]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Remove incorrect code inadvertently introduced through commit 59669b6ab.
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 1.0.2 and 1.1.0 [xx XXX xxxx]
6
7 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
8 done while fixing the error code for the key-too-small case.
9 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
10
11 *) Experimental support for a new, fast, unbiased prime candidate generator,
12 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
13 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
14
15 *) New output format NSS in the sess_id command line tool. This allows
16 exporting the session id and the master key in NSS keylog format.
17 [Martin Kaiser <martin@kaiser.cx>]
18
19 *) Harmonize version and its documentation. -f flag is used to display
20 compilation flags.
21 [mancha <mancha1@zoho.com>]
22
23 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
24 in i2d_ECPrivateKey.
25 [mancha <mancha1@zoho.com>]
26
27 *) Fix some double frees. These are not thought to be exploitable.
28 [mancha <mancha1@zoho.com>]
29
30 *) A missing bounds check in the handling of the TLS heartbeat extension
31 can be used to reveal up to 64k of memory to a connected client or
32 server.
33
34 Thanks for Neel Mehta of Google Security for discovering this bug and to
35 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
36 preparing the fix (CVE-2014-0160)
37 [Adam Langley, Bodo Moeller]
38
39 *) Fix for the attack described in the paper "Recovering OpenSSL
40 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
41 by Yuval Yarom and Naomi Benger. Details can be obtained from:
42 http://eprint.iacr.org/2014/140
43
44 Thanks to Yuval Yarom and Naomi Benger for discovering this
45 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
46 [Yuval Yarom and Naomi Benger]
47
48 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
49 this fixes a limitation in previous versions of OpenSSL.
50 [Steve Henson]
51
52 *) Experimental encrypt-then-mac support.
53
54 Experimental support for encrypt then mac from
55 draft-gutmann-tls-encrypt-then-mac-02.txt
56
57 To enable it set the appropriate extension number (0x42 for the test
58 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
59
60 For non-compliant peers (i.e. just about everything) this should have no
61 effect.
62
63 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
64
65 [Steve Henson]
66
67 *) Add EVP support for key wrapping algorithms, to avoid problems with
68 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
69 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
70 algorithms and include tests cases.
71 [Steve Henson]
72
73 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
74 enveloped data.
75 [Steve Henson]
76
77 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
78 MGF1 digest and OAEP label.
79 [Steve Henson]
80
81 *) Make openssl verify return errors.
82 [Chris Palmer <palmer@google.com> and Ben Laurie]
83
84 *) New function ASN1_TIME_diff to calculate the difference between two
85 ASN1_TIME structures or one structure and the current time.
86 [Steve Henson]
87
88 *) Update fips_test_suite to support multiple command line options. New
89 test to induce all self test errors in sequence and check expected
90 failures.
91 [Steve Henson]
92
93 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
94 sign or verify all in one operation.
95 [Steve Henson]
96
97 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
98 test programs and fips_test_suite. Includes functionality to parse
99 the minimal script output of fipsalgest.pl directly.
100 [Steve Henson]
101
102 *) Add authorisation parameter to FIPS_module_mode_set().
103 [Steve Henson]
104
105 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
106 [Steve Henson]
107
108 *) Use separate DRBG fields for internal and external flags. New function
109 FIPS_drbg_health_check() to perform on demand health checking. Add
110 generation tests to fips_test_suite with reduced health check interval to
111 demonstrate periodic health checking. Add "nodh" option to
112 fips_test_suite to skip very slow DH test.
113 [Steve Henson]
114
115 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
116 based on NID.
117 [Steve Henson]
118
119 *) More extensive health check for DRBG checking many more failure modes.
120 New function FIPS_selftest_drbg_all() to handle every possible DRBG
121 combination: call this in fips_test_suite.
122 [Steve Henson]
123
124 *) Add support for Dual EC DRBG from SP800-90. Update DRBG algorithm test
125 and POST to handle Dual EC cases.
126 [Steve Henson]
127
128 *) Add support for canonical generation of DSA parameter 'g'. See
129 FIPS 186-3 A.2.3.
130
131 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
132 POST to handle HMAC cases.
133 [Steve Henson]
134
135 *) Add functions FIPS_module_version() and FIPS_module_version_text()
136 to return numerical and string versions of the FIPS module number.
137 [Steve Henson]
138
139 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
140 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
141 outside the validated module in the FIPS capable OpenSSL.
142 [Steve Henson]
143
144 *) Minor change to DRBG entropy callback semantics. In some cases
145 there is no multiple of the block length between min_len and
146 max_len. Allow the callback to return more than max_len bytes
147 of entropy but discard any extra: it is the callback's responsibility
148 to ensure that the extra data discarded does not impact the
149 requested amount of entropy.
150 [Steve Henson]
151
152 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
153 information in FIPS186-3, SP800-57 and SP800-131A.
154 [Steve Henson]
155
156 *) CCM support via EVP. Interface is very similar to GCM case except we
157 must supply all data in one chunk (i.e. no update, final) and the
158 message length must be supplied if AAD is used. Add algorithm test
159 support.
160 [Steve Henson]
161
162 *) Initial version of POST overhaul. Add POST callback to allow the status
163 of POST to be monitored and/or failures induced. Modify fips_test_suite
164 to use callback. Always run all selftests even if one fails.
165 [Steve Henson]
166
167 *) XTS support including algorithm test driver in the fips_gcmtest program.
168 Note: this does increase the maximum key length from 32 to 64 bytes but
169 there should be no binary compatibility issues as existing applications
170 will never use XTS mode.
171 [Steve Henson]
172
173 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
174 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
175 performs algorithm blocking for unapproved PRNG types. Also do not
176 set PRNG type in FIPS_mode_set(): leave this to the application.
177 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
178 the standard OpenSSL PRNG: set additional data to a date time vector.
179 [Steve Henson]
180
181 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
182 This shouldn't present any incompatibility problems because applications
183 shouldn't be using these directly and any that are will need to rethink
184 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
185 [Steve Henson]
186
187 *) Extensive self tests and health checking required by SP800-90 DRBG.
188 Remove strength parameter from FIPS_drbg_instantiate and always
189 instantiate at maximum supported strength.
190 [Steve Henson]
191
192 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
193 [Steve Henson]
194
195 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
196 [Steve Henson]
197
198 *) New function DH_compute_key_padded() to compute a DH key and pad with
199 leading zeroes if needed: this complies with SP800-56A et al.
200 [Steve Henson]
201
202 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
203 anything, incomplete, subject to change and largely untested at present.
204 [Steve Henson]
205
206 *) Modify fipscanisteronly build option to only build the necessary object
207 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
208 [Steve Henson]
209
210 *) Add experimental option FIPSSYMS to give all symbols in
211 fipscanister.o and FIPS or fips prefix. This will avoid
212 conflicts with future versions of OpenSSL. Add perl script
213 util/fipsas.pl to preprocess assembly language source files
214 and rename any affected symbols.
215 [Steve Henson]
216
217 *) Add selftest checks and algorithm block of non-fips algorithms in
218 FIPS mode. Remove DES2 from selftests.
219 [Steve Henson]
220
221 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
222 return internal method without any ENGINE dependencies. Add new
223 tiny fips sign and verify functions.
224 [Steve Henson]
225
226 *) New build option no-ec2m to disable characteristic 2 code.
227 [Steve Henson]
228
229 *) New build option "fipscanisteronly". This only builds fipscanister.o
230 and (currently) associated fips utilities. Uses the file Makefile.fips
231 instead of Makefile.org as the prototype.
232 [Steve Henson]
233
234 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
235 Update fips_gcmtest to use IV generator.
236 [Steve Henson]
237
238 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
239 setting output buffer to NULL. The *Final function must be
240 called although it will not retrieve any additional data. The tag
241 can be set or retrieved with a ctrl. The IV length is by default 12
242 bytes (96 bits) but can be set to an alternative value. If the IV
243 length exceeds the maximum IV length (currently 16 bytes) it cannot be
244 set before the key.
245 [Steve Henson]
246
247 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
248 underlying do_cipher function handles all cipher semantics itself
249 including padding and finalisation. This is useful if (for example)
250 an ENGINE cipher handles block padding itself. The behaviour of
251 do_cipher is subtly changed if this flag is set: the return value
252 is the number of characters written to the output buffer (zero is
253 no longer an error code) or a negative error code. Also if the
254 input buffer is NULL and length 0 finalisation should be performed.
255 [Steve Henson]
256
257 *) If a candidate issuer certificate is already part of the constructed
258 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
259 [Steve Henson]
260
261 *) Improve forward-security support: add functions
262
263 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
264 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
265
266 for use by SSL/TLS servers; the callback function will be called whenever a
267 new session is created, and gets to decide whether the session may be
268 cached to make it resumable (return 0) or not (return 1). (As by the
269 SSL/TLS protocol specifications, the session_id sent by the server will be
270 empty to indicate that the session is not resumable; also, the server will
271 not generate RFC 4507 (RFC 5077) session tickets.)
272
273 A simple reasonable callback implementation is to return is_forward_secure.
274 This parameter will be set to 1 or 0 depending on the ciphersuite selected
275 by the SSL/TLS server library, indicating whether it can provide forward
276 security.
277 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
278
279 *) New -verify_name option in command line utilities to set verification
280 parameters by name.
281 [Steve Henson]
282
283 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
284 Add CMAC pkey methods.
285 [Steve Henson]
286
287 *) Experimental renegotiation in s_server -www mode. If the client
288 browses /reneg connection is renegotiated. If /renegcert it is
289 renegotiated requesting a certificate.
290 [Steve Henson]
291
292 *) Add an "external" session cache for debugging purposes to s_server. This
293 should help trace issues which normally are only apparent in deployed
294 multi-process servers.
295 [Steve Henson]
296
297 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
298 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
299 BIO_set_cipher() and some obscure PEM functions were changed so they
300 can now return an error. The RAND changes required a change to the
301 RAND_METHOD structure.
302 [Steve Henson]
303
304 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
305 a gcc attribute to warn if the result of a function is ignored. This
306 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
307 whose return value is often ignored.
308 [Steve Henson]
309
310 Changes between 1.0.1k and 1.0.2 [xx XXX xxxx]
311
312 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
313 (other platforms pending).
314 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
315
316 *) Add support for the SignedCertificateTimestampList certificate and
317 OCSP response extensions from RFC6962.
318 [Rob Stradling]
319
320 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
321 for corner cases. (Certain input points at infinity could lead to
322 bogus results, with non-infinity inputs mapped to infinity too.)
323 [Bodo Moeller]
324
325 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
326 This covers AES, SHA256/512 and GHASH. "Initial" means that most
327 common cases are optimized and there still is room for further
328 improvements. Vector Permutation AES for Altivec is also added.
329 [Andy Polyakov]
330
331 *) Add support for little-endian ppc64 Linux target.
332 [Marcelo Cerri (IBM)]
333
334 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
335 SHA1, SHA256 and GHASH. "Initial" means that most common cases
336 are optimized and there still is room for further improvements.
337 Both 32- and 64-bit modes are supported.
338 [Andy Polyakov, Ard Biesheuvel (Linaro)]
339
340 *) Improved ARMv7 NEON support.
341 [Andy Polyakov]
342
343 *) Support for SPARC Architecture 2011 crypto extensions, first
344 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
345 SHA256/512, MD5, GHASH and modular exponentiation.
346 [Andy Polyakov, David Miller]
347
348 *) Accelerated modular exponentiation for Intel processors, a.k.a.
349 RSAZ.
350 [Shay Gueron & Vlad Krasnov (Intel Corp)]
351
352 *) Support for new and upcoming Intel processors, including AVX2,
353 BMI and SHA ISA extensions. This includes additional "stitched"
354 implementations, AESNI-SHA256 and GCM, and multi-buffer support
355 for TLS encrypt.
356
357 This work was sponsored by Intel Corp.
358 [Andy Polyakov]
359
360 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
361 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
362 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
363 [Steve Henson]
364
365 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
366 this fixes a limiation in previous versions of OpenSSL.
367 [Steve Henson]
368
369 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
370 MGF1 digest and OAEP label.
371 [Steve Henson]
372
373 *) Add EVP support for key wrapping algorithms, to avoid problems with
374 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
375 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
376 algorithms and include tests cases.
377 [Steve Henson]
378
379 *) Add functions to allocate and set the fields of an ECDSA_METHOD
380 structure.
381 [Douglas E. Engert, Steve Henson]
382
383 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
384 difference in days and seconds between two tm or ASN1_TIME structures.
385 [Steve Henson]
386
387 *) Add -rev test option to s_server to just reverse order of characters
388 received by client and send back to server. Also prints an abbreviated
389 summary of the connection parameters.
390 [Steve Henson]
391
392 *) New option -brief for s_client and s_server to print out a brief summary
393 of connection parameters.
394 [Steve Henson]
395
396 *) Add callbacks for arbitrary TLS extensions.
397 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
398
399 *) New option -crl_download in several openssl utilities to download CRLs
400 from CRLDP extension in certificates.
401 [Steve Henson]
402
403 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
404 [Steve Henson]
405
406 *) New function X509_CRL_diff to generate a delta CRL from the difference
407 of two full CRLs. Add support to "crl" utility.
408 [Steve Henson]
409
410 *) New functions to set lookup_crls function and to retrieve
411 X509_STORE from X509_STORE_CTX.
412 [Steve Henson]
413
414 *) Print out deprecated issuer and subject unique ID fields in
415 certificates.
416 [Steve Henson]
417
418 *) Extend OCSP I/O functions so they can be used for simple general purpose
419 HTTP as well as OCSP. New wrapper function which can be used to download
420 CRLs using the OCSP API.
421 [Steve Henson]
422
423 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
424 [Steve Henson]
425
426 *) SSL_CONF* functions. These provide a common framework for application
427 configuration using configuration files or command lines.
428 [Steve Henson]
429
430 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
431 message callback and prints the results. Needs compile time option
432 "enable-ssl-trace". New options to s_client and s_server to enable
433 tracing.
434 [Steve Henson]
435
436 *) New ctrl and macro to retrieve supported points extensions.
437 Print out extension in s_server and s_client.
438 [Steve Henson]
439
440 *) New functions to retrieve certificate signature and signature
441 OID NID.
442 [Steve Henson]
443
444 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
445 client to OpenSSL.
446 [Steve Henson]
447
448 *) New Suite B modes for TLS code. These use and enforce the requirements
449 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
450 only use Suite B curves. The Suite B modes can be set by using the
451 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
452 [Steve Henson]
453
454 *) New chain verification flags for Suite B levels of security. Check
455 algorithms are acceptable when flags are set in X509_verify_cert.
456 [Steve Henson]
457
458 *) Make tls1_check_chain return a set of flags indicating checks passed
459 by a certificate chain. Add additional tests to handle client
460 certificates: checks for matching certificate type and issuer name
461 comparison.
462 [Steve Henson]
463
464 *) If an attempt is made to use a signature algorithm not in the peer
465 preference list abort the handshake. If client has no suitable
466 signature algorithms in response to a certificate request do not
467 use the certificate.
468 [Steve Henson]
469
470 *) If server EC tmp key is not in client preference list abort handshake.
471 [Steve Henson]
472
473 *) Add support for certificate stores in CERT structure. This makes it
474 possible to have different stores per SSL structure or one store in
475 the parent SSL_CTX. Include distint stores for certificate chain
476 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
477 to build and store a certificate chain in CERT structure: returing
478 an error if the chain cannot be built: this will allow applications
479 to test if a chain is correctly configured.
480
481 Note: if the CERT based stores are not set then the parent SSL_CTX
482 store is used to retain compatibility with existing behaviour.
483
484 [Steve Henson]
485
486 *) New function ssl_set_client_disabled to set a ciphersuite disabled
487 mask based on the current session, check mask when sending client
488 hello and checking the requested ciphersuite.
489 [Steve Henson]
490
491 *) New ctrls to retrieve and set certificate types in a certificate
492 request message. Print out received values in s_client. If certificate
493 types is not set with custom values set sensible values based on
494 supported signature algorithms.
495 [Steve Henson]
496
497 *) Support for distinct client and server supported signature algorithms.
498 [Steve Henson]
499
500 *) Add certificate callback. If set this is called whenever a certificate
501 is required by client or server. An application can decide which
502 certificate chain to present based on arbitrary criteria: for example
503 supported signature algorithms. Add very simple example to s_server.
504 This fixes many of the problems and restrictions of the existing client
505 certificate callback: for example you can now clear an existing
506 certificate and specify the whole chain.
507 [Steve Henson]
508
509 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
510 the certificate can be used for (if anything). Set valid_flags field
511 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
512 to have similar checks in it.
513
514 Add new "cert_flags" field to CERT structure and include a "strict mode".
515 This enforces some TLS certificate requirements (such as only permitting
516 certificate signature algorithms contained in the supported algorithms
517 extension) which some implementations ignore: this option should be used
518 with caution as it could cause interoperability issues.
519 [Steve Henson]
520
521 *) Update and tidy signature algorithm extension processing. Work out
522 shared signature algorithms based on preferences and peer algorithms
523 and print them out in s_client and s_server. Abort handshake if no
524 shared signature algorithms.
525 [Steve Henson]
526
527 *) Add new functions to allow customised supported signature algorithms
528 for SSL and SSL_CTX structures. Add options to s_client and s_server
529 to support them.
530 [Steve Henson]
531
532 *) New function SSL_certs_clear() to delete all references to certificates
533 from an SSL structure. Before this once a certificate had been added
534 it couldn't be removed.
535 [Steve Henson]
536
537 *) Integrate hostname, email address and IP address checking with certificate
538 verification. New verify options supporting checking in opensl utility.
539 [Steve Henson]
540
541 *) Fixes and wildcard matching support to hostname and email checking
542 functions. Add manual page.
543 [Florian Weimer (Red Hat Product Security Team)]
544
545 *) New functions to check a hostname email or IP address against a
546 certificate. Add options x509 utility to print results of checks against
547 a certificate.
548 [Steve Henson]
549
550 *) Fix OCSP checking.
551 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
552
553 *) Initial experimental support for explicitly trusted non-root CAs.
554 OpenSSL still tries to build a complete chain to a root but if an
555 intermediate CA has a trust setting included that is used. The first
556 setting is used: whether to trust (e.g., -addtrust option to the x509
557 utility) or reject.
558 [Steve Henson]
559
560 *) Add -trusted_first option which attempts to find certificates in the
561 trusted store even if an untrusted chain is also supplied.
562 [Steve Henson]
563
564 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
565 platform support for Linux and Android.
566 [Andy Polyakov]
567
568 *) Support for linux-x32, ILP32 environment in x86_64 framework.
569 [Andy Polyakov]
570
571 *) Experimental multi-implementation support for FIPS capable OpenSSL.
572 When in FIPS mode the approved implementations are used as normal,
573 when not in FIPS mode the internal unapproved versions are used instead.
574 This means that the FIPS capable OpenSSL isn't forced to use the
575 (often lower performance) FIPS implementations outside FIPS mode.
576 [Steve Henson]
577
578 *) Transparently support X9.42 DH parameters when calling
579 PEM_read_bio_DHparameters. This means existing applications can handle
580 the new parameter format automatically.
581 [Steve Henson]
582
583 *) Initial experimental support for X9.42 DH parameter format: mainly
584 to support use of 'q' parameter for RFC5114 parameters.
585 [Steve Henson]
586
587 *) Add DH parameters from RFC5114 including test data to dhtest.
588 [Steve Henson]
589
590 *) Support for automatic EC temporary key parameter selection. If enabled
591 the most preferred EC parameters are automatically used instead of
592 hardcoded fixed parameters. Now a server just has to call:
593 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
594 support ECDH and use the most appropriate parameters.
595 [Steve Henson]
596
597 *) Enhance and tidy EC curve and point format TLS extension code. Use
598 static structures instead of allocation if default values are used.
599 New ctrls to set curves we wish to support and to retrieve shared curves.
600 Print out shared curves in s_server. New options to s_server and s_client
601 to set list of supported curves.
602 [Steve Henson]
603
604 *) New ctrls to retrieve supported signature algorithms and
605 supported curve values as an array of NIDs. Extend openssl utility
606 to print out received values.
607 [Steve Henson]
608
609 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
610 between NIDs and the more common NIST names such as "P-256". Enhance
611 ecparam utility and ECC method to recognise the NIST names for curves.
612 [Steve Henson]
613
614 *) Enhance SSL/TLS certificate chain handling to support different
615 chains for each certificate instead of one chain in the parent SSL_CTX.
616 [Steve Henson]
617
618 *) Support for fixed DH ciphersuite client authentication: where both
619 server and client use DH certificates with common parameters.
620 [Steve Henson]
621
622 *) Support for fixed DH ciphersuites: those requiring DH server
623 certificates.
624 [Steve Henson]
625
626 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
627 the certificate.
628 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
629 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
630 X509_CINF_get_signature were reverted post internal team review.
631
632 Changes between 1.0.1j and 1.0.1k [xx XXX xxxx]
633
634 *) Do not resume sessions on the server if the negotiated protocol
635 version does not match the session's version. Resuming with a different
636 version, while not strictly forbidden by the RFC, is of questionable
637 sanity and breaks all known clients.
638 [David Benjamin, Emilia Käsper]
639
640 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
641 early CCS messages during renegotiation. (Note that because
642 renegotiation is encrypted, this early CCS was not exploitable.)
643 [Emilia Käsper]
644
645 *) Tighten client-side session ticket handling during renegotiation:
646 ensure that the client only accepts a session ticket if the server sends
647 the extension anew in the ServerHello. Previously, a TLS client would
648 reuse the old extension state and thus accept a session ticket if one was
649 announced in the initial ServerHello.
650
651 Similarly, ensure that the client requires a session ticket if one
652 was advertised in the ServerHello. Previously, a TLS client would
653 ignore a missing NewSessionTicket message.
654 [Emilia Käsper]
655
656 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
657
658 *) SRTP Memory Leak.
659
660 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
661 sends a carefully crafted handshake message, to cause OpenSSL to fail
662 to free up to 64k of memory causing a memory leak. This could be
663 exploited in a Denial Of Service attack. This issue affects OpenSSL
664 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
665 whether SRTP is used or configured. Implementations of OpenSSL that
666 have been compiled with OPENSSL_NO_SRTP defined are not affected.
667
668 The fix was developed by the OpenSSL team.
669 (CVE-2014-3513)
670 [OpenSSL team]
671
672 *) Session Ticket Memory Leak.
673
674 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
675 integrity of that ticket is first verified. In the event of a session
676 ticket integrity check failing, OpenSSL will fail to free memory
677 causing a memory leak. By sending a large number of invalid session
678 tickets an attacker could exploit this issue in a Denial Of Service
679 attack.
680 (CVE-2014-3567)
681 [Steve Henson]
682
683 *) Build option no-ssl3 is incomplete.
684
685 When OpenSSL is configured with "no-ssl3" as a build option, servers
686 could accept and complete a SSL 3.0 handshake, and clients could be
687 configured to send them.
688 (CVE-2014-3568)
689 [Akamai and the OpenSSL team]
690
691 *) Add support for TLS_FALLBACK_SCSV.
692 Client applications doing fallback retries should call
693 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
694 (CVE-2014-3566)
695 [Adam Langley, Bodo Moeller]
696
697 *) Add additional DigestInfo checks.
698
699 Reencode DigestInto in DER and check against the original when
700 verifying RSA signature: this will reject any improperly encoded
701 DigestInfo structures.
702
703 Note: this is a precautionary measure and no attacks are currently known.
704
705 [Steve Henson]
706
707 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
708
709 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
710 SRP code can be overrun an internal buffer. Add sanity check that
711 g, A, B < N to SRP code.
712
713 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
714 Group for discovering this issue.
715 (CVE-2014-3512)
716 [Steve Henson]
717
718 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
719 TLS 1.0 instead of higher protocol versions when the ClientHello message
720 is badly fragmented. This allows a man-in-the-middle attacker to force a
721 downgrade to TLS 1.0 even if both the server and the client support a
722 higher protocol version, by modifying the client's TLS records.
723
724 Thanks to David Benjamin and Adam Langley (Google) for discovering and
725 researching this issue.
726 (CVE-2014-3511)
727 [David Benjamin]
728
729 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
730 to a denial of service attack. A malicious server can crash the client
731 with a null pointer dereference (read) by specifying an anonymous (EC)DH
732 ciphersuite and sending carefully crafted handshake messages.
733
734 Thanks to Felix Gröbert (Google) for discovering and researching this
735 issue.
736 (CVE-2014-3510)
737 [Emilia Käsper]
738
739 *) By sending carefully crafted DTLS packets an attacker could cause openssl
740 to leak memory. This can be exploited through a Denial of Service attack.
741 Thanks to Adam Langley for discovering and researching this issue.
742 (CVE-2014-3507)
743 [Adam Langley]
744
745 *) An attacker can force openssl to consume large amounts of memory whilst
746 processing DTLS handshake messages. This can be exploited through a
747 Denial of Service attack.
748 Thanks to Adam Langley for discovering and researching this issue.
749 (CVE-2014-3506)
750 [Adam Langley]
751
752 *) An attacker can force an error condition which causes openssl to crash
753 whilst processing DTLS packets due to memory being freed twice. This
754 can be exploited through a Denial of Service attack.
755 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
756 this issue.
757 (CVE-2014-3505)
758 [Adam Langley]
759
760 *) If a multithreaded client connects to a malicious server using a resumed
761 session and the server sends an ec point format extension it could write
762 up to 255 bytes to freed memory.
763
764 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
765 issue.
766 (CVE-2014-3509)
767 [Gabor Tyukasz]
768
769 *) A malicious server can crash an OpenSSL client with a null pointer
770 dereference (read) by specifying an SRP ciphersuite even though it was not
771 properly negotiated with the client. This can be exploited through a
772 Denial of Service attack.
773
774 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
775 discovering and researching this issue.
776 (CVE-2014-5139)
777 [Steve Henson]
778
779 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
780 X509_name_oneline, X509_name_print_ex et al. to leak some information
781 from the stack. Applications may be affected if they echo pretty printing
782 output to the attacker.
783
784 Thanks to Ivan Fratric (Google) for discovering this issue.
785 (CVE-2014-3508)
786 [Emilia Käsper, and Steve Henson]
787
788 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
789 for corner cases. (Certain input points at infinity could lead to
790 bogus results, with non-infinity inputs mapped to infinity too.)
791 [Bodo Moeller]
792
793 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
794
795 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
796 handshake can force the use of weak keying material in OpenSSL
797 SSL/TLS clients and servers.
798
799 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
800 researching this issue. (CVE-2014-0224)
801 [KIKUCHI Masashi, Steve Henson]
802
803 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
804 OpenSSL DTLS client the code can be made to recurse eventually crashing
805 in a DoS attack.
806
807 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
808 (CVE-2014-0221)
809 [Imre Rad, Steve Henson]
810
811 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
812 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
813 client or server. This is potentially exploitable to run arbitrary
814 code on a vulnerable client or server.
815
816 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
817 [Jüri Aedla, Steve Henson]
818
819 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
820 are subject to a denial of service attack.
821
822 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
823 this issue. (CVE-2014-3470)
824 [Felix Gröbert, Ivan Fratric, Steve Henson]
825
826 *) Harmonize version and its documentation. -f flag is used to display
827 compilation flags.
828 [mancha <mancha1@zoho.com>]
829
830 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
831 in i2d_ECPrivateKey.
832 [mancha <mancha1@zoho.com>]
833
834 *) Fix some double frees. These are not thought to be exploitable.
835 [mancha <mancha1@zoho.com>]
836
837 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
838
839 *) A missing bounds check in the handling of the TLS heartbeat extension
840 can be used to reveal up to 64k of memory to a connected client or
841 server.
842
843 Thanks for Neel Mehta of Google Security for discovering this bug and to
844 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
845 preparing the fix (CVE-2014-0160)
846 [Adam Langley, Bodo Moeller]
847
848 *) Fix for the attack described in the paper "Recovering OpenSSL
849 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
850 by Yuval Yarom and Naomi Benger. Details can be obtained from:
851 http://eprint.iacr.org/2014/140
852
853 Thanks to Yuval Yarom and Naomi Benger for discovering this
854 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
855 [Yuval Yarom and Naomi Benger]
856
857 *) TLS pad extension: draft-agl-tls-padding-03
858
859 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
860 TLS client Hello record length value would otherwise be > 255 and
861 less that 512 pad with a dummy extension containing zeroes so it
862 is at least 512 bytes long.
863
864 [Adam Langley, Steve Henson]
865
866 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
867
868 *) Fix for TLS record tampering bug. A carefully crafted invalid
869 handshake could crash OpenSSL with a NULL pointer exception.
870 Thanks to Anton Johansson for reporting this issues.
871 (CVE-2013-4353)
872
873 *) Keep original DTLS digest and encryption contexts in retransmission
874 structures so we can use the previous session parameters if they need
875 to be resent. (CVE-2013-6450)
876 [Steve Henson]
877
878 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
879 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
880 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
881 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
882 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
883 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
884 [Rob Stradling, Adam Langley]
885
886 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
887
888 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
889 supporting platforms or when small records were transferred.
890 [Andy Polyakov, Steve Henson]
891
892 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
893
894 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
895
896 This addresses the flaw in CBC record processing discovered by
897 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
898 at: http://www.isg.rhul.ac.uk/tls/
899
900 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
901 Security Group at Royal Holloway, University of London
902 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
903 Emilia Käsper for the initial patch.
904 (CVE-2013-0169)
905 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
906
907 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
908 ciphersuites which can be exploited in a denial of service attack.
909 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
910 and detecting this bug and to Wolfgang Ettlinger
911 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
912 (CVE-2012-2686)
913 [Adam Langley]
914
915 *) Return an error when checking OCSP signatures when key is NULL.
916 This fixes a DoS attack. (CVE-2013-0166)
917 [Steve Henson]
918
919 *) Make openssl verify return errors.
920 [Chris Palmer <palmer@google.com> and Ben Laurie]
921
922 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
923 the right response is stapled. Also change SSL_get_certificate()
924 so it returns the certificate actually sent.
925 See http://rt.openssl.org/Ticket/Display.html?id=2836.
926 [Rob Stradling <rob.stradling@comodo.com>]
927
928 *) Fix possible deadlock when decoding public keys.
929 [Steve Henson]
930
931 *) Don't use TLS 1.0 record version number in initial client hello
932 if renegotiating.
