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