]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Fix d2i_SSL_SESSION.
[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 support for canonical generation of DSA parameter 'g'. See
8 FIPS 186-3 A.2.3.
9
10 *) Add support for HMAC DRBG from SP800-90. Update algorithm and POST
11 to handle HMAC cases.
12 [Steve Henson]
13
14 *) Add functions FIPS_module_version() and FIPS_module_version_text()
15 to return numberical and string versions of the FIPS module number.
16 [Steve Henson]
17
18 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
19 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implmeneted
20 outside the validated module in the FIPS capable OpenSSL.
21 [Steve Henson]
22
23 *) Minor change to DRBG entropy callback semantics. In some cases
24 there is no mutiple of the block length between min_len and
25 max_len. Allow the callback to return more than max_len bytes
26 of entropy but discard any extra: it is the callback's responsibility
27 to ensure that the extra data discarded does not impact the
28 requested amount of entropy.
29 [Steve Henson]
30
31 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
32 information in FIPS186-3, SP800-57 and SP800-131A.
33 [Steve Henson]
34
35 *) CCM support via EVP. Interface is very similar to GCM case except we
36 must supply all data in one chunk (i.e. no update, final) and the
37 message length must be supplied if AAD is used. Add algorithm test
38 support.
39 [Steve Henson]
40
41 *) Initial version of POST overhaul. Add POST callback to allow the status
42 of POST to be monitored and/or failures induced. Modify fips_test_suite
43 to use callback. Always run all selftests even if one fails.
44 [Steve Henson]
45
46 *) XTS support including algorithm test driver in the fips_gcmtest program.
47 Note: this does increase the maximum key length from 32 to 64 bytes but
48 there should be no binary compatibility issues as existing applications
49 will never use XTS mode.
50 [Steve Henson]
51
52 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
53 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
54 performs algorithm blocking for unapproved PRNG types. Also do not
55 set PRNG type in FIPS_mode_set(): leave this to the application.
56 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
57 the standard OpenSSL PRNG: set additional data to a date time vector.
58 [Steve Henson]
59
60 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
61 This shouldn't present any incompatibility problems because applications
62 shouldn't be using these directly and any that are will need to rethink
63 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
64 [Steve Henson]
65
66 *) Extensive self tests and health checking required by SP800-90 DRBG.
67 Remove strength parameter from FIPS_drbg_instantiate and always
68 instantiate at maximum supported strength.
69 [Steve Henson]
70
71 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
72 [Steve Henson]
73
74 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
75 [Steve Henson]
76
77 *) New function DH_compute_key_padded() to compute a DH key and pad with
78 leading zeroes if needed: this complies with SP800-56A et al.
79 [Steve Henson]
80
81 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
82 anything, incomplete, subject to change and largely untested at present.
83 [Steve Henson]
84
85 *) Modify fipscanisteronly build option to only build the necessary object
86 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
87 [Steve Henson]
88
89 *) Add experimental option FIPSSYMS to give all symbols in
90 fipscanister.o and FIPS or fips prefix. This will avoid
91 conflicts with future versions of OpenSSL. Add perl script
92 util/fipsas.pl to preprocess assembly language source files
93 and rename any affected symbols.
94 [Steve Henson]
95
96 *) Add selftest checks and algorithm block of non-fips algorithms in
97 FIPS mode. Remove DES2 from selftests.
98 [Steve Henson]
99
100 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
101 return internal method without any ENGINE dependencies. Add new
102 tiny fips sign and verify functions.
103 [Steve Henson]
104
105 *) New build option no-ec2m to disable characteristic 2 code.
106 [Steve Henson]
107
108 *) New build option "fipscanisteronly". This only builds fipscanister.o
109 and (currently) associated fips utilities. Uses the file Makefile.fips
110 instead of Makefile.org as the prototype.
111 [Steve Henson]
112
113 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
114 Update fips_gcmtest to use IV generator.
115 [Steve Henson]
116
117 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
118 setting output buffer to NULL. The *Final function must be
119 called although it will not retrieve any additional data. The tag
120 can be set or retrieved with a ctrl. The IV length is by default 12
121 bytes (96 bits) but can be set to an alternative value. If the IV
122 length exceeds the maximum IV length (currently 16 bytes) it cannot be
123 set before the key.
124 [Steve Henson]
125
126 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
127 underlying do_cipher function handles all cipher semantics itself
128 including padding and finalisation. This is useful if (for example)
129 an ENGINE cipher handles block padding itself. The behaviour of
130 do_cipher is subtly changed if this flag is set: the return value
131 is the number of characters written to the output buffer (zero is
132 no longer an error code) or a negative error code. Also if the
133 input buffer is NULL and length 0 finalisation should be performed.
134 [Steve Henson]
135
136 *) If a candidate issuer certificate is already part of the constructed
137 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
138 [Steve Henson]
139
140 *) Improve forward-security support: add functions
141
142 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
143 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
144
145 for use by SSL/TLS servers; the callback function will be called whenever a
146 new session is created, and gets to decide whether the session may be
147 cached to make it resumable (return 0) or not (return 1). (As by the
148 SSL/TLS protocol specifications, the session_id sent by the server will be
149 empty to indicate that the session is not resumable; also, the server will
150 not generate RFC 4507 (RFC 5077) session tickets.)
151
152 A simple reasonable callback implementation is to return is_forward_secure.
153 This parameter will be set to 1 or 0 depending on the ciphersuite selected
154 by the SSL/TLS server library, indicating whether it can provide forward
155 security.
156 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
157
158 *) Add Next Protocol Negotiation,
159 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
160 disabled with a no-npn flag to config or Configure. Code donated
161 by Google.
162 [Adam Langley <agl@google.com> and Ben Laurie]
163
164 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
165 all platforms. Move ssize_t definition from e_os.h to the public
166 header file e_os2.h as it now appears in public header file cms.h
167 [Steve Henson]
168
169 *) New function OPENSSL_gmtime_diff to find the difference in days
170 and seconds between two tm structures. This will be used to provide
171 additional functionality for ASN1_TIME.
172 [Steve Henson]
173
174 *) New -sigopt option to the ca, req and x509 utilities. Additional
175 signature parameters can be passed using this option and in
176 particular PSS.
177 [Steve Henson]
178
179 *) Add RSA PSS signing function. This will generate and set the
180 appropriate AlgorithmIdentifiers for PSS based on those in the
181 corresponding EVP_MD_CTX structure. No application support yet.
182 [Steve Henson]
183
184 *) Support for companion algorithm specific ASN1 signing routines.
185 New function ASN1_item_sign_ctx() signs a pre-initialised
186 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
187 the appropriate parameters.
188 [Steve Henson]
189
190 *) Add new algorithm specific ASN1 verification initialisation function
191 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
192 handling will be the same no matter what EVP_PKEY_METHOD is used.
193 Add a PSS handler to support verification of PSS signatures: checked
194 against a number of sample certificates.
195 [Steve Henson]
196
197 *) Add signature printing for PSS. Add PSS OIDs.
198 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
199
200 *) Add algorithm specific signature printing. An individual ASN1 method
201 can now print out signatures instead of the standard hex dump.
202
203 More complex signatures (e.g. PSS) can print out more meaningful
204 information. Include DSA version that prints out the signature
205 parameters r, s.
206 [Steve Henson]
207
208 *) Add -trusted_first option which attempts to find certificates in the
209 trusted store even if an untrusted chain is also supplied.
210 [Steve Henson]
211
212 *) Initial experimental support for explicitly trusted non-root CAs.
213 OpenSSL still tries to build a complete chain to a root but if an
214 intermediate CA has a trust setting included that is used. The first
215 setting is used: whether to trust or reject.
216 [Steve Henson]
217
218 *) New -verify_name option in command line utilities to set verification
219 parameters by name.
220 [Steve Henson]
221
222 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
223 Add CMAC pkey methods.
224 [Steve Henson]
225
226 *) Experiemental regnegotiation in s_server -www mode. If the client
227 browses /reneg connection is renegotiated. If /renegcert it is
228 renegotiated requesting a certificate.
229 [Steve Henson]
230
231 *) Add an "external" session cache for debugging purposes to s_server. This
232 should help trace issues which normally are only apparent in deployed
233 multi-process servers.
234 [Steve Henson]
235
236 *) Experiemental password based recipient info support for CMS library:
237 implementing RFC3211.
238 [Steve Henson]
239
240 *) Split password based encryption into PBES2 and PBKDF2 functions. This
241 neatly separates the code into cipher and PBE sections and is required
242 for some algorithms that split PBES2 into separate pieces (such as
243 password based CMS).
244 [Steve Henson]
245
246 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
247 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
248 BIO_set_cipher() and some obscure PEM functions were changed so they
249 can now return an error. The RAND changes required a change to the
250 RAND_METHOD structure.
251 [Steve Henson]
252
253 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
254 a gcc attribute to warn if the result of a function is ignored. This
255 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
256 whose return value is often ignored.
257 [Steve Henson]
258
259 Changes between 1.0.0e and 1.0.1 [xx XXX xxxx]
260
261 *) Fix PSK session representation.
262 [Bodo Moeller]
263
264 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
265
266 This work was sponsored by Intel.
267 [Andy Polyakov]
268
269 *) Add GCM support to TLS library. Some custom code is needed to split
270 the IV between the fixed (from PRF) and explicit (from TLS record)
271 portions. This adds all GCM ciphersuites supported by RFC5288 and
272 RFC5289. Generalise some AES* cipherstrings to inlclude GCM and
273 add a special AESGCM string for GCM only.
274 [Steve Henson]
275
276 *) Expand range of ctrls for AES GCM. Permit setting invocation
277 field on decrypt and retrieval of invocation field only on encrypt.
278 [Steve Henson]
279
280 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
281 As required by RFC5289 these ciphersuites cannot be used if for
282 versions of TLS earlier than 1.2.
283 [Steve Henson]
284
285 *) For FIPS capable OpenSSL interpret a NULL default public key method
286 as unset and return the appopriate default but do *not* set the default.
287 This means we can return the appopriate method in applications that
288 swicth between FIPS and non-FIPS modes.
289 [Steve Henson]
290
291 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
292 ENGINE is used then we cannot handle that in the FIPS module so we
293 keep original code iff non-FIPS operations are allowed.
294 [Steve Henson]
295
296 *) Add -attime option to openssl verify.
297 [Peter Eckersley <pde@eff.org> and Ben Laurie]
298
299 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
300 [Steve Henson]
301
302 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
303 FIPS EC methods unconditionally for now.
304 [Steve Henson]
305
306 *) New build option no-ec2m to disable characteristic 2 code.
307 [Steve Henson]
308
309 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
310 all cases can be covered as some introduce binary incompatibilities.
311 [Steve Henson]
312
313 *) Redirect RSA operations to FIPS module including keygen,
314 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
315 [Steve Henson]
316
317 *) Add similar low level API blocking to ciphers.
318 [Steve Henson]
319
320 *) Low level digest APIs are not approved in FIPS mode: any attempt
321 to use these will cause a fatal error. Applications that *really* want
322 to use them can use the private_* version instead.
323 [Steve Henson]
324
325 *) Redirect cipher operations to FIPS module for FIPS builds.
326 [Steve Henson]
327
328 *) Redirect digest operations to FIPS module for FIPS builds.
329 [Steve Henson]
330
331 *) Update build system to add "fips" flag which will link in fipscanister.o
332 for static and shared library builds embedding a signature if needed.
333 [Steve Henson]
334
335 *) Output TLS supported curves in preference order instead of numerical
336 order. This is currently hardcoded for the highest order curves first.
337 This should be configurable so applications can judge speed vs strength.
338 [Steve Henson]
339
340 *) Add TLS v1.2 server support for client authentication.
341 [Steve Henson]
342
343 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
344 and enable MD5.
345 [Steve Henson]
346
347 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
348 FIPS modules versions.
349 [Steve Henson]
350
351 *) Add TLS v1.2 client side support for client authentication. Keep cache
352 of handshake records longer as we don't know the hash algorithm to use
353 until after the certificate request message is received.
354 [Steve Henson]
355
356 *) Initial TLS v1.2 client support. Add a default signature algorithms
357 extension including all the algorithms we support. Parse new signature
358 format in client key exchange. Relax some ECC signing restrictions for
359 TLS v1.2 as indicated in RFC5246.
360 [Steve Henson]
361
362 *) Add server support for TLS v1.2 signature algorithms extension. Switch
363 to new signature format when needed using client digest preference.
364 All server ciphersuites should now work correctly in TLS v1.2. No client
365 support yet and no support for client certificates.
366 [Steve Henson]
367
368 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
369 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
370 ciphersuites. At present only RSA key exchange ciphersuites work with
371 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
372 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
373 and version checking.
374 [Steve Henson]
375
376 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
377 with this defined it will not be affected by any changes to ssl internal
378 structures. Add several utility functions to allow openssl application
379 to work with OPENSSL_NO_SSL_INTERN defined.
380 [Steve Henson]
381
382 *) Add SRP support.
383 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
384
385 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
386 [Steve Henson]
387
388 *) Add EC_GFp_nistp224_method(), a 64-bit optimized implementation for
389 elliptic curve NIST-P224 with constant-time single point multiplication on
390 typical inputs. EC_GROUP_new_by_curve_name() will automatically use this
391 (while EC_GROUP_new_curve_GFp() currently won't and prefers the more
392 flexible implementations).
393
394 The implementation requires support for the nonstandard type __uint128_t,
395 and so is disabled by default. To include this in your build of OpenSSL,
396 use -DEC_NISTP224_64_GCC_128 on the Configure (or config) command line,
397 and run "make depend" (or "make update").
398 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
399
400 *) Permit abbreviated handshakes when renegotiating using the function
401 SSL_renegotiate_abbreviated().
402 [Robin Seggelmann <seggelmann@fh-muenster.de>]
403
404 *) Add call to ENGINE_register_all_complete() to
405 ENGINE_load_builtin_engines(), so some implementations get used
406 automatically instead of needing explicit application support.
407 [Steve Henson]
408
409 *) Add support for TLS key exporter as described in RFC5705.
410 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
411
412 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
413 a few changes are required:
414
415 Add SSL_OP_NO_TLSv1_1 flag.
416 Add TLSv1_1 methods.
417 Update version checking logic to handle version 1.1.
418 Add explicit IV handling (ported from DTLS code).
419 Add command line options to s_client/s_server.
420 [Steve Henson]
421
422 Changes between 1.0.0d and 1.0.0e [xx XXX xxxx]
423
424 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
425 for multi-threaded use of ECDH.
426 [Adam Langley (Google)]
427
428 *) Fix x509_name_ex_d2i memory leak on bad inputs.
429 [Bodo Moeller]
430
431 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
432 signature public key algorithm by using OID xref utilities instead.
433 Before this you could only use some ECC ciphersuites with SHA1 only.
434 [Steve Henson]
435
436 *) Add protection against ECDSA timing attacks as mentioned in the paper
437 by Billy Bob Brumley and Nicola Tuveri, see:
438
439 http://eprint.iacr.org/2011/232.pdf
440
441 [Billy Bob Brumley and Nicola Tuveri]
442
443 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
444
445 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
446 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
447
448 *) Fix bug in string printing code: if *any* escaping is enabled we must
449 escape the escape character (backslash) or the resulting string is
450 ambiguous.
451 [Steve Henson]
452
453 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
454
455 *) Disable code workaround for ancient and obsolete Netscape browsers
456 and servers: an attacker can use it in a ciphersuite downgrade attack.
457 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
458 [Steve Henson]
459
460 *) Fixed J-PAKE implementation error, originally discovered by
461 Sebastien Martini, further info and confirmation from Stefan
462 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
463 [Ben Laurie]
464
465 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
466
467 *) Fix extension code to avoid race conditions which can result in a buffer
468 overrun vulnerability: resumed sessions must not be modified as they can
469 be shared by multiple threads. CVE-2010-3864
470 [Steve Henson]
471
472 *) Fix WIN32 build system to correctly link an ENGINE directory into
473 a DLL.
474 [Steve Henson]
475
476 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
477
478 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
479 (CVE-2010-1633)
480 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
481
482 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
483
484 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
485 context. The operation can be customised via the ctrl mechanism in
486 case ENGINEs want to include additional functionality.
487 [Steve Henson]
488
489 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
490 [Steve Henson]
491
492 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
493 output hashes compatible with older versions of OpenSSL.
494 [Willy Weisz <weisz@vcpc.univie.ac.at>]
495
496 *) Fix compression algorithm handling: if resuming a session use the
497 compression algorithm of the resumed session instead of determining
498 it from client hello again. Don't allow server to change algorithm.
499 [Steve Henson]
500
501 *) Add load_crls() function to apps tidying load_certs() too. Add option
502 to verify utility to allow additional CRLs to be included.
503 [Steve Henson]
504
505 *) Update OCSP request code to permit adding custom headers to the request:
506 some responders need this.
507 [Steve Henson]
508
509 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
510 correctly.
511 [Julia Lawall <julia@diku.dk>]
512
513 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
514 needlessly dereferenced structures, used obsolete functions and
515 didn't handle all updated verify codes correctly.
516 [Steve Henson]
517
518 *) Disable MD2 in the default configuration.
519 [Steve Henson]
520
521 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
522 indicate the initial BIO being pushed or popped. This makes it possible
523 to determine whether the BIO is the one explicitly called or as a result
524 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
525 it handles reference counts correctly and doesn't zero out the I/O bio
526 when it is not being explicitly popped. WARNING: applications which
527 included workarounds for the old buggy behaviour will need to be modified
528 or they could free up already freed BIOs.
529 [Steve Henson]
530
531 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
532 renaming to all platforms (within the 0.9.8 branch, this was
533 done conditionally on Netware platforms to avoid a name clash).
534 [Guenter <lists@gknw.net>]
535
536 *) Add ECDHE and PSK support to DTLS.
537 [Michael Tuexen <tuexen@fh-muenster.de>]
538
539 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
540 be used on C++.
541 [Steve Henson]
542
543 *) Add "missing" function EVP_MD_flags() (without this the only way to
544 retrieve a digest flags is by accessing the structure directly. Update
545 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
546 or cipher is registered as in the "from" argument. Print out all
547 registered digests in the dgst usage message instead of manually
548 attempting to work them out.
549 [Steve Henson]
550
551 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
552 this allows the use of compression and extensions. Change default cipher
553 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
554 by default unless an application cipher string requests it.
555 [Steve Henson]
556
557 *) Alter match criteria in PKCS12_parse(). It used to try to use local
558 key ids to find matching certificates and keys but some PKCS#12 files
559 don't follow the (somewhat unwritten) rules and this strategy fails.
560 Now just gather all certificates together and the first private key
561 then look for the first certificate that matches the key.
562 [Steve Henson]
563
564 *) Support use of registered digest and cipher names for dgst and cipher
565 commands instead of having to add each one as a special case. So now
566 you can do:
567
568 openssl sha256 foo
569
570 as well as:
571
572 openssl dgst -sha256 foo
573
574 and this works for ENGINE based algorithms too.
575
576 [Steve Henson]
577
578 *) Update Gost ENGINE to support parameter files.
579 [Victor B. Wagner <vitus@cryptocom.ru>]
580
581 *) Support GeneralizedTime in ca utility.
582 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
583
584 *) Enhance the hash format used for certificate directory links. The new
585 form uses the canonical encoding (meaning equivalent names will work
586 even if they aren't identical) and uses SHA1 instead of MD5. This form
587 is incompatible with the older format and as a result c_rehash should
588 be used to rebuild symbolic links.
589 [Steve Henson]
590
591 *) Make PKCS#8 the default write format for private keys, replacing the
592 traditional format. This form is standardised, more secure and doesn't
593 include an implicit MD5 dependency.
594 [Steve Henson]
595
596 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
597 committed to OpenSSL should pass this lot as a minimum.
598 [Steve Henson]
599
600 *) Add session ticket override functionality for use by EAP-FAST.
601 [Jouni Malinen <j@w1.fi>]
602
603 *) Modify HMAC functions to return a value. Since these can be implemented
604 in an ENGINE errors can occur.
605 [Steve Henson]
606
607 *) Type-checked OBJ_bsearch_ex.
608 [Ben Laurie]
609
610 *) Type-checked OBJ_bsearch. Also some constification necessitated
611 by type-checking. Still to come: TXT_DB, bsearch(?),
612 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
613 CONF_VALUE.
614 [Ben Laurie]
615
616 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
617 seconds to a tm structure directly, instead of going through OS
618 specific date routines. This avoids any issues with OS routines such
619 as the year 2038 bug. New *_adj() functions for ASN1 time structures
620 and X509_time_adj_ex() to cover the extended range. The existing
621 X509_time_adj() is still usable and will no longer have any date issues.
622 [Steve Henson]
623
624 *) Delta CRL support. New use deltas option which will attempt to locate
625 and search any appropriate delta CRLs available.
626
627 This work was sponsored by Google.
628 [Steve Henson]
629
630 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
631 code and add additional score elements. Validate alternate CRL paths
632 as part of the CRL checking and indicate a new error "CRL path validation
633 error" in this case. Applications wanting additional details can use
634 the verify callback and check the new "parent" field. If this is not
635 NULL CRL path validation is taking place. Existing applications wont
636 see this because it requires extended CRL support which is off by
637 default.
638
639 This work was sponsored by Google.
640 [Steve Henson]
641
642 *) Support for freshest CRL extension.
643
644 This work was sponsored by Google.
645 [Steve Henson]
646
647 *) Initial indirect CRL support. Currently only supported in the CRLs
648 passed directly and not via lookup. Process certificate issuer
649 CRL entry extension and lookup CRL entries by bother issuer name
650 and serial number. Check and process CRL issuer entry in IDP extension.
651
652 This work was sponsored by Google.
653 [Steve Henson]
654
655 *) Add support for distinct certificate and CRL paths. The CRL issuer
656 certificate is validated separately in this case. Only enabled if
657 an extended CRL support flag is set: this flag will enable additional
658 CRL functionality in future.
659
660 This work was sponsored by Google.
661 [Steve Henson]
662
663 *) Add support for policy mappings extension.
664
665 This work was sponsored by Google.
666 [Steve Henson]
667
668 *) Fixes to pathlength constraint, self issued certificate handling,
669 policy processing to align with RFC3280 and PKITS tests.
670
671 This work was sponsored by Google.
672 [Steve Henson]
673
674 *) Support for name constraints certificate extension. DN, email, DNS
675 and URI types are currently supported.
676
677 This work was sponsored by Google.
678 [Steve Henson]
679
680 *) To cater for systems that provide a pointer-based thread ID rather
681 than numeric, deprecate the current numeric thread ID mechanism and
682 replace it with a structure and associated callback type. This
683 mechanism allows a numeric "hash" to be extracted from a thread ID in
684 either case, and on platforms where pointers are larger than 'long',
685 mixing is done to help ensure the numeric 'hash' is usable even if it
686 can't be guaranteed unique. The default mechanism is to use "&errno"
687 as a pointer-based thread ID to distinguish between threads.
688
689 Applications that want to provide their own thread IDs should now use
690 CRYPTO_THREADID_set_callback() to register a callback that will call
691 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
692
693 Note that ERR_remove_state() is now deprecated, because it is tied
694 to the assumption that thread IDs are numeric. ERR_remove_state(0)
695 to free the current thread's error state should be replaced by
696 ERR_remove_thread_state(NULL).
697
698 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
699 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
700 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
701 application was previously providing a numeric thread callback that
702 was inappropriate for distinguishing threads, then uniqueness might
703 have been obtained with &errno that happened immediately in the
704 intermediate development versions of OpenSSL; this is no longer the
705 case, the numeric thread callback will now override the automatic use
706 of &errno.)
707 [Geoff Thorpe, with help from Bodo Moeller]
708
709 *) Initial support for different CRL issuing certificates. This covers a
710 simple case where the self issued certificates in the chain exist and
711 the real CRL issuer is higher in the existing chain.
712
713 This work was sponsored by Google.
714 [Steve Henson]
715
716 *) Removed effectively defunct crypto/store from the build.
717 [Ben Laurie]
718
719 *) Revamp of STACK to provide stronger type-checking. Still to come:
720 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
721 ASN1_STRING, CONF_VALUE.
722 [Ben Laurie]
723
724 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
725 RAM on SSL connections. This option can save about 34k per idle SSL.
726 [Nick Mathewson]
727
728 *) Revamp of LHASH to provide stronger type-checking. Still to come:
729 STACK, TXT_DB, bsearch, qsort.
730 [Ben Laurie]
731
732 *) Initial support for Cryptographic Message Syntax (aka CMS) based
733 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
734 support for data, signedData, compressedData, digestedData and
735 encryptedData, envelopedData types included. Scripts to check against
736 RFC4134 examples draft and interop and consistency checks of many
737 content types and variants.
738 [Steve Henson]
739
740 *) Add options to enc utility to support use of zlib compression BIO.
741 [Steve Henson]
742
743 *) Extend mk1mf to support importing of options and assembly language
744 files from Configure script, currently only included in VC-WIN32.
745 The assembly language rules can now optionally generate the source
746 files from the associated perl scripts.
747 [Steve Henson]
748
749 *) Implement remaining functionality needed to support GOST ciphersuites.
750 Interop testing has been performed using CryptoPro implementations.
751 [Victor B. Wagner <vitus@cryptocom.ru>]
752
753 *) s390x assembler pack.
754 [Andy Polyakov]
755
756 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
757 "family."
758 [Andy Polyakov]
759
760 *) Implement Opaque PRF Input TLS extension as specified in
761 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
762 official specification yet and no extension type assignment by
763 IANA exists, this extension (for now) will have to be explicitly
764 enabled when building OpenSSL by providing the extension number
765 to use. For example, specify an option
766
767 -DTLSEXT_TYPE_opaque_prf_input=0x9527
768
769 to the "config" or "Configure" script to enable the extension,
770 assuming extension number 0x9527 (which is a completely arbitrary
771 and unofficial assignment based on the MD5 hash of the Internet
772 Draft). Note that by doing so, you potentially lose
773 interoperability with other TLS implementations since these might
774 be using the same extension number for other purposes.
775
776 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
777 opaque PRF input value to use in the handshake. This will create
778 an interal copy of the length-'len' string at 'src', and will
779 return non-zero for success.
780
781 To get more control and flexibility, provide a callback function
782 by using
783
784 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
785 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
786
787 where
788
789 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
790 void *arg;
791
792 Callback function 'cb' will be called in handshakes, and is
793 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
794 Argument 'arg' is for application purposes (the value as given to
795 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
796 be provided to the callback function). The callback function
797 has to return non-zero to report success: usually 1 to use opaque
798 PRF input just if possible, or 2 to enforce use of the opaque PRF
799 input. In the latter case, the library will abort the handshake
800 if opaque PRF input is not successfully negotiated.
801
802 Arguments 'peerinput' and 'len' given to the callback function
803 will always be NULL and 0 in the case of a client. A server will
804 see the client's opaque PRF input through these variables if
805 available (NULL and 0 otherwise). Note that if the server
806 provides an opaque PRF input, the length must be the same as the
807 length of the client's opaque PRF input.
808
809 Note that the callback function will only be called when creating
810 a new session (session resumption can resume whatever was
811 previously negotiated), and will not be called in SSL 2.0
812 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
813 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
814 for applications that need to enforce opaque PRF input.
815
816 [Bodo Moeller]
817
818 *) Update ssl code to support digests other than SHA1+MD5 for handshake
819 MAC.
820
821 [Victor B. Wagner <vitus@cryptocom.ru>]
822
823 *) Add RFC4507 support to OpenSSL. This includes the corrections in
824 RFC4507bis. The encrypted ticket format is an encrypted encoded
825 SSL_SESSION structure, that way new session features are automatically
826 supported.
827
828 If a client application caches session in an SSL_SESSION structure
829 support is transparent because tickets are now stored in the encoded
830 SSL_SESSION.
831
832 The SSL_CTX structure automatically generates keys for ticket
833 protection in servers so again support should be possible
834 with no application modification.
835
836 If a client or server wishes to disable RFC4507 support then the option
837 SSL_OP_NO_TICKET can be set.
838
839 Add a TLS extension debugging callback to allow the contents of any client
840 or server extensions to be examined.
841
842 This work was sponsored by Google.
843 [Steve Henson]
844
845 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
846 OpenSSL should now compile cleanly on gcc 4.2
847 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
848
849 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
850 support including streaming MAC support: this is required for GOST
851 ciphersuite support.
852 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
853
854 *) Add option -stream to use PKCS#7 streaming in smime utility. New
855 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
856 to output in BER and PEM format.
857 [Steve Henson]
858
859 *) Experimental support for use of HMAC via EVP_PKEY interface. This
860 allows HMAC to be handled via the EVP_DigestSign*() interface. The
861 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
862 ENGINE support for HMAC keys which are unextractable. New -mac and
863 -macopt options to dgst utility.
864 [Steve Henson]
865
866 *) New option -sigopt to dgst utility. Update dgst to use
867 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
868 alternative signing paramaters such as X9.31 or PSS in the dgst
869 utility.
870 [Steve Henson]
871
872 *) Change ssl_cipher_apply_rule(), the internal function that does
873 the work each time a ciphersuite string requests enabling
874 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
875 removing ("!foo+bar") a class of ciphersuites: Now it maintains
876 the order of disabled ciphersuites such that those ciphersuites
877 that most recently went from enabled to disabled not only stay
878 in order with respect to each other, but also have higher priority
879 than other disabled ciphersuites the next time ciphersuites are
880 enabled again.
881
882 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
883 the same ciphersuites as with "HIGH" alone, but in a specific
884 order where the PSK ciphersuites come first (since they are the
885 most recently disabled ciphersuites when "HIGH" is parsed).
886
887 Also, change ssl_create_cipher_list() (using this new
888 funcionality) such that between otherwise identical
889 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
890 the default order.
891 [Bodo Moeller]
892
893 *) Change ssl_create_cipher_list() so that it automatically
894 arranges the ciphersuites in reasonable order before starting
895 to process the rule string. Thus, the definition for "DEFAULT"
896 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
897 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
898 This makes it much easier to arrive at a reasonable default order
899 in applications for which anonymous ciphers are OK (meaning
900 that you can't actually use DEFAULT).
901 [Bodo Moeller; suggested by Victor Duchovni]
902
903 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
904 processing) into multiple integers instead of setting
905 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
906 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
907 (These masks as well as the individual bit definitions are hidden
908 away into the non-exported interface ssl/ssl_locl.h, so this
909 change to the definition of the SSL_CIPHER structure shouldn't
910 affect applications.) This give us more bits for each of these
911 categories, so there is no longer a need to coagulate AES128 and
912 AES256 into a single algorithm bit, and to coagulate Camellia128
913 and Camellia256 into a single algorithm bit, which has led to all
914 kinds of kludges.
915
916 Thus, among other things, the kludge introduced in 0.9.7m and
917 0.9.8e for masking out AES256 independently of AES128 or masking
918 out Camellia256 independently of AES256 is not needed here in 0.9.9.
919
920 With the change, we also introduce new ciphersuite aliases that
921 so far were missing: "AES128", "AES256", "CAMELLIA128", and
922 "CAMELLIA256".
923 [Bodo Moeller]
924
925 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
926 Use the leftmost N bytes of the signature input if the input is
927 larger than the prime q (with N being the size in bytes of q).
928 [Nils Larsch]
929
930 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
931 it yet and it is largely untested.
932 [Steve Henson]
933
934 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
935 [Nils Larsch]
936
937 *) Initial incomplete changes to avoid need for function casts in OpenSSL
938 some compilers (gcc 4.2 and later) reject their use. Safestack is
939 reimplemented. Update ASN1 to avoid use of legacy functions.
940 [Steve Henson]
941
942 *) Win32/64 targets are linked with Winsock2.
943 [Andy Polyakov]
944
945 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
946 to external functions. This can be used to increase CRL handling
947 efficiency especially when CRLs are very large by (for example) storing
948 the CRL revoked certificates in a database.
949 [Steve Henson]
950
951 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
952 new CRLs added to a directory can be used. New command line option
953 -verify_return_error to s_client and s_server. This causes real errors
954 to be returned by the verify callback instead of carrying on no matter
955 what. This reflects the way a "real world" verify callback would behave.
956 [Steve Henson]
957
958 *) GOST engine, supporting several GOST algorithms and public key formats.
959 Kindly donated by Cryptocom.
960 [Cryptocom]
961
962 *) Partial support for Issuing Distribution Point CRL extension. CRLs
963 partitioned by DP are handled but no indirect CRL or reason partitioning
964 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
965 selected via a scoring technique which handles IDP and AKID in CRLs.
966 [Steve Henson]
967
968 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
969 will ultimately be used for all verify operations: this will remove the
970 X509_STORE dependency on certificate verification and allow alternative
971 lookup methods. X509_STORE based implementations of these two callbacks.
972 [Steve Henson]
973
974 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
975 Modify get_crl() to find a valid (unexpired) CRL if possible.
976 [Steve Henson]
977
978 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
979 this would be called X509_CRL_cmp() but that name is already used by
980 a function that just compares CRL issuer names. Cache several CRL
981 extensions in X509_CRL structure and cache CRLDP in X509.
982 [Steve Henson]
983
984 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
985 this maps equivalent X509_NAME structures into a consistent structure.
986 Name comparison can then be performed rapidly using memcmp().
987 [Steve Henson]
988
989 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
990 utility.
991 [Steve Henson]
992
993 *) Allow digests to supply their own micalg string for S/MIME type using
994 the ctrl EVP_MD_CTRL_MICALG.
995 [Steve Henson]
996
997 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
998 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
999 ctrl. It can then customise the structure before and/or after signing
1000 if necessary.
1001 [Steve Henson]
1002
1003 *) New function OBJ_add_sigid() to allow application defined signature OIDs
1004 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
1005 to free up any added signature OIDs.
1006 [Steve Henson]
1007
1008 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
1009 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
1010 digest and cipher tables. New options added to openssl utility:
1011 list-message-digest-algorithms and list-cipher-algorithms.
1012 [Steve Henson]
1013
1014 *) Change the array representation of binary polynomials: the list
1015 of degrees of non-zero coefficients is now terminated with -1.
1016 Previously it was terminated with 0, which was also part of the
1017 value; thus, the array representation was not applicable to
1018 polynomials where t^0 has coefficient zero. This change makes
1019 the array representation useful in a more general context.
1020 [Douglas Stebila]
1021
1022 *) Various modifications and fixes to SSL/TLS cipher string
1023 handling. For ECC, the code now distinguishes between fixed ECDH
1024 with RSA certificates on the one hand and with ECDSA certificates
1025 on the other hand, since these are separate ciphersuites. The
1026 unused code for Fortezza ciphersuites has been removed.
1027
1028 For consistency with EDH, ephemeral ECDH is now called "EECDH"
1029 (not "ECDHE"). For consistency with the code for DH
1030 certificates, use of ECDH certificates is now considered ECDH
1031 authentication, not RSA or ECDSA authentication (the latter is
1032 merely the CA's signing algorithm and not actively used in the
1033 protocol).
1034
1035 The temporary ciphersuite alias "ECCdraft" is no longer
1036 available, and ECC ciphersuites are no longer excluded from "ALL"
1037 and "DEFAULT". The following aliases now exist for RFC 4492
1038 ciphersuites, most of these by analogy with the DH case:
1039
1040 kECDHr - ECDH cert, signed with RSA
1041 kECDHe - ECDH cert, signed with ECDSA
1042 kECDH - ECDH cert (signed with either RSA or ECDSA)
1043 kEECDH - ephemeral ECDH
1044 ECDH - ECDH cert or ephemeral ECDH
1045
1046 aECDH - ECDH cert
1047 aECDSA - ECDSA cert
1048 ECDSA - ECDSA cert
1049
1050 AECDH - anonymous ECDH
1051 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
1052
1053 [Bodo Moeller]
1054
1055 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
1056 Use correct micalg parameters depending on digest(s) in signed message.
1057 [Steve Henson]
1058
1059 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
1060 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
1061 [Steve Henson]
1062
1063 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
1064 an engine to register a method. Add ENGINE lookups for methods and
1065 functional reference processing.
1066 [Steve Henson]
1067
1068 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
1069 EVP_{Sign,Verify}* which allow an application to customise the signature
1070 process.
1071 [Steve Henson]
1072
1073 *) New -resign option to smime utility. This adds one or more signers
1074 to an existing PKCS#7 signedData structure. Also -md option to use an
1075 alternative message digest algorithm for signing.
1076 [Steve Henson]
1077
1078 *) Tidy up PKCS#7 routines and add new functions to make it easier to
1079 create PKCS7 structures containing multiple signers. Update smime
1080 application to support multiple signers.
1081 [Steve Henson]
1082
1083 *) New -macalg option to pkcs12 utility to allow setting of an alternative
1084 digest MAC.
1085 [Steve Henson]
1086
1087 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
1088 Reorganize PBE internals to lookup from a static table using NIDs,
1089 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
1090 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
1091 PRF which will be automatically used with PBES2.
1092 [Steve Henson]
1093
1094 *) Replace the algorithm specific calls to generate keys in "req" with the
1095 new API.
1096 [Steve Henson]
1097
1098 *) Update PKCS#7 enveloped data routines to use new API. This is now
1099 supported by any public key method supporting the encrypt operation. A
1100 ctrl is added to allow the public key algorithm to examine or modify
1101 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
1102 a no op.
1103 [Steve Henson]
1104
1105 *) Add a ctrl to asn1 method to allow a public key algorithm to express
1106 a default digest type to use. In most cases this will be SHA1 but some
1107 algorithms (such as GOST) need to specify an alternative digest. The
1108 return value indicates how strong the prefernce is 1 means optional and
1109 2 is mandatory (that is it is the only supported type). Modify
1110 ASN1_item_sign() to accept a NULL digest argument to indicate it should
1111 use the default md. Update openssl utilities to use the default digest
1112 type for signing if it is not explicitly indicated.
1113 [Steve Henson]
1114
1115 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
1116 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
1117 signing method from the key type. This effectively removes the link
1118 between digests and public key types.
1119 [Steve Henson]
1120
1121 *) Add an OID cross reference table and utility functions. Its purpose is to
1122 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
1123 rsaEncryption. This will allow some of the algorithm specific hackery
1124 needed to use the correct OID to be removed.
1125 [Steve Henson]
1126
1127 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
1128 structures for PKCS7_sign(). They are now set up by the relevant public
1129 key ASN1 method.
1130 [Steve Henson]
1131
1132 *) Add provisional EC pkey method with support for ECDSA and ECDH.
1133 [Steve Henson]
1134
1135 *) Add support for key derivation (agreement) in the API, DH method and
1136 pkeyutl.
1137 [Steve Henson]
1138
1139 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
1140 public and private key formats. As a side effect these add additional
1141 command line functionality not previously available: DSA signatures can be
1142 generated and verified using pkeyutl and DH key support and generation in
1143 pkey, genpkey.
1144 [Steve Henson]
1145
1146 *) BeOS support.
1147 [Oliver Tappe <zooey@hirschkaefer.de>]
1148
1149 *) New make target "install_html_docs" installs HTML renditions of the
1150 manual pages.
1151 [Oliver Tappe <zooey@hirschkaefer.de>]
1152
1153 *) New utility "genpkey" this is analagous to "genrsa" etc except it can
1154 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
1155 support key and parameter generation and add initial key generation
1156 functionality for RSA.
1157 [Steve Henson]
1158
1159 *) Add functions for main EVP_PKEY_method operations. The undocumented
1160 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
1161 EVP_PKEY_{encrypt,decrypt}_old.
1162 [Steve Henson]
1163
1164 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
1165 key API, doesn't do much yet.
1166 [Steve Henson]
1167
1168 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
1169 public key algorithms. New option to openssl utility:
1170 "list-public-key-algorithms" to print out info.
1171 [Steve Henson]
1172
1173 *) Implement the Supported Elliptic Curves Extension for
1174 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1175 [Douglas Stebila]
1176
1177 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
1178 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
1179 [Steve Henson]
1180
1181 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
1182 utilities such as rsa, dsa, dsaparam etc except they process any key
1183 type.
1184 [Steve Henson]
1185
1186 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
1187 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
1188 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
1189 structure.
1190 [Steve Henson]
1191
1192 *) Initial support for pluggable public key ASN1.
1193 De-spaghettify the public key ASN1 handling. Move public and private
1194 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
1195 algorithm specific handling to a single module within the relevant
1196 algorithm directory. Add functions to allow (near) opaque processing
1197 of public and private key structures.
1198 [Steve Henson]
1199
1200 *) Implement the Supported Point Formats Extension for
1201 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
1202 [Douglas Stebila]
1203
1204 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
1205 for the psk identity [hint] and the psk callback functions to the
1206 SSL_SESSION, SSL and SSL_CTX structure.
1207
1208 New ciphersuites:
1209 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
1210 PSK-AES256-CBC-SHA
1211
1212 New functions:
1213 SSL_CTX_use_psk_identity_hint
1214 SSL_get_psk_identity_hint
1215 SSL_get_psk_identity
1216 SSL_use_psk_identity_hint
1217
1218 [Mika Kousa and Pasi Eronen of Nokia Corporation]
1219
1220 *) Add RFC 3161 compliant time stamp request creation, response generation
1221 and response verification functionality.
1222 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
1223
1224 *) Add initial support for TLS extensions, specifically for the server_name
1225 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1226 have new members for a host name. The SSL data structure has an
1227 additional member SSL_CTX *initial_ctx so that new sessions can be
1228 stored in that context to allow for session resumption, even after the
1229 SSL has been switched to a new SSL_CTX in reaction to a client's
1230 server_name extension.
1231
1232 New functions (subject to change):
1233
1234 SSL_get_servername()
1235 SSL_get_servername_type()
1236 SSL_set_SSL_CTX()
1237
1238 New CTRL codes and macros (subject to change):
1239
1240 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1241 - SSL_CTX_set_tlsext_servername_callback()
1242 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1243 - SSL_CTX_set_tlsext_servername_arg()
1244 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1245
1246 openssl s_client has a new '-servername ...' option.
1247
1248 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1249 '-key2 ...', '-servername_fatal' (subject to change). This allows
1250 testing the HostName extension for a specific single host name ('-cert'
1251 and '-key' remain fallbacks for handshakes without HostName
1252 negotiation). If the unrecogninzed_name alert has to be sent, this by
1253 default is a warning; it becomes fatal with the '-servername_fatal'
1254 option.
1255
1256 [Peter Sylvester, Remy Allais, Christophe Renou]
1257
1258 *) Whirlpool hash implementation is added.
1259 [Andy Polyakov]
1260
1261 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
1262 bn(64,32). Because of instruction set limitations it doesn't have
1263 any negative impact on performance. This was done mostly in order
1264 to make it possible to share assembler modules, such as bn_mul_mont
1265 implementations, between 32- and 64-bit builds without hassle.
1266 [Andy Polyakov]
1267
1268 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
1269 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
1270 macro.
1271 [Bodo Moeller]
1272
1273 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
1274 dedicated Montgomery multiplication procedure, is introduced.
1275 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
1276 "64-bit" performance on certain 32-bit targets.
1277 [Andy Polyakov]
1278
1279 *) New option SSL_OP_NO_COMP to disable use of compression selectively
1280 in SSL structures. New SSL ctrl to set maximum send fragment size.
1281 Save memory by seeting the I/O buffer sizes dynamically instead of
1282 using the maximum available value.
1283 [Steve Henson]
1284
1285 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
1286 in addition to the text details.
1287 [Bodo Moeller]
1288
1289 *) Very, very preliminary EXPERIMENTAL support for printing of general
1290 ASN1 structures. This currently produces rather ugly output and doesn't
1291 handle several customised structures at all.
1292 [Steve Henson]
1293
1294 *) Integrated support for PVK file format and some related formats such
1295 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
1296 these in the 'rsa' and 'dsa' utilities.
1297 [Steve Henson]
1298
1299 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
1300 [Steve Henson]
1301
1302 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
1303 place for the (very old) "NETSCAPE" format certificates which are now
1304 handled using new ASN1 code equivalents.
1305 [Steve Henson]
1306
1307 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
1308 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
1309 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
1310 [Nils Larsch]
1311
1312 *) Modify CRL distribution points extension code to print out previously
1313 unsupported fields. Enhance extension setting code to allow setting of
1314 all fields.