933 [Steve Henson]
934
935 Changes between 1.0.1b and 1.0.1c [10 May 2012]
936
937 *) Sanity check record length before skipping explicit IV in TLS
938 1.2, 1.1 and DTLS to fix DoS attack.
939
940 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
941 fuzzing as a service testing platform.
942 (CVE-2012-2333)
943 [Steve Henson]
944
945 *) Initialise tkeylen properly when encrypting CMS messages.
946 Thanks to Solar Designer of Openwall for reporting this issue.
947 [Steve Henson]
948
949 *) In FIPS mode don't try to use composite ciphers as they are not
950 approved.
951 [Steve Henson]
952
953 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
954
955 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
956 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
957 mean any application compiled against OpenSSL 1.0.0 headers setting
958 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng
959 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
960 0x10000000L Any application which was previously compiled against
961 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
962 will need to be recompiled as a result. Letting be results in
963 inability to disable specifically TLS 1.1 and in client context,
964 in unlike event, limit maximum offered version to TLS 1.0 [see below].
965 [Steve Henson]
966
967 *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not
968 disable just protocol X, but all protocols above X *if* there are
969 protocols *below* X still enabled. In more practical terms it means
970 that if application wants to disable TLS1.0 in favor of TLS1.1 and
971 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
972 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
973 client side.
974 [Andy Polyakov]
975
976 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
977
978 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
979 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
980 in CRYPTO_realloc_clean.
981
982 Thanks to Tavis Ormandy, Google Security Team, for discovering this
983 issue and to Adam Langley <agl@chromium.org> for fixing it.
984 (CVE-2012-2110)
985 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
986
987 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
988 [Adam Langley]
989
990 *) Workarounds for some broken servers that "hang" if a client hello
991 record length exceeds 255 bytes.
992
993 1. Do not use record version number > TLS 1.0 in initial client
994 hello: some (but not all) hanging servers will now work.
995 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
996 the number of ciphers sent in the client hello. This should be
997 set to an even number, such as 50, for example by passing:
998 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
999 Most broken servers should now work.
1000 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
1001 TLS 1.2 client support entirely.
1002 [Steve Henson]
1003
1004 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
1005 [Andy Polyakov]
1006
1007 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
1008
1009 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
1010 STRING form instead of a DigestInfo.
1011 [Steve Henson]
1012
1013 *) The format used for MDC2 RSA signatures is inconsistent between EVP
1014 and the RSA_sign/RSA_verify functions. This was made more apparent when
1015 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
1016 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
1017 the correct format in RSA_verify so both forms transparently work.
1018 [Steve Henson]
1019
1020 *) Some servers which support TLS 1.0 can choke if we initially indicate
1021 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
1022 encrypted premaster secret. As a workaround use the maximum permitted
1023 client version in client hello, this should keep such servers happy
1024 and still work with previous versions of OpenSSL.
1025 [Steve Henson]
1026
1027 *) Add support for TLS/DTLS heartbeats.
1028 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1029
1030 *) Add support for SCTP.
1031 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1032
1033 *) Improved PRNG seeding for VOS.
1034 [Paul Green <Paul.Green@stratus.com>]
1035
1036 *) Extensive assembler packs updates, most notably:
1037
1038 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
1039 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
1040 - x86_64: bit-sliced AES implementation;
1041 - ARM: NEON support, contemporary platforms optimizations;
1042 - s390x: z196 support;
1043 - *: GHASH and GF(2^m) multiplication implementations;
1044
1045 [Andy Polyakov]
1046
1047 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
1048 (removal of unnecessary code)
1049 [Peter Sylvester <peter.sylvester@edelweb.fr>]
1050
1051 *) Add TLS key material exporter from RFC 5705.
1052 [Eric Rescorla]
1053
1054 *) Add DTLS-SRTP negotiation from RFC 5764.
1055 [Eric Rescorla]
1056
1057 *) Add Next Protocol Negotiation,
1058 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
1059 disabled with a no-npn flag to config or Configure. Code donated
1060 by Google.
1061 [Adam Langley <agl@google.com> and Ben Laurie]
1062
1063 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
1064 NIST-P256, NIST-P521, with constant-time single point multiplication on
1065 typical inputs. Compiler support for the nonstandard type __uint128_t is
1066 required to use this (present in gcc 4.4 and later, for 64-bit builds).
1067 Code made available under Apache License version 2.0.
1068
1069 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
1070 line to include this in your build of OpenSSL, and run "make depend" (or
1071 "make update"). This enables the following EC_METHODs:
1072
1073 EC_GFp_nistp224_method()
1074 EC_GFp_nistp256_method()
1075 EC_GFp_nistp521_method()
1076
1077 EC_GROUP_new_by_curve_name() will automatically use these (while
1078 EC_GROUP_new_curve_GFp() currently prefers the more flexible
1079 implementations).
1080 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
1081
1082 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
1083 all platforms. Move ssize_t definition from e_os.h to the public
1084 header file e_os2.h as it now appears in public header file cms.h
1085 [Steve Henson]
1086
1087 *) New -sigopt option to the ca, req and x509 utilities. Additional
1088 signature parameters can be passed using this option and in
1089 particular PSS.
1090 [Steve Henson]
1091
1092 *) Add RSA PSS signing function. This will generate and set the
1093 appropriate AlgorithmIdentifiers for PSS based on those in the
1094 corresponding EVP_MD_CTX structure. No application support yet.
1095 [Steve Henson]
1096
1097 *) Support for companion algorithm specific ASN1 signing routines.
1098 New function ASN1_item_sign_ctx() signs a pre-initialised
1099 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
1100 the appropriate parameters.
1101 [Steve Henson]
1102
1103 *) Add new algorithm specific ASN1 verification initialisation function
1104 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
1105 handling will be the same no matter what EVP_PKEY_METHOD is used.
1106 Add a PSS handler to support verification of PSS signatures: checked
1107 against a number of sample certificates.
1108 [Steve Henson]
1109
1110 *) Add signature printing for PSS. Add PSS OIDs.
1111 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
1112
1113 *) Add algorithm specific signature printing. An individual ASN1 method
1114 can now print out signatures instead of the standard hex dump.
1115
1116 More complex signatures (e.g. PSS) can print out more meaningful
1117 information. Include DSA version that prints out the signature
1118 parameters r, s.
1119 [Steve Henson]
1120
1121 *) Password based recipient info support for CMS library: implementing
1122 RFC3211.
1123 [Steve Henson]
1124
1125 *) Split password based encryption into PBES2 and PBKDF2 functions. This
1126 neatly separates the code into cipher and PBE sections and is required
1127 for some algorithms that split PBES2 into separate pieces (such as
1128 password based CMS).
1129 [Steve Henson]
1130
1131 *) Session-handling fixes:
1132 - Fix handling of connections that are resuming with a session ID,
1133 but also support Session Tickets.
1134 - Fix a bug that suppressed issuing of a new ticket if the client
1135 presented a ticket with an expired session.
1136 - Try to set the ticket lifetime hint to something reasonable.
1137 - Make tickets shorter by excluding irrelevant information.
1138 - On the client side, don't ignore renewed tickets.
1139 [Adam Langley, Bodo Moeller (Google)]
1140
1141 *) Fix PSK session representation.
1142 [Bodo Moeller]
1143
1144 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
1145
1146 This work was sponsored by Intel.
1147 [Andy Polyakov]
1148
1149 *) Add GCM support to TLS library. Some custom code is needed to split
1150 the IV between the fixed (from PRF) and explicit (from TLS record)
1151 portions. This adds all GCM ciphersuites supported by RFC5288 and
1152 RFC5289. Generalise some AES* cipherstrings to include GCM and
1153 add a special AESGCM string for GCM only.
1154 [Steve Henson]
1155
1156 *) Expand range of ctrls for AES GCM. Permit setting invocation
1157 field on decrypt and retrieval of invocation field only on encrypt.
1158 [Steve Henson]
1159
1160 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
1161 As required by RFC5289 these ciphersuites cannot be used if for
1162 versions of TLS earlier than 1.2.
1163 [Steve Henson]
1164
1165 *) For FIPS capable OpenSSL interpret a NULL default public key method
1166 as unset and return the appropriate default but do *not* set the default.
1167 This means we can return the appropriate method in applications that
1168 switch between FIPS and non-FIPS modes.
1169 [Steve Henson]
1170
1171 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
1172 ENGINE is used then we cannot handle that in the FIPS module so we
1173 keep original code iff non-FIPS operations are allowed.
1174 [Steve Henson]
1175
1176 *) Add -attime option to openssl utilities.
1177 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
1178
1179 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
1180 [Steve Henson]
1181
1182 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
1183 FIPS EC methods unconditionally for now.
1184 [Steve Henson]
1185
1186 *) New build option no-ec2m to disable characteristic 2 code.
1187 [Steve Henson]
1188
1189 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
1190 all cases can be covered as some introduce binary incompatibilities.
1191 [Steve Henson]
1192
1193 *) Redirect RSA operations to FIPS module including keygen,
1194 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
1195 [Steve Henson]
1196
1197 *) Add similar low level API blocking to ciphers.
1198 [Steve Henson]
1199
1200 *) Low level digest APIs are not approved in FIPS mode: any attempt
1201 to use these will cause a fatal error. Applications that *really* want
1202 to use them can use the private_* version instead.
1203 [Steve Henson]
1204
1205 *) Redirect cipher operations to FIPS module for FIPS builds.
1206 [Steve Henson]
1207
1208 *) Redirect digest operations to FIPS module for FIPS builds.
1209 [Steve Henson]
1210
1211 *) Update build system to add "fips" flag which will link in fipscanister.o
1212 for static and shared library builds embedding a signature if needed.
1213 [Steve Henson]
1214
1215 *) Output TLS supported curves in preference order instead of numerical
1216 order. This is currently hardcoded for the highest order curves first.
1217 This should be configurable so applications can judge speed vs strength.
1218 [Steve Henson]
1219
1220 *) Add TLS v1.2 server support for client authentication.
1221 [Steve Henson]
1222
1223 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
1224 and enable MD5.
1225 [Steve Henson]
1226
1227 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
1228 FIPS modules versions.
1229 [Steve Henson]
1230
1231 *) Add TLS v1.2 client side support for client authentication. Keep cache
1232 of handshake records longer as we don't know the hash algorithm to use
1233 until after the certificate request message is received.
1234 [Steve Henson]
1235
1236 *) Initial TLS v1.2 client support. Add a default signature algorithms
1237 extension including all the algorithms we support. Parse new signature
1238 format in client key exchange. Relax some ECC signing restrictions for
1239 TLS v1.2 as indicated in RFC5246.
1240 [Steve Henson]
1241
1242 *) Add server support for TLS v1.2 signature algorithms extension. Switch
1243 to new signature format when needed using client digest preference.
1244 All server ciphersuites should now work correctly in TLS v1.2. No client
1245 support yet and no support for client certificates.
1246 [Steve Henson]
1247
1248 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
1249 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
1250 ciphersuites. At present only RSA key exchange ciphersuites work with
1251 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
1252 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
1253 and version checking.
1254 [Steve Henson]
1255
1256 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
1257 with this defined it will not be affected by any changes to ssl internal
1258 structures. Add several utility functions to allow openssl application
1259 to work with OPENSSL_NO_SSL_INTERN defined.
1260 [Steve Henson]
1261
1262 *) Add SRP support.
1263 [Tom Wu <tjw@cs.stanford.edu> and Ben Laurie]
1264
1265 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
1266 [Steve Henson]
1267
1268 *) Permit abbreviated handshakes when renegotiating using the function
1269 SSL_renegotiate_abbreviated().
1270 [Robin Seggelmann <seggelmann@fh-muenster.de>]
1271
1272 *) Add call to ENGINE_register_all_complete() to
1273 ENGINE_load_builtin_engines(), so some implementations get used
1274 automatically instead of needing explicit application support.
1275 [Steve Henson]
1276
1277 *) Add support for TLS key exporter as described in RFC5705.
1278 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
1279
1280 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
1281 a few changes are required:
1282
1283 Add SSL_OP_NO_TLSv1_1 flag.
1284 Add TLSv1_1 methods.
1285 Update version checking logic to handle version 1.1.
1286 Add explicit IV handling (ported from DTLS code).
1287 Add command line options to s_client/s_server.
1288 [Steve Henson]
1289
1290 Changes between 1.0.0j and 1.0.0k [5 Feb 2013]
1291
1292 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
1293
1294 This addresses the flaw in CBC record processing discovered by
1295 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
1296 at: http://www.isg.rhul.ac.uk/tls/
1297
1298 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1299 Security Group at Royal Holloway, University of London
1300 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
1301 Emilia Käsper for the initial patch.
1302 (CVE-2013-0169)
1303 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
1304
1305 *) Return an error when checking OCSP signatures when key is NULL.
1306 This fixes a DoS attack. (CVE-2013-0166)
1307 [Steve Henson]
1308
1309 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
1310 the right response is stapled. Also change SSL_get_certificate()
1311 so it returns the certificate actually sent.
1312 See http://rt.openssl.org/Ticket/Display.html?id=2836.
1313 (This is a backport)
1314 [Rob Stradling <rob.stradling@comodo.com>]
1315
1316 *) Fix possible deadlock when decoding public keys.
1317 [Steve Henson]
1318
1319 Changes between 1.0.0i and 1.0.0j [10 May 2012]
1320
1321 [NB: OpenSSL 1.0.0i and later 1.0.0 patch levels were released after
1322 OpenSSL 1.0.1.]
1323
1324 *) Sanity check record length before skipping explicit IV in DTLS
1325 to fix DoS attack.
1326
1327 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
1328 fuzzing as a service testing platform.
1329 (CVE-2012-2333)
1330 [Steve Henson]
1331
1332 *) Initialise tkeylen properly when encrypting CMS messages.
1333 Thanks to Solar Designer of Openwall for reporting this issue.
1334 [Steve Henson]
1335
1336 Changes between 1.0.0h and 1.0.0i [19 Apr 2012]
1337
1338 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
1339 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
1340 in CRYPTO_realloc_clean.
1341
1342 Thanks to Tavis Ormandy, Google Security Team, for discovering this
1343 issue and to Adam Langley <agl@chromium.org> for fixing it.
1344 (CVE-2012-2110)
1345 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
1346
1347 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
1348
1349 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
1350 in CMS and PKCS7 code. When RSA decryption fails use a random key for
1351 content decryption and always return the same error. Note: this attack
1352 needs on average 2^20 messages so it only affects automated senders. The
1353 old behaviour can be reenabled in the CMS code by setting the
1354 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
1355 an MMA defence is not necessary.
1356 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
1357 this issue. (CVE-2012-0884)
1358 [Steve Henson]
1359
1360 *) Fix CVE-2011-4619: make sure we really are receiving a
1361 client hello before rejecting multiple SGC restarts. Thanks to
1362 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
1363 [Steve Henson]
1364
1365 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
1366
1367 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
1368 Thanks to Antonio Martin, Enterprise Secure Access Research and
1369 Development, Cisco Systems, Inc. for discovering this bug and
1370 preparing a fix. (CVE-2012-0050)
1371 [Antonio Martin]
1372
1373 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
1374
1375 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
1376 of the Vaudenay padding oracle attack on CBC mode encryption
1377 which enables an efficient plaintext recovery attack against
1378 the OpenSSL implementation of DTLS. Their attack exploits timing
1379 differences arising during decryption processing. A research
1380 paper describing this attack can be found at:
1381 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
1382 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
1383 Security Group at Royal Holloway, University of London
1384 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
1385 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
1386 for preparing the fix. (CVE-2011-4108)
1387 [Robin Seggelmann, Michael Tuexen]
1388
1389 *) Clear bytes used for block padding of SSL 3.0 records.
1390 (CVE-2011-4576)
1391 [Adam Langley (Google)]
1392
1393 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
1394 Kadianakis <desnacked@gmail.com> for discovering this issue and
1395 Adam Langley for preparing the fix. (CVE-2011-4619)
1396 [Adam Langley (Google)]
1397
1398 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
1399 [Andrey Kulikov <amdeich@gmail.com>]
1400
1401 *) Prevent malformed RFC3779 data triggering an assertion failure.
1402 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
1403 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
1404 [Rob Austein <sra@hactrn.net>]
1405
1406 *) Improved PRNG seeding for VOS.
1407 [Paul Green <Paul.Green@stratus.com>]
1408
1409 *) Fix ssl_ciph.c set-up race.
1410 [Adam Langley (Google)]
1411
1412 *) Fix spurious failures in ecdsatest.c.
1413 [Emilia Käsper (Google)]
1414
1415 *) Fix the BIO_f_buffer() implementation (which was mixing different
1416 interpretations of the '..._len' fields).
1417 [Adam Langley (Google)]
1418
1419 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
1420 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
1421 threads won't reuse the same blinding coefficients.
1422
1423 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
1424 lock to call BN_BLINDING_invert_ex, and avoids one use of
1425 BN_BLINDING_update for each BN_BLINDING structure (previously,
1426 the last update always remained unused).
1427 [Emilia Käsper (Google)]
1428
1429 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
1430 [Bob Buckholz (Google)]
1431
1432 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
1433
1434 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
1435 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
1436 [Kaspar Brand <ossl@velox.ch>]
1437
1438 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
1439 for multi-threaded use of ECDH. (CVE-2011-3210)
1440 [Adam Langley (Google)]
1441
1442 *) Fix x509_name_ex_d2i memory leak on bad inputs.
1443 [Bodo Moeller]
1444
1445 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
1446 signature public key algorithm by using OID xref utilities instead.
1447 Before this you could only use some ECC ciphersuites with SHA1 only.
1448 [Steve Henson]
1449
1450 *) Add protection against ECDSA timing attacks as mentioned in the paper
1451 by Billy Bob Brumley and Nicola Tuveri, see:
1452
1453 http://eprint.iacr.org/2011/232.pdf
1454
1455 [Billy Bob Brumley and Nicola Tuveri]
1456
1457 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
1458
1459 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
1460 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
1461
1462 *) Fix bug in string printing code: if *any* escaping is enabled we must
1463 escape the escape character (backslash) or the resulting string is
1464 ambiguous.
1465 [Steve Henson]
1466
1467 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
1468
1469 *) Disable code workaround for ancient and obsolete Netscape browsers
1470 and servers: an attacker can use it in a ciphersuite downgrade attack.
1471 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
1472 [Steve Henson]
1473
1474 *) Fixed J-PAKE implementation error, originally discovered by
1475 Sebastien Martini, further info and confirmation from Stefan
1476 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
1477 [Ben Laurie]
1478
1479 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
1480
1481 *) Fix extension code to avoid race conditions which can result in a buffer
1482 overrun vulnerability: resumed sessions must not be modified as they can
1483 be shared by multiple threads. CVE-2010-3864
1484 [Steve Henson]
1485
1486 *) Fix WIN32 build system to correctly link an ENGINE directory into
1487 a DLL.
1488 [Steve Henson]
1489
1490 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
1491
1492 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
1493 (CVE-2010-1633)
1494 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
1495
1496 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
1497
1498 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
1499 context. The operation can be customised via the ctrl mechanism in
1500 case ENGINEs want to include additional functionality.
1501 [Steve Henson]
1502
1503 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
1504 [Steve Henson]
1505
1506 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
1507 output hashes compatible with older versions of OpenSSL.
1508 [Willy Weisz <weisz@vcpc.univie.ac.at>]
1509
1510 *) Fix compression algorithm handling: if resuming a session use the
1511 compression algorithm of the resumed session instead of determining
1512 it from client hello again. Don't allow server to change algorithm.
1513 [Steve Henson]
1514
1515 *) Add load_crls() function to apps tidying load_certs() too. Add option
1516 to verify utility to allow additional CRLs to be included.
1517 [Steve Henson]
1518
1519 *) Update OCSP request code to permit adding custom headers to the request:
1520 some responders need this.
1521 [Steve Henson]
1522
1523 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
1524 correctly.
1525 [Julia Lawall <julia@diku.dk>]
1526
1527 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
1528 needlessly dereferenced structures, used obsolete functions and
1529 didn't handle all updated verify codes correctly.
1530 [Steve Henson]
1531
1532 *) Disable MD2 in the default configuration.
1533 [Steve Henson]
1534
1535 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
1536 indicate the initial BIO being pushed or popped. This makes it possible
1537 to determine whether the BIO is the one explicitly called or as a result
1538 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
1539 it handles reference counts correctly and doesn't zero out the I/O bio
1540 when it is not being explicitly popped. WARNING: applications which
1541 included workarounds for the old buggy behaviour will need to be modified
1542 or they could free up already freed BIOs.
1543 [Steve Henson]
1544
1545 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
1546 renaming to all platforms (within the 0.9.8 branch, this was
1547 done conditionally on Netware platforms to avoid a name clash).
1548 [Guenter <lists@gknw.net>]
1549
1550 *) Add ECDHE and PSK support to DTLS.
1551 [Michael Tuexen <tuexen@fh-muenster.de>]
1552
1553 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
1554 be used on C++.
1555 [Steve Henson]
1556
1557 *) Add "missing" function EVP_MD_flags() (without this the only way to
1558 retrieve a digest flags is by accessing the structure directly. Update
1559 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
1560 or cipher is registered as in the "from" argument. Print out all
1561 registered digests in the dgst usage message instead of manually
1562 attempting to work them out.
1563 [Steve Henson]
1564
1565 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
1566 this allows the use of compression and extensions. Change default cipher
1567 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
1568 by default unless an application cipher string requests it.
1569 [Steve Henson]
1570
1571 *) Alter match criteria in PKCS12_parse(). It used to try to use local
1572 key ids to find matching certificates and keys but some PKCS#12 files
1573 don't follow the (somewhat unwritten) rules and this strategy fails.
1574 Now just gather all certificates together and the first private key
1575 then look for the first certificate that matches the key.
1576 [Steve Henson]
1577
1578 *) Support use of registered digest and cipher names for dgst and cipher
1579 commands instead of having to add each one as a special case. So now
1580 you can do:
1581
1582 openssl sha256 foo
1583
1584 as well as:
1585
1586 openssl dgst -sha256 foo
1587
1588 and this works for ENGINE based algorithms too.
1589
1590 [Steve Henson]
1591
1592 *) Update Gost ENGINE to support parameter files.
1593 [Victor B. Wagner <vitus@cryptocom.ru>]
1594
1595 *) Support GeneralizedTime in ca utility.
1596 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
1597
1598 *) Enhance the hash format used for certificate directory links. The new
1599 form uses the canonical encoding (meaning equivalent names will work
1600 even if they aren't identical) and uses SHA1 instead of MD5. This form
1601 is incompatible with the older format and as a result c_rehash should
1602 be used to rebuild symbolic links.
1603 [Steve Henson]
1604
1605 *) Make PKCS#8 the default write format for private keys, replacing the
1606 traditional format. This form is standardised, more secure and doesn't
1607 include an implicit MD5 dependency.
1608 [Steve Henson]
1609
1610 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
1611 committed to OpenSSL should pass this lot as a minimum.
1612 [Steve Henson]
1613
1614 *) Add session ticket override functionality for use by EAP-FAST.
1615 [Jouni Malinen <j@w1.fi>]
1616
1617 *) Modify HMAC functions to return a value. Since these can be implemented
1618 in an ENGINE errors can occur.
1619 [Steve Henson]
1620
1621 *) Type-checked OBJ_bsearch_ex.
1622 [Ben Laurie]
1623
1624 *) Type-checked OBJ_bsearch. Also some constification necessitated
1625 by type-checking. Still to come: TXT_DB, bsearch(?),
1626 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
1627 CONF_VALUE.
1628 [Ben Laurie]
1629
1630 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
1631 seconds to a tm structure directly, instead of going through OS
1632 specific date routines. This avoids any issues with OS routines such
1633 as the year 2038 bug. New *_adj() functions for ASN1 time structures
1634 and X509_time_adj_ex() to cover the extended range. The existing
1635 X509_time_adj() is still usable and will no longer have any date issues.
1636 [Steve Henson]
1637
1638 *) Delta CRL support. New use deltas option which will attempt to locate
1639 and search any appropriate delta CRLs available.
1640
1641 This work was sponsored by Google.
1642 [Steve Henson]
1643
1644 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
1645 code and add additional score elements. Validate alternate CRL paths
1646 as part of the CRL checking and indicate a new error "CRL path validation
1647 error" in this case. Applications wanting additional details can use
1648 the verify callback and check the new "parent" field. If this is not
1649 NULL CRL path validation is taking place. Existing applications wont
1650 see this because it requires extended CRL support which is off by
1651 default.
1652
1653 This work was sponsored by Google.
1654 [Steve Henson]
1655
1656 *) Support for freshest CRL extension.
1657
1658 This work was sponsored by Google.
1659 [Steve Henson]
1660
1661 *) Initial indirect CRL support. Currently only supported in the CRLs
1662 passed directly and not via lookup. Process certificate issuer
1663 CRL entry extension and lookup CRL entries by bother issuer name
1664 and serial number. Check and process CRL issuer entry in IDP extension.
1665
1666 This work was sponsored by Google.
1667 [Steve Henson]
1668
1669 *) Add support for distinct certificate and CRL paths. The CRL issuer
1670 certificate is validated separately in this case. Only enabled if
1671 an extended CRL support flag is set: this flag will enable additional
1672 CRL functionality in future.
1673
1674 This work was sponsored by Google.
1675 [Steve Henson]
1676
1677 *) Add support for policy mappings extension.
1678
1679 This work was sponsored by Google.
1680 [Steve Henson]
1681
1682 *) Fixes to pathlength constraint, self issued certificate handling,
1683 policy processing to align with RFC3280 and PKITS tests.
1684
1685 This work was sponsored by Google.
1686 [Steve Henson]
1687
1688 *) Support for name constraints certificate extension. DN, email, DNS
1689 and URI types are currently supported.
1690
1691 This work was sponsored by Google.
1692 [Steve Henson]
1693
1694 *) To cater for systems that provide a pointer-based thread ID rather
1695 than numeric, deprecate the current numeric thread ID mechanism and
1696 replace it with a structure and associated callback type. This
1697 mechanism allows a numeric "hash" to be extracted from a thread ID in
1698 either case, and on platforms where pointers are larger than 'long',
1699 mixing is done to help ensure the numeric 'hash' is usable even if it
1700 can't be guaranteed unique. The default mechanism is to use "&errno"
1701 as a pointer-based thread ID to distinguish between threads.
1702
1703 Applications that want to provide their own thread IDs should now use
1704 CRYPTO_THREADID_set_callback() to register a callback that will call
1705 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
1706
1707 Note that ERR_remove_state() is now deprecated, because it is tied
1708 to the assumption that thread IDs are numeric. ERR_remove_state(0)
1709 to free the current thread's error state should be replaced by
1710 ERR_remove_thread_state(NULL).
1711
1712 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
1713 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
1714 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
1715 application was previously providing a numeric thread callback that
1716 was inappropriate for distinguishing threads, then uniqueness might
1717 have been obtained with &errno that happened immediately in the
1718 intermediate development versions of OpenSSL; this is no longer the
1719 case, the numeric thread callback will now override the automatic use
1720 of &errno.)
1721 [Geoff Thorpe, with help from Bodo Moeller]
1722
1723 *) Initial support for different CRL issuing certificates. This covers a
1724 simple case where the self issued certificates in the chain exist and
1725 the real CRL issuer is higher in the existing chain.
1726
1727 This work was sponsored by Google.
1728 [Steve Henson]
1729
1730 *) Removed effectively defunct crypto/store from the build.
1731 [Ben Laurie]
1732
1733 *) Revamp of STACK to provide stronger type-checking. Still to come:
1734 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
1735 ASN1_STRING, CONF_VALUE.
1736 [Ben Laurie]
1737
1738 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
1739 RAM on SSL connections. This option can save about 34k per idle SSL.
1740 [Nick Mathewson]
1741
1742 *) Revamp of LHASH to provide stronger type-checking. Still to come:
1743 STACK, TXT_DB, bsearch, qsort.
1744 [Ben Laurie]
1745
1746 *) Initial support for Cryptographic Message Syntax (aka CMS) based
1747 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
1748 support for data, signedData, compressedData, digestedData and
1749 encryptedData, envelopedData types included. Scripts to check against
1750 RFC4134 examples draft and interop and consistency checks of many
1751 content types and variants.
1752 [Steve Henson]
1753
1754 *) Add options to enc utility to support use of zlib compression BIO.
1755 [Steve Henson]
1756
1757 *) Extend mk1mf to support importing of options and assembly language
1758 files from Configure script, currently only included in VC-WIN32.
1759 The assembly language rules can now optionally generate the source
1760 files from the associated perl scripts.
1761 [Steve Henson]
1762
1763 *) Implement remaining functionality needed to support GOST ciphersuites.
1764 Interop testing has been performed using CryptoPro implementations.
1765 [Victor B. Wagner <vitus@cryptocom.ru>]
1766
1767 *) s390x assembler pack.
1768 [Andy Polyakov]
1769
1770 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
1771 "family."
1772 [Andy Polyakov]
1773
1774 *) Implement Opaque PRF Input TLS extension as specified in
1775 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
1776 official specification yet and no extension type assignment by
1777 IANA exists, this extension (for now) will have to be explicitly
1778 enabled when building OpenSSL by providing the extension number
1779 to use. For example, specify an option
1780
1781 -DTLSEXT_TYPE_opaque_prf_input=0x9527
1782
1783 to the "config" or "Configure" script to enable the extension,
1784 assuming extension number 0x9527 (which is a completely arbitrary
1785 and unofficial assignment based on the MD5 hash of the Internet
1786 Draft). Note that by doing so, you potentially lose
1787 interoperability with other TLS implementations since these might
1788 be using the same extension number for other purposes.
1789
1790 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
1791 opaque PRF input value to use in the handshake. This will create
1792 an interal copy of the length-'len' string at 'src', and will
1793 return non-zero for success.
1794
1795 To get more control and flexibility, provide a callback function
1796 by using
1797
1798 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
1799 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
1800
1801 where
1802
1803 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
1804 void *arg;
1805
1806 Callback function 'cb' will be called in handshakes, and is
1807 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
1808 Argument 'arg' is for application purposes (the value as given to
1809 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
1810 be provided to the callback function). The callback function
1811 has to return non-zero to report success: usually 1 to use opaque
1812 PRF input just if possible, or 2 to enforce use of the opaque PRF
1813 input. In the latter case, the library will abort the handshake
1814 if opaque PRF input is not successfully negotiated.
