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