1315 [Steve Henson]
1316
1317 *) Add print and set support for Issuing Distribution Point CRL extension.
1318 [Steve Henson]
1319
1320 *) Change 'Configure' script to enable Camellia by default.
1321 [NTT]
1322
1323 Changes between 0.9.8r and 0.9.8s [xx XXX xxxx]
1324
1325 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1326 for multi-threaded use of ECDH.
1327 [Adam Langley (Google)]
1328
1329 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1330 [Bodo Moeller]
1331
1332 *) Add protection against ECDSA timing attacks as mentioned in the paper
1333 by Billy Bob Brumley and Nicola Tuveri, see:
1334
1335 http://eprint.iacr.org/2011/232.pdf
1336
1337 [Billy Bob Brumley and Nicola Tuveri]
1338
1339 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
1340
1341 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1342 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1343
1344 *) Fix bug in string printing code: if *any* escaping is enabled we must
1345 escape the escape character (backslash) or the resulting string is
1346 ambiguous.
1347 [Steve Henson]
1348
1349 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
1350
1351 *) Disable code workaround for ancient and obsolete Netscape browsers
1352 and servers: an attacker can use it in a ciphersuite downgrade attack.
1353 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1354 [Steve Henson]
1355
1356 *) Fixed J-PAKE implementation error, originally discovered by
1357 Sebastien Martini, further info and confirmation from Stefan
1358 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1359 [Ben Laurie]
1360
1361 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
1362
1363 *) Fix extension code to avoid race conditions which can result in a buffer
1364 overrun vulnerability: resumed sessions must not be modified as they can
1365 be shared by multiple threads. CVE-2010-3864
1366 [Steve Henson]
1367
1368 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
1369 [Steve Henson]
1370
1371 *) Don't reencode certificate when calculating signature: cache and use
1372 the original encoding instead. This makes signature verification of
1373 some broken encodings work correctly.
1374 [Steve Henson]
1375
1376 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
1377 is also one of the inputs.
1378 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
1379
1380 *) Don't repeatedly append PBE algorithms to table if they already exist.
1381 Sort table on each new add. This effectively makes the table read only
1382 after all algorithms are added and subsequent calls to PKCS12_pbe_add
1383 etc are non-op.
1384 [Steve Henson]
1385
1386 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
1387
1388 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
1389 OpenSSL 1.0.0.]
1390
1391 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
1392 access or freeing data twice (CVE-2010-0742)
1393 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
1394
1395 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
1396 common in certificates and some applications which only call
1397 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
1398 [Steve Henson]
1399
1400 *) VMS fixes:
1401 Reduce copying into .apps and .test in makevms.com
1402 Don't try to use blank CA certificate in CA.com
1403 Allow use of C files from original directories in maketests.com
1404 [Steven M. Schweda" <sms@antinode.info>]
1405
1406 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
1407
1408 *) When rejecting SSL/TLS records due to an incorrect version number, never
1409 update s->server with a new major version number. As of
1410 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
1411 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
1412 the previous behavior could result in a read attempt at NULL when
1413 receiving specific incorrect SSL/TLS records once record payload
1414 protection is active. (CVE-2010-0740)
1415 [Bodo Moeller, Adam Langley <agl@chromium.org>]
1416
1417 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
1418 could be crashed if the relevant tables were not present (e.g. chrooted).
1419 [Tomas Hoger <thoger@redhat.com>]
1420
1421 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
1422
1423 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
1424 [Martin Olsson, Neel Mehta]
1425
1426 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
1427 accommodate for stack sorting, always a write lock!).
1428 [Bodo Moeller]
1429
1430 *) On some versions of WIN32 Heap32Next is very slow. This can cause
1431 excessive delays in the RAND_poll(): over a minute. As a workaround
1432 include a time check in the inner Heap32Next loop too.
1433 [Steve Henson]
1434
1435 *) The code that handled flushing of data in SSL/TLS originally used the
1436 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
1437 the problem outlined in PR#1949. The fix suggested there however can
1438 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
1439 of Apache). So instead simplify the code to flush unconditionally.
1440 This should be fine since flushing with no data to flush is a no op.
1441 [Steve Henson]
1442
1443 *) Handle TLS versions 2.0 and later properly and correctly use the
1444 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
1445 off ancient servers have a habit of sticking around for a while...
1446 [Steve Henson]
1447
1448 *) Modify compression code so it frees up structures without using the
1449 ex_data callbacks. This works around a problem where some applications
1450 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
1451 restarting) then use compression (e.g. SSL with compression) later.
1452 This results in significant per-connection memory leaks and
1453 has caused some security issues including CVE-2008-1678 and
1454 CVE-2009-4355.
1455 [Steve Henson]
1456
1457 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
1458 change when encrypting or decrypting.
1459 [Bodo Moeller]
1460
1461 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
1462 connect and renegotiate with servers which do not support RI.
1463 Until RI is more widely deployed this option is enabled by default.
1464 [Steve Henson]
1465
1466 *) Add "missing" ssl ctrls to clear options and mode.
1467 [Steve Henson]
1468
1469 *) If client attempts to renegotiate and doesn't support RI respond with
1470 a no_renegotiation alert as required by RFC5746. Some renegotiating
1471 TLS clients will continue a connection gracefully when they receive
1472 the alert. Unfortunately OpenSSL mishandled this alert and would hang
1473 waiting for a server hello which it will never receive. Now we treat a
1474 received no_renegotiation alert as a fatal error. This is because
1475 applications requesting a renegotiation might well expect it to succeed
1476 and would have no code in place to handle the server denying it so the
1477 only safe thing to do is to terminate the connection.
1478 [Steve Henson]
1479
1480 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
1481 peer supports secure renegotiation and 0 otherwise. Print out peer
1482 renegotiation support in s_client/s_server.
1483 [Steve Henson]
1484
1485 *) Replace the highly broken and deprecated SPKAC certification method with
1486 the updated NID creation version. This should correctly handle UTF8.
1487 [Steve Henson]
1488
1489 *) Implement RFC5746. Re-enable renegotiation but require the extension
1490 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
1491 turns out to be a bad idea. It has been replaced by
1492 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
1493 SSL_CTX_set_options(). This is really not recommended unless you
1494 know what you are doing.
1495 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
1496
1497 *) Fixes to stateless session resumption handling. Use initial_ctx when
1498 issuing and attempting to decrypt tickets in case it has changed during
1499 servername handling. Use a non-zero length session ID when attempting
1500 stateless session resumption: this makes it possible to determine if
1501 a resumption has occurred immediately after receiving server hello
1502 (several places in OpenSSL subtly assume this) instead of later in
1503 the handshake.
1504 [Steve Henson]
1505
1506 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
1507 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
1508 fixes for a few places where the return code is not checked
1509 correctly.
1510 [Julia Lawall <julia@diku.dk>]
1511
1512 *) Add --strict-warnings option to Configure script to include devteam
1513 warnings in other configurations.
1514 [Steve Henson]
1515
1516 *) Add support for --libdir option and LIBDIR variable in makefiles. This
1517 makes it possible to install openssl libraries in locations which
1518 have names other than "lib", for example "/usr/lib64" which some
1519 systems need.
1520 [Steve Henson, based on patch from Jeremy Utley]
1521
1522 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
1523 X690 8.9.12 and can produce some misleading textual output of OIDs.
1524 [Steve Henson, reported by Dan Kaminsky]
1525
1526 *) Delete MD2 from algorithm tables. This follows the recommendation in
1527 several standards that it is not used in new applications due to
1528 several cryptographic weaknesses. For binary compatibility reasons
1529 the MD2 API is still compiled in by default.
1530 [Steve Henson]
1531
1532 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
1533 and restored.
1534 [Steve Henson]
1535
1536 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
1537 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
1538 clash.
1539 [Guenter <lists@gknw.net>]
1540
1541 *) Fix the server certificate chain building code to use X509_verify_cert(),
1542 it used to have an ad-hoc builder which was unable to cope with anything
1543 other than a simple chain.
1544 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
1545
1546 *) Don't check self signed certificate signatures in X509_verify_cert()
1547 by default (a flag can override this): it just wastes time without
1548 adding any security. As a useful side effect self signed root CAs
1549 with non-FIPS digests are now usable in FIPS mode.
1550 [Steve Henson]
1551
1552 *) In dtls1_process_out_of_seq_message() the check if the current message
1553 is already buffered was missing. For every new message was memory
1554 allocated, allowing an attacker to perform an denial of service attack
1555 with sending out of seq handshake messages until there is no memory
1556 left. Additionally every future messege was buffered, even if the
1557 sequence number made no sense and would be part of another handshake.
1558 So only messages with sequence numbers less than 10 in advance will be
1559 buffered. (CVE-2009-1378)
1560 [Robin Seggelmann, discovered by Daniel Mentz]
1561
1562 *) Records are buffered if they arrive with a future epoch to be
1563 processed after finishing the corresponding handshake. There is
1564 currently no limitation to this buffer allowing an attacker to perform
1565 a DOS attack with sending records with future epochs until there is no
1566 memory left. This patch adds the pqueue_size() function to detemine
1567 the size of a buffer and limits the record buffer to 100 entries.
1568 (CVE-2009-1377)
1569 [Robin Seggelmann, discovered by Daniel Mentz]
1570
1571 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
1572 parent structure is freed. (CVE-2009-1379)
1573 [Daniel Mentz]
1574
1575 *) Handle non-blocking I/O properly in SSL_shutdown() call.
1576 [Darryl Miles <darryl-mailinglists@netbauds.net>]
1577
1578 *) Add 2.5.4.* OIDs
1579 [Ilya O. <vrghost@gmail.com>]
1580
1581 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
1582
1583 *) Disable renegotiation completely - this fixes a severe security
1584 problem (CVE-2009-3555) at the cost of breaking all
1585 renegotiation. Renegotiation can be re-enabled by setting
1586 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
1587 run-time. This is really not recommended unless you know what
1588 you're doing.
1589 [Ben Laurie]
1590
1591 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
1592
1593 *) Don't set val to NULL when freeing up structures, it is freed up by
1594 underlying code. If sizeof(void *) > sizeof(long) this can result in
1595 zeroing past the valid field. (CVE-2009-0789)
1596 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
1597
1598 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
1599 checked correctly. This would allow some invalid signed attributes to
1600 appear to verify correctly. (CVE-2009-0591)
1601 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1602
1603 *) Reject UniversalString and BMPString types with invalid lengths. This
1604 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
1605 a legal length. (CVE-2009-0590)
1606 [Steve Henson]
1607
1608 *) Set S/MIME signing as the default purpose rather than setting it
1609 unconditionally. This allows applications to override it at the store
1610 level.
1611 [Steve Henson]
1612
1613 *) Permit restricted recursion of ASN1 strings. This is needed in practice
1614 to handle some structures.
1615 [Steve Henson]
1616
1617 *) Improve efficiency of mem_gets: don't search whole buffer each time
1618 for a '\n'
1619 [Jeremy Shapiro <jnshapir@us.ibm.com>]
1620
1621 *) New -hex option for openssl rand.
1622 [Matthieu Herrb]
1623
1624 *) Print out UTF8String and NumericString when parsing ASN1.
1625 [Steve Henson]
1626
1627 *) Support NumericString type for name components.
1628 [Steve Henson]
1629
1630 *) Allow CC in the environment to override the automatically chosen
1631 compiler. Note that nothing is done to ensure flags work with the
1632 chosen compiler.
1633 [Ben Laurie]
1634
1635 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
1636
1637 *) Properly check EVP_VerifyFinal() and similar return values
1638 (CVE-2008-5077).
1639 [Ben Laurie, Bodo Moeller, Google Security Team]
1640
1641 *) Enable TLS extensions by default.
1642 [Ben Laurie]
1643
1644 *) Allow the CHIL engine to be loaded, whether the application is
1645 multithreaded or not. (This does not release the developer from the
1646 obligation to set up the dynamic locking callbacks.)
1647 [Sander Temme <sander@temme.net>]
1648
1649 *) Use correct exit code if there is an error in dgst command.
1650 [Steve Henson; problem pointed out by Roland Dirlewanger]
1651
1652 *) Tweak Configure so that you need to say "experimental-jpake" to enable
1653 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
1654 [Bodo Moeller]
1655
1656 *) Add experimental JPAKE support, including demo authentication in
1657 s_client and s_server.
1658 [Ben Laurie]
1659
1660 *) Set the comparison function in v3_addr_canonize().
1661 [Rob Austein <sra@hactrn.net>]
1662
1663 *) Add support for XMPP STARTTLS in s_client.
1664 [Philip Paeps <philip@freebsd.org>]
1665
1666 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
1667 to ensure that even with this option, only ciphersuites in the
1668 server's preference list will be accepted. (Note that the option
1669 applies only when resuming a session, so the earlier behavior was
1670 just about the algorithm choice for symmetric cryptography.)
1671 [Bodo Moeller]
1672
1673 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
1674
1675 *) Fix NULL pointer dereference if a DTLS server received
1676 ChangeCipherSpec as first record (CVE-2009-1386).
1677 [PR #1679]
1678
1679 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
1680 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
1681 [Nagendra Modadugu]
1682
1683 *) The fix in 0.9.8c that supposedly got rid of unsafe
1684 double-checked locking was incomplete for RSA blinding,
1685 addressing just one layer of what turns out to have been
1686 doubly unsafe triple-checked locking.
1687
1688 So now fix this for real by retiring the MONT_HELPER macro
1689 in crypto/rsa/rsa_eay.c.
1690
1691 [Bodo Moeller; problem pointed out by Marius Schilder]
1692
1693 *) Various precautionary measures:
1694
1695 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
1696
1697 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
1698 (NB: This would require knowledge of the secret session ticket key
1699 to exploit, in which case you'd be SOL either way.)
1700
1701 - Change bn_nist.c so that it will properly handle input BIGNUMs
1702 outside the expected range.
1703
1704 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
1705 builds.
1706
1707 [Neel Mehta, Bodo Moeller]
1708
1709 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
1710 the load fails. Useful for distros.
1711 [Ben Laurie and the FreeBSD team]
1712
1713 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
1714 [Steve Henson]
1715
1716 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
1717 [Huang Ying]
1718
1719 *) Expand ENGINE to support engine supplied SSL client certificate functions.
1720
1721 This work was sponsored by Logica.
1722 [Steve Henson]
1723
1724 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
1725 keystores. Support for SSL/TLS client authentication too.
1726 Not compiled unless enable-capieng specified to Configure.
1727
1728 This work was sponsored by Logica.
1729 [Steve Henson]
1730
1731 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
1732 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
1733 attribute creation routines such as certifcate requests and PKCS#12
1734 files.
1735 [Steve Henson]
1736
1737 Changes between 0.9.8g and 0.9.8h [28 May 2008]
1738
1739 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
1740 handshake which could lead to a cilent crash as found using the
1741 Codenomicon TLS test suite (CVE-2008-1672)
1742 [Steve Henson, Mark Cox]
1743
1744 *) Fix double free in TLS server name extensions which could lead to
1745 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
1746 [Joe Orton]
1747
1748 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
1749
1750 Clear the error queue to ensure that error entries left from
1751 older function calls do not interfere with the correct operation.
1752 [Lutz Jaenicke, Erik de Castro Lopo]
1753
1754 *) Remove root CA certificates of commercial CAs:
1755
1756 The OpenSSL project does not recommend any specific CA and does not
1757 have any policy with respect to including or excluding any CA.
1758 Therefore it does not make any sense to ship an arbitrary selection
1759 of root CA certificates with the OpenSSL software.
1760 [Lutz Jaenicke]
1761
1762 *) RSA OAEP patches to fix two separate invalid memory reads.
1763 The first one involves inputs when 'lzero' is greater than
1764 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
1765 before the beginning of from). The second one involves inputs where
1766 the 'db' section contains nothing but zeroes (there is a one-byte
1767 invalid read after the end of 'db').
1768 [Ivan Nestlerode <inestlerode@us.ibm.com>]
1769
1770 *) Partial backport from 0.9.9-dev:
1771
1772 Introduce bn_mul_mont (dedicated Montgomery multiplication
1773 procedure) as a candidate for BIGNUM assembler implementation.
1774 While 0.9.9-dev uses assembler for various architectures, only
1775 x86_64 is available by default here in the 0.9.8 branch, and
1776 32-bit x86 is available through a compile-time setting.
1777
1778 To try the 32-bit x86 assembler implementation, use Configure
1779 option "enable-montasm" (which exists only for this backport).
1780
1781 As "enable-montasm" for 32-bit x86 disclaims code stability
1782 anyway, in this constellation we activate additional code
1783 backported from 0.9.9-dev for further performance improvements,
1784 namely BN_from_montgomery_word. (To enable this otherwise,
1785 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
1786
1787 [Andy Polyakov (backport partially by Bodo Moeller)]
1788
1789 *) Add TLS session ticket callback. This allows an application to set
1790 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
1791 values. This is useful for key rollover for example where several key
1792 sets may exist with different names.
1793 [Steve Henson]
1794
1795 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
1796 This was broken until now in 0.9.8 releases, such that the only way
1797 a registered ENGINE could be used (assuming it initialises
1798 successfully on the host) was to explicitly set it as the default
1799 for the relevant algorithms. This is in contradiction with 0.9.7
1800 behaviour and the documentation. With this fix, when an ENGINE is
1801 registered into a given algorithm's table of implementations, the
1802 'uptodate' flag is reset so that auto-discovery will be used next
1803 time a new context for that algorithm attempts to select an
1804 implementation.
1805 [Ian Lister (tweaked by Geoff Thorpe)]
1806
1807 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
1808 implemention in the following ways:
1809
1810 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
1811 hard coded.
1812
1813 Lack of BER streaming support means one pass streaming processing is
1814 only supported if data is detached: setting the streaming flag is
1815 ignored for embedded content.
1816
1817 CMS support is disabled by default and must be explicitly enabled
1818 with the enable-cms configuration option.
1819 [Steve Henson]
1820
1821 *) Update the GMP engine glue to do direct copies between BIGNUM and
1822 mpz_t when openssl and GMP use the same limb size. Otherwise the
1823 existing "conversion via a text string export" trick is still used.
1824 [Paul Sheer <paulsheer@gmail.com>]
1825
1826 *) Zlib compression BIO. This is a filter BIO which compressed and
1827 uncompresses any data passed through it.
1828 [Steve Henson]
1829
1830 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
1831 RFC3394 compatible AES key wrapping.
1832 [Steve Henson]
1833
1834 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
1835 sets string data without copying. X509_ALGOR_set0() and
1836 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
1837 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
1838 from an X509_ATTRIBUTE structure optionally checking it occurs only
1839 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
1840 data.
1841 [Steve Henson]
1842
1843 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
1844 to get the expected BN_FLG_CONSTTIME behavior.
1845 [Bodo Moeller (Google)]
1846
1847 *) Netware support:
1848
1849 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
1850 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
1851 - added some more tests to do_tests.pl
1852 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
1853 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
1854 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
1855 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
1856 - various changes to netware.pl to enable gcc-cross builds on Win32
1857 platform
1858 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
1859 - various changes to fix missing prototype warnings
1860 - fixed x86nasm.pl to create correct asm files for NASM COFF output
1861 - added AES, WHIRLPOOL and CPUID assembler code to build files
1862 - added missing AES assembler make rules to mk1mf.pl
1863 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
1864 [Guenter Knauf <eflash@gmx.net>]
1865
1866 *) Implement certificate status request TLS extension defined in RFC3546.
1867 A client can set the appropriate parameters and receive the encoded
1868 OCSP response via a callback. A server can query the supplied parameters
1869 and set the encoded OCSP response in the callback. Add simplified examples
1870 to s_client and s_server.
1871 [Steve Henson]
1872
1873 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
1874
1875 *) Fix various bugs:
1876 + Binary incompatibility of ssl_ctx_st structure
1877 + DTLS interoperation with non-compliant servers
1878 + Don't call get_session_cb() without proposed session
1879 + Fix ia64 assembler code
1880 [Andy Polyakov, Steve Henson]
1881
1882 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
1883
1884 *) DTLS Handshake overhaul. There were longstanding issues with
1885 OpenSSL DTLS implementation, which were making it impossible for
1886 RFC 4347 compliant client to communicate with OpenSSL server.
1887 Unfortunately just fixing these incompatibilities would "cut off"
1888 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
1889 server keeps tolerating non RFC compliant syntax. The opposite is
1890 not true, 0.9.8f client can not communicate with earlier server.
1891 This update even addresses CVE-2007-4995.
1892 [Andy Polyakov]
1893
1894 *) Changes to avoid need for function casts in OpenSSL: some compilers
1895 (gcc 4.2 and later) reject their use.
1896 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
1897 Steve Henson]
1898
1899 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1900 RFC4507bis. The encrypted ticket format is an encrypted encoded
1901 SSL_SESSION structure, that way new session features are automatically
1902 supported.
1903
1904 If a client application caches session in an SSL_SESSION structure
1905 support is transparent because tickets are now stored in the encoded
1906 SSL_SESSION.
1907
1908 The SSL_CTX structure automatically generates keys for ticket
1909 protection in servers so again support should be possible
1910 with no application modification.
1911
1912 If a client or server wishes to disable RFC4507 support then the option
1913 SSL_OP_NO_TICKET can be set.
1914
1915 Add a TLS extension debugging callback to allow the contents of any client
1916 or server extensions to be examined.
1917
1918 This work was sponsored by Google.
1919 [Steve Henson]
1920
1921 *) Add initial support for TLS extensions, specifically for the server_name
1922 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
1923 have new members for a host name. The SSL data structure has an
1924 additional member SSL_CTX *initial_ctx so that new sessions can be
1925 stored in that context to allow for session resumption, even after the
1926 SSL has been switched to a new SSL_CTX in reaction to a client's
1927 server_name extension.
1928
1929 New functions (subject to change):
1930
1931 SSL_get_servername()
1932 SSL_get_servername_type()
1933 SSL_set_SSL_CTX()
1934
1935 New CTRL codes and macros (subject to change):
1936
1937 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
1938 - SSL_CTX_set_tlsext_servername_callback()
1939 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
1940 - SSL_CTX_set_tlsext_servername_arg()
1941 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
1942
1943 openssl s_client has a new '-servername ...' option.
1944
1945 openssl s_server has new options '-servername_host ...', '-cert2 ...',
1946 '-key2 ...', '-servername_fatal' (subject to change). This allows
1947 testing the HostName extension for a specific single host name ('-cert'
1948 and '-key' remain fallbacks for handshakes without HostName
1949 negotiation). If the unrecogninzed_name alert has to be sent, this by
1950 default is a warning; it becomes fatal with the '-servername_fatal'
1951 option.
1952
1953 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1954
1955 *) Add AES and SSE2 assembly language support to VC++ build.
1956 [Steve Henson]
1957
1958 *) Mitigate attack on final subtraction in Montgomery reduction.
1959 [Andy Polyakov]
1960
1961 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1962 (which previously caused an internal error).
1963 [Bodo Moeller]
1964
1965 *) Squeeze another 10% out of IGE mode when in != out.
1966 [Ben Laurie]
1967
1968 *) AES IGE mode speedup.
1969 [Dean Gaudet (Google)]
1970
1971 *) Add the Korean symmetric 128-bit cipher SEED (see
1972 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1973 add SEED ciphersuites from RFC 4162:
1974
1975 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1976 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1977 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1978 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1979
1980 To minimize changes between patchlevels in the OpenSSL 0.9.8
1981 series, SEED remains excluded from compilation unless OpenSSL
1982 is configured with 'enable-seed'.
1983 [KISA, Bodo Moeller]
1984
1985 *) Mitigate branch prediction attacks, which can be practical if a
1986 single processor is shared, allowing a spy process to extract
1987 information. For detailed background information, see
1988 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1989 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1990 and Necessary Software Countermeasures"). The core of the change
1991 are new versions BN_div_no_branch() and
1992 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1993 respectively, which are slower, but avoid the security-relevant
1994 conditional branches. These are automatically called by BN_div()
1995 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1996 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1997 remove a conditional branch.
1998
1999 BN_FLG_CONSTTIME is the new name for the previous
2000 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
2001 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
2002 in the exponent causes BN_mod_exp_mont() to use the alternative
2003 implementation in BN_mod_exp_mont_consttime().) The old name
2004 remains as a deprecated alias.
2005
2006 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
2007 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
2008 constant-time implementations for more than just exponentiation.
2009 Here too the old name is kept as a deprecated alias.
2010
2011 BN_BLINDING_new() will now use BN_dup() for the modulus so that
2012 the BN_BLINDING structure gets an independent copy of the
2013 modulus. This means that the previous "BIGNUM *m" argument to
2014 BN_BLINDING_new() and to BN_BLINDING_create_param() now
2015 essentially becomes "const BIGNUM *m", although we can't actually
2016 change this in the header file before 0.9.9. It allows
2017 RSA_setup_blinding() to use BN_with_flags() on the modulus to
2018 enable BN_FLG_CONSTTIME.
2019
2020 [Matthew D Wood (Intel Corp)]
2021
2022 *) In the SSL/TLS server implementation, be strict about session ID
2023 context matching (which matters if an application uses a single
2024 external cache for different purposes). Previously,
2025 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
2026 set. This did ensure strict client verification, but meant that,
2027 with applications using a single external cache for quite
2028 different requirements, clients could circumvent ciphersuite
2029 restrictions for a given session ID context by starting a session
2030 in a different context.
2031 [Bodo Moeller]
2032
2033 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2034 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2035 authentication-only ciphersuites.
2036 [Bodo Moeller]
2037
2038 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
2039 not complete and could lead to a possible single byte overflow
2040 (CVE-2007-5135) [Ben Laurie]
2041
2042 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
2043
2044 *) Since AES128 and AES256 (and similarly Camellia128 and
2045 Camellia256) share a single mask bit in the logic of
2046 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2047 kludge to work properly if AES128 is available and AES256 isn't
2048 (or if Camellia128 is available and Camellia256 isn't).
2049 [Victor Duchovni]
2050
2051 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
2052 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
2053 When a point or a seed is encoded in a BIT STRING, we need to
2054 prevent the removal of trailing zero bits to get the proper DER
2055 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
2056 of a NamedBitList, for which trailing 0 bits need to be removed.)
2057 [Bodo Moeller]
2058
2059 *) Have SSL/TLS server implementation tolerate "mismatched" record
2060 protocol version while receiving ClientHello even if the
2061 ClientHello is fragmented. (The server can't insist on the
2062 particular protocol version it has chosen before the ServerHello
2063 message has informed the client about his choice.)
2064 [Bodo Moeller]
2065
2066 *) Add RFC 3779 support.
2067 [Rob Austein for ARIN, Ben Laurie]
2068
2069 *) Load error codes if they are not already present instead of using a
2070 static variable. This allows them to be cleanly unloaded and reloaded.
2071 Improve header file function name parsing.
2072 [Steve Henson]
2073
2074 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
2075 or CAPABILITY handshake as required by RFCs.
2076 [Goetz Babin-Ebell]
2077
2078 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
2079
2080 *) Introduce limits to prevent malicious keys being able to
2081 cause a denial of service. (CVE-2006-2940)
2082 [Steve Henson, Bodo Moeller]
2083
2084 *) Fix ASN.1 parsing of certain invalid structures that can result
2085 in a denial of service. (CVE-2006-2937) [Steve Henson]
2086
2087 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2088 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2089
2090 *) Fix SSL client code which could crash if connecting to a
2091 malicious SSLv2 server. (CVE-2006-4343)
2092 [Tavis Ormandy and Will Drewry, Google Security Team]
2093
2094 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
2095 match only those. Before that, "AES256-SHA" would be interpreted
2096 as a pattern and match "AES128-SHA" too (since AES128-SHA got
2097 the same strength classification in 0.9.7h) as we currently only
2098 have a single AES bit in the ciphersuite description bitmap.
2099 That change, however, also applied to ciphersuite strings such as
2100 "RC4-MD5" that intentionally matched multiple ciphersuites --
2101 namely, SSL 2.0 ciphersuites in addition to the more common ones
2102 from SSL 3.0/TLS 1.0.
2103
2104 So we change the selection algorithm again: Naming an explicit
2105 ciphersuite selects this one ciphersuite, and any other similar
2106 ciphersuite (same bitmap) from *other* protocol versions.
2107 Thus, "RC4-MD5" again will properly select both the SSL 2.0
2108 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
2109
2110 Since SSL 2.0 does not have any ciphersuites for which the
2111 128/256 bit distinction would be relevant, this works for now.
2112 The proper fix will be to use different bits for AES128 and
2113 AES256, which would have avoided the problems from the beginning;
2114 however, bits are scarce, so we can only do this in a new release
2115 (not just a patchlevel) when we can change the SSL_CIPHER
2116 definition to split the single 'unsigned long mask' bitmap into
2117 multiple values to extend the available space.
2118
2119 [Bodo Moeller]
2120
2121 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
2122
2123 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2124 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2125
2126 *) Add AES IGE and biIGE modes.
2127 [Ben Laurie]
2128
2129 *) Change the Unix randomness entropy gathering to use poll() when
2130 possible instead of select(), since the latter has some
2131 undesirable limitations.
2132 [Darryl Miles via Richard Levitte and Bodo Moeller]
2133
2134 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
2135 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
2136 cannot be implicitly activated as part of, e.g., the "AES" alias.
2137 However, please upgrade to OpenSSL 0.9.9[-dev] for
2138 non-experimental use of the ECC ciphersuites to get TLS extension
2139 support, which is required for curve and point format negotiation
2140 to avoid potential handshake problems.
2141 [Bodo Moeller]
2142
2143 *) Disable rogue ciphersuites:
2144
2145 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2146 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2147 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2148
2149 The latter two were purportedly from
2150 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2151 appear there.
2152
2153 Also deactivate the remaining ciphersuites from
2154 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2155 unofficial, and the ID has long expired.
2156 [Bodo Moeller]
2157
2158 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2159 dual-core machines) and other potential thread-safety issues.
2160 [Bodo Moeller]
2161
2162 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
2163 versions), which is now available for royalty-free use
2164 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
2165 Also, add Camellia TLS ciphersuites from RFC 4132.
2166
2167 To minimize changes between patchlevels in the OpenSSL 0.9.8
2168 series, Camellia remains excluded from compilation unless OpenSSL
2169 is configured with 'enable-camellia'.
2170 [NTT]
2171
2172 *) Disable the padding bug check when compression is in use. The padding
2173 bug check assumes the first packet is of even length, this is not
2174 necessarily true if compresssion is enabled and can result in false
2175 positives causing handshake failure. The actual bug test is ancient
2176 code so it is hoped that implementations will either have fixed it by
2177 now or any which still have the bug do not support compression.
2178 [Steve Henson]
2179
2180 Changes between 0.9.8a and 0.9.8b [04 May 2006]
2181
2182 *) When applying a cipher rule check to see if string match is an explicit
2183 cipher suite and only match that one cipher suite if it is.
2184 [Steve Henson]
2185
2186 *) Link in manifests for VC++ if needed.
2187 [Austin Ziegler <halostatue@gmail.com>]
2188
2189 *) Update support for ECC-based TLS ciphersuites according to
2190 draft-ietf-tls-ecc-12.txt with proposed changes (but without
2191 TLS extensions, which are supported starting with the 0.9.9
2192 branch, not in the OpenSSL 0.9.8 branch).
2193 [Douglas Stebila]
2194
2195 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
2196 opaque EVP_CIPHER_CTX handling.
2197 [Steve Henson]
2198
2199 *) Fixes and enhancements to zlib compression code. We now only use
2200 "zlib1.dll" and use the default __cdecl calling convention on Win32
2201 to conform with the standards mentioned here:
2202 http://www.zlib.net/DLL_FAQ.txt
2203 Static zlib linking now works on Windows and the new --with-zlib-include
2204 --with-zlib-lib options to Configure can be used to supply the location
2205 of the headers and library. Gracefully handle case where zlib library
2206 can't be loaded.
2207 [Steve Henson]
2208
2209 *) Several fixes and enhancements to the OID generation code. The old code
2210 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
2211 handle numbers larger than ULONG_MAX, truncated printing and had a
2212 non standard OBJ_obj2txt() behaviour.
2213 [Steve Henson]
2214
2215 *) Add support for building of engines under engine/ as shared libraries
2216 under VC++ build system.
2217 [Steve Henson]
2218
2219 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
2220 Hopefully, we will not see any false combination of paths any more.
2221 [Richard Levitte]
2222
2223 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
2224
2225 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2226 (part of SSL_OP_ALL). This option used to disable the
2227 countermeasure against man-in-the-middle protocol-version
2228 rollback in the SSL 2.0 server implementation, which is a bad
2229 idea. (CVE-2005-2969)
2230
2231 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2232 for Information Security, National Institute of Advanced Industrial
2233 Science and Technology [AIST], Japan)]
2234
2235 *) Add two function to clear and return the verify parameter flags.
2236 [Steve Henson]
2237
2238 *) Keep cipherlists sorted in the source instead of sorting them at
2239 runtime, thus removing the need for a lock.
2240 [Nils Larsch]
2241
2242 *) Avoid some small subgroup attacks in Diffie-Hellman.
2243 [Nick Mathewson and Ben Laurie]
2244
2245 *) Add functions for well-known primes.
2246 [Nick Mathewson]
2247
2248 *) Extended Windows CE support.
2249 [Satoshi Nakamura and Andy Polyakov]
2250
2251 *) Initialize SSL_METHOD structures at compile time instead of during
2252 runtime, thus removing the need for a lock.
2253 [Steve Henson]
2254
2255 *) Make PKCS7_decrypt() work even if no certificate is supplied by
2256 attempting to decrypt each encrypted key in turn. Add support to
2257 smime utility.
2258 [Steve Henson]
2259
2260 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
2261
2262 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
2263 OpenSSL 0.9.8.]
2264
2265 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
2266 [Richard Levitte]
2267
2268 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
2269 key into the same file any more.
2270 [Richard Levitte]
2271
2272 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
2273 [Andy Polyakov]
2274
2275 *) Add -utf8 command line and config file option to 'ca'.
2276 [Stefan <stf@udoma.org]
2277
2278 *) Removed the macro des_crypt(), as it seems to conflict with some
2279 libraries. Use DES_crypt().
2280 [Richard Levitte]
2281
2282 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
2283 involves renaming the source and generated shared-libs for
2284 both. The engines will accept the corrected or legacy ids
2285 ('ncipher' and '4758_cca' respectively) when binding. NB,
2286 this only applies when building 'shared'.
2287 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
2288
2289 *) Add attribute functions to EVP_PKEY structure. Modify
2290 PKCS12_create() to recognize a CSP name attribute and
2291 use it. Make -CSP option work again in pkcs12 utility.
2292 [Steve Henson]
2293
2294 *) Add new functionality to the bn blinding code:
2295 - automatic re-creation of the BN_BLINDING parameters after
2296 a fixed number of uses (currently 32)
2297 - add new function for parameter creation
2298 - introduce flags to control the update behaviour of the
2299 BN_BLINDING parameters
2300 - hide BN_BLINDING structure
2301 Add a second BN_BLINDING slot to the RSA structure to improve
2302 performance when a single RSA object is shared among several
2303 threads.
2304 [Nils Larsch]
2305
2306 *) Add support for DTLS.
2307 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
2308
2309 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
2310 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
2311 [Walter Goulet]
2312
2313 *) Remove buggy and incompletet DH cert support from
2314 ssl/ssl_rsa.c and ssl/s3_both.c
2315 [Nils Larsch]
2316
2317 *) Use SHA-1 instead of MD5 as the default digest algorithm for
2318 the apps/openssl applications.
2319 [Nils Larsch]
2320
2321 *) Compile clean with "-Wall -Wmissing-prototypes
2322 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
2323 DEBUG_SAFESTACK must also be set.
2324 [Ben Laurie]
2325
2326 *) Change ./Configure so that certain algorithms can be disabled by default.
2327 The new counterpiece to "no-xxx" is "enable-xxx".
2328
2329 The patented RC5 and MDC2 algorithms will now be disabled unless
2330 "enable-rc5" and "enable-mdc2", respectively, are specified.
2331
2332 (IDEA remains enabled despite being patented. This is because IDEA
2333 is frequently required for interoperability, and there is no license
2334 fee for non-commercial use. As before, "no-idea" can be used to
2335 avoid this algorithm.)
2336
2337 [Bodo Moeller]
2338
2339 *) Add processing of proxy certificates (see RFC 3820). This work was
2340 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
2341 EGEE (Enabling Grids for E-science in Europe).
2342 [Richard Levitte]
2343
2344 *) RC4 performance overhaul on modern architectures/implementations, such
2345 as Intel P4, IA-64 and AMD64.
2346 [Andy Polyakov]
2347
2348 *) New utility extract-section.pl. This can be used specify an alternative
2349 section number in a pod file instead of having to treat each file as
2350 a separate case in Makefile. This can be done by adding two lines to the
2351 pod file:
2352
2353 =for comment openssl_section:XXX
2354
2355 The blank line is mandatory.
2356
2357 [Steve Henson]
2358
2359 *) New arguments -certform, -keyform and -pass for s_client and s_server
2360 to allow alternative format key and certificate files and passphrase
2361 sources.
2362 [Steve Henson]
2363
2364 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
2365 update associated structures and add various utility functions.
2366
2367 Add new policy related verify parameters, include policy checking in
2368 standard verify code. Enhance 'smime' application with extra parameters
2369 to support policy checking and print out.
2370 [Steve Henson]
2371
2372 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
2373 Nehemiah processors. These extensions support AES encryption in hardware
2374 as well as RNG (though RNG support is currently disabled).
2375 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
2376
2377 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
2378 [Geoff Thorpe]
2379
2380 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
2381 [Andy Polyakov and a number of other people]
2382
2383 *) Improved PowerPC platform support. Most notably BIGNUM assembler
2384 implementation contributed by IBM.
2385 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
2386
2387 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
2388 exponent rather than 'unsigned long'. There is a corresponding change to
2389 the new 'rsa_keygen' element of the RSA_METHOD structure.
2390 [Jelte Jansen, Geoff Thorpe]
2391
2392 *) Functionality for creating the initial serial number file is now
2393 moved from CA.pl to the 'ca' utility with a new option -create_serial.
2394
2395 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
2396 number file to 1, which is bound to cause problems. To avoid
2397 the problems while respecting compatibility between different 0.9.7
2398 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
2399 CA.pl for serial number initialization. With the new release 0.9.8,
2400 we can fix the problem directly in the 'ca' utility.)
2401 [Steve Henson]
2402
2403 *) Reduced header interdepencies by declaring more opaque objects in
2404 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
2405 give fewer recursive includes, which could break lazy source code - so
2406 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
2407 developers should define this symbol when building and using openssl to
2408 ensure they track the recommended behaviour, interfaces, [etc], but
2409 backwards-compatible behaviour prevails when this isn't defined.
2410 [Geoff Thorpe]
2411
2412 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
2413 [Steve Henson]
2414
2415 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
2416 This will generate a random key of the appropriate length based on the
2417 cipher context. The EVP_CIPHER can provide its own random key generation
2418 routine to support keys of a specific form. This is used in the des and
2419 3des routines to generate a key of the correct parity. Update S/MIME
2420 code to use new functions and hence generate correct parity DES keys.
2421 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
2422 valid (weak or incorrect parity).
2423 [Steve Henson]
2424
2425 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
2426 as looking them up. This is useful when the verified structure may contain
2427 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
2428 present unless the new PKCS7_NO_CRL flag is asserted.
2429 [Steve Henson]
2430
2431 *) Extend ASN1 oid configuration module. It now additionally accepts the
2432 syntax:
2433
2434 shortName = some long name, 1.2.3.4
2435 [Steve Henson]
2436
2437 *) Reimplemented the BN_CTX implementation. There is now no more static
2438 limitation on the number of variables it can handle nor the depth of the
2439 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
2440 information can now expand as required, and rather than having a single
2441 static array of bignums, BN_CTX now uses a linked-list of such arrays
2442 allowing it to expand on demand whilst maintaining the usefulness of
2443 BN_CTX's "bundling".
2444 [Geoff Thorpe]
2445
2446 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
2447 to allow all RSA operations to function using a single BN_CTX.
2448 [Geoff Thorpe]
2449
2450 *) Preliminary support for certificate policy evaluation and checking. This
2451 is initially intended to pass the tests outlined in "Conformance Testing
2452 of Relying Party Client Certificate Path Processing Logic" v1.07.
2453 [Steve Henson]
2454
2455 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
2456 remained unused and not that useful. A variety of other little bignum
2457 tweaks and fixes have also been made continuing on from the audit (see
2458 below).
2459 [Geoff Thorpe]
2460
2461 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
2462 associated ASN1, EVP and SSL functions and old ASN1 macros.
2463 [Richard Levitte]
2464
2465 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
2466 and this should never fail. So the return value from the use of
2467 BN_set_word() (which can fail due to needless expansion) is now deprecated;
2468 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
2469 [Geoff Thorpe]
2470
2471 *) BN_CTX_get() should return zero-valued bignums, providing the same
2472 initialised value as BN_new().
2473 [Geoff Thorpe, suggested by Ulf Möller]
2474
2475 *) Support for inhibitAnyPolicy certificate extension.
2476 [Steve Henson]
2477
2478 *) An audit of the BIGNUM code is underway, for which debugging code is
2479 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
2480 is considered valid when processing BIGNUMs, and causes execution to
2481 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
2482 further steps are taken to deliberately pollute unused data in BIGNUM
2483 structures to try and expose faulty code further on. For now, openssl will
2484 (in its default mode of operation) continue to tolerate the inconsistent
2485 forms that it has tolerated in the past, but authors and packagers should
2486 consider trying openssl and their own applications when compiled with
2487 these debugging symbols defined. It will help highlight potential bugs in
2488 their own code, and will improve the test coverage for OpenSSL itself. At
2489 some point, these tighter rules will become openssl's default to improve
2490 maintainability, though the assert()s and other overheads will remain only
2491 in debugging configurations. See bn.h for more details.
2492 [Geoff Thorpe, Nils Larsch, Ulf Möller]
2493
2494 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
2495 that can only be obtained through BN_CTX_new() (which implicitly
2496 initialises it). The presence of this function only made it possible
2497 to overwrite an existing structure (and cause memory leaks).
2498 [Geoff Thorpe]
2499
2500 *) Because of the callback-based approach for implementing LHASH as a
2501 template type, lh_insert() adds opaque objects to hash-tables and
2502 lh_doall() or lh_doall_arg() are typically used with a destructor callback
2503 to clean up those corresponding objects before destroying the hash table
2504 (and losing the object pointers). So some over-zealous constifications in
2505 LHASH have been relaxed so that lh_insert() does not take (nor store) the
2506 objects as "const" and the lh_doall[_arg] callback wrappers are not
2507 prototyped to have "const" restrictions on the object pointers they are
2508 given (and so aren't required to cast them away any more).
2509 [Geoff Thorpe]
2510
2511 *) The tmdiff.h API was so ugly and minimal that our own timing utility
2512 (speed) prefers to use its own implementation. The two implementations
2513 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
2514 its object type properly exposed (MS_TM) instead of casting to/from "char
2515 *". This may still change yet if someone realises MS_TM and "ms_time_***"
2516 aren't necessarily the greatest nomenclatures - but this is what was used
2517 internally to the implementation so I've used that for now.
2518 [Geoff Thorpe]
2519
2520 *) Ensure that deprecated functions do not get compiled when
2521 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
2522 the self-tests were still using deprecated key-generation functions so
2523 these have been updated also.
2524 [Geoff Thorpe]
2525
2526 *) Reorganise PKCS#7 code to separate the digest location functionality
2527 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
2528 New function PKCS7_set_digest() to set the digest type for PKCS#7
2529 digestedData type. Add additional code to correctly generate the
2530 digestedData type and add support for this type in PKCS7 initialization
2531 functions.
2532 [Steve Henson]
2533
2534 *) New function PKCS7_set0_type_other() this initializes a PKCS7
2535 structure of type "other".
2536 [Steve Henson]
2537
2538 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
2539 sure the loop does correctly stop and breaking ("division by zero")
2540 modulus operations are not performed. The (pre-generated) prime
2541 table crypto/bn/bn_prime.h was already correct, but it could not be
2542 re-generated on some platforms because of the "division by zero"
2543 situation in the script.