1815
1816 Arguments 'peerinput' and 'len' given to the callback function
1817 will always be NULL and 0 in the case of a client. A server will
1818 see the client's opaque PRF input through these variables if
1819 available (NULL and 0 otherwise). Note that if the server
1820 provides an opaque PRF input, the length must be the same as the
1821 length of the client's opaque PRF input.
1822
1823 Note that the callback function will only be called when creating
1824 a new session (session resumption can resume whatever was
1825 previously negotiated), and will not be called in SSL 2.0
1826 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
1827 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
1828 for applications that need to enforce opaque PRF input.
1829
1830 [Bodo Moeller]
1831
1832 *) Update ssl code to support digests other than SHA1+MD5 for handshake
1833 MAC.
1834
1835 [Victor B. Wagner <vitus@cryptocom.ru>]
1836
1837 *) Add RFC4507 support to OpenSSL. This includes the corrections in
1838 RFC4507bis. The encrypted ticket format is an encrypted encoded
1839 SSL_SESSION structure, that way new session features are automatically
1840 supported.
1841
1842 If a client application caches session in an SSL_SESSION structure
1843 support is transparent because tickets are now stored in the encoded
1844 SSL_SESSION.
1845
1846 The SSL_CTX structure automatically generates keys for ticket
1847 protection in servers so again support should be possible
1848 with no application modification.
1849
1850 If a client or server wishes to disable RFC4507 support then the option
1851 SSL_OP_NO_TICKET can be set.
1852
1853 Add a TLS extension debugging callback to allow the contents of any client
1854 or server extensions to be examined.
1855
1856 This work was sponsored by Google.
1857 [Steve Henson]
1858
1859 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
1860 OpenSSL should now compile cleanly on gcc 4.2
1861 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
1862
1863 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
1864 support including streaming MAC support: this is required for GOST
1865 ciphersuite support.
1866 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
1867
1868 *) Add option -stream to use PKCS#7 streaming in smime utility. New
1869 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
1870 to output in BER and PEM format.
1871 [Steve Henson]
1872
1873 *) Experimental support for use of HMAC via EVP_PKEY interface. This
1874 allows HMAC to be handled via the EVP_DigestSign*() interface. The
1875 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
1876 ENGINE support for HMAC keys which are unextractable. New -mac and
1877 -macopt options to dgst utility.
1878 [Steve Henson]
1879
1880 *) New option -sigopt to dgst utility. Update dgst to use
1881 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
1882 alternative signing paramaters such as X9.31 or PSS in the dgst
1883 utility.
1884 [Steve Henson]
1885
1886 *) Change ssl_cipher_apply_rule(), the internal function that does
1887 the work each time a ciphersuite string requests enabling
1888 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
1889 removing ("!foo+bar") a class of ciphersuites: Now it maintains
1890 the order of disabled ciphersuites such that those ciphersuites
1891 that most recently went from enabled to disabled not only stay
1892 in order with respect to each other, but also have higher priority
1893 than other disabled ciphersuites the next time ciphersuites are
1894 enabled again.
1895
1896 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
1897 the same ciphersuites as with "HIGH" alone, but in a specific
1898 order where the PSK ciphersuites come first (since they are the
1899 most recently disabled ciphersuites when "HIGH" is parsed).
1900
1901 Also, change ssl_create_cipher_list() (using this new
1902 funcionality) such that between otherwise identical
1903 cihpersuites, ephemeral ECDH is preferred over ephemeral DH in
1904 the default order.
1905 [Bodo Moeller]
1906
1907 *) Change ssl_create_cipher_list() so that it automatically
1908 arranges the ciphersuites in reasonable order before starting
1909 to process the rule string. Thus, the definition for "DEFAULT"
1910 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
1911 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
1912 This makes it much easier to arrive at a reasonable default order
1913 in applications for which anonymous ciphers are OK (meaning
1914 that you can't actually use DEFAULT).
1915 [Bodo Moeller; suggested by Victor Duchovni]
1916
1917 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
1918 processing) into multiple integers instead of setting
1919 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
1920 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
1921 (These masks as well as the individual bit definitions are hidden
1922 away into the non-exported interface ssl/ssl_locl.h, so this
1923 change to the definition of the SSL_CIPHER structure shouldn't
1924 affect applications.) This give us more bits for each of these
1925 categories, so there is no longer a need to coagulate AES128 and
1926 AES256 into a single algorithm bit, and to coagulate Camellia128
1927 and Camellia256 into a single algorithm bit, which has led to all
1928 kinds of kludges.
1929
1930 Thus, among other things, the kludge introduced in 0.9.7m and
1931 0.9.8e for masking out AES256 independently of AES128 or masking
1932 out Camellia256 independently of AES256 is not needed here in 0.9.9.
1933
1934 With the change, we also introduce new ciphersuite aliases that
1935 so far were missing: "AES128", "AES256", "CAMELLIA128", and
1936 "CAMELLIA256".
1937 [Bodo Moeller]
1938
1939 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
1940 Use the leftmost N bytes of the signature input if the input is
1941 larger than the prime q (with N being the size in bytes of q).
1942 [Nils Larsch]
1943
1944 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
1945 it yet and it is largely untested.
1946 [Steve Henson]
1947
1948 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
1949 [Nils Larsch]
1950
1951 *) Initial incomplete changes to avoid need for function casts in OpenSSL
1952 some compilers (gcc 4.2 and later) reject their use. Safestack is
1953 reimplemented. Update ASN1 to avoid use of legacy functions.
1954 [Steve Henson]
1955
1956 *) Win32/64 targets are linked with Winsock2.
1957 [Andy Polyakov]
1958
1959 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
1960 to external functions. This can be used to increase CRL handling
1961 efficiency especially when CRLs are very large by (for example) storing
1962 the CRL revoked certificates in a database.
1963 [Steve Henson]
1964
1965 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
1966 new CRLs added to a directory can be used. New command line option
1967 -verify_return_error to s_client and s_server. This causes real errors
1968 to be returned by the verify callback instead of carrying on no matter
1969 what. This reflects the way a "real world" verify callback would behave.
1970 [Steve Henson]
1971
1972 *) GOST engine, supporting several GOST algorithms and public key formats.
1973 Kindly donated by Cryptocom.
1974 [Cryptocom]
1975
1976 *) Partial support for Issuing Distribution Point CRL extension. CRLs
1977 partitioned by DP are handled but no indirect CRL or reason partitioning
1978 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
1979 selected via a scoring technique which handles IDP and AKID in CRLs.
1980 [Steve Henson]
1981
1982 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
1983 will ultimately be used for all verify operations: this will remove the
1984 X509_STORE dependency on certificate verification and allow alternative
1985 lookup methods. X509_STORE based implementations of these two callbacks.
1986 [Steve Henson]
1987
1988 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
1989 Modify get_crl() to find a valid (unexpired) CRL if possible.
1990 [Steve Henson]
1991
1992 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
1993 this would be called X509_CRL_cmp() but that name is already used by
1994 a function that just compares CRL issuer names. Cache several CRL
1995 extensions in X509_CRL structure and cache CRLDP in X509.
1996 [Steve Henson]
1997
1998 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
1999 this maps equivalent X509_NAME structures into a consistent structure.
2000 Name comparison can then be performed rapidly using memcmp().
2001 [Steve Henson]
2002
2003 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
2004 utility.
2005 [Steve Henson]
2006
2007 *) Allow digests to supply their own micalg string for S/MIME type using
2008 the ctrl EVP_MD_CTRL_MICALG.
2009 [Steve Henson]
2010
2011 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
2012 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
2013 ctrl. It can then customise the structure before and/or after signing
2014 if necessary.
2015 [Steve Henson]
2016
2017 *) New function OBJ_add_sigid() to allow application defined signature OIDs
2018 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
2019 to free up any added signature OIDs.
2020 [Steve Henson]
2021
2022 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
2023 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
2024 digest and cipher tables. New options added to openssl utility:
2025 list-message-digest-algorithms and list-cipher-algorithms.
2026 [Steve Henson]
2027
2028 *) Change the array representation of binary polynomials: the list
2029 of degrees of non-zero coefficients is now terminated with -1.
2030 Previously it was terminated with 0, which was also part of the
2031 value; thus, the array representation was not applicable to
2032 polynomials where t^0 has coefficient zero. This change makes
2033 the array representation useful in a more general context.
2034 [Douglas Stebila]
2035
2036 *) Various modifications and fixes to SSL/TLS cipher string
2037 handling. For ECC, the code now distinguishes between fixed ECDH
2038 with RSA certificates on the one hand and with ECDSA certificates
2039 on the other hand, since these are separate ciphersuites. The
2040 unused code for Fortezza ciphersuites has been removed.
2041
2042 For consistency with EDH, ephemeral ECDH is now called "EECDH"
2043 (not "ECDHE"). For consistency with the code for DH
2044 certificates, use of ECDH certificates is now considered ECDH
2045 authentication, not RSA or ECDSA authentication (the latter is
2046 merely the CA's signing algorithm and not actively used in the
2047 protocol).
2048
2049 The temporary ciphersuite alias "ECCdraft" is no longer
2050 available, and ECC ciphersuites are no longer excluded from "ALL"
2051 and "DEFAULT". The following aliases now exist for RFC 4492
2052 ciphersuites, most of these by analogy with the DH case:
2053
2054 kECDHr - ECDH cert, signed with RSA
2055 kECDHe - ECDH cert, signed with ECDSA
2056 kECDH - ECDH cert (signed with either RSA or ECDSA)
2057 kEECDH - ephemeral ECDH
2058 ECDH - ECDH cert or ephemeral ECDH
2059
2060 aECDH - ECDH cert
2061 aECDSA - ECDSA cert
2062 ECDSA - ECDSA cert
2063
2064 AECDH - anonymous ECDH
2065 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
2066
2067 [Bodo Moeller]
2068
2069 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
2070 Use correct micalg parameters depending on digest(s) in signed message.
2071 [Steve Henson]
2072
2073 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
2074 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
2075 [Steve Henson]
2076
2077 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
2078 an engine to register a method. Add ENGINE lookups for methods and
2079 functional reference processing.
2080 [Steve Henson]
2081
2082 *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of
2083 EVP_{Sign,Verify}* which allow an application to customise the signature
2084 process.
2085 [Steve Henson]
2086
2087 *) New -resign option to smime utility. This adds one or more signers
2088 to an existing PKCS#7 signedData structure. Also -md option to use an
2089 alternative message digest algorithm for signing.
2090 [Steve Henson]
2091
2092 *) Tidy up PKCS#7 routines and add new functions to make it easier to
2093 create PKCS7 structures containing multiple signers. Update smime
2094 application to support multiple signers.
2095 [Steve Henson]
2096
2097 *) New -macalg option to pkcs12 utility to allow setting of an alternative
2098 digest MAC.
2099 [Steve Henson]
2100
2101 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
2102 Reorganize PBE internals to lookup from a static table using NIDs,
2103 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
2104 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
2105 PRF which will be automatically used with PBES2.
2106 [Steve Henson]
2107
2108 *) Replace the algorithm specific calls to generate keys in "req" with the
2109 new API.
2110 [Steve Henson]
2111
2112 *) Update PKCS#7 enveloped data routines to use new API. This is now
2113 supported by any public key method supporting the encrypt operation. A
2114 ctrl is added to allow the public key algorithm to examine or modify
2115 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
2116 a no op.
2117 [Steve Henson]
2118
2119 *) Add a ctrl to asn1 method to allow a public key algorithm to express
2120 a default digest type to use. In most cases this will be SHA1 but some
2121 algorithms (such as GOST) need to specify an alternative digest. The
2122 return value indicates how strong the preference is 1 means optional and
2123 2 is mandatory (that is it is the only supported type). Modify
2124 ASN1_item_sign() to accept a NULL digest argument to indicate it should
2125 use the default md. Update openssl utilities to use the default digest
2126 type for signing if it is not explicitly indicated.
2127 [Steve Henson]
2128
2129 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
2130 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
2131 signing method from the key type. This effectively removes the link
2132 between digests and public key types.
2133 [Steve Henson]
2134
2135 *) Add an OID cross reference table and utility functions. Its purpose is to
2136 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
2137 rsaEncryption. This will allow some of the algorithm specific hackery
2138 needed to use the correct OID to be removed.
2139 [Steve Henson]
2140
2141 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
2142 structures for PKCS7_sign(). They are now set up by the relevant public
2143 key ASN1 method.
2144 [Steve Henson]
2145
2146 *) Add provisional EC pkey method with support for ECDSA and ECDH.
2147 [Steve Henson]
2148
2149 *) Add support for key derivation (agreement) in the API, DH method and
2150 pkeyutl.
2151 [Steve Henson]
2152
2153 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
2154 public and private key formats. As a side effect these add additional
2155 command line functionality not previously available: DSA signatures can be
2156 generated and verified using pkeyutl and DH key support and generation in
2157 pkey, genpkey.
2158 [Steve Henson]
2159
2160 *) BeOS support.
2161 [Oliver Tappe <zooey@hirschkaefer.de>]
2162
2163 *) New make target "install_html_docs" installs HTML renditions of the
2164 manual pages.
2165 [Oliver Tappe <zooey@hirschkaefer.de>]
2166
2167 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
2168 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
2169 support key and parameter generation and add initial key generation
2170 functionality for RSA.
2171 [Steve Henson]
2172
2173 *) Add functions for main EVP_PKEY_method operations. The undocumented
2174 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
2175 EVP_PKEY_{encrypt,decrypt}_old.
2176 [Steve Henson]
2177
2178 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
2179 key API, doesn't do much yet.
2180 [Steve Henson]
2181
2182 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
2183 public key algorithms. New option to openssl utility:
2184 "list-public-key-algorithms" to print out info.
2185 [Steve Henson]
2186
2187 *) Implement the Supported Elliptic Curves Extension for
2188 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2189 [Douglas Stebila]
2190
2191 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
2192 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
2193 [Steve Henson]
2194
2195 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
2196 utilities such as rsa, dsa, dsaparam etc except they process any key
2197 type.
2198 [Steve Henson]
2199
2200 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
2201 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
2202 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
2203 structure.
2204 [Steve Henson]
2205
2206 *) Initial support for pluggable public key ASN1.
2207 De-spaghettify the public key ASN1 handling. Move public and private
2208 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
2209 algorithm specific handling to a single module within the relevant
2210 algorithm directory. Add functions to allow (near) opaque processing
2211 of public and private key structures.
2212 [Steve Henson]
2213
2214 *) Implement the Supported Point Formats Extension for
2215 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
2216 [Douglas Stebila]
2217
2218 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
2219 for the psk identity [hint] and the psk callback functions to the
2220 SSL_SESSION, SSL and SSL_CTX structure.
2221
2222 New ciphersuites:
2223 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
2224 PSK-AES256-CBC-SHA
2225
2226 New functions:
2227 SSL_CTX_use_psk_identity_hint
2228 SSL_get_psk_identity_hint
2229 SSL_get_psk_identity
2230 SSL_use_psk_identity_hint
2231
2232 [Mika Kousa and Pasi Eronen of Nokia Corporation]
2233
2234 *) Add RFC 3161 compliant time stamp request creation, response generation
2235 and response verification functionality.
2236 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
2237
2238 *) Add initial support for TLS extensions, specifically for the server_name
2239 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
2240 have new members for a host name. The SSL data structure has an
2241 additional member SSL_CTX *initial_ctx so that new sessions can be
2242 stored in that context to allow for session resumption, even after the
2243 SSL has been switched to a new SSL_CTX in reaction to a client's
2244 server_name extension.
2245
2246 New functions (subject to change):
2247
2248 SSL_get_servername()
2249 SSL_get_servername_type()
2250 SSL_set_SSL_CTX()
2251
2252 New CTRL codes and macros (subject to change):
2253
2254 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
2255 - SSL_CTX_set_tlsext_servername_callback()
2256 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
2257 - SSL_CTX_set_tlsext_servername_arg()
2258 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
2259
2260 openssl s_client has a new '-servername ...' option.
2261
2262 openssl s_server has new options '-servername_host ...', '-cert2 ...',
2263 '-key2 ...', '-servername_fatal' (subject to change). This allows
2264 testing the HostName extension for a specific single host name ('-cert'
2265 and '-key' remain fallbacks for handshakes without HostName
2266 negotiation). If the unrecognized_name alert has to be sent, this by
2267 default is a warning; it becomes fatal with the '-servername_fatal'
2268 option.
2269
2270 [Peter Sylvester, Remy Allais, Christophe Renou]
2271
2272 *) Whirlpool hash implementation is added.
2273 [Andy Polyakov]
2274
2275 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
2276 bn(64,32). Because of instruction set limitations it doesn't have
2277 any negative impact on performance. This was done mostly in order
2278 to make it possible to share assembler modules, such as bn_mul_mont
2279 implementations, between 32- and 64-bit builds without hassle.
2280 [Andy Polyakov]
2281
2282 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
2283 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
2284 macro.
2285 [Bodo Moeller]
2286
2287 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
2288 dedicated Montgomery multiplication procedure, is introduced.
2289 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
2290 "64-bit" performance on certain 32-bit targets.
2291 [Andy Polyakov]
2292
2293 *) New option SSL_OP_NO_COMP to disable use of compression selectively
2294 in SSL structures. New SSL ctrl to set maximum send fragment size.
2295 Save memory by seeting the I/O buffer sizes dynamically instead of
2296 using the maximum available value.
2297 [Steve Henson]
2298
2299 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
2300 in addition to the text details.
2301 [Bodo Moeller]
2302
2303 *) Very, very preliminary EXPERIMENTAL support for printing of general
2304 ASN1 structures. This currently produces rather ugly output and doesn't
2305 handle several customised structures at all.
2306 [Steve Henson]
2307
2308 *) Integrated support for PVK file format and some related formats such
2309 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
2310 these in the 'rsa' and 'dsa' utilities.
2311 [Steve Henson]
2312
2313 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
2314 [Steve Henson]
2315
2316 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
2317 place for the (very old) "NETSCAPE" format certificates which are now
2318 handled using new ASN1 code equivalents.
2319 [Steve Henson]
2320
2321 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
2322 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
2323 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
2324 [Nils Larsch]
2325
2326 *) Modify CRL distribution points extension code to print out previously
2327 unsupported fields. Enhance extension setting code to allow setting of
2328 all fields.
2329 [Steve Henson]
2330
2331 *) Add print and set support for Issuing Distribution Point CRL extension.
2332 [Steve Henson]
2333
2334 *) Change 'Configure' script to enable Camellia by default.
2335 [NTT]
2336
2337 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
2338
2339 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
2340
2341 This addresses the flaw in CBC record processing discovered by
2342 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
2343 at: http://www.isg.rhul.ac.uk/tls/
2344
2345 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2346 Security Group at Royal Holloway, University of London
2347 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
2348 Emilia Käsper for the initial patch.
2349 (CVE-2013-0169)
2350 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
2351
2352 *) Return an error when checking OCSP signatures when key is NULL.
2353 This fixes a DoS attack. (CVE-2013-0166)
2354 [Steve Henson]
2355
2356 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
2357 the right response is stapled. Also change SSL_get_certificate()
2358 so it returns the certificate actually sent.
2359 See http://rt.openssl.org/Ticket/Display.html?id=2836.
2360 (This is a backport)
2361 [Rob Stradling <rob.stradling@comodo.com>]
2362
2363 *) Fix possible deadlock when decoding public keys.
2364 [Steve Henson]
2365
2366 Changes between 0.9.8w and 0.9.8x [10 May 2012]
2367
2368 *) Sanity check record length before skipping explicit IV in DTLS
2369 to fix DoS attack.
2370
2371 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
2372 fuzzing as a service testing platform.
2373 (CVE-2012-2333)
2374 [Steve Henson]
2375
2376 *) Initialise tkeylen properly when encrypting CMS messages.
2377 Thanks to Solar Designer of Openwall for reporting this issue.
2378 [Steve Henson]
2379
2380 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
2381
2382 *) The fix for CVE-2012-2110 did not take into account that the
2383 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
2384 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
2385 rejecting negative len parameter. (CVE-2012-2131)
2386 [Tomas Hoger <thoger@redhat.com>]
2387
2388 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
2389
2390 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
2391 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
2392 in CRYPTO_realloc_clean.
2393
2394 Thanks to Tavis Ormandy, Google Security Team, for discovering this
2395 issue and to Adam Langley <agl@chromium.org> for fixing it.
2396 (CVE-2012-2110)
2397 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
2398
2399 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
2400
2401 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
2402 in CMS and PKCS7 code. When RSA decryption fails use a random key for
2403 content decryption and always return the same error. Note: this attack
2404 needs on average 2^20 messages so it only affects automated senders. The
2405 old behaviour can be reenabled in the CMS code by setting the
2406 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
2407 an MMA defence is not necessary.
2408 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
2409 this issue. (CVE-2012-0884)
2410 [Steve Henson]
2411
2412 *) Fix CVE-2011-4619: make sure we really are receiving a
2413 client hello before rejecting multiple SGC restarts. Thanks to
2414 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
2415 [Steve Henson]
2416
2417 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
2418
2419 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
2420 Thanks to Antonio Martin, Enterprise Secure Access Research and
2421 Development, Cisco Systems, Inc. for discovering this bug and
2422 preparing a fix. (CVE-2012-0050)
2423 [Antonio Martin]
2424
2425 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
2426
2427 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
2428 of the Vaudenay padding oracle attack on CBC mode encryption
2429 which enables an efficient plaintext recovery attack against
2430 the OpenSSL implementation of DTLS. Their attack exploits timing
2431 differences arising during decryption processing. A research
2432 paper describing this attack can be found at:
2433 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
2434 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
2435 Security Group at Royal Holloway, University of London
2436 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
2437 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
2438 for preparing the fix. (CVE-2011-4108)
2439 [Robin Seggelmann, Michael Tuexen]
2440
2441 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
2442 [Ben Laurie, Kasper <ekasper@google.com>]
2443
2444 *) Clear bytes used for block padding of SSL 3.0 records.
2445 (CVE-2011-4576)
2446 [Adam Langley (Google)]
2447
2448 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
2449 Kadianakis <desnacked@gmail.com> for discovering this issue and
2450 Adam Langley for preparing the fix. (CVE-2011-4619)
2451 [Adam Langley (Google)]
2452
2453 *) Prevent malformed RFC3779 data triggering an assertion failure.
2454 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
2455 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
2456 [Rob Austein <sra@hactrn.net>]
2457
2458 *) Fix ssl_ciph.c set-up race.
2459 [Adam Langley (Google)]
2460
2461 *) Fix spurious failures in ecdsatest.c.
2462 [Emilia Käsper (Google)]
2463
2464 *) Fix the BIO_f_buffer() implementation (which was mixing different
2465 interpretations of the '..._len' fields).
2466 [Adam Langley (Google)]
2467
2468 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
2469 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
2470 threads won't reuse the same blinding coefficients.
2471
2472 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
2473 lock to call BN_BLINDING_invert_ex, and avoids one use of
2474 BN_BLINDING_update for each BN_BLINDING structure (previously,
2475 the last update always remained unused).
2476 [Emilia Käsper (Google)]
2477
2478 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
2479 for multi-threaded use of ECDH.
2480 [Adam Langley (Google)]
2481
2482 *) Fix x509_name_ex_d2i memory leak on bad inputs.
2483 [Bodo Moeller]
2484
2485 *) Add protection against ECDSA timing attacks as mentioned in the paper
2486 by Billy Bob Brumley and Nicola Tuveri, see:
2487
2488 http://eprint.iacr.org/2011/232.pdf
2489
2490 [Billy Bob Brumley and Nicola Tuveri]
2491
2492 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
2493
2494 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
2495 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
2496
2497 *) Fix bug in string printing code: if *any* escaping is enabled we must
2498 escape the escape character (backslash) or the resulting string is
2499 ambiguous.
2500 [Steve Henson]
2501
2502 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
2503
2504 *) Disable code workaround for ancient and obsolete Netscape browsers
2505 and servers: an attacker can use it in a ciphersuite downgrade attack.
2506 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
2507 [Steve Henson]
2508
2509 *) Fixed J-PAKE implementation error, originally discovered by
2510 Sebastien Martini, further info and confirmation from Stefan
2511 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
2512 [Ben Laurie]
2513
2514 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
2515
2516 *) Fix extension code to avoid race conditions which can result in a buffer
2517 overrun vulnerability: resumed sessions must not be modified as they can
2518 be shared by multiple threads. CVE-2010-3864
2519 [Steve Henson]
2520
2521 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
2522 [Steve Henson]
2523
2524 *) Don't reencode certificate when calculating signature: cache and use
2525 the original encoding instead. This makes signature verification of
2526 some broken encodings work correctly.
2527 [Steve Henson]
2528
2529 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
2530 is also one of the inputs.
2531 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2532
2533 *) Don't repeatedly append PBE algorithms to table if they already exist.
2534 Sort table on each new add. This effectively makes the table read only
2535 after all algorithms are added and subsequent calls to PKCS12_pbe_add
2536 etc are non-op.
2537 [Steve Henson]
2538
2539 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
2540
2541 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
2542 OpenSSL 1.0.0.]
2543
2544 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
2545 access or freeing data twice (CVE-2010-0742)
2546 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
2547
2548 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
2549 common in certificates and some applications which only call
2550 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
2551 [Steve Henson]
2552
2553 *) VMS fixes:
2554 Reduce copying into .apps and .test in makevms.com
2555 Don't try to use blank CA certificate in CA.com
2556 Allow use of C files from original directories in maketests.com
2557 [Steven M. Schweda" <sms@antinode.info>]
2558
2559 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
2560
2561 *) When rejecting SSL/TLS records due to an incorrect version number, never
2562 update s->server with a new major version number. As of
2563 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
2564 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
2565 the previous behavior could result in a read attempt at NULL when
2566 receiving specific incorrect SSL/TLS records once record payload
2567 protection is active. (CVE-2010-0740)
2568 [Bodo Moeller, Adam Langley <agl@chromium.org>]
2569
2570 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
2571 could be crashed if the relevant tables were not present (e.g. chrooted).
2572 [Tomas Hoger <thoger@redhat.com>]
2573
2574 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
2575
2576 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
2577 [Martin Olsson, Neel Mehta]
2578
2579 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
2580 accommodate for stack sorting, always a write lock!).
2581 [Bodo Moeller]
2582
2583 *) On some versions of WIN32 Heap32Next is very slow. This can cause
2584 excessive delays in the RAND_poll(): over a minute. As a workaround
2585 include a time check in the inner Heap32Next loop too.
2586 [Steve Henson]
2587
2588 *) The code that handled flushing of data in SSL/TLS originally used the
2589 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
2590 the problem outlined in PR#1949. The fix suggested there however can
2591 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
2592 of Apache). So instead simplify the code to flush unconditionally.
2593 This should be fine since flushing with no data to flush is a no op.
2594 [Steve Henson]
2595
2596 *) Handle TLS versions 2.0 and later properly and correctly use the
2597 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
2598 off ancient servers have a habit of sticking around for a while...
2599 [Steve Henson]
2600
2601 *) Modify compression code so it frees up structures without using the
2602 ex_data callbacks. This works around a problem where some applications
2603 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
2604 restarting) then use compression (e.g. SSL with compression) later.
2605 This results in significant per-connection memory leaks and
2606 has caused some security issues including CVE-2008-1678 and
2607 CVE-2009-4355.
2608 [Steve Henson]
2609
2610 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
2611 change when encrypting or decrypting.
2612 [Bodo Moeller]
2613
2614 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
2615 connect and renegotiate with servers which do not support RI.
2616 Until RI is more widely deployed this option is enabled by default.
2617 [Steve Henson]
2618
2619 *) Add "missing" ssl ctrls to clear options and mode.
2620 [Steve Henson]
2621
2622 *) If client attempts to renegotiate and doesn't support RI respond with
2623 a no_renegotiation alert as required by RFC5746. Some renegotiating
2624 TLS clients will continue a connection gracefully when they receive
2625 the alert. Unfortunately OpenSSL mishandled this alert and would hang
2626 waiting for a server hello which it will never receive. Now we treat a
2627 received no_renegotiation alert as a fatal error. This is because
2628 applications requesting a renegotiation might well expect it to succeed
2629 and would have no code in place to handle the server denying it so the
2630 only safe thing to do is to terminate the connection.
2631 [Steve Henson]
2632
2633 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
2634 peer supports secure renegotiation and 0 otherwise. Print out peer
2635 renegotiation support in s_client/s_server.
2636 [Steve Henson]
2637
2638 *) Replace the highly broken and deprecated SPKAC certification method with
2639 the updated NID creation version. This should correctly handle UTF8.
2640 [Steve Henson]
2641
2642 *) Implement RFC5746. Re-enable renegotiation but require the extension
2643 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
2644 turns out to be a bad idea. It has been replaced by
2645 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
2646 SSL_CTX_set_options(). This is really not recommended unless you
2647 know what you are doing.
2648 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
2649
2650 *) Fixes to stateless session resumption handling. Use initial_ctx when
2651 issuing and attempting to decrypt tickets in case it has changed during
2652 servername handling. Use a non-zero length session ID when attempting
2653 stateless session resumption: this makes it possible to determine if
2654 a resumption has occurred immediately after receiving server hello
2655 (several places in OpenSSL subtly assume this) instead of later in
2656 the handshake.
2657 [Steve Henson]
2658
2659 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
2660 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
2661 fixes for a few places where the return code is not checked
2662 correctly.
2663 [Julia Lawall <julia@diku.dk>]
2664
2665 *) Add --strict-warnings option to Configure script to include devteam
2666 warnings in other configurations.
2667 [Steve Henson]
2668
2669 *) Add support for --libdir option and LIBDIR variable in makefiles. This
2670 makes it possible to install openssl libraries in locations which
2671 have names other than "lib", for example "/usr/lib64" which some
2672 systems need.
2673 [Steve Henson, based on patch from Jeremy Utley]
2674
2675 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
2676 X690 8.9.12 and can produce some misleading textual output of OIDs.
2677 [Steve Henson, reported by Dan Kaminsky]
2678
2679 *) Delete MD2 from algorithm tables. This follows the recommendation in
2680 several standards that it is not used in new applications due to
2681 several cryptographic weaknesses. For binary compatibility reasons
2682 the MD2 API is still compiled in by default.
2683 [Steve Henson]
2684
2685 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
2686 and restored.
2687 [Steve Henson]
2688
2689 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
2690 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
2691 clash.
