, arg)
4364
4365 to request calling a callback function
4366
4367 void cb(int write_p, int version, int content_type,
4368 const void *buf, size_t len, SSL *ssl, void *arg)
4369
4370 whenever a protocol message has been completely received
4371 (write_p == 0) or sent (write_p == 1). Here 'version' is the
4372 protocol version according to which the SSL library interprets
4373 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
4374 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
4375 the content type as defined in the SSL 3.0/TLS 1.0 protocol
4376 specification (change_cipher_spec(20), alert(21), handshake(22)).
4377 'buf' and 'len' point to the actual message, 'ssl' to the
4378 SSL object, and 'arg' is the application-defined value set by
4379 SSL[_CTX]_set_msg_callback_arg().
4380
4381 'openssl s_client' and 'openssl s_server' have new '-msg' options
4382 to enable a callback that displays all protocol messages.
4383 [Bodo Moeller]
4384
4385 *) Change the shared library support so shared libraries are built as
4386 soon as the corresponding static library is finished, and thereby get
4387 openssl and the test programs linked against the shared library.
4388 This still only happens when the keyword "shard" has been given to
4389 the configuration scripts.
4390
4391 NOTE: shared library support is still an experimental thing, and
4392 backward binary compatibility is still not guaranteed.
4393 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
4394
4395 *) Add support for Subject Information Access extension.
4396 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4397
4398 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
4399 additional bytes when new memory had to be allocated, not just
4400 when reusing an existing buffer.
4401 [Bodo Moeller]
4402
4403 *) New command line and configuration option 'utf8' for the req command.
4404 This allows field values to be specified as UTF8 strings.
4405 [Steve Henson]
4406
4407 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
4408 runs for the former and machine-readable output for the latter.
4409 [Ben Laurie]
4410
4411 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
4412 of the e-mail address in the DN (i.e., it will go into a certificate
4413 extension only). The new configuration file option 'email_in_dn = no'
4414 has the same effect.
4415 [Massimiliano Pala madwolf@openca.org]
4416
4417 *) Change all functions with names starting with des_ to be starting
4418 with DES_ instead. Add wrappers that are compatible with libdes,
4419 but are named _ossl_old_des_*. Finally, add macros that map the
4420 des_* symbols to the corresponding _ossl_old_des_* if libdes
4421 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4422 desired, the des_* symbols will be mapped to DES_*, with one
4423 exception.
4424
4425 Since we provide two compatibility mappings, the user needs to
4426 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4427 compatibility is desired. The default (i.e., when that macro
4428 isn't defined) is OpenSSL 0.9.6c compatibility.
4429
4430 There are also macros that enable and disable the support of old
4431 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4432 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4433 are defined, the default will apply: to support the old des routines.
4434
4435 In either case, one must include openssl/des.h to get the correct
4436 definitions. Do not try to just include openssl/des_old.h, that
4437 won't work.
4438
4439 NOTE: This is a major break of an old API into a new one. Software
4440 authors are encouraged to switch to the DES_ style functions. Some
4441 time in the future, des_old.h and the libdes compatibility functions
4442 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4443 default), and then completely removed.
4444 [Richard Levitte]
4445
4446 *) Test for certificates which contain unsupported critical extensions.
4447 If such a certificate is found during a verify operation it is
4448 rejected by default: this behaviour can be overridden by either
4449 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4450 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4451 X509_supported_extension() has also been added which returns 1 if a
4452 particular extension is supported.
4453 [Steve Henson]
4454
4455 *) Modify the behaviour of EVP cipher functions in similar way to digests
4456 to retain compatibility with existing code.
4457 [Steve Henson]
4458
4459 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4460 compatibility with existing code. In particular the 'ctx' parameter does
4461 not have to be to be initialized before the call to EVP_DigestInit() and
4462 it is tidied up after a call to EVP_DigestFinal(). New function
4463 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4464 EVP_MD_CTX_copy() changed to not require the destination to be
4465 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4466 requires the destination to be valid.
4467
4468 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4469 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4470 [Steve Henson]
4471
4472 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4473 so that complete 'Handshake' protocol structures are kept in memory
4474 instead of overwriting 'msg_type' and 'length' with 'body' data.
4475 [Bodo Moeller]
4476
4477 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4478 [Massimo Santin via Richard Levitte]
4479
4480 *) Major restructuring to the underlying ENGINE code. This includes
4481 reduction of linker bloat, separation of pure "ENGINE" manipulation
4482 (initialisation, etc) from functionality dealing with implementations
4483 of specific crypto iterfaces. This change also introduces integrated
4484 support for symmetric ciphers and digest implementations - so ENGINEs
4485 can now accelerate these by providing EVP_CIPHER and EVP_MD
4486 implementations of their own. This is detailed in crypto/engine/README
4487 as it couldn't be adequately described here. However, there are a few
4488 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4489 were changed in the original introduction of ENGINE code have now
4490 reverted back - the hooking from this code to ENGINE is now a good
4491 deal more passive and at run-time, operations deal directly with
4492 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4493 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4494 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4495 they were not being used by the framework as there is no concept of a
4496 BIGNUM_METHOD and they could not be generalised to the new
4497 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4498 ENGINE_cpy() has been removed as it cannot be consistently defined in
4499 the new code.
4500 [Geoff Thorpe]
4501
4502 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4503 [Steve Henson]
4504
4505 *) Change mkdef.pl to sort symbols that get the same entry number,
4506 and make sure the automatically generated functions ERR_load_*
4507 become part of libeay.num as well.
4508 [Richard Levitte]
4509
4510 *) New function SSL_renegotiate_pending(). This returns true once
4511 renegotiation has been requested (either SSL_renegotiate() call
4512 or HelloRequest/ClientHello receveived from the peer) and becomes
4513 false once a handshake has been completed.
4514 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4515 sends a HelloRequest, but does not ensure that a handshake takes
4516 place. SSL_renegotiate_pending() is useful for checking if the
4517 client has followed the request.)
4518 [Bodo Moeller]
4519
4520 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4521 By default, clients may request session resumption even during
4522 renegotiation (if session ID contexts permit); with this option,
4523 session resumption is possible only in the first handshake.
4524
4525 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4526 more bits available for options that should not be part of
4527 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4528 [Bodo Moeller]
4529
4530 *) Add some demos for certificate and certificate request creation.
4531 [Steve Henson]
4532
4533 *) Make maximum certificate chain size accepted from the peer application
4534 settable (SSL*_get/set_max_cert_list()), as proposed by
4535 "Douglas E. Engert" <deengert@anl.gov>.
4536 [Lutz Jaenicke]
4537
4538 *) Add support for shared libraries for Unixware-7
4539 (Boyd Lynn Gerber <gerberb@zenez.com>).
4540 [Lutz Jaenicke]
4541
4542 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4543 be done prior to destruction. Use this to unload error strings from
4544 ENGINEs that load their own error strings. NB: This adds two new API
4545 functions to "get" and "set" this destroy handler in an ENGINE.
4546 [Geoff Thorpe]
4547
4548 *) Alter all existing ENGINE implementations (except "openssl" and
4549 "openbsd") to dynamically instantiate their own error strings. This
4550 makes them more flexible to be built both as statically-linked ENGINEs
4551 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4552 Also, add stub code to each that makes building them as self-contained
4553 shared-libraries easier (see README.ENGINE).
4554 [Geoff Thorpe]
4555
4556 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4557 implementations into applications that are completely implemented in
4558 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4559 commands that can be used to configure what shared-library to load and
4560 to control aspects of the way it is handled. Also, made an update to
4561 the README.ENGINE file that brings its information up-to-date and
4562 provides some information and instructions on the "dynamic" ENGINE
4563 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4564 [Geoff Thorpe]
4565
4566 *) Make it possible to unload ranges of ERR strings with a new
4567 "ERR_unload_strings" function.
4568 [Geoff Thorpe]
4569
4570 *) Add a copy() function to EVP_MD.
4571 [Ben Laurie]
4572
4573 *) Make EVP_MD routines take a context pointer instead of just the
4574 md_data void pointer.
4575 [Ben Laurie]
4576
4577 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4578 that the digest can only process a single chunk of data
4579 (typically because it is provided by a piece of
4580 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4581 is only going to provide a single chunk of data, and hence the
4582 framework needn't accumulate the data for oneshot drivers.
4583 [Ben Laurie]
4584
4585 *) As with "ERR", make it possible to replace the underlying "ex_data"
4586 functions. This change also alters the storage and management of global
4587 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4588 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4589 index counters. The API functions that use this state have been changed
4590 to take a "class_index" rather than pointers to the class's local STACK
4591 and counter, and there is now an API function to dynamically create new
4592 classes. This centralisation allows us to (a) plug a lot of the
4593 thread-safety problems that existed, and (b) makes it possible to clean
4594 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4595 such data would previously have always leaked in application code and
4596 workarounds were in place to make the memory debugging turn a blind eye
4597 to it. Application code that doesn't use this new function will still
4598 leak as before, but their memory debugging output will announce it now
4599 rather than letting it slide.
4600
4601 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4602 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4603 has a return value to indicate success or failure.
4604 [Geoff Thorpe]
4605
4606 *) Make it possible to replace the underlying "ERR" functions such that the
4607 global state (2 LHASH tables and 2 locks) is only used by the "default"
4608 implementation. This change also adds two functions to "get" and "set"
4609 the implementation prior to it being automatically set the first time
4610 any other ERR function takes place. Ie. an application can call "get",
4611 pass the return value to a module it has just loaded, and that module
4612 can call its own "set" function using that value. This means the
4613 module's "ERR" operations will use (and modify) the error state in the
4614 application and not in its own statically linked copy of OpenSSL code.
4615 [Geoff Thorpe]
4616
4617 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4618 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4619 the operation, and provides a more encapsulated way for external code
4620 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4621 to use these functions rather than manually incrementing the counts.
4622
4623 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4624 [Geoff Thorpe]
4625
4626 *) Add EVP test program.
4627 [Ben Laurie]
4628
4629 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4630 [Ben Laurie]
4631
4632 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4633 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4634 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4635 These allow a CRL to be built without having to access X509_CRL fields
4636 directly. Modify 'ca' application to use new functions.
4637 [Steve Henson]
4638
4639 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4640 bug workarounds. Rollback attack detection is a security feature.
4641 The problem will only arise on OpenSSL servers when TLSv1 is not
4642 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4643 Software authors not wanting to support TLSv1 will have special reasons
4644 for their choice and can explicitly enable this option.
4645 [Bodo Moeller, Lutz Jaenicke]
4646
4647 *) Rationalise EVP so it can be extended: don't include a union of
4648 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4649 (similar to those existing for EVP_CIPHER_CTX).
4650 Usage example:
4651
4652 EVP_MD_CTX md;
4653
4654 EVP_MD_CTX_init(&md); /* new function call */
4655 EVP_DigestInit(&md, EVP_sha1());
4656 EVP_DigestUpdate(&md, in, len);
4657 EVP_DigestFinal(&md, out, NULL);
4658 EVP_MD_CTX_cleanup(&md); /* new function call */
4659
4660 [Ben Laurie]
4661
4662 *) Make DES key schedule conform to the usual scheme, as well as
4663 correcting its structure. This means that calls to DES functions
4664 now have to pass a pointer to a des_key_schedule instead of a
4665 plain des_key_schedule (which was actually always a pointer
4666 anyway): E.g.,
4667
4668 des_key_schedule ks;
4669
4670 des_set_key_checked(..., &ks);
4671 des_ncbc_encrypt(..., &ks, ...);
4672
4673 (Note that a later change renames 'des_...' into 'DES_...'.)
4674 [Ben Laurie]
4675
4676 *) Initial reduction of linker bloat: the use of some functions, such as
4677 PEM causes large amounts of unused functions to be linked in due to
4678 poor organisation. For example pem_all.c contains every PEM function
4679 which has a knock on effect of linking in large amounts of (unused)
4680 ASN1 code. Grouping together similar functions and splitting unrelated
4681 functions prevents this.
4682 [Steve Henson]
4683
4684 *) Cleanup of EVP macros.
4685 [Ben Laurie]
4686
4687 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4688 correct _ecb suffix.
4689 [Ben Laurie]
4690
4691 *) Add initial OCSP responder support to ocsp application. The
4692 revocation information is handled using the text based index
4693 use by the ca application. The responder can either handle
4694 requests generated internally, supplied in files (for example
4695 via a CGI script) or using an internal minimal server.
4696 [Steve Henson]
4697
4698 *) Add configuration choices to get zlib compression for TLS.
4699 [Richard Levitte]
4700
4701 *) Changes to Kerberos SSL for RFC 2712 compliance:
4702 1. Implemented real KerberosWrapper, instead of just using
4703 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4704 2. Implemented optional authenticator field of KerberosWrapper.
4705
4706 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4707 and authenticator structs; see crypto/krb5/.
4708
4709 Generalized Kerberos calls to support multiple Kerberos libraries.
4710 [Vern Staats <staatsvr@asc.hpc.mil>,
4711 Jeffrey Altman <jaltman@columbia.edu>
4712 via Richard Levitte]
4713
4714 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4715 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4716 values for each of the key sizes rather than having just
4717 parameters (and 'speed' generating keys each time).
4718 [Geoff Thorpe]
4719
4720 *) Speed up EVP routines.
4721 Before:
4722 encrypt
4723 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4724 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4725 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4726 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4727 decrypt
4728 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4729 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4730 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4731 After:
4732 encrypt
4733 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4734 decrypt
4735 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4736 [Ben Laurie]
4737
4738 *) Added the OS2-EMX target.
4739 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4740
4741 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4742 to support NCONF routines in extension code. New function CONF_set_nconf()
4743 to allow functions which take an NCONF to also handle the old LHASH
4744 structure: this means that the old CONF compatible routines can be
4745 retained (in particular wrt extensions) without having to duplicate the
4746 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4747 [Steve Henson]
4748
4749 *) Enhance the general user interface with mechanisms for inner control
4750 and with possibilities to have yes/no kind of prompts.
4751 [Richard Levitte]
4752
4753 *) Change all calls to low level digest routines in the library and
4754 applications to use EVP. Add missing calls to HMAC_cleanup() and
4755 don't assume HMAC_CTX can be copied using memcpy().
4756 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4757
4758 *) Add the possibility to control engines through control names but with
4759 arbitrary arguments instead of just a string.
4760 Change the key loaders to take a UI_METHOD instead of a callback
4761 function pointer. NOTE: this breaks binary compatibility with earlier
4762 versions of OpenSSL [engine].
4763 Adapt the nCipher code for these new conditions and add a card insertion
4764 callback.
4765 [Richard Levitte]
4766
4767 *) Enhance the general user interface with mechanisms to better support
4768 dialog box interfaces, application-defined prompts, the possibility
4769 to use defaults (for example default passwords from somewhere else)
4770 and interrupts/cancellations.
4771 [Richard Levitte]
4772
4773 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4774 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4775 [Steve Henson]
4776
4777 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4778 tidy up some unnecessarily weird code in 'sk_new()').
4779 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4780
4781 *) Change the key loading routines for ENGINEs to use the same kind
4782 callback (pem_password_cb) as all other routines that need this
4783 kind of callback.
4784 [Richard Levitte]
4785
4786 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4787 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4788 than this minimum value is recommended.
4789 [Lutz Jaenicke]
4790
4791 *) New random seeder for OpenVMS, using the system process statistics
4792 that are easily reachable.
4793 [Richard Levitte]
4794
4795 *) Windows apparently can't transparently handle global
4796 variables defined in DLLs. Initialisations such as:
4797
4798 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4799
4800 wont compile. This is used by the any applications that need to
4801 declare their own ASN1 modules. This was fixed by adding the option
4802 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4803 needed for static libraries under Win32.
4804 [Steve Henson]
4805
4806 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4807 setting of purpose and trust fields. New X509_STORE trust and
4808 purpose functions and tidy up setting in other SSL functions.
4809 [Steve Henson]
4810
4811 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4812 structure. These are inherited by X509_STORE_CTX when it is
4813 initialised. This allows various defaults to be set in the
4814 X509_STORE structure (such as flags for CRL checking and custom
4815 purpose or trust settings) for functions which only use X509_STORE_CTX
4816 internally such as S/MIME.
4817
4818 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4819 trust settings if they are not set in X509_STORE. This allows X509_STORE
4820 purposes and trust (in S/MIME for example) to override any set by default.
4821
4822 Add command line options for CRL checking to smime, s_client and s_server
4823 applications.
4824 [Steve Henson]
4825
4826 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4827 are set then the CRL is looked up in the X509_STORE structure and
4828 its validity and signature checked, then if the certificate is found
4829 in the CRL the verify fails with a revoked error.
4830
4831 Various new CRL related callbacks added to X509_STORE_CTX structure.
4832
4833 Command line options added to 'verify' application to support this.
4834
4835 This needs some additional work, such as being able to handle multiple
4836 CRLs with different times, extension based lookup (rather than just
4837 by subject name) and ultimately more complete V2 CRL extension
4838 handling.
4839 [Steve Henson]
4840
4841 *) Add a general user interface API (crypto/ui/). This is designed
4842 to replace things like des_read_password and friends (backward
4843 compatibility functions using this new API are provided).
4844 The purpose is to remove prompting functions from the DES code
4845 section as well as provide for prompting through dialog boxes in
4846 a window system and the like.
4847 [Richard Levitte]
4848
4849 *) Add "ex_data" support to ENGINE so implementations can add state at a
4850 per-structure level rather than having to store it globally.
4851 [Geoff]
4852
4853 *) Make it possible for ENGINE structures to be copied when retrieved by
4854 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4855 This causes the "original" ENGINE structure to act like a template,
4856 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4857 operational state can be localised to each ENGINE structure, despite the
4858 fact they all share the same "methods". New ENGINE structures returned in
4859 this case have no functional references and the return value is the single
4860 structural reference. This matches the single structural reference returned
4861 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4862 ENGINE structure.
4863 [Geoff]
4864
4865 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4866 needs to match any other type at all we need to manually clear the
4867 tag cache.
4868 [Steve Henson]
4869
4870 *) Changes to the "openssl engine" utility to include;
4871 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4872 about an ENGINE's available control commands.
4873 - executing control commands from command line arguments using the
4874 '-pre' and '-post' switches. '-post' is only used if '-t' is
4875 specified and the ENGINE is successfully initialised. The syntax for
4876 the individual commands are colon-separated, for example;
4877 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4878 [Geoff]
4879
4880 *) New dynamic control command support for ENGINEs. ENGINEs can now
4881 declare their own commands (numbers), names (strings), descriptions,
4882 and input types for run-time discovery by calling applications. A
4883 subset of these commands are implicitly classed as "executable"
4884 depending on their input type, and only these can be invoked through
4885 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4886 can be based on user input, config files, etc). The distinction is
4887 that "executable" commands cannot return anything other than a boolean
4888 result and can only support numeric or string input, whereas some
4889 discoverable commands may only be for direct use through
4890 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4891 pointers, or other custom uses. The "executable" commands are to
4892 support parameterisations of ENGINE behaviour that can be
4893 unambiguously defined by ENGINEs and used consistently across any
4894 OpenSSL-based application. Commands have been added to all the
4895 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4896 control over shared-library paths without source code alterations.
4897 [Geoff]
4898
4899 *) Changed all ENGINE implementations to dynamically allocate their
4900 ENGINEs rather than declaring them statically. Apart from this being
4901 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4902 this also allows the implementations to compile without using the
4903 internal engine_int.h header.
4904 [Geoff]
4905
4906 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4907 'const' value. Any code that should be able to modify a RAND_METHOD
4908 should already have non-const pointers to it (ie. they should only
4909 modify their own ones).
4910 [Geoff]
4911
4912 *) Made a variety of little tweaks to the ENGINE code.
4913 - "atalla" and "ubsec" string definitions were moved from header files
4914 to C code. "nuron" string definitions were placed in variables
4915 rather than hard-coded - allowing parameterisation of these values
4916 later on via ctrl() commands.
4917 - Removed unused "#if 0"'d code.
4918 - Fixed engine list iteration code so it uses ENGINE_free() to release
4919 structural references.
4920 - Constified the RAND_METHOD element of ENGINE structures.
4921 - Constified various get/set functions as appropriate and added
4922 missing functions (including a catch-all ENGINE_cpy that duplicates
4923 all ENGINE values onto a new ENGINE except reference counts/state).
4924 - Removed NULL parameter checks in get/set functions. Setting a method
4925 or function to NULL is a way of cancelling out a previously set
4926 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4927 and doesn't justify the extra error symbols and code.
4928 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4929 flags from engine_int.h to engine.h.
4930 - Changed prototypes for ENGINE handler functions (init(), finish(),
4931 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4932 [Geoff]
4933
4934 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4935 to the algorithm using long division. The binary algorithm can be
4936 used only if the modulus is odd. On 32-bit systems, it is faster
4937 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4938 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4939 up to 450 bits. In 64-bit environments, the binary algorithm
4940 appears to be advantageous for much longer moduli; here we use it
4941 for moduli up to 2048 bits.
4942 [Bodo Moeller]
4943
4944 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4945 could not support the combine flag in choice fields.
4946 [Steve Henson]
4947
4948 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4949 extensions from a certificate request to the certificate.
4950 [Steve Henson]
4951
4952 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4953 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4954 file: this allows the display of the certificate about to be
4955 signed to be customised, to allow certain fields to be included
4956 or excluded and extension details. The old system didn't display
4957 multicharacter strings properly, omitted fields not in the policy
4958 and couldn't display additional details such as extensions.
4959 [Steve Henson]
4960
4961 *) Function EC_POINTs_mul for multiple scalar multiplication
4962 of an arbitrary number of elliptic curve points
4963 \sum scalars[i]*points[i],
4964 optionally including the generator defined for the EC_GROUP:
4965 scalar*generator + \sum scalars[i]*points[i].
4966
4967 EC_POINT_mul is a simple wrapper function for the typical case
4968 that the point list has just one item (besides the optional
4969 generator).
4970 [Bodo Moeller]
4971
4972 *) First EC_METHODs for curves over GF(p):
4973
4974 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4975 operations and provides various method functions that can also
4976 operate with faster implementations of modular arithmetic.
4977
4978 EC_GFp_mont_method() reuses most functions that are part of
4979 EC_GFp_simple_method, but uses Montgomery arithmetic.
4980
4981 [Bodo Moeller; point addition and point doubling
4982 implementation directly derived from source code provided by
4983 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4984
4985 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4986 crypto/ec/ec_lib.c):
4987
4988 Curves are EC_GROUP objects (with an optional group generator)
4989 based on EC_METHODs that are built into the library.
4990
4991 Points are EC_POINT objects based on EC_GROUP objects.
4992
4993 Most of the framework would be able to handle curves over arbitrary
4994 finite fields, but as there are no obvious types for fields other
4995 than GF(p), some functions are limited to that for now.
4996 [Bodo Moeller]
4997
4998 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4999 that the file contains a complete HTTP response.
5000 [Richard Levitte]
5001
5002 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5003 change the def and num file printf format specifier from "%-40sXXX"
5004 to "%-39s XXX". The latter will always guarantee a space after the
5005 field while the former will cause them to run together if the field
5006 is 40 of more characters long.
5007 [Steve Henson]
5008
5009 *) Constify the cipher and digest 'method' functions and structures
5010 and modify related functions to take constant EVP_MD and EVP_CIPHER
5011 pointers.
5012 [Steve Henson]
5013
5014 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5015 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5016 [Bodo Moeller]
5017
5018 *) Modify EVP_Digest*() routines so they now return values. Although the
5019 internal software routines can never fail additional hardware versions
5020 might.
5021 [Steve Henson]
5022
5023 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5024
5025 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5026 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5027
5028 ASN1 error codes
5029 ERR_R_NESTED_ASN1_ERROR
5030 ...
5031 ERR_R_MISSING_ASN1_EOS
5032 were 4 .. 9, conflicting with
5033 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5034 ...
5035 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5036 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5037
5038 Add new error code 'ERR_R_INTERNAL_ERROR'.
5039 [Bodo Moeller]
5040
5041 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5042 suffices.
5043 [Bodo Moeller]
5044
5045 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5046 sets the subject name for a new request or supersedes the
5047 subject name in a given request. Formats that can be parsed are
5048 'CN=Some Name, OU=myOU, C=IT'
5049 and
5050 'CN=Some Name/OU=myOU/C=IT'.
5051
5052 Add options '-batch' and '-verbose' to 'openssl req'.
5053 [Massimiliano Pala <madwolf@hackmasters.net>]
5054
5055 *) Introduce the possibility to access global variables through
5056 functions on platform were that's the best way to handle exporting
5057 global variables in shared libraries. To enable this functionality,
5058 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5059 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5060 is normally done by Configure or something similar).
5061
5062 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5063 in the source file (foo.c) like this:
5064
5065 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5066 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5067
5068 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5069 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5070
5071 OPENSSL_DECLARE_GLOBAL(int,foo);
5072 #define foo OPENSSL_GLOBAL_REF(foo)
5073 OPENSSL_DECLARE_GLOBAL(double,bar);
5074 #define bar OPENSSL_GLOBAL_REF(bar)
5075
5076 The #defines are very important, and therefore so is including the
5077 header file everywhere where the defined globals are used.
5078
5079 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5080 of ASN.1 items, but that structure is a bit different.
5081
5082 The largest change is in util/mkdef.pl which has been enhanced with
5083 better and easier to understand logic to choose which symbols should
5084 go into the Windows .def files as well as a number of fixes and code
5085 cleanup (among others, algorithm keywords are now sorted
5086 lexicographically to avoid constant rewrites).
5087 [Richard Levitte]
5088
5089 *) In BN_div() keep a copy of the sign of 'num' before writing the
5090 result to 'rm' because if rm==num the value will be overwritten
5091 and produce the wrong result if 'num' is negative: this caused
5092 problems with BN_mod() and BN_nnmod().