2544 [Ralf S. Engelschall]
2545
2546 *) Update support for ECC-based TLS ciphersuites according to
2547 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
2548 SHA-1 now is only used for "small" curves (where the
2549 representation of a field element takes up to 24 bytes); for
2550 larger curves, the field element resulting from ECDH is directly
2551 used as premaster secret.
2552 [Douglas Stebila (Sun Microsystems Laboratories)]
2553
2554 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
2555 curve secp160r1 to the tests.
2556 [Douglas Stebila (Sun Microsystems Laboratories)]
2557
2558 *) Add the possibility to load symbols globally with DSO.
2559 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
2560
2561 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
2562 control of the error stack.
2563 [Richard Levitte]
2564
2565 *) Add support for STORE in ENGINE.
2566 [Richard Levitte]
2567
2568 *) Add the STORE type. The intention is to provide a common interface
2569 to certificate and key stores, be they simple file-based stores, or
2570 HSM-type store, or LDAP stores, or...
2571 NOTE: The code is currently UNTESTED and isn't really used anywhere.
2572 [Richard Levitte]
2573
2574 *) Add a generic structure called OPENSSL_ITEM. This can be used to
2575 pass a list of arguments to any function as well as provide a way
2576 for a function to pass data back to the caller.
2577 [Richard Levitte]
2578
2579 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
2580 works like BUF_strdup() but can be used to duplicate a portion of
2581 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
2582 a memory area.
2583 [Richard Levitte]
2584
2585 *) Add the function sk_find_ex() which works like sk_find(), but will
2586 return an index to an element even if an exact match couldn't be
2587 found. The index is guaranteed to point at the element where the
2588 searched-for key would be inserted to preserve sorting order.
2589 [Richard Levitte]
2590
2591 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
2592 takes an extra flags argument for optional functionality. Currently,
2593 the following flags are defined:
2594
2595 OBJ_BSEARCH_VALUE_ON_NOMATCH
2596 This one gets OBJ_bsearch_ex() to return a pointer to the first
2597 element where the comparing function returns a negative or zero
2598 number.
2599
2600 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
2601 This one gets OBJ_bsearch_ex() to return a pointer to the first
2602 element where the comparing function returns zero. This is useful
2603 if there are more than one element where the comparing function
2604 returns zero.
2605 [Richard Levitte]
2606
2607 *) Make it possible to create self-signed certificates with 'openssl ca'
2608 in such a way that the self-signed certificate becomes part of the
2609 CA database and uses the same mechanisms for serial number generation
2610 as all other certificate signing. The new flag '-selfsign' enables
2611 this functionality. Adapt CA.sh and CA.pl.in.
2612 [Richard Levitte]
2613
2614 *) Add functionality to check the public key of a certificate request
2615 against a given private. This is useful to check that a certificate
2616 request can be signed by that key (self-signing).
2617 [Richard Levitte]
2618
2619 *) Make it possible to have multiple active certificates with the same
2620 subject in the CA index file. This is done only if the keyword
2621 'unique_subject' is set to 'no' in the main CA section (default
2622 if 'CA_default') of the configuration file. The value is saved
2623 with the database itself in a separate index attribute file,
2624 named like the index file with '.attr' appended to the name.
2625 [Richard Levitte]
2626
2627 *) Generate muti valued AVAs using '+' notation in config files for
2628 req and dirName.
2629 [Steve Henson]
2630
2631 *) Support for nameConstraints certificate extension.
2632 [Steve Henson]
2633
2634 *) Support for policyConstraints certificate extension.
2635 [Steve Henson]
2636
2637 *) Support for policyMappings certificate extension.
2638 [Steve Henson]
2639
2640 *) Make sure the default DSA_METHOD implementation only uses its
2641 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
2642 and change its own handlers to be NULL so as to remove unnecessary
2643 indirection. This lets alternative implementations fallback to the
2644 default implementation more easily.
2645 [Geoff Thorpe]
2646
2647 *) Support for directoryName in GeneralName related extensions
2648 in config files.
2649 [Steve Henson]
2650
2651 *) Make it possible to link applications using Makefile.shared.
2652 Make that possible even when linking against static libraries!
2653 [Richard Levitte]
2654
2655 *) Support for single pass processing for S/MIME signing. This now
2656 means that S/MIME signing can be done from a pipe, in addition
2657 cleartext signing (multipart/signed type) is effectively streaming
2658 and the signed data does not need to be all held in memory.
2659
2660 This is done with a new flag PKCS7_STREAM. When this flag is set
2661 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
2662 is done after the data is output (and digests calculated) in
2663 SMIME_write_PKCS7().
2664 [Steve Henson]
2665
2666 *) Add full support for -rpath/-R, both in shared libraries and
2667 applications, at least on the platforms where it's known how
2668 to do it.
2669 [Richard Levitte]
2670
2671 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
2672 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
2673 will now compute a table of multiples of the generator that
2674 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
2675 faster (notably in the case of a single point multiplication,
2676 scalar * generator).
2677 [Nils Larsch, Bodo Moeller]
2678
2679 *) IPv6 support for certificate extensions. The various extensions
2680 which use the IP:a.b.c.d can now take IPv6 addresses using the
2681 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
2682 correctly.
2683 [Steve Henson]
2684
2685 *) Added an ENGINE that implements RSA by performing private key
2686 exponentiations with the GMP library. The conversions to and from
2687 GMP's mpz_t format aren't optimised nor are any montgomery forms
2688 cached, and on x86 it appears OpenSSL's own performance has caught up.
2689 However there are likely to be other architectures where GMP could
2690 provide a boost. This ENGINE is not built in by default, but it can be
2691 specified at Configure time and should be accompanied by the necessary
2692 linker additions, eg;
2693 ./config -DOPENSSL_USE_GMP -lgmp
2694 [Geoff Thorpe]
2695
2696 *) "openssl engine" will not display ENGINE/DSO load failure errors when
2697 testing availability of engines with "-t" - the old behaviour is
2698 produced by increasing the feature's verbosity with "-tt".
2699 [Geoff Thorpe]
2700
2701 *) ECDSA routines: under certain error conditions uninitialized BN objects
2702 could be freed. Solution: make sure initialization is performed early
2703 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
2704 via PR#459)
2705 [Lutz Jaenicke]
2706
2707 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
2708 and DH_METHOD (eg. by ENGINE implementations) to override the normal
2709 software implementations. For DSA and DH, parameter generation can
2710 also be overriden by providing the appropriate method callbacks.
2711 [Geoff Thorpe]
2712
2713 *) Change the "progress" mechanism used in key-generation and
2714 primality testing to functions that take a new BN_GENCB pointer in
2715 place of callback/argument pairs. The new API functions have "_ex"
2716 postfixes and the older functions are reimplemented as wrappers for
2717 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
2718 declarations of the old functions to help (graceful) attempts to
2719 migrate to the new functions. Also, the new key-generation API
2720 functions operate on a caller-supplied key-structure and return
2721 success/failure rather than returning a key or NULL - this is to
2722 help make "keygen" another member function of RSA_METHOD etc.
2723
2724 Example for using the new callback interface:
2725
2726 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
2727 void *my_arg = ...;
2728 BN_GENCB my_cb;
2729
2730 BN_GENCB_set(&my_cb, my_callback, my_arg);
2731
2732 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
2733 /* For the meaning of a, b in calls to my_callback(), see the
2734 * documentation of the function that calls the callback.
2735 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
2736 * my_callback should return 1 if it wants BN_is_prime_ex()
2737 * to continue, or 0 to stop.
2738 */
2739
2740 [Geoff Thorpe]
2741
2742 *) Change the ZLIB compression method to be stateful, and make it
2743 available to TLS with the number defined in
2744 draft-ietf-tls-compression-04.txt.
2745 [Richard Levitte]
2746
2747 *) Add the ASN.1 structures and functions for CertificatePair, which
2748 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
2749
2750 CertificatePair ::= SEQUENCE {
2751 forward [0] Certificate OPTIONAL,
2752 reverse [1] Certificate OPTIONAL,
2753 -- at least one of the pair shall be present -- }
2754
2755 Also implement the PEM functions to read and write certificate
2756 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
2757
2758 This needed to be defined, mostly for the sake of the LDAP
2759 attribute crossCertificatePair, but may prove useful elsewhere as
2760 well.
2761 [Richard Levitte]
2762
2763 *) Make it possible to inhibit symlinking of shared libraries in
2764 Makefile.shared, for Cygwin's sake.
2765 [Richard Levitte]
2766
2767 *) Extend the BIGNUM API by creating a function
2768 void BN_set_negative(BIGNUM *a, int neg);
2769 and a macro that behave like
2770 int BN_is_negative(const BIGNUM *a);
2771
2772 to avoid the need to access 'a->neg' directly in applications.
2773 [Nils Larsch]
2774
2775 *) Implement fast modular reduction for pseudo-Mersenne primes
2776 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
2777 EC_GROUP_new_curve_GFp() will now automatically use this
2778 if applicable.
2779 [Nils Larsch <nla@trustcenter.de>]
2780
2781 *) Add new lock type (CRYPTO_LOCK_BN).
2782 [Bodo Moeller]
2783
2784 *) Change the ENGINE framework to automatically load engines
2785 dynamically from specific directories unless they could be
2786 found to already be built in or loaded. Move all the
2787 current engines except for the cryptodev one to a new
2788 directory engines/.
2789 The engines in engines/ are built as shared libraries if
2790 the "shared" options was given to ./Configure or ./config.
2791 Otherwise, they are inserted in libcrypto.a.
2792 /usr/local/ssl/engines is the default directory for dynamic
2793 engines, but that can be overriden at configure time through
2794 the usual use of --prefix and/or --openssldir, and at run
2795 time with the environment variable OPENSSL_ENGINES.
2796 [Geoff Thorpe and Richard Levitte]
2797
2798 *) Add Makefile.shared, a helper makefile to build shared
2799 libraries. Addapt Makefile.org.
2800 [Richard Levitte]
2801
2802 *) Add version info to Win32 DLLs.
2803 [Peter 'Luna' Runestig" <peter@runestig.com>]
2804
2805 *) Add new 'medium level' PKCS#12 API. Certificates and keys
2806 can be added using this API to created arbitrary PKCS#12
2807 files while avoiding the low level API.
2808
2809 New options to PKCS12_create(), key or cert can be NULL and
2810 will then be omitted from the output file. The encryption
2811 algorithm NIDs can be set to -1 for no encryption, the mac
2812 iteration count can be set to 0 to omit the mac.
2813
2814 Enhance pkcs12 utility by making the -nokeys and -nocerts
2815 options work when creating a PKCS#12 file. New option -nomac
2816 to omit the mac, NONE can be set for an encryption algorithm.
2817 New code is modified to use the enhanced PKCS12_create()
2818 instead of the low level API.
2819 [Steve Henson]
2820
2821 *) Extend ASN1 encoder to support indefinite length constructed
2822 encoding. This can output sequences tags and octet strings in
2823 this form. Modify pk7_asn1.c to support indefinite length
2824 encoding. This is experimental and needs additional code to
2825 be useful, such as an ASN1 bio and some enhanced streaming
2826 PKCS#7 code.
2827
2828 Extend template encode functionality so that tagging is passed
2829 down to the template encoder.
2830 [Steve Henson]
2831
2832 *) Let 'openssl req' fail if an argument to '-newkey' is not
2833 recognized instead of using RSA as a default.
2834 [Bodo Moeller]
2835
2836 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
2837 As these are not official, they are not included in "ALL";
2838 the "ECCdraft" ciphersuite group alias can be used to select them.
2839 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
2840
2841 *) Add ECDH engine support.
2842 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
2843
2844 *) Add ECDH in new directory crypto/ecdh/.
2845 [Douglas Stebila (Sun Microsystems Laboratories)]
2846
2847 *) Let BN_rand_range() abort with an error after 100 iterations
2848 without success (which indicates a broken PRNG).
2849 [Bodo Moeller]
2850
2851 *) Change BN_mod_sqrt() so that it verifies that the input value
2852 is really the square of the return value. (Previously,
2853 BN_mod_sqrt would show GIGO behaviour.)
2854 [Bodo Moeller]
2855
2856 *) Add named elliptic curves over binary fields from X9.62, SECG,
2857 and WAP/WTLS; add OIDs that were still missing.
2858
2859 [Sheueling Chang Shantz and Douglas Stebila
2860 (Sun Microsystems Laboratories)]
2861
2862 *) Extend the EC library for elliptic curves over binary fields
2863 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
2864 New EC_METHOD:
2865
2866 EC_GF2m_simple_method
2867
2868 New API functions:
2869
2870 EC_GROUP_new_curve_GF2m
2871 EC_GROUP_set_curve_GF2m
2872 EC_GROUP_get_curve_GF2m
2873 EC_POINT_set_affine_coordinates_GF2m
2874 EC_POINT_get_affine_coordinates_GF2m
2875 EC_POINT_set_compressed_coordinates_GF2m
2876
2877 Point compression for binary fields is disabled by default for
2878 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
2879 enable it).
2880
2881 As binary polynomials are represented as BIGNUMs, various members
2882 of the EC_GROUP and EC_POINT data structures can be shared
2883 between the implementations for prime fields and binary fields;
2884 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
2885 are essentially identical to their ..._GFp counterparts.
2886 (For simplicity, the '..._GFp' prefix has been dropped from
2887 various internal method names.)
2888
2889 An internal 'field_div' method (similar to 'field_mul' and
2890 'field_sqr') has been added; this is used only for binary fields.
2891
2892 [Sheueling Chang Shantz and Douglas Stebila
2893 (Sun Microsystems Laboratories)]
2894
2895 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
2896 through methods ('mul', 'precompute_mult').
2897
2898 The generic implementations (now internally called 'ec_wNAF_mul'
2899 and 'ec_wNAF_precomputed_mult') remain the default if these
2900 methods are undefined.
2901
2902 [Sheueling Chang Shantz and Douglas Stebila
2903 (Sun Microsystems Laboratories)]
2904
2905 *) New function EC_GROUP_get_degree, which is defined through
2906 EC_METHOD. For curves over prime fields, this returns the bit
2907 length of the modulus.
2908
2909 [Sheueling Chang Shantz and Douglas Stebila
2910 (Sun Microsystems Laboratories)]
2911
2912 *) New functions EC_GROUP_dup, EC_POINT_dup.
2913 (These simply call ..._new and ..._copy).
2914
2915 [Sheueling Chang Shantz and Douglas Stebila
2916 (Sun Microsystems Laboratories)]
2917
2918 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
2919 Polynomials are represented as BIGNUMs (where the sign bit is not
2920 used) in the following functions [macros]:
2921
2922 BN_GF2m_add
2923 BN_GF2m_sub [= BN_GF2m_add]
2924 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
2925 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
2926 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
2927 BN_GF2m_mod_inv
2928 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
2929 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
2930 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
2931 BN_GF2m_cmp [= BN_ucmp]
2932
2933 (Note that only the 'mod' functions are actually for fields GF(2^m).
2934 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
2935
2936 For some functions, an the irreducible polynomial defining a
2937 field can be given as an 'unsigned int[]' with strictly
2938 decreasing elements giving the indices of those bits that are set;
2939 i.e., p[] represents the polynomial
2940 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
2941 where
2942 p[0] > p[1] > ... > p[k] = 0.
2943 This applies to the following functions:
2944
2945 BN_GF2m_mod_arr
2946 BN_GF2m_mod_mul_arr
2947 BN_GF2m_mod_sqr_arr
2948 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
2949 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
2950 BN_GF2m_mod_exp_arr
2951 BN_GF2m_mod_sqrt_arr
2952 BN_GF2m_mod_solve_quad_arr
2953 BN_GF2m_poly2arr
2954 BN_GF2m_arr2poly
2955
2956 Conversion can be performed by the following functions:
2957
2958 BN_GF2m_poly2arr
2959 BN_GF2m_arr2poly
2960
2961 bntest.c has additional tests for binary polynomial arithmetic.
2962
2963 Two implementations for BN_GF2m_mod_div() are available.
2964 The default algorithm simply uses BN_GF2m_mod_inv() and
2965 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2966 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2967 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2968
2969 [Sheueling Chang Shantz and Douglas Stebila
2970 (Sun Microsystems Laboratories)]
2971
2972 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2973 functionality is disabled at compile-time.
2974 [Douglas Stebila <douglas.stebila@sun.com>]
2975
2976 *) Change default behaviour of 'openssl asn1parse' so that more
2977 information is visible when viewing, e.g., a certificate:
2978
2979 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2980 mode the content of non-printable OCTET STRINGs is output in a
2981 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2982 avoid the appearance of a printable string.
2983 [Nils Larsch <nla@trustcenter.de>]
2984
2985 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2986 functions
2987 EC_GROUP_set_asn1_flag()
2988 EC_GROUP_get_asn1_flag()
2989 EC_GROUP_set_point_conversion_form()
2990 EC_GROUP_get_point_conversion_form()
2991 These control ASN1 encoding details:
2992 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2993 has been set to OPENSSL_EC_NAMED_CURVE.
2994 - Points are encoded in uncompressed form by default; options for
2995 asn1_for are as for point2oct, namely
2996 POINT_CONVERSION_COMPRESSED
2997 POINT_CONVERSION_UNCOMPRESSED
2998 POINT_CONVERSION_HYBRID
2999
3000 Also add 'seed' and 'seed_len' members to EC_GROUP with access
3001 functions
3002 EC_GROUP_set_seed()
3003 EC_GROUP_get0_seed()
3004 EC_GROUP_get_seed_len()
3005 This is used only for ASN1 purposes (so far).
3006 [Nils Larsch <nla@trustcenter.de>]
3007
3008 *) Add 'field_type' member to EC_METHOD, which holds the NID
3009 of the appropriate field type OID. The new function
3010 EC_METHOD_get_field_type() returns this value.
3011 [Nils Larsch <nla@trustcenter.de>]
3012
3013 *) Add functions
3014 EC_POINT_point2bn()
3015 EC_POINT_bn2point()
3016 EC_POINT_point2hex()
3017 EC_POINT_hex2point()
3018 providing useful interfaces to EC_POINT_point2oct() and
3019 EC_POINT_oct2point().
3020 [Nils Larsch <nla@trustcenter.de>]
3021
3022 *) Change internals of the EC library so that the functions
3023 EC_GROUP_set_generator()
3024 EC_GROUP_get_generator()
3025 EC_GROUP_get_order()
3026 EC_GROUP_get_cofactor()
3027 are implemented directly in crypto/ec/ec_lib.c and not dispatched
3028 to methods, which would lead to unnecessary code duplication when
3029 adding different types of curves.
3030 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
3031
3032 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
3033 arithmetic, and such that modified wNAFs are generated
3034 (which avoid length expansion in many cases).
3035 [Bodo Moeller]
3036
3037 *) Add a function EC_GROUP_check_discriminant() (defined via
3038 EC_METHOD) that verifies that the curve discriminant is non-zero.
3039
3040 Add a function EC_GROUP_check() that makes some sanity tests
3041 on a EC_GROUP, its generator and order. This includes
3042 EC_GROUP_check_discriminant().
3043 [Nils Larsch <nla@trustcenter.de>]
3044
3045 *) Add ECDSA in new directory crypto/ecdsa/.
3046
3047 Add applications 'openssl ecparam' and 'openssl ecdsa'
3048 (these are based on 'openssl dsaparam' and 'openssl dsa').
3049
3050 ECDSA support is also included in various other files across the
3051 library. Most notably,
3052 - 'openssl req' now has a '-newkey ecdsa:file' option;
3053 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
3054 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
3055 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
3056 them suitable for ECDSA where domain parameters must be
3057 extracted before the specific public key;
3058 - ECDSA engine support has been added.
3059 [Nils Larsch <nla@trustcenter.de>]
3060
3061 *) Include some named elliptic curves, and add OIDs from X9.62,
3062 SECG, and WAP/WTLS. Each curve can be obtained from the new
3063 function
3064 EC_GROUP_new_by_curve_name(),
3065 and the list of available named curves can be obtained with
3066 EC_get_builtin_curves().
3067 Also add a 'curve_name' member to EC_GROUP objects, which can be
3068 accessed via
3069 EC_GROUP_set_curve_name()
3070 EC_GROUP_get_curve_name()
3071 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
3072
3073 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
3074 was actually never needed) and in BN_mul(). The removal in BN_mul()
3075 required a small change in bn_mul_part_recursive() and the addition
3076 of the functions bn_cmp_part_words(), bn_sub_part_words() and
3077 bn_add_part_words(), which do the same thing as bn_cmp_words(),
3078 bn_sub_words() and bn_add_words() except they take arrays with
3079 differing sizes.
3080 [Richard Levitte]
3081
3082 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
3083
3084 *) Cleanse PEM buffers before freeing them since they may contain
3085 sensitive data.
3086 [Benjamin Bennett <ben@psc.edu>]
3087
3088 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3089 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3090 authentication-only ciphersuites.
3091 [Bodo Moeller]
3092
3093 *) Since AES128 and AES256 share a single mask bit in the logic of
3094 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3095 kludge to work properly if AES128 is available and AES256 isn't.
3096 [Victor Duchovni]
3097
3098 *) Expand security boundary to match 1.1.1 module.
3099 [Steve Henson]
3100
3101 *) Remove redundant features: hash file source, editing of test vectors
3102 modify fipsld to use external fips_premain.c signature.
3103 [Steve Henson]
3104
3105 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
3106 run algorithm test programs.
3107 [Steve Henson]
3108
3109 *) Make algorithm test programs more tolerant of whitespace.
3110 [Steve Henson]
3111
3112 *) Have SSL/TLS server implementation tolerate "mismatched" record
3113 protocol version while receiving ClientHello even if the
3114 ClientHello is fragmented. (The server can't insist on the
3115 particular protocol version it has chosen before the ServerHello
3116 message has informed the client about his choice.)
3117 [Bodo Moeller]
3118
3119 *) Load error codes if they are not already present instead of using a
3120 static variable. This allows them to be cleanly unloaded and reloaded.
3121 [Steve Henson]
3122
3123 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
3124
3125 *) Introduce limits to prevent malicious keys being able to
3126 cause a denial of service. (CVE-2006-2940)
3127 [Steve Henson, Bodo Moeller]
3128
3129 *) Fix ASN.1 parsing of certain invalid structures that can result
3130 in a denial of service. (CVE-2006-2937) [Steve Henson]
3131
3132 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3133 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3134
3135 *) Fix SSL client code which could crash if connecting to a
3136 malicious SSLv2 server. (CVE-2006-4343)
3137 [Tavis Ormandy and Will Drewry, Google Security Team]
3138
3139 *) Change ciphersuite string processing so that an explicit
3140 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
3141 will no longer include "AES128-SHA"), and any other similar
3142 ciphersuite (same bitmap) from *other* protocol versions (so that
3143 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
3144 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
3145 changes from 0.9.8b and 0.9.8d.
3146 [Bodo Moeller]
3147
3148 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
3149
3150 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3151 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3152
3153 *) Change the Unix randomness entropy gathering to use poll() when
3154 possible instead of select(), since the latter has some
3155 undesirable limitations.
3156 [Darryl Miles via Richard Levitte and Bodo Moeller]
3157
3158 *) Disable rogue ciphersuites:
3159
3160 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3161 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3162 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3163
3164 The latter two were purportedly from
3165 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3166 appear there.
3167
3168 Also deactive the remaining ciphersuites from
3169 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3170 unofficial, and the ID has long expired.
3171 [Bodo Moeller]
3172
3173 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3174 dual-core machines) and other potential thread-safety issues.
3175 [Bodo Moeller]
3176
3177 Changes between 0.9.7i and 0.9.7j [04 May 2006]
3178
3179 *) Adapt fipsld and the build system to link against the validated FIPS
3180 module in FIPS mode.
3181 [Steve Henson]
3182
3183 *) Fixes for VC++ 2005 build under Windows.
3184 [Steve Henson]
3185
3186 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
3187 from a Windows bash shell such as MSYS. It is autodetected from the
3188 "config" script when run from a VC++ environment. Modify standard VC++
3189 build to use fipscanister.o from the GNU make build.
3190 [Steve Henson]
3191
3192 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
3193
3194 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
3195 The value now differs depending on if you build for FIPS or not.
3196 BEWARE! A program linked with a shared FIPSed libcrypto can't be
3197 safely run with a non-FIPSed libcrypto, as it may crash because of
3198 the difference induced by this change.
3199 [Andy Polyakov]
3200
3201 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
3202
3203 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3204 (part of SSL_OP_ALL). This option used to disable the
3205 countermeasure against man-in-the-middle protocol-version
3206 rollback in the SSL 2.0 server implementation, which is a bad
3207 idea. (CVE-2005-2969)
3208
3209 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3210 for Information Security, National Institute of Advanced Industrial
3211 Science and Technology [AIST], Japan)]
3212
3213 *) Minimal support for X9.31 signatures and PSS padding modes. This is
3214 mainly for FIPS compliance and not fully integrated at this stage.
3215 [Steve Henson]
3216
3217 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
3218 the exponentiation using a fixed-length exponent. (Otherwise,
3219 the information leaked through timing could expose the secret key
3220 after many signatures; cf. Bleichenbacher's attack on DSA with
3221 biased k.)
3222 [Bodo Moeller]
3223
3224 *) Make a new fixed-window mod_exp implementation the default for
3225 RSA, DSA, and DH private-key operations so that the sequence of
3226 squares and multiplies and the memory access pattern are
3227 independent of the particular secret key. This will mitigate
3228 cache-timing and potential related attacks.
3229
3230 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
3231 and this is automatically used by BN_mod_exp_mont() if the new flag
3232 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
3233 will use this BN flag for private exponents unless the flag
3234 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
3235 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
3236
3237 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
3238
3239 *) Change the client implementation for SSLv23_method() and
3240 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
3241 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
3242 (Previously, the SSL 2.0 backwards compatible Client Hello
3243 message format would be used even with SSL_OP_NO_SSLv2.)
3244 [Bodo Moeller]
3245
3246 *) Add support for smime-type MIME parameter in S/MIME messages which some
3247 clients need.
3248 [Steve Henson]
3249
3250 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
3251 a threadsafe manner. Modify rsa code to use new function and add calls
3252 to dsa and dh code (which had race conditions before).
3253 [Steve Henson]
3254
3255 *) Include the fixed error library code in the C error file definitions
3256 instead of fixing them up at runtime. This keeps the error code
3257 structures constant.
3258 [Steve Henson]
3259
3260 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
3261
3262 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
3263 OpenSSL 0.9.8.]
3264
3265 *) Fixes for newer kerberos headers. NB: the casts are needed because
3266 the 'length' field is signed on one version and unsigned on another
3267 with no (?) obvious way to tell the difference, without these VC++
3268 complains. Also the "definition" of FAR (blank) is no longer included
3269 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
3270 some needed definitions.
3271 [Steve Henson]
3272
3273 *) Undo Cygwin change.
3274 [Ulf Möller]
3275
3276 *) Added support for proxy certificates according to RFC 3820.
3277 Because they may be a security thread to unaware applications,
3278 they must be explicitely allowed in run-time. See
3279 docs/HOWTO/proxy_certificates.txt for further information.
3280 [Richard Levitte]
3281
3282 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
3283
3284 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
3285 server and client random values. Previously
3286 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
3287 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
3288
3289 This change has negligible security impact because:
3290
3291 1. Server and client random values still have 24 bytes of pseudo random
3292 data.
3293
3294 2. Server and client random values are sent in the clear in the initial
3295 handshake.
3296
3297 3. The master secret is derived using the premaster secret (48 bytes in
3298 size for static RSA ciphersuites) as well as client server and random
3299 values.
3300
3301 The OpenSSL team would like to thank the UK NISCC for bringing this issue
3302 to our attention.
3303
3304 [Stephen Henson, reported by UK NISCC]
3305
3306 *) Use Windows randomness collection on Cygwin.
3307 [Ulf Möller]
3308
3309 *) Fix hang in EGD/PRNGD query when communication socket is closed
3310 prematurely by EGD/PRNGD.
3311 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
3312
3313 *) Prompt for pass phrases when appropriate for PKCS12 input format.
3314 [Steve Henson]
3315
3316 *) Back-port of selected performance improvements from development
3317 branch, as well as improved support for PowerPC platforms.
3318 [Andy Polyakov]
3319
3320 *) Add lots of checks for memory allocation failure, error codes to indicate
3321 failure and freeing up memory if a failure occurs.
3322 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
3323
3324 *) Add new -passin argument to dgst.
3325 [Steve Henson]
3326
3327 *) Perform some character comparisons of different types in X509_NAME_cmp:
3328 this is needed for some certificates that reencode DNs into UTF8Strings
3329 (in violation of RFC3280) and can't or wont issue name rollover
3330 certificates.
3331 [Steve Henson]
3332
3333 *) Make an explicit check during certificate validation to see that
3334 the CA setting in each certificate on the chain is correct. As a
3335 side effect always do the following basic checks on extensions,
3336 not just when there's an associated purpose to the check:
3337
3338 - if there is an unhandled critical extension (unless the user
3339 has chosen to ignore this fault)
3340 - if the path length has been exceeded (if one is set at all)
3341 - that certain extensions fit the associated purpose (if one has
3342 been given)
3343 [Richard Levitte]
3344
3345 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
3346
3347 *) Avoid a race condition when CRLs are checked in a multi threaded
3348 environment. This would happen due to the reordering of the revoked
3349 entries during signature checking and serial number lookup. Now the
3350 encoding is cached and the serial number sort performed under a lock.
3351 Add new STACK function sk_is_sorted().
3352 [Steve Henson]
3353
3354 *) Add Delta CRL to the extension code.
3355 [Steve Henson]
3356
3357 *) Various fixes to s3_pkt.c so alerts are sent properly.
3358 [David Holmes <d.holmes@f5.com>]
3359
3360 *) Reduce the chances of duplicate issuer name and serial numbers (in
3361 violation of RFC3280) using the OpenSSL certificate creation utilities.
3362 This is done by creating a random 64 bit value for the initial serial
3363 number when a serial number file is created or when a self signed
3364 certificate is created using 'openssl req -x509'. The initial serial
3365 number file is created using 'openssl x509 -next_serial' in CA.pl
3366 rather than being initialized to 1.
3367 [Steve Henson]
3368
3369 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
3370
3371 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
3372 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
3373 [Joe Orton, Steve Henson]
3374
3375 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
3376 (CVE-2004-0112)
3377 [Joe Orton, Steve Henson]
3378
3379 *) Make it possible to have multiple active certificates with the same
3380 subject in the CA index file. This is done only if the keyword
3381 'unique_subject' is set to 'no' in the main CA section (default
3382 if 'CA_default') of the configuration file. The value is saved
3383 with the database itself in a separate index attribute file,
3384 named like the index file with '.attr' appended to the name.
3385 [Richard Levitte]
3386
3387 *) X509 verify fixes. Disable broken certificate workarounds when
3388 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
3389 keyUsage extension present. Don't accept CRLs with unhandled critical
3390 extensions: since verify currently doesn't process CRL extensions this
3391 rejects a CRL with *any* critical extensions. Add new verify error codes
3392 for these cases.
3393 [Steve Henson]
3394
3395 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
3396 A clarification of RFC2560 will require the use of OCTET STRINGs and
3397 some implementations cannot handle the current raw format. Since OpenSSL
3398 copies and compares OCSP nonces as opaque blobs without any attempt at
3399 parsing them this should not create any compatibility issues.
3400 [Steve Henson]
3401
3402 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
3403 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
3404 this HMAC (and other) operations are several times slower than OpenSSL
3405 < 0.9.7.
3406 [Steve Henson]
3407
3408 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
3409 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3410
3411 *) Use the correct content when signing type "other".
3412 [Steve Henson]
3413
3414 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
3415
3416 *) Fix various bugs revealed by running the NISCC test suite:
3417
3418 Stop out of bounds reads in the ASN1 code when presented with
3419 invalid tags (CVE-2003-0543 and CVE-2003-0544).
3420
3421 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
3422
3423 If verify callback ignores invalid public key errors don't try to check
3424 certificate signature with the NULL public key.
3425
3426 [Steve Henson]
3427
3428 *) New -ignore_err option in ocsp application to stop the server
3429 exiting on the first error in a request.
3430 [Steve Henson]
3431
3432 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
3433 if the server requested one: as stated in TLS 1.0 and SSL 3.0
3434 specifications.
3435 [Steve Henson]
3436
3437 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
3438 extra data after the compression methods not only for TLS 1.0
3439 but also for SSL 3.0 (as required by the specification).
3440 [Bodo Moeller; problem pointed out by Matthias Loepfe]
3441
3442 *) Change X509_certificate_type() to mark the key as exported/exportable
3443 when it's 512 *bits* long, not 512 bytes.
3444 [Richard Levitte]
3445
3446 *) Change AES_cbc_encrypt() so it outputs exact multiple of
3447 blocks during encryption.
3448 [Richard Levitte]
3449
3450 *) Various fixes to base64 BIO and non blocking I/O. On write
3451 flushes were not handled properly if the BIO retried. On read
3452 data was not being buffered properly and had various logic bugs.
3453 This also affects blocking I/O when the data being decoded is a
3454 certain size.
3455 [Steve Henson]
3456
3457 *) Various S/MIME bugfixes and compatibility changes:
3458 output correct application/pkcs7 MIME type if
3459 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
3460 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
3461 of files as .eml work). Correctly handle very long lines in MIME
3462 parser.
3463 [Steve Henson]
3464
3465 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
3466
3467 *) Countermeasure against the Klima-Pokorny-Rosa extension of
3468 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
3469 a protocol version number mismatch like a decryption error
3470 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
3471 [Bodo Moeller]
3472
3473 *) Turn on RSA blinding by default in the default implementation
3474 to avoid a timing attack. Applications that don't want it can call
3475 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
3476 They would be ill-advised to do so in most cases.
3477 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
3478
3479 *) Change RSA blinding code so that it works when the PRNG is not
3480 seeded (in this case, the secret RSA exponent is abused as
3481 an unpredictable seed -- if it is not unpredictable, there
3482 is no point in blinding anyway). Make RSA blinding thread-safe
3483 by remembering the creator's thread ID in rsa->blinding and
3484 having all other threads use local one-time blinding factors
3485 (this requires more computation than sharing rsa->blinding, but
3486 avoids excessive locking; and if an RSA object is not shared
3487 between threads, blinding will still be very fast).
3488 [Bodo Moeller]
3489
3490 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
3491 ENGINE as defaults for all supported algorithms irrespective of
3492 the 'flags' parameter. 'flags' is now honoured, so applications
3493 should make sure they are passing it correctly.
3494 [Geoff Thorpe]
3495
3496 *) Target "mingw" now allows native Windows code to be generated in
3497 the Cygwin environment as well as with the MinGW compiler.
3498 [Ulf Moeller]
3499
3500 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
3501
3502 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
3503 via timing by performing a MAC computation even if incorrrect
3504 block cipher padding has been found. This is a countermeasure
3505 against active attacks where the attacker has to distinguish
3506 between bad padding and a MAC verification error. (CVE-2003-0078)
3507
3508 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
3509 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
3510 Martin Vuagnoux (EPFL, Ilion)]
3511
3512 *) Make the no-err option work as intended. The intention with no-err
3513 is not to have the whole error stack handling routines removed from
3514 libcrypto, it's only intended to remove all the function name and
3515 reason texts, thereby removing some of the footprint that may not
3516 be interesting if those errors aren't displayed anyway.
3517
3518 NOTE: it's still possible for any application or module to have it's
3519 own set of error texts inserted. The routines are there, just not
3520 used by default when no-err is given.
3521 [Richard Levitte]
3522
3523 *) Add support for FreeBSD on IA64.
3524 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
3525
3526 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
3527 Kerberos function mit_des_cbc_cksum(). Before this change,
3528 the value returned by DES_cbc_cksum() was like the one from
3529 mit_des_cbc_cksum(), except the bytes were swapped.
3530 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
3531
3532 *) Allow an application to disable the automatic SSL chain building.
3533 Before this a rather primitive chain build was always performed in
3534 ssl3_output_cert_chain(): an application had no way to send the
3535 correct chain if the automatic operation produced an incorrect result.
3536
3537 Now the chain builder is disabled if either:
3538
3539 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
3540
3541 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
3542
3543 The reasoning behind this is that an application would not want the
3544 auto chain building to take place if extra chain certificates are
3545 present and it might also want a means of sending no additional
3546 certificates (for example the chain has two certificates and the
3547 root is omitted).
3548 [Steve Henson]
3549
3550 *) Add the possibility to build without the ENGINE framework.
3551 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3552
3553 *) Under Win32 gmtime() can return NULL: check return value in
3554 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
3555 [Steve Henson]
3556
3557 *) DSA routines: under certain error conditions uninitialized BN objects
3558 could be freed. Solution: make sure initialization is performed early
3559 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
3560 Nils Larsch <nla@trustcenter.de> via PR#459)
3561 [Lutz Jaenicke]
3562
3563 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
3564 checked on reconnect on the client side, therefore session resumption
3565 could still fail with a "ssl session id is different" error. This
3566 behaviour is masked when SSL_OP_ALL is used due to
3567 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
3568 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3569 followup to PR #377.
3570 [Lutz Jaenicke]
3571
3572 *) IA-32 assembler support enhancements: unified ELF targets, support
3573 for SCO/Caldera platforms, fix for Cygwin shared build.
3574 [Andy Polyakov]
3575
3576 *) Add support for FreeBSD on sparc64. As a consequence, support for
3577 FreeBSD on non-x86 processors is separate from x86 processors on
3578 the config script, much like the NetBSD support.
3579 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
3580
3581 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
3582
3583 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
3584 OpenSSL 0.9.7.]
3585
3586 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
3587 code (06) was taken as the first octet of the session ID and the last
3588 octet was ignored consequently. As a result SSLv2 client side session
3589 caching could not have worked due to the session ID mismatch between
3590 client and server.
3591 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
3592 PR #377.
3593 [Lutz Jaenicke]
3594
3595 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
3596 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
3597 removed entirely.
3598 [Richard Levitte]
3599
3600 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
3601 seems that in spite of existing for more than a year, many application
3602 author have done nothing to provide the necessary callbacks, which
3603 means that this particular engine will not work properly anywhere.
3604 This is a very unfortunate situation which forces us, in the name
3605 of usability, to give the hw_ncipher.c a static lock, which is part
3606 of libcrypto.
3607 NOTE: This is for the 0.9.7 series ONLY. This hack will never
3608 appear in 0.9.8 or later. We EXPECT application authors to have
3609 dealt properly with this when 0.9.8 is released (unless we actually
3610 make such changes in the libcrypto locking code that changes will
3611 have to be made anyway).
3612 [Richard Levitte]
3613
3614 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
3615 octets have been read, EOF or an error occurs. Without this change
3616 some truncated ASN1 structures will not produce an error.
3617 [Steve Henson]
3618
3619 *) Disable Heimdal support, since it hasn't been fully implemented.
3620 Still give the possibility to force the use of Heimdal, but with
3621 warnings and a request that patches get sent to openssl-dev.
3622 [Richard Levitte]
3623
3624 *) Add the VC-CE target, introduce the WINCE sysname, and add
3625 INSTALL.WCE and appropriate conditionals to make it build.
3626 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
3627
3628 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
3629 cygssl-x.y.z.dll, where x, y and z are the major, minor and
3630 edit numbers of the version.
3631 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
3632
3633 *) Introduce safe string copy and catenation functions
3634 (BUF_strlcpy() and BUF_strlcat()).
3635 [Ben Laurie (CHATS) and Richard Levitte]
3636
3637 *) Avoid using fixed-size buffers for one-line DNs.
3638 [Ben Laurie (CHATS)]
3639
3640 *) Add BUF_MEM_grow_clean() to avoid information leakage when
3641 resizing buffers containing secrets, and use where appropriate.
3642 [Ben Laurie (CHATS)]
3643
3644 *) Avoid using fixed size buffers for configuration file location.
3645 [Ben Laurie (CHATS)]
3646
3647 *) Avoid filename truncation for various CA files.
3648 [Ben Laurie (CHATS)]
3649
3650 *) Use sizeof in preference to magic numbers.
3651 [Ben Laurie (CHATS)]
3652
3653 *) Avoid filename truncation in cert requests.
3654 [Ben Laurie (CHATS)]
3655
3656 *) Add assertions to check for (supposedly impossible) buffer
3657 overflows.
3658 [Ben Laurie (CHATS)]
3659
3660 *) Don't cache truncated DNS entries in the local cache (this could
3661 potentially lead to a spoofing attack).
3662 [Ben Laurie (CHATS)]
3663
3664 *) Fix various buffers to be large enough for hex/decimal
3665 representations in a platform independent manner.
3666 [Ben Laurie (CHATS)]
3667
3668 *) Add CRYPTO_realloc_clean() to avoid information leakage when
3669 resizing buffers containing secrets, and use where appropriate.
3670 [Ben Laurie (CHATS)]
3671
3672 *) Add BIO_indent() to avoid much slightly worrying code to do
3673 indents.
3674 [Ben Laurie (CHATS)]
3675
3676 *) Convert sprintf()/BIO_puts() to BIO_printf().
3677 [Ben Laurie (CHATS)]
3678
3679 *) buffer_gets() could terminate with the buffer only half
3680 full. Fixed.
3681 [Ben Laurie (CHATS)]
3682
3683 *) Add assertions to prevent user-supplied crypto functions from
3684 overflowing internal buffers by having large block sizes, etc.
3685 [Ben Laurie (CHATS)]
3686
3687 *) New OPENSSL_assert() macro (similar to assert(), but enabled
3688 unconditionally).
3689 [Ben Laurie (CHATS)]
3690
3691 *) Eliminate unused copy of key in RC4.
3692 [Ben Laurie (CHATS)]
3693
3694 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
3695 [Ben Laurie (CHATS)]
3696
3697 *) Fix off-by-one error in EGD path.
3698 [Ben Laurie (CHATS)]
3699
3700 *) If RANDFILE path is too long, ignore instead of truncating.
3701 [Ben Laurie (CHATS)]
3702
3703 *) Eliminate unused and incorrectly sized X.509 structure
3704 CBCParameter.
3705 [Ben Laurie (CHATS)]
3706
3707 *) Eliminate unused and dangerous function knumber().
3708 [Ben Laurie (CHATS)]
3709
3710 *) Eliminate unused and dangerous structure, KSSL_ERR.
3711 [Ben Laurie (CHATS)]
3712
3713 *) Protect against overlong session ID context length in an encoded
3714 session object. Since these are local, this does not appear to be
3715 exploitable.
3716 [Ben Laurie (CHATS)]
3717
3718 *) Change from security patch (see 0.9.6e below) that did not affect
3719 the 0.9.6 release series:
3720
3721 Remote buffer overflow in SSL3 protocol - an attacker could
3722 supply an oversized master key in Kerberos-enabled versions.
3723 (CVE-2002-0657)
3724 [Ben Laurie (CHATS)]
3725
3726 *) Change the SSL kerb5 codes to match RFC 2712.
3727 [Richard Levitte]
3728
3729 *) Make -nameopt work fully for req and add -reqopt switch.
3730 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
3731
3732 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
3733 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
3734
3735 *) Make sure tests can be performed even if the corresponding algorithms
3736 have been removed entirely. This was also the last step to make
3737 OpenSSL compilable with DJGPP under all reasonable conditions.
3738 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
3739
3740 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
3741 to allow version independent disabling of normally unselected ciphers,
3742 which may be activated as a side-effect of selecting a single cipher.
3743
3744 (E.g., cipher list string "RSA" enables ciphersuites that are left
3745 out of "ALL" because they do not provide symmetric encryption.
3746 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
3747 [Lutz Jaenicke, Bodo Moeller]
3748
3749 *) Add appropriate support for separate platform-dependent build
3750 directories. The recommended way to make a platform-dependent
3751 build directory is the following (tested on Linux), maybe with
3752 some local tweaks:
3753
3754 # Place yourself outside of the OpenSSL source tree. In
3755 # this example, the environment variable OPENSSL_SOURCE
3756 # is assumed to contain the absolute OpenSSL source directory.
3757 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
3758 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
3759 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
3760 mkdir -p `dirname $F`
3761 ln -s $OPENSSL_SOURCE/$F $F
3762 done
3763
3764 To be absolutely sure not to disturb the source tree, a "make clean"
3765 is a good thing. If it isn't successfull, don't worry about it,
3766 it probably means the source directory is very clean.