2692 [Guenter <lists@gknw.net>]
2693
2694 *) Fix the server certificate chain building code to use X509_verify_cert(),
2695 it used to have an ad-hoc builder which was unable to cope with anything
2696 other than a simple chain.
2697 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
2698
2699 *) Don't check self signed certificate signatures in X509_verify_cert()
2700 by default (a flag can override this): it just wastes time without
2701 adding any security. As a useful side effect self signed root CAs
2702 with non-FIPS digests are now usable in FIPS mode.
2703 [Steve Henson]
2704
2705 *) In dtls1_process_out_of_seq_message() the check if the current message
2706 is already buffered was missing. For every new message was memory
2707 allocated, allowing an attacker to perform an denial of service attack
2708 with sending out of seq handshake messages until there is no memory
2709 left. Additionally every future messege was buffered, even if the
2710 sequence number made no sense and would be part of another handshake.
2711 So only messages with sequence numbers less than 10 in advance will be
2712 buffered. (CVE-2009-1378)
2713 [Robin Seggelmann, discovered by Daniel Mentz]
2714
2715 *) Records are buffered if they arrive with a future epoch to be
2716 processed after finishing the corresponding handshake. There is
2717 currently no limitation to this buffer allowing an attacker to perform
2718 a DOS attack with sending records with future epochs until there is no
2719 memory left. This patch adds the pqueue_size() function to determine
2720 the size of a buffer and limits the record buffer to 100 entries.
2721 (CVE-2009-1377)
2722 [Robin Seggelmann, discovered by Daniel Mentz]
2723
2724 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
2725 parent structure is freed. (CVE-2009-1379)
2726 [Daniel Mentz]
2727
2728 *) Handle non-blocking I/O properly in SSL_shutdown() call.
2729 [Darryl Miles <darryl-mailinglists@netbauds.net>]
2730
2731 *) Add 2.5.4.* OIDs
2732 [Ilya O. <vrghost@gmail.com>]
2733
2734 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
2735
2736 *) Disable renegotiation completely - this fixes a severe security
2737 problem (CVE-2009-3555) at the cost of breaking all
2738 renegotiation. Renegotiation can be re-enabled by setting
2739 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
2740 run-time. This is really not recommended unless you know what
2741 you're doing.
2742 [Ben Laurie]
2743
2744 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
2745
2746 *) Don't set val to NULL when freeing up structures, it is freed up by
2747 underlying code. If sizeof(void *) > sizeof(long) this can result in
2748 zeroing past the valid field. (CVE-2009-0789)
2749 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
2750
2751 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
2752 checked correctly. This would allow some invalid signed attributes to
2753 appear to verify correctly. (CVE-2009-0591)
2754 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2755
2756 *) Reject UniversalString and BMPString types with invalid lengths. This
2757 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
2758 a legal length. (CVE-2009-0590)
2759 [Steve Henson]
2760
2761 *) Set S/MIME signing as the default purpose rather than setting it
2762 unconditionally. This allows applications to override it at the store
2763 level.
2764 [Steve Henson]
2765
2766 *) Permit restricted recursion of ASN1 strings. This is needed in practice
2767 to handle some structures.
2768 [Steve Henson]
2769
2770 *) Improve efficiency of mem_gets: don't search whole buffer each time
2771 for a '\n'
2772 [Jeremy Shapiro <jnshapir@us.ibm.com>]
2773
2774 *) New -hex option for openssl rand.
2775 [Matthieu Herrb]
2776
2777 *) Print out UTF8String and NumericString when parsing ASN1.
2778 [Steve Henson]
2779
2780 *) Support NumericString type for name components.
2781 [Steve Henson]
2782
2783 *) Allow CC in the environment to override the automatically chosen
2784 compiler. Note that nothing is done to ensure flags work with the
2785 chosen compiler.
2786 [Ben Laurie]
2787
2788 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
2789
2790 *) Properly check EVP_VerifyFinal() and similar return values
2791 (CVE-2008-5077).
2792 [Ben Laurie, Bodo Moeller, Google Security Team]
2793
2794 *) Enable TLS extensions by default.
2795 [Ben Laurie]
2796
2797 *) Allow the CHIL engine to be loaded, whether the application is
2798 multithreaded or not. (This does not release the developer from the
2799 obligation to set up the dynamic locking callbacks.)
2800 [Sander Temme <sander@temme.net>]
2801
2802 *) Use correct exit code if there is an error in dgst command.
2803 [Steve Henson; problem pointed out by Roland Dirlewanger]
2804
2805 *) Tweak Configure so that you need to say "experimental-jpake" to enable
2806 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
2807 [Bodo Moeller]
2808
2809 *) Add experimental JPAKE support, including demo authentication in
2810 s_client and s_server.
2811 [Ben Laurie]
2812
2813 *) Set the comparison function in v3_addr_canonize().
2814 [Rob Austein <sra@hactrn.net>]
2815
2816 *) Add support for XMPP STARTTLS in s_client.
2817 [Philip Paeps <philip@freebsd.org>]
2818
2819 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
2820 to ensure that even with this option, only ciphersuites in the
2821 server's preference list will be accepted. (Note that the option
2822 applies only when resuming a session, so the earlier behavior was
2823 just about the algorithm choice for symmetric cryptography.)
2824 [Bodo Moeller]
2825
2826 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
2827
2828 *) Fix NULL pointer dereference if a DTLS server received
2829 ChangeCipherSpec as first record (CVE-2009-1386).
2830 [PR #1679]
2831
2832 *) Fix a state transition in s3_srvr.c and d1_srvr.c
2833 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
2834 [Nagendra Modadugu]
2835
2836 *) The fix in 0.9.8c that supposedly got rid of unsafe
2837 double-checked locking was incomplete for RSA blinding,
2838 addressing just one layer of what turns out to have been
2839 doubly unsafe triple-checked locking.
2840
2841 So now fix this for real by retiring the MONT_HELPER macro
2842 in crypto/rsa/rsa_eay.c.
2843
2844 [Bodo Moeller; problem pointed out by Marius Schilder]
2845
2846 *) Various precautionary measures:
2847
2848 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
2849
2850 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
2851 (NB: This would require knowledge of the secret session ticket key
2852 to exploit, in which case you'd be SOL either way.)
2853
2854 - Change bn_nist.c so that it will properly handle input BIGNUMs
2855 outside the expected range.
2856
2857 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
2858 builds.
2859
2860 [Neel Mehta, Bodo Moeller]
2861
2862 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
2863 the load fails. Useful for distros.
2864 [Ben Laurie and the FreeBSD team]
2865
2866 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
2867 [Steve Henson]
2868
2869 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
2870 [Huang Ying]
2871
2872 *) Expand ENGINE to support engine supplied SSL client certificate functions.
2873
2874 This work was sponsored by Logica.
2875 [Steve Henson]
2876
2877 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
2878 keystores. Support for SSL/TLS client authentication too.
2879 Not compiled unless enable-capieng specified to Configure.
2880
2881 This work was sponsored by Logica.
2882 [Steve Henson]
2883
2884 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
2885 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
2886 attribute creation routines such as certifcate requests and PKCS#12
2887 files.
2888 [Steve Henson]
2889
2890 Changes between 0.9.8g and 0.9.8h [28 May 2008]
2891
2892 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
2893 handshake which could lead to a cilent crash as found using the
2894 Codenomicon TLS test suite (CVE-2008-1672)
2895 [Steve Henson, Mark Cox]
2896
2897 *) Fix double free in TLS server name extensions which could lead to
2898 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
2899 [Joe Orton]
2900
2901 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
2902
2903 Clear the error queue to ensure that error entries left from
2904 older function calls do not interfere with the correct operation.
2905 [Lutz Jaenicke, Erik de Castro Lopo]
2906
2907 *) Remove root CA certificates of commercial CAs:
2908
2909 The OpenSSL project does not recommend any specific CA and does not
2910 have any policy with respect to including or excluding any CA.
2911 Therefore it does not make any sense to ship an arbitrary selection
2912 of root CA certificates with the OpenSSL software.
2913 [Lutz Jaenicke]
2914
2915 *) RSA OAEP patches to fix two separate invalid memory reads.
2916 The first one involves inputs when 'lzero' is greater than
2917 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
2918 before the beginning of from). The second one involves inputs where
2919 the 'db' section contains nothing but zeroes (there is a one-byte
2920 invalid read after the end of 'db').
2921 [Ivan Nestlerode <inestlerode@us.ibm.com>]
2922
2923 *) Partial backport from 0.9.9-dev:
2924
2925 Introduce bn_mul_mont (dedicated Montgomery multiplication
2926 procedure) as a candidate for BIGNUM assembler implementation.
2927 While 0.9.9-dev uses assembler for various architectures, only
2928 x86_64 is available by default here in the 0.9.8 branch, and
2929 32-bit x86 is available through a compile-time setting.
2930
2931 To try the 32-bit x86 assembler implementation, use Configure
2932 option "enable-montasm" (which exists only for this backport).
2933
2934 As "enable-montasm" for 32-bit x86 disclaims code stability
2935 anyway, in this constellation we activate additional code
2936 backported from 0.9.9-dev for further performance improvements,
2937 namely BN_from_montgomery_word. (To enable this otherwise,
2938 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
2939
2940 [Andy Polyakov (backport partially by Bodo Moeller)]
2941
2942 *) Add TLS session ticket callback. This allows an application to set
2943 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
2944 values. This is useful for key rollover for example where several key
2945 sets may exist with different names.
2946 [Steve Henson]
2947
2948 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
2949 This was broken until now in 0.9.8 releases, such that the only way
2950 a registered ENGINE could be used (assuming it initialises
2951 successfully on the host) was to explicitly set it as the default
2952 for the relevant algorithms. This is in contradiction with 0.9.7
2953 behaviour and the documentation. With this fix, when an ENGINE is
2954 registered into a given algorithm's table of implementations, the
2955 'uptodate' flag is reset so that auto-discovery will be used next
2956 time a new context for that algorithm attempts to select an
2957 implementation.
2958 [Ian Lister (tweaked by Geoff Thorpe)]
2959
2960 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
2961 implemention in the following ways:
2962
2963 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
2964 hard coded.
2965
2966 Lack of BER streaming support means one pass streaming processing is
2967 only supported if data is detached: setting the streaming flag is
2968 ignored for embedded content.
2969
2970 CMS support is disabled by default and must be explicitly enabled
2971 with the enable-cms configuration option.
2972 [Steve Henson]
2973
2974 *) Update the GMP engine glue to do direct copies between BIGNUM and
2975 mpz_t when openssl and GMP use the same limb size. Otherwise the
2976 existing "conversion via a text string export" trick is still used.
2977 [Paul Sheer <paulsheer@gmail.com>]
2978
2979 *) Zlib compression BIO. This is a filter BIO which compressed and
2980 uncompresses any data passed through it.
2981 [Steve Henson]
2982
2983 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
2984 RFC3394 compatible AES key wrapping.
2985 [Steve Henson]
2986
2987 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
2988 sets string data without copying. X509_ALGOR_set0() and
2989 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
2990 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
2991 from an X509_ATTRIBUTE structure optionally checking it occurs only
2992 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
2993 data.
2994 [Steve Henson]
2995
2996 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
2997 to get the expected BN_FLG_CONSTTIME behavior.
2998 [Bodo Moeller (Google)]
2999
3000 *) Netware support:
3001
3002 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
3003 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
3004 - added some more tests to do_tests.pl
3005 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
3006 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
3007 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
3008 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
3009 - various changes to netware.pl to enable gcc-cross builds on Win32
3010 platform
3011 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
3012 - various changes to fix missing prototype warnings
3013 - fixed x86nasm.pl to create correct asm files for NASM COFF output
3014 - added AES, WHIRLPOOL and CPUID assembler code to build files
3015 - added missing AES assembler make rules to mk1mf.pl
3016 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
3017 [Guenter Knauf <eflash@gmx.net>]
3018
3019 *) Implement certificate status request TLS extension defined in RFC3546.
3020 A client can set the appropriate parameters and receive the encoded
3021 OCSP response via a callback. A server can query the supplied parameters
3022 and set the encoded OCSP response in the callback. Add simplified examples
3023 to s_client and s_server.
3024 [Steve Henson]
3025
3026 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
3027
3028 *) Fix various bugs:
3029 + Binary incompatibility of ssl_ctx_st structure
3030 + DTLS interoperation with non-compliant servers
3031 + Don't call get_session_cb() without proposed session
3032 + Fix ia64 assembler code
3033 [Andy Polyakov, Steve Henson]
3034
3035 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
3036
3037 *) DTLS Handshake overhaul. There were longstanding issues with
3038 OpenSSL DTLS implementation, which were making it impossible for
3039 RFC 4347 compliant client to communicate with OpenSSL server.
3040 Unfortunately just fixing these incompatibilities would "cut off"
3041 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
3042 server keeps tolerating non RFC compliant syntax. The opposite is
3043 not true, 0.9.8f client can not communicate with earlier server.
3044 This update even addresses CVE-2007-4995.
3045 [Andy Polyakov]
3046
3047 *) Changes to avoid need for function casts in OpenSSL: some compilers
3048 (gcc 4.2 and later) reject their use.
3049 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
3050 Steve Henson]
3051
3052 *) Add RFC4507 support to OpenSSL. This includes the corrections in
3053 RFC4507bis. The encrypted ticket format is an encrypted encoded
3054 SSL_SESSION structure, that way new session features are automatically
3055 supported.
3056
3057 If a client application caches session in an SSL_SESSION structure
3058 support is transparent because tickets are now stored in the encoded
3059 SSL_SESSION.
3060
3061 The SSL_CTX structure automatically generates keys for ticket
3062 protection in servers so again support should be possible
3063 with no application modification.
3064
3065 If a client or server wishes to disable RFC4507 support then the option
3066 SSL_OP_NO_TICKET can be set.
3067
3068 Add a TLS extension debugging callback to allow the contents of any client
3069 or server extensions to be examined.
3070
3071 This work was sponsored by Google.
3072 [Steve Henson]
3073
3074 *) Add initial support for TLS extensions, specifically for the server_name
3075 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
3076 have new members for a host name. The SSL data structure has an
3077 additional member SSL_CTX *initial_ctx so that new sessions can be
3078 stored in that context to allow for session resumption, even after the
3079 SSL has been switched to a new SSL_CTX in reaction to a client's
3080 server_name extension.
3081
3082 New functions (subject to change):
3083
3084 SSL_get_servername()
3085 SSL_get_servername_type()
3086 SSL_set_SSL_CTX()
3087
3088 New CTRL codes and macros (subject to change):
3089
3090 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
3091 - SSL_CTX_set_tlsext_servername_callback()
3092 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
3093 - SSL_CTX_set_tlsext_servername_arg()
3094 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
3095
3096 openssl s_client has a new '-servername ...' option.
3097
3098 openssl s_server has new options '-servername_host ...', '-cert2 ...',
3099 '-key2 ...', '-servername_fatal' (subject to change). This allows
3100 testing the HostName extension for a specific single host name ('-cert'
3101 and '-key' remain fallbacks for handshakes without HostName
3102 negotiation). If the unrecognized_name alert has to be sent, this by
3103 default is a warning; it becomes fatal with the '-servername_fatal'
3104 option.
3105
3106 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
3107
3108 *) Add AES and SSE2 assembly language support to VC++ build.
3109 [Steve Henson]
3110
3111 *) Mitigate attack on final subtraction in Montgomery reduction.
3112 [Andy Polyakov]
3113
3114 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
3115 (which previously caused an internal error).
3116 [Bodo Moeller]
3117
3118 *) Squeeze another 10% out of IGE mode when in != out.
3119 [Ben Laurie]
3120
3121 *) AES IGE mode speedup.
3122 [Dean Gaudet (Google)]
3123
3124 *) Add the Korean symmetric 128-bit cipher SEED (see
3125 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
3126 add SEED ciphersuites from RFC 4162:
3127
3128 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
3129 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
3130 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
3131 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
3132
3133 To minimize changes between patchlevels in the OpenSSL 0.9.8
3134 series, SEED remains excluded from compilation unless OpenSSL
3135 is configured with 'enable-seed'.
3136 [KISA, Bodo Moeller]
3137
3138 *) Mitigate branch prediction attacks, which can be practical if a
3139 single processor is shared, allowing a spy process to extract
3140 information. For detailed background information, see
3141 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
3142 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
3143 and Necessary Software Countermeasures"). The core of the change
3144 are new versions BN_div_no_branch() and
3145 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
3146 respectively, which are slower, but avoid the security-relevant
3147 conditional branches. These are automatically called by BN_div()
3148 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
3149 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
3150 remove a conditional branch.
3151
3152 BN_FLG_CONSTTIME is the new name for the previous
3153 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
3154 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
3155 in the exponent causes BN_mod_exp_mont() to use the alternative
3156 implementation in BN_mod_exp_mont_consttime().) The old name
3157 remains as a deprecated alias.
3158
3159 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
3160 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
3161 constant-time implementations for more than just exponentiation.
3162 Here too the old name is kept as a deprecated alias.
3163
3164 BN_BLINDING_new() will now use BN_dup() for the modulus so that
3165 the BN_BLINDING structure gets an independent copy of the
3166 modulus. This means that the previous "BIGNUM *m" argument to
3167 BN_BLINDING_new() and to BN_BLINDING_create_param() now
3168 essentially becomes "const BIGNUM *m", although we can't actually
3169 change this in the header file before 0.9.9. It allows
3170 RSA_setup_blinding() to use BN_with_flags() on the modulus to
3171 enable BN_FLG_CONSTTIME.
3172
3173 [Matthew D Wood (Intel Corp)]
3174
3175 *) In the SSL/TLS server implementation, be strict about session ID
3176 context matching (which matters if an application uses a single
3177 external cache for different purposes). Previously,
3178 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
3179 set. This did ensure strict client verification, but meant that,
3180 with applications using a single external cache for quite
3181 different requirements, clients could circumvent ciphersuite
3182 restrictions for a given session ID context by starting a session
3183 in a different context.
3184 [Bodo Moeller]
3185
3186 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
3187 a ciphersuite string such as "DEFAULT:RSA" cannot enable
3188 authentication-only ciphersuites.
3189 [Bodo Moeller]
3190
3191 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
3192 not complete and could lead to a possible single byte overflow
3193 (CVE-2007-5135) [Ben Laurie]
3194
3195 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
3196
3197 *) Since AES128 and AES256 (and similarly Camellia128 and
3198 Camellia256) share a single mask bit in the logic of
3199 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
3200 kludge to work properly if AES128 is available and AES256 isn't
3201 (or if Camellia128 is available and Camellia256 isn't).
3202 [Victor Duchovni]
3203
3204 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
3205 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
3206 When a point or a seed is encoded in a BIT STRING, we need to
3207 prevent the removal of trailing zero bits to get the proper DER
3208 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
3209 of a NamedBitList, for which trailing 0 bits need to be removed.)
3210 [Bodo Moeller]
3211
3212 *) Have SSL/TLS server implementation tolerate "mismatched" record
3213 protocol version while receiving ClientHello even if the
3214 ClientHello is fragmented. (The server can't insist on the
3215 particular protocol version it has chosen before the ServerHello
3216 message has informed the client about his choice.)
3217 [Bodo Moeller]
3218
3219 *) Add RFC 3779 support.
3220 [Rob Austein for ARIN, Ben Laurie]
3221
3222 *) Load error codes if they are not already present instead of using a
3223 static variable. This allows them to be cleanly unloaded and reloaded.
3224 Improve header file function name parsing.
3225 [Steve Henson]
3226
3227 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
3228 or CAPABILITY handshake as required by RFCs.
3229 [Goetz Babin-Ebell]
3230
3231 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
3232
3233 *) Introduce limits to prevent malicious keys being able to
3234 cause a denial of service. (CVE-2006-2940)
3235 [Steve Henson, Bodo Moeller]
3236
3237 *) Fix ASN.1 parsing of certain invalid structures that can result
3238 in a denial of service. (CVE-2006-2937) [Steve Henson]
3239
3240 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
3241 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
3242
3243 *) Fix SSL client code which could crash if connecting to a
3244 malicious SSLv2 server. (CVE-2006-4343)
3245 [Tavis Ormandy and Will Drewry, Google Security Team]
3246
3247 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
3248 match only those. Before that, "AES256-SHA" would be interpreted
3249 as a pattern and match "AES128-SHA" too (since AES128-SHA got
3250 the same strength classification in 0.9.7h) as we currently only
3251 have a single AES bit in the ciphersuite description bitmap.
3252 That change, however, also applied to ciphersuite strings such as
3253 "RC4-MD5" that intentionally matched multiple ciphersuites --
3254 namely, SSL 2.0 ciphersuites in addition to the more common ones
3255 from SSL 3.0/TLS 1.0.
3256
3257 So we change the selection algorithm again: Naming an explicit
3258 ciphersuite selects this one ciphersuite, and any other similar
3259 ciphersuite (same bitmap) from *other* protocol versions.
3260 Thus, "RC4-MD5" again will properly select both the SSL 2.0
3261 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
3262
3263 Since SSL 2.0 does not have any ciphersuites for which the
3264 128/256 bit distinction would be relevant, this works for now.
3265 The proper fix will be to use different bits for AES128 and
3266 AES256, which would have avoided the problems from the beginning;
3267 however, bits are scarce, so we can only do this in a new release
3268 (not just a patchlevel) when we can change the SSL_CIPHER
3269 definition to split the single 'unsigned long mask' bitmap into
3270 multiple values to extend the available space.
3271
3272 [Bodo Moeller]
3273
3274 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
3275
3276 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
3277 (CVE-2006-4339) [Ben Laurie and Google Security Team]
3278
3279 *) Add AES IGE and biIGE modes.
3280 [Ben Laurie]
3281
3282 *) Change the Unix randomness entropy gathering to use poll() when
3283 possible instead of select(), since the latter has some
3284 undesirable limitations.
3285 [Darryl Miles via Richard Levitte and Bodo Moeller]
3286
3287 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
3288 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
3289 cannot be implicitly activated as part of, e.g., the "AES" alias.
3290 However, please upgrade to OpenSSL 0.9.9[-dev] for
3291 non-experimental use of the ECC ciphersuites to get TLS extension
3292 support, which is required for curve and point format negotiation
3293 to avoid potential handshake problems.
3294 [Bodo Moeller]
3295
3296 *) Disable rogue ciphersuites:
3297
3298 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
3299 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
3300 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
3301
3302 The latter two were purportedly from
3303 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
3304 appear there.
3305
3306 Also deactivate the remaining ciphersuites from
3307 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
3308 unofficial, and the ID has long expired.
3309 [Bodo Moeller]
3310
3311 *) Fix RSA blinding Heisenbug (problems sometimes occured on
3312 dual-core machines) and other potential thread-safety issues.
3313 [Bodo Moeller]
3314
3315 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
3316 versions), which is now available for royalty-free use
3317 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
3318 Also, add Camellia TLS ciphersuites from RFC 4132.
3319
3320 To minimize changes between patchlevels in the OpenSSL 0.9.8
3321 series, Camellia remains excluded from compilation unless OpenSSL
3322 is configured with 'enable-camellia'.
3323 [NTT]
3324
3325 *) Disable the padding bug check when compression is in use. The padding
3326 bug check assumes the first packet is of even length, this is not
3327 necessarily true if compresssion is enabled and can result in false
3328 positives causing handshake failure. The actual bug test is ancient
3329 code so it is hoped that implementations will either have fixed it by
3330 now or any which still have the bug do not support compression.
3331 [Steve Henson]
3332
3333 Changes between 0.9.8a and 0.9.8b [04 May 2006]
3334
3335 *) When applying a cipher rule check to see if string match is an explicit
3336 cipher suite and only match that one cipher suite if it is.
3337 [Steve Henson]
3338
3339 *) Link in manifests for VC++ if needed.
3340 [Austin Ziegler <halostatue@gmail.com>]
3341
3342 *) Update support for ECC-based TLS ciphersuites according to
3343 draft-ietf-tls-ecc-12.txt with proposed changes (but without
3344 TLS extensions, which are supported starting with the 0.9.9
3345 branch, not in the OpenSSL 0.9.8 branch).
3346 [Douglas Stebila]
3347
3348 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
3349 opaque EVP_CIPHER_CTX handling.
3350 [Steve Henson]
3351
3352 *) Fixes and enhancements to zlib compression code. We now only use
3353 "zlib1.dll" and use the default __cdecl calling convention on Win32
3354 to conform with the standards mentioned here:
3355 http://www.zlib.net/DLL_FAQ.txt
3356 Static zlib linking now works on Windows and the new --with-zlib-include
3357 --with-zlib-lib options to Configure can be used to supply the location
3358 of the headers and library. Gracefully handle case where zlib library
3359 can't be loaded.
3360 [Steve Henson]
3361
3362 *) Several fixes and enhancements to the OID generation code. The old code
3363 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
3364 handle numbers larger than ULONG_MAX, truncated printing and had a
3365 non standard OBJ_obj2txt() behaviour.
3366 [Steve Henson]
3367
3368 *) Add support for building of engines under engine/ as shared libraries
3369 under VC++ build system.
3370 [Steve Henson]
3371
3372 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
3373 Hopefully, we will not see any false combination of paths any more.
3374 [Richard Levitte]
3375
3376 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
3377
3378 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
3379 (part of SSL_OP_ALL). This option used to disable the
3380 countermeasure against man-in-the-middle protocol-version
3381 rollback in the SSL 2.0 server implementation, which is a bad
3382 idea. (CVE-2005-2969)
3383
3384 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
3385 for Information Security, National Institute of Advanced Industrial
3386 Science and Technology [AIST], Japan)]
3387
3388 *) Add two function to clear and return the verify parameter flags.
3389 [Steve Henson]
3390
3391 *) Keep cipherlists sorted in the source instead of sorting them at
3392 runtime, thus removing the need for a lock.
3393 [Nils Larsch]
3394
3395 *) Avoid some small subgroup attacks in Diffie-Hellman.
3396 [Nick Mathewson and Ben Laurie]
3397
3398 *) Add functions for well-known primes.
3399 [Nick Mathewson]
3400
3401 *) Extended Windows CE support.
3402 [Satoshi Nakamura and Andy Polyakov]
3403
3404 *) Initialize SSL_METHOD structures at compile time instead of during
3405 runtime, thus removing the need for a lock.
3406 [Steve Henson]
3407
3408 *) Make PKCS7_decrypt() work even if no certificate is supplied by
3409 attempting to decrypt each encrypted key in turn. Add support to
3410 smime utility.
3411 [Steve Henson]
3412
3413 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
3414
3415 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
3416 OpenSSL 0.9.8.]
3417
3418 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
3419 [Richard Levitte]
3420
3421 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
3422 key into the same file any more.
3423 [Richard Levitte]
3424
3425 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
3426 [Andy Polyakov]
3427
3428 *) Add -utf8 command line and config file option to 'ca'.
3429 [Stefan <stf@udoma.org]
3430
3431 *) Removed the macro des_crypt(), as it seems to conflict with some
3432 libraries. Use DES_crypt().
3433 [Richard Levitte]
3434
3435 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
3436 involves renaming the source and generated shared-libs for
3437 both. The engines will accept the corrected or legacy ids
3438 ('ncipher' and '4758_cca' respectively) when binding. NB,
3439 this only applies when building 'shared'.
3440 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
3441
3442 *) Add attribute functions to EVP_PKEY structure. Modify
3443 PKCS12_create() to recognize a CSP name attribute and
3444 use it. Make -CSP option work again in pkcs12 utility.
3445 [Steve Henson]
3446
3447 *) Add new functionality to the bn blinding code:
3448 - automatic re-creation of the BN_BLINDING parameters after
3449 a fixed number of uses (currently 32)
3450 - add new function for parameter creation
3451 - introduce flags to control the update behaviour of the
3452 BN_BLINDING parameters
3453 - hide BN_BLINDING structure
3454 Add a second BN_BLINDING slot to the RSA structure to improve
3455 performance when a single RSA object is shared among several
3456 threads.
3457 [Nils Larsch]
3458
3459 *) Add support for DTLS.
3460 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
3461
3462 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
3463 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
3464 [Walter Goulet]
3465
3466 *) Remove buggy and incomplete DH cert support from
3467 ssl/ssl_rsa.c and ssl/s3_both.c
3468 [Nils Larsch]
3469
3470 *) Use SHA-1 instead of MD5 as the default digest algorithm for
3471 the apps/openssl applications.
3472 [Nils Larsch]
3473
3474 *) Compile clean with "-Wall -Wmissing-prototypes
3475 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
3476 DEBUG_SAFESTACK must also be set.
3477 [Ben Laurie]
3478
3479 *) Change ./Configure so that certain algorithms can be disabled by default.
3480 The new counterpiece to "no-xxx" is "enable-xxx".
3481
3482 The patented RC5 and MDC2 algorithms will now be disabled unless
3483 "enable-rc5" and "enable-mdc2", respectively, are specified.
3484
3485 (IDEA remains enabled despite being patented. This is because IDEA
3486 is frequently required for interoperability, and there is no license
3487 fee for non-commercial use. As before, "no-idea" can be used to
3488 avoid this algorithm.)
3489
3490 [Bodo Moeller]
3491
3492 *) Add processing of proxy certificates (see RFC 3820). This work was
3493 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
3494 EGEE (Enabling Grids for E-science in Europe).
3495 [Richard Levitte]
3496
3497 *) RC4 performance overhaul on modern architectures/implementations, such
3498 as Intel P4, IA-64 and AMD64.
3499 [Andy Polyakov]
3500
3501 *) New utility extract-section.pl. This can be used specify an alternative
3502 section number in a pod file instead of having to treat each file as
3503 a separate case in Makefile. This can be done by adding two lines to the
3504 pod file:
3505
3506 =for comment openssl_section:XXX
3507
3508 The blank line is mandatory.
3509
3510 [Steve Henson]
3511
3512 *) New arguments -certform, -keyform and -pass for s_client and s_server
3513 to allow alternative format key and certificate files and passphrase
3514 sources.
3515 [Steve Henson]
3516
3517 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
3518 update associated structures and add various utility functions.
3519
3520 Add new policy related verify parameters, include policy checking in
3521 standard verify code. Enhance 'smime' application with extra parameters
3522 to support policy checking and print out.
3523 [Steve Henson]
3524
3525 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
3526 Nehemiah processors. These extensions support AES encryption in hardware
3527 as well as RNG (though RNG support is currently disabled).