5093 [Steve Henson]
5094
5095 *) Function OCSP_request_verify(). This checks the signature on an
5096 OCSP request and verifies the signer certificate. The signer
5097 certificate is just checked for a generic purpose and OCSP request
5098 trust settings.
5099 [Steve Henson]
5100
5101 *) Add OCSP_check_validity() function to check the validity of OCSP
5102 responses. OCSP responses are prepared in real time and may only
5103 be a few seconds old. Simply checking that the current time lies
5104 between thisUpdate and nextUpdate max reject otherwise valid responses
5105 caused by either OCSP responder or client clock inaccuracy. Instead
5106 we allow thisUpdate and nextUpdate to fall within a certain period of
5107 the current time. The age of the response can also optionally be
5108 checked. Two new options -validity_period and -status_age added to
5109 ocsp utility.
5110 [Steve Henson]
5111
5112 *) If signature or public key algorithm is unrecognized print out its
5113 OID rather that just UNKNOWN.
5114 [Steve Henson]
5115
5116 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5117 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5118 ID to be generated from the issuer certificate alone which can then be
5119 passed to OCSP_id_issuer_cmp().
5120 [Steve Henson]
5121
5122 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5123 ASN1 modules to export functions returning ASN1_ITEM pointers
5124 instead of the ASN1_ITEM structures themselves. This adds several
5125 new macros which allow the underlying ASN1 function/structure to
5126 be accessed transparently. As a result code should not use ASN1_ITEM
5127 references directly (such as &X509_it) but instead use the relevant
5128 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5129 use of the new ASN1 code on platforms where exporting structures
5130 is problematical (for example in shared libraries) but exporting
5131 functions returning pointers to structures is not.
5132 [Steve Henson]
5133
5134 *) Add support for overriding the generation of SSL/TLS session IDs.
5135 These callbacks can be registered either in an SSL_CTX or per SSL.
5136 The purpose of this is to allow applications to control, if they wish,
5137 the arbitrary values chosen for use as session IDs, particularly as it
5138 can be useful for session caching in multiple-server environments. A
5139 command-line switch for testing this (and any client code that wishes
5140 to use such a feature) has been added to "s_server".
5141 [Geoff Thorpe, Lutz Jaenicke]
5142
5143 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5144 of the form '#if defined(...) || defined(...) || ...' and
5145 '#if !defined(...) && !defined(...) && ...'. This also avoids
5146 the growing number of special cases it was previously handling.
5147 [Richard Levitte]
5148
5149 *) Make all configuration macros available for application by making
5150 sure they are available in opensslconf.h, by giving them names starting
5151 with "OPENSSL_" to avoid conflicts with other packages and by making
5152 sure e_os2.h will cover all platform-specific cases together with
5153 opensslconf.h.
5154 Additionally, it is now possible to define configuration/platform-
5155 specific names (called "system identities"). In the C code, these
5156 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5157 macro with the name beginning with "OPENSSL_SYS_", which is determined
5158 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5159 what is available.
5160 [Richard Levitte]
5161
5162 *) New option -set_serial to 'req' and 'x509' this allows the serial
5163 number to use to be specified on the command line. Previously self
5164 signed certificates were hard coded with serial number 0 and the
5165 CA options of 'x509' had to use a serial number in a file which was
5166 auto incremented.
5167 [Steve Henson]
5168
5169 *) New options to 'ca' utility to support V2 CRL entry extensions.
5170 Currently CRL reason, invalidity date and hold instruction are
5171 supported. Add new CRL extensions to V3 code and some new objects.
5172 [Steve Henson]
5173
5174 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5175 disable standard block padding (aka PKCS#5 padding) in the EVP
5176 API, which was previously mandatory. This means that the data is
5177 not padded in any way and so the total length much be a multiple
5178 of the block size, otherwise an error occurs.
5179 [Steve Henson]
5180
5181 *) Initial (incomplete) OCSP SSL support.
5182 [Steve Henson]
5183
5184 *) New function OCSP_parse_url(). This splits up a URL into its host,
5185 port and path components: primarily to parse OCSP URLs. New -url
5186 option to ocsp utility.
5187 [Steve Henson]
5188
5189 *) New nonce behavior. The return value of OCSP_check_nonce() now
5190 reflects the various checks performed. Applications can decide
5191 whether to tolerate certain situations such as an absent nonce
5192 in a response when one was present in a request: the ocsp application
5193 just prints out a warning. New function OCSP_add1_basic_nonce()
5194 this is to allow responders to include a nonce in a response even if
5195 the request is nonce-less.
5196 [Steve Henson]
5197
5198 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5199 skipped when using openssl x509 multiple times on a single input file,
5200 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5201 [Bodo Moeller]
5202
5203 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5204 set string type: to handle setting ASN1_TIME structures. Fix ca
5205 utility to correctly initialize revocation date of CRLs.
5206 [Steve Henson]
5207
5208 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5209 the clients preferred ciphersuites and rather use its own preferences.
5210 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5211 Internet Explorer by ensuring unchanged hash method during stepup.
5212 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5213 [Lutz Jaenicke]
5214
5215 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5216 to aes and add a new 'exist' option to print out symbols that don't
5217 appear to exist.
5218 [Steve Henson]
5219
5220 *) Additional options to ocsp utility to allow flags to be set and
5221 additional certificates supplied.
5222 [Steve Henson]
5223
5224 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5225 OCSP client a number of certificate to only verify the response
5226 signature against.
5227 [Richard Levitte]
5228
5229 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5230 handle the new API. Currently only ECB, CBC modes supported. Add new
5231 AES OIDs.
5232
5233 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5234 Encryption Standard (AES) Ciphersuites for Transport Layer
5235 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5236 not enabled by default and were not part of the "ALL" ciphersuite
5237 alias because they were not yet official; they could be
5238 explicitly requested by specifying the "AESdraft" ciphersuite
5239 group alias. In the final release of OpenSSL 0.9.7, the group
5240 alias is called "AES" and is part of "ALL".)
5241 [Ben Laurie, Steve Henson, Bodo Moeller]
5242
5243 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5244 request to response.
5245 [Steve Henson]
5246
5247 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5248 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5249 extract information from a certificate request. OCSP_response_create()
5250 creates a response and optionally adds a basic response structure.
5251 OCSP_basic_add1_status() adds a complete single response to a basic
5252 response and returns the OCSP_SINGLERESP structure just added (to allow
5253 extensions to be included for example). OCSP_basic_add1_cert() adds a
5254 certificate to a basic response and OCSP_basic_sign() signs a basic
5255 response with various flags. New helper functions ASN1_TIME_check()
5256 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5257 (converts ASN1_TIME to GeneralizedTime).
5258 [Steve Henson]
5259
5260 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
5261 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
5262 structure from a certificate. X509_pubkey_digest() digests the public_key
5263 contents: this is used in various key identifiers.
5264 [Steve Henson]
5265
5266 *) Make sk_sort() tolerate a NULL argument.
5267 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
5268
5269 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
5270 passed by the function are trusted implicitly. If any of them signed the
5271 response then it is assumed to be valid and is not verified.
5272 [Steve Henson]
5273
5274 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
5275 to data. This was previously part of the PKCS7 ASN1 code. This
5276 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
5277 [Steve Henson, reported by Kenneth R. Robinette
5278 <support@securenetterm.com>]
5279
5280 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
5281 routines: without these tracing memory leaks is very painful.
5282 Fix leaks in PKCS12 and PKCS7 routines.
5283 [Steve Henson]
5284
5285 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
5286 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
5287 effectively meant GeneralizedTime would never be used. Now it
5288 is initialised to -1 but X509_time_adj() now has to check the value
5289 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
5290 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
5291 [Steve Henson, reported by Kenneth R. Robinette
5292 <support@securenetterm.com>]
5293
5294 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
5295 result in a zero length in the ASN1_INTEGER structure which was
5296 not consistent with the structure when d2i_ASN1_INTEGER() was used
5297 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
5298 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
5299 where it did not print out a minus for negative ASN1_INTEGER.
5300 [Steve Henson]
5301
5302 *) Add summary printout to ocsp utility. The various functions which
5303 convert status values to strings have been renamed to:
5304 OCSP_response_status_str(), OCSP_cert_status_str() and
5305 OCSP_crl_reason_str() and are no longer static. New options
5306 to verify nonce values and to disable verification. OCSP response
5307 printout format cleaned up.
5308 [Steve Henson]
5309
5310 *) Add additional OCSP certificate checks. These are those specified
5311 in RFC2560. This consists of two separate checks: the CA of the
5312 certificate being checked must either be the OCSP signer certificate
5313 or the issuer of the OCSP signer certificate. In the latter case the
5314 OCSP signer certificate must contain the OCSP signing extended key
5315 usage. This check is performed by attempting to match the OCSP
5316 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
5317 in the OCSP_CERTID structures of the response.
5318 [Steve Henson]
5319
5320 *) Initial OCSP certificate verification added to OCSP_basic_verify()
5321 and related routines. This uses the standard OpenSSL certificate
5322 verify routines to perform initial checks (just CA validity) and
5323 to obtain the certificate chain. Then additional checks will be
5324 performed on the chain. Currently the root CA is checked to see
5325 if it is explicitly trusted for OCSP signing. This is used to set
5326 a root CA as a global signing root: that is any certificate that
5327 chains to that CA is an acceptable OCSP signing certificate.
5328 [Steve Henson]
5329
5330 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
5331 extensions from a separate configuration file.
5332 As when reading extensions from the main configuration file,
5333 the '-extensions ...' option may be used for specifying the
5334 section to use.
5335 [Massimiliano Pala <madwolf@comune.modena.it>]
5336
5337 *) New OCSP utility. Allows OCSP requests to be generated or
5338 read. The request can be sent to a responder and the output
5339 parsed, outputed or printed in text form. Not complete yet:
5340 still needs to check the OCSP response validity.
5341 [Steve Henson]
5342
5343 *) New subcommands for 'openssl ca':
5344 'openssl ca -status <serial>' prints the status of the cert with
5345 the given serial number (according to the index file).
5346 'openssl ca -updatedb' updates the expiry status of certificates
5347 in the index file.
5348 [Massimiliano Pala <madwolf@comune.modena.it>]
5349
5350 *) New '-newreq-nodes' command option to CA.pl. This is like
5351 '-newreq', but calls 'openssl req' with the '-nodes' option
5352 so that the resulting key is not encrypted.
5353 [Damien Miller <djm@mindrot.org>]
5354
5355 *) New configuration for the GNU Hurd.
5356 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
5357
5358 *) Initial code to implement OCSP basic response verify. This
5359 is currently incomplete. Currently just finds the signer's
5360 certificate and verifies the signature on the response.
5361 [Steve Henson]
5362
5363 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
5364 value of OPENSSLDIR. This is available via the new '-d' option
5365 to 'openssl version', and is also included in 'openssl version -a'.
5366 [Bodo Moeller]
5367
5368 *) Allowing defining memory allocation callbacks that will be given
5369 file name and line number information in additional arguments
5370 (a const char* and an int). The basic functionality remains, as
5371 well as the original possibility to just replace malloc(),
5372 realloc() and free() by functions that do not know about these
5373 additional arguments. To register and find out the current
5374 settings for extended allocation functions, the following
5375 functions are provided:
5376
5377 CRYPTO_set_mem_ex_functions
5378 CRYPTO_set_locked_mem_ex_functions
5379 CRYPTO_get_mem_ex_functions
5380 CRYPTO_get_locked_mem_ex_functions
5381
5382 These work the same way as CRYPTO_set_mem_functions and friends.
5383 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
5384 extended allocation function is enabled.
5385 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
5386 a conventional allocation function is enabled.
5387 [Richard Levitte, Bodo Moeller]
5388
5389 *) Finish off removing the remaining LHASH function pointer casts.
5390 There should no longer be any prototype-casting required when using
5391 the LHASH abstraction, and any casts that remain are "bugs". See
5392 the callback types and macros at the head of lhash.h for details
5393 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
5394 [Geoff Thorpe]
5395
5396 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
5397 If /dev/[u]random devices are not available or do not return enough
5398 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
5399 be queried.
5400 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
5401 /etc/entropy will be queried once each in this sequence, quering stops
5402 when enough entropy was collected without querying more sockets.
5403 [Lutz Jaenicke]
5404
5405 *) Change the Unix RAND_poll() variant to be able to poll several
5406 random devices, as specified by DEVRANDOM, until a sufficient amount
5407 of data has been collected. We spend at most 10 ms on each file
5408 (select timeout) and read in non-blocking mode. DEVRANDOM now
5409 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
5410 (previously it was just the string "/dev/urandom"), so on typical
5411 platforms the 10 ms delay will never occur.
5412 Also separate out the Unix variant to its own file, rand_unix.c.
5413 For VMS, there's a currently-empty rand_vms.c.
5414 [Richard Levitte]
5415
5416 *) Move OCSP client related routines to ocsp_cl.c. These
5417 provide utility functions which an application needing
5418 to issue a request to an OCSP responder and analyse the
5419 response will typically need: as opposed to those which an
5420 OCSP responder itself would need which will be added later.
5421
5422 OCSP_request_sign() signs an OCSP request with an API similar
5423 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5424 response. OCSP_response_get1_basic() extracts basic response
5425 from response. OCSP_resp_find_status(): finds and extracts status
5426 information from an OCSP_CERTID structure (which will be created
5427 when the request structure is built). These are built from lower
5428 level functions which work on OCSP_SINGLERESP structures but
5429 wont normally be used unless the application wishes to examine
5430 extensions in the OCSP response for example.
5431
5432 Replace nonce routines with a pair of functions.
5433 OCSP_request_add1_nonce() adds a nonce value and optionally
5434 generates a random value. OCSP_check_nonce() checks the
5435 validity of the nonce in an OCSP response.
5436 [Steve Henson]
5437
5438 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5439 This doesn't copy the supplied OCSP_CERTID and avoids the
5440 need to free up the newly created id. Change return type
5441 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5442 This can then be used to add extensions to the request.
5443 Deleted OCSP_request_new(), since most of its functionality
5444 is now in OCSP_REQUEST_new() (and the case insensitive name
5445 clash) apart from the ability to set the request name which
5446 will be added elsewhere.
5447 [Steve Henson]
5448
5449 *) Update OCSP API. Remove obsolete extensions argument from
5450 various functions. Extensions are now handled using the new
5451 OCSP extension code. New simple OCSP HTTP function which
5452 can be used to send requests and parse the response.
5453 [Steve Henson]
5454
5455 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5456 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5457 uses the special reorder version of SET OF to sort the attributes
5458 and reorder them to match the encoded order. This resolves a long
5459 standing problem: a verify on a PKCS7 structure just after signing
5460 it used to fail because the attribute order did not match the
5461 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5462 it uses the received order. This is necessary to tolerate some broken
5463 software that does not order SET OF. This is handled by encoding
5464 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5465 to produce the required SET OF.
5466 [Steve Henson]
5467
5468 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5469 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5470 files to get correct declarations of the ASN.1 item variables.
5471 [Richard Levitte]
5472
5473 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5474 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5475 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5476 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5477 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5478 ASN1_ITEM and no wrapper functions.
5479 [Steve Henson]
5480
5481 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5482 replace the old function pointer based I/O routines. Change most of
5483 the *_d2i_bio() and *_d2i_fp() functions to use these.
5484 [Steve Henson]
5485
5486 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5487 lines, recognice more "algorithms" that can be deselected, and make
5488 it complain about algorithm deselection that isn't recognised.
5489 [Richard Levitte]
5490
5491 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5492 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5493 to use new functions. Add NO_ASN1_OLD which can be set to remove
5494 some old style ASN1 functions: this can be used to determine if old
5495 code will still work when these eventually go away.
5496 [Steve Henson]
5497
5498 *) New extension functions for OCSP structures, these follow the
5499 same conventions as certificates and CRLs.
5500 [Steve Henson]
5501
5502 *) New function X509V3_add1_i2d(). This automatically encodes and
5503 adds an extension. Its behaviour can be customised with various
5504 flags to append, replace or delete. Various wrappers added for
5505 certifcates and CRLs.
5506 [Steve Henson]
5507
5508 *) Fix to avoid calling the underlying ASN1 print routine when
5509 an extension cannot be parsed. Correct a typo in the
5510 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5511 [Steve Henson]
5512
5513 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5514 entries for variables.
5515 [Steve Henson]
5516
5517 *) Add functionality to apps/openssl.c for detecting locking
5518 problems: As the program is single-threaded, all we have
5519 to do is register a locking callback using an array for
5520 storing which locks are currently held by the program.
5521 [Bodo Moeller]
5522
5523 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5524 SSL_get_ex_data_X509_STORE_idx(), which is used in
5525 ssl_verify_cert_chain() and thus can be called at any time
5526 during TLS/SSL handshakes so that thread-safety is essential.
5527 Unfortunately, the ex_data design is not at all suited
5528 for multi-threaded use, so it probably should be abolished.
5529 [Bodo Moeller]
5530
5531 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5532 [Broadcom, tweaked and integrated by Geoff Thorpe]
5533
5534 *) Move common extension printing code to new function
5535 X509V3_print_extensions(). Reorganise OCSP print routines and
5536 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5537 [Steve Henson]
5538
5539 *) New function X509_signature_print() to remove duplication in some
5540 print routines.
5541 [Steve Henson]
5542
5543 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5544 set (this was treated exactly the same as SET OF previously). This
5545 is used to reorder the STACK representing the structure to match the
5546 encoding. This will be used to get round a problem where a PKCS7
5547 structure which was signed could not be verified because the STACK
5548 order did not reflect the encoded order.
5549 [Steve Henson]
5550
5551 *) Reimplement the OCSP ASN1 module using the new code.
5552 [Steve Henson]
5553
5554 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5555 for its ASN1 operations. The old style function pointers still exist
5556 for now but they will eventually go away.
5557 [Steve Henson]
5558
5559 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5560 completely replaces the old ASN1 functionality with a table driven
5561 encoder and decoder which interprets an ASN1_ITEM structure describing
5562 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5563 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5564 has also been converted to the new form.
5565 [Steve Henson]
5566
5567 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5568 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5569 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5570 for negative moduli.
5571 [Bodo Moeller]
5572
5573 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5574 of not touching the result's sign bit.
5575 [Bodo Moeller]
5576
5577 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5578 set.
5579 [Bodo Moeller]
5580
5581 *) Changed the LHASH code to use prototypes for callbacks, and created
5582 macros to declare and implement thin (optionally static) functions
5583 that provide type-safety and avoid function pointer casting for the
5584 type-specific callbacks.
5585 [Geoff Thorpe]
5586
5587 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5588 RFC 2712.
5589 [Veers Staats <staatsvr@asc.hpc.mil>,
5590 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5591
5592 *) Reformat the FAQ so the different questions and answers can be divided
5593 in sections depending on the subject.
5594 [Richard Levitte]
5595
5596 *) Have the zlib compression code load ZLIB.DLL dynamically under
5597 Windows.
5598 [Richard Levitte]
5599
5600 *) New function BN_mod_sqrt for computing square roots modulo a prime
5601 (using the probabilistic Tonelli-Shanks algorithm unless
5602 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5603 be handled deterministically).
5604 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5605
5606 *) Make BN_mod_inverse faster by explicitly handling small quotients
5607 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5608 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5609 [Bodo Moeller]
5610
5611 *) New function BN_kronecker.
5612 [Bodo Moeller]
5613
5614 *) Fix BN_gcd so that it works on negative inputs; the result is
5615 positive unless both parameters are zero.
5616 Previously something reasonably close to an infinite loop was
5617 possible because numbers could be growing instead of shrinking
5618 in the implementation of Euclid's algorithm.
5619 [Bodo Moeller]
5620
5621 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5622 sign of the number in question.
5623
5624 Fix BN_is_word(a,w) to work correctly for w == 0.
5625
5626 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5627 because its test if the absolute value of 'a' equals 'w'.
5628 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5629 it exists mostly for use in the implementations of BN_is_zero(),
5630 BN_is_one(), and BN_is_word().
5631 [Bodo Moeller]
5632
5633 *) New function BN_swap.
5634 [Bodo Moeller]
5635
5636 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5637 the exponentiation functions are more likely to produce reasonable
5638 results on negative inputs.
5639 [Bodo Moeller]
5640
5641 *) Change BN_mod_mul so that the result is always non-negative.
5642 Previously, it could be negative if one of the factors was negative;
5643 I don't think anyone really wanted that behaviour.
5644 [Bodo Moeller]
5645
5646 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5647 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5648 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5649 and add new functions:
5650
5651 BN_nnmod
5652 BN_mod_sqr
5653 BN_mod_add
5654 BN_mod_add_quick
5655 BN_mod_sub
5656 BN_mod_sub_quick
5657 BN_mod_lshift1
5658 BN_mod_lshift1_quick
5659 BN_mod_lshift
5660 BN_mod_lshift_quick
5661
5662 These functions always generate non-negative results.
5663
5664 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5665 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5666
5667 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5668 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5669 be reduced modulo m.
5670 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5671
5672 #if 0
5673 The following entry accidentily appeared in the CHANGES file
5674 distributed with OpenSSL 0.9.7. The modifications described in
5675 it do *not* apply to OpenSSL 0.9.7.
5676
5677 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5678 was actually never needed) and in BN_mul(). The removal in BN_mul()
5679 required a small change in bn_mul_part_recursive() and the addition
5680 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5681 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5682 bn_sub_words() and bn_add_words() except they take arrays with
5683 differing sizes.
5684 [Richard Levitte]
5685 #endif
5686
5687 *) In 'openssl passwd', verify passwords read from the terminal
5688 unless the '-salt' option is used (which usually means that
5689 verification would just waste user's time since the resulting
5690 hash is going to be compared with some given password hash)
5691 or the new '-noverify' option is used.
5692
5693 This is an incompatible change, but it does not affect
5694 non-interactive use of 'openssl passwd' (passwords on the command
5695 line, '-stdin' option, '-in ...' option) and thus should not
5696 cause any problems.
5697 [Bodo Moeller]
5698
5699 *) Remove all references to RSAref, since there's no more need for it.
5700 [Richard Levitte]
5701
5702 *) Make DSO load along a path given through an environment variable
5703 (SHLIB_PATH) with shl_load().
5704 [Richard Levitte]
5705
5706 *) Constify the ENGINE code as a result of BIGNUM constification.
5707 Also constify the RSA code and most things related to it. In a
5708 few places, most notable in the depth of the ASN.1 code, ugly
5709 casts back to non-const were required (to be solved at a later
5710 time)
5711 [Richard Levitte]
5712
5713 *) Make it so the openssl application has all engines loaded by default.
5714 [Richard Levitte]
5715
5716 *) Constify the BIGNUM routines a little more.
5717 [Richard Levitte]
5718
5719 *) Add the following functions:
5720
5721 ENGINE_load_cswift()
5722 ENGINE_load_chil()
5723 ENGINE_load_atalla()
5724 ENGINE_load_nuron()
5725 ENGINE_load_builtin_engines()
5726
5727 That way, an application can itself choose if external engines that
5728 are built-in in OpenSSL shall ever be used or not. The benefit is
5729 that applications won't have to be linked with libdl or other dso
5730 libraries unless it's really needed.
5731
5732 Changed 'openssl engine' to load all engines on demand.
5733 Changed the engine header files to avoid the duplication of some
5734 declarations (they differed!).
5735 [Richard Levitte]
5736
5737 *) 'openssl engine' can now list capabilities.
5738 [Richard Levitte]
5739
5740 *) Better error reporting in 'openssl engine'.
5741 [Richard Levitte]
5742
5743 *) Never call load_dh_param(NULL) in s_server.
5744 [Bodo Moeller]
5745
5746 *) Add engine application. It can currently list engines by name and
5747 identity, and test if they are actually available.
5748 [Richard Levitte]
5749
5750 *) Improve RPM specification file by forcing symbolic linking and making
5751 sure the installed documentation is also owned by root.root.
5752 [Damien Miller <djm@mindrot.org>]
5753
5754 *) Give the OpenSSL applications more possibilities to make use of
5755 keys (public as well as private) handled by engines.
5756 [Richard Levitte]
5757
5758 *) Add OCSP code that comes from CertCo.
5759 [Richard Levitte]
5760
5761 *) Add VMS support for the Rijndael code.
5762 [Richard Levitte]
5763
5764 *) Added untested support for Nuron crypto accelerator.
5765 [Ben Laurie]
5766
5767 *) Add support for external cryptographic devices. This code was
5768 previously distributed separately as the "engine" branch.
5769 [Geoff Thorpe, Richard Levitte]
5770
5771 *) Rework the filename-translation in the DSO code. It is now possible to
5772 have far greater control over how a "name" is turned into a filename
5773 depending on the operating environment and any oddities about the
5774 different shared library filenames on each system.
5775 [Geoff Thorpe]
5776
5777 *) Support threads on FreeBSD-elf in Configure.
5778 [Richard Levitte]
5779
5780 *) Fix for SHA1 assembly problem with MASM: it produces
5781 warnings about corrupt line number information when assembling
5782 with debugging information. This is caused by the overlapping
5783 of two sections.
5784 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5785
5786 *) NCONF changes.
5787 NCONF_get_number() has no error checking at all. As a replacement,
5788 NCONF_get_number_e() is defined (_e for "error checking") and is
5789 promoted strongly. The old NCONF_get_number is kept around for
5790 binary backward compatibility.
5791 Make it possible for methods to load from something other than a BIO,
5792 by providing a function pointer that is given a name instead of a BIO.