3767 [Richard Levitte]
3768
3769 *) Make sure any ENGINE control commands make local copies of string
3770 pointers passed to them whenever necessary. Otherwise it is possible
3771 the caller may have overwritten (or deallocated) the original string
3772 data when a later ENGINE operation tries to use the stored values.
3773 [Götz Babin-Ebell <babinebell@trustcenter.de>]
3774
3775 *) Improve diagnostics in file reading and command-line digests.
3776 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
3777
3778 *) Add AES modes CFB and OFB to the object database. Correct an
3779 error in AES-CFB decryption.
3780 [Richard Levitte]
3781
3782 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
3783 allows existing EVP_CIPHER_CTX structures to be reused after
3784 calling EVP_*Final(). This behaviour is used by encryption
3785 BIOs and some applications. This has the side effect that
3786 applications must explicitly clean up cipher contexts with
3787 EVP_CIPHER_CTX_cleanup() or they will leak memory.
3788 [Steve Henson]
3789
3790 *) Check the values of dna and dnb in bn_mul_recursive before calling
3791 bn_mul_comba (a non zero value means the a or b arrays do not contain
3792 n2 elements) and fallback to bn_mul_normal if either is not zero.
3793 [Steve Henson]
3794
3795 *) Fix escaping of non-ASCII characters when using the -subj option
3796 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
3797 [Lutz Jaenicke]
3798
3799 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
3800 form for "surname", serialNumber has no short form.
3801 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
3802 therefore remove "mail" short name for "internet 7".
3803 The OID for unique identifiers in X509 certificates is
3804 x500UniqueIdentifier, not uniqueIdentifier.
3805 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
3806 [Lutz Jaenicke]
3807
3808 *) Add an "init" command to the ENGINE config module and auto initialize
3809 ENGINEs. Without any "init" command the ENGINE will be initialized
3810 after all ctrl commands have been executed on it. If init=1 the
3811 ENGINE is initailized at that point (ctrls before that point are run
3812 on the uninitialized ENGINE and after on the initialized one). If
3813 init=0 then the ENGINE will not be iniatialized at all.
3814 [Steve Henson]
3815
3816 *) Fix the 'app_verify_callback' interface so that the user-defined
3817 argument is actually passed to the callback: In the
3818 SSL_CTX_set_cert_verify_callback() prototype, the callback
3819 declaration has been changed from
3820 int (*cb)()
3821 into
3822 int (*cb)(X509_STORE_CTX *,void *);
3823 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
3824 i=s->ctx->app_verify_callback(&ctx)
3825 has been changed into
3826 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
3827
3828 To update applications using SSL_CTX_set_cert_verify_callback(),
3829 a dummy argument can be added to their callback functions.
3830 [D. K. Smetters <smetters@parc.xerox.com>]
3831
3832 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
3833 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
3834
3835 *) Add and OPENSSL_LOAD_CONF define which will cause
3836 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
3837 This allows older applications to transparently support certain
3838 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
3839 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
3840 load the config file and OPENSSL_add_all_algorithms_conf() which will
3841 always load it have also been added.
3842 [Steve Henson]
3843
3844 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
3845 Adjust NIDs and EVP layer.
3846 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3847
3848 *) Config modules support in openssl utility.
3849
3850 Most commands now load modules from the config file,
3851 though in a few (such as version) this isn't done
3852 because it couldn't be used for anything.
3853
3854 In the case of ca and req the config file used is
3855 the same as the utility itself: that is the -config
3856 command line option can be used to specify an
3857 alternative file.
3858 [Steve Henson]
3859
3860 *) Move default behaviour from OPENSSL_config(). If appname is NULL
3861 use "openssl_conf" if filename is NULL use default openssl config file.
3862 [Steve Henson]
3863
3864 *) Add an argument to OPENSSL_config() to allow the use of an alternative
3865 config section name. Add a new flag to tolerate a missing config file
3866 and move code to CONF_modules_load_file().
3867 [Steve Henson]
3868
3869 *) Support for crypto accelerator cards from Accelerated Encryption
3870 Processing, www.aep.ie. (Use engine 'aep')
3871 The support was copied from 0.9.6c [engine] and adapted/corrected
3872 to work with the new engine framework.
3873 [AEP Inc. and Richard Levitte]
3874
3875 *) Support for SureWare crypto accelerator cards from Baltimore
3876 Technologies. (Use engine 'sureware')
3877 The support was copied from 0.9.6c [engine] and adapted
3878 to work with the new engine framework.
3879 [Richard Levitte]
3880
3881 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
3882 make the newer ENGINE framework commands for the CHIL engine work.
3883 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
3884
3885 *) Make it possible to produce shared libraries on ReliantUNIX.
3886 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
3887
3888 *) Add the configuration target debug-linux-ppro.
3889 Make 'openssl rsa' use the general key loading routines
3890 implemented in apps.c, and make those routines able to
3891 handle the key format FORMAT_NETSCAPE and the variant
3892 FORMAT_IISSGC.
3893 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3894
3895 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
3896 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
3897
3898 *) Add -keyform to rsautl, and document -engine.
3899 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
3900
3901 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
3902 BIO_R_NO_SUCH_FILE error code rather than the generic
3903 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
3904 [Ben Laurie]
3905
3906 *) Add new functions
3907 ERR_peek_last_error
3908 ERR_peek_last_error_line
3909 ERR_peek_last_error_line_data.
3910 These are similar to
3911 ERR_peek_error
3912 ERR_peek_error_line
3913 ERR_peek_error_line_data,
3914 but report on the latest error recorded rather than the first one
3915 still in the error queue.
3916 [Ben Laurie, Bodo Moeller]
3917
3918 *) default_algorithms option in ENGINE config module. This allows things
3919 like:
3920 default_algorithms = ALL
3921 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
3922 [Steve Henson]
3923
3924 *) Prelminary ENGINE config module.
3925 [Steve Henson]
3926
3927 *) New experimental application configuration code.
3928 [Steve Henson]
3929
3930 *) Change the AES code to follow the same name structure as all other
3931 symmetric ciphers, and behave the same way. Move everything to
3932 the directory crypto/aes, thereby obsoleting crypto/rijndael.
3933 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
3934
3935 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
3936 [Ben Laurie and Theo de Raadt]
3937
3938 *) Add option to output public keys in req command.
3939 [Massimiliano Pala madwolf@openca.org]
3940
3941 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
3942 (up to about 10% better than before for P-192 and P-224).
3943 [Bodo Moeller]
3944
3945 *) New functions/macros
3946
3947 SSL_CTX_set_msg_callback(ctx, cb)
3948 SSL_CTX_set_msg_callback_arg(ctx, arg)
3949 SSL_set_msg_callback(ssl, cb)
3950 SSL_set_msg_callback_arg(ssl, arg)
3951
3952 to request calling a callback function
3953
3954 void cb(int write_p, int version, int content_type,
3955 const void *buf, size_t len, SSL *ssl, void *arg)
3956
3957 whenever a protocol message has been completely received
3958 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3959 protocol version according to which the SSL library interprets
3960 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3961 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3962 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3963 specification (change_cipher_spec(20), alert(21), handshake(22)).
3964 'buf' and 'len' point to the actual message, 'ssl' to the
3965 SSL object, and 'arg' is the application-defined value set by
3966 SSL[_CTX]_set_msg_callback_arg().
3967
3968 'openssl s_client' and 'openssl s_server' have new '-msg' options
3969 to enable a callback that displays all protocol messages.
3970 [Bodo Moeller]
3971
3972 *) Change the shared library support so shared libraries are built as
3973 soon as the corresponding static library is finished, and thereby get
3974 openssl and the test programs linked against the shared library.
3975 This still only happens when the keyword "shard" has been given to
3976 the configuration scripts.
3977
3978 NOTE: shared library support is still an experimental thing, and
3979 backward binary compatibility is still not guaranteed.
3980 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3981
3982 *) Add support for Subject Information Access extension.
3983 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3984
3985 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3986 additional bytes when new memory had to be allocated, not just
3987 when reusing an existing buffer.
3988 [Bodo Moeller]
3989
3990 *) New command line and configuration option 'utf8' for the req command.
3991 This allows field values to be specified as UTF8 strings.
3992 [Steve Henson]
3993
3994 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3995 runs for the former and machine-readable output for the latter.
3996 [Ben Laurie]
3997
3998 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3999 of the e-mail address in the DN (i.e., it will go into a certificate
4000 extension only). The new configuration file option 'email_in_dn = no'
4001 has the same effect.
4002 [Massimiliano Pala madwolf@openca.org]
4003
4004 *) Change all functions with names starting with des_ to be starting
4005 with DES_ instead. Add wrappers that are compatible with libdes,
4006 but are named _ossl_old_des_*. Finally, add macros that map the
4007 des_* symbols to the corresponding _ossl_old_des_* if libdes
4008 compatibility is desired. If OpenSSL 0.9.6c compatibility is
4009 desired, the des_* symbols will be mapped to DES_*, with one
4010 exception.
4011
4012 Since we provide two compatibility mappings, the user needs to
4013 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
4014 compatibility is desired. The default (i.e., when that macro
4015 isn't defined) is OpenSSL 0.9.6c compatibility.
4016
4017 There are also macros that enable and disable the support of old
4018 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
4019 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
4020 are defined, the default will apply: to support the old des routines.
4021
4022 In either case, one must include openssl/des.h to get the correct
4023 definitions. Do not try to just include openssl/des_old.h, that
4024 won't work.
4025
4026 NOTE: This is a major break of an old API into a new one. Software
4027 authors are encouraged to switch to the DES_ style functions. Some
4028 time in the future, des_old.h and the libdes compatibility functions
4029 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
4030 default), and then completely removed.
4031 [Richard Levitte]
4032
4033 *) Test for certificates which contain unsupported critical extensions.
4034 If such a certificate is found during a verify operation it is
4035 rejected by default: this behaviour can be overridden by either
4036 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
4037 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
4038 X509_supported_extension() has also been added which returns 1 if a
4039 particular extension is supported.
4040 [Steve Henson]
4041
4042 *) Modify the behaviour of EVP cipher functions in similar way to digests
4043 to retain compatibility with existing code.
4044 [Steve Henson]
4045
4046 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
4047 compatibility with existing code. In particular the 'ctx' parameter does
4048 not have to be to be initialized before the call to EVP_DigestInit() and
4049 it is tidied up after a call to EVP_DigestFinal(). New function
4050 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
4051 EVP_MD_CTX_copy() changed to not require the destination to be
4052 initialized valid and new function EVP_MD_CTX_copy_ex() added which
4053 requires the destination to be valid.
4054
4055 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
4056 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
4057 [Steve Henson]
4058
4059 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
4060 so that complete 'Handshake' protocol structures are kept in memory
4061 instead of overwriting 'msg_type' and 'length' with 'body' data.
4062 [Bodo Moeller]
4063
4064 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
4065 [Massimo Santin via Richard Levitte]
4066
4067 *) Major restructuring to the underlying ENGINE code. This includes
4068 reduction of linker bloat, separation of pure "ENGINE" manipulation
4069 (initialisation, etc) from functionality dealing with implementations
4070 of specific crypto iterfaces. This change also introduces integrated
4071 support for symmetric ciphers and digest implementations - so ENGINEs
4072 can now accelerate these by providing EVP_CIPHER and EVP_MD
4073 implementations of their own. This is detailed in crypto/engine/README
4074 as it couldn't be adequately described here. However, there are a few
4075 API changes worth noting - some RSA, DSA, DH, and RAND functions that
4076 were changed in the original introduction of ENGINE code have now
4077 reverted back - the hooking from this code to ENGINE is now a good
4078 deal more passive and at run-time, operations deal directly with
4079 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
4080 dereferencing through an ENGINE pointer any more. Also, the ENGINE
4081 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
4082 they were not being used by the framework as there is no concept of a
4083 BIGNUM_METHOD and they could not be generalised to the new
4084 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
4085 ENGINE_cpy() has been removed as it cannot be consistently defined in
4086 the new code.
4087 [Geoff Thorpe]
4088
4089 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
4090 [Steve Henson]
4091
4092 *) Change mkdef.pl to sort symbols that get the same entry number,
4093 and make sure the automatically generated functions ERR_load_*
4094 become part of libeay.num as well.
4095 [Richard Levitte]
4096
4097 *) New function SSL_renegotiate_pending(). This returns true once
4098 renegotiation has been requested (either SSL_renegotiate() call
4099 or HelloRequest/ClientHello receveived from the peer) and becomes
4100 false once a handshake has been completed.
4101 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
4102 sends a HelloRequest, but does not ensure that a handshake takes
4103 place. SSL_renegotiate_pending() is useful for checking if the
4104 client has followed the request.)
4105 [Bodo Moeller]
4106
4107 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
4108 By default, clients may request session resumption even during
4109 renegotiation (if session ID contexts permit); with this option,
4110 session resumption is possible only in the first handshake.
4111
4112 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
4113 more bits available for options that should not be part of
4114 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
4115 [Bodo Moeller]
4116
4117 *) Add some demos for certificate and certificate request creation.
4118 [Steve Henson]
4119
4120 *) Make maximum certificate chain size accepted from the peer application
4121 settable (SSL*_get/set_max_cert_list()), as proposed by
4122 "Douglas E. Engert" <deengert@anl.gov>.
4123 [Lutz Jaenicke]
4124
4125 *) Add support for shared libraries for Unixware-7
4126 (Boyd Lynn Gerber <gerberb@zenez.com>).
4127 [Lutz Jaenicke]
4128
4129 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
4130 be done prior to destruction. Use this to unload error strings from
4131 ENGINEs that load their own error strings. NB: This adds two new API
4132 functions to "get" and "set" this destroy handler in an ENGINE.
4133 [Geoff Thorpe]
4134
4135 *) Alter all existing ENGINE implementations (except "openssl" and
4136 "openbsd") to dynamically instantiate their own error strings. This
4137 makes them more flexible to be built both as statically-linked ENGINEs
4138 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
4139 Also, add stub code to each that makes building them as self-contained
4140 shared-libraries easier (see README.ENGINE).
4141 [Geoff Thorpe]
4142
4143 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
4144 implementations into applications that are completely implemented in
4145 self-contained shared-libraries. The "dynamic" ENGINE exposes control
4146 commands that can be used to configure what shared-library to load and
4147 to control aspects of the way it is handled. Also, made an update to
4148 the README.ENGINE file that brings its information up-to-date and
4149 provides some information and instructions on the "dynamic" ENGINE
4150 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
4151 [Geoff Thorpe]
4152
4153 *) Make it possible to unload ranges of ERR strings with a new
4154 "ERR_unload_strings" function.
4155 [Geoff Thorpe]
4156
4157 *) Add a copy() function to EVP_MD.
4158 [Ben Laurie]
4159
4160 *) Make EVP_MD routines take a context pointer instead of just the
4161 md_data void pointer.
4162 [Ben Laurie]
4163
4164 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
4165 that the digest can only process a single chunk of data
4166 (typically because it is provided by a piece of
4167 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
4168 is only going to provide a single chunk of data, and hence the
4169 framework needn't accumulate the data for oneshot drivers.
4170 [Ben Laurie]
4171
4172 *) As with "ERR", make it possible to replace the underlying "ex_data"
4173 functions. This change also alters the storage and management of global
4174 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
4175 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
4176 index counters. The API functions that use this state have been changed
4177 to take a "class_index" rather than pointers to the class's local STACK
4178 and counter, and there is now an API function to dynamically create new
4179 classes. This centralisation allows us to (a) plug a lot of the
4180 thread-safety problems that existed, and (b) makes it possible to clean
4181 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
4182 such data would previously have always leaked in application code and
4183 workarounds were in place to make the memory debugging turn a blind eye
4184 to it. Application code that doesn't use this new function will still
4185 leak as before, but their memory debugging output will announce it now
4186 rather than letting it slide.
4187
4188 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
4189 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
4190 has a return value to indicate success or failure.
4191 [Geoff Thorpe]
4192
4193 *) Make it possible to replace the underlying "ERR" functions such that the
4194 global state (2 LHASH tables and 2 locks) is only used by the "default"
4195 implementation. This change also adds two functions to "get" and "set"
4196 the implementation prior to it being automatically set the first time
4197 any other ERR function takes place. Ie. an application can call "get",
4198 pass the return value to a module it has just loaded, and that module
4199 can call its own "set" function using that value. This means the
4200 module's "ERR" operations will use (and modify) the error state in the
4201 application and not in its own statically linked copy of OpenSSL code.
4202 [Geoff Thorpe]
4203
4204 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
4205 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
4206 the operation, and provides a more encapsulated way for external code
4207 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
4208 to use these functions rather than manually incrementing the counts.
4209
4210 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
4211 [Geoff Thorpe]
4212
4213 *) Add EVP test program.
4214 [Ben Laurie]
4215
4216 *) Add symmetric cipher support to ENGINE. Expect the API to change!
4217 [Ben Laurie]
4218
4219 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
4220 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
4221 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
4222 These allow a CRL to be built without having to access X509_CRL fields
4223 directly. Modify 'ca' application to use new functions.
4224 [Steve Henson]
4225
4226 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
4227 bug workarounds. Rollback attack detection is a security feature.
4228 The problem will only arise on OpenSSL servers when TLSv1 is not
4229 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
4230 Software authors not wanting to support TLSv1 will have special reasons
4231 for their choice and can explicitly enable this option.
4232 [Bodo Moeller, Lutz Jaenicke]
4233
4234 *) Rationalise EVP so it can be extended: don't include a union of
4235 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
4236 (similar to those existing for EVP_CIPHER_CTX).
4237 Usage example:
4238
4239 EVP_MD_CTX md;
4240
4241 EVP_MD_CTX_init(&md); /* new function call */
4242 EVP_DigestInit(&md, EVP_sha1());
4243 EVP_DigestUpdate(&md, in, len);
4244 EVP_DigestFinal(&md, out, NULL);
4245 EVP_MD_CTX_cleanup(&md); /* new function call */
4246
4247 [Ben Laurie]
4248
4249 *) Make DES key schedule conform to the usual scheme, as well as
4250 correcting its structure. This means that calls to DES functions
4251 now have to pass a pointer to a des_key_schedule instead of a
4252 plain des_key_schedule (which was actually always a pointer
4253 anyway): E.g.,
4254
4255 des_key_schedule ks;
4256
4257 des_set_key_checked(..., &ks);
4258 des_ncbc_encrypt(..., &ks, ...);
4259
4260 (Note that a later change renames 'des_...' into 'DES_...'.)
4261 [Ben Laurie]
4262
4263 *) Initial reduction of linker bloat: the use of some functions, such as
4264 PEM causes large amounts of unused functions to be linked in due to
4265 poor organisation. For example pem_all.c contains every PEM function
4266 which has a knock on effect of linking in large amounts of (unused)
4267 ASN1 code. Grouping together similar functions and splitting unrelated
4268 functions prevents this.
4269 [Steve Henson]
4270
4271 *) Cleanup of EVP macros.
4272 [Ben Laurie]
4273
4274 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
4275 correct _ecb suffix.
4276 [Ben Laurie]
4277
4278 *) Add initial OCSP responder support to ocsp application. The
4279 revocation information is handled using the text based index
4280 use by the ca application. The responder can either handle
4281 requests generated internally, supplied in files (for example
4282 via a CGI script) or using an internal minimal server.
4283 [Steve Henson]
4284
4285 *) Add configuration choices to get zlib compression for TLS.
4286 [Richard Levitte]
4287
4288 *) Changes to Kerberos SSL for RFC 2712 compliance:
4289 1. Implemented real KerberosWrapper, instead of just using
4290 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
4291 2. Implemented optional authenticator field of KerberosWrapper.
4292
4293 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
4294 and authenticator structs; see crypto/krb5/.
4295
4296 Generalized Kerberos calls to support multiple Kerberos libraries.
4297 [Vern Staats <staatsvr@asc.hpc.mil>,
4298 Jeffrey Altman <jaltman@columbia.edu>
4299 via Richard Levitte]
4300
4301 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
4302 already does with RSA. testdsa.h now has 'priv_key/pub_key'
4303 values for each of the key sizes rather than having just
4304 parameters (and 'speed' generating keys each time).
4305 [Geoff Thorpe]
4306
4307 *) Speed up EVP routines.
4308 Before:
4309 encrypt
4310 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
4311 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
4312 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
4313 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
4314 decrypt
4315 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
4316 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
4317 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
4318 After:
4319 encrypt
4320 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
4321 decrypt
4322 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
4323 [Ben Laurie]
4324
4325 *) Added the OS2-EMX target.
4326 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
4327
4328 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
4329 to support NCONF routines in extension code. New function CONF_set_nconf()
4330 to allow functions which take an NCONF to also handle the old LHASH
4331 structure: this means that the old CONF compatible routines can be
4332 retained (in particular wrt extensions) without having to duplicate the
4333 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
4334 [Steve Henson]
4335
4336 *) Enhance the general user interface with mechanisms for inner control
4337 and with possibilities to have yes/no kind of prompts.
4338 [Richard Levitte]
4339
4340 *) Change all calls to low level digest routines in the library and
4341 applications to use EVP. Add missing calls to HMAC_cleanup() and
4342 don't assume HMAC_CTX can be copied using memcpy().
4343 [Verdon Walker <VWalker@novell.com>, Steve Henson]
4344
4345 *) Add the possibility to control engines through control names but with
4346 arbitrary arguments instead of just a string.
4347 Change the key loaders to take a UI_METHOD instead of a callback
4348 function pointer. NOTE: this breaks binary compatibility with earlier
4349 versions of OpenSSL [engine].
4350 Adapt the nCipher code for these new conditions and add a card insertion
4351 callback.
4352 [Richard Levitte]
4353
4354 *) Enhance the general user interface with mechanisms to better support
4355 dialog box interfaces, application-defined prompts, the possibility
4356 to use defaults (for example default passwords from somewhere else)
4357 and interrupts/cancellations.
4358 [Richard Levitte]
4359
4360 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
4361 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
4362 [Steve Henson]
4363
4364 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
4365 tidy up some unnecessarily weird code in 'sk_new()').
4366 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
4367
4368 *) Change the key loading routines for ENGINEs to use the same kind
4369 callback (pem_password_cb) as all other routines that need this
4370 kind of callback.
4371 [Richard Levitte]
4372
4373 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
4374 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
4375 than this minimum value is recommended.
4376 [Lutz Jaenicke]
4377
4378 *) New random seeder for OpenVMS, using the system process statistics
4379 that are easily reachable.
4380 [Richard Levitte]
4381
4382 *) Windows apparently can't transparently handle global
4383 variables defined in DLLs. Initialisations such as:
4384
4385 const ASN1_ITEM *it = &ASN1_INTEGER_it;
4386
4387 wont compile. This is used by the any applications that need to
4388 declare their own ASN1 modules. This was fixed by adding the option
4389 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
4390 needed for static libraries under Win32.
4391 [Steve Henson]
4392
4393 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
4394 setting of purpose and trust fields. New X509_STORE trust and
4395 purpose functions and tidy up setting in other SSL functions.
4396 [Steve Henson]
4397
4398 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
4399 structure. These are inherited by X509_STORE_CTX when it is
4400 initialised. This allows various defaults to be set in the
4401 X509_STORE structure (such as flags for CRL checking and custom
4402 purpose or trust settings) for functions which only use X509_STORE_CTX
4403 internally such as S/MIME.
4404
4405 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
4406 trust settings if they are not set in X509_STORE. This allows X509_STORE
4407 purposes and trust (in S/MIME for example) to override any set by default.
4408
4409 Add command line options for CRL checking to smime, s_client and s_server
4410 applications.
4411 [Steve Henson]
4412
4413 *) Initial CRL based revocation checking. If the CRL checking flag(s)
4414 are set then the CRL is looked up in the X509_STORE structure and
4415 its validity and signature checked, then if the certificate is found
4416 in the CRL the verify fails with a revoked error.
4417
4418 Various new CRL related callbacks added to X509_STORE_CTX structure.
4419
4420 Command line options added to 'verify' application to support this.
4421
4422 This needs some additional work, such as being able to handle multiple
4423 CRLs with different times, extension based lookup (rather than just
4424 by subject name) and ultimately more complete V2 CRL extension
4425 handling.
4426 [Steve Henson]
4427
4428 *) Add a general user interface API (crypto/ui/). This is designed
4429 to replace things like des_read_password and friends (backward
4430 compatibility functions using this new API are provided).
4431 The purpose is to remove prompting functions from the DES code
4432 section as well as provide for prompting through dialog boxes in
4433 a window system and the like.
4434 [Richard Levitte]
4435
4436 *) Add "ex_data" support to ENGINE so implementations can add state at a
4437 per-structure level rather than having to store it globally.
4438 [Geoff]
4439
4440 *) Make it possible for ENGINE structures to be copied when retrieved by
4441 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
4442 This causes the "original" ENGINE structure to act like a template,
4443 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
4444 operational state can be localised to each ENGINE structure, despite the
4445 fact they all share the same "methods". New ENGINE structures returned in
4446 this case have no functional references and the return value is the single
4447 structural reference. This matches the single structural reference returned
4448 by ENGINE_by_id() normally, when it is incremented on the pre-existing
4449 ENGINE structure.
4450 [Geoff]
4451
4452 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
4453 needs to match any other type at all we need to manually clear the
4454 tag cache.
4455 [Steve Henson]
4456
4457 *) Changes to the "openssl engine" utility to include;
4458 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
4459 about an ENGINE's available control commands.
4460 - executing control commands from command line arguments using the
4461 '-pre' and '-post' switches. '-post' is only used if '-t' is
4462 specified and the ENGINE is successfully initialised. The syntax for
4463 the individual commands are colon-separated, for example;
4464 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
4465 [Geoff]
4466
4467 *) New dynamic control command support for ENGINEs. ENGINEs can now
4468 declare their own commands (numbers), names (strings), descriptions,
4469 and input types for run-time discovery by calling applications. A
4470 subset of these commands are implicitly classed as "executable"
4471 depending on their input type, and only these can be invoked through
4472 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
4473 can be based on user input, config files, etc). The distinction is
4474 that "executable" commands cannot return anything other than a boolean
4475 result and can only support numeric or string input, whereas some
4476 discoverable commands may only be for direct use through
4477 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
4478 pointers, or other custom uses. The "executable" commands are to
4479 support parameterisations of ENGINE behaviour that can be
4480 unambiguously defined by ENGINEs and used consistently across any
4481 OpenSSL-based application. Commands have been added to all the
4482 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
4483 control over shared-library paths without source code alterations.
4484 [Geoff]
4485
4486 *) Changed all ENGINE implementations to dynamically allocate their
4487 ENGINEs rather than declaring them statically. Apart from this being
4488 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
4489 this also allows the implementations to compile without using the
4490 internal engine_int.h header.
4491 [Geoff]
4492
4493 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
4494 'const' value. Any code that should be able to modify a RAND_METHOD
4495 should already have non-const pointers to it (ie. they should only
4496 modify their own ones).
4497 [Geoff]
4498
4499 *) Made a variety of little tweaks to the ENGINE code.
4500 - "atalla" and "ubsec" string definitions were moved from header files
4501 to C code. "nuron" string definitions were placed in variables
4502 rather than hard-coded - allowing parameterisation of these values
4503 later on via ctrl() commands.
4504 - Removed unused "#if 0"'d code.
4505 - Fixed engine list iteration code so it uses ENGINE_free() to release
4506 structural references.
4507 - Constified the RAND_METHOD element of ENGINE structures.
4508 - Constified various get/set functions as appropriate and added
4509 missing functions (including a catch-all ENGINE_cpy that duplicates
4510 all ENGINE values onto a new ENGINE except reference counts/state).
4511 - Removed NULL parameter checks in get/set functions. Setting a method
4512 or function to NULL is a way of cancelling out a previously set
4513 value. Passing a NULL ENGINE parameter is just plain stupid anyway
4514 and doesn't justify the extra error symbols and code.
4515 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
4516 flags from engine_int.h to engine.h.
4517 - Changed prototypes for ENGINE handler functions (init(), finish(),
4518 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
4519 [Geoff]
4520
4521 *) Implement binary inversion algorithm for BN_mod_inverse in addition
4522 to the algorithm using long division. The binary algorithm can be
4523 used only if the modulus is odd. On 32-bit systems, it is faster
4524 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
4525 roughly 5-15% for 256-bit moduli), so we use it only for moduli
4526 up to 450 bits. In 64-bit environments, the binary algorithm
4527 appears to be advantageous for much longer moduli; here we use it
4528 for moduli up to 2048 bits.
4529 [Bodo Moeller]
4530
4531 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
4532 could not support the combine flag in choice fields.
4533 [Steve Henson]
4534
4535 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
4536 extensions from a certificate request to the certificate.
4537 [Steve Henson]
4538
4539 *) Allow multiple 'certopt' and 'nameopt' options to be separated
4540 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
4541 file: this allows the display of the certificate about to be
4542 signed to be customised, to allow certain fields to be included
4543 or excluded and extension details. The old system didn't display
4544 multicharacter strings properly, omitted fields not in the policy
4545 and couldn't display additional details such as extensions.
4546 [Steve Henson]
4547
4548 *) Function EC_POINTs_mul for multiple scalar multiplication
4549 of an arbitrary number of elliptic curve points
4550 \sum scalars[i]*points[i],
4551 optionally including the generator defined for the EC_GROUP:
4552 scalar*generator + \sum scalars[i]*points[i].
4553
4554 EC_POINT_mul is a simple wrapper function for the typical case
4555 that the point list has just one item (besides the optional
4556 generator).
4557 [Bodo Moeller]
4558
4559 *) First EC_METHODs for curves over GF(p):
4560
4561 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
4562 operations and provides various method functions that can also
4563 operate with faster implementations of modular arithmetic.
4564
4565 EC_GFp_mont_method() reuses most functions that are part of
4566 EC_GFp_simple_method, but uses Montgomery arithmetic.
4567
4568 [Bodo Moeller; point addition and point doubling
4569 implementation directly derived from source code provided by
4570 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
4571
4572 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
4573 crypto/ec/ec_lib.c):
4574
4575 Curves are EC_GROUP objects (with an optional group generator)
4576 based on EC_METHODs that are built into the library.
4577
4578 Points are EC_POINT objects based on EC_GROUP objects.
4579
4580 Most of the framework would be able to handle curves over arbitrary
4581 finite fields, but as there are no obvious types for fields other
4582 than GF(p), some functions are limited to that for now.
4583 [Bodo Moeller]
4584
4585 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
4586 that the file contains a complete HTTP response.
4587 [Richard Levitte]
4588
4589 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
4590 change the def and num file printf format specifier from "%-40sXXX"
4591 to "%-39s XXX". The latter will always guarantee a space after the
4592 field while the former will cause them to run together if the field
4593 is 40 of more characters long.
4594 [Steve Henson]
4595
4596 *) Constify the cipher and digest 'method' functions and structures
4597 and modify related functions to take constant EVP_MD and EVP_CIPHER
4598 pointers.
4599 [Steve Henson]
4600
4601 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
4602 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
4603 [Bodo Moeller]
4604
4605 *) Modify EVP_Digest*() routines so they now return values. Although the
4606 internal software routines can never fail additional hardware versions
4607 might.
4608 [Steve Henson]
4609
4610 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
4611
4612 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
4613 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
4614
4615 ASN1 error codes
4616 ERR_R_NESTED_ASN1_ERROR
4617 ...
4618 ERR_R_MISSING_ASN1_EOS
4619 were 4 .. 9, conflicting with
4620 ERR_LIB_RSA (= ERR_R_RSA_LIB)
4621 ...
4622 ERR_LIB_PEM (= ERR_R_PEM_LIB).
4623 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
4624
4625 Add new error code 'ERR_R_INTERNAL_ERROR'.
4626 [Bodo Moeller]
4627
4628 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
4629 suffices.
4630 [Bodo Moeller]
4631
4632 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
4633 sets the subject name for a new request or supersedes the
4634 subject name in a given request. Formats that can be parsed are
4635 'CN=Some Name, OU=myOU, C=IT'
4636 and
4637 'CN=Some Name/OU=myOU/C=IT'.
4638
4639 Add options '-batch' and '-verbose' to 'openssl req'.
4640 [Massimiliano Pala <madwolf@hackmasters.net>]
4641
4642 *) Introduce the possibility to access global variables through
4643 functions on platform were that's the best way to handle exporting
4644 global variables in shared libraries. To enable this functionality,
4645 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
4646 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
4647 is normally done by Configure or something similar).
4648
4649 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
4650 in the source file (foo.c) like this:
4651
4652 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
4653 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
4654
4655 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
4656 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
4657
4658 OPENSSL_DECLARE_GLOBAL(int,foo);
4659 #define foo OPENSSL_GLOBAL_REF(foo)
4660 OPENSSL_DECLARE_GLOBAL(double,bar);
4661 #define bar OPENSSL_GLOBAL_REF(bar)
4662
4663 The #defines are very important, and therefore so is including the
4664 header file everywhere where the defined globals are used.
4665
4666 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
4667 of ASN.1 items, but that structure is a bit different.
4668
4669 The largest change is in util/mkdef.pl which has been enhanced with
4670 better and easier to understand logic to choose which symbols should
4671 go into the Windows .def files as well as a number of fixes and code
4672 cleanup (among others, algorithm keywords are now sorted
4673 lexicographically to avoid constant rewrites).
4674 [Richard Levitte]
4675
4676 *) In BN_div() keep a copy of the sign of 'num' before writing the
4677 result to 'rm' because if rm==num the value will be overwritten
4678 and produce the wrong result if 'num' is negative: this caused
4679 problems with BN_mod() and BN_nnmod().
4680 [Steve Henson]
4681
4682 *) Function OCSP_request_verify(). This checks the signature on an
4683 OCSP request and verifies the signer certificate. The signer
4684 certificate is just checked for a generic purpose and OCSP request
4685 trust settings.
4686 [Steve Henson]
4687
4688 *) Add OCSP_check_validity() function to check the validity of OCSP
4689 responses. OCSP responses are prepared in real time and may only
4690 be a few seconds old. Simply checking that the current time lies
4691 between thisUpdate and nextUpdate max reject otherwise valid responses
4692 caused by either OCSP responder or client clock inaccuracy. Instead
4693 we allow thisUpdate and nextUpdate to fall within a certain period of
4694 the current time. The age of the response can also optionally be
4695 checked. Two new options -validity_period and -status_age added to
4696 ocsp utility.
4697 [Steve Henson]
4698
4699 *) If signature or public key algorithm is unrecognized print out its
4700 OID rather that just UNKNOWN.
4701 [Steve Henson]
4702
4703 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
4704 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
4705 ID to be generated from the issuer certificate alone which can then be
4706 passed to OCSP_id_issuer_cmp().
4707 [Steve Henson]
4708
4709 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
4710 ASN1 modules to export functions returning ASN1_ITEM pointers
4711 instead of the ASN1_ITEM structures themselves. This adds several
4712 new macros which allow the underlying ASN1 function/structure to
4713 be accessed transparently. As a result code should not use ASN1_ITEM
4714 references directly (such as &X509_it) but instead use the relevant
4715 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
4716 use of the new ASN1 code on platforms where exporting structures
4717 is problematical (for example in shared libraries) but exporting
4718 functions returning pointers to structures is not.
4719 [Steve Henson]
4720
4721 *) Add support for overriding the generation of SSL/TLS session IDs.
4722 These callbacks can be registered either in an SSL_CTX or per SSL.
4723 The purpose of this is to allow applications to control, if they wish,
4724 the arbitrary values chosen for use as session IDs, particularly as it
4725 can be useful for session caching in multiple-server environments. A
4726 command-line switch for testing this (and any client code that wishes
4727 to use such a feature) has been added to "s_server".
4728 [Geoff Thorpe, Lutz Jaenicke]
4729
4730 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
4731 of the form '#if defined(...) || defined(...) || ...' and
4732 '#if !defined(...) && !defined(...) && ...'. This also avoids
4733 the growing number of special cases it was previously handling.
4734 [Richard Levitte]
4735
4736 *) Make all configuration macros available for application by making
4737 sure they are available in opensslconf.h, by giving them names starting
4738 with "OPENSSL_" to avoid conflicts with other packages and by making
4739 sure e_os2.h will cover all platform-specific cases together with
4740 opensslconf.h.
4741 Additionally, it is now possible to define configuration/platform-
4742 specific names (called "system identities"). In the C code, these
4743 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
4744 macro with the name beginning with "OPENSSL_SYS_", which is determined
4745 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
4746 what is available.
4747 [Richard Levitte]
4748
4749 *) New option -set_serial to 'req' and 'x509' this allows the serial
4750 number to use to be specified on the command line. Previously self
4751 signed certificates were hard coded with serial number 0 and the
4752 CA options of 'x509' had to use a serial number in a file which was
4753 auto incremented.
4754 [Steve Henson]
4755
4756 *) New options to 'ca' utility to support V2 CRL entry extensions.
4757 Currently CRL reason, invalidity date and hold instruction are
4758 supported. Add new CRL extensions to V3 code and some new objects.
4759 [Steve Henson]
4760
4761 *) New function EVP_CIPHER_CTX_set_padding() this is used to
4762 disable standard block padding (aka PKCS#5 padding) in the EVP
4763 API, which was previously mandatory. This means that the data is
4764 not padded in any way and so the total length much be a multiple
4765 of the block size, otherwise an error occurs.
4766 [Steve Henson]
4767
4768 *) Initial (incomplete) OCSP SSL support.
4769 [Steve Henson]
4770
4771 *) New function OCSP_parse_url(). This splits up a URL into its host,
4772 port and path components: primarily to parse OCSP URLs. New -url
4773 option to ocsp utility.
4774 [Steve Henson]
4775
4776 *) New nonce behavior. The return value of OCSP_check_nonce() now
4777 reflects the various checks performed. Applications can decide
4778 whether to tolerate certain situations such as an absent nonce
4779 in a response when one was present in a request: the ocsp application
4780 just prints out a warning. New function OCSP_add1_basic_nonce()
4781 this is to allow responders to include a nonce in a response even if
4782 the request is nonce-less.
4783 [Steve Henson]
4784
4785 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
4786 skipped when using openssl x509 multiple times on a single input file,
4787 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
4788 [Bodo Moeller]
4789
4790 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
4791 set string type: to handle setting ASN1_TIME structures. Fix ca
4792 utility to correctly initialize revocation date of CRLs.
4793 [Steve Henson]
4794
4795 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
4796 the clients preferred ciphersuites and rather use its own preferences.
4797 Should help to work around M$ SGC (Server Gated Cryptography) bug in
4798 Internet Explorer by ensuring unchanged hash method during stepup.
4799 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
4800 [Lutz Jaenicke]
4801
4802 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
4803 to aes and add a new 'exist' option to print out symbols that don't
4804 appear to exist.
4805 [Steve Henson]
4806
4807 *) Additional options to ocsp utility to allow flags to be set and
4808 additional certificates supplied.
4809 [Steve Henson]
4810
4811 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
4812 OCSP client a number of certificate to only verify the response
4813 signature against.
4814 [Richard Levitte]
4815
4816 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
4817 handle the new API. Currently only ECB, CBC modes supported. Add new
4818 AES OIDs.
4819
4820 Add TLS AES ciphersuites as described in RFC3268, "Advanced
4821 Encryption Standard (AES) Ciphersuites for Transport Layer
4822 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
4823 not enabled by default and were not part of the "ALL" ciphersuite
4824 alias because they were not yet official; they could be
4825 explicitly requested by specifying the "AESdraft" ciphersuite
4826 group alias. In the final release of OpenSSL 0.9.7, the group
4827 alias is called "AES" and is part of "ALL".)
4828 [Ben Laurie, Steve Henson, Bodo Moeller]
4829
4830 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
4831 request to response.
4832 [Steve Henson]
4833
4834 *) Functions for OCSP responders. OCSP_request_onereq_count(),
4835 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
4836 extract information from a certificate request. OCSP_response_create()
4837 creates a response and optionally adds a basic response structure.
4838 OCSP_basic_add1_status() adds a complete single response to a basic
4839 response and returns the OCSP_SINGLERESP structure just added (to allow
4840 extensions to be included for example). OCSP_basic_add1_cert() adds a
4841 certificate to a basic response and OCSP_basic_sign() signs a basic
4842 response with various flags. New helper functions ASN1_TIME_check()
4843 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
4844 (converts ASN1_TIME to GeneralizedTime).
4845 [Steve Henson]
4846
4847 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
4848 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
4849 structure from a certificate. X509_pubkey_digest() digests the public_key
4850 contents: this is used in various key identifiers.
4851 [Steve Henson]
4852
4853 *) Make sk_sort() tolerate a NULL argument.
4854 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
4855
4856 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
4857 passed by the function are trusted implicitly. If any of them signed the
4858 response then it is assumed to be valid and is not verified.
4859 [Steve Henson]
4860
4861 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
4862 to data. This was previously part of the PKCS7 ASN1 code. This
4863 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
4864 [Steve Henson, reported by Kenneth R. Robinette
4865 <support@securenetterm.com>]
4866
4867 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
4868 routines: without these tracing memory leaks is very painful.
4869 Fix leaks in PKCS12 and PKCS7 routines.
4870 [Steve Henson]
4871
4872 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
4873 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
4874 effectively meant GeneralizedTime would never be used. Now it
4875 is initialised to -1 but X509_time_adj() now has to check the value
4876 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
4877 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
4878 [Steve Henson, reported by Kenneth R. Robinette
4879 <support@securenetterm.com>]
4880
4881 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
4882 result in a zero length in the ASN1_INTEGER structure which was
4883 not consistent with the structure when d2i_ASN1_INTEGER() was used
4884 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
4885 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
4886 where it did not print out a minus for negative ASN1_INTEGER.
4887 [Steve Henson]
4888
4889 *) Add summary printout to ocsp utility. The various functions which
4890 convert status values to strings have been renamed to:
4891 OCSP_response_status_str(), OCSP_cert_status_str() and
4892 OCSP_crl_reason_str() and are no longer static. New options
4893 to verify nonce values and to disable verification. OCSP response
4894 printout format cleaned up.
4895 [Steve Henson]
4896
4897 *) Add additional OCSP certificate checks. These are those specified
4898 in RFC2560. This consists of two separate checks: the CA of the
4899 certificate being checked must either be the OCSP signer certificate
4900 or the issuer of the OCSP signer certificate. In the latter case the
4901 OCSP signer certificate must contain the OCSP signing extended key
4902 usage. This check is performed by attempting to match the OCSP
4903 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
4904 in the OCSP_CERTID structures of the response.
4905 [Steve Henson]
4906
4907 *) Initial OCSP certificate verification added to OCSP_basic_verify()
4908 and related routines. This uses the standard OpenSSL certificate
4909 verify routines to perform initial checks (just CA validity) and
4910 to obtain the certificate chain. Then additional checks will be
4911 performed on the chain. Currently the root CA is checked to see
4912 if it is explicitly trusted for OCSP signing. This is used to set
4913 a root CA as a global signing root: that is any certificate that
4914 chains to that CA is an acceptable OCSP signing certificate.
4915 [Steve Henson]
4916
4917 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
4918 extensions from a separate configuration file.
4919 As when reading extensions from the main configuration file,
4920 the '-extensions ...' option may be used for specifying the
4921 section to use.
4922 [Massimiliano Pala <madwolf@comune.modena.it>]
4923
4924 *) New OCSP utility. Allows OCSP requests to be generated or
4925 read. The request can be sent to a responder and the output
4926 parsed, outputed or printed in text form. Not complete yet:
4927 still needs to check the OCSP response validity.
4928 [Steve Henson]
4929
4930 *) New subcommands for 'openssl ca':
4931 'openssl ca -status <serial>' prints the status of the cert with
4932 the given serial number (according to the index file).
4933 'openssl ca -updatedb' updates the expiry status of certificates
4934 in the index file.
4935 [Massimiliano Pala <madwolf@comune.modena.it>]
4936
4937 *) New '-newreq-nodes' command option to CA.pl. This is like
4938 '-newreq', but calls 'openssl req' with the '-nodes' option
4939 so that the resulting key is not encrypted.
4940 [Damien Miller <djm@mindrot.org>]
4941
4942 *) New configuration for the GNU Hurd.