3528 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
3529
3530 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
3531 [Geoff Thorpe]
3532
3533 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
3534 [Andy Polyakov and a number of other people]
3535
3536 *) Improved PowerPC platform support. Most notably BIGNUM assembler
3537 implementation contributed by IBM.
3538 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
3539
3540 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
3541 exponent rather than 'unsigned long'. There is a corresponding change to
3542 the new 'rsa_keygen' element of the RSA_METHOD structure.
3543 [Jelte Jansen, Geoff Thorpe]
3544
3545 *) Functionality for creating the initial serial number file is now
3546 moved from CA.pl to the 'ca' utility with a new option -create_serial.
3547
3548 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
3549 number file to 1, which is bound to cause problems. To avoid
3550 the problems while respecting compatibility between different 0.9.7
3551 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
3552 CA.pl for serial number initialization. With the new release 0.9.8,
3553 we can fix the problem directly in the 'ca' utility.)
3554 [Steve Henson]
3555
3556 *) Reduced header interdepencies by declaring more opaque objects in
3557 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
3558 give fewer recursive includes, which could break lazy source code - so
3559 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
3560 developers should define this symbol when building and using openssl to
3561 ensure they track the recommended behaviour, interfaces, [etc], but
3562 backwards-compatible behaviour prevails when this isn't defined.
3563 [Geoff Thorpe]
3564
3565 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
3566 [Steve Henson]
3567
3568 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
3569 This will generate a random key of the appropriate length based on the
3570 cipher context. The EVP_CIPHER can provide its own random key generation
3571 routine to support keys of a specific form. This is used in the des and
3572 3des routines to generate a key of the correct parity. Update S/MIME
3573 code to use new functions and hence generate correct parity DES keys.
3574 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
3575 valid (weak or incorrect parity).
3576 [Steve Henson]
3577
3578 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
3579 as looking them up. This is useful when the verified structure may contain
3580 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
3581 present unless the new PKCS7_NO_CRL flag is asserted.
3582 [Steve Henson]
3583
3584 *) Extend ASN1 oid configuration module. It now additionally accepts the
3585 syntax:
3586
3587 shortName = some long name, 1.2.3.4
3588 [Steve Henson]
3589
3590 *) Reimplemented the BN_CTX implementation. There is now no more static
3591 limitation on the number of variables it can handle nor the depth of the
3592 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
3593 information can now expand as required, and rather than having a single
3594 static array of bignums, BN_CTX now uses a linked-list of such arrays
3595 allowing it to expand on demand whilst maintaining the usefulness of
3596 BN_CTX's "bundling".
3597 [Geoff Thorpe]
3598
3599 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
3600 to allow all RSA operations to function using a single BN_CTX.
3601 [Geoff Thorpe]
3602
3603 *) Preliminary support for certificate policy evaluation and checking. This
3604 is initially intended to pass the tests outlined in "Conformance Testing
3605 of Relying Party Client Certificate Path Processing Logic" v1.07.
3606 [Steve Henson]
3607
3608 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
3609 remained unused and not that useful. A variety of other little bignum
3610 tweaks and fixes have also been made continuing on from the audit (see
3611 below).
3612 [Geoff Thorpe]
3613
3614 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
3615 associated ASN1, EVP and SSL functions and old ASN1 macros.
3616 [Richard Levitte]
3617
3618 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
3619 and this should never fail. So the return value from the use of
3620 BN_set_word() (which can fail due to needless expansion) is now deprecated;
3621 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
3622 [Geoff Thorpe]
3623
3624 *) BN_CTX_get() should return zero-valued bignums, providing the same
3625 initialised value as BN_new().
3626 [Geoff Thorpe, suggested by Ulf Möller]
3627
3628 *) Support for inhibitAnyPolicy certificate extension.
3629 [Steve Henson]
3630
3631 *) An audit of the BIGNUM code is underway, for which debugging code is
3632 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
3633 is considered valid when processing BIGNUMs, and causes execution to
3634 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
3635 further steps are taken to deliberately pollute unused data in BIGNUM
3636 structures to try and expose faulty code further on. For now, openssl will
3637 (in its default mode of operation) continue to tolerate the inconsistent
3638 forms that it has tolerated in the past, but authors and packagers should
3639 consider trying openssl and their own applications when compiled with
3640 these debugging symbols defined. It will help highlight potential bugs in
3641 their own code, and will improve the test coverage for OpenSSL itself. At
3642 some point, these tighter rules will become openssl's default to improve
3643 maintainability, though the assert()s and other overheads will remain only
3644 in debugging configurations. See bn.h for more details.
3645 [Geoff Thorpe, Nils Larsch, Ulf Möller]
3646
3647 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
3648 that can only be obtained through BN_CTX_new() (which implicitly
3649 initialises it). The presence of this function only made it possible
3650 to overwrite an existing structure (and cause memory leaks).
3651 [Geoff Thorpe]
3652
3653 *) Because of the callback-based approach for implementing LHASH as a
3654 template type, lh_insert() adds opaque objects to hash-tables and
3655 lh_doall() or lh_doall_arg() are typically used with a destructor callback
3656 to clean up those corresponding objects before destroying the hash table
3657 (and losing the object pointers). So some over-zealous constifications in
3658 LHASH have been relaxed so that lh_insert() does not take (nor store) the
3659 objects as "const" and the lh_doall[_arg] callback wrappers are not
3660 prototyped to have "const" restrictions on the object pointers they are
3661 given (and so aren't required to cast them away any more).
3662 [Geoff Thorpe]
3663
3664 *) The tmdiff.h API was so ugly and minimal that our own timing utility
3665 (speed) prefers to use its own implementation. The two implementations
3666 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
3667 its object type properly exposed (MS_TM) instead of casting to/from "char
3668 *". This may still change yet if someone realises MS_TM and "ms_time_***"
3669 aren't necessarily the greatest nomenclatures - but this is what was used
3670 internally to the implementation so I've used that for now.
3671 [Geoff Thorpe]
3672
3673 *) Ensure that deprecated functions do not get compiled when
3674 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
3675 the self-tests were still using deprecated key-generation functions so
3676 these have been updated also.
3677 [Geoff Thorpe]
3678
3679 *) Reorganise PKCS#7 code to separate the digest location functionality
3680 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
3681 New function PKCS7_set_digest() to set the digest type for PKCS#7
3682 digestedData type. Add additional code to correctly generate the
3683 digestedData type and add support for this type in PKCS7 initialization
3684 functions.
3685 [Steve Henson]
3686
3687 *) New function PKCS7_set0_type_other() this initializes a PKCS7
3688 structure of type "other".
3689 [Steve Henson]
3690
3691 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
3692 sure the loop does correctly stop and breaking ("division by zero")
3693 modulus operations are not performed. The (pre-generated) prime
3694 table crypto/bn/bn_prime.h was already correct, but it could not be
3695 re-generated on some platforms because of the "division by zero"
3696 situation in the script.
3697 [Ralf S. Engelschall]
3698
3699 *) Update support for ECC-based TLS ciphersuites according to
3700 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
3701 SHA-1 now is only used for "small" curves (where the
3702 representation of a field element takes up to 24 bytes); for
3703 larger curves, the field element resulting from ECDH is directly
3704 used as premaster secret.
3705 [Douglas Stebila (Sun Microsystems Laboratories)]
3706
3707 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
3708 curve secp160r1 to the tests.
3709 [Douglas Stebila (Sun Microsystems Laboratories)]
3710
3711 *) Add the possibility to load symbols globally with DSO.
3712 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
3713
3714 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
3715 control of the error stack.
3716 [Richard Levitte]
3717
3718 *) Add support for STORE in ENGINE.
3719 [Richard Levitte]
3720
3721 *) Add the STORE type. The intention is to provide a common interface
3722 to certificate and key stores, be they simple file-based stores, or
3723 HSM-type store, or LDAP stores, or...
3724 NOTE: The code is currently UNTESTED and isn't really used anywhere.
3725 [Richard Levitte]
3726
3727 *) Add a generic structure called OPENSSL_ITEM. This can be used to
3728 pass a list of arguments to any function as well as provide a way
3729 for a function to pass data back to the caller.
3730 [Richard Levitte]
3731
3732 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
3733 works like BUF_strdup() but can be used to duplicate a portion of
3734 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
3735 a memory area.
3736 [Richard Levitte]
3737
3738 *) Add the function sk_find_ex() which works like sk_find(), but will
3739 return an index to an element even if an exact match couldn't be
3740 found. The index is guaranteed to point at the element where the
3741 searched-for key would be inserted to preserve sorting order.
3742 [Richard Levitte]
3743
3744 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
3745 takes an extra flags argument for optional functionality. Currently,
3746 the following flags are defined:
3747
3748 OBJ_BSEARCH_VALUE_ON_NOMATCH
3749 This one gets OBJ_bsearch_ex() to return a pointer to the first
3750 element where the comparing function returns a negative or zero
3751 number.
3752
3753 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
3754 This one gets OBJ_bsearch_ex() to return a pointer to the first
3755 element where the comparing function returns zero. This is useful
3756 if there are more than one element where the comparing function
3757 returns zero.
3758 [Richard Levitte]
3759
3760 *) Make it possible to create self-signed certificates with 'openssl ca'
3761 in such a way that the self-signed certificate becomes part of the
3762 CA database and uses the same mechanisms for serial number generation
3763 as all other certificate signing. The new flag '-selfsign' enables
3764 this functionality. Adapt CA.sh and CA.pl.in.
3765 [Richard Levitte]
3766
3767 *) Add functionality to check the public key of a certificate request
3768 against a given private. This is useful to check that a certificate
3769 request can be signed by that key (self-signing).
3770 [Richard Levitte]
3771
3772 *) Make it possible to have multiple active certificates with the same
3773 subject in the CA index file. This is done only if the keyword
3774 'unique_subject' is set to 'no' in the main CA section (default
3775 if 'CA_default') of the configuration file. The value is saved
3776 with the database itself in a separate index attribute file,
3777 named like the index file with '.attr' appended to the name.
3778 [Richard Levitte]
3779
3780 *) Generate muti valued AVAs using '+' notation in config files for
3781 req and dirName.
3782 [Steve Henson]
3783
3784 *) Support for nameConstraints certificate extension.
3785 [Steve Henson]
3786
3787 *) Support for policyConstraints certificate extension.
3788 [Steve Henson]
3789
3790 *) Support for policyMappings certificate extension.
3791 [Steve Henson]
3792
3793 *) Make sure the default DSA_METHOD implementation only uses its
3794 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
3795 and change its own handlers to be NULL so as to remove unnecessary
3796 indirection. This lets alternative implementations fallback to the
3797 default implementation more easily.
3798 [Geoff Thorpe]
3799
3800 *) Support for directoryName in GeneralName related extensions
3801 in config files.
3802 [Steve Henson]
3803
3804 *) Make it possible to link applications using Makefile.shared.
3805 Make that possible even when linking against static libraries!
3806 [Richard Levitte]
3807
3808 *) Support for single pass processing for S/MIME signing. This now
3809 means that S/MIME signing can be done from a pipe, in addition
3810 cleartext signing (multipart/signed type) is effectively streaming
3811 and the signed data does not need to be all held in memory.
3812
3813 This is done with a new flag PKCS7_STREAM. When this flag is set
3814 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
3815 is done after the data is output (and digests calculated) in
3816 SMIME_write_PKCS7().
3817 [Steve Henson]
3818
3819 *) Add full support for -rpath/-R, both in shared libraries and
3820 applications, at least on the platforms where it's known how
3821 to do it.
3822 [Richard Levitte]
3823
3824 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
3825 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
3826 will now compute a table of multiples of the generator that
3827 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
3828 faster (notably in the case of a single point multiplication,
3829 scalar * generator).
3830 [Nils Larsch, Bodo Moeller]
3831
3832 *) IPv6 support for certificate extensions. The various extensions
3833 which use the IP:a.b.c.d can now take IPv6 addresses using the
3834 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
3835 correctly.
3836 [Steve Henson]
3837
3838 *) Added an ENGINE that implements RSA by performing private key
3839 exponentiations with the GMP library. The conversions to and from
3840 GMP's mpz_t format aren't optimised nor are any montgomery forms
3841 cached, and on x86 it appears OpenSSL's own performance has caught up.
3842 However there are likely to be other architectures where GMP could
3843 provide a boost. This ENGINE is not built in by default, but it can be
3844 specified at Configure time and should be accompanied by the necessary
3845 linker additions, eg;
3846 ./config -DOPENSSL_USE_GMP -lgmp
3847 [Geoff Thorpe]
3848
3849 *) "openssl engine" will not display ENGINE/DSO load failure errors when
3850 testing availability of engines with "-t" - the old behaviour is
3851 produced by increasing the feature's verbosity with "-tt".
3852 [Geoff Thorpe]
3853
3854 *) ECDSA routines: under certain error conditions uninitialized BN objects
3855 could be freed. Solution: make sure initialization is performed early
3856 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
3857 via PR#459)
3858 [Lutz Jaenicke]
3859
3860 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
3861 and DH_METHOD (eg. by ENGINE implementations) to override the normal
3862 software implementations. For DSA and DH, parameter generation can
3863 also be overriden by providing the appropriate method callbacks.
3864 [Geoff Thorpe]
3865
3866 *) Change the "progress" mechanism used in key-generation and
3867 primality testing to functions that take a new BN_GENCB pointer in
3868 place of callback/argument pairs. The new API functions have "_ex"
3869 postfixes and the older functions are reimplemented as wrappers for
3870 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
3871 declarations of the old functions to help (graceful) attempts to
3872 migrate to the new functions. Also, the new key-generation API
3873 functions operate on a caller-supplied key-structure and return
3874 success/failure rather than returning a key or NULL - this is to
3875 help make "keygen" another member function of RSA_METHOD etc.
3876
3877 Example for using the new callback interface:
3878
3879 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
3880 void *my_arg = ...;
3881 BN_GENCB my_cb;
3882
3883 BN_GENCB_set(&my_cb, my_callback, my_arg);
3884
3885 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
3886 /* For the meaning of a, b in calls to my_callback(), see the
3887 * documentation of the function that calls the callback.
3888 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
3889 * my_callback should return 1 if it wants BN_is_prime_ex()
3890 * to continue, or 0 to stop.
3891 */
3892
3893 [Geoff Thorpe]
3894
3895 *) Change the ZLIB compression method to be stateful, and make it
3896 available to TLS with the number defined in
3897 draft-ietf-tls-compression-04.txt.
3898 [Richard Levitte]
3899
3900 *) Add the ASN.1 structures and functions for CertificatePair, which
3901 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
3902
3903 CertificatePair ::= SEQUENCE {
3904 forward [0] Certificate OPTIONAL,
3905 reverse [1] Certificate OPTIONAL,
3906 -- at least one of the pair shall be present -- }
3907
3908 Also implement the PEM functions to read and write certificate
3909 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
3910
3911 This needed to be defined, mostly for the sake of the LDAP
3912 attribute crossCertificatePair, but may prove useful elsewhere as
3913 well.
3914 [Richard Levitte]
3915
3916 *) Make it possible to inhibit symlinking of shared libraries in
3917 Makefile.shared, for Cygwin's sake.
3918 [Richard Levitte]
3919
3920 *) Extend the BIGNUM API by creating a function
3921 void BN_set_negative(BIGNUM *a, int neg);
3922 and a macro that behave like
3923 int BN_is_negative(const BIGNUM *a);
3924
3925 to avoid the need to access 'a->neg' directly in applications.
3926 [Nils Larsch]
3927
3928 *) Implement fast modular reduction for pseudo-Mersenne primes
3929 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
3930 EC_GROUP_new_curve_GFp() will now automatically use this
3931 if applicable.
3932 [Nils Larsch <nla@trustcenter.de>]
3933
3934 *) Add new lock type (CRYPTO_LOCK_BN).
3935 [Bodo Moeller]
3936
3937 *) Change the ENGINE framework to automatically load engines
3938 dynamically from specific directories unless they could be
3939 found to already be built in or loaded. Move all the
3940 current engines except for the cryptodev one to a new
3941 directory engines/.
3942 The engines in engines/ are built as shared libraries if
3943 the "shared" options was given to ./Configure or ./config.
3944 Otherwise, they are inserted in libcrypto.a.
3945 /usr/local/ssl/engines is the default directory for dynamic
3946 engines, but that can be overriden at configure time through
3947 the usual use of --prefix and/or --openssldir, and at run
3948 time with the environment variable OPENSSL_ENGINES.
3949 [Geoff Thorpe and Richard Levitte]
3950
3951 *) Add Makefile.shared, a helper makefile to build shared
3952 libraries. Adapt Makefile.org.
3953 [Richard Levitte]
3954
3955 *) Add version info to Win32 DLLs.
3956 [Peter 'Luna' Runestig" <peter@runestig.com>]
3957
3958 *) Add new 'medium level' PKCS#12 API. Certificates and keys
3959 can be added using this API to created arbitrary PKCS#12
3960 files while avoiding the low level API.
3961
3962 New options to PKCS12_create(), key or cert can be NULL and
3963 will then be omitted from the output file. The encryption
3964 algorithm NIDs can be set to -1 for no encryption, the mac
3965 iteration count can be set to 0 to omit the mac.
3966
3967 Enhance pkcs12 utility by making the -nokeys and -nocerts
3968 options work when creating a PKCS#12 file. New option -nomac
3969 to omit the mac, NONE can be set for an encryption algorithm.
3970 New code is modified to use the enhanced PKCS12_create()
3971 instead of the low level API.
3972 [Steve Henson]
3973
3974 *) Extend ASN1 encoder to support indefinite length constructed
3975 encoding. This can output sequences tags and octet strings in
3976 this form. Modify pk7_asn1.c to support indefinite length
3977 encoding. This is experimental and needs additional code to
3978 be useful, such as an ASN1 bio and some enhanced streaming
3979 PKCS#7 code.
3980
3981 Extend template encode functionality so that tagging is passed
3982 down to the template encoder.
3983 [Steve Henson]
3984
3985 *) Let 'openssl req' fail if an argument to '-newkey' is not
3986 recognized instead of using RSA as a default.
3987 [Bodo Moeller]
3988
3989 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
3990 As these are not official, they are not included in "ALL";
3991 the "ECCdraft" ciphersuite group alias can be used to select them.
3992 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
3993
3994 *) Add ECDH engine support.
3995 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
3996
3997 *) Add ECDH in new directory crypto/ecdh/.
3998 [Douglas Stebila (Sun Microsystems Laboratories)]
3999
4000 *) Let BN_rand_range() abort with an error after 100 iterations
4001 without success (which indicates a broken PRNG).
4002 [Bodo Moeller]
4003
4004 *) Change BN_mod_sqrt() so that it verifies that the input value
4005 is really the square of the return value. (Previously,
4006 BN_mod_sqrt would show GIGO behaviour.)
4007 [Bodo Moeller]
4008
4009 *) Add named elliptic curves over binary fields from X9.62, SECG,
4010 and WAP/WTLS; add OIDs that were still missing.
4011
4012 [Sheueling Chang Shantz and Douglas Stebila
4013 (Sun Microsystems Laboratories)]
4014
4015 *) Extend the EC library for elliptic curves over binary fields
4016 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
4017 New EC_METHOD:
4018
4019 EC_GF2m_simple_method
4020
4021 New API functions:
4022
4023 EC_GROUP_new_curve_GF2m
4024 EC_GROUP_set_curve_GF2m
4025 EC_GROUP_get_curve_GF2m
4026 EC_POINT_set_affine_coordinates_GF2m
4027 EC_POINT_get_affine_coordinates_GF2m
4028 EC_POINT_set_compressed_coordinates_GF2m
4029
4030 Point compression for binary fields is disabled by default for
4031 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
4032 enable it).
4033
4034 As binary polynomials are represented as BIGNUMs, various members
4035 of the EC_GROUP and EC_POINT data structures can be shared
4036 between the implementations for prime fields and binary fields;
4037 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
4038 are essentially identical to their ..._GFp counterparts.
4039 (For simplicity, the '..._GFp' prefix has been dropped from
4040 various internal method names.)
4041
4042 An internal 'field_div' method (similar to 'field_mul' and
4043 'field_sqr') has been added; this is used only for binary fields.
4044
4045 [Sheueling Chang Shantz and Douglas Stebila
4046 (Sun Microsystems Laboratories)]
4047
4048 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
4049 through methods ('mul', 'precompute_mult').
4050
4051 The generic implementations (now internally called 'ec_wNAF_mul'
4052 and 'ec_wNAF_precomputed_mult') remain the default if these
4053 methods are undefined.
4054
4055 [Sheueling Chang Shantz and Douglas Stebila
4056 (Sun Microsystems Laboratories)]
4057
4058 *) New function EC_GROUP_get_degree, which is defined through
4059 EC_METHOD. For curves over prime fields, this returns the bit
4060 length of the modulus.
4061
4062 [Sheueling Chang Shantz and Douglas Stebila
4063 (Sun Microsystems Laboratories)]
4064
4065 *) New functions EC_GROUP_dup, EC_POINT_dup.
4066 (These simply call ..._new and ..._copy).
4067
4068 [Sheueling Chang Shantz and Douglas Stebila
4069 (Sun Microsystems Laboratories)]
4070
4071 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
4072 Polynomials are represented as BIGNUMs (where the sign bit is not
4073 used) in the following functions [macros]:
4074
4075 BN_GF2m_add
4076 BN_GF2m_sub [= BN_GF2m_add]
4077 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
4078 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
4079 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
4080 BN_GF2m_mod_inv
4081 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
4082 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
4083 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
4084 BN_GF2m_cmp [= BN_ucmp]
4085
4086 (Note that only the 'mod' functions are actually for fields GF(2^m).
4087 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
4088
4089 For some functions, an the irreducible polynomial defining a
4090 field can be given as an 'unsigned int[]' with strictly
4091 decreasing elements giving the indices of those bits that are set;
4092 i.e., p[] represents the polynomial
4093 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
4094 where
4095 p[0] > p[1] > ... > p[k] = 0.
4096 This applies to the following functions:
4097
4098 BN_GF2m_mod_arr
4099 BN_GF2m_mod_mul_arr
4100 BN_GF2m_mod_sqr_arr
4101 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
4102 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
4103 BN_GF2m_mod_exp_arr
4104 BN_GF2m_mod_sqrt_arr
4105 BN_GF2m_mod_solve_quad_arr
4106 BN_GF2m_poly2arr
4107 BN_GF2m_arr2poly
4108
4109 Conversion can be performed by the following functions:
4110
4111 BN_GF2m_poly2arr
4112 BN_GF2m_arr2poly
4113
4114 bntest.c has additional tests for binary polynomial arithmetic.
4115
4116 Two implementations for BN_GF2m_mod_div() are available.
4117 The default algorithm simply uses BN_GF2m_mod_inv() and
4118 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
4119 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
4120 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
4121
4122 [Sheueling Chang Shantz and Douglas Stebila
4123 (Sun Microsystems Laboratories)]
4124
4125 *) Add new error code 'ERR_R_DISABLED' that can be used when some
4126 functionality is disabled at compile-time.
4127 [Douglas Stebila <douglas.stebila@sun.com>]
4128
4129 *) Change default behaviour of 'openssl asn1parse' so that more
4130 information is visible when viewing, e.g., a certificate:
4131
4132 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
4133 mode the content of non-printable OCTET STRINGs is output in a
4134 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
4135 avoid the appearance of a printable string.
4136 [Nils Larsch <nla@trustcenter.de>]
4137
4138 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
4139 functions
4140 EC_GROUP_set_asn1_flag()
4141 EC_GROUP_get_asn1_flag()
4142 EC_GROUP_set_point_conversion_form()
4143 EC_GROUP_get_point_conversion_form()
4144 These control ASN1 encoding details:
4145 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
4146 has been set to OPENSSL_EC_NAMED_CURVE.
4147 - Points are encoded in uncompressed form by default; options for
4148 asn1_for are as for point2oct, namely
4149 POINT_CONVERSION_COMPRESSED
4150 POINT_CONVERSION_UNCOMPRESSED
4151 POINT_CONVERSION_HYBRID
4152
4153 Also add 'seed' and 'seed_len' members to EC_GROUP with access
4154 functions
4155 EC_GROUP_set_seed()
4156 EC_GROUP_get0_seed()
4157 EC_GROUP_get_seed_len()
4158 This is used only for ASN1 purposes (so far).
4159 [Nils Larsch <nla@trustcenter.de>]
4160
4161 *) Add 'field_type' member to EC_METHOD, which holds the NID
4162 of the appropriate field type OID. The new function
4163 EC_METHOD_get_field_type() returns this value.
4164 [Nils Larsch <nla@trustcenter.de>]
4165
4166 *) Add functions
4167 EC_POINT_point2bn()
4168 EC_POINT_bn2point()
4169 EC_POINT_point2hex()
4170 EC_POINT_hex2point()
4171 providing useful interfaces to EC_POINT_point2oct() and
4172 EC_POINT_oct2point().
4173 [Nils Larsch <nla@trustcenter.de>]
4174
4175 *) Change internals of the EC library so that the functions
4176 EC_GROUP_set_generator()
4177 EC_GROUP_get_generator()
4178 EC_GROUP_get_order()
4179 EC_GROUP_get_cofactor()
4180 are implemented directly in crypto/ec/ec_lib.c and not dispatched
4181 to methods, which would lead to unnecessary code duplication when
4182 adding different types of curves.
4183 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
4184
4185 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
4186 arithmetic, and such that modified wNAFs are generated
4187 (which avoid length expansion in many cases).
4188 [Bodo Moeller]
4189
4190 *) Add a function EC_GROUP_check_discriminant() (defined via
4191 EC_METHOD) that verifies that the curve discriminant is non-zero.
4192
4193 Add a function EC_GROUP_check() that makes some sanity tests
4194 on a EC_GROUP, its generator and order. This includes
4195 EC_GROUP_check_discriminant().
4196 [Nils Larsch <nla@trustcenter.de>]
4197
4198 *) Add ECDSA in new directory crypto/ecdsa/.
4199
4200 Add applications 'openssl ecparam' and 'openssl ecdsa'
4201 (these are based on 'openssl dsaparam' and 'openssl dsa').
4202
4203 ECDSA support is also included in various other files across the
4204 library. Most notably,
4205 - 'openssl req' now has a '-newkey ecdsa:file' option;
4206 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
4207 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
4208 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
4209 them suitable for ECDSA where domain parameters must be
4210 extracted before the specific public key;
4211 - ECDSA engine support has been added.
4212 [Nils Larsch <nla@trustcenter.de>]
4213
4214 *) Include some named elliptic curves, and add OIDs from X9.62,
4215 SECG, and WAP/WTLS. Each curve can be obtained from the new
4216 function
4217 EC_GROUP_new_by_curve_name(),
4218 and the list of available named curves can be obtained with
4219 EC_get_builtin_curves().
4220 Also add a 'curve_name' member to EC_GROUP objects, which can be
4221 accessed via
4222 EC_GROUP_set_curve_name()
4223 EC_GROUP_get_curve_name()
4224 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
4225
4226 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4227 was actually never needed) and in BN_mul(). The removal in BN_mul()
4228 required a small change in bn_mul_part_recursive() and the addition
4229 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4230 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4231 bn_sub_words() and bn_add_words() except they take arrays with
4232 differing sizes.
4233 [Richard Levitte]
4234
4235 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
4236
4237 *) Cleanse PEM buffers before freeing them since they may contain
4238 sensitive data.
4239 [Benjamin Bennett <ben@psc.edu>]
4240
4241 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
4242 a ciphersuite string such as "DEFAULT:RSA" cannot enable
4243 authentication-only ciphersuites.
4244 [Bodo Moeller]
4245
4246 *) Since AES128 and AES256 share a single mask bit in the logic of
4247 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
4248 kludge to work properly if AES128 is available and AES256 isn't.
4249 [Victor Duchovni]
4250
4251 *) Expand security boundary to match 1.1.1 module.
4252 [Steve Henson]
4253
4254 *) Remove redundant features: hash file source, editing of test vectors
4255 modify fipsld to use external fips_premain.c signature.
4256 [Steve Henson]
4257
4258 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
4259 run algorithm test programs.
4260 [Steve Henson]
4261
4262 *) Make algorithm test programs more tolerant of whitespace.
4263 [Steve Henson]
4264
4265 *) Have SSL/TLS server implementation tolerate "mismatched" record
4266 protocol version while receiving ClientHello even if the
4267 ClientHello is fragmented. (The server can't insist on the
4268 particular protocol version it has chosen before the ServerHello
4269 message has informed the client about his choice.)
4270 [Bodo Moeller]
4271
4272 *) Load error codes if they are not already present instead of using a
4273 static variable. This allows them to be cleanly unloaded and reloaded.
4274 [Steve Henson]
4275
4276 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
4277
4278 *) Introduce limits to prevent malicious keys being able to
4279 cause a denial of service. (CVE-2006-2940)
4280 [Steve Henson, Bodo Moeller]
4281
4282 *) Fix ASN.1 parsing of certain invalid structures that can result
4283 in a denial of service. (CVE-2006-2937) [Steve Henson]
4284
4285 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
4286 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
4287
4288 *) Fix SSL client code which could crash if connecting to a
4289 malicious SSLv2 server. (CVE-2006-4343)
4290 [Tavis Ormandy and Will Drewry, Google Security Team]
4291
4292 *) Change ciphersuite string processing so that an explicit
4293 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
4294 will no longer include "AES128-SHA"), and any other similar
4295 ciphersuite (same bitmap) from *other* protocol versions (so that
4296 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
4297 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
4298 changes from 0.9.8b and 0.9.8d.
4299 [Bodo Moeller]
4300
4301 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
4302
4303 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
4304 (CVE-2006-4339) [Ben Laurie and Google Security Team]
4305
4306 *) Change the Unix randomness entropy gathering to use poll() when
4307 possible instead of select(), since the latter has some
4308 undesirable limitations.
4309 [Darryl Miles via Richard Levitte and Bodo Moeller]
4310
4311 *) Disable rogue ciphersuites:
4312
4313 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
4314 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
4315 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
4316
4317 The latter two were purportedly from
4318 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
4319 appear there.
4320
4321 Also deactive the remaining ciphersuites from
4322 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
4323 unofficial, and the ID has long expired.
4324 [Bodo Moeller]
4325
4326 *) Fix RSA blinding Heisenbug (problems sometimes occured on
4327 dual-core machines) and other potential thread-safety issues.
4328 [Bodo Moeller]
4329
4330 Changes between 0.9.7i and 0.9.7j [04 May 2006]
4331
4332 *) Adapt fipsld and the build system to link against the validated FIPS
4333 module in FIPS mode.