5793 For example, this could be used to load configuration data from an
5794 LDAP server.
5795 [Richard Levitte]
5796
5797 *) Fix for non blocking accept BIOs. Added new I/O special reason
5798 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5799 with non blocking I/O was not possible because no retry code was
5800 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5801 this case.
5802 [Steve Henson]
5803
5804 *) Added the beginnings of Rijndael support.
5805 [Ben Laurie]
5806
5807 *) Fix for bug in DirectoryString mask setting. Add support for
5808 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5809 to allow certificate printing to more controllable, additional
5810 'certopt' option to 'x509' to allow new printing options to be
5811 set.
5812 [Steve Henson]
5813
5814 *) Clean old EAY MD5 hack from e_os.h.
5815 [Richard Levitte]
5816
5817 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5818
5819 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5820 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5821 [Joe Orton, Steve Henson]
5822
5823 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5824
5825 *) Fix additional bug revealed by the NISCC test suite:
5826
5827 Stop bug triggering large recursion when presented with
5828 certain ASN.1 tags (CVE-2003-0851)
5829 [Steve Henson]
5830
5831 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5832
5833 *) Fix various bugs revealed by running the NISCC test suite:
5834
5835 Stop out of bounds reads in the ASN1 code when presented with
5836 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5837
5838 If verify callback ignores invalid public key errors don't try to check
5839 certificate signature with the NULL public key.
5840
5841 [Steve Henson]
5842
5843 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5844 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5845 specifications.
5846 [Steve Henson]
5847
5848 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5849 extra data after the compression methods not only for TLS 1.0
5850 but also for SSL 3.0 (as required by the specification).
5851 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5852
5853 *) Change X509_certificate_type() to mark the key as exported/exportable
5854 when it's 512 *bits* long, not 512 bytes.
5855 [Richard Levitte]
5856
5857 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5858
5859 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5860 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5861 a protocol version number mismatch like a decryption error
5862 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5863 [Bodo Moeller]
5864
5865 *) Turn on RSA blinding by default in the default implementation
5866 to avoid a timing attack. Applications that don't want it can call
5867 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5868 They would be ill-advised to do so in most cases.
5869 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5870
5871 *) Change RSA blinding code so that it works when the PRNG is not
5872 seeded (in this case, the secret RSA exponent is abused as
5873 an unpredictable seed -- if it is not unpredictable, there
5874 is no point in blinding anyway). Make RSA blinding thread-safe
5875 by remembering the creator's thread ID in rsa->blinding and
5876 having all other threads use local one-time blinding factors
5877 (this requires more computation than sharing rsa->blinding, but
5878 avoids excessive locking; and if an RSA object is not shared
5879 between threads, blinding will still be very fast).
5880 [Bodo Moeller]
5881
5882 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5883
5884 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5885 via timing by performing a MAC computation even if incorrrect
5886 block cipher padding has been found. This is a countermeasure
5887 against active attacks where the attacker has to distinguish
5888 between bad padding and a MAC verification error. (CVE-2003-0078)
5889
5890 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5891 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5892 Martin Vuagnoux (EPFL, Ilion)]
5893
5894 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5895
5896 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5897 memory from it's contents. This is done with a counter that will
5898 place alternating values in each byte. This can be used to solve
5899 two issues: 1) the removal of calls to memset() by highly optimizing
5900 compilers, and 2) cleansing with other values than 0, since those can
5901 be read through on certain media, for example a swap space on disk.
5902 [Geoff Thorpe]
5903
5904 *) Bugfix: client side session caching did not work with external caching,
5905 because the session->cipher setting was not restored when reloading
5906 from the external cache. This problem was masked, when
5907 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5908 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5909 [Lutz Jaenicke]
5910
5911 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5912 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5913 [Zeev Lieber <zeev-l@yahoo.com>]
5914
5915 *) Undo an undocumented change introduced in 0.9.6e which caused
5916 repeated calls to OpenSSL_add_all_ciphers() and
5917 OpenSSL_add_all_digests() to be ignored, even after calling
5918 EVP_cleanup().
5919 [Richard Levitte]
5920
5921 *) Change the default configuration reader to deal with last line not
5922 being properly terminated.
5923 [Richard Levitte]
5924
5925 *) Change X509_NAME_cmp() so it applies the special rules on handling
5926 DN values that are of type PrintableString, as well as RDNs of type
5927 emailAddress where the value has the type ia5String.
5928 [stefank@valicert.com via Richard Levitte]
5929
5930 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5931 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5932 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5933 the bitwise-OR of the two for use by the majority of applications
5934 wanting this behaviour, and update the docs. The documented
5935 behaviour and actual behaviour were inconsistent and had been
5936 changing anyway, so this is more a bug-fix than a behavioural
5937 change.
5938 [Geoff Thorpe, diagnosed by Nadav Har'El]
5939
5940 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5941 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5942 [Bodo Moeller]
5943
5944 *) Fix initialization code race conditions in
5945 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5946 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5947 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5948 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5949 ssl2_get_cipher_by_char(),
5950 ssl3_get_cipher_by_char().
5951 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5952
5953 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5954 the cached sessions are flushed, as the remove_cb() might use ex_data
5955 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5956 (see [openssl.org #212]).
5957 [Geoff Thorpe, Lutz Jaenicke]
5958
5959 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5960 length, instead of the encoding length to d2i_ASN1_OBJECT.
5961 [Steve Henson]
5962
5963 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5964
5965 *) [In 0.9.6g-engine release:]
5966 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5967 [Lynn Gazis <lgazis@rainbow.com>]
5968
5969 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5970
5971 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5972 and get fix the header length calculation.
5973 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5974 Alon Kantor <alonk@checkpoint.com> (and others),
5975 Steve Henson]
5976
5977 *) Use proper error handling instead of 'assertions' in buffer
5978 overflow checks added in 0.9.6e. This prevents DoS (the
5979 assertions could call abort()).
5980 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5981
5982 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5983
5984 *) Add various sanity checks to asn1_get_length() to reject
5985 the ASN1 length bytes if they exceed sizeof(long), will appear
5986 negative or the content length exceeds the length of the
5987 supplied buffer.
5988 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5989
5990 *) Fix cipher selection routines: ciphers without encryption had no flags
5991 for the cipher strength set and where therefore not handled correctly
5992 by the selection routines (PR #130).
5993 [Lutz Jaenicke]
5994
5995 *) Fix EVP_dsa_sha macro.
5996 [Nils Larsch]
5997
5998 *) New option
5999 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6000 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6001 that was added in OpenSSL 0.9.6d.
6002
6003 As the countermeasure turned out to be incompatible with some
6004 broken SSL implementations, the new option is part of SSL_OP_ALL.
6005 SSL_OP_ALL is usually employed when compatibility with weird SSL
6006 implementations is desired (e.g. '-bugs' option to 's_client' and
6007 's_server'), so the new option is automatically set in many
6008 applications.
6009 [Bodo Moeller]
6010
6011 *) Changes in security patch:
6012
6013 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6014 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6015 Air Force Materiel Command, USAF, under agreement number
6016 F30602-01-2-0537.
6017
6018 *) Add various sanity checks to asn1_get_length() to reject
6019 the ASN1 length bytes if they exceed sizeof(long), will appear
6020 negative or the content length exceeds the length of the
6021 supplied buffer. (CVE-2002-0659)
6022 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6023
6024 *) Assertions for various potential buffer overflows, not known to
6025 happen in practice.
6026 [Ben Laurie (CHATS)]
6027
6028 *) Various temporary buffers to hold ASCII versions of integers were
6029 too small for 64 bit platforms. (CVE-2002-0655)
6030 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6031
6032 *) Remote buffer overflow in SSL3 protocol - an attacker could
6033 supply an oversized session ID to a client. (CVE-2002-0656)
6034 [Ben Laurie (CHATS)]
6035
6036 *) Remote buffer overflow in SSL2 protocol - an attacker could
6037 supply an oversized client master key. (CVE-2002-0656)
6038 [Ben Laurie (CHATS)]
6039
6040 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6041
6042 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6043 encoded as NULL) with id-dsa-with-sha1.
6044 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6045
6046 *) Check various X509_...() return values in apps/req.c.
6047 [Nils Larsch <nla@trustcenter.de>]
6048
6049 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6050 an end-of-file condition would erronously be flagged, when the CRLF
6051 was just at the end of a processed block. The bug was discovered when
6052 processing data through a buffering memory BIO handing the data to a
6053 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6054 <ptsekov@syntrex.com> and Nedelcho Stanev.
6055 [Lutz Jaenicke]
6056
6057 *) Implement a countermeasure against a vulnerability recently found
6058 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6059 before application data chunks to avoid the use of known IVs
6060 with data potentially chosen by the attacker.
6061 [Bodo Moeller]
6062
6063 *) Fix length checks in ssl3_get_client_hello().
6064 [Bodo Moeller]
6065
6066 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6067 to prevent ssl3_read_internal() from incorrectly assuming that
6068 ssl3_read_bytes() found application data while handshake
6069 processing was enabled when in fact s->s3->in_read_app_data was
6070 merely automatically cleared during the initial handshake.
6071 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6072
6073 *) Fix object definitions for Private and Enterprise: they were not
6074 recognized in their shortname (=lowercase) representation. Extend
6075 obj_dat.pl to issue an error when using undefined keywords instead
6076 of silently ignoring the problem (Svenning Sorensen
6077 <sss@sss.dnsalias.net>).
6078 [Lutz Jaenicke]
6079
6080 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6081 generators, i.e. generators other than 2 and 5. (Previously, the
6082 code did not properly initialise the 'add' and 'rem' values to
6083 BN_generate_prime().)
6084
6085 In the new general case, we do not insist that 'generator' is
6086 actually a primitive root: This requirement is rather pointless;
6087 a generator of the order-q subgroup is just as good, if not
6088 better.
6089 [Bodo Moeller]
6090
6091 *) Map new X509 verification errors to alerts. Discovered and submitted by
6092 Tom Wu <tom@arcot.com>.
6093 [Lutz Jaenicke]
6094
6095 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6096 returning non-zero before the data has been completely received
6097 when using non-blocking I/O.
6098 [Bodo Moeller; problem pointed out by John Hughes]
6099
6100 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6101 [Ben Laurie, Lutz Jaenicke]
6102
6103 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6104 Yoram Zahavi <YoramZ@gilian.com>).
6105 [Lutz Jaenicke]
6106
6107 *) Add information about CygWin 1.3 and on, and preserve proper
6108 configuration for the versions before that.
6109 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6110
6111 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6112 check whether we deal with a copy of a session and do not delete from
6113 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6114 <izhar@checkpoint.com>.
6115 [Lutz Jaenicke]
6116
6117 *) Do not store session data into the internal session cache, if it
6118 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6119 flag is set). Proposed by Aslam <aslam@funk.com>.
6120 [Lutz Jaenicke]
6121
6122 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6123 value is 0.
6124 [Richard Levitte]
6125
6126 *) [In 0.9.6d-engine release:]
6127 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6128 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6129
6130 *) Add the configuration target linux-s390x.
6131 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6132
6133 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6134 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6135 variable as an indication that a ClientHello message has been
6136 received. As the flag value will be lost between multiple
6137 invocations of ssl3_accept when using non-blocking I/O, the
6138 function may not be aware that a handshake has actually taken
6139 place, thus preventing a new session from being added to the
6140 session cache.
6141
6142 To avoid this problem, we now set s->new_session to 2 instead of
6143 using a local variable.
6144 [Lutz Jaenicke, Bodo Moeller]
6145
6146 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6147 if the SSL_R_LENGTH_MISMATCH error is detected.
6148 [Geoff Thorpe, Bodo Moeller]
6149
6150 *) New 'shared_ldflag' column in Configure platform table.
6151 [Richard Levitte]
6152
6153 *) Fix EVP_CIPHER_mode macro.
6154 ["Dan S. Camper" <dan@bti.net>]
6155
6156 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6157 type, we must throw them away by setting rr->length to 0.
6158 [D P Chang <dpc@qualys.com>]
6159
6160 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6161
6162 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6163 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6164 worked incorrectly for those cases where range = 10..._2 and
6165 3*range is two bits longer than range.)
6166 [Bodo Moeller]
6167
6168 *) Only add signing time to PKCS7 structures if it is not already
6169 present.
6170 [Steve Henson]
6171
6172 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6173 OBJ_ld_ce should be OBJ_id_ce.
6174 Also some ip-pda OIDs in crypto/objects/objects.txt were
6175 incorrect (cf. RFC 3039).
6176 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6177
6178 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6179 returns early because it has nothing to do.
6180 [Andy Schneider <andy.schneider@bjss.co.uk>]
6181
6182 *) [In 0.9.6c-engine release:]
6183 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6184 [Andy Schneider <andy.schneider@bjss.co.uk>]
6185
6186 *) [In 0.9.6c-engine release:]
6187 Add support for Cryptographic Appliance's keyserver technology.
6188 (Use engine 'keyclient')
6189 [Cryptographic Appliances and Geoff Thorpe]
6190
6191 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6192 is called via tools/c89.sh because arguments have to be
6193 rearranged (all '-L' options must appear before the first object
6194 modules).
6195 [Richard Shapiro <rshapiro@abinitio.com>]
6196
6197 *) [In 0.9.6c-engine release:]
6198 Add support for Broadcom crypto accelerator cards, backported
6199 from 0.9.7.
6200 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6201
6202 *) [In 0.9.6c-engine release:]
6203 Add support for SureWare crypto accelerator cards from
6204 Baltimore Technologies. (Use engine 'sureware')
6205 [Baltimore Technologies and Mark Cox]
6206
6207 *) [In 0.9.6c-engine release:]
6208 Add support for crypto accelerator cards from Accelerated
6209 Encryption Processing, www.aep.ie. (Use engine 'aep')
6210 [AEP Inc. and Mark Cox]
6211
6212 *) Add a configuration entry for gcc on UnixWare.
6213 [Gary Benson <gbenson@redhat.com>]
6214
6215 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6216 messages are stored in a single piece (fixed-length part and
6217 variable-length part combined) and fix various bugs found on the way.
6218 [Bodo Moeller]
6219
6220 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6221 instead. BIO_gethostbyname() does not know what timeouts are
6222 appropriate, so entries would stay in cache even when they have
6223 become invalid.
6224 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6225
6226 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6227 faced with a pathologically small ClientHello fragment that does
6228 not contain client_version: Instead of aborting with an error,
6229 simply choose the highest available protocol version (i.e.,
6230 TLS 1.0 unless it is disabled). In practice, ClientHello
6231 messages are never sent like this, but this change gives us
6232 strictly correct behaviour at least for TLS.
6233 [Bodo Moeller]
6234
6235 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6236 never resets s->method to s->ctx->method when called from within
6237 one of the SSL handshake functions.
6238 [Bodo Moeller; problem pointed out by Niko Baric]
6239
6240 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6241 (sent using the client's version number) if client_version is
6242 smaller than the protocol version in use. Also change
6243 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6244 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6245 the client will at least see that alert.
6246 [Bodo Moeller]
6247
6248 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6249 correctly.
6250 [Bodo Moeller]
6251
6252 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6253 client receives HelloRequest while in a handshake.
6254 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6255
6256 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6257 should end in 'break', not 'goto end' which circuments various
6258 cleanups done in state SSL_ST_OK. But session related stuff
6259 must be disabled for SSL_ST_OK in the case that we just sent a
6260 HelloRequest.
6261
6262 Also avoid some overhead by not calling ssl_init_wbio_buffer()
6263 before just sending a HelloRequest.
6264 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
6265
6266 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
6267 reveal whether illegal block cipher padding was found or a MAC
6268 verification error occured. (Neither SSLerr() codes nor alerts
6269 are directly visible to potential attackers, but the information
6270 may leak via logfiles.)
6271
6272 Similar changes are not required for the SSL 2.0 implementation
6273 because the number of padding bytes is sent in clear for SSL 2.0,
6274 and the extra bytes are just ignored. However ssl/s2_pkt.c
6275 failed to verify that the purported number of padding bytes is in
6276 the legal range.
6277 [Bodo Moeller]
6278
6279 *) Add OpenUNIX-8 support including shared libraries
6280 (Boyd Lynn Gerber <gerberb@zenez.com>).
6281 [Lutz Jaenicke]
6282
6283 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
6284 'wristwatch attack' using huge encoding parameters (cf.
6285 James H. Manger's CRYPTO 2001 paper). Note that the
6286 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
6287 encoding parameters and hence was not vulnerable.
6288 [Bodo Moeller]
6289
6290 *) BN_sqr() bug fix.
6291 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
6292
6293 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
6294 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
6295 followed by modular reduction.
6296 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
6297
6298 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
6299 equivalent based on BN_pseudo_rand() instead of BN_rand().
6300 [Bodo Moeller]
6301
6302 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
6303 This function was broken, as the check for a new client hello message
6304 to handle SGC did not allow these large messages.
6305 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
6306 [Lutz Jaenicke]
6307
6308 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
6309 [Lutz Jaenicke]
6310
6311 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
6312 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
6313 [Lutz Jaenicke]
6314
6315 *) Rework the configuration and shared library support for Tru64 Unix.
6316 The configuration part makes use of modern compiler features and
6317 still retains old compiler behavior for those that run older versions
6318 of the OS. The shared library support part includes a variant that
6319 uses the RPATH feature, and is available through the special
6320 configuration target "alpha-cc-rpath", which will never be selected
6321 automatically.
6322 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
6323
6324 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
6325 with the same message size as in ssl3_get_certificate_request().
6326 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
6327 messages might inadvertently be reject as too long.
6328 [Petr Lampa <lampa@fee.vutbr.cz>]
6329
6330 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
6331 [Andy Polyakov]
6332
6333 *) Modified SSL library such that the verify_callback that has been set
6334 specificly for an SSL object with SSL_set_verify() is actually being
6335 used. Before the change, a verify_callback set with this function was
6336 ignored and the verify_callback() set in the SSL_CTX at the time of
6337 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
6338 to allow the necessary settings.
6339 [Lutz Jaenicke]
6340
6341 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
6342 explicitly to NULL, as at least on Solaris 8 this seems not always to be
6343 done automatically (in contradiction to the requirements of the C
6344 standard). This made problems when used from OpenSSH.
6345 [Lutz Jaenicke]
6346
6347 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
6348 dh->length and always used
6349
6350 BN_rand_range(priv_key, dh->p).
6351
6352 BN_rand_range() is not necessary for Diffie-Hellman, and this
6353 specific range makes Diffie-Hellman unnecessarily inefficient if
6354 dh->length (recommended exponent length) is much smaller than the
6355 length of dh->p. We could use BN_rand_range() if the order of
6356 the subgroup was stored in the DH structure, but we only have
6357 dh->length.
6358
6359 So switch back to
6360
6361 BN_rand(priv_key, l, ...)
6362
6363 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
6364 otherwise.
6365 [Bodo Moeller]
6366
6367 *) In
6368
6369 RSA_eay_public_encrypt
6370 RSA_eay_private_decrypt
6371 RSA_eay_private_encrypt (signing)
6372 RSA_eay_public_decrypt (signature verification)
6373
6374 (default implementations for RSA_public_encrypt,
6375 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
6376 always reject numbers >= n.
6377 [Bodo Moeller]
6378
6379 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
6380 to synchronize access to 'locking_thread'. This is necessary on
6381 systems where access to 'locking_thread' (an 'unsigned long'
6382 variable) is not atomic.
6383 [Bodo Moeller]
6384
6385 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
6386 *before* setting the 'crypto_lock_rand' flag. The previous code had
6387 a race condition if 0 is a valid thread ID.
6388 [Travis Vitek <vitek@roguewave.com>]
6389
6390 *) Add support for shared libraries under Irix.
6391 [Albert Chin-A-Young <china@thewrittenword.com>]
6392
6393 *) Add configuration option to build on Linux on both big-endian and
6394 little-endian MIPS.
6395 [Ralf Baechle <ralf@uni-koblenz.de>]
6396
6397 *) Add the possibility to create shared libraries on HP-UX.
6398 [Richard Levitte]
6399
6400 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
6401
6402 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
6403 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
6404 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
6405 PRNG state recovery was possible based on the output of
6406 one PRNG request appropriately sized to gain knowledge on
6407 'md' followed by enough consecutive 1-byte PRNG requests
6408 to traverse all of 'state'.
6409
6410 1. When updating 'md_local' (the current thread's copy of 'md')
6411 during PRNG output generation, hash all of the previous
6412 'md_local' value, not just the half used for PRNG output.
6413
6414 2. Make the number of bytes from 'state' included into the hash
6415 independent from the number of PRNG bytes requested.
6416
6417 The first measure alone would be sufficient to avoid
6418 Markku-Juhani's attack. (Actually it had never occurred
6419 to me that the half of 'md_local' used for chaining was the
6420 half from which PRNG output bytes were taken -- I had always
6421 assumed that the secret half would be used.) The second
6422 measure makes sure that additional data from 'state' is never
6423 mixed into 'md_local' in small portions; this heuristically
6424 further strengthens the PRNG.
6425 [Bodo Moeller]
6426
6427 *) Fix crypto/bn/asm/mips3.s.
6428 [Andy Polyakov]
6429
6430 *) When only the key is given to "enc", the IV is undefined. Print out
6431 an error message in this case.
6432 [Lutz Jaenicke]
6433
6434 *) Handle special case when X509_NAME is empty in X509 printing routines.
6435 [Steve Henson]
6436
6437 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6438 positive and less than q.
6439 [Bodo Moeller]
6440
6441 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6442 used: it isn't thread safe and the add_lock_callback should handle
6443 that itself.
6444 [Paul Rose <Paul.Rose@bridge.com>]
6445
6446 *) Verify that incoming data obeys the block size in
6447 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6448 [Bodo Moeller]
6449
6450 *) Fix OAEP check.
6451 [Ulf Möller, Bodo Möller]
6452
6453 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6454 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6455 when fixing the server behaviour for backwards-compatible 'client
6456 hello' messages. (Note that the attack is impractical against
6457 SSL 3.0 and TLS 1.0 anyway because length and version checking
6458 means that the probability of guessing a valid ciphertext is
6459 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6460 paper.)
6461
6462 Before 0.9.5, the countermeasure (hide the error by generating a
6463 random 'decryption result') did not work properly because
6464 ERR_clear_error() was missing, meaning that SSL_get_error() would
6465 detect the supposedly ignored error.
6466
6467 Both problems are now fixed.
6468 [Bodo Moeller]
6469
6470 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6471 (previously it was 1024).
6472 [Bodo Moeller]
6473
6474 *) Fix for compatibility mode trust settings: ignore trust settings
6475 unless some valid trust or reject settings are present.
6476 [Steve Henson]
6477
6478 *) Fix for blowfish EVP: its a variable length cipher.
6479 [Steve Henson]
6480
6481 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6482 parameters in DSA public key structures and return an error in the
6483 DSA routines if parameters are absent.
6484 [Steve Henson]
6485
6486 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6487 in the current directory if neither $RANDFILE nor $HOME was set.
6488 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6489 caused some confusion to Windows users who haven't defined $HOME.
6490 Thus RAND_file_name() is changed again: e_os.h can define a
6491 DEFAULT_HOME, which will be used if $HOME is not set.
6492 For Windows, we use "C:"; on other platforms, we still require
6493 environment variables.
6494
6495 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6496 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6497 having multiple threads call RAND_poll() concurrently.
6498 [Bodo Moeller]
6499
6500 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6501 combination of a flag and a thread ID variable.
6502 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6503 flag), *other* threads can enter ssleay_add_bytes without obeying
6504 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6505 that they do not hold after the first thread unsets add_do_not_lock).
6506 [Bodo Moeller]
6507
6508 *) Change bctest again: '-x' expressions are not available in all
6509 versions of 'test'.
6510 [Bodo Moeller]
6511
6512 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6513
6514 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6515 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6516
6517 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6518 the default extension for executables, if any. Also, make the perl
6519 scripts that use symlink() to test if it really exists and use "cp"
6520 if it doesn't. All this made OpenSSL compilable and installable in
6521 CygWin.
6522 [Richard Levitte]
6523
6524 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6525 If SEQUENCE is length is indefinite just set c->slen to the total
6526 amount of data available.
6527 [Steve Henson, reported by shige@FreeBSD.org]
6528 [This change does not apply to 0.9.7.]
6529
6530 *) Change bctest to avoid here-documents inside command substitution
6531 (workaround for FreeBSD /bin/sh bug).
6532 For compatibility with Ultrix, avoid shell functions (introduced
6533 in the bctest version that searches along $PATH).
6534 [Bodo Moeller]
6535
6536 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6537 with des_encrypt() defined on some operating systems, like Solaris
6538 and UnixWare.
6539 [Richard Levitte]
6540
6541 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6542 On the Importance of Eliminating Errors in Cryptographic
6543 Computations, J. Cryptology 14 (2001) 2, 101-119,
6544 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6545 [Ulf Moeller]
6546
6547 *) MIPS assembler BIGNUM division bug fix.
6548 [Andy Polyakov]
6549
6550 *) Disabled incorrect Alpha assembler code.
6551 [Richard Levitte]
6552
6553 *) Fix PKCS#7 decode routines so they correctly update the length
6554 after reading an EOC for the EXPLICIT tag.
6555 [Steve Henson]
6556 [This change does not apply to 0.9.7.]
6557
6558 *) Fix bug in PKCS#12 key generation routines. This was triggered
6559 if a 3DES key was generated with a 0 initial byte. Include
6560 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6561 (but broken) behaviour.
6562 [Steve Henson]
6563
6564 *) Enhance bctest to search for a working bc along $PATH and print
6565 it when found.
6566 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6567
6568 *) Fix memory leaks in err.c: free err_data string if necessary;
6569 don't write to the wrong index in ERR_set_error_data.
6570 [Bodo Moeller]
6571
6572 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6573 did not exist.