4943 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
4944
4945 *) Initial code to implement OCSP basic response verify. This
4946 is currently incomplete. Currently just finds the signer's
4947 certificate and verifies the signature on the response.
4948 [Steve Henson]
4949
4950 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
4951 value of OPENSSLDIR. This is available via the new '-d' option
4952 to 'openssl version', and is also included in 'openssl version -a'.
4953 [Bodo Moeller]
4954
4955 *) Allowing defining memory allocation callbacks that will be given
4956 file name and line number information in additional arguments
4957 (a const char* and an int). The basic functionality remains, as
4958 well as the original possibility to just replace malloc(),
4959 realloc() and free() by functions that do not know about these
4960 additional arguments. To register and find out the current
4961 settings for extended allocation functions, the following
4962 functions are provided:
4963
4964 CRYPTO_set_mem_ex_functions
4965 CRYPTO_set_locked_mem_ex_functions
4966 CRYPTO_get_mem_ex_functions
4967 CRYPTO_get_locked_mem_ex_functions
4968
4969 These work the same way as CRYPTO_set_mem_functions and friends.
4970 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4971 extended allocation function is enabled.
4972 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4973 a conventional allocation function is enabled.
4974 [Richard Levitte, Bodo Moeller]
4975
4976 *) Finish off removing the remaining LHASH function pointer casts.
4977 There should no longer be any prototype-casting required when using
4978 the LHASH abstraction, and any casts that remain are "bugs". See
4979 the callback types and macros at the head of lhash.h for details
4980 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4981 [Geoff Thorpe]
4982
4983 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4984 If /dev/[u]random devices are not available or do not return enough
4985 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4986 be queried.
4987 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4988 /etc/entropy will be queried once each in this sequence, quering stops
4989 when enough entropy was collected without querying more sockets.
4990 [Lutz Jaenicke]
4991
4992 *) Change the Unix RAND_poll() variant to be able to poll several
4993 random devices, as specified by DEVRANDOM, until a sufficient amount
4994 of data has been collected. We spend at most 10 ms on each file
4995 (select timeout) and read in non-blocking mode. DEVRANDOM now
4996 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4997 (previously it was just the string "/dev/urandom"), so on typical
4998 platforms the 10 ms delay will never occur.
4999 Also separate out the Unix variant to its own file, rand_unix.c.
5000 For VMS, there's a currently-empty rand_vms.c.
5001 [Richard Levitte]
5002
5003 *) Move OCSP client related routines to ocsp_cl.c. These
5004 provide utility functions which an application needing
5005 to issue a request to an OCSP responder and analyse the
5006 response will typically need: as opposed to those which an
5007 OCSP responder itself would need which will be added later.
5008
5009 OCSP_request_sign() signs an OCSP request with an API similar
5010 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
5011 response. OCSP_response_get1_basic() extracts basic response
5012 from response. OCSP_resp_find_status(): finds and extracts status
5013 information from an OCSP_CERTID structure (which will be created
5014 when the request structure is built). These are built from lower
5015 level functions which work on OCSP_SINGLERESP structures but
5016 wont normally be used unless the application wishes to examine
5017 extensions in the OCSP response for example.
5018
5019 Replace nonce routines with a pair of functions.
5020 OCSP_request_add1_nonce() adds a nonce value and optionally
5021 generates a random value. OCSP_check_nonce() checks the
5022 validity of the nonce in an OCSP response.
5023 [Steve Henson]
5024
5025 *) Change function OCSP_request_add() to OCSP_request_add0_id().
5026 This doesn't copy the supplied OCSP_CERTID and avoids the
5027 need to free up the newly created id. Change return type
5028 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
5029 This can then be used to add extensions to the request.
5030 Deleted OCSP_request_new(), since most of its functionality
5031 is now in OCSP_REQUEST_new() (and the case insensitive name
5032 clash) apart from the ability to set the request name which
5033 will be added elsewhere.
5034 [Steve Henson]
5035
5036 *) Update OCSP API. Remove obsolete extensions argument from
5037 various functions. Extensions are now handled using the new
5038 OCSP extension code. New simple OCSP HTTP function which
5039 can be used to send requests and parse the response.
5040 [Steve Henson]
5041
5042 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
5043 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
5044 uses the special reorder version of SET OF to sort the attributes
5045 and reorder them to match the encoded order. This resolves a long
5046 standing problem: a verify on a PKCS7 structure just after signing
5047 it used to fail because the attribute order did not match the
5048 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
5049 it uses the received order. This is necessary to tolerate some broken
5050 software that does not order SET OF. This is handled by encoding
5051 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
5052 to produce the required SET OF.
5053 [Steve Henson]
5054
5055 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
5056 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
5057 files to get correct declarations of the ASN.1 item variables.
5058 [Richard Levitte]
5059
5060 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
5061 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
5062 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
5063 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
5064 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
5065 ASN1_ITEM and no wrapper functions.
5066 [Steve Henson]
5067
5068 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
5069 replace the old function pointer based I/O routines. Change most of
5070 the *_d2i_bio() and *_d2i_fp() functions to use these.
5071 [Steve Henson]
5072
5073 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
5074 lines, recognice more "algorithms" that can be deselected, and make
5075 it complain about algorithm deselection that isn't recognised.
5076 [Richard Levitte]
5077
5078 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
5079 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
5080 to use new functions. Add NO_ASN1_OLD which can be set to remove
5081 some old style ASN1 functions: this can be used to determine if old
5082 code will still work when these eventually go away.
5083 [Steve Henson]
5084
5085 *) New extension functions for OCSP structures, these follow the
5086 same conventions as certificates and CRLs.
5087 [Steve Henson]
5088
5089 *) New function X509V3_add1_i2d(). This automatically encodes and
5090 adds an extension. Its behaviour can be customised with various
5091 flags to append, replace or delete. Various wrappers added for
5092 certifcates and CRLs.
5093 [Steve Henson]
5094
5095 *) Fix to avoid calling the underlying ASN1 print routine when
5096 an extension cannot be parsed. Correct a typo in the
5097 OCSP_SERVICELOC extension. Tidy up print OCSP format.
5098 [Steve Henson]
5099
5100 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
5101 entries for variables.
5102 [Steve Henson]
5103
5104 *) Add functionality to apps/openssl.c for detecting locking
5105 problems: As the program is single-threaded, all we have
5106 to do is register a locking callback using an array for
5107 storing which locks are currently held by the program.
5108 [Bodo Moeller]
5109
5110 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
5111 SSL_get_ex_data_X509_STORE_idx(), which is used in
5112 ssl_verify_cert_chain() and thus can be called at any time
5113 during TLS/SSL handshakes so that thread-safety is essential.
5114 Unfortunately, the ex_data design is not at all suited
5115 for multi-threaded use, so it probably should be abolished.
5116 [Bodo Moeller]
5117
5118 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
5119 [Broadcom, tweaked and integrated by Geoff Thorpe]
5120
5121 *) Move common extension printing code to new function
5122 X509V3_print_extensions(). Reorganise OCSP print routines and
5123 implement some needed OCSP ASN1 functions. Add OCSP extensions.
5124 [Steve Henson]
5125
5126 *) New function X509_signature_print() to remove duplication in some
5127 print routines.
5128 [Steve Henson]
5129
5130 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
5131 set (this was treated exactly the same as SET OF previously). This
5132 is used to reorder the STACK representing the structure to match the
5133 encoding. This will be used to get round a problem where a PKCS7
5134 structure which was signed could not be verified because the STACK
5135 order did not reflect the encoded order.
5136 [Steve Henson]
5137
5138 *) Reimplement the OCSP ASN1 module using the new code.
5139 [Steve Henson]
5140
5141 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
5142 for its ASN1 operations. The old style function pointers still exist
5143 for now but they will eventually go away.
5144 [Steve Henson]
5145
5146 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
5147 completely replaces the old ASN1 functionality with a table driven
5148 encoder and decoder which interprets an ASN1_ITEM structure describing
5149 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
5150 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
5151 has also been converted to the new form.
5152 [Steve Henson]
5153
5154 *) Change BN_mod_exp_recp so that negative moduli are tolerated
5155 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
5156 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
5157 for negative moduli.
5158 [Bodo Moeller]
5159
5160 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
5161 of not touching the result's sign bit.
5162 [Bodo Moeller]
5163
5164 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
5165 set.
5166 [Bodo Moeller]
5167
5168 *) Changed the LHASH code to use prototypes for callbacks, and created
5169 macros to declare and implement thin (optionally static) functions
5170 that provide type-safety and avoid function pointer casting for the
5171 type-specific callbacks.
5172 [Geoff Thorpe]
5173
5174 *) Added Kerberos Cipher Suites to be used with TLS, as written in
5175 RFC 2712.
5176 [Veers Staats <staatsvr@asc.hpc.mil>,
5177 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
5178
5179 *) Reformat the FAQ so the different questions and answers can be divided
5180 in sections depending on the subject.
5181 [Richard Levitte]
5182
5183 *) Have the zlib compression code load ZLIB.DLL dynamically under
5184 Windows.
5185 [Richard Levitte]
5186
5187 *) New function BN_mod_sqrt for computing square roots modulo a prime
5188 (using the probabilistic Tonelli-Shanks algorithm unless
5189 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
5190 be handled deterministically).
5191 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5192
5193 *) Make BN_mod_inverse faster by explicitly handling small quotients
5194 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
5195 512 bits], about 30% for larger ones [1024 or 2048 bits].)
5196 [Bodo Moeller]
5197
5198 *) New function BN_kronecker.
5199 [Bodo Moeller]
5200
5201 *) Fix BN_gcd so that it works on negative inputs; the result is
5202 positive unless both parameters are zero.
5203 Previously something reasonably close to an infinite loop was
5204 possible because numbers could be growing instead of shrinking
5205 in the implementation of Euclid's algorithm.
5206 [Bodo Moeller]
5207
5208 *) Fix BN_is_word() and BN_is_one() macros to take into account the
5209 sign of the number in question.
5210
5211 Fix BN_is_word(a,w) to work correctly for w == 0.
5212
5213 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
5214 because its test if the absolute value of 'a' equals 'w'.
5215 Note that BN_abs_is_word does *not* handle w == 0 reliably;
5216 it exists mostly for use in the implementations of BN_is_zero(),
5217 BN_is_one(), and BN_is_word().
5218 [Bodo Moeller]
5219
5220 *) New function BN_swap.
5221 [Bodo Moeller]
5222
5223 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
5224 the exponentiation functions are more likely to produce reasonable
5225 results on negative inputs.
5226 [Bodo Moeller]
5227
5228 *) Change BN_mod_mul so that the result is always non-negative.
5229 Previously, it could be negative if one of the factors was negative;
5230 I don't think anyone really wanted that behaviour.
5231 [Bodo Moeller]
5232
5233 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
5234 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
5235 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
5236 and add new functions:
5237
5238 BN_nnmod
5239 BN_mod_sqr
5240 BN_mod_add
5241 BN_mod_add_quick
5242 BN_mod_sub
5243 BN_mod_sub_quick
5244 BN_mod_lshift1
5245 BN_mod_lshift1_quick
5246 BN_mod_lshift
5247 BN_mod_lshift_quick
5248
5249 These functions always generate non-negative results.
5250
5251 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
5252 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
5253
5254 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
5255 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
5256 be reduced modulo m.
5257 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
5258
5259 #if 0
5260 The following entry accidentily appeared in the CHANGES file
5261 distributed with OpenSSL 0.9.7. The modifications described in
5262 it do *not* apply to OpenSSL 0.9.7.
5263
5264 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
5265 was actually never needed) and in BN_mul(). The removal in BN_mul()
5266 required a small change in bn_mul_part_recursive() and the addition
5267 of the functions bn_cmp_part_words(), bn_sub_part_words() and
5268 bn_add_part_words(), which do the same thing as bn_cmp_words(),
5269 bn_sub_words() and bn_add_words() except they take arrays with
5270 differing sizes.
5271 [Richard Levitte]
5272 #endif
5273
5274 *) In 'openssl passwd', verify passwords read from the terminal
5275 unless the '-salt' option is used (which usually means that
5276 verification would just waste user's time since the resulting
5277 hash is going to be compared with some given password hash)
5278 or the new '-noverify' option is used.
5279
5280 This is an incompatible change, but it does not affect
5281 non-interactive use of 'openssl passwd' (passwords on the command
5282 line, '-stdin' option, '-in ...' option) and thus should not
5283 cause any problems.
5284 [Bodo Moeller]
5285
5286 *) Remove all references to RSAref, since there's no more need for it.
5287 [Richard Levitte]
5288
5289 *) Make DSO load along a path given through an environment variable
5290 (SHLIB_PATH) with shl_load().
5291 [Richard Levitte]
5292
5293 *) Constify the ENGINE code as a result of BIGNUM constification.
5294 Also constify the RSA code and most things related to it. In a
5295 few places, most notable in the depth of the ASN.1 code, ugly
5296 casts back to non-const were required (to be solved at a later
5297 time)
5298 [Richard Levitte]
5299
5300 *) Make it so the openssl application has all engines loaded by default.
5301 [Richard Levitte]
5302
5303 *) Constify the BIGNUM routines a little more.
5304 [Richard Levitte]
5305
5306 *) Add the following functions:
5307
5308 ENGINE_load_cswift()
5309 ENGINE_load_chil()
5310 ENGINE_load_atalla()
5311 ENGINE_load_nuron()
5312 ENGINE_load_builtin_engines()
5313
5314 That way, an application can itself choose if external engines that
5315 are built-in in OpenSSL shall ever be used or not. The benefit is
5316 that applications won't have to be linked with libdl or other dso
5317 libraries unless it's really needed.
5318
5319 Changed 'openssl engine' to load all engines on demand.
5320 Changed the engine header files to avoid the duplication of some
5321 declarations (they differed!).
5322 [Richard Levitte]
5323
5324 *) 'openssl engine' can now list capabilities.
5325 [Richard Levitte]
5326
5327 *) Better error reporting in 'openssl engine'.
5328 [Richard Levitte]
5329
5330 *) Never call load_dh_param(NULL) in s_server.
5331 [Bodo Moeller]
5332
5333 *) Add engine application. It can currently list engines by name and
5334 identity, and test if they are actually available.
5335 [Richard Levitte]
5336
5337 *) Improve RPM specification file by forcing symbolic linking and making
5338 sure the installed documentation is also owned by root.root.
5339 [Damien Miller <djm@mindrot.org>]
5340
5341 *) Give the OpenSSL applications more possibilities to make use of
5342 keys (public as well as private) handled by engines.
5343 [Richard Levitte]
5344
5345 *) Add OCSP code that comes from CertCo.
5346 [Richard Levitte]
5347
5348 *) Add VMS support for the Rijndael code.
5349 [Richard Levitte]
5350
5351 *) Added untested support for Nuron crypto accelerator.
5352 [Ben Laurie]
5353
5354 *) Add support for external cryptographic devices. This code was
5355 previously distributed separately as the "engine" branch.
5356 [Geoff Thorpe, Richard Levitte]
5357
5358 *) Rework the filename-translation in the DSO code. It is now possible to
5359 have far greater control over how a "name" is turned into a filename
5360 depending on the operating environment and any oddities about the
5361 different shared library filenames on each system.
5362 [Geoff Thorpe]
5363
5364 *) Support threads on FreeBSD-elf in Configure.
5365 [Richard Levitte]
5366
5367 *) Fix for SHA1 assembly problem with MASM: it produces
5368 warnings about corrupt line number information when assembling
5369 with debugging information. This is caused by the overlapping
5370 of two sections.
5371 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
5372
5373 *) NCONF changes.
5374 NCONF_get_number() has no error checking at all. As a replacement,
5375 NCONF_get_number_e() is defined (_e for "error checking") and is
5376 promoted strongly. The old NCONF_get_number is kept around for
5377 binary backward compatibility.
5378 Make it possible for methods to load from something other than a BIO,
5379 by providing a function pointer that is given a name instead of a BIO.
5380 For example, this could be used to load configuration data from an
5381 LDAP server.
5382 [Richard Levitte]
5383
5384 *) Fix for non blocking accept BIOs. Added new I/O special reason
5385 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
5386 with non blocking I/O was not possible because no retry code was
5387 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
5388 this case.
5389 [Steve Henson]
5390
5391 *) Added the beginnings of Rijndael support.
5392 [Ben Laurie]
5393
5394 *) Fix for bug in DirectoryString mask setting. Add support for
5395 X509_NAME_print_ex() in 'req' and X509_print_ex() function
5396 to allow certificate printing to more controllable, additional
5397 'certopt' option to 'x509' to allow new printing options to be
5398 set.
5399 [Steve Henson]
5400
5401 *) Clean old EAY MD5 hack from e_os.h.
5402 [Richard Levitte]
5403
5404 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
5405
5406 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
5407 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
5408 [Joe Orton, Steve Henson]
5409
5410 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
5411
5412 *) Fix additional bug revealed by the NISCC test suite:
5413
5414 Stop bug triggering large recursion when presented with
5415 certain ASN.1 tags (CVE-2003-0851)
5416 [Steve Henson]
5417
5418 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
5419
5420 *) Fix various bugs revealed by running the NISCC test suite:
5421
5422 Stop out of bounds reads in the ASN1 code when presented with
5423 invalid tags (CVE-2003-0543 and CVE-2003-0544).
5424
5425 If verify callback ignores invalid public key errors don't try to check
5426 certificate signature with the NULL public key.
5427
5428 [Steve Henson]
5429
5430 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
5431 if the server requested one: as stated in TLS 1.0 and SSL 3.0
5432 specifications.
5433 [Steve Henson]
5434
5435 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
5436 extra data after the compression methods not only for TLS 1.0
5437 but also for SSL 3.0 (as required by the specification).
5438 [Bodo Moeller; problem pointed out by Matthias Loepfe]
5439
5440 *) Change X509_certificate_type() to mark the key as exported/exportable
5441 when it's 512 *bits* long, not 512 bytes.
5442 [Richard Levitte]
5443
5444 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
5445
5446 *) Countermeasure against the Klima-Pokorny-Rosa extension of
5447 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
5448 a protocol version number mismatch like a decryption error
5449 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
5450 [Bodo Moeller]
5451
5452 *) Turn on RSA blinding by default in the default implementation
5453 to avoid a timing attack. Applications that don't want it can call
5454 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
5455 They would be ill-advised to do so in most cases.
5456 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
5457
5458 *) Change RSA blinding code so that it works when the PRNG is not
5459 seeded (in this case, the secret RSA exponent is abused as
5460 an unpredictable seed -- if it is not unpredictable, there
5461 is no point in blinding anyway). Make RSA blinding thread-safe
5462 by remembering the creator's thread ID in rsa->blinding and
5463 having all other threads use local one-time blinding factors
5464 (this requires more computation than sharing rsa->blinding, but
5465 avoids excessive locking; and if an RSA object is not shared
5466 between threads, blinding will still be very fast).
5467 [Bodo Moeller]
5468
5469 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
5470
5471 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
5472 via timing by performing a MAC computation even if incorrrect
5473 block cipher padding has been found. This is a countermeasure
5474 against active attacks where the attacker has to distinguish
5475 between bad padding and a MAC verification error. (CVE-2003-0078)
5476
5477 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
5478 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
5479 Martin Vuagnoux (EPFL, Ilion)]
5480
5481 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
5482
5483 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
5484 memory from it's contents. This is done with a counter that will
5485 place alternating values in each byte. This can be used to solve
5486 two issues: 1) the removal of calls to memset() by highly optimizing
5487 compilers, and 2) cleansing with other values than 0, since those can
5488 be read through on certain media, for example a swap space on disk.
5489 [Geoff Thorpe]
5490
5491 *) Bugfix: client side session caching did not work with external caching,
5492 because the session->cipher setting was not restored when reloading
5493 from the external cache. This problem was masked, when
5494 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
5495 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
5496 [Lutz Jaenicke]
5497
5498 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
5499 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
5500 [Zeev Lieber <zeev-l@yahoo.com>]
5501
5502 *) Undo an undocumented change introduced in 0.9.6e which caused
5503 repeated calls to OpenSSL_add_all_ciphers() and
5504 OpenSSL_add_all_digests() to be ignored, even after calling
5505 EVP_cleanup().
5506 [Richard Levitte]
5507
5508 *) Change the default configuration reader to deal with last line not
5509 being properly terminated.
5510 [Richard Levitte]
5511
5512 *) Change X509_NAME_cmp() so it applies the special rules on handling
5513 DN values that are of type PrintableString, as well as RDNs of type
5514 emailAddress where the value has the type ia5String.
5515 [stefank@valicert.com via Richard Levitte]
5516
5517 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
5518 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
5519 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
5520 the bitwise-OR of the two for use by the majority of applications
5521 wanting this behaviour, and update the docs. The documented
5522 behaviour and actual behaviour were inconsistent and had been
5523 changing anyway, so this is more a bug-fix than a behavioural
5524 change.
5525 [Geoff Thorpe, diagnosed by Nadav Har'El]
5526
5527 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
5528 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
5529 [Bodo Moeller]
5530
5531 *) Fix initialization code race conditions in
5532 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
5533 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
5534 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
5535 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
5536 ssl2_get_cipher_by_char(),
5537 ssl3_get_cipher_by_char().
5538 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
5539
5540 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
5541 the cached sessions are flushed, as the remove_cb() might use ex_data
5542 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
5543 (see [openssl.org #212]).
5544 [Geoff Thorpe, Lutz Jaenicke]
5545
5546 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
5547 length, instead of the encoding length to d2i_ASN1_OBJECT.
5548 [Steve Henson]
5549
5550 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
5551
5552 *) [In 0.9.6g-engine release:]
5553 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
5554 [Lynn Gazis <lgazis@rainbow.com>]
5555
5556 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
5557
5558 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
5559 and get fix the header length calculation.
5560 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
5561 Alon Kantor <alonk@checkpoint.com> (and others),
5562 Steve Henson]
5563
5564 *) Use proper error handling instead of 'assertions' in buffer
5565 overflow checks added in 0.9.6e. This prevents DoS (the
5566 assertions could call abort()).
5567 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
5568
5569 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
5570
5571 *) Add various sanity checks to asn1_get_length() to reject
5572 the ASN1 length bytes if they exceed sizeof(long), will appear
5573 negative or the content length exceeds the length of the
5574 supplied buffer.
5575 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5576
5577 *) Fix cipher selection routines: ciphers without encryption had no flags
5578 for the cipher strength set and where therefore not handled correctly
5579 by the selection routines (PR #130).
5580 [Lutz Jaenicke]
5581
5582 *) Fix EVP_dsa_sha macro.
5583 [Nils Larsch]
5584
5585 *) New option
5586 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
5587 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
5588 that was added in OpenSSL 0.9.6d.
5589
5590 As the countermeasure turned out to be incompatible with some
5591 broken SSL implementations, the new option is part of SSL_OP_ALL.
5592 SSL_OP_ALL is usually employed when compatibility with weird SSL
5593 implementations is desired (e.g. '-bugs' option to 's_client' and
5594 's_server'), so the new option is automatically set in many
5595 applications.
5596 [Bodo Moeller]
5597
5598 *) Changes in security patch:
5599
5600 Changes marked "(CHATS)" were sponsored by the Defense Advanced
5601 Research Projects Agency (DARPA) and Air Force Research Laboratory,
5602 Air Force Materiel Command, USAF, under agreement number
5603 F30602-01-2-0537.
5604
5605 *) Add various sanity checks to asn1_get_length() to reject
5606 the ASN1 length bytes if they exceed sizeof(long), will appear
5607 negative or the content length exceeds the length of the
5608 supplied buffer. (CVE-2002-0659)
5609 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
5610
5611 *) Assertions for various potential buffer overflows, not known to
5612 happen in practice.
5613 [Ben Laurie (CHATS)]
5614
5615 *) Various temporary buffers to hold ASCII versions of integers were
5616 too small for 64 bit platforms. (CVE-2002-0655)
5617 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
5618
5619 *) Remote buffer overflow in SSL3 protocol - an attacker could
5620 supply an oversized session ID to a client. (CVE-2002-0656)
5621 [Ben Laurie (CHATS)]
5622
5623 *) Remote buffer overflow in SSL2 protocol - an attacker could
5624 supply an oversized client master key. (CVE-2002-0656)
5625 [Ben Laurie (CHATS)]
5626
5627 Changes between 0.9.6c and 0.9.6d [9 May 2002]
5628
5629 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
5630 encoded as NULL) with id-dsa-with-sha1.
5631 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
5632
5633 *) Check various X509_...() return values in apps/req.c.
5634 [Nils Larsch <nla@trustcenter.de>]
5635
5636 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
5637 an end-of-file condition would erronously be flagged, when the CRLF
5638 was just at the end of a processed block. The bug was discovered when
5639 processing data through a buffering memory BIO handing the data to a
5640 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
5641 <ptsekov@syntrex.com> and Nedelcho Stanev.
5642 [Lutz Jaenicke]
5643
5644 *) Implement a countermeasure against a vulnerability recently found
5645 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
5646 before application data chunks to avoid the use of known IVs
5647 with data potentially chosen by the attacker.
5648 [Bodo Moeller]
5649
5650 *) Fix length checks in ssl3_get_client_hello().
5651 [Bodo Moeller]
5652
5653 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
5654 to prevent ssl3_read_internal() from incorrectly assuming that
5655 ssl3_read_bytes() found application data while handshake
5656 processing was enabled when in fact s->s3->in_read_app_data was
5657 merely automatically cleared during the initial handshake.
5658 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
5659
5660 *) Fix object definitions for Private and Enterprise: they were not
5661 recognized in their shortname (=lowercase) representation. Extend
5662 obj_dat.pl to issue an error when using undefined keywords instead
5663 of silently ignoring the problem (Svenning Sorensen
5664 <sss@sss.dnsalias.net>).
5665 [Lutz Jaenicke]
5666
5667 *) Fix DH_generate_parameters() so that it works for 'non-standard'
5668 generators, i.e. generators other than 2 and 5. (Previously, the
5669 code did not properly initialise the 'add' and 'rem' values to
5670 BN_generate_prime().)
5671
5672 In the new general case, we do not insist that 'generator' is
5673 actually a primitive root: This requirement is rather pointless;
5674 a generator of the order-q subgroup is just as good, if not
5675 better.
5676 [Bodo Moeller]
5677
5678 *) Map new X509 verification errors to alerts. Discovered and submitted by
5679 Tom Wu <tom@arcot.com>.
5680 [Lutz Jaenicke]
5681
5682 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
5683 returning non-zero before the data has been completely received
5684 when using non-blocking I/O.
5685 [Bodo Moeller; problem pointed out by John Hughes]
5686
5687 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
5688 [Ben Laurie, Lutz Jaenicke]
5689
5690 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
5691 Yoram Zahavi <YoramZ@gilian.com>).
5692 [Lutz Jaenicke]
5693
5694 *) Add information about CygWin 1.3 and on, and preserve proper
5695 configuration for the versions before that.
5696 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
5697
5698 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
5699 check whether we deal with a copy of a session and do not delete from
5700 the cache in this case. Problem reported by "Izhar Shoshani Levi"
5701 <izhar@checkpoint.com>.
5702 [Lutz Jaenicke]
5703
5704 *) Do not store session data into the internal session cache, if it
5705 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
5706 flag is set). Proposed by Aslam <aslam@funk.com>.
5707 [Lutz Jaenicke]
5708
5709 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
5710 value is 0.
5711 [Richard Levitte]
5712
5713 *) [In 0.9.6d-engine release:]
5714 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5715 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5716
5717 *) Add the configuration target linux-s390x.
5718 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
5719
5720 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
5721 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
5722 variable as an indication that a ClientHello message has been
5723 received. As the flag value will be lost between multiple
5724 invocations of ssl3_accept when using non-blocking I/O, the
5725 function may not be aware that a handshake has actually taken
5726 place, thus preventing a new session from being added to the
5727 session cache.
5728
5729 To avoid this problem, we now set s->new_session to 2 instead of
5730 using a local variable.
5731 [Lutz Jaenicke, Bodo Moeller]
5732
5733 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
5734 if the SSL_R_LENGTH_MISMATCH error is detected.
5735 [Geoff Thorpe, Bodo Moeller]
5736
5737 *) New 'shared_ldflag' column in Configure platform table.
5738 [Richard Levitte]
5739
5740 *) Fix EVP_CIPHER_mode macro.
5741 ["Dan S. Camper" <dan@bti.net>]
5742
5743 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
5744 type, we must throw them away by setting rr->length to 0.
5745 [D P Chang <dpc@qualys.com>]
5746
5747 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
5748
5749 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
5750 <Dominikus.Scherkl@biodata.com>. (The previous implementation
5751 worked incorrectly for those cases where range = 10..._2 and
5752 3*range is two bits longer than range.)
5753 [Bodo Moeller]
5754
5755 *) Only add signing time to PKCS7 structures if it is not already
5756 present.
5757 [Steve Henson]
5758
5759 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
5760 OBJ_ld_ce should be OBJ_id_ce.
5761 Also some ip-pda OIDs in crypto/objects/objects.txt were
5762 incorrect (cf. RFC 3039).
5763 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
5764
5765 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
5766 returns early because it has nothing to do.
5767 [Andy Schneider <andy.schneider@bjss.co.uk>]
5768
5769 *) [In 0.9.6c-engine release:]
5770 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
5771 [Andy Schneider <andy.schneider@bjss.co.uk>]
5772
5773 *) [In 0.9.6c-engine release:]
5774 Add support for Cryptographic Appliance's keyserver technology.
5775 (Use engine 'keyclient')
5776 [Cryptographic Appliances and Geoff Thorpe]
5777
5778 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
5779 is called via tools/c89.sh because arguments have to be
5780 rearranged (all '-L' options must appear before the first object
5781 modules).
5782 [Richard Shapiro <rshapiro@abinitio.com>]
5783
5784 *) [In 0.9.6c-engine release:]
5785 Add support for Broadcom crypto accelerator cards, backported
5786 from 0.9.7.
5787 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
5788
5789 *) [In 0.9.6c-engine release:]
5790 Add support for SureWare crypto accelerator cards from
5791 Baltimore Technologies. (Use engine 'sureware')
5792 [Baltimore Technologies and Mark Cox]
5793
5794 *) [In 0.9.6c-engine release:]
5795 Add support for crypto accelerator cards from Accelerated
5796 Encryption Processing, www.aep.ie. (Use engine 'aep')
5797 [AEP Inc. and Mark Cox]
5798
5799 *) Add a configuration entry for gcc on UnixWare.
5800 [Gary Benson <gbenson@redhat.com>]
5801
5802 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
5803 messages are stored in a single piece (fixed-length part and
5804 variable-length part combined) and fix various bugs found on the way.
5805 [Bodo Moeller]
5806
5807 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
5808 instead. BIO_gethostbyname() does not know what timeouts are
5809 appropriate, so entries would stay in cache even when they have
5810 become invalid.
5811 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
5812
5813 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
5814 faced with a pathologically small ClientHello fragment that does
5815 not contain client_version: Instead of aborting with an error,
5816 simply choose the highest available protocol version (i.e.,
5817 TLS 1.0 unless it is disabled). In practice, ClientHello
5818 messages are never sent like this, but this change gives us
5819 strictly correct behaviour at least for TLS.
5820 [Bodo Moeller]
5821
5822 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
5823 never resets s->method to s->ctx->method when called from within
5824 one of the SSL handshake functions.
5825 [Bodo Moeller; problem pointed out by Niko Baric]
5826
5827 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
5828 (sent using the client's version number) if client_version is
5829 smaller than the protocol version in use. Also change
5830 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
5831 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
5832 the client will at least see that alert.
5833 [Bodo Moeller]
5834
5835 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
5836 correctly.
5837 [Bodo Moeller]
5838
5839 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
5840 client receives HelloRequest while in a handshake.
5841 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
5842
5843 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
5844 should end in 'break', not 'goto end' which circuments various
5845 cleanups done in state SSL_ST_OK. But session related stuff
5846 must be disabled for SSL_ST_OK in the case that we just sent a
5847 HelloRequest.
5848
5849 Also avoid some overhead by not calling ssl_init_wbio_buffer()
5850 before just sending a HelloRequest.
5851 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
5852
5853 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
5854 reveal whether illegal block cipher padding was found or a MAC
5855 verification error occured. (Neither SSLerr() codes nor alerts
5856 are directly visible to potential attackers, but the information
5857 may leak via logfiles.)
5858
5859 Similar changes are not required for the SSL 2.0 implementation
5860 because the number of padding bytes is sent in clear for SSL 2.0,
5861 and the extra bytes are just ignored. However ssl/s2_pkt.c
5862 failed to verify that the purported number of padding bytes is in
5863 the legal range.
5864 [Bodo Moeller]
5865
5866 *) Add OpenUNIX-8 support including shared libraries
5867 (Boyd Lynn Gerber <gerberb@zenez.com>).
5868 [Lutz Jaenicke]
5869
5870 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
5871 'wristwatch attack' using huge encoding parameters (cf.
5872 James H. Manger's CRYPTO 2001 paper). Note that the
5873 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
5874 encoding parameters and hence was not vulnerable.
5875 [Bodo Moeller]
5876
5877 *) BN_sqr() bug fix.
5878 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
5879
5880 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
5881 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
5882 followed by modular reduction.
5883 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
5884
5885 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
5886 equivalent based on BN_pseudo_rand() instead of BN_rand().
5887 [Bodo Moeller]
5888
5889 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
5890 This function was broken, as the check for a new client hello message
5891 to handle SGC did not allow these large messages.
5892 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
5893 [Lutz Jaenicke]
5894
5895 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
5896 [Lutz Jaenicke]
5897
5898 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
5899 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
5900 [Lutz Jaenicke]
5901
5902 *) Rework the configuration and shared library support for Tru64 Unix.
5903 The configuration part makes use of modern compiler features and
5904 still retains old compiler behavior for those that run older versions
5905 of the OS. The shared library support part includes a variant that
5906 uses the RPATH feature, and is available through the special
5907 configuration target "alpha-cc-rpath", which will never be selected
5908 automatically.
5909 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
5910
5911 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
5912 with the same message size as in ssl3_get_certificate_request().
5913 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
5914 messages might inadvertently be reject as too long.
5915 [Petr Lampa <lampa@fee.vutbr.cz>]
5916
5917 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
5918 [Andy Polyakov]
5919
5920 *) Modified SSL library such that the verify_callback that has been set
5921 specificly for an SSL object with SSL_set_verify() is actually being
5922 used. Before the change, a verify_callback set with this function was
5923 ignored and the verify_callback() set in the SSL_CTX at the time of
5924 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
5925 to allow the necessary settings.
5926 [Lutz Jaenicke]
5927
5928 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
5929 explicitly to NULL, as at least on Solaris 8 this seems not always to be
5930 done automatically (in contradiction to the requirements of the C
5931 standard). This made problems when used from OpenSSH.
5932 [Lutz Jaenicke]
5933
5934 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
5935 dh->length and always used
5936
5937 BN_rand_range(priv_key, dh->p).
5938
5939 BN_rand_range() is not necessary for Diffie-Hellman, and this
5940 specific range makes Diffie-Hellman unnecessarily inefficient if
5941 dh->length (recommended exponent length) is much smaller than the
5942 length of dh->p. We could use BN_rand_range() if the order of
5943 the subgroup was stored in the DH structure, but we only have
5944 dh->length.
5945
5946 So switch back to
5947
5948 BN_rand(priv_key, l, ...)
5949
5950 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5951 otherwise.
5952 [Bodo Moeller]
5953
5954 *) In
5955
5956 RSA_eay_public_encrypt
5957 RSA_eay_private_decrypt
5958 RSA_eay_private_encrypt (signing)
5959 RSA_eay_public_decrypt (signature verification)
5960
5961 (default implementations for RSA_public_encrypt,
5962 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5963 always reject numbers >= n.
5964 [Bodo Moeller]
5965
5966 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5967 to synchronize access to 'locking_thread'. This is necessary on
5968 systems where access to 'locking_thread' (an 'unsigned long'
5969 variable) is not atomic.
5970 [Bodo Moeller]
5971
5972 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5973 *before* setting the 'crypto_lock_rand' flag. The previous code had
5974 a race condition if 0 is a valid thread ID.
5975 [Travis Vitek <vitek@roguewave.com>]
5976
5977 *) Add support for shared libraries under Irix.
5978 [Albert Chin-A-Young <china@thewrittenword.com>]
5979
5980 *) Add configuration option to build on Linux on both big-endian and
5981 little-endian MIPS.
5982 [Ralf Baechle <ralf@uni-koblenz.de>]
5983
5984 *) Add the possibility to create shared libraries on HP-UX.
5985 [Richard Levitte]
5986
5987 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5988
5989 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5990 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5991 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5992 PRNG state recovery was possible based on the output of
5993 one PRNG request appropriately sized to gain knowledge on
5994 'md' followed by enough consecutive 1-byte PRNG requests
5995 to traverse all of 'state'.
5996
5997 1. When updating 'md_local' (the current thread's copy of 'md')
5998 during PRNG output generation, hash all of the previous
5999 'md_local' value, not just the half used for PRNG output.
6000
6001 2. Make the number of bytes from 'state' included into the hash
6002 independent from the number of PRNG bytes requested.
6003
6004 The first measure alone would be sufficient to avoid
6005 Markku-Juhani's attack. (Actually it had never occurred
6006 to me that the half of 'md_local' used for chaining was the
6007 half from which PRNG output bytes were taken -- I had always
6008 assumed that the secret half would be used.) The second
6009 measure makes sure that additional data from 'state' is never
6010 mixed into 'md_local' in small portions; this heuristically
6011 further strengthens the PRNG.
6012 [Bodo Moeller]
6013
6014 *) Fix crypto/bn/asm/mips3.s.
6015 [Andy Polyakov]
6016
6017 *) When only the key is given to "enc", the IV is undefined. Print out
6018 an error message in this case.
6019 [Lutz Jaenicke]
6020
6021 *) Handle special case when X509_NAME is empty in X509 printing routines.
6022 [Steve Henson]
6023
6024 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
6025 positive and less than q.
6026 [Bodo Moeller]
6027
6028 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
6029 used: it isn't thread safe and the add_lock_callback should handle
6030 that itself.
6031 [Paul Rose <Paul.Rose@bridge.com>]
6032
6033 *) Verify that incoming data obeys the block size in
6034 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
6035 [Bodo Moeller]
6036
6037 *) Fix OAEP check.
6038 [Ulf Möller, Bodo Möller]
6039
6040 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
6041 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
6042 when fixing the server behaviour for backwards-compatible 'client
6043 hello' messages. (Note that the attack is impractical against
6044 SSL 3.0 and TLS 1.0 anyway because length and version checking
6045 means that the probability of guessing a valid ciphertext is
6046 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
6047 paper.)
6048
6049 Before 0.9.5, the countermeasure (hide the error by generating a
6050 random 'decryption result') did not work properly because
6051 ERR_clear_error() was missing, meaning that SSL_get_error() would
6052 detect the supposedly ignored error.
6053
6054 Both problems are now fixed.
6055 [Bodo Moeller]
6056
6057 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
6058 (previously it was 1024).
6059 [Bodo Moeller]
6060
6061 *) Fix for compatibility mode trust settings: ignore trust settings
6062 unless some valid trust or reject settings are present.
6063 [Steve Henson]
6064
6065 *) Fix for blowfish EVP: its a variable length cipher.
6066 [Steve Henson]
6067
6068 *) Fix various bugs related to DSA S/MIME verification. Handle missing
6069 parameters in DSA public key structures and return an error in the
6070 DSA routines if parameters are absent.
6071 [Steve Henson]
6072
6073 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
6074 in the current directory if neither $RANDFILE nor $HOME was set.
6075 RAND_file_name() in 0.9.6a returned NULL in this case. This has
6076 caused some confusion to Windows users who haven't defined $HOME.
6077 Thus RAND_file_name() is changed again: e_os.h can define a
6078 DEFAULT_HOME, which will be used if $HOME is not set.
6079 For Windows, we use "C:"; on other platforms, we still require
6080 environment variables.
6081
6082 *) Move 'if (!initialized) RAND_poll()' into regions protected by
6083 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
6084 having multiple threads call RAND_poll() concurrently.
6085 [Bodo Moeller]
6086
6087 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
6088 combination of a flag and a thread ID variable.
6089 Otherwise while one thread is in ssleay_rand_bytes (which sets the
6090 flag), *other* threads can enter ssleay_add_bytes without obeying
6091 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
6092 that they do not hold after the first thread unsets add_do_not_lock).
6093 [Bodo Moeller]
6094
6095 *) Change bctest again: '-x' expressions are not available in all
6096 versions of 'test'.
6097 [Bodo Moeller]
6098
6099 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
6100
6101 *) Fix a couple of memory leaks in PKCS7_dataDecode()
6102 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
6103
6104 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
6105 the default extension for executables, if any. Also, make the perl
6106 scripts that use symlink() to test if it really exists and use "cp"
6107 if it doesn't. All this made OpenSSL compilable and installable in
6108 CygWin.
6109 [Richard Levitte]
6110
6111 *) Fix for asn1_GetSequence() for indefinite length constructed data.
6112 If SEQUENCE is length is indefinite just set c->slen to the total
6113 amount of data available.
6114 [Steve Henson, reported by shige@FreeBSD.org]
6115 [This change does not apply to 0.9.7.]
6116
6117 *) Change bctest to avoid here-documents inside command substitution
6118 (workaround for FreeBSD /bin/sh bug).
6119 For compatibility with Ultrix, avoid shell functions (introduced
6120 in the bctest version that searches along $PATH).
6121 [Bodo Moeller]
6122
6123 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
6124 with des_encrypt() defined on some operating systems, like Solaris
6125 and UnixWare.
6126 [Richard Levitte]
6127
6128 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
6129 On the Importance of Eliminating Errors in Cryptographic
6130 Computations, J. Cryptology 14 (2001) 2, 101-119,
6131 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
6132 [Ulf Moeller]
6133
6134 *) MIPS assembler BIGNUM division bug fix.
6135 [Andy Polyakov]
6136
6137 *) Disabled incorrect Alpha assembler code.
6138 [Richard Levitte]
6139
6140 *) Fix PKCS#7 decode routines so they correctly update the length
6141 after reading an EOC for the EXPLICIT tag.
6142 [Steve Henson]
6143 [This change does not apply to 0.9.7.]
6144
6145 *) Fix bug in PKCS#12 key generation routines. This was triggered
6146 if a 3DES key was generated with a 0 initial byte. Include
6147 PKCS12_BROKEN_KEYGEN compilation option to retain the old
6148 (but broken) behaviour.
6149 [Steve Henson]
6150
6151 *) Enhance bctest to search for a working bc along $PATH and print
6152 it when found.
6153 [Tim Rice <tim@multitalents.net> via Richard Levitte]
6154
6155 *) Fix memory leaks in err.c: free err_data string if necessary;
6156 don't write to the wrong index in ERR_set_error_data.
6157 [Bodo Moeller]
6158
6159 *) Implement ssl23_peek (analogous to ssl23_read), which previously
6160 did not exist.
6161 [Bodo Moeller]
6162
6163 *) Replace rdtsc with _emit statements for VC++ version 5.
6164 [Jeremy Cooper <jeremy@baymoo.org>]
6165
6166 *) Make it possible to reuse SSLv2 sessions.
6167 [Richard Levitte]
6168
6169 *) In copy_email() check for >= 0 as a return value for
6170 X509_NAME_get_index_by_NID() since 0 is a valid index.
6171 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
6172
6173 *) Avoid coredump with unsupported or invalid public keys by checking if
6174 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
6175 PKCS7_verify() fails with non detached data.
6176 [Steve Henson]
6177
6178 *) Don't use getenv in library functions when run as setuid/setgid.
6179 New function OPENSSL_issetugid().
6180 [Ulf Moeller]
6181
6182 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
6183 due to incorrect handling of multi-threading:
6184
6185 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
6186
6187 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
6188
6189 3. Count how many times MemCheck_off() has been called so that
6190 nested use can be treated correctly. This also avoids
6191 inband-signalling in the previous code (which relied on the
6192 assumption that thread ID 0 is impossible).
6193 [Bodo Moeller]
6194
6195 *) Add "-rand" option also to s_client and s_server.