4334 [Steve Henson]
4335
4336 *) Fixes for VC++ 2005 build under Windows.
4337 [Steve Henson]
4338
4339 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
4340 from a Windows bash shell such as MSYS. It is autodetected from the
4341 "config" script when run from a VC++ environment. Modify standard VC++
4342 build to use fipscanister.o from the GNU make build.
4343 [Steve Henson]
4344
4345 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
4346
4347 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
4348 The value now differs depending on if you build for FIPS or not.
4349 BEWARE! A program linked with a shared FIPSed libcrypto can't be
4350 safely run with a non-FIPSed libcrypto, as it may crash because of
4351 the difference induced by this change.
4352 [Andy Polyakov]
4353
4354 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
4355
4356 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
4357 (part of SSL_OP_ALL). This option used to disable the
4358 countermeasure against man-in-the-middle protocol-version
4359 rollback in the SSL 2.0 server implementation, which is a bad
4360 idea. (CVE-2005-2969)
4361
4362 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
4363 for Information Security, National Institute of Advanced Industrial
4364 Science and Technology [AIST], Japan)]
4365
4366 *) Minimal support for X9.31 signatures and PSS padding modes. This is
4367 mainly for FIPS compliance and not fully integrated at this stage.
4368 [Steve Henson]
4369
4370 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
4371 the exponentiation using a fixed-length exponent. (Otherwise,
4372 the information leaked through timing could expose the secret key
4373 after many signatures; cf. Bleichenbacher's attack on DSA with
4374 biased k.)
4375 [Bodo Moeller]
4376
4377 *) Make a new fixed-window mod_exp implementation the default for
4378 RSA, DSA, and DH private-key operations so that the sequence of
4379 squares and multiplies and the memory access pattern are
4380 independent of the particular secret key. This will mitigate
4381 cache-timing and potential related attacks.
4382
4383 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
4384 and this is automatically used by BN_mod_exp_mont() if the new flag
4385 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
4386 will use this BN flag for private exponents unless the flag
4387 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
4388 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
4389
4390 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
4391
4392 *) Change the client implementation for SSLv23_method() and
4393 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
4394 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
4395 (Previously, the SSL 2.0 backwards compatible Client Hello
4396 message format would be used even with SSL_OP_NO_SSLv2.)
4397 [Bodo Moeller]
4398
4399 *) Add support for smime-type MIME parameter in S/MIME messages which some
4400 clients need.
4401 [Steve Henson]
4402
4403 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
4404 a threadsafe manner. Modify rsa code to use new function and add calls
4405 to dsa and dh code (which had race conditions before).
4406 [Steve Henson]
4407
4408 *) Include the fixed error library code in the C error file definitions
4409 instead of fixing them up at runtime. This keeps the error code
4410 structures constant.
4411 [Steve Henson]
4412
4413 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
4414
4415 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
4416 OpenSSL 0.9.8.]
4417
4418 *) Fixes for newer kerberos headers. NB: the casts are needed because
4419 the 'length' field is signed on one version and unsigned on another
4420 with no (?) obvious way to tell the difference, without these VC++
4421 complains. Also the "definition" of FAR (blank) is no longer included
4422 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
4423 some needed definitions.
4424 [Steve Henson]
4425
4426 *) Undo Cygwin change.
4427 [Ulf Möller]
4428
4429 *) Added support for proxy certificates according to RFC 3820.
4430 Because they may be a security thread to unaware applications,
4431 they must be explicitely allowed in run-time. See
4432 docs/HOWTO/proxy_certificates.txt for further information.
4433 [Richard Levitte]
4434
4435 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
4436
4437 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
4438 server and client random values. Previously
4439 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
4440 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
4441
4442 This change has negligible security impact because:
4443
4444 1. Server and client random values still have 24 bytes of pseudo random
4445 data.
4446
4447 2. Server and client random values are sent in the clear in the initial
4448 handshake.
4449
4450 3. The master secret is derived using the premaster secret (48 bytes in
4451 size for static RSA ciphersuites) as well as client server and random
4452 values.
4453
4454 The OpenSSL team would like to thank the UK NISCC for bringing this issue
4455 to our attention.
4456
4457 [Stephen Henson, reported by UK NISCC]
4458
4459 *) Use Windows randomness collection on Cygwin.
4460 [Ulf Möller]
4461
4462 *) Fix hang in EGD/PRNGD query when communication socket is closed
4463 prematurely by EGD/PRNGD.
4464 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
4465
4466 *) Prompt for pass phrases when appropriate for PKCS12 input format.
4467 [Steve Henson]
4468
4469 *) Back-port of selected performance improvements from development
4470 branch, as well as improved support for PowerPC platforms.
4471 [Andy Polyakov]
4472
4473 *) Add lots of checks for memory allocation failure, error codes to indicate
4474 failure and freeing up memory if a failure occurs.
4475 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
4476
4477 *) Add new -passin argument to dgst.
4478 [Steve Henson]
4479
4480 *) Perform some character comparisons of different types in X509_NAME_cmp:
4481 this is needed for some certificates that reencode DNs into UTF8Strings
4482 (in violation of RFC3280) and can't or wont issue name rollover
4483 certificates.
4484 [Steve Henson]
4485
4486 *) Make an explicit check during certificate validation to see that
4487 the CA setting in each certificate on the chain is correct. As a
4488 side effect always do the following basic checks on extensions,
4489 not just when there's an associated purpose to the check:
4490
4491 - if there is an unhandled critical extension (unless the user
4492 has chosen to ignore this fault)
4493 - if the path length has been exceeded (if one is set at all)
4494 - that certain extensions fit the associated purpose (if one has
4495 been given)
4496 [Richard Levitte]
4497
4498 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
4499
4500 *) Avoid a race condition when CRLs are checked in a multi threaded
4501 environment. This would happen due to the reordering of the revoked
4502 entries during signature checking and serial number lookup. Now the
4503 encoding is cached and the serial number sort performed under a lock.
4504 Add new STACK function sk_is_sorted().
4505 [Steve Henson]
4506
4507 *) Add Delta CRL to the extension code.
4508 [Steve Henson]
4509
4510 *) Various fixes to s3_pkt.c so alerts are sent properly.
4511 [David Holmes <d.holmes@f5.com>]
4512
4513 *) Reduce the chances of duplicate issuer name and serial numbers (in
4514 violation of RFC3280) using the OpenSSL certificate creation utilities.
4515 This is done by creating a random 64 bit value for the initial serial
4516 number when a serial number file is created or when a self signed
4517 certificate is created using 'openssl req -x509'. The initial serial
4518 number file is created using 'openssl x509 -next_serial' in CA.pl
4519 rather than being initialized to 1.
4520 [Steve Henson]
4521
4522 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
4523
4524 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4525 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4526 [Joe Orton, Steve Henson]
4527
4528 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
4529 (CVE-2004-0112)
4530 [Joe Orton, Steve Henson]
4531
4532 *) Make it possible to have multiple active certificates with the same
4533 subject in the CA index file. This is done only if the keyword
4534 'unique_subject' is set to 'no' in the main CA section (default
4535 if 'CA_default') of the configuration file. The value is saved
4536 with the database itself in a separate index attribute file,
4537 named like the index file with '.attr' appended to the name.
4538 [Richard Levitte]
4539
4540 *) X509 verify fixes. Disable broken certificate workarounds when
4541 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
4542 keyUsage extension present. Don't accept CRLs with unhandled critical
4543 extensions: since verify currently doesn't process CRL extensions this
4544 rejects a CRL with *any* critical extensions. Add new verify error codes
4545 for these cases.
4546 [Steve Henson]
4547
4548 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
4549 A clarification of RFC2560 will require the use of OCTET STRINGs and
4550 some implementations cannot handle the current raw format. Since OpenSSL
4551 copies and compares OCSP nonces as opaque blobs without any attempt at
4552 parsing them this should not create any compatibility issues.
4553 [Steve Henson]
4554
4555 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
4556 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
4557 this HMAC (and other) operations are several times slower than OpenSSL
4558 < 0.9.7.
4559 [Steve Henson]
4560
4561 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
4562 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
4563
4564 *) Use the correct content when signing type "other".
4565 [Steve Henson]
4566
4567 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
4568
4569 *) Fix various bugs revealed by running the NISCC test suite:
4570
4571 Stop out of bounds reads in the ASN1 code when presented with
4572 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4573
4574 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
4575
4576 If verify callback ignores invalid public key errors don't try to check
4577 certificate signature with the NULL public key.
4578
4579 [Steve Henson]
4580
4581 *) New -ignore_err option in ocsp application to stop the server
4582 exiting on the first error in a request.
4583 [Steve Henson]
4584
4585 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4586 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4587 specifications.
4588 [Steve Henson]
4589
4590 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4591 extra data after the compression methods not only for TLS 1.0
4592 but also for SSL 3.0 (as required by the specification).
4593 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4594
4595 *) Change X509_certificate_type() to mark the key as exported/exportable
4596 when it's 512 *bits* long, not 512 bytes.
4597 [Richard Levitte]
4598
4599 *) Change AES_cbc_encrypt() so it outputs exact multiple of
4600 blocks during encryption.
4601 [Richard Levitte]
4602
4603 *) Various fixes to base64 BIO and non blocking I/O. On write
4604 flushes were not handled properly if the BIO retried. On read
4605 data was not being buffered properly and had various logic bugs.
4606 This also affects blocking I/O when the data being decoded is a
4607 certain size.
4608 [Steve Henson]
4609
4610 *) Various S/MIME bugfixes and compatibility changes:
4611 output correct application/pkcs7 MIME type if
4612 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
4613 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
4614 of files as .eml work). Correctly handle very long lines in MIME
4615 parser.
4616 [Steve Henson]
4617
4618 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
4619
4620 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4621 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4622 a protocol version number mismatch like a decryption error
4623 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4624 [Bodo Moeller]
4625
4626 *) Turn on RSA blinding by default in the default implementation
4627 to avoid a timing attack. Applications that don't want it can call
4628 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4629 They would be ill-advised to do so in most cases.
4630 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4631
4632 *) Change RSA blinding code so that it works when the PRNG is not
4633 seeded (in this case, the secret RSA exponent is abused as
4634 an unpredictable seed -- if it is not unpredictable, there
4635 is no point in blinding anyway). Make RSA blinding thread-safe
4636 by remembering the creator's thread ID in rsa->blinding and
4637 having all other threads use local one-time blinding factors
4638 (this requires more computation than sharing rsa->blinding, but
4639 avoids excessive locking; and if an RSA object is not shared
4640 between threads, blinding will still be very fast).
4641 [Bodo Moeller]
4642
4643 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
4644 ENGINE as defaults for all supported algorithms irrespective of
4645 the 'flags' parameter. 'flags' is now honoured, so applications
4646 should make sure they are passing it correctly.
4647 [Geoff Thorpe]
4648
4649 *) Target "mingw" now allows native Windows code to be generated in
4650 the Cygwin environment as well as with the MinGW compiler.
4651 [Ulf Moeller]
4652
4653 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
4654
4655 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4656 via timing by performing a MAC computation even if incorrect
4657 block cipher padding has been found. This is a countermeasure
4658 against active attacks where the attacker has to distinguish
4659 between bad padding and a MAC verification error. (CVE-2003-0078)
4660
4661 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4662 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4663 Martin Vuagnoux (EPFL, Ilion)]
4664
4665 *) Make the no-err option work as intended. The intention with no-err
4666 is not to have the whole error stack handling routines removed from
4667 libcrypto, it's only intended to remove all the function name and
4668 reason texts, thereby removing some of the footprint that may not
4669 be interesting if those errors aren't displayed anyway.
4670
4671 NOTE: it's still possible for any application or module to have it's
4672 own set of error texts inserted. The routines are there, just not
4673 used by default when no-err is given.
4674 [Richard Levitte]
4675
4676 *) Add support for FreeBSD on IA64.
4677 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
4678
4679 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
4680 Kerberos function mit_des_cbc_cksum(). Before this change,
4681 the value returned by DES_cbc_cksum() was like the one from
4682 mit_des_cbc_cksum(), except the bytes were swapped.
4683 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
4684
4685 *) Allow an application to disable the automatic SSL chain building.
4686 Before this a rather primitive chain build was always performed in
4687 ssl3_output_cert_chain(): an application had no way to send the
4688 correct chain if the automatic operation produced an incorrect result.
4689
4690 Now the chain builder is disabled if either:
4691
4692 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
4693
4694 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
4695
4696 The reasoning behind this is that an application would not want the
4697 auto chain building to take place if extra chain certificates are
4698 present and it might also want a means of sending no additional
4699 certificates (for example the chain has two certificates and the
4700 root is omitted).
4701 [Steve Henson]
4702
4703 *) Add the possibility to build without the ENGINE framework.
4704 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4705
4706 *) Under Win32 gmtime() can return NULL: check return value in
4707 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
4708 [Steve Henson]
4709
4710 *) DSA routines: under certain error conditions uninitialized BN objects
4711 could be freed. Solution: make sure initialization is performed early
4712 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
4713 Nils Larsch <nla@trustcenter.de> via PR#459)
4714 [Lutz Jaenicke]
4715
4716 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
4717 checked on reconnect on the client side, therefore session resumption
4718 could still fail with a "ssl session id is different" error. This
4719 behaviour is masked when SSL_OP_ALL is used due to
4720 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
4721 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4722 followup to PR #377.
4723 [Lutz Jaenicke]
4724
4725 *) IA-32 assembler support enhancements: unified ELF targets, support
4726 for SCO/Caldera platforms, fix for Cygwin shared build.
4727 [Andy Polyakov]
4728
4729 *) Add support for FreeBSD on sparc64. As a consequence, support for
4730 FreeBSD on non-x86 processors is separate from x86 processors on
4731 the config script, much like the NetBSD support.
4732 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
4733
4734 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
4735
4736 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
4737 OpenSSL 0.9.7.]
4738
4739 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
4740 code (06) was taken as the first octet of the session ID and the last
4741 octet was ignored consequently. As a result SSLv2 client side session
4742 caching could not have worked due to the session ID mismatch between
4743 client and server.
4744 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
4745 PR #377.
4746 [Lutz Jaenicke]
4747
4748 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
4749 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
4750 removed entirely.
4751 [Richard Levitte]
4752
4753 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
4754 seems that in spite of existing for more than a year, many application
4755 author have done nothing to provide the necessary callbacks, which
4756 means that this particular engine will not work properly anywhere.
4757 This is a very unfortunate situation which forces us, in the name
4758 of usability, to give the hw_ncipher.c a static lock, which is part
4759 of libcrypto.
4760 NOTE: This is for the 0.9.7 series ONLY. This hack will never
4761 appear in 0.9.8 or later. We EXPECT application authors to have
4762 dealt properly with this when 0.9.8 is released (unless we actually
4763 make such changes in the libcrypto locking code that changes will
4764 have to be made anyway).
4765 [Richard Levitte]
4766
4767 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
4768 octets have been read, EOF or an error occurs. Without this change
4769 some truncated ASN1 structures will not produce an error.
4770 [Steve Henson]
4771
4772 *) Disable Heimdal support, since it hasn't been fully implemented.
4773 Still give the possibility to force the use of Heimdal, but with
4774 warnings and a request that patches get sent to openssl-dev.
4775 [Richard Levitte]
4776
4777 *) Add the VC-CE target, introduce the WINCE sysname, and add
4778 INSTALL.WCE and appropriate conditionals to make it build.
4779 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
4780
4781 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
4782 cygssl-x.y.z.dll, where x, y and z are the major, minor and
4783 edit numbers of the version.
4784 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4785
4786 *) Introduce safe string copy and catenation functions
4787 (BUF_strlcpy() and BUF_strlcat()).
4788 [Ben Laurie (CHATS) and Richard Levitte]
4789
4790 *) Avoid using fixed-size buffers for one-line DNs.
4791 [Ben Laurie (CHATS)]
4792
4793 *) Add BUF_MEM_grow_clean() to avoid information leakage when
4794 resizing buffers containing secrets, and use where appropriate.
4795 [Ben Laurie (CHATS)]
4796
4797 *) Avoid using fixed size buffers for configuration file location.
4798 [Ben Laurie (CHATS)]
4799
4800 *) Avoid filename truncation for various CA files.
4801 [Ben Laurie (CHATS)]
4802
4803 *) Use sizeof in preference to magic numbers.
4804 [Ben Laurie (CHATS)]
4805
4806 *) Avoid filename truncation in cert requests.
4807 [Ben Laurie (CHATS)]
4808
4809 *) Add assertions to check for (supposedly impossible) buffer
4810 overflows.
4811 [Ben Laurie (CHATS)]
4812
4813 *) Don't cache truncated DNS entries in the local cache (this could
4814 potentially lead to a spoofing attack).
4815 [Ben Laurie (CHATS)]
4816
4817 *) Fix various buffers to be large enough for hex/decimal
4818 representations in a platform independent manner.
4819 [Ben Laurie (CHATS)]
4820
4821 *) Add CRYPTO_realloc_clean() to avoid information leakage when
4822 resizing buffers containing secrets, and use where appropriate.
4823 [Ben Laurie (CHATS)]
4824
4825 *) Add BIO_indent() to avoid much slightly worrying code to do
4826 indents.
4827 [Ben Laurie (CHATS)]
4828
4829 *) Convert sprintf()/BIO_puts() to BIO_printf().
4830 [Ben Laurie (CHATS)]
4831
4832 *) buffer_gets() could terminate with the buffer only half
4833 full. Fixed.
4834 [Ben Laurie (CHATS)]
4835
4836 *) Add assertions to prevent user-supplied crypto functions from
4837 overflowing internal buffers by having large block sizes, etc.
4838 [Ben Laurie (CHATS)]
4839
4840 *) New OPENSSL_assert() macro (similar to assert(), but enabled
4841 unconditionally).
4842 [Ben Laurie (CHATS)]
4843
4844 *) Eliminate unused copy of key in RC4.
4845 [Ben Laurie (CHATS)]
4846
4847 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
4848 [Ben Laurie (CHATS)]
4849
4850 *) Fix off-by-one error in EGD path.
4851 [Ben Laurie (CHATS)]
4852
4853 *) If RANDFILE path is too long, ignore instead of truncating.
4854 [Ben Laurie (CHATS)]
4855
4856 *) Eliminate unused and incorrectly sized X.509 structure
4857 CBCParameter.
4858 [Ben Laurie (CHATS)]
4859
4860 *) Eliminate unused and dangerous function knumber().
4861 [Ben Laurie (CHATS)]
4862
4863 *) Eliminate unused and dangerous structure, KSSL_ERR.
4864 [Ben Laurie (CHATS)]
4865
4866 *) Protect against overlong session ID context length in an encoded
4867 session object. Since these are local, this does not appear to be
4868 exploitable.
4869 [Ben Laurie (CHATS)]
4870
4871 *) Change from security patch (see 0.9.6e below) that did not affect
4872 the 0.9.6 release series:
4873
4874 Remote buffer overflow in SSL3 protocol - an attacker could
4875 supply an oversized master key in Kerberos-enabled versions.
4876 (CVE-2002-0657)
4877 [Ben Laurie (CHATS)]
4878
4879 *) Change the SSL kerb5 codes to match RFC 2712.
4880 [Richard Levitte]
4881
4882 *) Make -nameopt work fully for req and add -reqopt switch.
4883 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
4884
4885 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
4886 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
4887
4888 *) Make sure tests can be performed even if the corresponding algorithms
4889 have been removed entirely. This was also the last step to make
4890 OpenSSL compilable with DJGPP under all reasonable conditions.
4891 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
4892
4893 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
4894 to allow version independent disabling of normally unselected ciphers,
4895 which may be activated as a side-effect of selecting a single cipher.
4896
4897 (E.g., cipher list string "RSA" enables ciphersuites that are left
4898 out of "ALL" because they do not provide symmetric encryption.
4899 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
4900 [Lutz Jaenicke, Bodo Moeller]
4901
4902 *) Add appropriate support for separate platform-dependent build
4903 directories. The recommended way to make a platform-dependent
4904 build directory is the following (tested on Linux), maybe with
4905 some local tweaks:
4906
4907 # Place yourself outside of the OpenSSL source tree. In
4908 # this example, the environment variable OPENSSL_SOURCE
4909 # is assumed to contain the absolute OpenSSL source directory.
4910 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
4911 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
4912 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
4913 mkdir -p `dirname $F`
4914 ln -s $OPENSSL_SOURCE/$F $F
4915 done
4916
4917 To be absolutely sure not to disturb the source tree, a "make clean"
4918 is a good thing. If it isn't successfull, don't worry about it,
4919 it probably means the source directory is very clean.
4920 [Richard Levitte]
4921
4922 *) Make sure any ENGINE control commands make local copies of string
4923 pointers passed to them whenever necessary. Otherwise it is possible
4924 the caller may have overwritten (or deallocated) the original string
4925 data when a later ENGINE operation tries to use the stored values.
4926 [Götz Babin-Ebell <babinebell@trustcenter.de>]
4927
4928 *) Improve diagnostics in file reading and command-line digests.
4929 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
4930
4931 *) Add AES modes CFB and OFB to the object database. Correct an
4932 error in AES-CFB decryption.
4933 [Richard Levitte]
4934
4935 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
4936 allows existing EVP_CIPHER_CTX structures to be reused after
4937 calling EVP_*Final(). This behaviour is used by encryption
4938 BIOs and some applications. This has the side effect that
4939 applications must explicitly clean up cipher contexts with
4940 EVP_CIPHER_CTX_cleanup() or they will leak memory.
4941 [Steve Henson]
4942
4943 *) Check the values of dna and dnb in bn_mul_recursive before calling
4944 bn_mul_comba (a non zero value means the a or b arrays do not contain
4945 n2 elements) and fallback to bn_mul_normal if either is not zero.
4946 [Steve Henson]
4947
4948 *) Fix escaping of non-ASCII characters when using the -subj option
4949 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
4950 [Lutz Jaenicke]
4951
4952 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
4953 form for "surname", serialNumber has no short form.
4954 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
4955 therefore remove "mail" short name for "internet 7".
4956 The OID for unique identifiers in X509 certificates is
4957 x500UniqueIdentifier, not uniqueIdentifier.
4958 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
4959 [Lutz Jaenicke]
4960
4961 *) Add an "init" command to the ENGINE config module and auto initialize
4962 ENGINEs. Without any "init" command the ENGINE will be initialized
4963 after all ctrl commands have been executed on it. If init=1 the
4964 ENGINE is initailized at that point (ctrls before that point are run
4965 on the uninitialized ENGINE and after on the initialized one). If
4966 init=0 then the ENGINE will not be iniatialized at all.
4967 [Steve Henson]
4968
4969 *) Fix the 'app_verify_callback' interface so that the user-defined
4970 argument is actually passed to the callback: In the
4971 SSL_CTX_set_cert_verify_callback() prototype, the callback
4972 declaration has been changed from
4973 int (*cb)()
4974 into
4975 int (*cb)(X509_STORE_CTX *,void *);
4976 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
4977 i=s->ctx->app_verify_callback(&ctx)
4978 has been changed into
4979 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
4980
4981 To update applications using SSL_CTX_set_cert_verify_callback(),
4982 a dummy argument can be added to their callback functions.
4983 [D. K. Smetters <smetters@parc.xerox.com>]
4984
4985 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
4986 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
4987
4988 *) Add and OPENSSL_LOAD_CONF define which will cause
4989 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
4990 This allows older applications to transparently support certain
4991 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
4992 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
4993 load the config file and OPENSSL_add_all_algorithms_conf() which will
4994 always load it have also been added.
4995 [Steve Henson]
4996
4997 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
4998 Adjust NIDs and EVP layer.
4999 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5000
5001 *) Config modules support in openssl utility.
5002
5003 Most commands now load modules from the config file,
5004 though in a few (such as version) this isn't done
5005 because it couldn't be used for anything.
5006
5007 In the case of ca and req the config file used is
5008 the same as the utility itself: that is the -config
5009 command line option can be used to specify an
5010 alternative file.
5011 [Steve Henson]
5012
5013 *) Move default behaviour from OPENSSL_config(). If appname is NULL
5014 use "openssl_conf" if filename is NULL use default openssl config file.
5015 [Steve Henson]
5016
5017 *) Add an argument to OPENSSL_config() to allow the use of an alternative
5018 config section name. Add a new flag to tolerate a missing config file
5019 and move code to CONF_modules_load_file().
5020 [Steve Henson]
5021
5022 *) Support for crypto accelerator cards from Accelerated Encryption
5023 Processing, www.aep.ie. (Use engine 'aep')
5024 The support was copied from 0.9.6c [engine] and adapted/corrected
5025 to work with the new engine framework.
5026 [AEP Inc. and Richard Levitte]
5027
5028 *) Support for SureWare crypto accelerator cards from Baltimore
5029 Technologies. (Use engine 'sureware')
5030 The support was copied from 0.9.6c [engine] and adapted
5031 to work with the new engine framework.
5032 [Richard Levitte]
5033
5034 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
5035 make the newer ENGINE framework commands for the CHIL engine work.
5036 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
5037
5038 *) Make it possible to produce shared libraries on ReliantUNIX.
5039 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
5040
5041 *) Add the configuration target debug-linux-ppro.
5042 Make 'openssl rsa' use the general key loading routines
5043 implemented in apps.c, and make those routines able to
5044 handle the key format FORMAT_NETSCAPE and the variant
5045 FORMAT_IISSGC.
5046 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5047
5048 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
5049 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
5050
5051 *) Add -keyform to rsautl, and document -engine.
5052 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
5053
5054 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
5055 BIO_R_NO_SUCH_FILE error code rather than the generic
5056 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
5057 [Ben Laurie]
5058
5059 *) Add new functions
5060 ERR_peek_last_error
5061 ERR_peek_last_error_line
5062 ERR_peek_last_error_line_data.
5063 These are similar to
5064 ERR_peek_error
5065 ERR_peek_error_line
5066 ERR_peek_error_line_data,
5067 but report on the latest error recorded rather than the first one
5068 still in the error queue.
5069 [Ben Laurie, Bodo Moeller]
5070
5071 *) default_algorithms option in ENGINE config module. This allows things
5072 like:
5073 default_algorithms = ALL
5074 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
5075 [Steve Henson]
5076
5077 *) Preliminary ENGINE config module.
5078 [Steve Henson]
5079
5080 *) New experimental application configuration code.
5081 [Steve Henson]
5082
5083 *) Change the AES code to follow the same name structure as all other
5084 symmetric ciphers, and behave the same way. Move everything to
5085 the directory crypto/aes, thereby obsoleting crypto/rijndael.
5086 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
5087
5088 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
5089 [Ben Laurie and Theo de Raadt]
5090
5091 *) Add option to output public keys in req command.
5092 [Massimiliano Pala madwolf@openca.org]
5093
5094 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
5095 (up to about 10% better than before for P-192 and P-224).
5096 [Bodo Moeller]
5097
5098 *) New functions/macros
5099
5100 SSL_CTX_set_msg_callback(ctx, cb)
5101 SSL_CTX_set_msg_callback_arg(ctx, arg)
5102 SSL_set_msg_callback(ssl, cb)
5103 SSL_set_msg_callback_arg(ssl, arg)
5104
5105 to request calling a callback function
5106
5107 void cb(int write_p, int version, int content_type,
5108 const void *buf, size_t len, SSL *ssl, void *arg)
5109
5110 whenever a protocol message has been completely received
5111 (write_p == 0) or sent (write_p == 1). Here 'version' is the
5112 protocol version according to which the SSL library interprets
5113 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
5114 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
5115 the content type as defined in the SSL 3.0/TLS 1.0 protocol
5116 specification (change_cipher_spec(20), alert(21), handshake(22)).
5117 'buf' and 'len' point to the actual message, 'ssl' to the
5118 SSL object, and 'arg' is the application-defined value set by
5119 SSL[_CTX]_set_msg_callback_arg().
5120
5121 'openssl s_client' and 'openssl s_server' have new '-msg' options
5122 to enable a callback that displays all protocol messages.
5123 [Bodo Moeller]
5124
5125 *) Change the shared library support so shared libraries are built as
5126 soon as the corresponding static library is finished, and thereby get
5127 openssl and the test programs linked against the shared library.
5128 This still only happens when the keyword "shard" has been given to
5129 the configuration scripts.
5130
5131 NOTE: shared library support is still an experimental thing, and
5132 backward binary compatibility is still not guaranteed.
5133 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
5134
5135 *) Add support for Subject Information Access extension.
5136 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
5137
5138 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
5139 additional bytes when new memory had to be allocated, not just
5140 when reusing an existing buffer.
5141 [Bodo Moeller]
5142
5143 *) New command line and configuration option 'utf8' for the req command.
5144 This allows field values to be specified as UTF8 strings.
5145 [Steve Henson]
5146
5147 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
5148 runs for the former and machine-readable output for the latter.
5149 [Ben Laurie]
5150
5151 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
5152 of the e-mail address in the DN (i.e., it will go into a certificate
5153 extension only). The new configuration file option 'email_in_dn = no'
5154 has the same effect.
5155 [Massimiliano Pala madwolf@openca.org]
5156
5157 *) Change all functions with names starting with des_ to be starting
5158 with DES_ instead. Add wrappers that are compatible with libdes,
5159 but are named _ossl_old_des_*. Finally, add macros that map the
5160 des_* symbols to the corresponding _ossl_old_des_* if libdes
5161 compatibility is desired. If OpenSSL 0.9.6c compatibility is
5162 desired, the des_* symbols will be mapped to DES_*, with one
5163 exception.
5164
5165 Since we provide two compatibility mappings, the user needs to
5166 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
5167 compatibility is desired. The default (i.e., when that macro
5168 isn't defined) is OpenSSL 0.9.6c compatibility.
5169
5170 There are also macros that enable and disable the support of old
5171 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
5172 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
5173 are defined, the default will apply: to support the old des routines.
5174
5175 In either case, one must include openssl/des.h to get the correct
5176 definitions. Do not try to just include openssl/des_old.h, that
5177 won't work.
5178
5179 NOTE: This is a major break of an old API into a new one. Software
5180 authors are encouraged to switch to the DES_ style functions. Some
5181 time in the future, des_old.h and the libdes compatibility functions
5182 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
5183 default), and then completely removed.
5184 [Richard Levitte]
5185
5186 *) Test for certificates which contain unsupported critical extensions.