6574 [Bodo Moeller]
6575
6576 *) Replace rdtsc with _emit statements for VC++ version 5.
6577 [Jeremy Cooper <jeremy@baymoo.org>]
6578
6579 *) Make it possible to reuse SSLv2 sessions.
6580 [Richard Levitte]
6581
6582 *) In copy_email() check for >= 0 as a return value for
6583 X509_NAME_get_index_by_NID() since 0 is a valid index.
6584 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6585
6586 *) Avoid coredump with unsupported or invalid public keys by checking if
6587 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6588 PKCS7_verify() fails with non detached data.
6589 [Steve Henson]
6590
6591 *) Don't use getenv in library functions when run as setuid/setgid.
6592 New function OPENSSL_issetugid().
6593 [Ulf Moeller]
6594
6595 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6596 due to incorrect handling of multi-threading:
6597
6598 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6599
6600 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6601
6602 3. Count how many times MemCheck_off() has been called so that
6603 nested use can be treated correctly. This also avoids
6604 inband-signalling in the previous code (which relied on the
6605 assumption that thread ID 0 is impossible).
6606 [Bodo Moeller]
6607
6608 *) Add "-rand" option also to s_client and s_server.
6609 [Lutz Jaenicke]
6610
6611 *) Fix CPU detection on Irix 6.x.
6612 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6613 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6614
6615 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6616 was empty.
6617 [Steve Henson]
6618 [This change does not apply to 0.9.7.]
6619
6620 *) Use the cached encoding of an X509_NAME structure rather than
6621 copying it. This is apparently the reason for the libsafe "errors"
6622 but the code is actually correct.
6623 [Steve Henson]
6624
6625 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6626 Bleichenbacher's DSA attack.
6627 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6628 to be set and top=0 forces the highest bit to be set; top=-1 is new
6629 and leaves the highest bit random.
6630 [Ulf Moeller, Bodo Moeller]
6631
6632 *) In the NCONF_...-based implementations for CONF_... queries
6633 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6634 a temporary CONF structure with the data component set to NULL
6635 (which gives segmentation faults in lh_retrieve).
6636 Instead, use NULL for the CONF pointer in CONF_get_string and
6637 CONF_get_number (which may use environment variables) and directly
6638 return NULL from CONF_get_section.
6639 [Bodo Moeller]
6640
6641 *) Fix potential buffer overrun for EBCDIC.
6642 [Ulf Moeller]
6643
6644 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6645 keyUsage if basicConstraints absent for a CA.
6646 [Steve Henson]
6647
6648 *) Make SMIME_write_PKCS7() write mail header values with a format that
6649 is more generally accepted (no spaces before the semicolon), since
6650 some programs can't parse those values properly otherwise. Also make
6651 sure BIO's that break lines after each write do not create invalid
6652 headers.
6653 [Richard Levitte]
6654
6655 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6656 macros previously used would not encode an empty SEQUENCE OF
6657 and break the signature.
6658 [Steve Henson]
6659 [This change does not apply to 0.9.7.]
6660
6661 *) Zero the premaster secret after deriving the master secret in
6662 DH ciphersuites.
6663 [Steve Henson]
6664
6665 *) Add some EVP_add_digest_alias registrations (as found in
6666 OpenSSL_add_all_digests()) to SSL_library_init()
6667 aka OpenSSL_add_ssl_algorithms(). This provides improved
6668 compatibility with peers using X.509 certificates
6669 with unconventional AlgorithmIdentifier OIDs.
6670 [Bodo Moeller]
6671
6672 *) Fix for Irix with NO_ASM.
6673 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6674
6675 *) ./config script fixes.
6676 [Ulf Moeller, Richard Levitte]
6677
6678 *) Fix 'openssl passwd -1'.
6679 [Bodo Moeller]
6680
6681 *) Change PKCS12_key_gen_asc() so it can cope with non null
6682 terminated strings whose length is passed in the passlen
6683 parameter, for example from PEM callbacks. This was done
6684 by adding an extra length parameter to asc2uni().
6685 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6686
6687 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6688 call failed, free the DSA structure.
6689 [Bodo Moeller]
6690
6691 *) Fix to uni2asc() to cope with zero length Unicode strings.
6692 These are present in some PKCS#12 files.
6693 [Steve Henson]
6694
6695 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6696 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6697 when writing a 32767 byte record.
6698 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6699
6700 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6701 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6702
6703 (RSA objects have a reference count access to which is protected
6704 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6705 so they are meant to be shared between threads.)
6706 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6707 "Reddie, Steven" <Steven.Reddie@ca.com>]
6708
6709 *) Fix a deadlock in CRYPTO_mem_leaks().
6710 [Bodo Moeller]
6711
6712 *) Use better test patterns in bntest.
6713 [Ulf Möller]
6714
6715 *) rand_win.c fix for Borland C.
6716 [Ulf Möller]
6717
6718 *) BN_rshift bugfix for n == 0.
6719 [Bodo Moeller]
6720
6721 *) Add a 'bctest' script that checks for some known 'bc' bugs
6722 so that 'make test' does not abort just because 'bc' is broken.
6723 [Bodo Moeller]
6724
6725 *) Store verify_result within SSL_SESSION also for client side to
6726 avoid potential security hole. (Re-used sessions on the client side
6727 always resulted in verify_result==X509_V_OK, not using the original
6728 result of the server certificate verification.)
6729 [Lutz Jaenicke]
6730
6731 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6732 SSL3_RT_APPLICATION_DATA, return 0.
6733 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6734 [Bodo Moeller]
6735
6736 *) Fix SSL_peek:
6737 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6738 releases, have been re-implemented by renaming the previous
6739 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6740 and ssl3_read_internal, respectively, and adding 'peek' parameters
6741 to them. The new ssl[23]_{read,peek} functions are calls to
6742 ssl[23]_read_internal with the 'peek' flag set appropriately.
6743 A 'peek' parameter has also been added to ssl3_read_bytes, which
6744 does the actual work for ssl3_read_internal.
6745 [Bodo Moeller]
6746
6747 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6748 the method-specific "init()" handler. Also clean up ex_data after
6749 calling the method-specific "finish()" handler. Previously, this was
6750 happening the other way round.
6751 [Geoff Thorpe]
6752
6753 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6754 The previous value, 12, was not always sufficient for BN_mod_exp().
6755 [Bodo Moeller]
6756
6757 *) Make sure that shared libraries get the internal name engine with
6758 the full version number and not just 0. This should mark the
6759 shared libraries as not backward compatible. Of course, this should
6760 be changed again when we can guarantee backward binary compatibility.
6761 [Richard Levitte]
6762
6763 *) Fix typo in get_cert_by_subject() in by_dir.c
6764 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6765
6766 *) Rework the system to generate shared libraries:
6767
6768 - Make note of the expected extension for the shared libraries and
6769 if there is a need for symbolic links from for example libcrypto.so.0
6770 to libcrypto.so.0.9.7. There is extended info in Configure for
6771 that.
6772
6773 - Make as few rebuilds of the shared libraries as possible.
6774
6775 - Still avoid linking the OpenSSL programs with the shared libraries.
6776
6777 - When installing, install the shared libraries separately from the
6778 static ones.
6779 [Richard Levitte]
6780
6781 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6782
6783 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6784 and not in SSL_clear because the latter is also used by the
6785 accept/connect functions; previously, the settings made by
6786 SSL_set_read_ahead would be lost during the handshake.
6787 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6788
6789 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6790 Previously, it would create entries for disableed algorithms no
6791 matter what.
6792 [Richard Levitte]
6793
6794 *) Added several new manual pages for SSL_* function.
6795 [Lutz Jaenicke]
6796
6797 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6798
6799 *) In ssl23_get_client_hello, generate an error message when faced
6800 with an initial SSL 3.0/TLS record that is too small to contain the
6801 first two bytes of the ClientHello message, i.e. client_version.
6802 (Note that this is a pathologic case that probably has never happened
6803 in real life.) The previous approach was to use the version number
6804 from the record header as a substitute; but our protocol choice
6805 should not depend on that one because it is not authenticated
6806 by the Finished messages.
6807 [Bodo Moeller]
6808
6809 *) More robust randomness gathering functions for Windows.
6810 [Jeffrey Altman <jaltman@columbia.edu>]
6811
6812 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6813 not set then we don't setup the error code for issuer check errors
6814 to avoid possibly overwriting other errors which the callback does
6815 handle. If an application does set the flag then we assume it knows
6816 what it is doing and can handle the new informational codes
6817 appropriately.
6818 [Steve Henson]
6819
6820 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6821 a general "ANY" type, as such it should be able to decode anything
6822 including tagged types. However it didn't check the class so it would
6823 wrongly interpret tagged types in the same way as their universal
6824 counterpart and unknown types were just rejected. Changed so that the
6825 tagged and unknown types are handled in the same way as a SEQUENCE:
6826 that is the encoding is stored intact. There is also a new type
6827 "V_ASN1_OTHER" which is used when the class is not universal, in this
6828 case we have no idea what the actual type is so we just lump them all
6829 together.
6830 [Steve Henson]
6831
6832 *) On VMS, stdout may very well lead to a file that is written to
6833 in a record-oriented fashion. That means that every write() will
6834 write a separate record, which will be read separately by the
6835 programs trying to read from it. This can be very confusing.
6836
6837 The solution is to put a BIO filter in the way that will buffer
6838 text until a linefeed is reached, and then write everything a
6839 line at a time, so every record written will be an actual line,
6840 not chunks of lines and not (usually doesn't happen, but I've
6841 seen it once) several lines in one record. BIO_f_linebuffer() is
6842 the answer.
6843
6844 Currently, it's a VMS-only method, because that's where it has
6845 been tested well enough.
6846 [Richard Levitte]
6847
6848 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6849 it can return incorrect results.
6850 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6851 but it was in 0.9.6-beta[12].)
6852 [Bodo Moeller]
6853
6854 *) Disable the check for content being present when verifying detached
6855 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6856 include zero length content when signing messages.
6857 [Steve Henson]
6858
6859 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6860 BIO_ctrl (for BIO pairs).
6861 [Bodo Möller]
6862
6863 *) Add DSO method for VMS.
6864 [Richard Levitte]
6865
6866 *) Bug fix: Montgomery multiplication could produce results with the
6867 wrong sign.
6868 [Ulf Möller]
6869
6870 *) Add RPM specification openssl.spec and modify it to build three
6871 packages. The default package contains applications, application
6872 documentation and run-time libraries. The devel package contains
6873 include files, static libraries and function documentation. The
6874 doc package contains the contents of the doc directory. The original
6875 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6876 [Richard Levitte]
6877
6878 *) Add a large number of documentation files for many SSL routines.
6879 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6880
6881 *) Add a configuration entry for Sony News 4.
6882 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6883
6884 *) Don't set the two most significant bits to one when generating a
6885 random number < q in the DSA library.
6886 [Ulf Möller]
6887
6888 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6889 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6890 the underlying transport is blocking) if a handshake took place.
6891 (The default behaviour is needed by applications such as s_client
6892 and s_server that use select() to determine when to use SSL_read;
6893 but for applications that know in advance when to expect data, it
6894 just makes things more complicated.)
6895 [Bodo Moeller]
6896
6897 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6898 from EGD.
6899 [Ben Laurie]
6900
6901 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6902 work better on such systems.
6903 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6904
6905 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6906 Update PKCS12_parse() so it copies the friendlyName and the
6907 keyid to the certificates aux info.
6908 [Steve Henson]
6909
6910 *) Fix bug in PKCS7_verify() which caused an infinite loop
6911 if there was more than one signature.
6912 [Sven Uszpelkat <su@celocom.de>]
6913
6914 *) Major change in util/mkdef.pl to include extra information
6915 about each symbol, as well as presentig variables as well
6916 as functions. This change means that there's n more need
6917 to rebuild the .num files when some algorithms are excluded.
6918 [Richard Levitte]
6919
6920 *) Allow the verify time to be set by an application,
6921 rather than always using the current time.
6922 [Steve Henson]
6923
6924 *) Phase 2 verify code reorganisation. The certificate
6925 verify code now looks up an issuer certificate by a
6926 number of criteria: subject name, authority key id
6927 and key usage. It also verifies self signed certificates
6928 by the same criteria. The main comparison function is
6929 X509_check_issued() which performs these checks.
6930
6931 Lot of changes were necessary in order to support this
6932 without completely rewriting the lookup code.
6933
6934 Authority and subject key identifier are now cached.
6935
6936 The LHASH 'certs' is X509_STORE has now been replaced
6937 by a STACK_OF(X509_OBJECT). This is mainly because an
6938 LHASH can't store or retrieve multiple objects with
6939 the same hash value.
6940
6941 As a result various functions (which were all internal
6942 use only) have changed to handle the new X509_STORE
6943 structure. This will break anything that messed round
6944 with X509_STORE internally.
6945
6946 The functions X509_STORE_add_cert() now checks for an
6947 exact match, rather than just subject name.
6948
6949 The X509_STORE API doesn't directly support the retrieval
6950 of multiple certificates matching a given criteria, however
6951 this can be worked round by performing a lookup first
6952 (which will fill the cache with candidate certificates)
6953 and then examining the cache for matches. This is probably
6954 the best we can do without throwing out X509_LOOKUP
6955 entirely (maybe later...).
6956
6957 The X509_VERIFY_CTX structure has been enhanced considerably.
6958
6959 All certificate lookup operations now go via a get_issuer()
6960 callback. Although this currently uses an X509_STORE it
6961 can be replaced by custom lookups. This is a simple way
6962 to bypass the X509_STORE hackery necessary to make this
6963 work and makes it possible to use more efficient techniques
6964 in future. A very simple version which uses a simple
6965 STACK for its trusted certificate store is also provided
6966 using X509_STORE_CTX_trusted_stack().
6967
6968 The verify_cb() and verify() callbacks now have equivalents
6969 in the X509_STORE_CTX structure.
6970
6971 X509_STORE_CTX also has a 'flags' field which can be used
6972 to customise the verify behaviour.
6973 [Steve Henson]
6974
6975 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6976 excludes S/MIME capabilities.
6977 [Steve Henson]
6978
6979 *) When a certificate request is read in keep a copy of the
6980 original encoding of the signed data and use it when outputing
6981 again. Signatures then use the original encoding rather than
6982 a decoded, encoded version which may cause problems if the
6983 request is improperly encoded.
6984 [Steve Henson]
6985
6986 *) For consistency with other BIO_puts implementations, call
6987 buffer_write(b, ...) directly in buffer_puts instead of calling
6988 BIO_write(b, ...).
6989
6990 In BIO_puts, increment b->num_write as in BIO_write.
6991 [Peter.Sylvester@EdelWeb.fr]
6992
6993 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6994 BN_zero, we may not return a BIGNUM with an array consisting of
6995 words set to zero.)
6996 [Bodo Moeller]
6997
6998 *) Avoid calling abort() from within the library when problems are
6999 detected, except if preprocessor symbols have been defined
7000 (such as REF_CHECK, BN_DEBUG etc.).
7001 [Bodo Moeller]
7002
7003 *) New openssl application 'rsautl'. This utility can be
7004 used for low level RSA operations. DER public key
7005 BIO/fp routines also added.
7006 [Steve Henson]
7007
7008 *) New Configure entry and patches for compiling on QNX 4.
7009 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7010
7011 *) A demo state-machine implementation was sponsored by
7012 Nuron (http://www.nuron.com/) and is now available in
7013 demos/state_machine.
7014 [Ben Laurie]
7015
7016 *) New options added to the 'dgst' utility for signature
7017 generation and verification.
7018 [Steve Henson]
7019
7020 *) Unrecognized PKCS#7 content types are now handled via a
7021 catch all ASN1_TYPE structure. This allows unsupported
7022 types to be stored as a "blob" and an application can
7023 encode and decode it manually.
7024 [Steve Henson]
7025
7026 *) Fix various signed/unsigned issues to make a_strex.c
7027 compile under VC++.
7028 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7029
7030 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7031 length if passed a buffer. ASN1_INTEGER_to_BN failed
7032 if passed a NULL BN and its argument was negative.
7033 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7034
7035 *) Modification to PKCS#7 encoding routines to output definite
7036 length encoding. Since currently the whole structures are in
7037 memory there's not real point in using indefinite length
7038 constructed encoding. However if OpenSSL is compiled with
7039 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7040 [Steve Henson]
7041
7042 *) Added BIO_vprintf() and BIO_vsnprintf().
7043 [Richard Levitte]
7044
7045 *) Added more prefixes to parse for in the the strings written
7046 through a logging bio, to cover all the levels that are available
7047 through syslog. The prefixes are now:
7048
7049 PANIC, EMERG, EMR => LOG_EMERG
7050 ALERT, ALR => LOG_ALERT
7051 CRIT, CRI => LOG_CRIT
7052 ERROR, ERR => LOG_ERR
7053 WARNING, WARN, WAR => LOG_WARNING
7054 NOTICE, NOTE, NOT => LOG_NOTICE
7055 INFO, INF => LOG_INFO
7056 DEBUG, DBG => LOG_DEBUG
7057
7058 and as before, if none of those prefixes are present at the
7059 beginning of the string, LOG_ERR is chosen.
7060
7061 On Win32, the LOG_* levels are mapped according to this:
7062
7063 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7064 LOG_WARNING => EVENTLOG_WARNING_TYPE
7065 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7066
7067 [Richard Levitte]
7068
7069 *) Made it possible to reconfigure with just the configuration
7070 argument "reconf" or "reconfigure". The command line arguments
7071 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7072 and are retrieved from there when reconfiguring.
7073 [Richard Levitte]
7074
7075 *) MD4 implemented.
7076 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7077
7078 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7079 [Richard Levitte]
7080
7081 *) The obj_dat.pl script was messing up the sorting of object
7082 names. The reason was that it compared the quoted version
7083 of strings as a result "OCSP" > "OCSP Signing" because
7084 " > SPACE. Changed script to store unquoted versions of
7085 names and add quotes on output. It was also omitting some
7086 names from the lookup table if they were given a default
7087 value (that is if SN is missing it is given the same
7088 value as LN and vice versa), these are now added on the
7089 grounds that if an object has a name we should be able to
7090 look it up. Finally added warning output when duplicate
7091 short or long names are found.
7092 [Steve Henson]
7093
7094 *) Changes needed for Tandem NSK.
7095 [Scott Uroff <scott@xypro.com>]
7096
7097 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7098 RSA_padding_check_SSLv23(), special padding was never detected
7099 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7100 version rollback attacks was not effective.
7101
7102 In s23_clnt.c, don't use special rollback-attack detection padding
7103 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7104 client; similarly, in s23_srvr.c, don't do the rollback check if
7105 SSL 2.0 is the only protocol enabled in the server.
7106 [Bodo Moeller]
7107
7108 *) Make it possible to get hexdumps of unprintable data with 'openssl
7109 asn1parse'. By implication, the functions ASN1_parse_dump() and
7110 BIO_dump_indent() are added.
7111 [Richard Levitte]
7112
7113 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7114 these print out strings and name structures based on various
7115 flags including RFC2253 support and proper handling of
7116 multibyte characters. Added options to the 'x509' utility
7117 to allow the various flags to be set.
7118 [Steve Henson]
7119
7120 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7121 Also change the functions X509_cmp_current_time() and
7122 X509_gmtime_adj() work with an ASN1_TIME structure,
7123 this will enable certificates using GeneralizedTime in validity
7124 dates to be checked.
7125 [Steve Henson]
7126
7127 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7128 negative public key encodings) on by default,
7129 NO_NEG_PUBKEY_BUG can be set to disable it.
7130 [Steve Henson]
7131
7132 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7133 content octets. An i2c_ASN1_OBJECT is unnecessary because
7134 the encoding can be trivially obtained from the structure.
7135 [Steve Henson]
7136
7137 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7138 not read locks (CRYPTO_r_[un]lock).
7139 [Bodo Moeller]
7140
7141 *) A first attempt at creating official support for shared
7142 libraries through configuration. I've kept it so the
7143 default is static libraries only, and the OpenSSL programs
7144 are always statically linked for now, but there are
7145 preparations for dynamic linking in place.
7146 This has been tested on Linux and Tru64.
7147 [Richard Levitte]
7148
7149 *) Randomness polling function for Win9x, as described in:
7150 Peter Gutmann, Software Generation of Practically Strong
7151 Random Numbers.
7152 [Ulf Möller]
7153
7154 *) Fix so PRNG is seeded in req if using an already existing
7155 DSA key.
7156 [Steve Henson]
7157
7158 *) New options to smime application. -inform and -outform
7159 allow alternative formats for the S/MIME message including
7160 PEM and DER. The -content option allows the content to be
7161 specified separately. This should allow things like Netscape
7162 form signing output easier to verify.
7163 [Steve Henson]
7164
7165 *) Fix the ASN1 encoding of tags using the 'long form'.
7166 [Steve Henson]
7167
7168 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7169 STRING types. These convert content octets to and from the
7170 underlying type. The actual tag and length octets are
7171 already assumed to have been read in and checked. These
7172 are needed because all other string types have virtually
7173 identical handling apart from the tag. By having versions
7174 of the ASN1 functions that just operate on content octets
7175 IMPLICIT tagging can be handled properly. It also allows
7176 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7177 and ASN1_INTEGER are identical apart from the tag.
7178 [Steve Henson]
7179
7180 *) Change the handling of OID objects as follows:
7181
7182 - New object identifiers are inserted in objects.txt, following
7183 the syntax given in objects.README.
7184 - objects.pl is used to process obj_mac.num and create a new
7185 obj_mac.h.
7186 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7187 obj_mac.h.
7188
7189 This is currently kind of a hack, and the perl code in objects.pl
7190 isn't very elegant, but it works as I intended. The simplest way
7191 to check that it worked correctly is to look in obj_dat.h and
7192 check the array nid_objs and make sure the objects haven't moved
7193 around (this is important!). Additions are OK, as well as
7194 consistent name changes.
7195 [Richard Levitte]
7196
7197 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7198 [Bodo Moeller]
7199
7200 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7201 The given file adds to whatever has already been seeded into the
7202 random pool through the RANDFILE configuration file option or
7203 environment variable, or the default random state file.
7204 [Richard Levitte]
7205
7206 *) mkstack.pl now sorts each macro group into lexical order.
7207 Previously the output order depended on the order the files
7208 appeared in the directory, resulting in needless rewriting
7209 of safestack.h .
7210 [Steve Henson]
7211
7212 *) Patches to make OpenSSL compile under Win32 again. Mostly
7213 work arounds for the VC++ problem that it treats func() as
7214 func(void). Also stripped out the parts of mkdef.pl that
7215 added extra typesafe functions: these no longer exist.
7216 [Steve Henson]
7217
7218 *) Reorganisation of the stack code. The macros are now all
7219 collected in safestack.h . Each macro is defined in terms of
7220 a "stack macro" of the form SKM_<name>(type, a, b). The
7221 DEBUG_SAFESTACK is now handled in terms of function casts,
7222 this has the advantage of retaining type safety without the
7223 use of additional functions. If DEBUG_SAFESTACK is not defined
7224 then the non typesafe macros are used instead. Also modified the
7225 mkstack.pl script to handle the new form. Needs testing to see
7226 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7227 the default if no major problems. Similar behaviour for ASN1_SET_OF
7228 and PKCS12_STACK_OF.
7229 [Steve Henson]
7230
7231 *) When some versions of IIS use the 'NET' form of private key the
7232 key derivation algorithm is different. Normally MD5(password) is
7233 used as a 128 bit RC4 key. In the modified case
7234 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
7235 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7236 as the old Netscape_RSA functions except they have an additional
7237 'sgckey' parameter which uses the modified algorithm. Also added
7238 an -sgckey command line option to the rsa utility. Thanks to
7239 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7240 algorithm to openssl-dev.
7241 [Steve Henson]
7242
7243 *) The evp_local.h macros were using 'c.##kname' which resulted in
7244 invalid expansion on some systems (SCO 5.0.5 for example).
7245 Corrected to 'c.kname'.
7246 [Phillip Porch <root@theporch.com>]
7247
7248 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7249 a STACK of email addresses from a certificate or request, these look
7250 in the subject name and the subject alternative name extensions and
7251 omit any duplicate addresses.
7252 [Steve Henson]
7253
7254 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7255 This makes DSA verification about 2 % faster.
7256 [Bodo Moeller]
7257
7258 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7259 (meaning that now 2^5 values will be precomputed, which is only 4 KB
7260 plus overhead for 1024 bit moduli).
7261 This makes exponentiations about 0.5 % faster for 1024 bit
7262 exponents (as measured by "openssl speed rsa2048").
7263 [Bodo Moeller]
7264
7265 *) Rename memory handling macros to avoid conflicts with other
7266 software:
7267 Malloc => OPENSSL_malloc
7268 Malloc_locked => OPENSSL_malloc_locked
7269 Realloc => OPENSSL_realloc
7270 Free => OPENSSL_free
7271 [Richard Levitte]
7272
7273 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
7274 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
7275 [Bodo Moeller]
7276
7277 *) CygWin32 support.
7278 [John Jarvie <jjarvie@newsguy.com>]
7279
7280 *) The type-safe stack code has been rejigged. It is now only compiled
7281 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
7282 by default all type-specific stack functions are "#define"d back to
7283 standard stack functions. This results in more streamlined output
7284 but retains the type-safety checking possibilities of the original
7285 approach.
7286 [Geoff Thorpe]
7287
7288 *) The STACK code has been cleaned up, and certain type declarations
7289 that didn't make a lot of sense have been brought in line. This has
7290 also involved a cleanup of sorts in safestack.h to more correctly
7291 map type-safe stack functions onto their plain stack counterparts.
7292 This work has also resulted in a variety of "const"ifications of
7293 lots of the code, especially "_cmp" operations which should normally
7294 be prototyped with "const" parameters anyway.