6196 [Lutz Jaenicke]
6197
6198 *) Fix CPU detection on Irix 6.x.
6199 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
6200 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6201
6202 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
6203 was empty.
6204 [Steve Henson]
6205 [This change does not apply to 0.9.7.]
6206
6207 *) Use the cached encoding of an X509_NAME structure rather than
6208 copying it. This is apparently the reason for the libsafe "errors"
6209 but the code is actually correct.
6210 [Steve Henson]
6211
6212 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
6213 Bleichenbacher's DSA attack.
6214 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
6215 to be set and top=0 forces the highest bit to be set; top=-1 is new
6216 and leaves the highest bit random.
6217 [Ulf Moeller, Bodo Moeller]
6218
6219 *) In the NCONF_...-based implementations for CONF_... queries
6220 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
6221 a temporary CONF structure with the data component set to NULL
6222 (which gives segmentation faults in lh_retrieve).
6223 Instead, use NULL for the CONF pointer in CONF_get_string and
6224 CONF_get_number (which may use environment variables) and directly
6225 return NULL from CONF_get_section.
6226 [Bodo Moeller]
6227
6228 *) Fix potential buffer overrun for EBCDIC.
6229 [Ulf Moeller]
6230
6231 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
6232 keyUsage if basicConstraints absent for a CA.
6233 [Steve Henson]
6234
6235 *) Make SMIME_write_PKCS7() write mail header values with a format that
6236 is more generally accepted (no spaces before the semicolon), since
6237 some programs can't parse those values properly otherwise. Also make
6238 sure BIO's that break lines after each write do not create invalid
6239 headers.
6240 [Richard Levitte]
6241
6242 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
6243 macros previously used would not encode an empty SEQUENCE OF
6244 and break the signature.
6245 [Steve Henson]
6246 [This change does not apply to 0.9.7.]
6247
6248 *) Zero the premaster secret after deriving the master secret in
6249 DH ciphersuites.
6250 [Steve Henson]
6251
6252 *) Add some EVP_add_digest_alias registrations (as found in
6253 OpenSSL_add_all_digests()) to SSL_library_init()
6254 aka OpenSSL_add_ssl_algorithms(). This provides improved
6255 compatibility with peers using X.509 certificates
6256 with unconventional AlgorithmIdentifier OIDs.
6257 [Bodo Moeller]
6258
6259 *) Fix for Irix with NO_ASM.
6260 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
6261
6262 *) ./config script fixes.
6263 [Ulf Moeller, Richard Levitte]
6264
6265 *) Fix 'openssl passwd -1'.
6266 [Bodo Moeller]
6267
6268 *) Change PKCS12_key_gen_asc() so it can cope with non null
6269 terminated strings whose length is passed in the passlen
6270 parameter, for example from PEM callbacks. This was done
6271 by adding an extra length parameter to asc2uni().
6272 [Steve Henson, reported by <oddissey@samsung.co.kr>]
6273
6274 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
6275 call failed, free the DSA structure.
6276 [Bodo Moeller]
6277
6278 *) Fix to uni2asc() to cope with zero length Unicode strings.
6279 These are present in some PKCS#12 files.
6280 [Steve Henson]
6281
6282 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
6283 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
6284 when writing a 32767 byte record.
6285 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
6286
6287 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
6288 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
6289
6290 (RSA objects have a reference count access to which is protected
6291 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
6292 so they are meant to be shared between threads.)
6293 [Bodo Moeller, Geoff Thorpe; original patch submitted by
6294 "Reddie, Steven" <Steven.Reddie@ca.com>]
6295
6296 *) Fix a deadlock in CRYPTO_mem_leaks().
6297 [Bodo Moeller]
6298
6299 *) Use better test patterns in bntest.
6300 [Ulf Möller]
6301
6302 *) rand_win.c fix for Borland C.
6303 [Ulf Möller]
6304
6305 *) BN_rshift bugfix for n == 0.
6306 [Bodo Moeller]
6307
6308 *) Add a 'bctest' script that checks for some known 'bc' bugs
6309 so that 'make test' does not abort just because 'bc' is broken.
6310 [Bodo Moeller]
6311
6312 *) Store verify_result within SSL_SESSION also for client side to
6313 avoid potential security hole. (Re-used sessions on the client side
6314 always resulted in verify_result==X509_V_OK, not using the original
6315 result of the server certificate verification.)
6316 [Lutz Jaenicke]
6317
6318 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
6319 SSL3_RT_APPLICATION_DATA, return 0.
6320 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
6321 [Bodo Moeller]
6322
6323 *) Fix SSL_peek:
6324 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
6325 releases, have been re-implemented by renaming the previous
6326 implementations of ssl2_read and ssl3_read to ssl2_read_internal
6327 and ssl3_read_internal, respectively, and adding 'peek' parameters
6328 to them. The new ssl[23]_{read,peek} functions are calls to
6329 ssl[23]_read_internal with the 'peek' flag set appropriately.
6330 A 'peek' parameter has also been added to ssl3_read_bytes, which
6331 does the actual work for ssl3_read_internal.
6332 [Bodo Moeller]
6333
6334 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
6335 the method-specific "init()" handler. Also clean up ex_data after
6336 calling the method-specific "finish()" handler. Previously, this was
6337 happening the other way round.
6338 [Geoff Thorpe]
6339
6340 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
6341 The previous value, 12, was not always sufficient for BN_mod_exp().
6342 [Bodo Moeller]
6343
6344 *) Make sure that shared libraries get the internal name engine with
6345 the full version number and not just 0. This should mark the
6346 shared libraries as not backward compatible. Of course, this should
6347 be changed again when we can guarantee backward binary compatibility.
6348 [Richard Levitte]
6349
6350 *) Fix typo in get_cert_by_subject() in by_dir.c
6351 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
6352
6353 *) Rework the system to generate shared libraries:
6354
6355 - Make note of the expected extension for the shared libraries and
6356 if there is a need for symbolic links from for example libcrypto.so.0
6357 to libcrypto.so.0.9.7. There is extended info in Configure for
6358 that.
6359
6360 - Make as few rebuilds of the shared libraries as possible.
6361
6362 - Still avoid linking the OpenSSL programs with the shared libraries.
6363
6364 - When installing, install the shared libraries separately from the
6365 static ones.
6366 [Richard Levitte]
6367
6368 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
6369
6370 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
6371 and not in SSL_clear because the latter is also used by the
6372 accept/connect functions; previously, the settings made by
6373 SSL_set_read_ahead would be lost during the handshake.
6374 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
6375
6376 *) Correct util/mkdef.pl to be selective about disabled algorithms.
6377 Previously, it would create entries for disableed algorithms no
6378 matter what.
6379 [Richard Levitte]
6380
6381 *) Added several new manual pages for SSL_* function.
6382 [Lutz Jaenicke]
6383
6384 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
6385
6386 *) In ssl23_get_client_hello, generate an error message when faced
6387 with an initial SSL 3.0/TLS record that is too small to contain the
6388 first two bytes of the ClientHello message, i.e. client_version.
6389 (Note that this is a pathologic case that probably has never happened
6390 in real life.) The previous approach was to use the version number
6391 from the record header as a substitute; but our protocol choice
6392 should not depend on that one because it is not authenticated
6393 by the Finished messages.
6394 [Bodo Moeller]
6395
6396 *) More robust randomness gathering functions for Windows.
6397 [Jeffrey Altman <jaltman@columbia.edu>]
6398
6399 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
6400 not set then we don't setup the error code for issuer check errors
6401 to avoid possibly overwriting other errors which the callback does
6402 handle. If an application does set the flag then we assume it knows
6403 what it is doing and can handle the new informational codes
6404 appropriately.
6405 [Steve Henson]
6406
6407 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
6408 a general "ANY" type, as such it should be able to decode anything
6409 including tagged types. However it didn't check the class so it would
6410 wrongly interpret tagged types in the same way as their universal
6411 counterpart and unknown types were just rejected. Changed so that the
6412 tagged and unknown types are handled in the same way as a SEQUENCE:
6413 that is the encoding is stored intact. There is also a new type
6414 "V_ASN1_OTHER" which is used when the class is not universal, in this
6415 case we have no idea what the actual type is so we just lump them all
6416 together.
6417 [Steve Henson]
6418
6419 *) On VMS, stdout may very well lead to a file that is written to
6420 in a record-oriented fashion. That means that every write() will
6421 write a separate record, which will be read separately by the
6422 programs trying to read from it. This can be very confusing.
6423
6424 The solution is to put a BIO filter in the way that will buffer
6425 text until a linefeed is reached, and then write everything a
6426 line at a time, so every record written will be an actual line,
6427 not chunks of lines and not (usually doesn't happen, but I've
6428 seen it once) several lines in one record. BIO_f_linebuffer() is
6429 the answer.
6430
6431 Currently, it's a VMS-only method, because that's where it has
6432 been tested well enough.
6433 [Richard Levitte]
6434
6435 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
6436 it can return incorrect results.
6437 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
6438 but it was in 0.9.6-beta[12].)
6439 [Bodo Moeller]
6440
6441 *) Disable the check for content being present when verifying detached
6442 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
6443 include zero length content when signing messages.
6444 [Steve Henson]
6445
6446 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
6447 BIO_ctrl (for BIO pairs).
6448 [Bodo Möller]
6449
6450 *) Add DSO method for VMS.
6451 [Richard Levitte]
6452
6453 *) Bug fix: Montgomery multiplication could produce results with the
6454 wrong sign.
6455 [Ulf Möller]
6456
6457 *) Add RPM specification openssl.spec and modify it to build three
6458 packages. The default package contains applications, application
6459 documentation and run-time libraries. The devel package contains
6460 include files, static libraries and function documentation. The
6461 doc package contains the contents of the doc directory. The original
6462 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
6463 [Richard Levitte]
6464
6465 *) Add a large number of documentation files for many SSL routines.
6466 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6467
6468 *) Add a configuration entry for Sony News 4.
6469 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
6470
6471 *) Don't set the two most significant bits to one when generating a
6472 random number < q in the DSA library.
6473 [Ulf Möller]
6474
6475 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
6476 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
6477 the underlying transport is blocking) if a handshake took place.
6478 (The default behaviour is needed by applications such as s_client
6479 and s_server that use select() to determine when to use SSL_read;
6480 but for applications that know in advance when to expect data, it
6481 just makes things more complicated.)
6482 [Bodo Moeller]
6483
6484 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
6485 from EGD.
6486 [Ben Laurie]
6487
6488 *) Add a few more EBCDIC conditionals that make `req' and `x509'
6489 work better on such systems.
6490 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
6491
6492 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
6493 Update PKCS12_parse() so it copies the friendlyName and the
6494 keyid to the certificates aux info.
6495 [Steve Henson]
6496
6497 *) Fix bug in PKCS7_verify() which caused an infinite loop
6498 if there was more than one signature.
6499 [Sven Uszpelkat <su@celocom.de>]
6500
6501 *) Major change in util/mkdef.pl to include extra information
6502 about each symbol, as well as presentig variables as well
6503 as functions. This change means that there's n more need
6504 to rebuild the .num files when some algorithms are excluded.
6505 [Richard Levitte]
6506
6507 *) Allow the verify time to be set by an application,
6508 rather than always using the current time.
6509 [Steve Henson]
6510
6511 *) Phase 2 verify code reorganisation. The certificate
6512 verify code now looks up an issuer certificate by a
6513 number of criteria: subject name, authority key id
6514 and key usage. It also verifies self signed certificates
6515 by the same criteria. The main comparison function is
6516 X509_check_issued() which performs these checks.
6517
6518 Lot of changes were necessary in order to support this
6519 without completely rewriting the lookup code.
6520
6521 Authority and subject key identifier are now cached.
6522
6523 The LHASH 'certs' is X509_STORE has now been replaced
6524 by a STACK_OF(X509_OBJECT). This is mainly because an
6525 LHASH can't store or retrieve multiple objects with
6526 the same hash value.
6527
6528 As a result various functions (which were all internal
6529 use only) have changed to handle the new X509_STORE
6530 structure. This will break anything that messed round
6531 with X509_STORE internally.
6532
6533 The functions X509_STORE_add_cert() now checks for an
6534 exact match, rather than just subject name.
6535
6536 The X509_STORE API doesn't directly support the retrieval
6537 of multiple certificates matching a given criteria, however
6538 this can be worked round by performing a lookup first
6539 (which will fill the cache with candidate certificates)
6540 and then examining the cache for matches. This is probably
6541 the best we can do without throwing out X509_LOOKUP
6542 entirely (maybe later...).
6543
6544 The X509_VERIFY_CTX structure has been enhanced considerably.
6545
6546 All certificate lookup operations now go via a get_issuer()
6547 callback. Although this currently uses an X509_STORE it
6548 can be replaced by custom lookups. This is a simple way
6549 to bypass the X509_STORE hackery necessary to make this
6550 work and makes it possible to use more efficient techniques
6551 in future. A very simple version which uses a simple
6552 STACK for its trusted certificate store is also provided
6553 using X509_STORE_CTX_trusted_stack().
6554
6555 The verify_cb() and verify() callbacks now have equivalents
6556 in the X509_STORE_CTX structure.
6557
6558 X509_STORE_CTX also has a 'flags' field which can be used
6559 to customise the verify behaviour.
6560 [Steve Henson]
6561
6562 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
6563 excludes S/MIME capabilities.
6564 [Steve Henson]
6565
6566 *) When a certificate request is read in keep a copy of the
6567 original encoding of the signed data and use it when outputing
6568 again. Signatures then use the original encoding rather than
6569 a decoded, encoded version which may cause problems if the
6570 request is improperly encoded.
6571 [Steve Henson]
6572
6573 *) For consistency with other BIO_puts implementations, call
6574 buffer_write(b, ...) directly in buffer_puts instead of calling
6575 BIO_write(b, ...).
6576
6577 In BIO_puts, increment b->num_write as in BIO_write.
6578 [Peter.Sylvester@EdelWeb.fr]
6579
6580 *) Fix BN_mul_word for the case where the word is 0. (We have to use
6581 BN_zero, we may not return a BIGNUM with an array consisting of
6582 words set to zero.)
6583 [Bodo Moeller]
6584
6585 *) Avoid calling abort() from within the library when problems are
6586 detected, except if preprocessor symbols have been defined
6587 (such as REF_CHECK, BN_DEBUG etc.).
6588 [Bodo Moeller]
6589
6590 *) New openssl application 'rsautl'. This utility can be
6591 used for low level RSA operations. DER public key
6592 BIO/fp routines also added.
6593 [Steve Henson]
6594
6595 *) New Configure entry and patches for compiling on QNX 4.
6596 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
6597
6598 *) A demo state-machine implementation was sponsored by
6599 Nuron (http://www.nuron.com/) and is now available in
6600 demos/state_machine.
6601 [Ben Laurie]
6602
6603 *) New options added to the 'dgst' utility for signature
6604 generation and verification.
6605 [Steve Henson]
6606
6607 *) Unrecognized PKCS#7 content types are now handled via a
6608 catch all ASN1_TYPE structure. This allows unsupported
6609 types to be stored as a "blob" and an application can
6610 encode and decode it manually.
6611 [Steve Henson]
6612
6613 *) Fix various signed/unsigned issues to make a_strex.c
6614 compile under VC++.
6615 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
6616
6617 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
6618 length if passed a buffer. ASN1_INTEGER_to_BN failed
6619 if passed a NULL BN and its argument was negative.
6620 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
6621
6622 *) Modification to PKCS#7 encoding routines to output definite
6623 length encoding. Since currently the whole structures are in
6624 memory there's not real point in using indefinite length
6625 constructed encoding. However if OpenSSL is compiled with
6626 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
6627 [Steve Henson]
6628
6629 *) Added BIO_vprintf() and BIO_vsnprintf().
6630 [Richard Levitte]
6631
6632 *) Added more prefixes to parse for in the the strings written
6633 through a logging bio, to cover all the levels that are available
6634 through syslog. The prefixes are now:
6635
6636 PANIC, EMERG, EMR => LOG_EMERG
6637 ALERT, ALR => LOG_ALERT
6638 CRIT, CRI => LOG_CRIT
6639 ERROR, ERR => LOG_ERR
6640 WARNING, WARN, WAR => LOG_WARNING
6641 NOTICE, NOTE, NOT => LOG_NOTICE
6642 INFO, INF => LOG_INFO
6643 DEBUG, DBG => LOG_DEBUG
6644
6645 and as before, if none of those prefixes are present at the
6646 beginning of the string, LOG_ERR is chosen.
6647
6648 On Win32, the LOG_* levels are mapped according to this:
6649
6650 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
6651 LOG_WARNING => EVENTLOG_WARNING_TYPE
6652 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
6653
6654 [Richard Levitte]
6655
6656 *) Made it possible to reconfigure with just the configuration
6657 argument "reconf" or "reconfigure". The command line arguments
6658 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
6659 and are retrieved from there when reconfiguring.
6660 [Richard Levitte]
6661
6662 *) MD4 implemented.
6663 [Assar Westerlund <assar@sics.se>, Richard Levitte]
6664
6665 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
6666 [Richard Levitte]
6667
6668 *) The obj_dat.pl script was messing up the sorting of object
6669 names. The reason was that it compared the quoted version
6670 of strings as a result "OCSP" > "OCSP Signing" because
6671 " > SPACE. Changed script to store unquoted versions of
6672 names and add quotes on output. It was also omitting some
6673 names from the lookup table if they were given a default
6674 value (that is if SN is missing it is given the same
6675 value as LN and vice versa), these are now added on the
6676 grounds that if an object has a name we should be able to
6677 look it up. Finally added warning output when duplicate
6678 short or long names are found.
6679 [Steve Henson]
6680
6681 *) Changes needed for Tandem NSK.
6682 [Scott Uroff <scott@xypro.com>]
6683
6684 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
6685 RSA_padding_check_SSLv23(), special padding was never detected
6686 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
6687 version rollback attacks was not effective.
6688
6689 In s23_clnt.c, don't use special rollback-attack detection padding
6690 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
6691 client; similarly, in s23_srvr.c, don't do the rollback check if
6692 SSL 2.0 is the only protocol enabled in the server.
6693 [Bodo Moeller]
6694
6695 *) Make it possible to get hexdumps of unprintable data with 'openssl
6696 asn1parse'. By implication, the functions ASN1_parse_dump() and
6697 BIO_dump_indent() are added.
6698 [Richard Levitte]
6699
6700 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
6701 these print out strings and name structures based on various
6702 flags including RFC2253 support and proper handling of
6703 multibyte characters. Added options to the 'x509' utility
6704 to allow the various flags to be set.
6705 [Steve Henson]
6706
6707 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
6708 Also change the functions X509_cmp_current_time() and
6709 X509_gmtime_adj() work with an ASN1_TIME structure,
6710 this will enable certificates using GeneralizedTime in validity
6711 dates to be checked.
6712 [Steve Henson]
6713
6714 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
6715 negative public key encodings) on by default,
6716 NO_NEG_PUBKEY_BUG can be set to disable it.
6717 [Steve Henson]
6718
6719 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
6720 content octets. An i2c_ASN1_OBJECT is unnecessary because
6721 the encoding can be trivially obtained from the structure.
6722 [Steve Henson]
6723
6724 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
6725 not read locks (CRYPTO_r_[un]lock).
6726 [Bodo Moeller]
6727
6728 *) A first attempt at creating official support for shared
6729 libraries through configuration. I've kept it so the
6730 default is static libraries only, and the OpenSSL programs
6731 are always statically linked for now, but there are
6732 preparations for dynamic linking in place.
6733 This has been tested on Linux and Tru64.
6734 [Richard Levitte]
6735
6736 *) Randomness polling function for Win9x, as described in:
6737 Peter Gutmann, Software Generation of Practically Strong
6738 Random Numbers.
6739 [Ulf Möller]
6740
6741 *) Fix so PRNG is seeded in req if using an already existing
6742 DSA key.
6743 [Steve Henson]
6744
6745 *) New options to smime application. -inform and -outform
6746 allow alternative formats for the S/MIME message including
6747 PEM and DER. The -content option allows the content to be
6748 specified separately. This should allow things like Netscape
6749 form signing output easier to verify.
6750 [Steve Henson]
6751
6752 *) Fix the ASN1 encoding of tags using the 'long form'.
6753 [Steve Henson]
6754
6755 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
6756 STRING types. These convert content octets to and from the
6757 underlying type. The actual tag and length octets are
6758 already assumed to have been read in and checked. These
6759 are needed because all other string types have virtually
6760 identical handling apart from the tag. By having versions
6761 of the ASN1 functions that just operate on content octets
6762 IMPLICIT tagging can be handled properly. It also allows
6763 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
6764 and ASN1_INTEGER are identical apart from the tag.
6765 [Steve Henson]
6766
6767 *) Change the handling of OID objects as follows:
6768
6769 - New object identifiers are inserted in objects.txt, following
6770 the syntax given in objects.README.
6771 - objects.pl is used to process obj_mac.num and create a new
6772 obj_mac.h.
6773 - obj_dat.pl is used to create a new obj_dat.h, using the data in
6774 obj_mac.h.
6775
6776 This is currently kind of a hack, and the perl code in objects.pl
6777 isn't very elegant, but it works as I intended. The simplest way
6778 to check that it worked correctly is to look in obj_dat.h and
6779 check the array nid_objs and make sure the objects haven't moved
6780 around (this is important!). Additions are OK, as well as
6781 consistent name changes.
6782 [Richard Levitte]
6783
6784 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
6785 [Bodo Moeller]
6786
6787 *) Addition of the command line parameter '-rand file' to 'openssl req'.
6788 The given file adds to whatever has already been seeded into the
6789 random pool through the RANDFILE configuration file option or
6790 environment variable, or the default random state file.
6791 [Richard Levitte]
6792
6793 *) mkstack.pl now sorts each macro group into lexical order.
6794 Previously the output order depended on the order the files
6795 appeared in the directory, resulting in needless rewriting
6796 of safestack.h .
6797 [Steve Henson]
6798
6799 *) Patches to make OpenSSL compile under Win32 again. Mostly
6800 work arounds for the VC++ problem that it treats func() as
6801 func(void). Also stripped out the parts of mkdef.pl that
6802 added extra typesafe functions: these no longer exist.
6803 [Steve Henson]
6804
6805 *) Reorganisation of the stack code. The macros are now all
6806 collected in safestack.h . Each macro is defined in terms of
6807 a "stack macro" of the form SKM_<name>(type, a, b). The
6808 DEBUG_SAFESTACK is now handled in terms of function casts,
6809 this has the advantage of retaining type safety without the
6810 use of additional functions. If DEBUG_SAFESTACK is not defined
6811 then the non typesafe macros are used instead. Also modified the
6812 mkstack.pl script to handle the new form. Needs testing to see
6813 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
6814 the default if no major problems. Similar behaviour for ASN1_SET_OF
6815 and PKCS12_STACK_OF.
6816 [Steve Henson]
6817
6818 *) When some versions of IIS use the 'NET' form of private key the
6819 key derivation algorithm is different. Normally MD5(password) is
6820 used as a 128 bit RC4 key. In the modified case
6821 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
6822 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
6823 as the old Netscape_RSA functions except they have an additional
6824 'sgckey' parameter which uses the modified algorithm. Also added
6825 an -sgckey command line option to the rsa utility. Thanks to
6826 Adrian Peck <bertie@ncipher.com> for posting details of the modified
6827 algorithm to openssl-dev.
6828 [Steve Henson]
6829
6830 *) The evp_local.h macros were using 'c.##kname' which resulted in
6831 invalid expansion on some systems (SCO 5.0.5 for example).
6832 Corrected to 'c.kname'.
6833 [Phillip Porch <root@theporch.com>]
6834
6835 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
6836 a STACK of email addresses from a certificate or request, these look
6837 in the subject name and the subject alternative name extensions and
6838 omit any duplicate addresses.
6839 [Steve Henson]
6840
6841 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
6842 This makes DSA verification about 2 % faster.
6843 [Bodo Moeller]
6844
6845 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
6846 (meaning that now 2^5 values will be precomputed, which is only 4 KB
6847 plus overhead for 1024 bit moduli).
6848 This makes exponentiations about 0.5 % faster for 1024 bit
6849 exponents (as measured by "openssl speed rsa2048").
6850 [Bodo Moeller]
6851
6852 *) Rename memory handling macros to avoid conflicts with other
6853 software:
6854 Malloc => OPENSSL_malloc
6855 Malloc_locked => OPENSSL_malloc_locked
6856 Realloc => OPENSSL_realloc
6857 Free => OPENSSL_free
6858 [Richard Levitte]
6859
6860 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
6861 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
6862 [Bodo Moeller]
6863
6864 *) CygWin32 support.
6865 [John Jarvie <jjarvie@newsguy.com>]
6866
6867 *) The type-safe stack code has been rejigged. It is now only compiled
6868 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
6869 by default all type-specific stack functions are "#define"d back to
6870 standard stack functions. This results in more streamlined output
6871 but retains the type-safety checking possibilities of the original
6872 approach.
6873 [Geoff Thorpe]
6874
6875 *) The STACK code has been cleaned up, and certain type declarations
6876 that didn't make a lot of sense have been brought in line. This has
6877 also involved a cleanup of sorts in safestack.h to more correctly
6878 map type-safe stack functions onto their plain stack counterparts.
6879 This work has also resulted in a variety of "const"ifications of
6880 lots of the code, especially "_cmp" operations which should normally
6881 be prototyped with "const" parameters anyway.
6882 [Geoff Thorpe]
6883
6884 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
6885 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
6886 (The PRNG state consists of two parts, the large pool 'state' and 'md',
6887 where all of 'md' is used each time the PRNG is used, but 'state'
6888 is used only indexed by a cyclic counter. As entropy may not be
6889 well distributed from the beginning, 'md' is important as a
6890 chaining variable. However, the output function chains only half
6891 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
6892 all of 'md', and seeding with STATE_SIZE dummy bytes will result
6893 in all of 'state' being rewritten, with the new values depending
6894 on virtually all of 'md'. This overcomes the 80 bit limitation.)
6895 [Bodo Moeller]
6896
6897 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
6898 the handshake is continued after ssl_verify_cert_chain();
6899 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
6900 can lead to 'unexplainable' connection aborts later.
6901 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
6902
6903 *) Major EVP API cipher revision.
6904 Add hooks for extra EVP features. This allows various cipher
6905 parameters to be set in the EVP interface. Support added for variable
6906 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
6907 setting of RC2 and RC5 parameters.
6908
6909 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
6910 ciphers.
6911
6912 Remove lots of duplicated code from the EVP library. For example *every*
6913 cipher init() function handles the 'iv' in the same way according to the
6914 cipher mode. They also all do nothing if the 'key' parameter is NULL and
6915 for CFB and OFB modes they zero ctx->num.
6916
6917 New functionality allows removal of S/MIME code RC2 hack.
6918
6919 Most of the routines have the same form and so can be declared in terms
6920 of macros.
6921
6922 By shifting this to the top level EVP_CipherInit() it can be removed from
6923 all individual ciphers. If the cipher wants to handle IVs or keys
6924 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
6925 flags.
6926
6927 Change lots of functions like EVP_EncryptUpdate() to now return a
6928 value: although software versions of the algorithms cannot fail
6929 any installed hardware versions can.
6930 [Steve Henson]
6931
6932 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
6933 this option is set, tolerate broken clients that send the negotiated
6934 protocol version number instead of the requested protocol version
6935 number.
6936 [Bodo Moeller]
6937
6938 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
6939 i.e. non-zero for export ciphersuites, zero otherwise.
6940 Previous versions had this flag inverted, inconsistent with
6941 rsa_tmp_cb (..._TMP_RSA_CB).
6942 [Bodo Moeller; problem reported by Amit Chopra]
6943
6944 *) Add missing DSA library text string. Work around for some IIS
6945 key files with invalid SEQUENCE encoding.
6946 [Steve Henson]
6947
6948 *) Add a document (doc/standards.txt) that list all kinds of standards
6949 and so on that are implemented in OpenSSL.
6950 [Richard Levitte]
6951
6952 *) Enhance c_rehash script. Old version would mishandle certificates
6953 with the same subject name hash and wouldn't handle CRLs at all.
6954 Added -fingerprint option to crl utility, to support new c_rehash
6955 features.
6956 [Steve Henson]
6957
6958 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6959 [Ulf Möller]
6960
6961 *) Fix for SSL server purpose checking. Server checking was
6962 rejecting certificates which had extended key usage present
6963 but no ssl client purpose.
6964 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6965
6966 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6967 is a little unclear about how a blank password is handled.
6968 Since the password in encoded as a BMPString with terminating
6969 double NULL a zero length password would end up as just the
6970 double NULL. However no password at all is different and is
6971 handled differently in the PKCS#12 key generation code. NS
6972 treats a blank password as zero length. MSIE treats it as no
6973 password on export: but it will try both on import. We now do
6974 the same: PKCS12_parse() tries zero length and no password if
6975 the password is set to "" or NULL (NULL is now a valid password:
6976 it wasn't before) as does the pkcs12 application.
6977 [Steve Henson]
6978
6979 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6980 perror when PEM_read_bio_X509_REQ fails, the error message must
6981 be obtained from the error queue.
6982 [Bodo Moeller]
6983
6984 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6985 it in ERR_remove_state if appropriate, and change ERR_get_state
6986 accordingly to avoid race conditions (this is necessary because
6987 thread_hash is no longer constant once set).
6988 [Bodo Moeller]
6989
6990 *) Bugfix for linux-elf makefile.one.
6991 [Ulf Möller]
6992
6993 *) RSA_get_default_method() will now cause a default
6994 RSA_METHOD to be chosen if one doesn't exist already.
6995 Previously this was only set during a call to RSA_new()
6996 or RSA_new_method(NULL) meaning it was possible for
6997 RSA_get_default_method() to return NULL.
6998 [Geoff Thorpe]
6999
7000 *) Added native name translation to the existing DSO code
7001 that will convert (if the flag to do so is set) filenames
7002 that are sufficiently small and have no path information
7003 into a canonical native form. Eg. "blah" converted to
7004 "libblah.so" or "blah.dll" etc.
7005 [Geoff Thorpe]
7006
7007 *) New function ERR_error_string_n(e, buf, len) which is like
7008 ERR_error_string(e, buf), but writes at most 'len' bytes
7009 including the 0 terminator. For ERR_error_string_n, 'buf'
7010 may not be NULL.
7011 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
7012
7013 *) CONF library reworked to become more general. A new CONF
7014 configuration file reader "class" is implemented as well as a
7015 new functions (NCONF_*, for "New CONF") to handle it. The now
7016 old CONF_* functions are still there, but are reimplemented to
7017 work in terms of the new functions. Also, a set of functions
7018 to handle the internal storage of the configuration data is
7019 provided to make it easier to write new configuration file
7020 reader "classes" (I can definitely see something reading a
7021 configuration file in XML format, for example), called _CONF_*,
7022 or "the configuration storage API"...
7023
7024 The new configuration file reading functions are:
7025
7026 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
7027 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
7028
7029 NCONF_default, NCONF_WIN32
7030
7031 NCONF_dump_fp, NCONF_dump_bio
7032
7033 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
7034 NCONF_new creates a new CONF object. This works in the same way
7035 as other interfaces in OpenSSL, like the BIO interface.
7036 NCONF_dump_* dump the internal storage of the configuration file,
7037 which is useful for debugging. All other functions take the same
7038 arguments as the old CONF_* functions wth the exception of the
7039 first that must be a `CONF *' instead of a `LHASH *'.
7040
7041 To make it easer to use the new classes with the old CONF_* functions,
7042 the function CONF_set_default_method is provided.
7043 [Richard Levitte]
7044
7045 *) Add '-tls1' option to 'openssl ciphers', which was already
7046 mentioned in the documentation but had not been implemented.
7047 (This option is not yet really useful because even the additional
7048 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
7049 [Bodo Moeller]
7050
7051 *) Initial DSO code added into libcrypto for letting OpenSSL (and
7052 OpenSSL-based applications) load shared libraries and bind to
7053 them in a portable way.
7054 [Geoff Thorpe, with contributions from Richard Levitte]
7055
7056 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
7057
7058 *) Make sure _lrotl and _lrotr are only used with MSVC.
7059
7060 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
7061 (the default implementation of RAND_status).
7062
7063 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
7064 to '-clrext' (= clear extensions), as intended and documented.
7065 [Bodo Moeller; inconsistency pointed out by Michael Attili
7066 <attili@amaxo.com>]
7067
7068 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
7069 was larger than the MD block size.
7070 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
7071
7072 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
7073 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
7074 using the passed key: if the passed key was a private key the result
7075 of X509_print(), for example, would be to print out all the private key
7076 components.
7077 [Steve Henson]
7078
7079 *) des_quad_cksum() byte order bug fix.
7080 [Ulf Möller, using the problem description in krb4-0.9.7, where
7081 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
7082
7083 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
7084 discouraged.
7085 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
7086
7087 *) For easily testing in shell scripts whether some command
7088 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
7089 returns with exit code 0 iff no command of the given name is available.
7090 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
7091 the output goes to stdout and nothing is printed to stderr.
7092 Additional arguments are always ignored.
7093
7094 Since for each cipher there is a command of the same name,
7095 the 'no-cipher' compilation switches can be tested this way.
7096
7097 ('openssl no-XXX' is not able to detect pseudo-commands such
7098 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
7099 [Bodo Moeller]
7100
7101 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
7102 [Bodo Moeller]
7103
7104 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
7105 is set; it will be thrown away anyway because each handshake creates
7106 its own key.
7107 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
7108 to parameters -- in previous versions (since OpenSSL 0.9.3) the
7109 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
7110 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
7111 [Bodo Moeller]
7112
7113 *) New s_client option -ign_eof: EOF at stdin is ignored, and
7114 'Q' and 'R' lose their special meanings (quit/renegotiate).
7115 This is part of what -quiet does; unlike -quiet, -ign_eof
7116 does not suppress any output.
7117 [Richard Levitte]
7118
7119 *) Add compatibility options to the purpose and trust code. The
7120 purpose X509_PURPOSE_ANY is "any purpose" which automatically
7121 accepts a certificate or CA, this was the previous behaviour,
7122 with all the associated security issues.
7123
7124 X509_TRUST_COMPAT is the old trust behaviour: only and
7125 automatically trust self signed roots in certificate store. A
7126 new trust setting X509_TRUST_DEFAULT is used to specify that
7127 a purpose has no associated trust setting and it should instead
7128 use the value in the default purpose.
7129 [Steve Henson]
7130
7131 *) Fix the PKCS#8 DSA private key code so it decodes keys again
7132 and fix a memory leak.
7133 [Steve Henson]
7134
7135 *) In util/mkerr.pl (which implements 'make errors'), preserve
7136 reason strings from the previous version of the .c file, as
7137 the default to have only downcase letters (and digits) in
7138 automatically generated reasons codes is not always appropriate.
7139 [Bodo Moeller]
7140
7141 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
7142 using strerror. Previously, ERR_reason_error_string() returned
7143 library names as reason strings for SYSerr; but SYSerr is a special
7144 case where small numbers are errno values, not library numbers.
7145 [Bodo Moeller]
7146
7147 *) Add '-dsaparam' option to 'openssl dhparam' application. This
7148 converts DSA parameters into DH parameters. (When creating parameters,
7149 DSA_generate_parameters is used.)
7150 [Bodo Moeller]
7151
7152 *) Include 'length' (recommended exponent length) in C code generated
7153 by 'openssl dhparam -C'.
7154 [Bodo Moeller]
7155
7156 *) The second argument to set_label in perlasm was already being used
7157 so couldn't be used as a "file scope" flag. Moved to third argument
7158 which was free.
7159 [Steve Henson]
7160
7161 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
7162 instead of RAND_bytes for encryption IVs and salts.
7163 [Bodo Moeller]
7164
7165 *) Include RAND_status() into RAND_METHOD instead of implementing
7166 it only for md_rand.c Otherwise replacing the PRNG by calling
7167 RAND_set_rand_method would be impossible.
7168 [Bodo Moeller]
7169
7170 *) Don't let DSA_generate_key() enter an infinite loop if the random
7171 number generation fails.
7172 [Bodo Moeller]
7173
7174 *) New 'rand' application for creating pseudo-random output.
7175 [Bodo Moeller]
7176
7177 *) Added configuration support for Linux/IA64
7178 [Rolf Haberrecker <rolf@suse.de>]
7179
7180 *) Assembler module support for Mingw32.
7181 [Ulf Möller]
7182
7183 *) Shared library support for HPUX (in shlib/).
7184 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
7185
7186 *) Shared library support for Solaris gcc.
7187 [Lutz Behnke <behnke@trustcenter.de>]
7188
7189 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
7190
7191 *) PKCS7_encrypt() was adding text MIME headers twice because they
7192 were added manually and by SMIME_crlf_copy().
7193 [Steve Henson]
7194
7195 *) In bntest.c don't call BN_rand with zero bits argument.
7196 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
7197
7198 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
7199 case was implemented. This caused BN_div_recp() to fail occasionally.
7200 [Ulf Möller]
7201
7202 *) Add an optional second argument to the set_label() in the perl
7203 assembly language builder. If this argument exists and is set
7204 to 1 it signals that the assembler should use a symbol whose
7205 scope is the entire file, not just the current function. This
7206 is needed with MASM which uses the format label:: for this scope.
7207 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
7208
7209 *) Change the ASN1 types so they are typedefs by default. Before
7210 almost all types were #define'd to ASN1_STRING which was causing
7211 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
7212 for example.
7213 [Steve Henson]
7214
7215 *) Change names of new functions to the new get1/get0 naming
7216 convention: After 'get1', the caller owns a reference count
7217 and has to call ..._free; 'get0' returns a pointer to some
7218 data structure without incrementing reference counters.
7219 (Some of the existing 'get' functions increment a reference
7220 counter, some don't.)
7221 Similarly, 'set1' and 'add1' functions increase reference
7222 counters or duplicate objects.
7223 [Steve Henson]
7224
7225 *) Allow for the possibility of temp RSA key generation failure:
7226 the code used to assume it always worked and crashed on failure.
7227 [Steve Henson]
7228
7229 *) Fix potential buffer overrun problem in BIO_printf().
7230 [Ulf Möller, using public domain code by Patrick Powell; problem
7231 pointed out by David Sacerdote <das33@cornell.edu>]
7232
7233 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
7234 RAND_egd() and RAND_status(). In the command line application,
7235 the EGD socket can be specified like a seed file using RANDFILE
7236 or -rand.
7237 [Ulf Möller]
7238
7239 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
7240 Some CAs (e.g. Verisign) distribute certificates in this form.
7241 [Steve Henson]
7242
7243 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
7244 list to exclude them. This means that no special compilation option
7245 is needed to use anonymous DH: it just needs to be included in the
7246 cipher list.
7247 [Steve Henson]
7248
7249 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
7250 EVP_MD_type. The old functionality is available in a new macro called
7251 EVP_MD_md(). Change code that uses it and update docs.
7252 [Steve Henson]
7253
7254 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
7255 where the 'void *' argument is replaced by a function pointer argument.
7256 Previously 'void *' was abused to point to functions, which works on
7257 many platforms, but is not correct. As these functions are usually
7258 called by macros defined in OpenSSL header files, most source code
7259 should work without changes.
7260 [Richard Levitte]
7261
7262 *) <openssl/opensslconf.h> (which is created by Configure) now contains
7263 sections with information on -D... compiler switches used for
7264 compiling the library so that applications can see them. To enable
7265 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
7266 must be defined. E.g.,
7267 #define OPENSSL_ALGORITHM_DEFINES
7268 #include <openssl/opensslconf.h>
7269 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
7270 [Richard Levitte, Ulf and Bodo Möller]
7271
7272 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
7273 record layer.
7274 [Bodo Moeller]
7275
7276 *) Change the 'other' type in certificate aux info to a STACK_OF
7277 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
7278 the required ASN1 format: arbitrary types determined by an OID.
7279 [Steve Henson]
7280
7281 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
7282 argument to 'req'. This is not because the function is newer or
7283 better than others it just uses the work 'NEW' in the certificate
7284 request header lines. Some software needs this.
7285 [Steve Henson]
7286
7287 *) Reorganise password command line arguments: now passwords can be
7288 obtained from various sources. Delete the PEM_cb function and make
7289 it the default behaviour: i.e. if the callback is NULL and the
7290 usrdata argument is not NULL interpret it as a null terminated pass
7291 phrase. If usrdata and the callback are NULL then the pass phrase
7292 is prompted for as usual.
7293 [Steve Henson]
7294
7295 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
7296 the support is automatically enabled. The resulting binaries will
7297 autodetect the card and use it if present.
7298 [Ben Laurie and Compaq Inc.]
7299
7300 *) Work around for Netscape hang bug. This sends certificate request
7301 and server done in one record. Since this is perfectly legal in the
7302 SSL/TLS protocol it isn't a "bug" option and is on by default. See
7303 the bugs/SSLv3 entry for more info.
7304 [Steve Henson]
7305
7306 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
7307 [Andy Polyakov]
7308
7309 *) Add -rand argument to smime and pkcs12 applications and read/write
7310 of seed file.
7311 [Steve Henson]
7312
7313 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
7314 [Bodo Moeller]
7315
7316 *) Add command line password options to the remaining applications.
7317 [Steve Henson]
7318
7319 *) Bug fix for BN_div_recp() for numerators with an even number of
7320 bits.
7321 [Ulf Möller]
7322
7323 *) More tests in bntest.c, and changed test_bn output.
7324 [Ulf Möller]
7325
7326 *) ./config recognizes MacOS X now.
7327 [Andy Polyakov]
7328
7329 *) Bug fix for BN_div() when the first words of num and divsor are
7330 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
7331 [Ulf Möller]
7332
7333 *) Add support for various broken PKCS#8 formats, and command line
7334 options to produce them.
7335 [Steve Henson]
7336
7337 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
7338 get temporary BIGNUMs from a BN_CTX.
7339 [Ulf Möller]
7340
7341 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
7342 for p == 0.
7343 [Ulf Möller]
7344
7345 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
7346 include a #define from the old name to the new. The original intent
7347 was that statically linked binaries could for example just call
7348 SSLeay_add_all_ciphers() to just add ciphers to the table and not
7349 link with digests. This never worked becayse SSLeay_add_all_digests()
7350 and SSLeay_add_all_ciphers() were in the same source file so calling
7351 one would link with the other. They are now in separate source files.
7352 [Steve Henson]
7353
7354 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
7355 [Steve Henson]
7356
7357 *) Use a less unusual form of the Miller-Rabin primality test (it used
7358 a binary algorithm for exponentiation integrated into the Miller-Rabin
7359 loop, our standard modexp algorithms are faster).
7360 [Bodo Moeller]
7361
7362 *) Support for the EBCDIC character set completed.
7363 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
7364
7365 *) Source code cleanups: use const where appropriate, eliminate casts,
7366 use void * instead of char * in lhash.
7367 [Ulf Möller]
7368
7369 *) Bugfix: ssl3_send_server_key_exchange was not restartable
7370 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
7371 this the server could overwrite ephemeral keys that the client
7372 has already seen).
7373 [Bodo Moeller]
7374
7375 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
7376 using 50 iterations of the Rabin-Miller test.
7377
7378 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
7379 iterations of the Rabin-Miller test as required by the appendix
7380 to FIPS PUB 186[-1]) instead of DSA_is_prime.
7381 As BN_is_prime_fasttest includes trial division, DSA parameter
7382 generation becomes much faster.
7383
7384 This implies a change for the callback functions in DSA_is_prime
7385 and DSA_generate_parameters: The callback function is called once
7386 for each positive witness in the Rabin-Miller test, not just
7387 occasionally in the inner loop; and the parameters to the
7388 callback function now provide an iteration count for the outer
7389 loop rather than for the current invocation of the inner loop.
7390 DSA_generate_parameters additionally can call the callback
7391 function with an 'iteration count' of -1, meaning that a
7392 candidate has passed the trial division test (when q is generated
7393 from an application-provided seed, trial division is skipped).
7394 [Bodo Moeller]
7395
7396 *) New function BN_is_prime_fasttest that optionally does trial
7397 division before starting the Rabin-Miller test and has
7398 an additional BN_CTX * argument (whereas BN_is_prime always
7399 has to allocate at least one BN_CTX).