5187 If such a certificate is found during a verify operation it is
5188 rejected by default: this behaviour can be overridden by either
5189 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
5190 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
5191 X509_supported_extension() has also been added which returns 1 if a
5192 particular extension is supported.
5193 [Steve Henson]
5194
5195 *) Modify the behaviour of EVP cipher functions in similar way to digests
5196 to retain compatibility with existing code.
5197 [Steve Henson]
5198
5199 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
5200 compatibility with existing code. In particular the 'ctx' parameter does
5201 not have to be to be initialized before the call to EVP_DigestInit() and
5202 it is tidied up after a call to EVP_DigestFinal(). New function
5203 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
5204 EVP_MD_CTX_copy() changed to not require the destination to be
5205 initialized valid and new function EVP_MD_CTX_copy_ex() added which
5206 requires the destination to be valid.
5207
5208 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
5209 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
5210 [Steve Henson]
5211
5212 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
5213 so that complete 'Handshake' protocol structures are kept in memory
5214 instead of overwriting 'msg_type' and 'length' with 'body' data.
5215 [Bodo Moeller]
5216
5217 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
5218 [Massimo Santin via Richard Levitte]
5219
5220 *) Major restructuring to the underlying ENGINE code. This includes
5221 reduction of linker bloat, separation of pure "ENGINE" manipulation
5222 (initialisation, etc) from functionality dealing with implementations
5223 of specific crypto iterfaces. This change also introduces integrated
5224 support for symmetric ciphers and digest implementations - so ENGINEs
5225 can now accelerate these by providing EVP_CIPHER and EVP_MD
5226 implementations of their own. This is detailed in crypto/engine/README
5227 as it couldn't be adequately described here. However, there are a few
5228 API changes worth noting - some RSA, DSA, DH, and RAND functions that
5229 were changed in the original introduction of ENGINE code have now
5230 reverted back - the hooking from this code to ENGINE is now a good
5231 deal more passive and at run-time, operations deal directly with
5232 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
5233 dereferencing through an ENGINE pointer any more. Also, the ENGINE
5234 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
5235 they were not being used by the framework as there is no concept of a
5236 BIGNUM_METHOD and they could not be generalised to the new
5237 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
5238 ENGINE_cpy() has been removed as it cannot be consistently defined in
5239 the new code.
5240 [Geoff Thorpe]
5241
5242 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
5243 [Steve Henson]
5244
5245 *) Change mkdef.pl to sort symbols that get the same entry number,
5246 and make sure the automatically generated functions ERR_load_*
5247 become part of libeay.num as well.
5248 [Richard Levitte]
5249
5250 *) New function SSL_renegotiate_pending(). This returns true once
5251 renegotiation has been requested (either SSL_renegotiate() call
5252 or HelloRequest/ClientHello received from the peer) and becomes
5253 false once a handshake has been completed.
5254 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
5255 sends a HelloRequest, but does not ensure that a handshake takes
5256 place. SSL_renegotiate_pending() is useful for checking if the
5257 client has followed the request.)
5258 [Bodo Moeller]
5259
5260 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
5261 By default, clients may request session resumption even during
5262 renegotiation (if session ID contexts permit); with this option,
5263 session resumption is possible only in the first handshake.
5264
5265 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
5266 more bits available for options that should not be part of
5267 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
5268 [Bodo Moeller]
5269
5270 *) Add some demos for certificate and certificate request creation.
5271 [Steve Henson]
5272
5273 *) Make maximum certificate chain size accepted from the peer application
5274 settable (SSL*_get/set_max_cert_list()), as proposed by
5275 "Douglas E. Engert" <deengert@anl.gov>.
5276 [Lutz Jaenicke]
5277
5278 *) Add support for shared libraries for Unixware-7
5279 (Boyd Lynn Gerber <gerberb@zenez.com>).
5280 [Lutz Jaenicke]
5281
5282 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
5283 be done prior to destruction. Use this to unload error strings from
5284 ENGINEs that load their own error strings. NB: This adds two new API
5285 functions to "get" and "set" this destroy handler in an ENGINE.
5286 [Geoff Thorpe]
5287
5288 *) Alter all existing ENGINE implementations (except "openssl" and
5289 "openbsd") to dynamically instantiate their own error strings. This
5290 makes them more flexible to be built both as statically-linked ENGINEs
5291 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
5292 Also, add stub code to each that makes building them as self-contained
5293 shared-libraries easier (see README.ENGINE).
5294 [Geoff Thorpe]
5295
5296 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
5297 implementations into applications that are completely implemented in
5298 self-contained shared-libraries. The "dynamic" ENGINE exposes control
5299 commands that can be used to configure what shared-library to load and
5300 to control aspects of the way it is handled. Also, made an update to
5301 the README.ENGINE file that brings its information up-to-date and
5302 provides some information and instructions on the "dynamic" ENGINE
5303 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
5304 [Geoff Thorpe]
5305
5306 *) Make it possible to unload ranges of ERR strings with a new
5307 "ERR_unload_strings" function.
5308 [Geoff Thorpe]
5309
5310 *) Add a copy() function to EVP_MD.
5311 [Ben Laurie]
5312
5313 *) Make EVP_MD routines take a context pointer instead of just the
5314 md_data void pointer.
5315 [Ben Laurie]
5316
5317 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
5318 that the digest can only process a single chunk of data
5319 (typically because it is provided by a piece of
5320 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
5321 is only going to provide a single chunk of data, and hence the
5322 framework needn't accumulate the data for oneshot drivers.
5323 [Ben Laurie]
5324
5325 *) As with "ERR", make it possible to replace the underlying "ex_data"
5326 functions. This change also alters the storage and management of global
5327 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
5328 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
5329 index counters. The API functions that use this state have been changed
5330 to take a "class_index" rather than pointers to the class's local STACK
5331 and counter, and there is now an API function to dynamically create new
5332 classes. This centralisation allows us to (a) plug a lot of the
5333 thread-safety problems that existed, and (b) makes it possible to clean
5334 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
5335 such data would previously have always leaked in application code and
5336 workarounds were in place to make the memory debugging turn a blind eye
5337 to it. Application code that doesn't use this new function will still
5338 leak as before, but their memory debugging output will announce it now
5339 rather than letting it slide.
5340
5341 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
5342 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
5343 has a return value to indicate success or failure.
5344 [Geoff Thorpe]
5345
5346 *) Make it possible to replace the underlying "ERR" functions such that the
5347 global state (2 LHASH tables and 2 locks) is only used by the "default"
5348 implementation. This change also adds two functions to "get" and "set"
5349 the implementation prior to it being automatically set the first time
5350 any other ERR function takes place. Ie. an application can call "get",
5351 pass the return value to a module it has just loaded, and that module
5352 can call its own "set" function using that value. This means the
5353 module's "ERR" operations will use (and modify) the error state in the
5354 application and not in its own statically linked copy of OpenSSL code.
5355 [Geoff Thorpe]
5356
5357 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
5358 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
5359 the operation, and provides a more encapsulated way for external code
5360 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
5361 to use these functions rather than manually incrementing the counts.
5362
5363 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
5364 [Geoff Thorpe]
5365
5366 *) Add EVP test program.
5367 [Ben Laurie]
5368
5369 *) Add symmetric cipher support to ENGINE. Expect the API to change!
5370 [Ben Laurie]
5371
5372 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
5373 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
5374 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
5375 These allow a CRL to be built without having to access X509_CRL fields
5376 directly. Modify 'ca' application to use new functions.
5377 [Steve Henson]
5378
5379 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
5380 bug workarounds. Rollback attack detection is a security feature.
5381 The problem will only arise on OpenSSL servers when TLSv1 is not
5382 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
5383 Software authors not wanting to support TLSv1 will have special reasons
5384 for their choice and can explicitly enable this option.
5385 [Bodo Moeller, Lutz Jaenicke]
5386
5387 *) Rationalise EVP so it can be extended: don't include a union of
5388 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
5389 (similar to those existing for EVP_CIPHER_CTX).
5390 Usage example:
5391
5392 EVP_MD_CTX md;
5393
5394 EVP_MD_CTX_init(&md); /* new function call */
5395 EVP_DigestInit(&md, EVP_sha1());
5396 EVP_DigestUpdate(&md, in, len);
5397 EVP_DigestFinal(&md, out, NULL);
5398 EVP_MD_CTX_cleanup(&md); /* new function call */
5399
5400 [Ben Laurie]
5401
5402 *) Make DES key schedule conform to the usual scheme, as well as
5403 correcting its structure. This means that calls to DES functions
5404 now have to pass a pointer to a des_key_schedule instead of a
5405 plain des_key_schedule (which was actually always a pointer
5406 anyway): E.g.,
5407
5408 des_key_schedule ks;
5409
5410 des_set_key_checked(..., &ks);
5411 des_ncbc_encrypt(..., &ks, ...);
5412
5413 (Note that a later change renames 'des_...' into 'DES_...'.)
5414 [Ben Laurie]
5415
5416 *) Initial reduction of linker bloat: the use of some functions, such as
5417 PEM causes large amounts of unused functions to be linked in due to
5418 poor organisation. For example pem_all.c contains every PEM function
5419 which has a knock on effect of linking in large amounts of (unused)
5420 ASN1 code. Grouping together similar functions and splitting unrelated
5421 functions prevents this.
5422 [Steve Henson]
5423
5424 *) Cleanup of EVP macros.
5425 [Ben Laurie]
5426
5427 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
5428 correct _ecb suffix.
5429 [Ben Laurie]
5430
5431 *) Add initial OCSP responder support to ocsp application. The
5432 revocation information is handled using the text based index
5433 use by the ca application. The responder can either handle
5434 requests generated internally, supplied in files (for example
5435 via a CGI script) or using an internal minimal server.
5436 [Steve Henson]
5437
5438 *) Add configuration choices to get zlib compression for TLS.
5439 [Richard Levitte]
5440
5441 *) Changes to Kerberos SSL for RFC 2712 compliance:
5442 1. Implemented real KerberosWrapper, instead of just using
5443 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
5444 2. Implemented optional authenticator field of KerberosWrapper.
5445
5446 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
5447 and authenticator structs; see crypto/krb5/.
5448
5449 Generalized Kerberos calls to support multiple Kerberos libraries.
5450 [Vern Staats <staatsvr@asc.hpc.mil>,
5451 Jeffrey Altman <jaltman@columbia.edu>
5452 via Richard Levitte]
5453
5454 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
5455 already does with RSA. testdsa.h now has 'priv_key/pub_key'
5456 values for each of the key sizes rather than having just
5457 parameters (and 'speed' generating keys each time).
5458 [Geoff Thorpe]
5459
5460 *) Speed up EVP routines.
5461 Before:
5462 encrypt
5463 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
5464 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
5465 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
5466 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
5467 decrypt
5468 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
5469 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
5470 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
5471 After:
5472 encrypt
5473 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
5474 decrypt
5475 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
5476 [Ben Laurie]
5477
5478 *) Added the OS2-EMX target.
5479 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
5480
5481 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
5482 to support NCONF routines in extension code. New function CONF_set_nconf()
5483 to allow functions which take an NCONF to also handle the old LHASH
5484 structure: this means that the old CONF compatible routines can be
5485 retained (in particular wrt extensions) without having to duplicate the
5486 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
5487 [Steve Henson]
5488
5489 *) Enhance the general user interface with mechanisms for inner control
5490 and with possibilities to have yes/no kind of prompts.
5491 [Richard Levitte]
5492
5493 *) Change all calls to low level digest routines in the library and
5494 applications to use EVP. Add missing calls to HMAC_cleanup() and
5495 don't assume HMAC_CTX can be copied using memcpy().
5496 [Verdon Walker <VWalker@novell.com>, Steve Henson]
5497
5498 *) Add the possibility to control engines through control names but with
5499 arbitrary arguments instead of just a string.
5500 Change the key loaders to take a UI_METHOD instead of a callback
5501 function pointer. NOTE: this breaks binary compatibility with earlier
5502 versions of OpenSSL [engine].
5503 Adapt the nCipher code for these new conditions and add a card insertion
5504 callback.
5505 [Richard Levitte]
5506
5507 *) Enhance the general user interface with mechanisms to better support
5508 dialog box interfaces, application-defined prompts, the possibility
5509 to use defaults (for example default passwords from somewhere else)
5510 and interrupts/cancellations.
5511 [Richard Levitte]
5512
5513 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
5514 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
5515 [Steve Henson]
5516
5517 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
5518 tidy up some unnecessarily weird code in 'sk_new()').
5519 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
5520
5521 *) Change the key loading routines for ENGINEs to use the same kind
5522 callback (pem_password_cb) as all other routines that need this
5523 kind of callback.
5524 [Richard Levitte]
5525
5526 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
5527 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
5528 than this minimum value is recommended.
5529 [Lutz Jaenicke]
5530
5531 *) New random seeder for OpenVMS, using the system process statistics
5532 that are easily reachable.
5533 [Richard Levitte]
5534
5535 *) Windows apparently can't transparently handle global
5536 variables defined in DLLs. Initialisations such as:
5537
5538 const ASN1_ITEM *it = &ASN1_INTEGER_it;
5539
5540 wont compile. This is used by the any applications that need to
5541 declare their own ASN1 modules. This was fixed by adding the option
5542 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
5543 needed for static libraries under Win32.
5544 [Steve Henson]
5545
5546 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
5547 setting of purpose and trust fields. New X509_STORE trust and
5548 purpose functions and tidy up setting in other SSL functions.
5549 [Steve Henson]
5550
5551 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
5552 structure. These are inherited by X509_STORE_CTX when it is
5553 initialised. This allows various defaults to be set in the
5554 X509_STORE structure (such as flags for CRL checking and custom
5555 purpose or trust settings) for functions which only use X509_STORE_CTX
5556 internally such as S/MIME.
5557
5558 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
5559 trust settings if they are not set in X509_STORE. This allows X509_STORE
5560 purposes and trust (in S/MIME for example) to override any set by default.
5561
5562 Add command line options for CRL checking to smime, s_client and s_server
5563 applications.
5564 [Steve Henson]
5565
5566 *) Initial CRL based revocation checking. If the CRL checking flag(s)
5567 are set then the CRL is looked up in the X509_STORE structure and
5568 its validity and signature checked, then if the certificate is found
5569 in the CRL the verify fails with a revoked error.
5570
5571 Various new CRL related callbacks added to X509_STORE_CTX structure.
5572
5573 Command line options added to 'verify' application to support this.
5574
5575 This needs some additional work, such as being able to handle multiple
5576 CRLs with different times, extension based lookup (rather than just
5577 by subject name) and ultimately more complete V2 CRL extension
5578 handling.
5579 [Steve Henson]
5580
5581 *) Add a general user interface API (crypto/ui/). This is designed
5582 to replace things like des_read_password and friends (backward
5583 compatibility functions using this new API are provided).
5584 The purpose is to remove prompting functions from the DES code
5585 section as well as provide for prompting through dialog boxes in
5586 a window system and the like.
5587 [Richard Levitte]
5588
5589 *) Add "ex_data" support to ENGINE so implementations can add state at a
5590 per-structure level rather than having to store it globally.
5591 [Geoff]
5592
5593 *) Make it possible for ENGINE structures to be copied when retrieved by
5594 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
5595 This causes the "original" ENGINE structure to act like a template,
5596 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
5597 operational state can be localised to each ENGINE structure, despite the
5598 fact they all share the same "methods". New ENGINE structures returned in
5599 this case have no functional references and the return value is the single
5600 structural reference. This matches the single structural reference returned
5601 by ENGINE_by_id() normally, when it is incremented on the pre-existing
5602 ENGINE structure.
5603 [Geoff]
5604
5605 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
5606 needs to match any other type at all we need to manually clear the
5607 tag cache.
5608 [Steve Henson]
5609
5610 *) Changes to the "openssl engine" utility to include;
5611 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
5612 about an ENGINE's available control commands.
5613 - executing control commands from command line arguments using the
5614 '-pre' and '-post' switches. '-post' is only used if '-t' is
5615 specified and the ENGINE is successfully initialised. The syntax for
5616 the individual commands are colon-separated, for example;
5617 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
5618 [Geoff]
5619
5620 *) New dynamic control command support for ENGINEs. ENGINEs can now
5621 declare their own commands (numbers), names (strings), descriptions,
5622 and input types for run-time discovery by calling applications. A
5623 subset of these commands are implicitly classed as "executable"
5624 depending on their input type, and only these can be invoked through
5625 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
5626 can be based on user input, config files, etc). The distinction is
5627 that "executable" commands cannot return anything other than a boolean
5628 result and can only support numeric or string input, whereas some
5629 discoverable commands may only be for direct use through
5630 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
5631 pointers, or other custom uses. The "executable" commands are to
5632 support parameterisations of ENGINE behaviour that can be
5633 unambiguously defined by ENGINEs and used consistently across any
5634 OpenSSL-based application. Commands have been added to all the
5635 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
5636 control over shared-library paths without source code alterations.
5637 [Geoff]
5638
5639 *) Changed all ENGINE implementations to dynamically allocate their
5640 ENGINEs rather than declaring them statically. Apart from this being
5641 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
5642 this also allows the implementations to compile without using the
5643 internal engine_int.h header.
5644 [Geoff]
5645
5646 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
5647 'const' value. Any code that should be able to modify a RAND_METHOD
5648 should already have non-const pointers to it (ie. they should only
5649 modify their own ones).
5650 [Geoff]
5651
5652 *) Made a variety of little tweaks to the ENGINE code.
5653 - "atalla" and "ubsec" string definitions were moved from header files
5654 to C code. "nuron" string definitions were placed in variables
5655 rather than hard-coded - allowing parameterisation of these values
5656 later on via ctrl() commands.
5657 - Removed unused "#if 0"'d code.
5658 - Fixed engine list iteration code so it uses ENGINE_free() to release
5659 structural references.
5660 - Constified the RAND_METHOD element of ENGINE structures.
5661 - Constified various get/set functions as appropriate and added
5662 missing functions (including a catch-all ENGINE_cpy that duplicates
5663 all ENGINE values onto a new ENGINE except reference counts/state).
5664 - Removed NULL parameter checks in get/set functions. Setting a method
5665 or function to NULL is a way of cancelling out a previously set
5666 value. Passing a NULL ENGINE parameter is just plain stupid anyway
5667 and doesn't justify the extra error symbols and code.
5668 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
5669 flags from engine_int.h to engine.h.
5670 - Changed prototypes for ENGINE handler functions (init(), finish(),
5671 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
5672 [Geoff]
5673
5674 *) Implement binary inversion algorithm for BN_mod_inverse in addition
5675 to the algorithm using long division. The binary algorithm can be
5676 used only if the modulus is odd. On 32-bit systems, it is faster
5677 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
5678 roughly 5-15% for 256-bit moduli), so we use it only for moduli
5679 up to 450 bits. In 64-bit environments, the binary algorithm
5680 appears to be advantageous for much longer moduli; here we use it
5681 for moduli up to 2048 bits.
5682 [Bodo Moeller]
5683
5684 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
5685 could not support the combine flag in choice fields.
5686 [Steve Henson]
5687
5688 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
5689 extensions from a certificate request to the certificate.
5690 [Steve Henson]
5691
5692 *) Allow multiple 'certopt' and 'nameopt' options to be separated
5693 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
5694 file: this allows the display of the certificate about to be
5695 signed to be customised, to allow certain fields to be included
5696 or excluded and extension details. The old system didn't display
5697 multicharacter strings properly, omitted fields not in the policy
5698 and couldn't display additional details such as extensions.
5699 [Steve Henson]
5700
5701 *) Function EC_POINTs_mul for multiple scalar multiplication
5702 of an arbitrary number of elliptic curve points
5703 \sum scalars[i]*points[i],
5704 optionally including the generator defined for the EC_GROUP:
5705 scalar*generator + \sum scalars[i]*points[i].
5706
5707 EC_POINT_mul is a simple wrapper function for the typical case
5708 that the point list has just one item (besides the optional
5709 generator).
5710 [Bodo Moeller]
5711
5712 *) First EC_METHODs for curves over GF(p):
5713
5714 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
5715 operations and provides various method functions that can also
5716 operate with faster implementations of modular arithmetic.
5717
5718 EC_GFp_mont_method() reuses most functions that are part of
5719 EC_GFp_simple_method, but uses Montgomery arithmetic.
5720
5721 [Bodo Moeller; point addition and point doubling
5722 implementation directly derived from source code provided by
5723 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
5724
5725 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
5726 crypto/ec/ec_lib.c):
5727
5728 Curves are EC_GROUP objects (with an optional group generator)
5729 based on EC_METHODs that are built into the library.
5730
5731 Points are EC_POINT objects based on EC_GROUP objects.
5732
5733 Most of the framework would be able to handle curves over arbitrary
5734 finite fields, but as there are no obvious types for fields other
5735 than GF(p), some functions are limited to that for now.
5736 [Bodo Moeller]
5737
5738 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
5739 that the file contains a complete HTTP response.
5740 [Richard Levitte]
5741
5742 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
5743 change the def and num file printf format specifier from "%-40sXXX"
5744 to "%-39s XXX". The latter will always guarantee a space after the
5745 field while the former will cause them to run together if the field
5746 is 40 of more characters long.
5747 [Steve Henson]
5748
5749 *) Constify the cipher and digest 'method' functions and structures
5750 and modify related functions to take constant EVP_MD and EVP_CIPHER
5751 pointers.
5752 [Steve Henson]
5753
5754 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
5755 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
5756 [Bodo Moeller]
5757
5758 *) Modify EVP_Digest*() routines so they now return values. Although the
5759 internal software routines can never fail additional hardware versions
5760 might.
5761 [Steve Henson]
5762
5763 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
5764
5765 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
5766 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
5767
5768 ASN1 error codes
5769 ERR_R_NESTED_ASN1_ERROR
5770 ...
5771 ERR_R_MISSING_ASN1_EOS
5772 were 4 .. 9, conflicting with
5773 ERR_LIB_RSA (= ERR_R_RSA_LIB)
5774 ...
5775 ERR_LIB_PEM (= ERR_R_PEM_LIB).
5776 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
5777
5778 Add new error code 'ERR_R_INTERNAL_ERROR'.
5779 [Bodo Moeller]
5780
5781 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
5782 suffices.
5783 [Bodo Moeller]
5784
5785 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
5786 sets the subject name for a new request or supersedes the
5787 subject name in a given request. Formats that can be parsed are
5788 'CN=Some Name, OU=myOU, C=IT'
5789 and
5790 'CN=Some Name/OU=myOU/C=IT'.
5791
5792 Add options '-batch' and '-verbose' to 'openssl req'.
5793 [Massimiliano Pala <madwolf@hackmasters.net>]
5794
5795 *) Introduce the possibility to access global variables through
5796 functions on platform were that's the best way to handle exporting
5797 global variables in shared libraries. To enable this functionality,
5798 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
5799 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
5800 is normally done by Configure or something similar).
5801
5802 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
5803 in the source file (foo.c) like this:
5804
5805 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
5806 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
5807
5808 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
5809 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
5810
5811 OPENSSL_DECLARE_GLOBAL(int,foo);
5812 #define foo OPENSSL_GLOBAL_REF(foo)
5813 OPENSSL_DECLARE_GLOBAL(double,bar);
5814 #define bar OPENSSL_GLOBAL_REF(bar)
5815
5816 The #defines are very important, and therefore so is including the
5817 header file everywhere where the defined globals are used.
5818
5819 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
5820 of ASN.1 items, but that structure is a bit different.
5821
5822 The largest change is in util/mkdef.pl which has been enhanced with
5823 better and easier to understand logic to choose which symbols should
5824 go into the Windows .def files as well as a number of fixes and code
5825 cleanup (among others, algorithm keywords are now sorted
5826 lexicographically to avoid constant rewrites).
5827 [Richard Levitte]
5828
5829 *) In BN_div() keep a copy of the sign of 'num' before writing the
5830 result to 'rm' because if rm==num the value will be overwritten
5831 and produce the wrong result if 'num' is negative: this caused
5832 problems with BN_mod() and BN_nnmod().
5833 [Steve Henson]
5834
5835 *) Function OCSP_request_verify(). This checks the signature on an
5836 OCSP request and verifies the signer certificate. The signer
5837 certificate is just checked for a generic purpose and OCSP request
5838 trust settings.
5839 [Steve Henson]
5840
5841 *) Add OCSP_check_validity() function to check the validity of OCSP
5842 responses. OCSP responses are prepared in real time and may only
5843 be a few seconds old. Simply checking that the current time lies
5844 between thisUpdate and nextUpdate max reject otherwise valid responses
5845 caused by either OCSP responder or client clock inaccuracy. Instead
5846 we allow thisUpdate and nextUpdate to fall within a certain period of
5847 the current time. The age of the response can also optionally be
5848 checked. Two new options -validity_period and -status_age added to
5849 ocsp utility.
5850 [Steve Henson]
5851
5852 *) If signature or public key algorithm is unrecognized print out its
5853 OID rather that just UNKNOWN.
5854 [Steve Henson]
5855
5856 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
5857 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
5858 ID to be generated from the issuer certificate alone which can then be
5859 passed to OCSP_id_issuer_cmp().
5860 [Steve Henson]
5861
5862 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
5863 ASN1 modules to export functions returning ASN1_ITEM pointers
5864 instead of the ASN1_ITEM structures themselves. This adds several
5865 new macros which allow the underlying ASN1 function/structure to
5866 be accessed transparently. As a result code should not use ASN1_ITEM
5867 references directly (such as &X509_it) but instead use the relevant
5868 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
5869 use of the new ASN1 code on platforms where exporting structures
5870 is problematical (for example in shared libraries) but exporting
5871 functions returning pointers to structures is not.
5872 [Steve Henson]
5873
5874 *) Add support for overriding the generation of SSL/TLS session IDs.
5875 These callbacks can be registered either in an SSL_CTX or per SSL.
5876 The purpose of this is to allow applications to control, if they wish,
5877 the arbitrary values chosen for use as session IDs, particularly as it
5878 can be useful for session caching in multiple-server environments. A
5879 command-line switch for testing this (and any client code that wishes
5880 to use such a feature) has been added to "s_server".
5881 [Geoff Thorpe, Lutz Jaenicke]
5882
5883 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
5884 of the form '#if defined(...) || defined(...) || ...' and
5885 '#if !defined(...) && !defined(...) && ...'. This also avoids
5886 the growing number of special cases it was previously handling.
5887 [Richard Levitte]
5888
5889 *) Make all configuration macros available for application by making
5890 sure they are available in opensslconf.h, by giving them names starting
5891 with "OPENSSL_" to avoid conflicts with other packages and by making
5892 sure e_os2.h will cover all platform-specific cases together with
5893 opensslconf.h.
5894 Additionally, it is now possible to define configuration/platform-
5895 specific names (called "system identities"). In the C code, these
5896 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
5897 macro with the name beginning with "OPENSSL_SYS_", which is determined
5898 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
5899 what is available.
5900 [Richard Levitte]
5901
5902 *) New option -set_serial to 'req' and 'x509' this allows the serial
5903 number to use to be specified on the command line. Previously self
5904 signed certificates were hard coded with serial number 0 and the
5905 CA options of 'x509' had to use a serial number in a file which was
5906 auto incremented.
5907 [Steve Henson]
5908
5909 *) New options to 'ca' utility to support V2 CRL entry extensions.
5910 Currently CRL reason, invalidity date and hold instruction are
5911 supported. Add new CRL extensions to V3 code and some new objects.
5912 [Steve Henson]
5913
5914 *) New function EVP_CIPHER_CTX_set_padding() this is used to
5915 disable standard block padding (aka PKCS#5 padding) in the EVP
5916 API, which was previously mandatory. This means that the data is
5917 not padded in any way and so the total length much be a multiple
5918 of the block size, otherwise an error occurs.
5919 [Steve Henson]
5920
5921 *) Initial (incomplete) OCSP SSL support.
5922 [Steve Henson]
5923
5924 *) New function OCSP_parse_url(). This splits up a URL into its host,
5925 port and path components: primarily to parse OCSP URLs. New -url
5926 option to ocsp utility.
5927 [Steve Henson]
5928
5929 *) New nonce behavior. The return value of OCSP_check_nonce() now
5930 reflects the various checks performed. Applications can decide
5931 whether to tolerate certain situations such as an absent nonce
5932 in a response when one was present in a request: the ocsp application
5933 just prints out a warning. New function OCSP_add1_basic_nonce()
5934 this is to allow responders to include a nonce in a response even if
5935 the request is nonce-less.
5936 [Steve Henson]
5937
5938 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
5939 skipped when using openssl x509 multiple times on a single input file,
5940 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
5941 [Bodo Moeller]
5942
5943 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
5944 set string type: to handle setting ASN1_TIME structures. Fix ca
5945 utility to correctly initialize revocation date of CRLs.
5946 [Steve Henson]
5947
5948 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
5949 the clients preferred ciphersuites and rather use its own preferences.
5950 Should help to work around M$ SGC (Server Gated Cryptography) bug in
5951 Internet Explorer by ensuring unchanged hash method during stepup.
5952 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
5953 [Lutz Jaenicke]
5954
5955 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
5956 to aes and add a new 'exist' option to print out symbols that don't
5957 appear to exist.
5958 [Steve Henson]
5959
5960 *) Additional options to ocsp utility to allow flags to be set and
5961 additional certificates supplied.
5962 [Steve Henson]
5963
5964 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
5965 OCSP client a number of certificate to only verify the response
5966 signature against.
5967 [Richard Levitte]
5968
5969 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
5970 handle the new API. Currently only ECB, CBC modes supported. Add new
5971 AES OIDs.
5972
5973 Add TLS AES ciphersuites as described in RFC3268, "Advanced
5974 Encryption Standard (AES) Ciphersuites for Transport Layer
5975 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
5976 not enabled by default and were not part of the "ALL" ciphersuite
5977 alias because they were not yet official; they could be
5978 explicitly requested by specifying the "AESdraft" ciphersuite
5979 group alias. In the final release of OpenSSL 0.9.7, the group
5980 alias is called "AES" and is part of "ALL".)
5981 [Ben Laurie, Steve Henson, Bodo Moeller]
5982
5983 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
5984 request to response.
5985 [Steve Henson]
5986
5987 *) Functions for OCSP responders. OCSP_request_onereq_count(),
5988 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
5989 extract information from a certificate request. OCSP_response_create()
5990 creates a response and optionally adds a basic response structure.