7295 [Geoff Thorpe]
7296
7297 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
7298 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
7299 (The PRNG state consists of two parts, the large pool 'state' and 'md',
7300 where all of 'md' is used each time the PRNG is used, but 'state'
7301 is used only indexed by a cyclic counter. As entropy may not be
7302 well distributed from the beginning, 'md' is important as a
7303 chaining variable. However, the output function chains only half
7304 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
7305 all of 'md', and seeding with STATE_SIZE dummy bytes will result
7306 in all of 'state' being rewritten, with the new values depending
7307 on virtually all of 'md'. This overcomes the 80 bit limitation.)
7308 [Bodo Moeller]
7309
7310 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
7311 the handshake is continued after ssl_verify_cert_chain();
7312 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
7313 can lead to 'unexplainable' connection aborts later.
7314 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
7315
7316 *) Major EVP API cipher revision.
7317 Add hooks for extra EVP features. This allows various cipher
7318 parameters to be set in the EVP interface. Support added for variable
7319 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
7320 setting of RC2 and RC5 parameters.
7321
7322 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
7323 ciphers.
7324
7325 Remove lots of duplicated code from the EVP library. For example *every*
7326 cipher init() function handles the 'iv' in the same way according to the
7327 cipher mode. They also all do nothing if the 'key' parameter is NULL and
7328 for CFB and OFB modes they zero ctx->num.
7329
7330 New functionality allows removal of S/MIME code RC2 hack.
7331
7332 Most of the routines have the same form and so can be declared in terms
7333 of macros.
7334
7335 By shifting this to the top level EVP_CipherInit() it can be removed from
7336 all individual ciphers. If the cipher wants to handle IVs or keys
7337 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
7338 flags.
7339
7340 Change lots of functions like EVP_EncryptUpdate() to now return a
7341 value: although software versions of the algorithms cannot fail
7342 any installed hardware versions can.
7343 [Steve Henson]
7344
7345 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
7346 this option is set, tolerate broken clients that send the negotiated
7347 protocol version number instead of the requested protocol version
7348 number.
7349 [Bodo Moeller]
7350
7351 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
7352 i.e. non-zero for export ciphersuites, zero otherwise.
7353 Previous versions had this flag inverted, inconsistent with
7354 rsa_tmp_cb (..._TMP_RSA_CB).
7355 [Bodo Moeller; problem reported by Amit Chopra]
7356
7357 *) Add missing DSA library text string. Work around for some IIS
7358 key files with invalid SEQUENCE encoding.
7359 [Steve Henson]
7360
7361 *) Add a document (doc/standards.txt) that list all kinds of standards
7362 and so on that are implemented in OpenSSL.
7363 [Richard Levitte]
7364
7365 *) Enhance c_rehash script. Old version would mishandle certificates
7366 with the same subject name hash and wouldn't handle CRLs at all.
7367 Added -fingerprint option to crl utility, to support new c_rehash
7368 features.
7369 [Steve Henson]
7370
7371 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
7372 [Ulf Möller]
7373
7374 *) Fix for SSL server purpose checking. Server checking was
7375 rejecting certificates which had extended key usage present
7376 but no ssl client purpose.
7377 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
7378
7379 *) Make PKCS#12 code work with no password. The PKCS#12 spec
7380 is a little unclear about how a blank password is handled.
7381 Since the password in encoded as a BMPString with terminating
7382 double NULL a zero length password would end up as just the
7383 double NULL. However no password at all is different and is
7384 handled differently in the PKCS#12 key generation code. NS
7385 treats a blank password as zero length. MSIE treats it as no
7386 password on export: but it will try both on import. We now do
7387 the same: PKCS12_parse() tries zero length and no password if
7388 the password is set to "" or NULL (NULL is now a valid password:
7389 it wasn't before) as does the pkcs12 application.
7390 [Steve Henson]
7391
7392 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
7393 perror when PEM_read_bio_X509_REQ fails, the error message must
7394 be obtained from the error queue.
7395 [Bodo Moeller]
7396
7397 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
7398 it in ERR_remove_state if appropriate, and change ERR_get_state
7399 accordingly to avoid race conditions (this is necessary because
7400 thread_hash is no longer constant once set).
7401 [Bodo Moeller]
7402
7403 *) Bugfix for linux-elf makefile.one.
7404 [Ulf Möller]
7405
7406 *) RSA_get_default_method() will now cause a default
7407 RSA_METHOD to be chosen if one doesn't exist already.
7408 Previously this was only set during a call to RSA_new()
7409 or RSA_new_method(NULL) meaning it was possible for
7410 RSA_get_default_method() to return NULL.
7411 [Geoff Thorpe]
7412
7413 *) Added native name translation to the existing DSO code
7414 that will convert (if the flag to do so is set) filenames
7415 that are sufficiently small and have no path information
7416 into a canonical native form. Eg. "blah" converted to
7417 "libblah.so" or "blah.dll" etc.
7418 [Geoff Thorpe]
7419
7420 *) New function ERR_error_string_n(e, buf, len) which is like
7421 ERR_error_string(e, buf), but writes at most 'len' bytes
7422 including the 0 terminator. For ERR_error_string_n, 'buf'
7423 may not be NULL.
7424 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7425
7426 *) CONF library reworked to become more general. A new CONF
7427 configuration file reader "class" is implemented as well as a
7428 new functions (NCONF_*, for "New CONF") to handle it. The now
7429 old CONF_* functions are still there, but are reimplemented to
7430 work in terms of the new functions. Also, a set of functions
7431 to handle the internal storage of the configuration data is
7432 provided to make it easier to write new configuration file
7433 reader "classes" (I can definitely see something reading a
7434 configuration file in XML format, for example), called _CONF_*,
7435 or "the configuration storage API"...
7436
7437 The new configuration file reading functions are:
7438
7439 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7440 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7441
7442 NCONF_default, NCONF_WIN32
7443
7444 NCONF_dump_fp, NCONF_dump_bio
7445
7446 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7447 NCONF_new creates a new CONF object. This works in the same way
7448 as other interfaces in OpenSSL, like the BIO interface.
7449 NCONF_dump_* dump the internal storage of the configuration file,
7450 which is useful for debugging. All other functions take the same
7451 arguments as the old CONF_* functions wth the exception of the
7452 first that must be a `CONF *' instead of a `LHASH *'.
7453
7454 To make it easer to use the new classes with the old CONF_* functions,
7455 the function CONF_set_default_method is provided.
7456 [Richard Levitte]
7457
7458 *) Add '-tls1' option to 'openssl ciphers', which was already
7459 mentioned in the documentation but had not been implemented.
7460 (This option is not yet really useful because even the additional
7461 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7462 [Bodo Moeller]
7463
7464 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7465 OpenSSL-based applications) load shared libraries and bind to
7466 them in a portable way.
7467 [Geoff Thorpe, with contributions from Richard Levitte]
7468
7469 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7470
7471 *) Make sure _lrotl and _lrotr are only used with MSVC.
7472
7473 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7474 (the default implementation of RAND_status).
7475
7476 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7477 to '-clrext' (= clear extensions), as intended and documented.
7478 [Bodo Moeller; inconsistency pointed out by Michael Attili
7479 <attili@amaxo.com>]
7480
7481 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7482 was larger than the MD block size.
7483 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7484
7485 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7486 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7487 using the passed key: if the passed key was a private key the result
7488 of X509_print(), for example, would be to print out all the private key
7489 components.
7490 [Steve Henson]
7491
7492 *) des_quad_cksum() byte order bug fix.
7493 [Ulf Möller, using the problem description in krb4-0.9.7, where
7494 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7495
7496 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7497 discouraged.
7498 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7499
7500 *) For easily testing in shell scripts whether some command
7501 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7502 returns with exit code 0 iff no command of the given name is available.
7503 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7504 the output goes to stdout and nothing is printed to stderr.
7505 Additional arguments are always ignored.
7506
7507 Since for each cipher there is a command of the same name,
7508 the 'no-cipher' compilation switches can be tested this way.
7509
7510 ('openssl no-XXX' is not able to detect pseudo-commands such
7511 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7512 [Bodo Moeller]
7513
7514 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7515 [Bodo Moeller]
7516
7517 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7518 is set; it will be thrown away anyway because each handshake creates
7519 its own key.
7520 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7521 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7522 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7523 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7524 [Bodo Moeller]
7525
7526 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7527 'Q' and 'R' lose their special meanings (quit/renegotiate).
7528 This is part of what -quiet does; unlike -quiet, -ign_eof
7529 does not suppress any output.
7530 [Richard Levitte]
7531
7532 *) Add compatibility options to the purpose and trust code. The
7533 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7534 accepts a certificate or CA, this was the previous behaviour,
7535 with all the associated security issues.
7536
7537 X509_TRUST_COMPAT is the old trust behaviour: only and
7538 automatically trust self signed roots in certificate store. A
7539 new trust setting X509_TRUST_DEFAULT is used to specify that
7540 a purpose has no associated trust setting and it should instead
7541 use the value in the default purpose.
7542 [Steve Henson]
7543
7544 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7545 and fix a memory leak.
7546 [Steve Henson]
7547
7548 *) In util/mkerr.pl (which implements 'make errors'), preserve
7549 reason strings from the previous version of the .c file, as
7550 the default to have only downcase letters (and digits) in
7551 automatically generated reasons codes is not always appropriate.
7552 [Bodo Moeller]
7553
7554 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7555 using strerror. Previously, ERR_reason_error_string() returned
7556 library names as reason strings for SYSerr; but SYSerr is a special
7557 case where small numbers are errno values, not library numbers.
7558 [Bodo Moeller]
7559
7560 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7561 converts DSA parameters into DH parameters. (When creating parameters,
7562 DSA_generate_parameters is used.)
7563 [Bodo Moeller]
7564
7565 *) Include 'length' (recommended exponent length) in C code generated
7566 by 'openssl dhparam -C'.
7567 [Bodo Moeller]
7568
7569 *) The second argument to set_label in perlasm was already being used
7570 so couldn't be used as a "file scope" flag. Moved to third argument
7571 which was free.
7572 [Steve Henson]
7573
7574 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7575 instead of RAND_bytes for encryption IVs and salts.
7576 [Bodo Moeller]
7577
7578 *) Include RAND_status() into RAND_METHOD instead of implementing
7579 it only for md_rand.c Otherwise replacing the PRNG by calling
7580 RAND_set_rand_method would be impossible.
7581 [Bodo Moeller]
7582
7583 *) Don't let DSA_generate_key() enter an infinite loop if the random
7584 number generation fails.
7585 [Bodo Moeller]
7586
7587 *) New 'rand' application for creating pseudo-random output.
7588 [Bodo Moeller]
7589
7590 *) Added configuration support for Linux/IA64
7591 [Rolf Haberrecker <rolf@suse.de>]
7592
7593 *) Assembler module support for Mingw32.
7594 [Ulf Möller]
7595
7596 *) Shared library support for HPUX (in shlib/).
7597 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7598
7599 *) Shared library support for Solaris gcc.
7600 [Lutz Behnke <behnke@trustcenter.de>]
7601
7602 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7603
7604 *) PKCS7_encrypt() was adding text MIME headers twice because they
7605 were added manually and by SMIME_crlf_copy().
7606 [Steve Henson]
7607
7608 *) In bntest.c don't call BN_rand with zero bits argument.
7609 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7610
7611 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7612 case was implemented. This caused BN_div_recp() to fail occasionally.
7613 [Ulf Möller]
7614
7615 *) Add an optional second argument to the set_label() in the perl
7616 assembly language builder. If this argument exists and is set
7617 to 1 it signals that the assembler should use a symbol whose
7618 scope is the entire file, not just the current function. This
7619 is needed with MASM which uses the format label:: for this scope.
7620 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7621
7622 *) Change the ASN1 types so they are typedefs by default. Before
7623 almost all types were #define'd to ASN1_STRING which was causing
7624 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7625 for example.
7626 [Steve Henson]
7627
7628 *) Change names of new functions to the new get1/get0 naming
7629 convention: After 'get1', the caller owns a reference count
7630 and has to call ..._free; 'get0' returns a pointer to some
7631 data structure without incrementing reference counters.
7632 (Some of the existing 'get' functions increment a reference
7633 counter, some don't.)
7634 Similarly, 'set1' and 'add1' functions increase reference
7635 counters or duplicate objects.
7636 [Steve Henson]
7637
7638 *) Allow for the possibility of temp RSA key generation failure:
7639 the code used to assume it always worked and crashed on failure.
7640 [Steve Henson]
7641
7642 *) Fix potential buffer overrun problem in BIO_printf().
7643 [Ulf Möller, using public domain code by Patrick Powell; problem
7644 pointed out by David Sacerdote <das33@cornell.edu>]
7645
7646 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7647 RAND_egd() and RAND_status(). In the command line application,
7648 the EGD socket can be specified like a seed file using RANDFILE
7649 or -rand.
7650 [Ulf Möller]
7651
7652 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7653 Some CAs (e.g. Verisign) distribute certificates in this form.
7654 [Steve Henson]
7655
7656 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7657 list to exclude them. This means that no special compilation option
7658 is needed to use anonymous DH: it just needs to be included in the
7659 cipher list.
7660 [Steve Henson]
7661
7662 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7663 EVP_MD_type. The old functionality is available in a new macro called
7664 EVP_MD_md(). Change code that uses it and update docs.
7665 [Steve Henson]
7666
7667 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7668 where the 'void *' argument is replaced by a function pointer argument.
7669 Previously 'void *' was abused to point to functions, which works on
7670 many platforms, but is not correct. As these functions are usually
7671 called by macros defined in OpenSSL header files, most source code
7672 should work without changes.
7673 [Richard Levitte]
7674
7675 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7676 sections with information on -D... compiler switches used for
7677 compiling the library so that applications can see them. To enable
7678 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7679 must be defined. E.g.,
7680 #define OPENSSL_ALGORITHM_DEFINES
7681 #include <openssl/opensslconf.h>
7682 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7683 [Richard Levitte, Ulf and Bodo Möller]
7684
7685 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7686 record layer.
7687 [Bodo Moeller]
7688
7689 *) Change the 'other' type in certificate aux info to a STACK_OF
7690 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7691 the required ASN1 format: arbitrary types determined by an OID.
7692 [Steve Henson]
7693
7694 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7695 argument to 'req'. This is not because the function is newer or
7696 better than others it just uses the work 'NEW' in the certificate
7697 request header lines. Some software needs this.
7698 [Steve Henson]
7699
7700 *) Reorganise password command line arguments: now passwords can be
7701 obtained from various sources. Delete the PEM_cb function and make
7702 it the default behaviour: i.e. if the callback is NULL and the
7703 usrdata argument is not NULL interpret it as a null terminated pass
7704 phrase. If usrdata and the callback are NULL then the pass phrase
7705 is prompted for as usual.
7706 [Steve Henson]
7707
7708 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7709 the support is automatically enabled. The resulting binaries will
7710 autodetect the card and use it if present.
7711 [Ben Laurie and Compaq Inc.]
7712
7713 *) Work around for Netscape hang bug. This sends certificate request
7714 and server done in one record. Since this is perfectly legal in the
7715 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7716 the bugs/SSLv3 entry for more info.
7717 [Steve Henson]
7718
7719 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7720 [Andy Polyakov]
7721
7722 *) Add -rand argument to smime and pkcs12 applications and read/write
7723 of seed file.
7724 [Steve Henson]
7725
7726 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7727 [Bodo Moeller]
7728
7729 *) Add command line password options to the remaining applications.
7730 [Steve Henson]
7731
7732 *) Bug fix for BN_div_recp() for numerators with an even number of
7733 bits.
7734 [Ulf Möller]
7735
7736 *) More tests in bntest.c, and changed test_bn output.
7737 [Ulf Möller]
7738
7739 *) ./config recognizes MacOS X now.
7740 [Andy Polyakov]
7741
7742 *) Bug fix for BN_div() when the first words of num and divsor are
7743 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7744 [Ulf Möller]
7745
7746 *) Add support for various broken PKCS#8 formats, and command line
7747 options to produce them.
7748 [Steve Henson]
7749
7750 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7751 get temporary BIGNUMs from a BN_CTX.
7752 [Ulf Möller]
7753
7754 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7755 for p == 0.
7756 [Ulf Möller]
7757
7758 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7759 include a #define from the old name to the new. The original intent
7760 was that statically linked binaries could for example just call
7761 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7762 link with digests. This never worked becayse SSLeay_add_all_digests()
7763 and SSLeay_add_all_ciphers() were in the same source file so calling
7764 one would link with the other. They are now in separate source files.
7765 [Steve Henson]
7766
7767 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7768 [Steve Henson]
7769
7770 *) Use a less unusual form of the Miller-Rabin primality test (it used
7771 a binary algorithm for exponentiation integrated into the Miller-Rabin
7772 loop, our standard modexp algorithms are faster).
7773 [Bodo Moeller]
7774
7775 *) Support for the EBCDIC character set completed.
7776 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7777
7778 *) Source code cleanups: use const where appropriate, eliminate casts,
7779 use void * instead of char * in lhash.
7780 [Ulf Möller]
7781
7782 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7783 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7784 this the server could overwrite ephemeral keys that the client
7785 has already seen).
7786 [Bodo Moeller]
7787
7788 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7789 using 50 iterations of the Rabin-Miller test.
7790
7791 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7792 iterations of the Rabin-Miller test as required by the appendix
7793 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7794 As BN_is_prime_fasttest includes trial division, DSA parameter
7795 generation becomes much faster.
7796
7797 This implies a change for the callback functions in DSA_is_prime
7798 and DSA_generate_parameters: The callback function is called once
7799 for each positive witness in the Rabin-Miller test, not just
7800 occasionally in the inner loop; and the parameters to the
7801 callback function now provide an iteration count for the outer
7802 loop rather than for the current invocation of the inner loop.
7803 DSA_generate_parameters additionally can call the callback
7804 function with an 'iteration count' of -1, meaning that a
7805 candidate has passed the trial division test (when q is generated
7806 from an application-provided seed, trial division is skipped).
7807 [Bodo Moeller]
7808
7809 *) New function BN_is_prime_fasttest that optionally does trial
7810 division before starting the Rabin-Miller test and has
7811 an additional BN_CTX * argument (whereas BN_is_prime always
7812 has to allocate at least one BN_CTX).
7813 'callback(1, -1, cb_arg)' is called when a number has passed the
7814 trial division stage.
7815 [Bodo Moeller]
7816
7817 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7818 as ASN1_TIME.
7819 [Steve Henson]
7820
7821 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7822 [Steve Henson]
7823
7824 *) New function BN_pseudo_rand().
7825 [Ulf Möller]
7826
7827 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7828 bignum version of BN_from_montgomery() with the working code from
7829 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7830 the comments.
7831 [Ulf Möller]
7832
7833 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7834 made it impossible to use the same SSL_SESSION data structure in
7835 SSL2 clients in multiple threads.
7836 [Bodo Moeller]
7837
7838 *) The return value of RAND_load_file() no longer counts bytes obtained
7839 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7840 to seed the PRNG (previously an explicit byte count was required).
7841 [Ulf Möller, Bodo Möller]
7842
7843 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7844 used (char *) instead of (void *) and had casts all over the place.
7845 [Steve Henson]
7846
7847 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7848 [Ulf Möller]
7849
7850 *) Retain source code compatibility for BN_prime_checks macro:
7851 BN_is_prime(..., BN_prime_checks, ...) now uses
7852 BN_prime_checks_for_size to determine the appropriate number of
7853 Rabin-Miller iterations.
7854 [Ulf Möller]
7855
7856 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7857 DH_CHECK_P_NOT_SAFE_PRIME.
7858 (Check if this is true? OpenPGP calls them "strong".)
7859 [Ulf Möller]
7860
7861 *) Merge the functionality of "dh" and "gendh" programs into a new program
7862 "dhparam". The old programs are retained for now but will handle DH keys
7863 (instead of parameters) in future.
7864 [Steve Henson]
7865
7866 *) Make the ciphers, s_server and s_client programs check the return values
7867 when a new cipher list is set.
7868 [Steve Henson]
7869
7870 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7871 ciphers. Before when the 56bit ciphers were enabled the sorting was
7872 wrong.
7873
7874 The syntax for the cipher sorting has been extended to support sorting by
7875 cipher-strength (using the strength_bits hard coded in the tables).
7876 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7877
7878 Fix a bug in the cipher-command parser: when supplying a cipher command
7879 string with an "undefined" symbol (neither command nor alphanumeric
7880 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7881 an error is flagged.
7882
7883 Due to the strength-sorting extension, the code of the
7884 ssl_create_cipher_list() function was completely rearranged. I hope that
7885 the readability was also increased :-)
7886 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7887
7888 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7889 for the first serial number and places 2 in the serial number file. This
7890 avoids problems when the root CA is created with serial number zero and
7891 the first user certificate has the same issuer name and serial number
7892 as the root CA.
7893 [Steve Henson]
7894
7895 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7896 the new code. Add documentation for this stuff.
7897 [Steve Henson]
7898
7899 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7900 X509_*() to X509at_*() on the grounds that they don't handle X509
7901 structures and behave in an analagous way to the X509v3 functions:
7902 they shouldn't be called directly but wrapper functions should be used
7903 instead.
7904
7905 So we also now have some wrapper functions that call the X509at functions
7906 when passed certificate requests. (TO DO: similar things can be done with
7907 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7908 things. Some of these need some d2i or i2d and print functionality
7909 because they handle more complex structures.)
7910 [Steve Henson]
7911
7912 *) Add missing #ifndefs that caused missing symbols when building libssl
7913 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7914 NO_RSA in ssl/s2*.c.
7915 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7916
7917 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7918 has a return value which indicates the quality of the random data
7919 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7920 error queue. New function RAND_pseudo_bytes() generates output that is
7921 guaranteed to be unique but not unpredictable. RAND_add is like
7922 RAND_seed, but takes an extra argument for an entropy estimate
7923 (RAND_seed always assumes full entropy).
7924 [Ulf Möller]
7925
7926 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7927 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7928 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7929 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7930 false-positive rate of at most 2^-80 for random input.
7931 [Bodo Moeller]
7932
7933 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7934 [Bodo Moeller]
7935
7936 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7937 in the 0.9.5 release), this returns the chain
7938 from an X509_CTX structure with a dup of the stack and all
7939 the X509 reference counts upped: so the stack will exist
7940 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7941 to use this.
7942
7943 Also make SSL_SESSION_print() print out the verify return
7944 code.
7945 [Steve Henson]
7946
7947 *) Add manpage for the pkcs12 command. Also change the default
7948 behaviour so MAC iteration counts are used unless the new
7949 -nomaciter option is used. This improves file security and
7950 only older versions of MSIE (4.0 for example) need it.
7951 [Steve Henson]
7952
7953 *) Honor the no-xxx Configure options when creating .DEF files.
7954 [Ulf Möller]
7955
7956 *) Add PKCS#10 attributes to field table: challengePassword,
7957 unstructuredName and unstructuredAddress. These are taken from
7958 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7959 international characters are used.
7960
7961 More changes to X509_ATTRIBUTE code: allow the setting of types
7962 based on strings. Remove the 'loc' parameter when adding
7963 attributes because these will be a SET OF encoding which is sorted
7964 in ASN1 order.
7965 [Steve Henson]
7966
7967 *) Initial changes to the 'req' utility to allow request generation
7968 automation. This will allow an application to just generate a template
7969 file containing all the field values and have req construct the
7970 request.
7971
7972 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7973 used all over the place including certificate requests and PKCS#7
7974 structures. They are currently handled manually where necessary with
7975 some primitive wrappers for PKCS#7. The new functions behave in a
7976 manner analogous to the X509 extension functions: they allow
7977 attributes to be looked up by NID and added.
7978
7979 Later something similar to the X509V3 code would be desirable to
7980 automatically handle the encoding, decoding and printing of the
7981 more complex types. The string types like challengePassword can
7982 be handled by the string table functions.
7983
7984 Also modified the multi byte string table handling. Now there is
7985 a 'global mask' which masks out certain types. The table itself
7986 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7987 is useful when for example there is only one permissible type
7988 (as in countryName) and using the mask might result in no valid
7989 types at all.
7990 [Steve Henson]
7991
7992 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7993 SSL_get_peer_finished to allow applications to obtain the latest
7994 Finished messages sent to the peer or expected from the peer,
7995 respectively. (SSL_get_peer_finished is usually the Finished message
7996 actually received from the peer, otherwise the protocol will be aborted.)
7997
7998 As the Finished message are message digests of the complete handshake
7999 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8000 be used for external authentication procedures when the authentication
8001 provided by SSL/TLS is not desired or is not enough.
8002 [Bodo Moeller]
8003
8004 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8005 the host supports BWX extension and if Compaq C is present on the
8006 $PATH. Just exploiting of the BWX extension results in 20-30%
8007 performance kick for some algorithms, e.g. DES and RC4 to mention
8008 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8009 SHA1.
8010 [Andy Polyakov]
8011
8012 *) Add support for MS "fast SGC". This is arguably a violation of the
8013 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8014 weak crypto and after checking the certificate is SGC a second one
8015 with strong crypto. MS SGC stops the first handshake after receiving
8016 the server certificate message and sends a second client hello. Since
8017 a server will typically do all the time consuming operations before
8018 expecting any further messages from the client (server key exchange
8019 is the most expensive) there is little difference between the two.
8020
8021 To get OpenSSL to support MS SGC we have to permit a second client
8022 hello message after we have sent server done. In addition we have to
8023 reset the MAC if we do get this second client hello.