7400 'callback(1, -1, cb_arg)' is called when a number has passed the
7401 trial division stage.
7402 [Bodo Moeller]
7403
7404 *) Fix for bug in CRL encoding. The validity dates weren't being handled
7405 as ASN1_TIME.
7406 [Steve Henson]
7407
7408 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
7409 [Steve Henson]
7410
7411 *) New function BN_pseudo_rand().
7412 [Ulf Möller]
7413
7414 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
7415 bignum version of BN_from_montgomery() with the working code from
7416 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
7417 the comments.
7418 [Ulf Möller]
7419
7420 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
7421 made it impossible to use the same SSL_SESSION data structure in
7422 SSL2 clients in multiple threads.
7423 [Bodo Moeller]
7424
7425 *) The return value of RAND_load_file() no longer counts bytes obtained
7426 by stat(). RAND_load_file(..., -1) is new and uses the complete file
7427 to seed the PRNG (previously an explicit byte count was required).
7428 [Ulf Möller, Bodo Möller]
7429
7430 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
7431 used (char *) instead of (void *) and had casts all over the place.
7432 [Steve Henson]
7433
7434 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
7435 [Ulf Möller]
7436
7437 *) Retain source code compatibility for BN_prime_checks macro:
7438 BN_is_prime(..., BN_prime_checks, ...) now uses
7439 BN_prime_checks_for_size to determine the appropriate number of
7440 Rabin-Miller iterations.
7441 [Ulf Möller]
7442
7443 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
7444 DH_CHECK_P_NOT_SAFE_PRIME.
7445 (Check if this is true? OpenPGP calls them "strong".)
7446 [Ulf Möller]
7447
7448 *) Merge the functionality of "dh" and "gendh" programs into a new program
7449 "dhparam". The old programs are retained for now but will handle DH keys
7450 (instead of parameters) in future.
7451 [Steve Henson]
7452
7453 *) Make the ciphers, s_server and s_client programs check the return values
7454 when a new cipher list is set.
7455 [Steve Henson]
7456
7457 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
7458 ciphers. Before when the 56bit ciphers were enabled the sorting was
7459 wrong.
7460
7461 The syntax for the cipher sorting has been extended to support sorting by
7462 cipher-strength (using the strength_bits hard coded in the tables).
7463 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
7464
7465 Fix a bug in the cipher-command parser: when supplying a cipher command
7466 string with an "undefined" symbol (neither command nor alphanumeric
7467 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
7468 an error is flagged.
7469
7470 Due to the strength-sorting extension, the code of the
7471 ssl_create_cipher_list() function was completely rearranged. I hope that
7472 the readability was also increased :-)
7473 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7474
7475 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
7476 for the first serial number and places 2 in the serial number file. This
7477 avoids problems when the root CA is created with serial number zero and
7478 the first user certificate has the same issuer name and serial number
7479 as the root CA.
7480 [Steve Henson]
7481
7482 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
7483 the new code. Add documentation for this stuff.
7484 [Steve Henson]
7485
7486 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
7487 X509_*() to X509at_*() on the grounds that they don't handle X509
7488 structures and behave in an analagous way to the X509v3 functions:
7489 they shouldn't be called directly but wrapper functions should be used
7490 instead.
7491
7492 So we also now have some wrapper functions that call the X509at functions
7493 when passed certificate requests. (TO DO: similar things can be done with
7494 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
7495 things. Some of these need some d2i or i2d and print functionality
7496 because they handle more complex structures.)
7497 [Steve Henson]
7498
7499 *) Add missing #ifndefs that caused missing symbols when building libssl
7500 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
7501 NO_RSA in ssl/s2*.c.
7502 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
7503
7504 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
7505 has a return value which indicates the quality of the random data
7506 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
7507 error queue. New function RAND_pseudo_bytes() generates output that is
7508 guaranteed to be unique but not unpredictable. RAND_add is like
7509 RAND_seed, but takes an extra argument for an entropy estimate
7510 (RAND_seed always assumes full entropy).
7511 [Ulf Möller]
7512
7513 *) Do more iterations of Rabin-Miller probable prime test (specifically,
7514 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
7515 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
7516 in crypto/bn/bn_prime.c for the complete table). This guarantees a
7517 false-positive rate of at most 2^-80 for random input.
7518 [Bodo Moeller]
7519
7520 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
7521 [Bodo Moeller]
7522
7523 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
7524 in the 0.9.5 release), this returns the chain
7525 from an X509_CTX structure with a dup of the stack and all
7526 the X509 reference counts upped: so the stack will exist
7527 after X509_CTX_cleanup() has been called. Modify pkcs12.c
7528 to use this.
7529
7530 Also make SSL_SESSION_print() print out the verify return
7531 code.
7532 [Steve Henson]
7533
7534 *) Add manpage for the pkcs12 command. Also change the default
7535 behaviour so MAC iteration counts are used unless the new
7536 -nomaciter option is used. This improves file security and
7537 only older versions of MSIE (4.0 for example) need it.
7538 [Steve Henson]
7539
7540 *) Honor the no-xxx Configure options when creating .DEF files.
7541 [Ulf Möller]
7542
7543 *) Add PKCS#10 attributes to field table: challengePassword,
7544 unstructuredName and unstructuredAddress. These are taken from
7545 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
7546 international characters are used.
7547
7548 More changes to X509_ATTRIBUTE code: allow the setting of types
7549 based on strings. Remove the 'loc' parameter when adding
7550 attributes because these will be a SET OF encoding which is sorted
7551 in ASN1 order.
7552 [Steve Henson]
7553
7554 *) Initial changes to the 'req' utility to allow request generation
7555 automation. This will allow an application to just generate a template
7556 file containing all the field values and have req construct the
7557 request.
7558
7559 Initial support for X509_ATTRIBUTE handling. Stacks of these are
7560 used all over the place including certificate requests and PKCS#7
7561 structures. They are currently handled manually where necessary with
7562 some primitive wrappers for PKCS#7. The new functions behave in a
7563 manner analogous to the X509 extension functions: they allow
7564 attributes to be looked up by NID and added.
7565
7566 Later something similar to the X509V3 code would be desirable to
7567 automatically handle the encoding, decoding and printing of the
7568 more complex types. The string types like challengePassword can
7569 be handled by the string table functions.
7570
7571 Also modified the multi byte string table handling. Now there is
7572 a 'global mask' which masks out certain types. The table itself
7573 can use the flag STABLE_NO_MASK to ignore the mask setting: this
7574 is useful when for example there is only one permissible type
7575 (as in countryName) and using the mask might result in no valid
7576 types at all.
7577 [Steve Henson]
7578
7579 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
7580 SSL_get_peer_finished to allow applications to obtain the latest
7581 Finished messages sent to the peer or expected from the peer,
7582 respectively. (SSL_get_peer_finished is usually the Finished message
7583 actually received from the peer, otherwise the protocol will be aborted.)
7584
7585 As the Finished message are message digests of the complete handshake
7586 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
7587 be used for external authentication procedures when the authentication
7588 provided by SSL/TLS is not desired or is not enough.
7589 [Bodo Moeller]
7590
7591 *) Enhanced support for Alpha Linux is added. Now ./config checks if
7592 the host supports BWX extension and if Compaq C is present on the
7593 $PATH. Just exploiting of the BWX extension results in 20-30%
7594 performance kick for some algorithms, e.g. DES and RC4 to mention
7595 a couple. Compaq C in turn generates ~20% faster code for MD5 and
7596 SHA1.
7597 [Andy Polyakov]
7598
7599 *) Add support for MS "fast SGC". This is arguably a violation of the
7600 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
7601 weak crypto and after checking the certificate is SGC a second one
7602 with strong crypto. MS SGC stops the first handshake after receiving
7603 the server certificate message and sends a second client hello. Since
7604 a server will typically do all the time consuming operations before
7605 expecting any further messages from the client (server key exchange
7606 is the most expensive) there is little difference between the two.
7607
7608 To get OpenSSL to support MS SGC we have to permit a second client
7609 hello message after we have sent server done. In addition we have to
7610 reset the MAC if we do get this second client hello.
7611 [Steve Henson]
7612
7613 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
7614 if a DER encoded private key is RSA or DSA traditional format. Changed
7615 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
7616 format DER encoded private key. Newer code should use PKCS#8 format which
7617 has the key type encoded in the ASN1 structure. Added DER private key
7618 support to pkcs8 application.
7619 [Steve Henson]
7620
7621 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
7622 ciphersuites has been selected (as required by the SSL 3/TLS 1
7623 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
7624 is set, we interpret this as a request to violate the specification
7625 (the worst that can happen is a handshake failure, and 'correct'
7626 behaviour would result in a handshake failure anyway).
7627 [Bodo Moeller]
7628
7629 *) In SSL_CTX_add_session, take into account that there might be multiple
7630 SSL_SESSION structures with the same session ID (e.g. when two threads
7631 concurrently obtain them from an external cache).
7632 The internal cache can handle only one SSL_SESSION with a given ID,
7633 so if there's a conflict, we now throw out the old one to achieve
7634 consistency.
7635 [Bodo Moeller]
7636
7637 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
7638 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
7639 some routines that use cipher OIDs: some ciphers do not have OIDs
7640 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
7641 example.
7642 [Steve Henson]
7643
7644 *) Simplify the trust setting structure and code. Now we just have
7645 two sequences of OIDs for trusted and rejected settings. These will
7646 typically have values the same as the extended key usage extension
7647 and any application specific purposes.
7648
7649 The trust checking code now has a default behaviour: it will just
7650 check for an object with the same NID as the passed id. Functions can
7651 be provided to override either the default behaviour or the behaviour
7652 for a given id. SSL client, server and email already have functions
7653 in place for compatibility: they check the NID and also return "trusted"
7654 if the certificate is self signed.
7655 [Steve Henson]
7656
7657 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
7658 traditional format into an EVP_PKEY structure.
7659 [Steve Henson]
7660
7661 *) Add a password callback function PEM_cb() which either prompts for
7662 a password if usr_data is NULL or otherwise assumes it is a null
7663 terminated password. Allow passwords to be passed on command line
7664 environment or config files in a few more utilities.
7665 [Steve Henson]
7666
7667 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
7668 keys. Add some short names for PKCS#8 PBE algorithms and allow them
7669 to be specified on the command line for the pkcs8 and pkcs12 utilities.
7670 Update documentation.
7671 [Steve Henson]
7672
7673 *) Support for ASN1 "NULL" type. This could be handled before by using
7674 ASN1_TYPE but there wasn't any function that would try to read a NULL
7675 and produce an error if it couldn't. For compatibility we also have
7676 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
7677 don't allocate anything because they don't need to.
7678 [Steve Henson]
7679
7680 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
7681 for details.
7682 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
7683
7684 *) Rebuild of the memory allocation routines used by OpenSSL code and
7685 possibly others as well. The purpose is to make an interface that
7686 provide hooks so anyone can build a separate set of allocation and
7687 deallocation routines to be used by OpenSSL, for example memory
7688 pool implementations, or something else, which was previously hard
7689 since Malloc(), Realloc() and Free() were defined as macros having
7690 the values malloc, realloc and free, respectively (except for Win32
7691 compilations). The same is provided for memory debugging code.
7692 OpenSSL already comes with functionality to find memory leaks, but
7693 this gives people a chance to debug other memory problems.
7694
7695 With these changes, a new set of functions and macros have appeared:
7696
7697 CRYPTO_set_mem_debug_functions() [F]
7698 CRYPTO_get_mem_debug_functions() [F]
7699 CRYPTO_dbg_set_options() [F]
7700 CRYPTO_dbg_get_options() [F]
7701 CRYPTO_malloc_debug_init() [M]
7702
7703 The memory debug functions are NULL by default, unless the library
7704 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
7705 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
7706 gives the standard debugging functions that come with OpenSSL) or
7707 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
7708 provided by the library user) must be used. When the standard
7709 debugging functions are used, CRYPTO_dbg_set_options can be used to
7710 request additional information:
7711 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
7712 the CRYPTO_MDEBUG_xxx macro when compiling the library.
7713
7714 Also, things like CRYPTO_set_mem_functions will always give the
7715 expected result (the new set of functions is used for allocation
7716 and deallocation) at all times, regardless of platform and compiler
7717 options.
7718
7719 To finish it up, some functions that were never use in any other
7720 way than through macros have a new API and new semantic:
7721
7722 CRYPTO_dbg_malloc()
7723 CRYPTO_dbg_realloc()
7724 CRYPTO_dbg_free()
7725
7726 All macros of value have retained their old syntax.
7727 [Richard Levitte and Bodo Moeller]
7728
7729 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
7730 ordering of SMIMECapabilities wasn't in "strength order" and there
7731 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
7732 algorithm.
7733 [Steve Henson]
7734
7735 *) Some ASN1 types with illegal zero length encoding (INTEGER,
7736 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
7737 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
7738
7739 *) Merge in my S/MIME library for OpenSSL. This provides a simple
7740 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
7741 functionality to handle multipart/signed properly) and a utility
7742 called 'smime' to call all this stuff. This is based on code I
7743 originally wrote for Celo who have kindly allowed it to be
7744 included in OpenSSL.
7745 [Steve Henson]
7746
7747 *) Add variants des_set_key_checked and des_set_key_unchecked of
7748 des_set_key (aka des_key_sched). Global variable des_check_key
7749 decides which of these is called by des_set_key; this way
7750 des_check_key behaves as it always did, but applications and
7751 the library itself, which was buggy for des_check_key == 1,
7752 have a cleaner way to pick the version they need.
7753 [Bodo Moeller]
7754
7755 *) New function PKCS12_newpass() which changes the password of a
7756 PKCS12 structure.
7757 [Steve Henson]
7758
7759 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
7760 dynamic mix. In both cases the ids can be used as an index into the
7761 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
7762 functions so they accept a list of the field values and the
7763 application doesn't need to directly manipulate the X509_TRUST
7764 structure.
7765 [Steve Henson]
7766
7767 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
7768 need initialising.
7769 [Steve Henson]
7770
7771 *) Modify the way the V3 extension code looks up extensions. This now
7772 works in a similar way to the object code: we have some "standard"
7773 extensions in a static table which is searched with OBJ_bsearch()
7774 and the application can add dynamic ones if needed. The file
7775 crypto/x509v3/ext_dat.h now has the info: this file needs to be
7776 updated whenever a new extension is added to the core code and kept
7777 in ext_nid order. There is a simple program 'tabtest.c' which checks
7778 this. New extensions are not added too often so this file can readily
7779 be maintained manually.
7780
7781 There are two big advantages in doing things this way. The extensions
7782 can be looked up immediately and no longer need to be "added" using
7783 X509V3_add_standard_extensions(): this function now does nothing.
7784 [Side note: I get *lots* of email saying the extension code doesn't
7785 work because people forget to call this function]
7786 Also no dynamic allocation is done unless new extensions are added:
7787 so if we don't add custom extensions there is no need to call
7788 X509V3_EXT_cleanup().
7789 [Steve Henson]
7790
7791 *) Modify enc utility's salting as follows: make salting the default. Add a
7792 magic header, so unsalted files fail gracefully instead of just decrypting
7793 to garbage. This is because not salting is a big security hole, so people
7794 should be discouraged from doing it.
7795 [Ben Laurie]
7796
7797 *) Fixes and enhancements to the 'x509' utility. It allowed a message
7798 digest to be passed on the command line but it only used this
7799 parameter when signing a certificate. Modified so all relevant
7800 operations are affected by the digest parameter including the
7801 -fingerprint and -x509toreq options. Also -x509toreq choked if a
7802 DSA key was used because it didn't fix the digest.
7803 [Steve Henson]
7804
7805 *) Initial certificate chain verify code. Currently tests the untrusted
7806 certificates for consistency with the verify purpose (which is set
7807 when the X509_STORE_CTX structure is set up) and checks the pathlength.
7808
7809 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
7810 this is because it will reject chains with invalid extensions whereas
7811 every previous version of OpenSSL and SSLeay made no checks at all.
7812
7813 Trust code: checks the root CA for the relevant trust settings. Trust
7814 settings have an initial value consistent with the verify purpose: e.g.
7815 if the verify purpose is for SSL client use it expects the CA to be
7816 trusted for SSL client use. However the default value can be changed to
7817 permit custom trust settings: one example of this would be to only trust
7818 certificates from a specific "secure" set of CAs.
7819
7820 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
7821 which should be used for version portability: especially since the
7822 verify structure is likely to change more often now.
7823
7824 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
7825 to set them. If not set then assume SSL clients will verify SSL servers
7826 and vice versa.
7827
7828 Two new options to the verify program: -untrusted allows a set of
7829 untrusted certificates to be passed in and -purpose which sets the
7830 intended purpose of the certificate. If a purpose is set then the
7831 new chain verify code is used to check extension consistency.
7832 [Steve Henson]
7833
7834 *) Support for the authority information access extension.
7835 [Steve Henson]
7836
7837 *) Modify RSA and DSA PEM read routines to transparently handle
7838 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
7839 public keys in a format compatible with certificate
7840 SubjectPublicKeyInfo structures. Unfortunately there were already
7841 functions called *_PublicKey_* which used various odd formats so
7842 these are retained for compatibility: however the DSA variants were
7843 never in a public release so they have been deleted. Changed dsa/rsa
7844 utilities to handle the new format: note no releases ever handled public
7845 keys so we should be OK.
7846
7847 The primary motivation for this change is to avoid the same fiasco
7848 that dogs private keys: there are several incompatible private key
7849 formats some of which are standard and some OpenSSL specific and
7850 require various evil hacks to allow partial transparent handling and
7851 even then it doesn't work with DER formats. Given the option anything
7852 other than PKCS#8 should be dumped: but the other formats have to
7853 stay in the name of compatibility.
7854
7855 With public keys and the benefit of hindsight one standard format
7856 is used which works with EVP_PKEY, RSA or DSA structures: though
7857 it clearly returns an error if you try to read the wrong kind of key.
7858
7859 Added a -pubkey option to the 'x509' utility to output the public key.
7860 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
7861 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
7862 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
7863 that do the same as the EVP_PKEY_assign_*() except they up the
7864 reference count of the added key (they don't "swallow" the
7865 supplied key).
7866 [Steve Henson]
7867
7868 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
7869 CRLs would fail if the file contained no certificates or no CRLs:
7870 added a new function to read in both types and return the number
7871 read: this means that if none are read it will be an error. The
7872 DER versions of the certificate and CRL reader would always fail
7873 because it isn't possible to mix certificates and CRLs in DER format
7874 without choking one or the other routine. Changed this to just read
7875 a certificate: this is the best we can do. Also modified the code
7876 in apps/verify.c to take notice of return codes: it was previously
7877 attempting to read in certificates from NULL pointers and ignoring
7878 any errors: this is one reason why the cert and CRL reader seemed
7879 to work. It doesn't check return codes from the default certificate
7880 routines: these may well fail if the certificates aren't installed.
7881 [Steve Henson]
7882
7883 *) Code to support otherName option in GeneralName.
7884 [Steve Henson]
7885
7886 *) First update to verify code. Change the verify utility
7887 so it warns if it is passed a self signed certificate:
7888 for consistency with the normal behaviour. X509_verify
7889 has been modified to it will now verify a self signed
7890 certificate if *exactly* the same certificate appears
7891 in the store: it was previously impossible to trust a
7892 single self signed certificate. This means that:
7893 openssl verify ss.pem
7894 now gives a warning about a self signed certificate but
7895 openssl verify -CAfile ss.pem ss.pem
7896 is OK.
7897 [Steve Henson]
7898
7899 *) For servers, store verify_result in SSL_SESSION data structure
7900 (and add it to external session representation).
7901 This is needed when client certificate verifications fails,
7902 but an application-provided verification callback (set by
7903 SSL_CTX_set_cert_verify_callback) allows accepting the session
7904 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
7905 but returns 1): When the session is reused, we have to set
7906 ssl->verify_result to the appropriate error code to avoid
7907 security holes.
7908 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
7909
7910 *) Fix a bug in the new PKCS#7 code: it didn't consider the
7911 case in PKCS7_dataInit() where the signed PKCS7 structure
7912 didn't contain any existing data because it was being created.
7913 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
7914
7915 *) Add a salt to the key derivation routines in enc.c. This
7916 forms the first 8 bytes of the encrypted file. Also add a
7917 -S option to allow a salt to be input on the command line.
7918 [Steve Henson]
7919
7920 *) New function X509_cmp(). Oddly enough there wasn't a function
7921 to compare two certificates. We do this by working out the SHA1
7922 hash and comparing that. X509_cmp() will be needed by the trust
7923 code.
7924 [Steve Henson]
7925
7926 *) SSL_get1_session() is like SSL_get_session(), but increments
7927 the reference count in the SSL_SESSION returned.
7928 [Geoff Thorpe <geoff@eu.c2.net>]
7929
7930 *) Fix for 'req': it was adding a null to request attributes.
7931 Also change the X509_LOOKUP and X509_INFO code to handle
7932 certificate auxiliary information.
7933 [Steve Henson]
7934
7935 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
7936 the 'enc' command.
7937 [Steve Henson]
7938
7939 *) Add the possibility to add extra information to the memory leak
7940 detecting output, to form tracebacks, showing from where each
7941 allocation was originated: CRYPTO_push_info("constant string") adds
7942 the string plus current file name and line number to a per-thread
7943 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
7944 is like calling CYRPTO_pop_info() until the stack is empty.
7945 Also updated memory leak detection code to be multi-thread-safe.
7946 [Richard Levitte]
7947
7948 *) Add options -text and -noout to pkcs7 utility and delete the
7949 encryption options which never did anything. Update docs.
7950 [Steve Henson]
7951
7952 *) Add options to some of the utilities to allow the pass phrase
7953 to be included on either the command line (not recommended on
7954 OSes like Unix) or read from the environment. Update the
7955 manpages and fix a few bugs.
7956 [Steve Henson]
7957
7958 *) Add a few manpages for some of the openssl commands.
7959 [Steve Henson]
7960
7961 *) Fix the -revoke option in ca. It was freeing up memory twice,
7962 leaking and not finding already revoked certificates.
7963 [Steve Henson]
7964
7965 *) Extensive changes to support certificate auxiliary information.
7966 This involves the use of X509_CERT_AUX structure and X509_AUX
7967 functions. An X509_AUX function such as PEM_read_X509_AUX()
7968 can still read in a certificate file in the usual way but it
7969 will also read in any additional "auxiliary information". By
7970 doing things this way a fair degree of compatibility can be
7971 retained: existing certificates can have this information added
7972 using the new 'x509' options.
7973
7974 Current auxiliary information includes an "alias" and some trust
7975 settings. The trust settings will ultimately be used in enhanced
7976 certificate chain verification routines: currently a certificate
7977 can only be trusted if it is self signed and then it is trusted
7978 for all purposes.
7979 [Steve Henson]
7980
7981 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7982 The problem was that one of the replacement routines had not been working
7983 since SSLeay releases. For now the offending routine has been replaced
7984 with non-optimised assembler. Even so, this now gives around 95%
7985 performance improvement for 1024 bit RSA signs.
7986 [Mark Cox]
7987
7988 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7989 handling. Most clients have the effective key size in bits equal to
7990 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7991 A few however don't do this and instead use the size of the decrypted key
7992 to determine the RC2 key length and the AlgorithmIdentifier to determine
7993 the effective key length. In this case the effective key length can still
7994 be 40 bits but the key length can be 168 bits for example. This is fixed
7995 by manually forcing an RC2 key into the EVP_PKEY structure because the
7996 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7997 the key length and effective key length are equal.
7998 [Steve Henson]
7999
8000 *) Add a bunch of functions that should simplify the creation of
8001 X509_NAME structures. Now you should be able to do:
8002 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
8003 and have it automatically work out the correct field type and fill in
8004 the structures. The more adventurous can try:
8005 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
8006 and it will (hopefully) work out the correct multibyte encoding.
8007 [Steve Henson]
8008
8009 *) Change the 'req' utility to use the new field handling and multibyte
8010 copy routines. Before the DN field creation was handled in an ad hoc
8011 way in req, ca, and x509 which was rather broken and didn't support
8012 BMPStrings or UTF8Strings. Since some software doesn't implement
8013 BMPStrings or UTF8Strings yet, they can be enabled using the config file
8014 using the dirstring_type option. See the new comment in the default
8015 openssl.cnf for more info.
8016 [Steve Henson]
8017
8018 *) Make crypto/rand/md_rand.c more robust:
8019 - Assure unique random numbers after fork().
8020 - Make sure that concurrent threads access the global counter and
8021 md serializably so that we never lose entropy in them
8022 or use exactly the same state in multiple threads.
8023 Access to the large state is not always serializable because
8024 the additional locking could be a performance killer, and
8025 md should be large enough anyway.
8026 [Bodo Moeller]
8027
8028 *) New file apps/app_rand.c with commonly needed functionality
8029 for handling the random seed file.
8030
8031 Use the random seed file in some applications that previously did not:
8032 ca,
8033 dsaparam -genkey (which also ignored its '-rand' option),
8034 s_client,
8035 s_server,
8036 x509 (when signing).
8037 Except on systems with /dev/urandom, it is crucial to have a random
8038 seed file at least for key creation, DSA signing, and for DH exchanges;
8039 for RSA signatures we could do without one.
8040
8041 gendh and gendsa (unlike genrsa) used to read only the first byte
8042 of each file listed in the '-rand' option. The function as previously
8043 found in genrsa is now in app_rand.c and is used by all programs
8044 that support '-rand'.
8045 [Bodo Moeller]
8046
8047 *) In RAND_write_file, use mode 0600 for creating files;
8048 don't just chmod when it may be too late.
8049 [Bodo Moeller]
8050
8051 *) Report an error from X509_STORE_load_locations
8052 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
8053 [Bill Perry]
8054
8055 *) New function ASN1_mbstring_copy() this copies a string in either
8056 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
8057 into an ASN1_STRING type. A mask of permissible types is passed
8058 and it chooses the "minimal" type to use or an error if not type
8059 is suitable.
8060 [Steve Henson]
8061
8062 *) Add function equivalents to the various macros in asn1.h. The old
8063 macros are retained with an M_ prefix. Code inside the library can
8064 use the M_ macros. External code (including the openssl utility)
8065 should *NOT* in order to be "shared library friendly".
8066 [Steve Henson]
8067
8068 *) Add various functions that can check a certificate's extensions
8069 to see if it usable for various purposes such as SSL client,
8070 server or S/MIME and CAs of these types. This is currently
8071 VERY EXPERIMENTAL but will ultimately be used for certificate chain
8072 verification. Also added a -purpose flag to x509 utility to
8073 print out all the purposes.
8074 [Steve Henson]
8075
8076 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
8077 functions.
8078 [Steve Henson]
8079
8080 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
8081 for, obtain and decode and extension and obtain its critical flag.
8082 This allows all the necessary extension code to be handled in a
8083 single function call.
8084 [Steve Henson]
8085
8086 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
8087 platforms. See crypto/rc4/rc4_enc.c for further details.
8088 [Andy Polyakov]
8089
8090 *) New -noout option to asn1parse. This causes no output to be produced
8091 its main use is when combined with -strparse and -out to extract data
8092 from a file (which may not be in ASN.1 format).
8093 [Steve Henson]
8094
8095 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
8096 when producing the local key id.
8097 [Richard Levitte <levitte@stacken.kth.se>]
8098
8099 *) New option -dhparam in s_server. This allows a DH parameter file to be
8100 stated explicitly. If it is not stated then it tries the first server
8101 certificate file. The previous behaviour hard coded the filename
8102 "server.pem".
8103 [Steve Henson]
8104
8105 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
8106 a public key to be input or output. For example:
8107 openssl rsa -in key.pem -pubout -out pubkey.pem
8108 Also added necessary DSA public key functions to handle this.
8109 [Steve Henson]
8110
8111 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
8112 in the message. This was handled by allowing
8113 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
8114 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
8115
8116 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
8117 to the end of the strings whereas this didn't. This would cause problems
8118 if strings read with d2i_ASN1_bytes() were later modified.
8119 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
8120
8121 *) Fix for base64 decode bug. When a base64 bio reads only one line of
8122 data and it contains EOF it will end up returning an error. This is
8123 caused by input 46 bytes long. The cause is due to the way base64
8124 BIOs find the start of base64 encoded data. They do this by trying a
8125 trial decode on each line until they find one that works. When they
8126 do a flag is set and it starts again knowing it can pass all the
8127 data directly through the decoder. Unfortunately it doesn't reset
8128 the context it uses. This means that if EOF is reached an attempt
8129 is made to pass two EOFs through the context and this causes the
8130 resulting error. This can also cause other problems as well. As is
8131 usual with these problems it takes *ages* to find and the fix is
8132 trivial: move one line.
8133 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
8134
8135 *) Ugly workaround to get s_client and s_server working under Windows. The
8136 old code wouldn't work because it needed to select() on sockets and the
8137 tty (for keypresses and to see if data could be written). Win32 only
8138 supports select() on sockets so we select() with a 1s timeout on the
8139 sockets and then see if any characters are waiting to be read, if none
8140 are present then we retry, we also assume we can always write data to
8141 the tty. This isn't nice because the code then blocks until we've
8142 received a complete line of data and it is effectively polling the
8143 keyboard at 1s intervals: however it's quite a bit better than not
8144 working at all :-) A dedicated Windows application might handle this
8145 with an event loop for example.
8146 [Steve Henson]
8147
8148 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
8149 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
8150 will be called when RSA_sign() and RSA_verify() are used. This is useful
8151 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
8152 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
8153 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
8154 This necessitated the support of an extra signature type NID_md5_sha1
8155 for SSL signatures and modifications to the SSL library to use it instead
8156 of calling RSA_public_decrypt() and RSA_private_encrypt().
8157 [Steve Henson]
8158
8159 *) Add new -verify -CAfile and -CApath options to the crl program, these
8160 will lookup a CRL issuers certificate and verify the signature in a
8161 similar way to the verify program. Tidy up the crl program so it
8162 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
8163 less strict. It will now permit CRL extensions even if it is not
8164 a V2 CRL: this will allow it to tolerate some broken CRLs.
8165 [Steve Henson]
8166
8167 *) Initialize all non-automatic variables each time one of the openssl
8168 sub-programs is started (this is necessary as they may be started
8169 multiple times from the "OpenSSL>" prompt).
8170 [Lennart Bang, Bodo Moeller]
8171
8172 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
8173 removing all other RSA functionality (this is what NO_RSA does). This
8174 is so (for example) those in the US can disable those operations covered
8175 by the RSA patent while allowing storage and parsing of RSA keys and RSA
8176 key generation.
8177 [Steve Henson]
8178
8179 *) Non-copying interface to BIO pairs.
8180 (still largely untested)
8181 [Bodo Moeller]
8182
8183 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
8184 ASCII string. This was handled independently in various places before.
8185 [Steve Henson]
8186
8187 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
8188 UTF8 strings a character at a time.
8189 [Steve Henson]
8190
8191 *) Use client_version from client hello to select the protocol
8192 (s23_srvr.c) and for RSA client key exchange verification
8193 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
8194 [Bodo Moeller]
8195
8196 *) Add various utility functions to handle SPKACs, these were previously
8197 handled by poking round in the structure internals. Added new function
8198 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
8199 print, verify and generate SPKACs. Based on an original idea from
8200 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
8201 [Steve Henson]
8202
8203 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
8204 [Andy Polyakov]
8205
8206 *) Allow the config file extension section to be overwritten on the
8207 command line. Based on an original idea from Massimiliano Pala
8208 <madwolf@comune.modena.it>. The new option is called -extensions
8209 and can be applied to ca, req and x509. Also -reqexts to override
8210 the request extensions in req and -crlexts to override the crl extensions
8211 in ca.
8212 [Steve Henson]
8213
8214 *) Add new feature to the SPKAC handling in ca. Now you can include
8215 the same field multiple times by preceding it by "XXXX." for example:
8216 1.OU="Unit name 1"
8217 2.OU="Unit name 2"
8218 this is the same syntax as used in the req config file.
8219 [Steve Henson]
8220
8221 *) Allow certificate extensions to be added to certificate requests. These
8222 are specified in a 'req_extensions' option of the req section of the
8223 config file. They can be printed out with the -text option to req but
8224 are otherwise ignored at present.
8225 [Steve Henson]
8226
8227 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
8228 data read consists of only the final block it would not decrypted because
8229 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
8230 A misplaced 'break' also meant the decrypted final block might not be
8231 copied until the next read.
8232 [Steve Henson]
8233
8234 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
8235 a few extra parameters to the DH structure: these will be useful if
8236 for example we want the value of 'q' or implement X9.42 DH.
8237 [Steve Henson]
8238
8239 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
8240 provides hooks that allow the default DSA functions or functions on a
8241 "per key" basis to be replaced. This allows hardware acceleration and
8242 hardware key storage to be handled without major modification to the
8243 library. Also added low level modexp hooks and CRYPTO_EX structure and
8244 associated functions.
8245 [Steve Henson]
8246
8247 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
8248 as "read only": it can't be written to and the buffer it points to will
8249 not be freed. Reading from a read only BIO is much more efficient than
8250 a normal memory BIO. This was added because there are several times when
8251 an area of memory needs to be read from a BIO. The previous method was
8252 to create a memory BIO and write the data to it, this results in two
8253 copies of the data and an O(n^2) reading algorithm. There is a new
8254 function BIO_new_mem_buf() which creates a read only memory BIO from
8255 an area of memory. Also modified the PKCS#7 routines to use read only
8256 memory BIOs.
8257 [Steve Henson]
8258
8259 *) Bugfix: ssl23_get_client_hello did not work properly when called in
8260 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
8261 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
8262 but a retry condition occured while trying to read the rest.
8263 [Bodo Moeller]
8264
8265 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
8266 NID_pkcs7_encrypted by default: this was wrong since this should almost
8267 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
8268 the encrypted data type: this is a more sensible place to put it and it
8269 allows the PKCS#12 code to be tidied up that duplicated this
8270 functionality.
8271 [Steve Henson]
8272
8273 *) Changed obj_dat.pl script so it takes its input and output files on
8274 the command line. This should avoid shell escape redirection problems
8275 under Win32.
8276 [Steve Henson]
8277
8278 *) Initial support for certificate extension requests, these are included
8279 in things like Xenroll certificate requests. Included functions to allow
8280 extensions to be obtained and added.
8281 [Steve Henson]
8282
8283 *) -crlf option to s_client and s_server for sending newlines as
8284 CRLF (as required by many protocols).
8285 [Bodo Moeller]
8286
8287 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
8288
8289 *) Install libRSAglue.a when OpenSSL is built with RSAref.
8290 [Ralf S. Engelschall]
8291
8292 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
8293 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
8294
8295 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
8296 program.
8297 [Steve Henson]
8298
8299 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
8300 DH parameters/keys (q is lost during that conversion, but the resulting
8301 DH parameters contain its length).
8302
8303 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
8304 much faster than DH_generate_parameters (which creates parameters
8305 where p = 2*q + 1), and also the smaller q makes DH computations
8306 much more efficient (160-bit exponentiation instead of 1024-bit
8307 exponentiation); so this provides a convenient way to support DHE
8308 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
8309 utter importance to use
8310 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8311 or
8312 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
8313 when such DH parameters are used, because otherwise small subgroup
8314 attacks may become possible!
8315 [Bodo Moeller]
8316
8317 *) Avoid memory leak in i2d_DHparams.
8318 [Bodo Moeller]
8319
8320 *) Allow the -k option to be used more than once in the enc program:
8321 this allows the same encrypted message to be read by multiple recipients.
8322 [Steve Henson]
8323
8324 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
8325 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
8326 it will always use the numerical form of the OID, even if it has a short
8327 or long name.
8328 [Steve Henson]
8329
8330 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
8331 method only got called if p,q,dmp1,dmq1,iqmp components were present,
8332 otherwise bn_mod_exp was called. In the case of hardware keys for example
8333 no private key components need be present and it might store extra data
8334 in the RSA structure, which cannot be accessed from bn_mod_exp.
8335 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
8336 private key operations.
8337 [Steve Henson]
8338
8339 *) Added support for SPARC Linux.
8340 [Andy Polyakov]
8341
8342 *) pem_password_cb function type incompatibly changed from
8343 typedef int pem_password_cb(char *buf, int size, int rwflag);
8344 to
8345 ....(char *buf, int size, int rwflag, void *userdata);
8346 so that applications can pass data to their callbacks:
8347 The PEM[_ASN1]_{read,write}... functions and macros now take an
8348 additional void * argument, which is just handed through whenever
8349 the password callback is called.
8350 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
8351
8352 New function SSL_CTX_set_default_passwd_cb_userdata.
8353
8354 Compatibility note: As many C implementations push function arguments
8355 onto the stack in reverse order, the new library version is likely to
8356 interoperate with programs that have been compiled with the old
8357 pem_password_cb definition (PEM_whatever takes some data that
8358 happens to be on the stack as its last argument, and the callback
8359 just ignores this garbage); but there is no guarantee whatsoever that
8360 this will work.
8361
8362 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
8363 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
8364 problems not only on Windows, but also on some Unix platforms.
8365 To avoid problematic command lines, these definitions are now in an
8366 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
8367 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
8368 [Bodo Moeller]
8369
8370 *) MIPS III/IV assembler module is reimplemented.
8371 [Andy Polyakov]
8372
8373 *) More DES library cleanups: remove references to srand/rand and
8374 delete an unused file.
8375 [Ulf Möller]
8376
8377 *) Add support for the the free Netwide assembler (NASM) under Win32,
8378 since not many people have MASM (ml) and it can be hard to obtain.
8379 This is currently experimental but it seems to work OK and pass all
8380 the tests. Check out INSTALL.W32 for info.
8381 [Steve Henson]
8382
8383 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
8384 without temporary keys kept an extra copy of the server key,
8385 and connections with temporary keys did not free everything in case
8386 of an error.
8387 [Bodo Moeller]
8388
8389 *) New function RSA_check_key and new openssl rsa option -check
8390 for verifying the consistency of RSA keys.
8391 [Ulf Moeller, Bodo Moeller]
8392
8393 *) Various changes to make Win32 compile work:
8394 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
8395 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
8396 comparison" warnings.
8397 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
8398 [Steve Henson]
8399
8400 *) Add a debugging option to PKCS#5 v2 key generation function: when
8401 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
8402 derived keys are printed to stderr.
8403 [Steve Henson]
8404
8405 *) Copy the flags in ASN1_STRING_dup().
8406 [Roman E. Pavlov <pre@mo.msk.ru>]
8407
8408 *) The x509 application mishandled signing requests containing DSA
8409 keys when the signing key was also DSA and the parameters didn't match.
8410
8411 It was supposed to omit the parameters when they matched the signing key:
8412 the verifying software was then supposed to automatically use the CA's
8413 parameters if they were absent from the end user certificate.
8414
8415 Omitting parameters is no longer recommended. The test was also
8416 the wrong way round! This was probably due to unusual behaviour in
8417 EVP_cmp_parameters() which returns 1 if the parameters match.
8418 This meant that parameters were omitted when they *didn't* match and
8419 the certificate was useless. Certificates signed with 'ca' didn't have
8420 this bug.
8421 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
8422
8423 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
8424 The interface is as follows:
8425 Applications can use
8426 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
8427 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
8428 "off" is now the default.
8429 The library internally uses
8430 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
8431 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
8432 to disable memory-checking temporarily.
8433
8434 Some inconsistent states that previously were possible (and were
8435 even the default) are now avoided.
8436
8437 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
8438 with each memory chunk allocated; this is occasionally more helpful
8439 than just having a counter.
8440
8441 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
8442
8443 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
8444 extensions.
8445 [Bodo Moeller]
8446
8447 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
8448 which largely parallels "options", but is for changing API behaviour,
8449 whereas "options" are about protocol behaviour.
8450 Initial "mode" flags are:
8451
8452 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
8453 a single record has been written.
8454 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
8455 retries use the same buffer location.
8456 (But all of the contents must be
8457 copied!)
8458 [Bodo Moeller]
8459
8460 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
8461 worked.
8462
8463 *) Fix problems with no-hmac etc.
8464 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
8465
8466 *) New functions RSA_get_default_method(), RSA_set_method() and
8467 RSA_get_method(). These allows replacement of RSA_METHODs without having
8468 to mess around with the internals of an RSA structure.
8469 [Steve Henson]
8470
8471 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
8472 Also really enable memory leak checks in openssl.c and in some
8473 test programs.
8474 [Chad C. Mulligan, Bodo Moeller]
8475
8476 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
8477 up the length of negative integers. This has now been simplified to just
8478 store the length when it is first determined and use it later, rather
8479 than trying to keep track of where data is copied and updating it to
8480 point to the end.
8481 [Steve Henson, reported by Brien Wheeler
8482 <bwheeler@authentica-security.com>]
8483
8484 *) Add a new function PKCS7_signatureVerify. This allows the verification
8485 of a PKCS#7 signature but with the signing certificate passed to the
8486 function itself. This contrasts with PKCS7_dataVerify which assumes the
8487 certificate is present in the PKCS#7 structure. This isn't always the
8488 case: certificates can be omitted from a PKCS#7 structure and be
8489 distributed by "out of band" means (such as a certificate database).
8490 [Steve Henson]
8491
8492 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
8493 function prototypes in pem.h, also change util/mkdef.pl to add the
8494 necessary function names.
8495 [Steve Henson]
8496
8497 *) mk1mf.pl (used by Windows builds) did not properly read the
8498 options set by Configure in the top level Makefile, and Configure
8499 was not even able to write more than one option correctly.
8500 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
8501 [Bodo Moeller]
8502
8503 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
8504 file to be loaded from a BIO or FILE pointer. The BIO version will
8505 for example allow memory BIOs to contain config info.
8506 [Steve Henson]
8507
8508 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
8509 Whoever hopes to achieve shared-library compatibility across versions
8510 must use this, not the compile-time macro.
8511 (Exercise 0.9.4: Which is the minimum library version required by
8512 such programs?)
8513 Note: All this applies only to multi-threaded programs, others don't
8514 need locks.
8515 [Bodo Moeller]
8516
8517 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
8518 through a BIO pair triggered the default case, i.e.
8519 SSLerr(...,SSL_R_UNKNOWN_STATE).
8520 [Bodo Moeller]
8521
8522 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
8523 can use the SSL library even if none of the specific BIOs is
8524 appropriate.
8525 [Bodo Moeller]
8526
8527 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
8528 for the encoded length.
8529 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
8530
8531 *) Add initial documentation of the X509V3 functions.
8532 [Steve Henson]
8533
8534 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
8535 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
8536 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
8537 secure PKCS#8 private key format with a high iteration count.
8538 [Steve Henson]
8539
8540 *) Fix determination of Perl interpreter: A perl or perl5
8541 _directory_ in $PATH was also accepted as the interpreter.
8542 [Ralf S. Engelschall]
8543
8544 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
8545 wrong with it but it was very old and did things like calling
8546 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
8547 unusual formatting.
8548 [Steve Henson]
8549
8550 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
8551 to use the new extension code.
8552 [Steve Henson]
8553
8554 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
8555 with macros. This should make it easier to change their form, add extra
8556 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
8557 constant.
8558 [Steve Henson]
8559
8560 *) Add to configuration table a new entry that can specify an alternative
8561 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
8562 according to Mark Crispin <MRC@Panda.COM>.
8563 [Bodo Moeller]
8564
8565 #if 0
8566 *) DES CBC did not update the IV. Weird.
8567 [Ben Laurie]
8568 #else
8569 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
8570 Changing the behaviour of the former might break existing programs --
8571 where IV updating is needed, des_ncbc_encrypt can be used.
8572 #endif
8573
8574 *) When bntest is run from "make test" it drives bc to check its
8575 calculations, as well as internally checking them. If an internal check
8576 fails, it needs to cause bc to give a non-zero result or make test carries
8577 on without noticing the failure. Fixed.
8578 [Ben Laurie]
8579
8580 *) DES library cleanups.
8581 [Ulf Möller]
8582
8583 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
8584 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
8585 ciphers. NOTE: although the key derivation function has been verified
8586 against some published test vectors it has not been extensively tested
8587 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
8588 of v2.0.