8024 [Steve Henson]
8025
8026 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8027 if a DER encoded private key is RSA or DSA traditional format. Changed
8028 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8029 format DER encoded private key. Newer code should use PKCS#8 format which
8030 has the key type encoded in the ASN1 structure. Added DER private key
8031 support to pkcs8 application.
8032 [Steve Henson]
8033
8034 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8035 ciphersuites has been selected (as required by the SSL 3/TLS 1
8036 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8037 is set, we interpret this as a request to violate the specification
8038 (the worst that can happen is a handshake failure, and 'correct'
8039 behaviour would result in a handshake failure anyway).
8040 [Bodo Moeller]
8041
8042 *) In SSL_CTX_add_session, take into account that there might be multiple
8043 SSL_SESSION structures with the same session ID (e.g. when two threads
8044 concurrently obtain them from an external cache).
8045 The internal cache can handle only one SSL_SESSION with a given ID,
8046 so if there's a conflict, we now throw out the old one to achieve
8047 consistency.
8048 [Bodo Moeller]
8049
8050 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8051 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8052 some routines that use cipher OIDs: some ciphers do not have OIDs
8053 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8054 example.
8055 [Steve Henson]
8056
8057 *) Simplify the trust setting structure and code. Now we just have
8058 two sequences of OIDs for trusted and rejected settings. These will
8059 typically have values the same as the extended key usage extension
8060 and any application specific purposes.
8061
8062 The trust checking code now has a default behaviour: it will just
8063 check for an object with the same NID as the passed id. Functions can
8064 be provided to override either the default behaviour or the behaviour
8065 for a given id. SSL client, server and email already have functions
8066 in place for compatibility: they check the NID and also return "trusted"
8067 if the certificate is self signed.
8068 [Steve Henson]
8069
8070 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8071 traditional format into an EVP_PKEY structure.
8072 [Steve Henson]
8073
8074 *) Add a password callback function PEM_cb() which either prompts for
8075 a password if usr_data is NULL or otherwise assumes it is a null
8076 terminated password. Allow passwords to be passed on command line
8077 environment or config files in a few more utilities.
8078 [Steve Henson]
8079
8080 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8081 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8082 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8083 Update documentation.
8084 [Steve Henson]
8085
8086 *) Support for ASN1 "NULL" type. This could be handled before by using
8087 ASN1_TYPE but there wasn't any function that would try to read a NULL
8088 and produce an error if it couldn't. For compatibility we also have
8089 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8090 don't allocate anything because they don't need to.
8091 [Steve Henson]
8092
8093 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8094 for details.
8095 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8096
8097 *) Rebuild of the memory allocation routines used by OpenSSL code and
8098 possibly others as well. The purpose is to make an interface that
8099 provide hooks so anyone can build a separate set of allocation and
8100 deallocation routines to be used by OpenSSL, for example memory
8101 pool implementations, or something else, which was previously hard
8102 since Malloc(), Realloc() and Free() were defined as macros having
8103 the values malloc, realloc and free, respectively (except for Win32
8104 compilations). The same is provided for memory debugging code.
8105 OpenSSL already comes with functionality to find memory leaks, but
8106 this gives people a chance to debug other memory problems.
8107
8108 With these changes, a new set of functions and macros have appeared:
8109
8110 CRYPTO_set_mem_debug_functions() [F]
8111 CRYPTO_get_mem_debug_functions() [F]
8112 CRYPTO_dbg_set_options() [F]
8113 CRYPTO_dbg_get_options() [F]
8114 CRYPTO_malloc_debug_init() [M]
8115
8116 The memory debug functions are NULL by default, unless the library
8117 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8118 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8119 gives the standard debugging functions that come with OpenSSL) or
8120 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8121 provided by the library user) must be used. When the standard
8122 debugging functions are used, CRYPTO_dbg_set_options can be used to
8123 request additional information:
8124 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8125 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8126
8127 Also, things like CRYPTO_set_mem_functions will always give the
8128 expected result (the new set of functions is used for allocation
8129 and deallocation) at all times, regardless of platform and compiler
8130 options.
8131
8132 To finish it up, some functions that were never use in any other
8133 way than through macros have a new API and new semantic:
8134
8135 CRYPTO_dbg_malloc()
8136 CRYPTO_dbg_realloc()
8137 CRYPTO_dbg_free()
8138
8139 All macros of value have retained their old syntax.
8140 [Richard Levitte and Bodo Moeller]
8141
8142 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8143 ordering of SMIMECapabilities wasn't in "strength order" and there
8144 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8145 algorithm.
8146 [Steve Henson]
8147
8148 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8149 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8150 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8151
8152 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8153 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8154 functionality to handle multipart/signed properly) and a utility
8155 called 'smime' to call all this stuff. This is based on code I
8156 originally wrote for Celo who have kindly allowed it to be
8157 included in OpenSSL.
8158 [Steve Henson]
8159
8160 *) Add variants des_set_key_checked and des_set_key_unchecked of
8161 des_set_key (aka des_key_sched). Global variable des_check_key
8162 decides which of these is called by des_set_key; this way
8163 des_check_key behaves as it always did, but applications and
8164 the library itself, which was buggy for des_check_key == 1,
8165 have a cleaner way to pick the version they need.
8166 [Bodo Moeller]
8167
8168 *) New function PKCS12_newpass() which changes the password of a
8169 PKCS12 structure.
8170 [Steve Henson]
8171
8172 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8173 dynamic mix. In both cases the ids can be used as an index into the
8174 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8175 functions so they accept a list of the field values and the
8176 application doesn't need to directly manipulate the X509_TRUST
8177 structure.
8178 [Steve Henson]
8179
8180 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8181 need initialising.
8182 [Steve Henson]
8183
8184 *) Modify the way the V3 extension code looks up extensions. This now
8185 works in a similar way to the object code: we have some "standard"
8186 extensions in a static table which is searched with OBJ_bsearch()
8187 and the application can add dynamic ones if needed. The file
8188 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8189 updated whenever a new extension is added to the core code and kept
8190 in ext_nid order. There is a simple program 'tabtest.c' which checks
8191 this. New extensions are not added too often so this file can readily
8192 be maintained manually.
8193
8194 There are two big advantages in doing things this way. The extensions
8195 can be looked up immediately and no longer need to be "added" using
8196 X509V3_add_standard_extensions(): this function now does nothing.
8197 [Side note: I get *lots* of email saying the extension code doesn't
8198 work because people forget to call this function]
8199 Also no dynamic allocation is done unless new extensions are added:
8200 so if we don't add custom extensions there is no need to call
8201 X509V3_EXT_cleanup().
8202 [Steve Henson]
8203
8204 *) Modify enc utility's salting as follows: make salting the default. Add a
8205 magic header, so unsalted files fail gracefully instead of just decrypting
8206 to garbage. This is because not salting is a big security hole, so people
8207 should be discouraged from doing it.
8208 [Ben Laurie]
8209
8210 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8211 digest to be passed on the command line but it only used this
8212 parameter when signing a certificate. Modified so all relevant
8213 operations are affected by the digest parameter including the
8214 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8215 DSA key was used because it didn't fix the digest.
8216 [Steve Henson]
8217
8218 *) Initial certificate chain verify code. Currently tests the untrusted
8219 certificates for consistency with the verify purpose (which is set
8220 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8221
8222 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8223 this is because it will reject chains with invalid extensions whereas
8224 every previous version of OpenSSL and SSLeay made no checks at all.
8225
8226 Trust code: checks the root CA for the relevant trust settings. Trust
8227 settings have an initial value consistent with the verify purpose: e.g.
8228 if the verify purpose is for SSL client use it expects the CA to be
8229 trusted for SSL client use. However the default value can be changed to
8230 permit custom trust settings: one example of this would be to only trust
8231 certificates from a specific "secure" set of CAs.
8232
8233 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8234 which should be used for version portability: especially since the
8235 verify structure is likely to change more often now.
8236
8237 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8238 to set them. If not set then assume SSL clients will verify SSL servers
8239 and vice versa.
8240
8241 Two new options to the verify program: -untrusted allows a set of
8242 untrusted certificates to be passed in and -purpose which sets the
8243 intended purpose of the certificate. If a purpose is set then the
8244 new chain verify code is used to check extension consistency.
8245 [Steve Henson]
8246
8247 *) Support for the authority information access extension.
8248 [Steve Henson]
8249
8250 *) Modify RSA and DSA PEM read routines to transparently handle
8251 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8252 public keys in a format compatible with certificate
8253 SubjectPublicKeyInfo structures. Unfortunately there were already
8254 functions called *_PublicKey_* which used various odd formats so
8255 these are retained for compatibility: however the DSA variants were
8256 never in a public release so they have been deleted. Changed dsa/rsa
8257 utilities to handle the new format: note no releases ever handled public
8258 keys so we should be OK.
8259
8260 The primary motivation for this change is to avoid the same fiasco
8261 that dogs private keys: there are several incompatible private key
8262 formats some of which are standard and some OpenSSL specific and
8263 require various evil hacks to allow partial transparent handling and
8264 even then it doesn't work with DER formats. Given the option anything
8265 other than PKCS#8 should be dumped: but the other formats have to
8266 stay in the name of compatibility.
8267
8268 With public keys and the benefit of hindsight one standard format
8269 is used which works with EVP_PKEY, RSA or DSA structures: though
8270 it clearly returns an error if you try to read the wrong kind of key.
8271
8272 Added a -pubkey option to the 'x509' utility to output the public key.
8273 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
8274 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
8275 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
8276 that do the same as the EVP_PKEY_assign_*() except they up the
8277 reference count of the added key (they don't "swallow" the
8278 supplied key).
8279 [Steve Henson]
8280
8281 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
8282 CRLs would fail if the file contained no certificates or no CRLs:
8283 added a new function to read in both types and return the number
8284 read: this means that if none are read it will be an error. The
8285 DER versions of the certificate and CRL reader would always fail
8286 because it isn't possible to mix certificates and CRLs in DER format
8287 without choking one or the other routine. Changed this to just read
8288 a certificate: this is the best we can do. Also modified the code
8289 in apps/verify.c to take notice of return codes: it was previously
8290 attempting to read in certificates from NULL pointers and ignoring
8291 any errors: this is one reason why the cert and CRL reader seemed
8292 to work. It doesn't check return codes from the default certificate
8293 routines: these may well fail if the certificates aren't installed.
8294 [Steve Henson]
8295
8296 *) Code to support otherName option in GeneralName.
8297 [Steve Henson]
8298
8299 *) First update to verify code. Change the verify utility
8300 so it warns if it is passed a self signed certificate:
8301 for consistency with the normal behaviour. X509_verify
8302 has been modified to it will now verify a self signed
8303 certificate if *exactly* the same certificate appears
8304 in the store: it was previously impossible to trust a
8305 single self signed certificate. This means that:
8306 openssl verify ss.pem
8307 now gives a warning about a self signed certificate but
8308 openssl verify -CAfile ss.pem ss.pem
8309 is OK.
8310 [Steve Henson]
8311
8312 *) For servers, store verify_result in SSL_SESSION data structure
8313 (and add it to external session representation).
8314 This is needed when client certificate verifications fails,
8315 but an application-provided verification callback (set by
8316 SSL_CTX_set_cert_verify_callback) allows accepting the session
8317 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
8318 but returns 1): When the session is reused, we have to set
8319 ssl->verify_result to the appropriate error code to avoid
8320 security holes.
8321 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
8322
8323 *) Fix a bug in the new PKCS#7 code: it didn't consider the
8324 case in PKCS7_dataInit() where the signed PKCS7 structure
8325 didn't contain any existing data because it was being created.
8326 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
8327
8328 *) Add a salt to the key derivation routines in enc.c. This
8329 forms the first 8 bytes of the encrypted file. Also add a
8330 -S option to allow a salt to be input on the command line.
8331 [Steve Henson]
8332
8333 *) New function X509_cmp(). Oddly enough there wasn't a function
8334 to compare two certificates. We do this by working out the SHA1
8335 hash and comparing that. X509_cmp() will be needed by the trust
8336 code.
8337 [Steve Henson]
8338
8339 *) SSL_get1_session() is like SSL_get_session(), but increments
8340 the reference count in the SSL_SESSION returned.
8341 [Geoff Thorpe <geoff@eu.c2.net>]
8342
8343 *) Fix for 'req': it was adding a null to request attributes.
8344 Also change the X509_LOOKUP and X509_INFO code to handle
8345 certificate auxiliary information.
8346 [Steve Henson]
8347
8348 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
8349 the 'enc' command.
8350 [Steve Henson]
8351
8352 *) Add the possibility to add extra information to the memory leak
8353 detecting output, to form tracebacks, showing from where each
8354 allocation was originated: CRYPTO_push_info("constant string") adds
8355 the string plus current file name and line number to a per-thread
8356 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
8357 is like calling CYRPTO_pop_info() until the stack is empty.
8358 Also updated memory leak detection code to be multi-thread-safe.
8359 [Richard Levitte]
8360
8361 *) Add options -text and -noout to pkcs7 utility and delete the
8362 encryption options which never did anything. Update docs.
8363 [Steve Henson]
8364
8365 *) Add options to some of the utilities to allow the pass phrase
8366 to be included on either the command line (not recommended on
8367 OSes like Unix) or read from the environment. Update the
8368 manpages and fix a few bugs.
8369 [Steve Henson]
8370
8371 *) Add a few manpages for some of the openssl commands.
8372 [Steve Henson]
8373
8374 *) Fix the -revoke option in ca. It was freeing up memory twice,
8375 leaking and not finding already revoked certificates.
8376 [Steve Henson]
8377
8378 *) Extensive changes to support certificate auxiliary information.
8379 This involves the use of X509_CERT_AUX structure and X509_AUX
8380 functions. An X509_AUX function such as PEM_read_X509_AUX()
8381 can still read in a certificate file in the usual way but it
8382 will also read in any additional "auxiliary information". By
8383 doing things this way a fair degree of compatibility can be
8384 retained: existing certificates can have this information added
8385 using the new 'x509' options.
8386
8387 Current auxiliary information includes an "alias" and some trust
8388 settings. The trust settings will ultimately be used in enhanced
8389 certificate chain verification routines: currently a certificate
8390 can only be trusted if it is self signed and then it is trusted
8391 for all purposes.
8392 [Steve Henson]
8393
8394 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
8395 The problem was that one of the replacement routines had not been working
8396 since SSLeay releases. For now the offending routine has been replaced
8397 with non-optimised assembler. Even so, this now gives around 95%
8398 performance improvement for 1024 bit RSA signs.
8399 [Mark Cox]
8400
8401 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
8402 handling. Most clients have the effective key size in bits equal to
8403 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
8404 A few however don't do this and instead use the size of the decrypted key
8405 to determine the RC2 key length and the AlgorithmIdentifier to determine
8406 the effective key length. In this case the effective key length can still
8407 be 40 bits but the key length can be 168 bits for example. This is fixed
8408 by manually forcing an RC2 key into the EVP_PKEY structure because the
8409 EVP code can't currently handle unusual RC2 key sizes: it always assumes
8410 the key length and effective key length are equal.
8411 [Steve Henson]
8412
8413 *) Add a bunch of functions that should simplify the creation of
8414 X509_NAME structures. Now you should be able to do:
8415 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8416 and have it automatically work out the correct field type and fill in
8417 the structures. The more adventurous can try:
8418 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8419 and it will (hopefully) work out the correct multibyte encoding.
8420 [Steve Henson]
8421
8422 *) Change the 'req' utility to use the new field handling and multibyte
8423 copy routines. Before the DN field creation was handled in an ad hoc
8424 way in req, ca, and x509 which was rather broken and didn't support
8425 BMPStrings or UTF8Strings. Since some software doesn't implement
8426 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8427 using the dirstring_type option. See the new comment in the default
8428 openssl.cnf for more info.
8429 [Steve Henson]
8430
8431 *) Make crypto/rand/md_rand.c more robust:
8432 - Assure unique random numbers after fork().
8433 - Make sure that concurrent threads access the global counter and
8434 md serializably so that we never lose entropy in them
8435 or use exactly the same state in multiple threads.
8436 Access to the large state is not always serializable because
8437 the additional locking could be a performance killer, and
8438 md should be large enough anyway.
8439 [Bodo Moeller]
8440
8441 *) New file apps/app_rand.c with commonly needed functionality
8442 for handling the random seed file.
8443
8444 Use the random seed file in some applications that previously did not:
8445 ca,
8446 dsaparam -genkey (which also ignored its '-rand' option),
8447 s_client,
8448 s_server,
8449 x509 (when signing).
8450 Except on systems with /dev/urandom, it is crucial to have a random
8451 seed file at least for key creation, DSA signing, and for DH exchanges;
8452 for RSA signatures we could do without one.
8453
8454 gendh and gendsa (unlike genrsa) used to read only the first byte
8455 of each file listed in the '-rand' option. The function as previously
8456 found in genrsa is now in app_rand.c and is used by all programs
8457 that support '-rand'.
8458 [Bodo Moeller]
8459
8460 *) In RAND_write_file, use mode 0600 for creating files;
8461 don't just chmod when it may be too late.
8462 [Bodo Moeller]
8463
8464 *) Report an error from X509_STORE_load_locations
8465 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8466 [Bill Perry]
8467
8468 *) New function ASN1_mbstring_copy() this copies a string in either
8469 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8470 into an ASN1_STRING type. A mask of permissible types is passed
8471 and it chooses the "minimal" type to use or an error if not type
8472 is suitable.
8473 [Steve Henson]
8474
8475 *) Add function equivalents to the various macros in asn1.h. The old
8476 macros are retained with an M_ prefix. Code inside the library can
8477 use the M_ macros. External code (including the openssl utility)
8478 should *NOT* in order to be "shared library friendly".
8479 [Steve Henson]
8480
8481 *) Add various functions that can check a certificate's extensions
8482 to see if it usable for various purposes such as SSL client,
8483 server or S/MIME and CAs of these types. This is currently
8484 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8485 verification. Also added a -purpose flag to x509 utility to
8486 print out all the purposes.
8487 [Steve Henson]
8488
8489 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8490 functions.
8491 [Steve Henson]
8492
8493 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8494 for, obtain and decode and extension and obtain its critical flag.
8495 This allows all the necessary extension code to be handled in a
8496 single function call.
8497 [Steve Henson]
8498
8499 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8500 platforms. See crypto/rc4/rc4_enc.c for further details.
8501 [Andy Polyakov]
8502
8503 *) New -noout option to asn1parse. This causes no output to be produced
8504 its main use is when combined with -strparse and -out to extract data
8505 from a file (which may not be in ASN.1 format).
8506 [Steve Henson]
8507
8508 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8509 when producing the local key id.
8510 [Richard Levitte <levitte@stacken.kth.se>]
8511
8512 *) New option -dhparam in s_server. This allows a DH parameter file to be
8513 stated explicitly. If it is not stated then it tries the first server
8514 certificate file. The previous behaviour hard coded the filename
8515 "server.pem".
8516 [Steve Henson]
8517
8518 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8519 a public key to be input or output. For example:
8520 openssl rsa -in key.pem -pubout -out pubkey.pem
8521 Also added necessary DSA public key functions to handle this.
8522 [Steve Henson]
8523
8524 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8525 in the message. This was handled by allowing
8526 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8527 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8528
8529 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8530 to the end of the strings whereas this didn't. This would cause problems
8531 if strings read with d2i_ASN1_bytes() were later modified.
8532 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8533
8534 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8535 data and it contains EOF it will end up returning an error. This is
8536 caused by input 46 bytes long. The cause is due to the way base64
8537 BIOs find the start of base64 encoded data. They do this by trying a
8538 trial decode on each line until they find one that works. When they
8539 do a flag is set and it starts again knowing it can pass all the
8540 data directly through the decoder. Unfortunately it doesn't reset
8541 the context it uses. This means that if EOF is reached an attempt
8542 is made to pass two EOFs through the context and this causes the
8543 resulting error. This can also cause other problems as well. As is
8544 usual with these problems it takes *ages* to find and the fix is
8545 trivial: move one line.
8546 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8547
8548 *) Ugly workaround to get s_client and s_server working under Windows. The
8549 old code wouldn't work because it needed to select() on sockets and the
8550 tty (for keypresses and to see if data could be written). Win32 only
8551 supports select() on sockets so we select() with a 1s timeout on the
8552 sockets and then see if any characters are waiting to be read, if none
8553 are present then we retry, we also assume we can always write data to
8554 the tty. This isn't nice because the code then blocks until we've
8555 received a complete line of data and it is effectively polling the
8556 keyboard at 1s intervals: however it's quite a bit better than not
8557 working at all :-) A dedicated Windows application might handle this
8558 with an event loop for example.
8559 [Steve Henson]
8560
8561 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8562 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8563 will be called when RSA_sign() and RSA_verify() are used. This is useful
8564 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8565 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8566 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8567 This necessitated the support of an extra signature type NID_md5_sha1
8568 for SSL signatures and modifications to the SSL library to use it instead
8569 of calling RSA_public_decrypt() and RSA_private_encrypt().
8570 [Steve Henson]
8571
8572 *) Add new -verify -CAfile and -CApath options to the crl program, these
8573 will lookup a CRL issuers certificate and verify the signature in a
8574 similar way to the verify program. Tidy up the crl program so it
8575 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8576 less strict. It will now permit CRL extensions even if it is not
8577 a V2 CRL: this will allow it to tolerate some broken CRLs.
8578 [Steve Henson]
8579
8580 *) Initialize all non-automatic variables each time one of the openssl
8581 sub-programs is started (this is necessary as they may be started
8582 multiple times from the "OpenSSL>" prompt).
8583 [Lennart Bang, Bodo Moeller]
8584
8585 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8586 removing all other RSA functionality (this is what NO_RSA does). This
8587 is so (for example) those in the US can disable those operations covered
8588 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8589 key generation.
8590 [Steve Henson]
8591
8592 *) Non-copying interface to BIO pairs.
8593 (still largely untested)
8594 [Bodo Moeller]
8595
8596 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8597 ASCII string. This was handled independently in various places before.
8598 [Steve Henson]
8599
8600 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8601 UTF8 strings a character at a time.
8602 [Steve Henson]
8603
8604 *) Use client_version from client hello to select the protocol
8605 (s23_srvr.c) and for RSA client key exchange verification
8606 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8607 [Bodo Moeller]
8608
8609 *) Add various utility functions to handle SPKACs, these were previously
8610 handled by poking round in the structure internals. Added new function
8611 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8612 print, verify and generate SPKACs. Based on an original idea from
8613 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8614 [Steve Henson]
8615
8616 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8617 [Andy Polyakov]
8618
8619 *) Allow the config file extension section to be overwritten on the
8620 command line. Based on an original idea from Massimiliano Pala
8621 <madwolf@comune.modena.it>. The new option is called -extensions
8622 and can be applied to ca, req and x509. Also -reqexts to override
8623 the request extensions in req and -crlexts to override the crl extensions
8624 in ca.
8625 [Steve Henson]
8626
8627 *) Add new feature to the SPKAC handling in ca. Now you can include
8628 the same field multiple times by preceding it by "XXXX." for example:
8629 1.OU="Unit name 1"
8630 2.OU="Unit name 2"
8631 this is the same syntax as used in the req config file.
8632 [Steve Henson]
8633
8634 *) Allow certificate extensions to be added to certificate requests. These
8635 are specified in a 'req_extensions' option of the req section of the
8636 config file. They can be printed out with the -text option to req but
8637 are otherwise ignored at present.
8638 [Steve Henson]
8639
8640 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8641 data read consists of only the final block it would not decrypted because
8642 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8643 A misplaced 'break' also meant the decrypted final block might not be
8644 copied until the next read.
8645 [Steve Henson]
8646
8647 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8648 a few extra parameters to the DH structure: these will be useful if
8649 for example we want the value of 'q' or implement X9.42 DH.
8650 [Steve Henson]
8651
8652 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8653 provides hooks that allow the default DSA functions or functions on a
8654 "per key" basis to be replaced. This allows hardware acceleration and
8655 hardware key storage to be handled without major modification to the
8656 library. Also added low level modexp hooks and CRYPTO_EX structure and
8657 associated functions.
8658 [Steve Henson]
8659
8660 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8661 as "read only": it can't be written to and the buffer it points to will
8662 not be freed. Reading from a read only BIO is much more efficient than
8663 a normal memory BIO. This was added because there are several times when
8664 an area of memory needs to be read from a BIO. The previous method was
8665 to create a memory BIO and write the data to it, this results in two
8666 copies of the data and an O(n^2) reading algorithm. There is a new
8667 function BIO_new_mem_buf() which creates a read only memory BIO from
8668 an area of memory. Also modified the PKCS#7 routines to use read only
8669 memory BIOs.
8670 [Steve Henson]
8671
8672 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8673 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8674 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8675 but a retry condition occured while trying to read the rest.
8676 [Bodo Moeller]
8677
8678 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8679 NID_pkcs7_encrypted by default: this was wrong since this should almost
8680 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8681 the encrypted data type: this is a more sensible place to put it and it
8682 allows the PKCS#12 code to be tidied up that duplicated this
8683 functionality.
8684 [Steve Henson]
8685
8686 *) Changed obj_dat.pl script so it takes its input and output files on
8687 the command line. This should avoid shell escape redirection problems
8688 under Win32.
8689 [Steve Henson]
8690
8691 *) Initial support for certificate extension requests, these are included
8692 in things like Xenroll certificate requests. Included functions to allow
8693 extensions to be obtained and added.
8694 [Steve Henson]
8695
8696 *) -crlf option to s_client and s_server for sending newlines as
8697 CRLF (as required by many protocols).
8698 [Bodo Moeller]
8699
8700 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8701
8702 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8703 [Ralf S. Engelschall]
8704
8705 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8706 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8707
8708 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8709 program.