8589 [Steve Henson]
8590
8591 *) Instead of "mkdir -p", which is not fully portable, use new
8592 Perl script "util/mkdir-p.pl".
8593 [Bodo Moeller]
8594
8595 *) Rewrite the way password based encryption (PBE) is handled. It used to
8596 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
8597 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
8598 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
8599 the 'parameter' field of the AlgorithmIdentifier is passed to the
8600 underlying key generation function so it must do its own ASN1 parsing.
8601 This has also changed the EVP_PBE_CipherInit() function which now has a
8602 'parameter' argument instead of literal salt and iteration count values
8603 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
8604 [Steve Henson]
8605
8606 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
8607 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
8608 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
8609 KEY" because this clashed with PKCS#8 unencrypted string. Since this
8610 value was just used as a "magic string" and not used directly its
8611 value doesn't matter.
8612 [Steve Henson]
8613
8614 *) Introduce some semblance of const correctness to BN. Shame C doesn't
8615 support mutable.
8616 [Ben Laurie]
8617
8618 *) "linux-sparc64" configuration (ultrapenguin).
8619 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
8620 "linux-sparc" configuration.
8621 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
8622
8623 *) config now generates no-xxx options for missing ciphers.
8624 [Ulf Möller]
8625
8626 *) Support the EBCDIC character set (work in progress).
8627 File ebcdic.c not yet included because it has a different license.
8628 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8629
8630 *) Support BS2000/OSD-POSIX.
8631 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
8632
8633 *) Make callbacks for key generation use void * instead of char *.
8634 [Ben Laurie]
8635
8636 *) Make S/MIME samples compile (not yet tested).
8637 [Ben Laurie]
8638
8639 *) Additional typesafe stacks.
8640 [Ben Laurie]
8641
8642 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
8643 [Bodo Moeller]
8644
8645
8646 Changes between 0.9.3 and 0.9.3a [29 May 1999]
8647
8648 *) New configuration variant "sco5-gcc".
8649
8650 *) Updated some demos.
8651 [Sean O Riordain, Wade Scholine]
8652
8653 *) Add missing BIO_free at exit of pkcs12 application.
8654 [Wu Zhigang]
8655
8656 *) Fix memory leak in conf.c.
8657 [Steve Henson]
8658
8659 *) Updates for Win32 to assembler version of MD5.
8660 [Steve Henson]
8661
8662 *) Set #! path to perl in apps/der_chop to where we found it
8663 instead of using a fixed path.
8664 [Bodo Moeller]
8665
8666 *) SHA library changes for irix64-mips4-cc.
8667 [Andy Polyakov]
8668
8669 *) Improvements for VMS support.
8670 [Richard Levitte]
8671
8672
8673 Changes between 0.9.2b and 0.9.3 [24 May 1999]
8674
8675 *) Bignum library bug fix. IRIX 6 passes "make test" now!
8676 This also avoids the problems with SC4.2 and unpatched SC5.
8677 [Andy Polyakov <appro@fy.chalmers.se>]
8678
8679 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
8680 These are required because of the typesafe stack would otherwise break
8681 existing code. If old code used a structure member which used to be STACK
8682 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
8683 sk_num or sk_value it would produce an error because the num, data members
8684 are not present in STACK_OF. Now it just produces a warning. sk_set
8685 replaces the old method of assigning a value to sk_value
8686 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
8687 that does this will no longer work (and should use sk_set instead) but
8688 this could be regarded as a "questionable" behaviour anyway.
8689 [Steve Henson]
8690
8691 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
8692 correctly handle encrypted S/MIME data.
8693 [Steve Henson]
8694
8695 *) Change type of various DES function arguments from des_cblock
8696 (which means, in function argument declarations, pointer to char)
8697 to des_cblock * (meaning pointer to array with 8 char elements),
8698 which allows the compiler to do more typechecking; it was like
8699 that back in SSLeay, but with lots of ugly casts.
8700
8701 Introduce new type const_des_cblock.
8702 [Bodo Moeller]
8703
8704 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
8705 problems: find RecipientInfo structure that matches recipient certificate
8706 and initialise the ASN1 structures properly based on passed cipher.
8707 [Steve Henson]
8708
8709 *) Belatedly make the BN tests actually check the results.
8710 [Ben Laurie]
8711
8712 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
8713 to and from BNs: it was completely broken. New compilation option
8714 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
8715 key elements as negative integers.
8716 [Steve Henson]
8717
8718 *) Reorganize and speed up MD5.
8719 [Andy Polyakov <appro@fy.chalmers.se>]
8720
8721 *) VMS support.
8722 [Richard Levitte <richard@levitte.org>]
8723
8724 *) New option -out to asn1parse to allow the parsed structure to be
8725 output to a file. This is most useful when combined with the -strparse
8726 option to examine the output of things like OCTET STRINGS.
8727 [Steve Henson]
8728
8729 *) Make SSL library a little more fool-proof by not requiring any longer
8730 that SSL_set_{accept,connect}_state be called before
8731 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
8732 in many applications because usually everything *appeared* to work as
8733 intended anyway -- now it really works as intended).
8734 [Bodo Moeller]
8735
8736 *) Move openssl.cnf out of lib/.
8737 [Ulf Möller]
8738
8739 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
8740 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
8741 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
8742 [Ralf S. Engelschall]
8743
8744 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
8745 handle PKCS#7 enveloped data properly.
8746 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
8747
8748 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
8749 copying pointers. The cert_st handling is changed by this in
8750 various ways (and thus what used to be known as ctx->default_cert
8751 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
8752 any longer when s->cert does not give us what we need).
8753 ssl_cert_instantiate becomes obsolete by this change.
8754 As soon as we've got the new code right (possibly it already is?),
8755 we have solved a couple of bugs of the earlier code where s->cert
8756 was used as if it could not have been shared with other SSL structures.
8757
8758 Note that using the SSL API in certain dirty ways now will result
8759 in different behaviour than observed with earlier library versions:
8760 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
8761 does not influence s as it used to.
8762
8763 In order to clean up things more thoroughly, inside SSL_SESSION
8764 we don't use CERT any longer, but a new structure SESS_CERT
8765 that holds per-session data (if available); currently, this is
8766 the peer's certificate chain and, for clients, the server's certificate
8767 and temporary key. CERT holds only those values that can have
8768 meaningful defaults in an SSL_CTX.
8769 [Bodo Moeller]
8770
8771 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
8772 from the internal representation. Various PKCS#7 fixes: remove some
8773 evil casts and set the enc_dig_alg field properly based on the signing
8774 key type.
8775 [Steve Henson]
8776
8777 *) Allow PKCS#12 password to be set from the command line or the
8778 environment. Let 'ca' get its config file name from the environment
8779 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
8780 and 'x509').
8781 [Steve Henson]
8782
8783 *) Allow certificate policies extension to use an IA5STRING for the
8784 organization field. This is contrary to the PKIX definition but
8785 VeriSign uses it and IE5 only recognises this form. Document 'x509'
8786 extension option.
8787 [Steve Henson]
8788
8789 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
8790 without disallowing inline assembler and the like for non-pedantic builds.
8791 [Ben Laurie]
8792
8793 *) Support Borland C++ builder.
8794 [Janez Jere <jj@void.si>, modified by Ulf Möller]
8795
8796 *) Support Mingw32.
8797 [Ulf Möller]
8798
8799 *) SHA-1 cleanups and performance enhancements.
8800 [Andy Polyakov <appro@fy.chalmers.se>]
8801
8802 *) Sparc v8plus assembler for the bignum library.
8803 [Andy Polyakov <appro@fy.chalmers.se>]
8804
8805 *) Accept any -xxx and +xxx compiler options in Configure.
8806 [Ulf Möller]
8807
8808 *) Update HPUX configuration.
8809 [Anonymous]
8810
8811 *) Add missing sk_<type>_unshift() function to safestack.h
8812 [Ralf S. Engelschall]
8813
8814 *) New function SSL_CTX_use_certificate_chain_file that sets the
8815 "extra_cert"s in addition to the certificate. (This makes sense
8816 only for "PEM" format files, as chains as a whole are not
8817 DER-encoded.)
8818 [Bodo Moeller]
8819
8820 *) Support verify_depth from the SSL API.
8821 x509_vfy.c had what can be considered an off-by-one-error:
8822 Its depth (which was not part of the external interface)
8823 was actually counting the number of certificates in a chain;
8824 now it really counts the depth.
8825 [Bodo Moeller]
8826
8827 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
8828 instead of X509err, which often resulted in confusing error
8829 messages since the error codes are not globally unique
8830 (e.g. an alleged error in ssl3_accept when a certificate
8831 didn't match the private key).
8832
8833 *) New function SSL_CTX_set_session_id_context that allows to set a default
8834 value (so that you don't need SSL_set_session_id_context for each
8835 connection using the SSL_CTX).
8836 [Bodo Moeller]
8837
8838 *) OAEP decoding bug fix.
8839 [Ulf Möller]
8840
8841 *) Support INSTALL_PREFIX for package builders, as proposed by
8842 David Harris.
8843 [Bodo Moeller]
8844
8845 *) New Configure options "threads" and "no-threads". For systems
8846 where the proper compiler options are known (currently Solaris
8847 and Linux), "threads" is the default.
8848 [Bodo Moeller]
8849
8850 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
8851 [Bodo Moeller]
8852
8853 *) Install various scripts to $(OPENSSLDIR)/misc, not to
8854 $(INSTALLTOP)/bin -- they shouldn't clutter directories
8855 such as /usr/local/bin.
8856 [Bodo Moeller]
8857
8858 *) "make linux-shared" to build shared libraries.
8859 [Niels Poppe <niels@netbox.org>]
8860
8861 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
8862 [Ulf Möller]
8863
8864 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
8865 extension adding in x509 utility.
8866 [Steve Henson]
8867
8868 *) Remove NOPROTO sections and error code comments.
8869 [Ulf Möller]
8870
8871 *) Partial rewrite of the DEF file generator to now parse the ANSI
8872 prototypes.
8873 [Steve Henson]
8874
8875 *) New Configure options --prefix=DIR and --openssldir=DIR.
8876 [Ulf Möller]
8877
8878 *) Complete rewrite of the error code script(s). It is all now handled
8879 by one script at the top level which handles error code gathering,
8880 header rewriting and C source file generation. It should be much better
8881 than the old method: it now uses a modified version of Ulf's parser to
8882 read the ANSI prototypes in all header files (thus the old K&R definitions
8883 aren't needed for error creation any more) and do a better job of
8884 translating function codes into names. The old 'ASN1 error code imbedded
8885 in a comment' is no longer necessary and it doesn't use .err files which
8886 have now been deleted. Also the error code call doesn't have to appear all
8887 on one line (which resulted in some large lines...).
8888 [Steve Henson]
8889
8890 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
8891 [Bodo Moeller]
8892
8893 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
8894 0 (which usually indicates a closed connection), but continue reading.
8895 [Bodo Moeller]
8896
8897 *) Fix some race conditions.
8898 [Bodo Moeller]
8899
8900 *) Add support for CRL distribution points extension. Add Certificate
8901 Policies and CRL distribution points documentation.
8902 [Steve Henson]
8903
8904 *) Move the autogenerated header file parts to crypto/opensslconf.h.
8905 [Ulf Möller]
8906
8907 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
8908 8 of keying material. Merlin has also confirmed interop with this fix
8909 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
8910 [Merlin Hughes <merlin@baltimore.ie>]
8911
8912 *) Fix lots of warnings.
8913 [Richard Levitte <levitte@stacken.kth.se>]
8914
8915 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
8916 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
8917 [Richard Levitte <levitte@stacken.kth.se>]
8918
8919 *) Fix problems with sizeof(long) == 8.
8920 [Andy Polyakov <appro@fy.chalmers.se>]
8921
8922 *) Change functions to ANSI C.
8923 [Ulf Möller]
8924
8925 *) Fix typos in error codes.
8926 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
8927
8928 *) Remove defunct assembler files from Configure.
8929 [Ulf Möller]
8930
8931 *) SPARC v8 assembler BIGNUM implementation.
8932 [Andy Polyakov <appro@fy.chalmers.se>]
8933
8934 *) Support for Certificate Policies extension: both print and set.
8935 Various additions to support the r2i method this uses.
8936 [Steve Henson]
8937
8938 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
8939 return a const string when you are expecting an allocated buffer.
8940 [Ben Laurie]
8941
8942 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
8943 types DirectoryString and DisplayText.
8944 [Steve Henson]
8945
8946 *) Add code to allow r2i extensions to access the configuration database,
8947 add an LHASH database driver and add several ctx helper functions.
8948 [Steve Henson]
8949
8950 *) Fix an evil bug in bn_expand2() which caused various BN functions to
8951 fail when they extended the size of a BIGNUM.
8952 [Steve Henson]
8953
8954 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8955 support typesafe stack.
8956 [Steve Henson]
8957
8958 *) Fix typo in SSL_[gs]et_options().
8959 [Nils Frostberg <nils@medcom.se>]
8960
8961 *) Delete various functions and files that belonged to the (now obsolete)
8962 old X509V3 handling code.
8963 [Steve Henson]
8964
8965 *) New Configure option "rsaref".
8966 [Ulf Möller]
8967
8968 *) Don't auto-generate pem.h.
8969 [Bodo Moeller]
8970
8971 *) Introduce type-safe ASN.1 SETs.
8972 [Ben Laurie]
8973
8974 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8975 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8976
8977 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8978 that links with OpenSSL (well at least cause lots of warnings), but fear
8979 not: the conversion is trivial, and it eliminates loads of evil casts. A
8980 few STACKed things have been converted already. Feel free to convert more.
8981 In the fullness of time, I'll do away with the STACK type altogether.
8982 [Ben Laurie]
8983
8984 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8985 specified in <certfile> by updating the entry in the index.txt file.
8986 This way one no longer has to edit the index.txt file manually for
8987 revoking a certificate. The -revoke option does the gory details now.
8988 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8989
8990 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8991 `-text' option at all and this way the `-noout -text' combination was
8992 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8993 [Ralf S. Engelschall]
8994
8995 *) Make sure a corresponding plain text error message exists for the
8996 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8997 verify callback function determined that a certificate was revoked.
8998 [Ralf S. Engelschall]
8999
9000 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
9001 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
9002 all available cipers including rc5, which was forgotten until now.
9003 In order to let the testing shell script know which algorithms
9004 are available, a new (up to now undocumented) command
9005 "openssl list-cipher-commands" is used.
9006 [Bodo Moeller]
9007
9008 *) Bugfix: s_client occasionally would sleep in select() when
9009 it should have checked SSL_pending() first.
9010 [Bodo Moeller]
9011
9012 *) New functions DSA_do_sign and DSA_do_verify to provide access to
9013 the raw DSA values prior to ASN.1 encoding.
9014 [Ulf Möller]
9015
9016 *) Tweaks to Configure
9017 [Niels Poppe <niels@netbox.org>]
9018
9019 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
9020 yet...
9021 [Steve Henson]
9022
9023 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
9024 [Ulf Möller]
9025
9026 *) New config option to avoid instructions that are illegal on the 80386.
9027 The default code is faster, but requires at least a 486.
9028 [Ulf Möller]
9029
9030 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
9031 SSL2_SERVER_VERSION (not used at all) macros, which are now the
9032 same as SSL2_VERSION anyway.
9033 [Bodo Moeller]
9034
9035 *) New "-showcerts" option for s_client.
9036 [Bodo Moeller]
9037
9038 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
9039 application. Various cleanups and fixes.
9040 [Steve Henson]
9041
9042 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
9043 modify error routines to work internally. Add error codes and PBE init
9044 to library startup routines.
9045 [Steve Henson]
9046
9047 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
9048 packing functions to asn1 and evp. Changed function names and error
9049 codes along the way.
9050 [Steve Henson]
9051
9052 *) PKCS12 integration: and so it begins... First of several patches to
9053 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
9054 objects to objects.h
9055 [Steve Henson]
9056
9057 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
9058 and display support for Thawte strong extranet extension.
9059 [Steve Henson]
9060
9061 *) Add LinuxPPC support.
9062 [Jeff Dubrule <igor@pobox.org>]
9063
9064 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
9065 bn_div_words in alpha.s.
9066 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
9067
9068 *) Make sure the RSA OAEP test is skipped under -DRSAref because
9069 OAEP isn't supported when OpenSSL is built with RSAref.
9070 [Ulf Moeller <ulf@fitug.de>]
9071
9072 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
9073 so they no longer are missing under -DNOPROTO.
9074 [Soren S. Jorvang <soren@t.dk>]
9075
9076
9077 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
9078
9079 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
9080 doesn't work when the session is reused. Coming soon!
9081 [Ben Laurie]
9082
9083 *) Fix a security hole, that allows sessions to be reused in the wrong
9084 context thus bypassing client cert protection! All software that uses
9085 client certs and session caches in multiple contexts NEEDS PATCHING to
9086 allow session reuse! A fuller solution is in the works.
9087 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
9088
9089 *) Some more source tree cleanups (removed obsolete files
9090 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
9091 permission on "config" script to be executable) and a fix for the INSTALL
9092 document.
9093 [Ulf Moeller <ulf@fitug.de>]
9094
9095 *) Remove some legacy and erroneous uses of malloc, free instead of
9096 Malloc, Free.
9097 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
9098
9099 *) Make rsa_oaep_test return non-zero on error.
9100 [Ulf Moeller <ulf@fitug.de>]
9101
9102 *) Add support for native Solaris shared libraries. Configure
9103 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
9104 if someone would make that last step automatic.
9105 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
9106
9107 *) ctx_size was not built with the right compiler during "make links". Fixed.
9108 [Ben Laurie]
9109
9110 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
9111 except NULL ciphers". This means the default cipher list will no longer
9112 enable NULL ciphers. They need to be specifically enabled e.g. with
9113 the string "DEFAULT:eNULL".
9114 [Steve Henson]
9115
9116 *) Fix to RSA private encryption routines: if p < q then it would
9117 occasionally produce an invalid result. This will only happen with
9118 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
9119 [Steve Henson]
9120
9121 *) Be less restrictive and allow also `perl util/perlpath.pl
9122 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
9123 because this way one can also use an interpreter named `perl5' (which is
9124 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
9125 installed as `perl').
9126 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9127
9128 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
9129 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9130
9131 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
9132 advapi32.lib to Win32 build and change the pem test comparision
9133 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
9134 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
9135 and crypto/des/ede_cbcm_enc.c.
9136 [Steve Henson]
9137
9138 *) DES quad checksum was broken on big-endian architectures. Fixed.
9139 [Ben Laurie]
9140
9141 *) Comment out two functions in bio.h that aren't implemented. Fix up the
9142 Win32 test batch file so it (might) work again. The Win32 test batch file
9143 is horrible: I feel ill....
9144 [Steve Henson]
9145
9146 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
9147 in e_os.h. Audit of header files to check ANSI and non ANSI
9148 sections: 10 functions were absent from non ANSI section and not exported
9149 from Windows DLLs. Fixed up libeay.num for new functions.
9150 [Steve Henson]
9151
9152 *) Make `openssl version' output lines consistent.
9153 [Ralf S. Engelschall]
9154
9155 *) Fix Win32 symbol export lists for BIO functions: Added
9156 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
9157 to ms/libeay{16,32}.def.
9158 [Ralf S. Engelschall]
9159
9160 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
9161 fine under Unix and passes some trivial tests I've now added. But the
9162 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
9163 added to make sure no one expects that this stuff really works in the
9164 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
9165 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
9166 openssl_bio.xs.
9167 [Ralf S. Engelschall]
9168
9169 *) Fix the generation of two part addresses in perl.
9170 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
9171
9172 *) Add config entry for Linux on MIPS.
9173 [John Tobey <jtobey@channel1.com>]
9174
9175 *) Make links whenever Configure is run, unless we are on Windoze.
9176 [Ben Laurie]
9177
9178 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
9179 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
9180 in CRLs.
9181 [Steve Henson]
9182
9183 *) Add a useful kludge to allow package maintainers to specify compiler and
9184 other platforms details on the command line without having to patch the
9185 Configure script everytime: One now can use ``perl Configure
9186 <id>:<details>'', i.e. platform ids are allowed to have details appended
9187 to them (seperated by colons). This is treated as there would be a static
9188 pre-configured entry in Configure's %table under key <id> with value
9189 <details> and ``perl Configure <id>'' is called. So, when you want to
9190 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
9191 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
9192 now, which overrides the FreeBSD-elf entry on-the-fly.
9193 [Ralf S. Engelschall]
9194
9195 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
9196 [Ben Laurie]
9197
9198 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
9199 on the `perl Configure ...' command line. This way one can compile
9200 OpenSSL libraries with Position Independent Code (PIC) which is needed
9201 for linking it into DSOs.
9202 [Ralf S. Engelschall]
9203
9204 *) Remarkably, export ciphers were totally broken and no-one had noticed!
9205 Fixed.
9206 [Ben Laurie]
9207
9208 *) Cleaned up the LICENSE document: The official contact for any license
9209 questions now is the OpenSSL core team under openssl-core@openssl.org.
9210 And add a paragraph about the dual-license situation to make sure people
9211 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
9212 to the OpenSSL toolkit.
9213 [Ralf S. Engelschall]
9214
9215 *) General source tree makefile cleanups: Made `making xxx in yyy...'
9216 display consistent in the source tree and replaced `/bin/rm' by `rm'.
9217 Additonally cleaned up the `make links' target: Remove unnecessary
9218 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
9219 to speed processing and no longer clutter the display with confusing
9220 stuff. Instead only the actually done links are displayed.
9221 [Ralf S. Engelschall]
9222
9223 *) Permit null encryption ciphersuites, used for authentication only. It used
9224 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
9225 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
9226 encryption.
9227 [Ben Laurie]
9228
9229 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
9230 signed attributes when verifying signatures (this would break them),
9231 the detached data encoding was wrong and public keys obtained using
9232 X509_get_pubkey() weren't freed.
9233 [Steve Henson]
9234
9235 *) Add text documentation for the BUFFER functions. Also added a work around
9236 to a Win95 console bug. This was triggered by the password read stuff: the
9237 last character typed gets carried over to the next fread(). If you were
9238 generating a new cert request using 'req' for example then the last
9239 character of the passphrase would be CR which would then enter the first
9240 field as blank.
9241 [Steve Henson]
9242
9243 *) Added the new `Includes OpenSSL Cryptography Software' button as
9244 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
9245 button and can be used by applications based on OpenSSL to show the
9246 relationship to the OpenSSL project.
9247 [Ralf S. Engelschall]
9248
9249 *) Remove confusing variables in function signatures in files
9250 ssl/ssl_lib.c and ssl/ssl.h.
9251 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9252
9253 *) Don't install bss_file.c under PREFIX/include/
9254 [Lennart Bong <lob@kulthea.stacken.kth.se>]
9255
9256 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
9257 functions that return function pointers and has support for NT specific
9258 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
9259 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
9260 unsigned to signed types: this was killing the Win32 compile.
9261 [Steve Henson]
9262
9263 *) Add new certificate file to stack functions,
9264 SSL_add_dir_cert_subjects_to_stack() and
9265 SSL_add_file_cert_subjects_to_stack(). These largely supplant
9266 SSL_load_client_CA_file(), and can be used to add multiple certs easily
9267 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
9268 This means that Apache-SSL and similar packages don't have to mess around
9269 to add as many CAs as they want to the preferred list.
9270 [Ben Laurie]
9271
9272 *) Experiment with doxygen documentation. Currently only partially applied to
9273 ssl/ssl_lib.c.
9274 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
9275 openssl.doxy as the configuration file.
9276 [Ben Laurie]
9277
9278 *) Get rid of remaining C++-style comments which strict C compilers hate.
9279 [Ralf S. Engelschall, pointed out by Carlos Amengual]
9280
9281 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
9282 compiled in by default: it has problems with large keys.
9283 [Steve Henson]
9284
9285 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
9286 DH private keys and/or callback functions which directly correspond to
9287 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
9288 is needed for applications which have to configure certificates on a
9289 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
9290 (e.g. s_server).
9291 For the RSA certificate situation is makes no difference, but
9292 for the DSA certificate situation this fixes the "no shared cipher"
9293 problem where the OpenSSL cipher selection procedure failed because the
9294 temporary keys were not overtaken from the context and the API provided
9295 no way to reconfigure them.
9296 The new functions now let applications reconfigure the stuff and they
9297 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
9298 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
9299 non-public-API function ssl_cert_instantiate() is used as a helper
9300 function and also to reduce code redundancy inside ssl_rsa.c.
9301 [Ralf S. Engelschall]
9302
9303 *) Move s_server -dcert and -dkey options out of the undocumented feature
9304 area because they are useful for the DSA situation and should be
9305 recognized by the users.
9306 [Ralf S. Engelschall]
9307
9308 *) Fix the cipher decision scheme for export ciphers: the export bits are
9309 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
9310 SSL_EXP_MASK. So, the original variable has to be used instead of the
9311 already masked variable.
9312 [Richard Levitte <levitte@stacken.kth.se>]
9313
9314 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
9315 [Richard Levitte <levitte@stacken.kth.se>]
9316
9317 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
9318 from `int' to `unsigned int' because it's a length and initialized by
9319 EVP_DigestFinal() which expects an `unsigned int *'.
9320 [Richard Levitte <levitte@stacken.kth.se>]
9321
9322 *) Don't hard-code path to Perl interpreter on shebang line of Configure
9323 script. Instead use the usual Shell->Perl transition trick.
9324 [Ralf S. Engelschall]
9325
9326 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
9327 (in addition to RSA certificates) to match the behaviour of `openssl dsa
9328 -noout -modulus' as it's already the case for `openssl rsa -noout
9329 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
9330 currently the public key is printed (a decision which was already done by
9331 `openssl dsa -modulus' in the past) which serves a similar purpose.
9332 Additionally the NO_RSA no longer completely removes the whole -modulus
9333 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
9334 now, too.
9335 [Ralf S. Engelschall]
9336
9337 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
9338 BIO. See the source (crypto/evp/bio_ok.c) for more info.
9339 [Arne Ansper <arne@ats.cyber.ee>]
9340
9341 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
9342 to be added. Now both 'req' and 'ca' can use new objects defined in the
9343 config file.
9344 [Steve Henson]
9345
9346 *) Add cool BIO that does syslog (or event log on NT).
9347 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
9348
9349 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
9350 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
9351 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
9352 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
9353 [Ben Laurie]
9354
9355 *) Add preliminary config info for new extension code.
9356 [Steve Henson]
9357
9358 *) Make RSA_NO_PADDING really use no padding.
9359 [Ulf Moeller <ulf@fitug.de>]
9360
9361 *) Generate errors when private/public key check is done.
9362 [Ben Laurie]
9363
9364 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
9365 for some CRL extensions and new objects added.
9366 [Steve Henson]
9367
9368 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
9369 key usage extension and fuller support for authority key id.
9370 [Steve Henson]
9371
9372 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
9373 padding method for RSA, which is recommended for new applications in PKCS
9374 #1 v2.0 (RFC 2437, October 1998).
9375 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
9376 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
9377 against Bleichbacher's attack on RSA.
9378 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
9379 Ben Laurie]
9380
9381 *) Updates to the new SSL compression code
9382 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9383
9384 *) Fix so that the version number in the master secret, when passed
9385 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
9386 (because the server will not accept higher), that the version number
9387 is 0x03,0x01, not 0x03,0x00
9388 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9389
9390 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
9391 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
9392 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
9393 [Steve Henson]
9394
9395 *) Support for RAW extensions where an arbitrary extension can be
9396 created by including its DER encoding. See apps/openssl.cnf for
9397 an example.
9398 [Steve Henson]
9399
9400 *) Make sure latest Perl versions don't interpret some generated C array
9401 code as Perl array code in the crypto/err/err_genc.pl script.
9402 [Lars Weber <3weber@informatik.uni-hamburg.de>]
9403
9404 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
9405 not many people have the assembler. Various Win32 compilation fixes and
9406 update to the INSTALL.W32 file with (hopefully) more accurate Win32
9407 build instructions.
9408 [Steve Henson]
9409
9410 *) Modify configure script 'Configure' to automatically create crypto/date.h
9411 file under Win32 and also build pem.h from pem.org. New script
9412 util/mkfiles.pl to create the MINFO file on environments that can't do a
9413 'make files': perl util/mkfiles.pl >MINFO should work.
9414 [Steve Henson]
9415
9416 *) Major rework of DES function declarations, in the pursuit of correctness
9417 and purity. As a result, many evil casts evaporated, and some weirdness,
9418 too. You may find this causes warnings in your code. Zapping your evil
9419 casts will probably fix them. Mostly.
9420 [Ben Laurie]
9421
9422 *) Fix for a typo in asn1.h. Bug fix to object creation script
9423 obj_dat.pl. It considered a zero in an object definition to mean
9424 "end of object": none of the objects in objects.h have any zeros
9425 so it wasn't spotted.
9426 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
9427
9428 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
9429 Masking (CBCM). In the absence of test vectors, the best I have been able
9430 to do is check that the decrypt undoes the encrypt, so far. Send me test
9431 vectors if you have them.
9432 [Ben Laurie]
9433
9434 *) Correct calculation of key length for export ciphers (too much space was
9435 allocated for null ciphers). This has not been tested!
9436 [Ben Laurie]
9437
9438 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
9439 message is now correct (it understands "crypto" and "ssl" on its
9440 command line). There is also now an "update" option. This will update
9441 the util/ssleay.num and util/libeay.num files with any new functions.
9442 If you do a:
9443 perl util/mkdef.pl crypto ssl update
9444 it will update them.
9445 [Steve Henson]
9446
9447 *) Overhauled the Perl interface (perl/*):
9448 - ported BN stuff to OpenSSL's different BN library
9449 - made the perl/ source tree CVS-aware
9450 - renamed the package from SSLeay to OpenSSL (the files still contain
9451 their history because I've copied them in the repository)
9452 - removed obsolete files (the test scripts will be replaced
9453 by better Test::Harness variants in the future)
9454 [Ralf S. Engelschall]
9455
9456 *) First cut for a very conservative source tree cleanup:
9457 1. merge various obsolete readme texts into doc/ssleay.txt
9458 where we collect the old documents and readme texts.
9459 2. remove the first part of files where I'm already sure that we no
9460 longer need them because of three reasons: either they are just temporary
9461 files which were left by Eric or they are preserved original files where
9462 I've verified that the diff is also available in the CVS via "cvs diff
9463 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
9464 the crypto/md/ stuff).
9465 [Ralf S. Engelschall]
9466
9467 *) More extension code. Incomplete support for subject and issuer alt
9468 name, issuer and authority key id. Change the i2v function parameters
9469 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
9470 what that's for :-) Fix to ASN1 macro which messed up
9471 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
9472 [Steve Henson]
9473
9474 *) Preliminary support for ENUMERATED type. This is largely copied from the
9475 INTEGER code.
9476 [Steve Henson]
9477
9478 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
9479 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9480
9481 *) Make sure `make rehash' target really finds the `openssl' program.
9482 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
9483
9484 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
9485 like to hear about it if this slows down other processors.
9486 [Ben Laurie]
9487
9488 *) Add CygWin32 platform information to Configure script.
9489 [Alan Batie <batie@aahz.jf.intel.com>]
9490
9491 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
9492 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
9493
9494 *) New program nseq to manipulate netscape certificate sequences
9495 [Steve Henson]
9496
9497 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
9498 few typos.
9499 [Steve Henson]
9500
9501 *) Fixes to BN code. Previously the default was to define BN_RECURSION
9502 but the BN code had some problems that would cause failures when
9503 doing certificate verification and some other functions.
9504 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
9505
9506 *) Add ASN1 and PEM code to support netscape certificate sequences.
9507 [Steve Henson]
9508
9509 *) Add ASN1 and PEM code to support netscape certificate sequences.
9510 [Steve Henson]
9511
9512 *) Add several PKIX and private extended key usage OIDs.
9513 [Steve Henson]
9514
9515 *) Modify the 'ca' program to handle the new extension code. Modify
9516 openssl.cnf for new extension format, add comments.
9517 [Steve Henson]
9518
9519 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
9520 and add a sample to openssl.cnf so req -x509 now adds appropriate
9521 CA extensions.
9522 [Steve Henson]
9523
9524 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
9525 error code, add initial support to X509_print() and x509 application.
9526 [Steve Henson]
9527
9528 *) Takes a deep breath and start addding X509 V3 extension support code. Add
9529 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
9530 stuff is currently isolated and isn't even compiled yet.
9531 [Steve Henson]
9532
9533 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
9534 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
9535 Removed the versions check from X509 routines when loading extensions:
9536 this allows certain broken certificates that don't set the version
9537 properly to be processed.
9538 [Steve Henson]
9539
9540 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
9541 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
9542 can still be regenerated with "make depend".
9543 [Ben Laurie]
9544
9545 *) Spelling mistake in C version of CAST-128.
9546 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
9547
9548 *) Changes to the error generation code. The perl script err-code.pl
9549 now reads in the old error codes and retains the old numbers, only
9550 adding new ones if necessary. It also only changes the .err files if new
9551 codes are added. The makefiles have been modified to only insert errors
9552 when needed (to avoid needlessly modifying header files). This is done
9553 by only inserting errors if the .err file is newer than the auto generated
9554 C file. To rebuild all the error codes from scratch (the old behaviour)
9555 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
9556 or delete all the .err files.
9557 [Steve Henson]
9558
9559 *) CAST-128 was incorrectly implemented for short keys. The C version has
9560 been fixed, but is untested. The assembler versions are also fixed, but
9561 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
9562 to regenerate it if needed.
9563 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
9564 Hagino <itojun@kame.net>]
9565
9566 *) File was opened incorrectly in randfile.c.
9567 [Ulf Möller <ulf@fitug.de>]
9568
9569 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
9570 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
9571 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
9572 al: it's just almost always a UTCTime. Note this patch adds new error
9573 codes so do a "make errors" if there are problems.
9574 [Steve Henson]
9575
9576 *) Correct Linux 1 recognition in config.
9577 [Ulf Möller <ulf@fitug.de>]
9578
9579 *) Remove pointless MD5 hash when using DSA keys in ca.
9580 [Anonymous <nobody@replay.com>]
9581
9582 *) Generate an error if given an empty string as a cert directory. Also
9583 generate an error if handed NULL (previously returned 0 to indicate an
9584 error, but didn't set one).
9585 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
9586
9587 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
9588 [Ben Laurie]
9589
9590 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
9591 parameters. This was causing a warning which killed off the Win32 compile.
9592 [Steve Henson]
9593
9594 *) Remove C++ style comments from crypto/bn/bn_local.h.
9595 [Neil Costigan <neil.costigan@celocom.com>]
9596
9597 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
9598 based on a text string, looking up short and long names and finally
9599 "dot" format. The "dot" format stuff didn't work. Added new function
9600 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
9601 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
9602 OID is not part of the table.
9603 [Steve Henson]
9604
9605 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
9606 X509_LOOKUP_by_alias().
9607 [Ben Laurie]
9608
9609 *) Sort openssl functions by name.
9610 [Ben Laurie]
9611
9612 *) Get the gendsa program working (hopefully) and add it to app list. Remove
9613 encryption from sample DSA keys (in case anyone is interested the password
9614 was "1234").
9615 [Steve Henson]
9616
9617 *) Make _all_ *_free functions accept a NULL pointer.
9618 [Frans Heymans <fheymans@isaserver.be>]
9619
9620 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
9621 NULL pointers.
9622 [Anonymous <nobody@replay.com>]
9623
9624 *) s_server should send the CAfile as acceptable CAs, not its own cert.
9625 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9626
9627 *) Don't blow it for numeric -newkey arguments to apps/req.
9628 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
9629
9630 *) Temp key "for export" tests were wrong in s3_srvr.c.
9631 [Anonymous <nobody@replay.com>]
9632
9633 *) Add prototype for temp key callback functions
9634 SSL_CTX_set_tmp_{rsa,dh}_callback().
9635 [Ben Laurie]
9636
9637 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
9638 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
9639 [Steve Henson]
9640
9641 *) X509_name_add_entry() freed the wrong thing after an error.
9642 [Arne Ansper <arne@ats.cyber.ee>]
9643
9644 *) rsa_eay.c would attempt to free a NULL context.
9645 [Arne Ansper <arne@ats.cyber.ee>]
9646
9647 *) BIO_s_socket() had a broken should_retry() on Windoze.
9648 [Arne Ansper <arne@ats.cyber.ee>]
9649
9650 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
9651 [Arne Ansper <arne@ats.cyber.ee>]
9652
9653 *) Make sure the already existing X509_STORE->depth variable is initialized
9654 in X509_STORE_new(), but document the fact that this variable is still
9655 unused in the certificate verification process.
9656 [Ralf S. Engelschall]
9657
9658 *) Fix the various library and apps files to free up pkeys obtained from
9659 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
9660 [Steve Henson]
9661
9662 *) Fix reference counting in X509_PUBKEY_get(). This makes
9663 demos/maurice/example2.c work, amongst others, probably.
9664 [Steve Henson and Ben Laurie]
9665
9666 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
9667 `openssl' and second, the shortcut symlinks for the `openssl <command>'
9668 are no longer created. This way we have a single and consistent command
9669 line interface `openssl <command>', similar to `cvs <command>'.
9670 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
9671
9672 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
9673 BIT STRING wrapper always have zero unused bits.
9674 [Steve Henson]
9675
9676 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
9677 [Steve Henson]
9678
9679 *) Make the top-level INSTALL documentation easier to understand.
9680 [Paul Sutton]
9681
9682 *) Makefiles updated to exit if an error occurs in a sub-directory
9683 make (including if user presses ^C) [Paul Sutton]
9684
9685 *) Make Montgomery context stuff explicit in RSA data structure.
9686 [Ben Laurie]
9687
9688 *) Fix build order of pem and err to allow for generated pem.h.
9689 [Ben Laurie]
9690
9691 *) Fix renumbering bug in X509_NAME_delete_entry().
9692 [Ben Laurie]
9693
9694 *) Enhanced the err-ins.pl script so it makes the error library number
9695 global and can add a library name. This is needed for external ASN1 and
9696 other error libraries.
9697 [Steve Henson]
9698
9699 *) Fixed sk_insert which never worked properly.
9700 [Steve Henson]
9701
9702 *) Fix ASN1 macros so they can handle indefinite length construted
9703 EXPLICIT tags. Some non standard certificates use these: they can now
9704 be read in.
9705 [Steve Henson]
9706
9707 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
9708 into a single doc/ssleay.txt bundle. This way the information is still
9709 preserved but no longer messes up this directory. Now it's new room for
9710 the new set of documenation files.
9711 [Ralf S. Engelschall]
9712
9713 *) SETs were incorrectly DER encoded. This was a major pain, because they
9714 shared code with SEQUENCEs, which aren't coded the same. This means that
9715 almost everything to do with SETs or SEQUENCEs has either changed name or
9716 number of arguments.
9717 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
9718
9719 *) Fix test data to work with the above.
9720 [Ben Laurie]
9721
9722 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
9723 was already fixed by Eric for 0.9.1 it seems.
9724 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
9725
9726 *) Autodetect FreeBSD3.
9727 [Ben Laurie]
9728
9729 *) Fix various bugs in Configure. This affects the following platforms:
9730 nextstep
9731 ncr-scde
9732 unixware-2.0
9733 unixware-2.0-pentium
9734 sco5-cc.
9735 [Ben Laurie]
9736
9737 *) Eliminate generated files from CVS. Reorder tests to regenerate files
9738 before they are needed.
9739 [Ben Laurie]
9740
9741 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
9742 [Ben Laurie]
9743
9744
9745 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
9746
9747 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
9748 changed SSLeay to OpenSSL in version strings.
9749 [Ralf S. Engelschall]
9750
9751 *) Some fixups to the top-level documents.
9752 [Paul Sutton]
9753
9754 *) Fixed the nasty bug where rsaref.h was not found under compile-time
9755 because the symlink to include/ was missing.
9756 [Ralf S. Engelschall]
9757
9758 *) Incorporated the popular no-RSA/DSA-only patches
9759 which allow to compile a RSA-free SSLeay.
9760 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
9761
9762 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
9763 when "ssleay" is still not found.
9764 [Ralf S. Engelschall]
9765
9766 *) Added more platforms to Configure: Cray T3E, HPUX 11,
9767 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
9768
9769 *) Updated the README file.
9770 [Ralf S. Engelschall]
9771
9772 *) Added various .cvsignore files in the CVS repository subdirs
9773 to make a "cvs update" really silent.
9774 [Ralf S. Engelschall]
9775
9776 *) Recompiled the error-definition header files and added
9777 missing symbols to the Win32 linker tables.
9778 [Ralf S. Engelschall]
9779
9780 *) Cleaned up the top-level documents;
9781 o new files: CHANGES and LICENSE
9782 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
9783 o merged COPYRIGHT into LICENSE
9784 o removed obsolete TODO file
9785 o renamed MICROSOFT to INSTALL.W32
9786 [Ralf S. Engelschall]
9787
9788 *) Removed dummy files from the 0.9.1b source tree:
9789 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
9790 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
9791 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
9792 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
9793 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
9794 [Ralf S. Engelschall]
9795
9796 *) Added various platform portability fixes.
9797 [Mark J. Cox]
9798
9799 *) The Genesis of the OpenSSL rpject:
9800 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
9801 Young and Tim J. Hudson created while they were working for C2Net until
9802 summer 1998.
9803 [The OpenSSL Project]
9804
9805
9806 Changes between 0.9.0b and 0.9.1b [not released]
9807
9808 *) Updated a few CA certificates under certs/
9809 [Eric A. Young]
9810
9811 *) Changed some BIGNUM api stuff.
9812 [Eric A. Young]
9813
9814 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
9815 DGUX x86, Linux Alpha, etc.
9816 [Eric A. Young]
9817
9818 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
9819 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
9820 available).
9821 [Eric A. Young]
9822
9823 *) Add -strparse option to asn1pars program which parses nested
9824 binary structures
9825 [Dr Stephen Henson <shenson@bigfoot.com>]
9826
9827 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
9828 [Eric A. Young]
9829
9830 *) DSA fix for "ca" program.
9831 [Eric A. Young]
9832
9833 *) Added "-genkey" option to "dsaparam" program.
9834 [Eric A. Young]
9835
9836 *) Added RIPE MD160 (rmd160) message digest.
9837 [Eric A. Young]
9838
9839 *) Added -a (all) option to "ssleay version" command.
9840 [Eric A. Young]
9841
9842 *) Added PLATFORM define which is the id given to Configure.
9843 [Eric A. Young]
9844
9845 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
9846 [Eric A. Young]
9847
9848 *) Extended the ASN.1 parser routines.
9849 [Eric A. Young]
9850
9851 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
9852 [Eric A. Young]
9853
9854 *) Added a BN_CTX to the BN library.
9855 [Eric A. Young]
9856
9857 *) Fixed the weak key values in DES library
9858 [Eric A. Young]
9859
9860 *) Changed API in EVP library for cipher aliases.
9861 [Eric A. Young]
9862
9863 *) Added support for RC2/64bit cipher.
9864 [Eric A. Young]
9865
9866 *) Converted the lhash library to the crypto/mem.c functions.
9867 [Eric A. Young]
9868
9869 *) Added more recognized ASN.1 object ids.
9870 [Eric A. Young]
9871
9872 *) Added more RSA padding checks for SSL/TLS.
9873 [Eric A. Young]
9874
9875 *) Added BIO proxy/filter functionality.
9876 [Eric A. Young]
9877
9878 *) Added extra_certs to SSL_CTX which can be used
9879 send extra CA certificates to the client in the CA cert chain sending
9880 process. It can be configured with SSL_CTX_add_extra_chain_cert().
9881 [Eric A. Young]
9882
9883 *) Now Fortezza is denied in the authentication phase because
9884 this is key exchange mechanism is not supported by SSLeay at all.
9885 [Eric A. Young]
9886
9887 *) Additional PKCS1 checks.
9888 [Eric A. Young]
9889
9890 *) Support the string "TLSv1" for all TLS v1 ciphers.
9891 [Eric A. Young]
9892
9893 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
9894 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
9895 [Eric A. Young]
9896
9897 *) Fixed a few memory leaks.
9898 [Eric A. Young]
9899
9900 *) Fixed various code and comment typos.
9901 [Eric A. Young]
9902
9903 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
9904 bytes sent in the client random.
9905 [Edward Bishop <ebishop@spyglass.com>]
9906