8710 [Steve Henson]
8711
8712 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8713 DH parameters/keys (q is lost during that conversion, but the resulting
8714 DH parameters contain its length).
8715
8716 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8717 much faster than DH_generate_parameters (which creates parameters
8718 where p = 2*q + 1), and also the smaller q makes DH computations
8719 much more efficient (160-bit exponentiation instead of 1024-bit
8720 exponentiation); so this provides a convenient way to support DHE
8721 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8722 utter importance to use
8723 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8724 or
8725 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8726 when such DH parameters are used, because otherwise small subgroup
8727 attacks may become possible!
8728 [Bodo Moeller]
8729
8730 *) Avoid memory leak in i2d_DHparams.
8731 [Bodo Moeller]
8732
8733 *) Allow the -k option to be used more than once in the enc program:
8734 this allows the same encrypted message to be read by multiple recipients.
8735 [Steve Henson]
8736
8737 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8738 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8739 it will always use the numerical form of the OID, even if it has a short
8740 or long name.
8741 [Steve Henson]
8742
8743 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8744 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8745 otherwise bn_mod_exp was called. In the case of hardware keys for example
8746 no private key components need be present and it might store extra data
8747 in the RSA structure, which cannot be accessed from bn_mod_exp.
8748 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8749 private key operations.
8750 [Steve Henson]
8751
8752 *) Added support for SPARC Linux.
8753 [Andy Polyakov]
8754
8755 *) pem_password_cb function type incompatibly changed from
8756 typedef int pem_password_cb(char *buf, int size, int rwflag);
8757 to
8758 ....(char *buf, int size, int rwflag, void *userdata);
8759 so that applications can pass data to their callbacks:
8760 The PEM[_ASN1]_{read,write}... functions and macros now take an
8761 additional void * argument, which is just handed through whenever
8762 the password callback is called.
8763 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8764
8765 New function SSL_CTX_set_default_passwd_cb_userdata.
8766
8767 Compatibility note: As many C implementations push function arguments
8768 onto the stack in reverse order, the new library version is likely to
8769 interoperate with programs that have been compiled with the old
8770 pem_password_cb definition (PEM_whatever takes some data that
8771 happens to be on the stack as its last argument, and the callback
8772 just ignores this garbage); but there is no guarantee whatsoever that
8773 this will work.
8774
8775 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8776 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8777 problems not only on Windows, but also on some Unix platforms.
8778 To avoid problematic command lines, these definitions are now in an
8779 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8780 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8781 [Bodo Moeller]
8782
8783 *) MIPS III/IV assembler module is reimplemented.
8784 [Andy Polyakov]
8785
8786 *) More DES library cleanups: remove references to srand/rand and
8787 delete an unused file.
8788 [Ulf Möller]
8789
8790 *) Add support for the the free Netwide assembler (NASM) under Win32,
8791 since not many people have MASM (ml) and it can be hard to obtain.
8792 This is currently experimental but it seems to work OK and pass all
8793 the tests. Check out INSTALL.W32 for info.
8794 [Steve Henson]
8795
8796 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8797 without temporary keys kept an extra copy of the server key,
8798 and connections with temporary keys did not free everything in case
8799 of an error.
8800 [Bodo Moeller]
8801
8802 *) New function RSA_check_key and new openssl rsa option -check
8803 for verifying the consistency of RSA keys.
8804 [Ulf Moeller, Bodo Moeller]
8805
8806 *) Various changes to make Win32 compile work:
8807 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8808 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8809 comparison" warnings.
8810 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8811 [Steve Henson]
8812
8813 *) Add a debugging option to PKCS#5 v2 key generation function: when
8814 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8815 derived keys are printed to stderr.
8816 [Steve Henson]
8817
8818 *) Copy the flags in ASN1_STRING_dup().
8819 [Roman E. Pavlov <pre@mo.msk.ru>]
8820
8821 *) The x509 application mishandled signing requests containing DSA
8822 keys when the signing key was also DSA and the parameters didn't match.
8823
8824 It was supposed to omit the parameters when they matched the signing key:
8825 the verifying software was then supposed to automatically use the CA's
8826 parameters if they were absent from the end user certificate.
8827
8828 Omitting parameters is no longer recommended. The test was also
8829 the wrong way round! This was probably due to unusual behaviour in
8830 EVP_cmp_parameters() which returns 1 if the parameters match.
8831 This meant that parameters were omitted when they *didn't* match and
8832 the certificate was useless. Certificates signed with 'ca' didn't have
8833 this bug.
8834 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8835
8836 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8837 The interface is as follows:
8838 Applications can use
8839 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8840 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8841 "off" is now the default.
8842 The library internally uses
8843 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8844 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8845 to disable memory-checking temporarily.
8846
8847 Some inconsistent states that previously were possible (and were
8848 even the default) are now avoided.
8849
8850 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8851 with each memory chunk allocated; this is occasionally more helpful
8852 than just having a counter.
8853
8854 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8855
8856 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8857 extensions.
8858 [Bodo Moeller]
8859
8860 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8861 which largely parallels "options", but is for changing API behaviour,
8862 whereas "options" are about protocol behaviour.
8863 Initial "mode" flags are:
8864
8865 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8866 a single record has been written.
8867 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8868 retries use the same buffer location.
8869 (But all of the contents must be
8870 copied!)
8871 [Bodo Moeller]
8872
8873 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8874 worked.
8875
8876 *) Fix problems with no-hmac etc.
8877 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8878
8879 *) New functions RSA_get_default_method(), RSA_set_method() and
8880 RSA_get_method(). These allows replacement of RSA_METHODs without having
8881 to mess around with the internals of an RSA structure.
8882 [Steve Henson]
8883
8884 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8885 Also really enable memory leak checks in openssl.c and in some
8886 test programs.
8887 [Chad C. Mulligan, Bodo Moeller]
8888
8889 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8890 up the length of negative integers. This has now been simplified to just
8891 store the length when it is first determined and use it later, rather
8892 than trying to keep track of where data is copied and updating it to
8893 point to the end.
8894 [Steve Henson, reported by Brien Wheeler
8895 <bwheeler@authentica-security.com>]
8896
8897 *) Add a new function PKCS7_signatureVerify. This allows the verification
8898 of a PKCS#7 signature but with the signing certificate passed to the
8899 function itself. This contrasts with PKCS7_dataVerify which assumes the
8900 certificate is present in the PKCS#7 structure. This isn't always the
8901 case: certificates can be omitted from a PKCS#7 structure and be
8902 distributed by "out of band" means (such as a certificate database).
8903 [Steve Henson]
8904
8905 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8906 function prototypes in pem.h, also change util/mkdef.pl to add the
8907 necessary function names.
8908 [Steve Henson]
8909
8910 *) mk1mf.pl (used by Windows builds) did not properly read the
8911 options set by Configure in the top level Makefile, and Configure
8912 was not even able to write more than one option correctly.
8913 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8914 [Bodo Moeller]
8915
8916 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8917 file to be loaded from a BIO or FILE pointer. The BIO version will
8918 for example allow memory BIOs to contain config info.
8919 [Steve Henson]
8920
8921 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8922 Whoever hopes to achieve shared-library compatibility across versions
8923 must use this, not the compile-time macro.
8924 (Exercise 0.9.4: Which is the minimum library version required by
8925 such programs?)
8926 Note: All this applies only to multi-threaded programs, others don't
8927 need locks.
8928 [Bodo Moeller]
8929
8930 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8931 through a BIO pair triggered the default case, i.e.
8932 SSLerr(...,SSL_R_UNKNOWN_STATE).
8933 [Bodo Moeller]
8934
8935 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8936 can use the SSL library even if none of the specific BIOs is
8937 appropriate.
8938 [Bodo Moeller]
8939
8940 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8941 for the encoded length.
8942 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8943
8944 *) Add initial documentation of the X509V3 functions.
8945 [Steve Henson]
8946
8947 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8948 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8949 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8950 secure PKCS#8 private key format with a high iteration count.
8951 [Steve Henson]
8952
8953 *) Fix determination of Perl interpreter: A perl or perl5
8954 _directory_ in $PATH was also accepted as the interpreter.
8955 [Ralf S. Engelschall]
8956
8957 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8958 wrong with it but it was very old and did things like calling
8959 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8960 unusual formatting.
8961 [Steve Henson]
8962
8963 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8964 to use the new extension code.
8965 [Steve Henson]
8966
8967 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8968 with macros. This should make it easier to change their form, add extra
8969 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8970 constant.
8971 [Steve Henson]
8972
8973 *) Add to configuration table a new entry that can specify an alternative
8974 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8975 according to Mark Crispin <MRC@Panda.COM>.
8976 [Bodo Moeller]
8977
8978 #if 0
8979 *) DES CBC did not update the IV. Weird.
8980 [Ben Laurie]
8981 #else
8982 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8983 Changing the behaviour of the former might break existing programs --
8984 where IV updating is needed, des_ncbc_encrypt can be used.
8985 #endif
8986
8987 *) When bntest is run from "make test" it drives bc to check its
8988 calculations, as well as internally checking them. If an internal check
8989 fails, it needs to cause bc to give a non-zero result or make test carries
8990 on without noticing the failure. Fixed.
8991 [Ben Laurie]
8992
8993 *) DES library cleanups.
8994 [Ulf Möller]
8995
8996 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8997 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8998 ciphers. NOTE: although the key derivation function has been verified
8999 against some published test vectors it has not been extensively tested
9000 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9001 of v2.0.
9002 [Steve Henson]
9003
9004 *) Instead of "mkdir -p", which is not fully portable, use new
9005 Perl script "util/mkdir-p.pl".
9006 [Bodo Moeller]
9007
9008 *) Rewrite the way password based encryption (PBE) is handled. It used to
9009 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9010 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9011 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9012 the 'parameter' field of the AlgorithmIdentifier is passed to the
9013 underlying key generation function so it must do its own ASN1 parsing.
9014 This has also changed the EVP_PBE_CipherInit() function which now has a
9015 'parameter' argument instead of literal salt and iteration count values
9016 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9017 [Steve Henson]
9018
9019 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9020 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9021 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9022 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9023 value was just used as a "magic string" and not used directly its
9024 value doesn't matter.
9025 [Steve Henson]
9026
9027 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9028 support mutable.
9029 [Ben Laurie]
9030
9031 *) "linux-sparc64" configuration (ultrapenguin).
9032 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9033 "linux-sparc" configuration.
9034 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9035
9036 *) config now generates no-xxx options for missing ciphers.
9037 [Ulf Möller]
9038
9039 *) Support the EBCDIC character set (work in progress).
9040 File ebcdic.c not yet included because it has a different license.
9041 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9042
9043 *) Support BS2000/OSD-POSIX.
9044 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9045
9046 *) Make callbacks for key generation use void * instead of char *.
9047 [Ben Laurie]
9048
9049 *) Make S/MIME samples compile (not yet tested).
9050 [Ben Laurie]
9051
9052 *) Additional typesafe stacks.
9053 [Ben Laurie]
9054
9055 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9056 [Bodo Moeller]
9057
9058
9059 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9060
9061 *) New configuration variant "sco5-gcc".
9062
9063 *) Updated some demos.
9064 [Sean O Riordain, Wade Scholine]
9065
9066 *) Add missing BIO_free at exit of pkcs12 application.
9067 [Wu Zhigang]
9068
9069 *) Fix memory leak in conf.c.
9070 [Steve Henson]
9071
9072 *) Updates for Win32 to assembler version of MD5.
9073 [Steve Henson]
9074
9075 *) Set #! path to perl in apps/der_chop to where we found it
9076 instead of using a fixed path.
9077 [Bodo Moeller]
9078
9079 *) SHA library changes for irix64-mips4-cc.
9080 [Andy Polyakov]
9081
9082 *) Improvements for VMS support.
9083 [Richard Levitte]
9084
9085
9086 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9087
9088 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9089 This also avoids the problems with SC4.2 and unpatched SC5.
9090 [Andy Polyakov <appro@fy.chalmers.se>]
9091
9092 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9093 These are required because of the typesafe stack would otherwise break
9094 existing code. If old code used a structure member which used to be STACK
9095 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9096 sk_num or sk_value it would produce an error because the num, data members
9097 are not present in STACK_OF. Now it just produces a warning. sk_set
9098 replaces the old method of assigning a value to sk_value
9099 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9100 that does this will no longer work (and should use sk_set instead) but
9101 this could be regarded as a "questionable" behaviour anyway.
9102 [Steve Henson]
9103
9104 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9105 correctly handle encrypted S/MIME data.
9106 [Steve Henson]
9107
9108 *) Change type of various DES function arguments from des_cblock
9109 (which means, in function argument declarations, pointer to char)
9110 to des_cblock * (meaning pointer to array with 8 char elements),
9111 which allows the compiler to do more typechecking; it was like
9112 that back in SSLeay, but with lots of ugly casts.
9113
9114 Introduce new type const_des_cblock.
9115 [Bodo Moeller]
9116
9117 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9118 problems: find RecipientInfo structure that matches recipient certificate
9119 and initialise the ASN1 structures properly based on passed cipher.
9120 [Steve Henson]
9121
9122 *) Belatedly make the BN tests actually check the results.
9123 [Ben Laurie]
9124
9125 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9126 to and from BNs: it was completely broken. New compilation option
9127 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9128 key elements as negative integers.
9129 [Steve Henson]
9130
9131 *) Reorganize and speed up MD5.
9132 [Andy Polyakov <appro@fy.chalmers.se>]
9133
9134 *) VMS support.
9135 [Richard Levitte <richard@levitte.org>]
9136
9137 *) New option -out to asn1parse to allow the parsed structure to be
9138 output to a file. This is most useful when combined with the -strparse
9139 option to examine the output of things like OCTET STRINGS.
9140 [Steve Henson]
9141
9142 *) Make SSL library a little more fool-proof by not requiring any longer
9143 that SSL_set_{accept,connect}_state be called before
9144 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9145 in many applications because usually everything *appeared* to work as
9146 intended anyway -- now it really works as intended).
9147 [Bodo Moeller]
9148
9149 *) Move openssl.cnf out of lib/.
9150 [Ulf Möller]
9151
9152 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9153 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9154 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9155 [Ralf S. Engelschall]
9156
9157 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9158 handle PKCS#7 enveloped data properly.
9159 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9160
9161 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9162 copying pointers. The cert_st handling is changed by this in
9163 various ways (and thus what used to be known as ctx->default_cert
9164 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9165 any longer when s->cert does not give us what we need).
9166 ssl_cert_instantiate becomes obsolete by this change.
9167 As soon as we've got the new code right (possibly it already is?),
9168 we have solved a couple of bugs of the earlier code where s->cert
9169 was used as if it could not have been shared with other SSL structures.
9170
9171 Note that using the SSL API in certain dirty ways now will result
9172 in different behaviour than observed with earlier library versions:
9173 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9174 does not influence s as it used to.
9175
9176 In order to clean up things more thoroughly, inside SSL_SESSION
9177 we don't use CERT any longer, but a new structure SESS_CERT
9178 that holds per-session data (if available); currently, this is
9179 the peer's certificate chain and, for clients, the server's certificate
9180 and temporary key. CERT holds only those values that can have
9181 meaningful defaults in an SSL_CTX.
9182 [Bodo Moeller]
9183
9184 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9185 from the internal representation. Various PKCS#7 fixes: remove some
9186 evil casts and set the enc_dig_alg field properly based on the signing
9187 key type.
9188 [Steve Henson]
9189
9190 *) Allow PKCS#12 password to be set from the command line or the
9191 environment. Let 'ca' get its config file name from the environment
9192 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9193 and 'x509').
9194 [Steve Henson]
9195
9196 *) Allow certificate policies extension to use an IA5STRING for the
9197 organization field. This is contrary to the PKIX definition but
9198 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9199 extension option.
9200 [Steve Henson]
9201
9202 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9203 without disallowing inline assembler and the like for non-pedantic builds.
9204 [Ben Laurie]
9205
9206 *) Support Borland C++ builder.
9207 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9208
9209 *) Support Mingw32.
9210 [Ulf Möller]
9211
9212 *) SHA-1 cleanups and performance enhancements.
9213 [Andy Polyakov <appro@fy.chalmers.se>]
9214
9215 *) Sparc v8plus assembler for the bignum library.
9216 [Andy Polyakov <appro@fy.chalmers.se>]
9217
9218 *) Accept any -xxx and +xxx compiler options in Configure.
9219 [Ulf Möller]
9220
9221 *) Update HPUX configuration.
9222 [Anonymous]
9223
9224 *) Add missing sk_<type>_unshift() function to safestack.h
9225 [Ralf S. Engelschall]
9226
9227 *) New function SSL_CTX_use_certificate_chain_file that sets the
9228 "extra_cert"s in addition to the certificate. (This makes sense
9229 only for "PEM" format files, as chains as a whole are not
9230 DER-encoded.)
9231 [Bodo Moeller]
9232
9233 *) Support verify_depth from the SSL API.
9234 x509_vfy.c had what can be considered an off-by-one-error:
9235 Its depth (which was not part of the external interface)
9236 was actually counting the number of certificates in a chain;
9237 now it really counts the depth.
9238 [Bodo Moeller]
9239
9240 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9241 instead of X509err, which often resulted in confusing error
9242 messages since the error codes are not globally unique
9243 (e.g. an alleged error in ssl3_accept when a certificate
9244 didn't match the private key).
9245
9246 *) New function SSL_CTX_set_session_id_context that allows to set a default
9247 value (so that you don't need SSL_set_session_id_context for each
9248 connection using the SSL_CTX).
9249 [Bodo Moeller]
9250
9251 *) OAEP decoding bug fix.
9252 [Ulf Möller]
9253
9254 *) Support INSTALL_PREFIX for package builders, as proposed by
9255 David Harris.
9256 [Bodo Moeller]
9257
9258 *) New Configure options "threads" and "no-threads". For systems
9259 where the proper compiler options are known (currently Solaris
9260 and Linux), "threads" is the default.
9261 [Bodo Moeller]
9262
9263 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
9264 [Bodo Moeller]
9265
9266 *) Install various scripts to $(OPENSSLDIR)/misc, not to
9267 $(INSTALLTOP)/bin -- they shouldn't clutter directories
9268 such as /usr/local/bin.
9269 [Bodo Moeller]
9270
9271 *) "make linux-shared" to build shared libraries.
9272 [Niels Poppe <niels@netbox.org>]
9273
9274 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
9275 [Ulf Möller]
9276
9277 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
9278 extension adding in x509 utility.
9279 [Steve Henson]
9280
9281 *) Remove NOPROTO sections and error code comments.
9282 [Ulf Möller]
9283
9284 *) Partial rewrite of the DEF file generator to now parse the ANSI
9285 prototypes.
9286 [Steve Henson]
9287
9288 *) New Configure options --prefix=DIR and --openssldir=DIR.
9289 [Ulf Möller]
9290
9291 *) Complete rewrite of the error code script(s). It is all now handled
9292 by one script at the top level which handles error code gathering,
9293 header rewriting and C source file generation. It should be much better
9294 than the old method: it now uses a modified version of Ulf's parser to
9295 read the ANSI prototypes in all header files (thus the old K&R definitions
9296 aren't needed for error creation any more) and do a better job of
9297 translating function codes into names. The old 'ASN1 error code imbedded
9298 in a comment' is no longer necessary and it doesn't use .err files which
9299 have now been deleted. Also the error code call doesn't have to appear all
9300 on one line (which resulted in some large lines...).
9301 [Steve Henson]
9302
9303 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
9304 [Bodo Moeller]
9305
9306 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
9307 0 (which usually indicates a closed connection), but continue reading.
9308 [Bodo Moeller]
9309
9310 *) Fix some race conditions.
9311 [Bodo Moeller]
9312
9313 *) Add support for CRL distribution points extension. Add Certificate
9314 Policies and CRL distribution points documentation.
9315 [Steve Henson]
9316
9317 *) Move the autogenerated header file parts to crypto/opensslconf.h.
9318 [Ulf Möller]
9319
9320 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
9321 8 of keying material. Merlin has also confirmed interop with this fix
9322 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
9323 [Merlin Hughes <merlin@baltimore.ie>]
9324
9325 *) Fix lots of warnings.
9326 [Richard Levitte <levitte@stacken.kth.se>]
9327
9328 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
9329 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
9330 [Richard Levitte <levitte@stacken.kth.se>]
9331
9332 *) Fix problems with sizeof(long) == 8.
9333 [Andy Polyakov <appro@fy.chalmers.se>]
9334
9335 *) Change functions to ANSI C.
9336 [Ulf Möller]
9337
9338 *) Fix typos in error codes.
9339 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
9340
9341 *) Remove defunct assembler files from Configure.
9342 [Ulf Möller]
9343
9344 *) SPARC v8 assembler BIGNUM implementation.
9345 [Andy Polyakov <appro@fy.chalmers.se>]
9346
9347 *) Support for Certificate Policies extension: both print and set.
9348 Various additions to support the r2i method this uses.
9349 [Steve Henson]
9350
9351 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
9352 return a const string when you are expecting an allocated buffer.
9353 [Ben Laurie]
9354
9355 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
9356 types DirectoryString and DisplayText.
9357 [Steve Henson]
9358
9359 *) Add code to allow r2i extensions to access the configuration database,
9360 add an LHASH database driver and add several ctx helper functions.
9361 [Steve Henson]
9362
9363 *) Fix an evil bug in bn_expand2() which caused various BN functions to
9364 fail when they extended the size of a BIGNUM.
9365 [Steve Henson]
9366
9367 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
9368 support typesafe stack.
9369 [Steve Henson]
9370
9371 *) Fix typo in SSL_[gs]et_options().
9372 [Nils Frostberg <nils@medcom.se>]
9373
9374 *) Delete various functions and files that belonged to the (now obsolete)
9375 old X509V3 handling code.
9376 [Steve Henson]
9377
9378 *) New Configure option "rsaref".
9379 [Ulf Möller]
9380
9381 *) Don't auto-generate pem.h.
9382 [Bodo Moeller]
9383
9384 *) Introduce type-safe ASN.1 SETs.
9385 [Ben Laurie]
9386
9387 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
9388 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
9389
9390 *) Introduce type-safe STACKs. This will almost certainly break lots of code
9391 that links with OpenSSL (well at least cause lots of warnings), but fear
9392 not: the conversion is trivial, and it eliminates loads of evil casts. A
9393 few STACKed things have been converted already. Feel free to convert more.
9394 In the fullness of time, I'll do away with the STACK type altogether.
9395 [Ben Laurie]
9396
9397 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
9398 specified in <certfile> by updating the entry in the index.txt file.
9399 This way one no longer has to edit the index.txt file manually for
9400 revoking a certificate. The -revoke option does the gory details now.
9401 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
9402
9403 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
9404 `-text' option at all and this way the `-noout -text' combination was
9405 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
9406 [Ralf S. Engelschall]
9407
9408 *) Make sure a corresponding plain text error message exists for the
9409 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
9410 verify callback function determined that a certificate was revoked.
9411 [Ralf S. Engelschall]
9412
9413 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9414 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9415 all available cipers including rc5, which was forgotten until now.
9416 In order to let the testing shell script know which algorithms
9417 are available, a new (up to now undocumented) command
9418 "openssl list-cipher-commands" is used.
9419 [Bodo Moeller]
9420
9421 *) Bugfix: s_client occasionally would sleep in select() when
9422 it should have checked SSL_pending() first.
9423 [Bodo Moeller]
9424
9425 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9426 the raw DSA values prior to ASN.1 encoding.
9427 [Ulf Möller]
9428
9429 *) Tweaks to Configure
9430 [Niels Poppe <niels@netbox.org>]
9431
9432 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9433 yet...
9434 [Steve Henson]
9435
9436 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9437 [Ulf Möller]
9438
9439 *) New config option to avoid instructions that are illegal on the 80386.
9440 The default code is faster, but requires at least a 486.
9441 [Ulf Möller]
9442
9443 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9444 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9445 same as SSL2_VERSION anyway.
9446 [Bodo Moeller]
9447
9448 *) New "-showcerts" option for s_client.
9449 [Bodo Moeller]
9450
9451 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9452 application. Various cleanups and fixes.
9453 [Steve Henson]
9454
9455 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9456 modify error routines to work internally. Add error codes and PBE init
9457 to library startup routines.
9458 [Steve Henson]
9459
9460 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9461 packing functions to asn1 and evp. Changed function names and error
9462 codes along the way.
9463 [Steve Henson]
9464
9465 *) PKCS12 integration: and so it begins... First of several patches to
9466 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9467 objects to objects.h
9468 [Steve Henson]
9469
9470 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9471 and display support for Thawte strong extranet extension.
9472 [Steve Henson]
9473
9474 *) Add LinuxPPC support.
9475 [Jeff Dubrule <igor@pobox.org>]
9476
9477 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9478 bn_div_words in alpha.s.
9479 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9480
9481 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9482 OAEP isn't supported when OpenSSL is built with RSAref.
9483 [Ulf Moeller <ulf@fitug.de>]
9484
9485 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9486 so they no longer are missing under -DNOPROTO.
9487 [Soren S. Jorvang <soren@t.dk>]
9488
9489
9490 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9491
9492 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9493 doesn't work when the session is reused. Coming soon!
9494 [Ben Laurie]
9495
9496 *) Fix a security hole, that allows sessions to be reused in the wrong
9497 context thus bypassing client cert protection! All software that uses
9498 client certs and session caches in multiple contexts NEEDS PATCHING to
9499 allow session reuse! A fuller solution is in the works.
9500 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9501
9502 *) Some more source tree cleanups (removed obsolete files
9503 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9504 permission on "config" script to be executable) and a fix for the INSTALL
9505 document.
9506 [Ulf Moeller <ulf@fitug.de>]
9507
9508 *) Remove some legacy and erroneous uses of malloc, free instead of
9509 Malloc, Free.
9510 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9511
9512 *) Make rsa_oaep_test return non-zero on error.
9513 [Ulf Moeller <ulf@fitug.de>]
9514
9515 *) Add support for native Solaris shared libraries. Configure
9516 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9517 if someone would make that last step automatic.
9518 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9519
9520 *) ctx_size was not built with the right compiler during "make links". Fixed.
9521 [Ben Laurie]
9522
9523 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9524 except NULL ciphers". This means the default cipher list will no longer
9525 enable NULL ciphers. They need to be specifically enabled e.g. with
9526 the string "DEFAULT:eNULL".
9527 [Steve Henson]
9528
9529 *) Fix to RSA private encryption routines: if p < q then it would
9530 occasionally produce an invalid result. This will only happen with
9531 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9532 [Steve Henson]
9533
9534 *) Be less restrictive and allow also `perl util/perlpath.pl
9535 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9536 because this way one can also use an interpreter named `perl5' (which is
9537 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9538 installed as `perl').
9539 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9540
9541 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9542 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9543
9544 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9545 advapi32.lib to Win32 build and change the pem test comparision
9546 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9547 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9548 and crypto/des/ede_cbcm_enc.c.
9549 [Steve Henson]
9550
9551 *) DES quad checksum was broken on big-endian architectures. Fixed.
9552 [Ben Laurie]
9553
9554 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9555 Win32 test batch file so it (might) work again. The Win32 test batch file
9556 is horrible: I feel ill....
9557 [Steve Henson]
9558
9559 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9560 in e_os.h. Audit of header files to check ANSI and non ANSI
9561 sections: 10 functions were absent from non ANSI section and not exported
9562 from Windows DLLs. Fixed up libeay.num for new functions.
9563 [Steve Henson]
9564
9565 *) Make `openssl version' output lines consistent.
9566 [Ralf S. Engelschall]
9567
9568 *) Fix Win32 symbol export lists for BIO functions: Added
9569 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9570 to ms/libeay{16,32}.def.
9571 [Ralf S. Engelschall]
9572
9573 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9574 fine under Unix and passes some trivial tests I've now added. But the
9575 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9576 added to make sure no one expects that this stuff really works in the
9577 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9578 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9579 openssl_bio.xs.
9580 [Ralf S. Engelschall]
9581
9582 *) Fix the generation of two part addresses in perl.
9583 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9584
9585 *) Add config entry for Linux on MIPS.
9586 [John Tobey <jtobey@channel1.com>]
9587
9588 *) Make links whenever Configure is run, unless we are on Windoze.
9589 [Ben Laurie]
9590
9591 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9592 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9593 in CRLs.
9594 [Steve Henson]
9595
9596 *) Add a useful kludge to allow package maintainers to specify compiler and
9597 other platforms details on the command line without having to patch the
9598 Configure script everytime: One now can use ``perl Configure
9599 <id>:<details>'', i.e. platform ids are allowed to have details appended
9600 to them (seperated by colons). This is treated as there would be a static
9601 pre-configured entry in Configure's %table under key <id> with value
9602 <details> and ``perl Configure <id>'' is called. So, when you want to
9603 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9604 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9605 now, which overrides the FreeBSD-elf entry on-the-fly.
9606 [Ralf S. Engelschall]
9607
9608 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9609 [Ben Laurie]
9610
9611 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9612 on the `perl Configure ...' command line. This way one can compile
9613 OpenSSL libraries with Position Independent Code (PIC) which is needed
9614 for linking it into DSOs.
9615 [Ralf S. Engelschall]
9616
9617 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9618 Fixed.
9619 [Ben Laurie]
9620
9621 *) Cleaned up the LICENSE document: The official contact for any license
9622 questions now is the OpenSSL core team under openssl-core@openssl.org.
9623 And add a paragraph about the dual-license situation to make sure people
9624 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9625 to the OpenSSL toolkit.
9626 [Ralf S. Engelschall]
9627
9628 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9629 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9630 Additonally cleaned up the `make links' target: Remove unnecessary
9631 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9632 to speed processing and no longer clutter the display with confusing
9633 stuff. Instead only the actually done links are displayed.
9634 [Ralf S. Engelschall]
9635
9636 *) Permit null encryption ciphersuites, used for authentication only. It used
9637 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9638 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9639 encryption.
9640 [Ben Laurie]
9641
9642 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9643 signed attributes when verifying signatures (this would break them),
9644 the detached data encoding was wrong and public keys obtained using
9645 X509_get_pubkey() weren't freed.
9646 [Steve Henson]
9647
9648 *) Add text documentation for the BUFFER functions. Also added a work around
9649 to a Win95 console bug. This was triggered by the password read stuff: the
9650 last character typed gets carried over to the next fread(). If you were
9651 generating a new cert request using 'req' for example then the last
9652 character of the passphrase would be CR which would then enter the first
9653 field as blank.
9654 [Steve Henson]
9655
9656 *) Added the new `Includes OpenSSL Cryptography Software' button as
9657 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9658 button and can be used by applications based on OpenSSL to show the
9659 relationship to the OpenSSL project.
9660 [Ralf S. Engelschall]
9661
9662 *) Remove confusing variables in function signatures in files
9663 ssl/ssl_lib.c and ssl/ssl.h.
9664 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9665
9666 *) Don't install bss_file.c under PREFIX/include/
9667 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9668
9669 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9670 functions that return function pointers and has support for NT specific
9671 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9672 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9673 unsigned to signed types: this was killing the Win32 compile.
9674 [Steve Henson]
9675
9676 *) Add new certificate file to stack functions,
9677 SSL_add_dir_cert_subjects_to_stack() and
9678 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9679 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9680 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9681 This means that Apache-SSL and similar packages don't have to mess around
9682 to add as many CAs as they want to the preferred list.
9683 [Ben Laurie]
9684
9685 *) Experiment with doxygen documentation. Currently only partially applied to
9686 ssl/ssl_lib.c.
9687 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9688 openssl.doxy as the configuration file.
9689 [Ben Laurie]
9690
9691 *) Get rid of remaining C++-style comments which strict C compilers hate.
9692 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9693
9694 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9695 compiled in by default: it has problems with large keys.
9696 [Steve Henson]
9697
9698 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9699 DH private keys and/or callback functions which directly correspond to
9700 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9701 is needed for applications which have to configure certificates on a
9702 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9703 (e.g. s_server).
9704 For the RSA certificate situation is makes no difference, but
9705 for the DSA certificate situation this fixes the "no shared cipher"
9706 problem where the OpenSSL cipher selection procedure failed because the
9707 temporary keys were not overtaken from the context and the API provided
9708 no way to reconfigure them.
9709 The new functions now let applications reconfigure the stuff and they
9710 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9711 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9712 non-public-API function ssl_cert_instantiate() is used as a helper
9713 function and also to reduce code redundancy inside ssl_rsa.c.
9714 [Ralf S. Engelschall]
9715
9716 *) Move s_server -dcert and -dkey options out of the undocumented feature
9717 area because they are useful for the DSA situation and should be
9718 recognized by the users.
9719 [Ralf S. Engelschall]
9720
9721 *) Fix the cipher decision scheme for export ciphers: the export bits are
9722 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9723 SSL_EXP_MASK. So, the original variable has to be used instead of the
9724 already masked variable.
9725 [Richard Levitte <levitte@stacken.kth.se>]
9726
9727 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9728 [Richard Levitte <levitte@stacken.kth.se>]
9729
9730 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9731 from `int' to `unsigned int' because it's a length and initialized by
9732 EVP_DigestFinal() which expects an `unsigned int *'.
9733 [Richard Levitte <levitte@stacken.kth.se>]
9734
9735 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9736 script. Instead use the usual Shell->Perl transition trick.
9737 [Ralf S. Engelschall]
9738
9739 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9740 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9741 -noout -modulus' as it's already the case for `openssl rsa -noout
9742 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9743 currently the public key is printed (a decision which was already done by
9744 `openssl dsa -modulus' in the past) which serves a similar purpose.
9745 Additionally the NO_RSA no longer completely removes the whole -modulus
9746 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9747 now, too.
9748 [Ralf S. Engelschall]
9749
9750 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9751 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9752 [Arne Ansper <arne@ats.cyber.ee>]
9753
9754 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9755 to be added. Now both 'req' and 'ca' can use new objects defined in the
9756 config file.
9757 [Steve Henson]
9758
9759 *) Add cool BIO that does syslog (or event log on NT).
9760 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9761
9762 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9763 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9764 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9765 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9766 [Ben Laurie]
9767
9768 *) Add preliminary config info for new extension code.
9769 [Steve Henson]
9770
9771 *) Make RSA_NO_PADDING really use no padding.
9772 [Ulf Moeller <ulf@fitug.de>]
9773
9774 *) Generate errors when private/public key check is done.
9775 [Ben Laurie]
9776
9777 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9778 for some CRL extensions and new objects added.
9779 [Steve Henson]
9780
9781 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9782 key usage extension and fuller support for authority key id.
9783 [Steve Henson]
9784
9785 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9786 padding method for RSA, which is recommended for new applications in PKCS
9787 #1 v2.0 (RFC 2437, October 1998).
9788 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9789 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9790 against Bleichbacher's attack on RSA.
9791 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9792 Ben Laurie]
9793
9794 *) Updates to the new SSL compression code
9795 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9796
9797 *) Fix so that the version number in the master secret, when passed
9798 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9799 (because the server will not accept higher), that the version number
9800 is 0x03,0x01, not 0x03,0x00
9801 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9802
9803 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9804 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9805 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9806 [Steve Henson]
9807
9808 *) Support for RAW extensions where an arbitrary extension can be
9809 created by including its DER encoding. See apps/openssl.cnf for
9810 an example.
9811 [Steve Henson]
9812
9813 *) Make sure latest Perl versions don't interpret some generated C array
9814 code as Perl array code in the crypto/err/err_genc.pl script.
9815 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9816
9817 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9818 not many people have the assembler. Various Win32 compilation fixes and
9819 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9820 build instructions.
9821 [Steve Henson]
9822
9823 *) Modify configure script 'Configure' to automatically create crypto/date.h
9824 file under Win32 and also build pem.h from pem.org. New script
9825 util/mkfiles.pl to create the MINFO file on environments that can't do a
9826 'make files': perl util/mkfiles.pl >MINFO should work.
9827 [Steve Henson]
9828
9829 *) Major rework of DES function declarations, in the pursuit of correctness
9830 and purity. As a result, many evil casts evaporated, and some weirdness,
9831 too. You may find this causes warnings in your code. Zapping your evil
9832 casts will probably fix them. Mostly.
9833 [Ben Laurie]
9834
9835 *) Fix for a typo in asn1.h. Bug fix to object creation script
9836 obj_dat.pl. It considered a zero in an object definition to mean
9837 "end of object": none of the objects in objects.h have any zeros
9838 so it wasn't spotted.
9839 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9840
9841 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9842 Masking (CBCM). In the absence of test vectors, the best I have been able
9843 to do is check that the decrypt undoes the encrypt, so far. Send me test
9844 vectors if you have them.
9845 [Ben Laurie]
9846
9847 *) Correct calculation of key length for export ciphers (too much space was
9848 allocated for null ciphers). This has not been tested!
9849 [Ben Laurie]
9850
9851 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9852 message is now correct (it understands "crypto" and "ssl" on its
9853 command line). There is also now an "update" option. This will update
9854 the util/ssleay.num and util/libeay.num files with any new functions.
9855 If you do a:
9856 perl util/mkdef.pl crypto ssl update
9857 it will update them.
9858 [Steve Henson]
9859
9860 *) Overhauled the Perl interface (perl/*):
9861 - ported BN stuff to OpenSSL's different BN library
9862 - made the perl/ source tree CVS-aware
9863 - renamed the package from SSLeay to OpenSSL (the files still contain
9864 their history because I've copied them in the repository)
9865 - removed obsolete files (the test scripts will be replaced
9866 by better Test::Harness variants in the future)
9867 [Ralf S. Engelschall]
9868
9869 *) First cut for a very conservative source tree cleanup:
9870 1. merge various obsolete readme texts into doc/ssleay.txt
9871 where we collect the old documents and readme texts.
9872 2. remove the first part of files where I'm already sure that we no
9873 longer need them because of three reasons: either they are just temporary
9874 files which were left by Eric or they are preserved original files where
9875 I've verified that the diff is also available in the CVS via "cvs diff
9876 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9877 the crypto/md/ stuff).
9878 [Ralf S. Engelschall]
9879
9880 *) More extension code. Incomplete support for subject and issuer alt
9881 name, issuer and authority key id. Change the i2v function parameters
9882 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9883 what that's for :-) Fix to ASN1 macro which messed up
9884 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9885 [Steve Henson]
9886
9887 *) Preliminary support for ENUMERATED type. This is largely copied from the
9888 INTEGER code.
9889 [Steve Henson]
9890
9891 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9892 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9893
9894 *) Make sure `make rehash' target really finds the `openssl' program.
9895 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9896
9897 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9898 like to hear about it if this slows down other processors.
9899 [Ben Laurie]
9900
9901 *) Add CygWin32 platform information to Configure script.
9902 [Alan Batie <batie@aahz.jf.intel.com>]
9903
9904 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9905 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9906
9907 *) New program nseq to manipulate netscape certificate sequences
9908 [Steve Henson]
9909
9910 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9911 few typos.
9912 [Steve Henson]
9913
9914 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9915 but the BN code had some problems that would cause failures when
9916 doing certificate verification and some other functions.
9917 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9918
9919 *) Add ASN1 and PEM code to support netscape certificate sequences.
9920 [Steve Henson]
9921
9922 *) Add ASN1 and PEM code to support netscape certificate sequences.
9923 [Steve Henson]
9924
9925 *) Add several PKIX and private extended key usage OIDs.
9926 [Steve Henson]
9927
9928 *) Modify the 'ca' program to handle the new extension code. Modify
9929 openssl.cnf for new extension format, add comments.
9930 [Steve Henson]
9931
9932 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9933 and add a sample to openssl.cnf so req -x509 now adds appropriate
9934 CA extensions.
9935 [Steve Henson]
9936
9937 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9938 error code, add initial support to X509_print() and x509 application.
9939 [Steve Henson]
9940
9941 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9942 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9943 stuff is currently isolated and isn't even compiled yet.
9944 [Steve Henson]
9945
9946 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9947 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9948 Removed the versions check from X509 routines when loading extensions:
9949 this allows certain broken certificates that don't set the version
9950 properly to be processed.
9951 [Steve Henson]
9952
9953 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9954 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9955 can still be regenerated with "make depend".
9956 [Ben Laurie]
9957
9958 *) Spelling mistake in C version of CAST-128.
9959 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9960
9961 *) Changes to the error generation code. The perl script err-code.pl
9962 now reads in the old error codes and retains the old numbers, only
9963 adding new ones if necessary. It also only changes the .err files if new
9964 codes are added. The makefiles have been modified to only insert errors
9965 when needed (to avoid needlessly modifying header files). This is done
9966 by only inserting errors if the .err file is newer than the auto generated
9967 C file. To rebuild all the error codes from scratch (the old behaviour)
9968 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9969 or delete all the .err files.
9970 [Steve Henson]
9971
9972 *) CAST-128 was incorrectly implemented for short keys. The C version has
9973 been fixed, but is untested. The assembler versions are also fixed, but
9974 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9975 to regenerate it if needed.
9976 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9977 Hagino <itojun@kame.net>]
9978
9979 *) File was opened incorrectly in randfile.c.
9980 [Ulf Möller <ulf@fitug.de>]
9981
9982 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9983 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9984 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9985 al: it's just almost always a UTCTime. Note this patch adds new error
9986 codes so do a "make errors" if there are problems.
9987 [Steve Henson]
9988
9989 *) Correct Linux 1 recognition in config.
9990 [Ulf Möller <ulf@fitug.de>]
9991
9992 *) Remove pointless MD5 hash when using DSA keys in ca.
9993 [Anonymous <nobody@replay.com>]
9994
9995 *) Generate an error if given an empty string as a cert directory. Also
9996 generate an error if handed NULL (previously returned 0 to indicate an
9997 error, but didn't set one).
9998 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9999
10000 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10001 [Ben Laurie]
10002
10003 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10004 parameters. This was causing a warning which killed off the Win32 compile.
10005 [Steve Henson]
10006
10007 *) Remove C++ style comments from crypto/bn/bn_local.h.
10008 [Neil Costigan <neil.costigan@celocom.com>]
10009
10010 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10011 based on a text string, looking up short and long names and finally
10012 "dot" format. The "dot" format stuff didn't work. Added new function
10013 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10014 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10015 OID is not part of the table.
10016 [Steve Henson]
10017
10018 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10019 X509_LOOKUP_by_alias().
10020 [Ben Laurie]
10021
10022 *) Sort openssl functions by name.
10023 [Ben Laurie]
10024
10025 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10026 encryption from sample DSA keys (in case anyone is interested the password
10027 was "1234").
10028 [Steve Henson]
10029
10030 *) Make _all_ *_free functions accept a NULL pointer.
10031 [Frans Heymans <fheymans@isaserver.be>]
10032
10033 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10034 NULL pointers.
10035 [Anonymous <nobody@replay.com>]
10036
10037 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10038 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10039
10040 *) Don't blow it for numeric -newkey arguments to apps/req.
10041 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10042
10043 *) Temp key "for export" tests were wrong in s3_srvr.c.
10044 [Anonymous <nobody@replay.com>]
10045
10046 *) Add prototype for temp key callback functions
10047 SSL_CTX_set_tmp_{rsa,dh}_callback().
10048 [Ben Laurie]
10049
10050 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10051 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10052 [Steve Henson]
10053
10054 *) X509_name_add_entry() freed the wrong thing after an error.
10055 [Arne Ansper <arne@ats.cyber.ee>]
10056
10057 *) rsa_eay.c would attempt to free a NULL context.
10058 [Arne Ansper <arne@ats.cyber.ee>]
10059
10060 *) BIO_s_socket() had a broken should_retry() on Windoze.
10061 [Arne Ansper <arne@ats.cyber.ee>]
10062
10063 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10064 [Arne Ansper <arne@ats.cyber.ee>]
10065
10066 *) Make sure the already existing X509_STORE->depth variable is initialized
10067 in X509_STORE_new(), but document the fact that this variable is still
10068 unused in the certificate verification process.
10069 [Ralf S. Engelschall]
10070
10071 *) Fix the various library and apps files to free up pkeys obtained from
10072 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10073 [Steve Henson]
10074
10075 *) Fix reference counting in X509_PUBKEY_get(). This makes
10076 demos/maurice/example2.c work, amongst others, probably.
10077 [Steve Henson and Ben Laurie]
10078
10079 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10080 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10081 are no longer created. This way we have a single and consistent command
10082 line interface `openssl <command>', similar to `cvs <command>'.
10083 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10084
10085 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10086 BIT STRING wrapper always have zero unused bits.
10087 [Steve Henson]
10088
10089 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10090 [Steve Henson]
10091
10092 *) Make the top-level INSTALL documentation easier to understand.
10093 [Paul Sutton]
10094
10095 *) Makefiles updated to exit if an error occurs in a sub-directory
10096 make (including if user presses ^C) [Paul Sutton]
10097
10098 *) Make Montgomery context stuff explicit in RSA data structure.
10099 [Ben Laurie]
10100
10101 *) Fix build order of pem and err to allow for generated pem.h.
10102 [Ben Laurie]
10103
10104 *) Fix renumbering bug in X509_NAME_delete_entry().
10105 [Ben Laurie]
10106
10107 *) Enhanced the err-ins.pl script so it makes the error library number
10108 global and can add a library name. This is needed for external ASN1 and
10109 other error libraries.
10110 [Steve Henson]
10111
10112 *) Fixed sk_insert which never worked properly.
10113 [Steve Henson]
10114
10115 *) Fix ASN1 macros so they can handle indefinite length construted
10116 EXPLICIT tags. Some non standard certificates use these: they can now
10117 be read in.
10118 [Steve Henson]
10119
10120 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10121 into a single doc/ssleay.txt bundle. This way the information is still
10122 preserved but no longer messes up this directory. Now it's new room for
10123 the new set of documenation files.
10124 [Ralf S. Engelschall]
10125
10126 *) SETs were incorrectly DER encoded. This was a major pain, because they
10127 shared code with SEQUENCEs, which aren't coded the same. This means that
10128 almost everything to do with SETs or SEQUENCEs has either changed name or
10129 number of arguments.
10130 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10131
10132 *) Fix test data to work with the above.
10133 [Ben Laurie]
10134
10135 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10136 was already fixed by Eric for 0.9.1 it seems.
10137 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10138
10139 *) Autodetect FreeBSD3.
10140 [Ben Laurie]
10141
10142 *) Fix various bugs in Configure. This affects the following platforms:
10143 nextstep
10144 ncr-scde
10145 unixware-2.0
10146 unixware-2.0-pentium
10147 sco5-cc.
10148 [Ben Laurie]
10149
10150 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10151 before they are needed.
10152 [Ben Laurie]
10153
10154 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10155 [Ben Laurie]
10156
10157
10158 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10159
10160 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10161 changed SSLeay to OpenSSL in version strings.
10162 [Ralf S. Engelschall]
10163
10164 *) Some fixups to the top-level documents.
10165 [Paul Sutton]
10166
10167 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10168 because the symlink to include/ was missing.
10169 [Ralf S. Engelschall]
10170
10171 *) Incorporated the popular no-RSA/DSA-only patches
10172 which allow to compile a RSA-free SSLeay.
10173 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10174
10175 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10176 when "ssleay" is still not found.
10177 [Ralf S. Engelschall]
10178
10179 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10180 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10181
10182 *) Updated the README file.
10183 [Ralf S. Engelschall]
10184
10185 *) Added various .cvsignore files in the CVS repository subdirs
10186 to make a "cvs update" really silent.
10187 [Ralf S. Engelschall]
10188
10189 *) Recompiled the error-definition header files and added
10190 missing symbols to the Win32 linker tables.
10191 [Ralf S. Engelschall]
10192
10193 *) Cleaned up the top-level documents;
10194 o new files: CHANGES and LICENSE
10195 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10196 o merged COPYRIGHT into LICENSE
10197 o removed obsolete TODO file
10198 o renamed MICROSOFT to INSTALL.W32
10199 [Ralf S. Engelschall]
10200
10201 *) Removed dummy files from the 0.9.1b source tree:
10202 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10203 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10204 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10205 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10206 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10207 [Ralf S. Engelschall]
10208
10209 *) Added various platform portability fixes.
10210 [Mark J. Cox]
10211
10212 *) The Genesis of the OpenSSL rpject:
10213 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10214 Young and Tim J. Hudson created while they were working for C2Net until
10215 summer 1998.
10216 [The OpenSSL Project]
10217
10218
10219 Changes between 0.9.0b and 0.9.1b [not released]
10220
10221 *) Updated a few CA certificates under certs/
10222 [Eric A. Young]
10223
10224 *) Changed some BIGNUM api stuff.
10225 [Eric A. Young]
10226
10227 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10228 DGUX x86, Linux Alpha, etc.
10229 [Eric A. Young]
10230
10231 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10232 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10233 available).
10234 [Eric A. Young]
10235
10236 *) Add -strparse option to asn1pars program which parses nested
10237 binary structures
10238 [Dr Stephen Henson <shenson@bigfoot.com>]
10239
10240 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10241 [Eric A. Young]
10242
10243 *) DSA fix for "ca" program.
10244 [Eric A. Young]
10245
10246 *) Added "-genkey" option to "dsaparam" program.
10247 [Eric A. Young]
10248
10249 *) Added RIPE MD160 (rmd160) message digest.
10250 [Eric A. Young]
10251
10252 *) Added -a (all) option to "ssleay version" command.
10253 [Eric A. Young]
10254
10255 *) Added PLATFORM define which is the id given to Configure.
10256 [Eric A. Young]
10257
10258 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10259 [Eric A. Young]
10260
10261 *) Extended the ASN.1 parser routines.
10262 [Eric A. Young]
10263
10264 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
10265 [Eric A. Young]
10266
10267 *) Added a BN_CTX to the BN library.
10268 [Eric A. Young]
10269
10270 *) Fixed the weak key values in DES library
10271 [Eric A. Young]
10272
10273 *) Changed API in EVP library for cipher aliases.
10274 [Eric A. Young]
10275
10276 *) Added support for RC2/64bit cipher.
10277 [Eric A. Young]
10278
10279 *) Converted the lhash library to the crypto/mem.c functions.
10280 [Eric A. Young]
10281
10282 *) Added more recognized ASN.1 object ids.
10283 [Eric A. Young]
10284
10285 *) Added more RSA padding checks for SSL/TLS.
10286 [Eric A. Young]
10287
10288 *) Added BIO proxy/filter functionality.
10289 [Eric A. Young]
10290
10291 *) Added extra_certs to SSL_CTX which can be used
10292 send extra CA certificates to the client in the CA cert chain sending
10293 process. It can be configured with SSL_CTX_add_extra_chain_cert().
10294 [Eric A. Young]
10295
10296 *) Now Fortezza is denied in the authentication phase because
10297 this is key exchange mechanism is not supported by SSLeay at all.
10298 [Eric A. Young]
10299
10300 *) Additional PKCS1 checks.
10301 [Eric A. Young]
10302
10303 *) Support the string "TLSv1" for all TLS v1 ciphers.
10304 [Eric A. Young]
10305
10306 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
10307 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
10308 [Eric A. Young]
10309
10310 *) Fixed a few memory leaks.
10311 [Eric A. Young]
10312
10313 *) Fixed various code and comment typos.
10314 [Eric A. Young]
10315
10316 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
10317 bytes sent in the client random.
10318 [Edward Bishop <ebishop@spyglass.com>]
10319