5991 OCSP_basic_add1_status() adds a complete single response to a basic
5992 response and returns the OCSP_SINGLERESP structure just added (to allow
5993 extensions to be included for example). OCSP_basic_add1_cert() adds a
5994 certificate to a basic response and OCSP_basic_sign() signs a basic
5995 response with various flags. New helper functions ASN1_TIME_check()
5996 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
5997 (converts ASN1_TIME to GeneralizedTime).
5998 [Steve Henson]
5999
6000 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
6001 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
6002 structure from a certificate. X509_pubkey_digest() digests the public_key
6003 contents: this is used in various key identifiers.
6004 [Steve Henson]
6005
6006 *) Make sk_sort() tolerate a NULL argument.
6007 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
6008
6009 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
6010 passed by the function are trusted implicitly. If any of them signed the
6011 response then it is assumed to be valid and is not verified.
6012 [Steve Henson]
6013
6014 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
6015 to data. This was previously part of the PKCS7 ASN1 code. This
6016 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
6017 [Steve Henson, reported by Kenneth R. Robinette
6018 <support@securenetterm.com>]
6019
6020 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
6021 routines: without these tracing memory leaks is very painful.
6022 Fix leaks in PKCS12 and PKCS7 routines.
6023 [Steve Henson]
6024
6025 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
6026 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
6027 effectively meant GeneralizedTime would never be used. Now it
6028 is initialised to -1 but X509_time_adj() now has to check the value
6029 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
6030 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
6031 [Steve Henson, reported by Kenneth R. Robinette
6032 <support@securenetterm.com>]
6033
6034 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
6035 result in a zero length in the ASN1_INTEGER structure which was
6036 not consistent with the structure when d2i_ASN1_INTEGER() was used
6037 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
6038 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
6039 where it did not print out a minus for negative ASN1_INTEGER.
6040 [Steve Henson]
6041
6042 *) Add summary printout to ocsp utility. The various functions which
6043 convert status values to strings have been renamed to:
6044 OCSP_response_status_str(), OCSP_cert_status_str() and
6045 OCSP_crl_reason_str() and are no longer static. New options
6046 to verify nonce values and to disable verification. OCSP response
6047 printout format cleaned up.
6048 [Steve Henson]
6049
6050 *) Add additional OCSP certificate checks. These are those specified
6051 in RFC2560. This consists of two separate checks: the CA of the
6052 certificate being checked must either be the OCSP signer certificate
6053 or the issuer of the OCSP signer certificate. In the latter case the
6054 OCSP signer certificate must contain the OCSP signing extended key
6055 usage. This check is performed by attempting to match the OCSP
6056 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
6057 in the OCSP_CERTID structures of the response.
6058 [Steve Henson]
6059
6060 *) Initial OCSP certificate verification added to OCSP_basic_verify()
6061 and related routines. This uses the standard OpenSSL certificate
6062 verify routines to perform initial checks (just CA validity) and
6063 to obtain the certificate chain. Then additional checks will be
6064 performed on the chain. Currently the root CA is checked to see
6065 if it is explicitly trusted for OCSP signing. This is used to set
6066 a root CA as a global signing root: that is any certificate that
6067 chains to that CA is an acceptable OCSP signing certificate.
6068 [Steve Henson]
6069
6070 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
6071 extensions from a separate configuration file.
6072 As when reading extensions from the main configuration file,
6073 the '-extensions ...' option may be used for specifying the
6074 section to use.
6075 [Massimiliano Pala <madwolf@comune.modena.it>]
6076
6077 *) New OCSP utility. Allows OCSP requests to be generated or
6078 read. The request can be sent to a responder and the output
6079 parsed, outputed or printed in text form. Not complete yet:
6080 still needs to check the OCSP response validity.
6081 [Steve Henson]
6082
6083 *) New subcommands for 'openssl ca':
6084 'openssl ca -status <serial>' prints the status of the cert with
6085 the given serial number (according to the index file).
6086 'openssl ca -updatedb' updates the expiry status of certificates
6087 in the index file.
6088 [Massimiliano Pala <madwolf@comune.modena.it>]
6089
6090 *) New '-newreq-nodes' command option to CA.pl. This is like
6091 '-newreq', but calls 'openssl req' with the '-nodes' option
6092 so that the resulting key is not encrypted.
6093 [Damien Miller <djm@mindrot.org>]
6094
6095 *) New configuration for the GNU Hurd.
6096 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
6097
6098 *) Initial code to implement OCSP basic response verify. This
6099 is currently incomplete. Currently just finds the signer's
6100 certificate and verifies the signature on the response.
6101 [Steve Henson]
6102
6103 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
6104 value of OPENSSLDIR. This is available via the new '-d' option
6105 to 'openssl version', and is also included in 'openssl version -a'.
6106 [Bodo Moeller]
6107
6108 *) Allowing defining memory allocation callbacks that will be given
6109 file name and line number information in additional arguments
6110 (a const char* and an int). The basic functionality remains, as
6111 well as the original possibility to just replace malloc(),
6112 realloc() and free() by functions that do not know about these
6113 additional arguments. To register and find out the current
6114 settings for extended allocation functions, the following
6115 functions are provided:
6116
6117 CRYPTO_set_mem_ex_functions
6118 CRYPTO_set_locked_mem_ex_functions
6119 CRYPTO_get_mem_ex_functions
6120 CRYPTO_get_locked_mem_ex_functions
6121
6122 These work the same way as CRYPTO_set_mem_functions and friends.
6123 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
6124 extended allocation function is enabled.
6125 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
6126 a conventional allocation function is enabled.
6127 [Richard Levitte, Bodo Moeller]
6128
6129 *) Finish off removing the remaining LHASH function pointer casts.
6130 There should no longer be any prototype-casting required when using
6131 the LHASH abstraction, and any casts that remain are "bugs". See
6132 the callback types and macros at the head of lhash.h for details
6133 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
6134 [Geoff Thorpe]
6135
6136 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
6137 If /dev/[u]random devices are not available or do not return enough
6138 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
6139 be queried.
6140 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
6141 /etc/entropy will be queried once each in this sequence, quering stops
6142 when enough entropy was collected without querying more sockets.
6143 [Lutz Jaenicke]
6144
6145 *) Change the Unix RAND_poll() variant to be able to poll several
6146 random devices, as specified by DEVRANDOM, until a sufficient amount
6147 of data has been collected. We spend at most 10 ms on each file
6148 (select timeout) and read in non-blocking mode. DEVRANDOM now
6149 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
6150 (previously it was just the string "/dev/urandom"), so on typical
6151 platforms the 10 ms delay will never occur.
6152 Also separate out the Unix variant to its own file, rand_unix.c.
6153 For VMS, there's a currently-empty rand_vms.c.
6154 [Richard Levitte]
6155
6156 *) Move OCSP client related routines to ocsp_cl.c. These
6157 provide utility functions which an application needing
6158 to issue a request to an OCSP responder and analyse the
6159 response will typically need: as opposed to those which an
6160 OCSP responder itself would need which will be added later.
6161
6162 OCSP_request_sign() signs an OCSP request with an API similar
6163 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
6164 response. OCSP_response_get1_basic() extracts basic response
6165 from response. OCSP_resp_find_status(): finds and extracts status
6166 information from an OCSP_CERTID structure (which will be created
6167 when the request structure is built). These are built from lower
6168 level functions which work on OCSP_SINGLERESP structures but
6169 wont normally be used unless the application wishes to examine
6170 extensions in the OCSP response for example.
6171
6172 Replace nonce routines with a pair of functions.
6173 OCSP_request_add1_nonce() adds a nonce value and optionally
6174 generates a random value. OCSP_check_nonce() checks the
6175 validity of the nonce in an OCSP response.
6176 [Steve Henson]
6177
6178 *) Change function OCSP_request_add() to OCSP_request_add0_id().
6179 This doesn't copy the supplied OCSP_CERTID and avoids the
6180 need to free up the newly created id. Change return type
6181 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
6182 This can then be used to add extensions to the request.
6183 Deleted OCSP_request_new(), since most of its functionality
6184 is now in OCSP_REQUEST_new() (and the case insensitive name
6185 clash) apart from the ability to set the request name which
6186 will be added elsewhere.
6187 [Steve Henson]
6188
6189 *) Update OCSP API. Remove obsolete extensions argument from
6190 various functions. Extensions are now handled using the new
6191 OCSP extension code. New simple OCSP HTTP function which
6192 can be used to send requests and parse the response.
6193 [Steve Henson]
6194
6195 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
6196 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
6197 uses the special reorder version of SET OF to sort the attributes
6198 and reorder them to match the encoded order. This resolves a long
6199 standing problem: a verify on a PKCS7 structure just after signing
6200 it used to fail because the attribute order did not match the
6201 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
6202 it uses the received order. This is necessary to tolerate some broken
6203 software that does not order SET OF. This is handled by encoding
6204 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
6205 to produce the required SET OF.
6206 [Steve Henson]
6207
6208 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
6209 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
6210 files to get correct declarations of the ASN.1 item variables.
6211 [Richard Levitte]
6212
6213 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
6214 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
6215 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
6216 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
6217 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
6218 ASN1_ITEM and no wrapper functions.
6219 [Steve Henson]
6220
6221 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
6222 replace the old function pointer based I/O routines. Change most of
6223 the *_d2i_bio() and *_d2i_fp() functions to use these.
6224 [Steve Henson]
6225
6226 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
6227 lines, recognice more "algorithms" that can be deselected, and make
6228 it complain about algorithm deselection that isn't recognised.
6229 [Richard Levitte]
6230
6231 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
6232 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
6233 to use new functions. Add NO_ASN1_OLD which can be set to remove
6234 some old style ASN1 functions: this can be used to determine if old
6235 code will still work when these eventually go away.
6236 [Steve Henson]
6237
6238 *) New extension functions for OCSP structures, these follow the
6239 same conventions as certificates and CRLs.
6240 [Steve Henson]
6241
6242 *) New function X509V3_add1_i2d(). This automatically encodes and
6243 adds an extension. Its behaviour can be customised with various
6244 flags to append, replace or delete. Various wrappers added for
6245 certifcates and CRLs.
6246 [Steve Henson]
6247
6248 *) Fix to avoid calling the underlying ASN1 print routine when
6249 an extension cannot be parsed. Correct a typo in the
6250 OCSP_SERVICELOC extension. Tidy up print OCSP format.
6251 [Steve Henson]
6252
6253 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
6254 entries for variables.
6255 [Steve Henson]
6256
6257 *) Add functionality to apps/openssl.c for detecting locking
6258 problems: As the program is single-threaded, all we have
6259 to do is register a locking callback using an array for
6260 storing which locks are currently held by the program.
6261 [Bodo Moeller]
6262
6263 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
6264 SSL_get_ex_data_X509_STORE_idx(), which is used in
6265 ssl_verify_cert_chain() and thus can be called at any time
6266 during TLS/SSL handshakes so that thread-safety is essential.
6267 Unfortunately, the ex_data design is not at all suited
6268 for multi-threaded use, so it probably should be abolished.
6269 [Bodo Moeller]
6270
6271 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
6272 [Broadcom, tweaked and integrated by Geoff Thorpe]
6273
6274 *) Move common extension printing code to new function
6275 X509V3_print_extensions(). Reorganise OCSP print routines and
6276 implement some needed OCSP ASN1 functions. Add OCSP extensions.
6277 [Steve Henson]
6278
6279 *) New function X509_signature_print() to remove duplication in some
6280 print routines.
6281 [Steve Henson]
6282
6283 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
6284 set (this was treated exactly the same as SET OF previously). This
6285 is used to reorder the STACK representing the structure to match the
6286 encoding. This will be used to get round a problem where a PKCS7
6287 structure which was signed could not be verified because the STACK
6288 order did not reflect the encoded order.
6289 [Steve Henson]
6290
6291 *) Reimplement the OCSP ASN1 module using the new code.
6292 [Steve Henson]
6293
6294 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
6295 for its ASN1 operations. The old style function pointers still exist
6296 for now but they will eventually go away.
6297 [Steve Henson]
6298
6299 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
6300 completely replaces the old ASN1 functionality with a table driven
6301 encoder and decoder which interprets an ASN1_ITEM structure describing
6302 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
6303 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
6304 has also been converted to the new form.
6305 [Steve Henson]
6306
6307 *) Change BN_mod_exp_recp so that negative moduli are tolerated
6308 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
6309 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
6310 for negative moduli.
6311 [Bodo Moeller]
6312
6313 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
6314 of not touching the result's sign bit.
6315 [Bodo Moeller]
6316
6317 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
6318 set.
6319 [Bodo Moeller]
6320
6321 *) Changed the LHASH code to use prototypes for callbacks, and created
6322 macros to declare and implement thin (optionally static) functions
6323 that provide type-safety and avoid function pointer casting for the
6324 type-specific callbacks.
6325 [Geoff Thorpe]
6326
6327 *) Added Kerberos Cipher Suites to be used with TLS, as written in
6328 RFC 2712.
6329 [Veers Staats <staatsvr@asc.hpc.mil>,
6330 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
6331
6332 *) Reformat the FAQ so the different questions and answers can be divided
6333 in sections depending on the subject.
6334 [Richard Levitte]
6335
6336 *) Have the zlib compression code load ZLIB.DLL dynamically under
6337 Windows.
6338 [Richard Levitte]
6339
6340 *) New function BN_mod_sqrt for computing square roots modulo a prime
6341 (using the probabilistic Tonelli-Shanks algorithm unless
6342 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
6343 be handled deterministically).
6344 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6345
6346 *) Make BN_mod_inverse faster by explicitly handling small quotients
6347 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
6348 512 bits], about 30% for larger ones [1024 or 2048 bits].)
6349 [Bodo Moeller]
6350
6351 *) New function BN_kronecker.
6352 [Bodo Moeller]
6353
6354 *) Fix BN_gcd so that it works on negative inputs; the result is
6355 positive unless both parameters are zero.
6356 Previously something reasonably close to an infinite loop was
6357 possible because numbers could be growing instead of shrinking
6358 in the implementation of Euclid's algorithm.
6359 [Bodo Moeller]
6360
6361 *) Fix BN_is_word() and BN_is_one() macros to take into account the
6362 sign of the number in question.
6363
6364 Fix BN_is_word(a,w) to work correctly for w == 0.
6365
6366 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
6367 because its test if the absolute value of 'a' equals 'w'.
6368 Note that BN_abs_is_word does *not* handle w == 0 reliably;
6369 it exists mostly for use in the implementations of BN_is_zero(),
6370 BN_is_one(), and BN_is_word().
6371 [Bodo Moeller]
6372
6373 *) New function BN_swap.
6374 [Bodo Moeller]
6375
6376 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
6377 the exponentiation functions are more likely to produce reasonable
6378 results on negative inputs.
6379 [Bodo Moeller]
6380
6381 *) Change BN_mod_mul so that the result is always non-negative.
6382 Previously, it could be negative if one of the factors was negative;
6383 I don't think anyone really wanted that behaviour.
6384 [Bodo Moeller]
6385
6386 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
6387 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
6388 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
6389 and add new functions:
6390
6391 BN_nnmod
6392 BN_mod_sqr
6393 BN_mod_add
6394 BN_mod_add_quick
6395 BN_mod_sub
6396 BN_mod_sub_quick
6397 BN_mod_lshift1
6398 BN_mod_lshift1_quick
6399 BN_mod_lshift
6400 BN_mod_lshift_quick
6401
6402 These functions always generate non-negative results.
6403
6404 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
6405 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
6406
6407 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
6408 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
6409 be reduced modulo m.
6410 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
6411
6412 #if 0
6413 The following entry accidentally appeared in the CHANGES file
6414 distributed with OpenSSL 0.9.7. The modifications described in
6415 it do *not* apply to OpenSSL 0.9.7.
6416
6417 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6418 was actually never needed) and in BN_mul(). The removal in BN_mul()
6419 required a small change in bn_mul_part_recursive() and the addition
6420 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6421 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6422 bn_sub_words() and bn_add_words() except they take arrays with
6423 differing sizes.
6424 [Richard Levitte]
6425 #endif
6426
6427 *) In 'openssl passwd', verify passwords read from the terminal
6428 unless the '-salt' option is used (which usually means that
6429 verification would just waste user's time since the resulting
6430 hash is going to be compared with some given password hash)
6431 or the new '-noverify' option is used.
6432
6433 This is an incompatible change, but it does not affect
6434 non-interactive use of 'openssl passwd' (passwords on the command
6435 line, '-stdin' option, '-in ...' option) and thus should not
6436 cause any problems.
6437 [Bodo Moeller]
6438
6439 *) Remove all references to RSAref, since there's no more need for it.
6440 [Richard Levitte]
6441
6442 *) Make DSO load along a path given through an environment variable
6443 (SHLIB_PATH) with shl_load().
6444 [Richard Levitte]
6445
6446 *) Constify the ENGINE code as a result of BIGNUM constification.
6447 Also constify the RSA code and most things related to it. In a
6448 few places, most notable in the depth of the ASN.1 code, ugly
6449 casts back to non-const were required (to be solved at a later
6450 time)
6451 [Richard Levitte]
6452
6453 *) Make it so the openssl application has all engines loaded by default.
6454 [Richard Levitte]
6455
6456 *) Constify the BIGNUM routines a little more.
6457 [Richard Levitte]
6458
6459 *) Add the following functions:
6460
6461 ENGINE_load_cswift()
6462 ENGINE_load_chil()
6463 ENGINE_load_atalla()
6464 ENGINE_load_nuron()
6465 ENGINE_load_builtin_engines()
6466
6467 That way, an application can itself choose if external engines that
6468 are built-in in OpenSSL shall ever be used or not. The benefit is
6469 that applications won't have to be linked with libdl or other dso
6470 libraries unless it's really needed.
6471
6472 Changed 'openssl engine' to load all engines on demand.
6473 Changed the engine header files to avoid the duplication of some
6474 declarations (they differed!).
6475 [Richard Levitte]
6476
6477 *) 'openssl engine' can now list capabilities.
6478 [Richard Levitte]
6479
6480 *) Better error reporting in 'openssl engine'.
6481 [Richard Levitte]
6482
6483 *) Never call load_dh_param(NULL) in s_server.
6484 [Bodo Moeller]
6485
6486 *) Add engine application. It can currently list engines by name and
6487 identity, and test if they are actually available.
6488 [Richard Levitte]
6489
6490 *) Improve RPM specification file by forcing symbolic linking and making
6491 sure the installed documentation is also owned by root.root.
6492 [Damien Miller <djm@mindrot.org>]
6493
6494 *) Give the OpenSSL applications more possibilities to make use of
6495 keys (public as well as private) handled by engines.
6496 [Richard Levitte]
6497
6498 *) Add OCSP code that comes from CertCo.
6499 [Richard Levitte]
6500
6501 *) Add VMS support for the Rijndael code.
6502 [Richard Levitte]
6503
6504 *) Added untested support for Nuron crypto accelerator.
6505 [Ben Laurie]
6506
6507 *) Add support for external cryptographic devices. This code was
6508 previously distributed separately as the "engine" branch.
6509 [Geoff Thorpe, Richard Levitte]
6510
6511 *) Rework the filename-translation in the DSO code. It is now possible to
6512 have far greater control over how a "name" is turned into a filename
6513 depending on the operating environment and any oddities about the
6514 different shared library filenames on each system.
6515 [Geoff Thorpe]
6516
6517 *) Support threads on FreeBSD-elf in Configure.
6518 [Richard Levitte]
6519
6520 *) Fix for SHA1 assembly problem with MASM: it produces
6521 warnings about corrupt line number information when assembling
6522 with debugging information. This is caused by the overlapping
6523 of two sections.
6524 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
6525
6526 *) NCONF changes.
6527 NCONF_get_number() has no error checking at all. As a replacement,
6528 NCONF_get_number_e() is defined (_e for "error checking") and is
6529 promoted strongly. The old NCONF_get_number is kept around for
6530 binary backward compatibility.
6531 Make it possible for methods to load from something other than a BIO,
6532 by providing a function pointer that is given a name instead of a BIO.
6533 For example, this could be used to load configuration data from an
6534 LDAP server.
6535 [Richard Levitte]
6536
6537 *) Fix for non blocking accept BIOs. Added new I/O special reason
6538 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
6539 with non blocking I/O was not possible because no retry code was
6540 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
6541 this case.
6542 [Steve Henson]
6543
6544 *) Added the beginnings of Rijndael support.
6545 [Ben Laurie]
6546
6547 *) Fix for bug in DirectoryString mask setting. Add support for
6548 X509_NAME_print_ex() in 'req' and X509_print_ex() function
6549 to allow certificate printing to more controllable, additional
6550 'certopt' option to 'x509' to allow new printing options to be
6551 set.
6552 [Steve Henson]
6553
6554 *) Clean old EAY MD5 hack from e_os.h.
6555 [Richard Levitte]
6556
6557 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
6558
6559 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6560 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6561 [Joe Orton, Steve Henson]
6562
6563 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
6564
6565 *) Fix additional bug revealed by the NISCC test suite:
6566
6567 Stop bug triggering large recursion when presented with
6568 certain ASN.1 tags (CVE-2003-0851)
6569 [Steve Henson]
6570
6571 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
6572
6573 *) Fix various bugs revealed by running the NISCC test suite:
6574
6575 Stop out of bounds reads in the ASN1 code when presented with
6576 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6577
6578 If verify callback ignores invalid public key errors don't try to check
6579 certificate signature with the NULL public key.
6580
6581 [Steve Henson]
6582
6583 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6584 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6585 specifications.
6586 [Steve Henson]
6587
6588 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6589 extra data after the compression methods not only for TLS 1.0
6590 but also for SSL 3.0 (as required by the specification).
6591 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6592
6593 *) Change X509_certificate_type() to mark the key as exported/exportable
6594 when it's 512 *bits* long, not 512 bytes.
6595 [Richard Levitte]
6596
6597 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
6598
6599 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6600 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6601 a protocol version number mismatch like a decryption error
6602 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6603 [Bodo Moeller]
6604
6605 *) Turn on RSA blinding by default in the default implementation
6606 to avoid a timing attack. Applications that don't want it can call
6607 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6608 They would be ill-advised to do so in most cases.
6609 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6610
6611 *) Change RSA blinding code so that it works when the PRNG is not
6612 seeded (in this case, the secret RSA exponent is abused as
6613 an unpredictable seed -- if it is not unpredictable, there
6614 is no point in blinding anyway). Make RSA blinding thread-safe
6615 by remembering the creator's thread ID in rsa->blinding and
6616 having all other threads use local one-time blinding factors
6617 (this requires more computation than sharing rsa->blinding, but
6618 avoids excessive locking; and if an RSA object is not shared
6619 between threads, blinding will still be very fast).
6620 [Bodo Moeller]
6621
6622 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
6623
6624 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6625 via timing by performing a MAC computation even if incorrrect
6626 block cipher padding has been found. This is a countermeasure
6627 against active attacks where the attacker has to distinguish
6628 between bad padding and a MAC verification error. (CVE-2003-0078)
6629
6630 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6631 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6632 Martin Vuagnoux (EPFL, Ilion)]
6633
6634 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
6635
6636 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
6637 memory from it's contents. This is done with a counter that will
6638 place alternating values in each byte. This can be used to solve
6639 two issues: 1) the removal of calls to memset() by highly optimizing
6640 compilers, and 2) cleansing with other values than 0, since those can
6641 be read through on certain media, for example a swap space on disk.
6642 [Geoff Thorpe]
6643
6644 *) Bugfix: client side session caching did not work with external caching,
6645 because the session->cipher setting was not restored when reloading
6646 from the external cache. This problem was masked, when
6647 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
6648 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
6649 [Lutz Jaenicke]
6650
6651 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
6652 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
6653 [Zeev Lieber <zeev-l@yahoo.com>]
6654
6655 *) Undo an undocumented change introduced in 0.9.6e which caused
6656 repeated calls to OpenSSL_add_all_ciphers() and
6657 OpenSSL_add_all_digests() to be ignored, even after calling
6658 EVP_cleanup().
6659 [Richard Levitte]
6660
6661 *) Change the default configuration reader to deal with last line not
6662 being properly terminated.
6663 [Richard Levitte]
6664
6665 *) Change X509_NAME_cmp() so it applies the special rules on handling
6666 DN values that are of type PrintableString, as well as RDNs of type
6667 emailAddress where the value has the type ia5String.
6668 [stefank@valicert.com via Richard Levitte]
6669
6670 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
6671 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
6672 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
6673 the bitwise-OR of the two for use by the majority of applications
6674 wanting this behaviour, and update the docs. The documented
6675 behaviour and actual behaviour were inconsistent and had been
6676 changing anyway, so this is more a bug-fix than a behavioural
6677 change.
6678 [Geoff Thorpe, diagnosed by Nadav Har'El]
6679
6680 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
6681 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
6682 [Bodo Moeller]
6683
6684 *) Fix initialization code race conditions in
6685 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
6686 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
6687 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
6688 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
6689 ssl2_get_cipher_by_char(),
6690 ssl3_get_cipher_by_char().
6691 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
6692
6693 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
6694 the cached sessions are flushed, as the remove_cb() might use ex_data
6695 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
6696 (see [openssl.org #212]).
6697 [Geoff Thorpe, Lutz Jaenicke]
6698
6699 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
6700 length, instead of the encoding length to d2i_ASN1_OBJECT.
6701 [Steve Henson]
6702
6703 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
6704
6705 *) [In 0.9.6g-engine release:]
6706 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
6707 [Lynn Gazis <lgazis@rainbow.com>]
6708
6709 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
6710
6711 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
6712 and get fix the header length calculation.
6713 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
6714 Alon Kantor <alonk@checkpoint.com> (and others),
6715 Steve Henson]
6716
6717 *) Use proper error handling instead of 'assertions' in buffer
6718 overflow checks added in 0.9.6e. This prevents DoS (the
6719 assertions could call abort()).
6720 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
6721
6722 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
6723
6724 *) Add various sanity checks to asn1_get_length() to reject
6725 the ASN1 length bytes if they exceed sizeof(long), will appear
6726 negative or the content length exceeds the length of the
6727 supplied buffer.
6728 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6729
6730 *) Fix cipher selection routines: ciphers without encryption had no flags
6731 for the cipher strength set and where therefore not handled correctly
6732 by the selection routines (PR #130).
6733 [Lutz Jaenicke]
6734
6735 *) Fix EVP_dsa_sha macro.
6736 [Nils Larsch]
6737
6738 *) New option
6739 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
6740 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
6741 that was added in OpenSSL 0.9.6d.
6742
6743 As the countermeasure turned out to be incompatible with some
6744 broken SSL implementations, the new option is part of SSL_OP_ALL.
6745 SSL_OP_ALL is usually employed when compatibility with weird SSL
6746 implementations is desired (e.g. '-bugs' option to 's_client' and
6747 's_server'), so the new option is automatically set in many
6748 applications.
6749 [Bodo Moeller]
6750
6751 *) Changes in security patch:
6752
6753 Changes marked "(CHATS)" were sponsored by the Defense Advanced
6754 Research Projects Agency (DARPA) and Air Force Research Laboratory,
6755 Air Force Materiel Command, USAF, under agreement number
6756 F30602-01-2-0537.
6757
6758 *) Add various sanity checks to asn1_get_length() to reject
6759 the ASN1 length bytes if they exceed sizeof(long), will appear
6760 negative or the content length exceeds the length of the
6761 supplied buffer. (CVE-2002-0659)
6762 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
6763
6764 *) Assertions for various potential buffer overflows, not known to
6765 happen in practice.
6766 [Ben Laurie (CHATS)]
6767
6768 *) Various temporary buffers to hold ASCII versions of integers were
6769 too small for 64 bit platforms. (CVE-2002-0655)
6770 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
6771
6772 *) Remote buffer overflow in SSL3 protocol - an attacker could
6773 supply an oversized session ID to a client. (CVE-2002-0656)
6774 [Ben Laurie (CHATS)]
6775
6776 *) Remote buffer overflow in SSL2 protocol - an attacker could
6777 supply an oversized client master key. (CVE-2002-0656)
6778 [Ben Laurie (CHATS)]
6779
6780 Changes between 0.9.6c and 0.9.6d [9 May 2002]
6781
6782 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
6783 encoded as NULL) with id-dsa-with-sha1.
6784 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
6785
6786 *) Check various X509_...() return values in apps/req.c.
6787 [Nils Larsch <nla@trustcenter.de>]
6788
6789 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
6790 an end-of-file condition would erronously be flagged, when the CRLF
6791 was just at the end of a processed block. The bug was discovered when
6792 processing data through a buffering memory BIO handing the data to a
6793 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
6794 <ptsekov@syntrex.com> and Nedelcho Stanev.
6795 [Lutz Jaenicke]
6796
6797 *) Implement a countermeasure against a vulnerability recently found
6798 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
6799 before application data chunks to avoid the use of known IVs
6800 with data potentially chosen by the attacker.
6801 [Bodo Moeller]
6802
6803 *) Fix length checks in ssl3_get_client_hello().
6804 [Bodo Moeller]
6805
6806 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
6807 to prevent ssl3_read_internal() from incorrectly assuming that
6808 ssl3_read_bytes() found application data while handshake
6809 processing was enabled when in fact s->s3->in_read_app_data was
6810 merely automatically cleared during the initial handshake.
6811 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
6812
6813 *) Fix object definitions for Private and Enterprise: they were not
6814 recognized in their shortname (=lowercase) representation. Extend
6815 obj_dat.pl to issue an error when using undefined keywords instead
6816 of silently ignoring the problem (Svenning Sorensen
6817 <sss@sss.dnsalias.net>).
6818 [Lutz Jaenicke]
6819
6820 *) Fix DH_generate_parameters() so that it works for 'non-standard'
6821 generators, i.e. generators other than 2 and 5. (Previously, the
6822 code did not properly initialise the 'add' and 'rem' values to
6823 BN_generate_prime().)
6824
6825 In the new general case, we do not insist that 'generator' is
6826 actually a primitive root: This requirement is rather pointless;
6827 a generator of the order-q subgroup is just as good, if not
6828 better.
6829 [Bodo Moeller]
6830
6831 *) Map new X509 verification errors to alerts. Discovered and submitted by
6832 Tom Wu <tom@arcot.com>.
6833 [Lutz Jaenicke]
6834
6835 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
6836 returning non-zero before the data has been completely received
6837 when using non-blocking I/O.
6838 [Bodo Moeller; problem pointed out by John Hughes]
6839
6840 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
6841 [Ben Laurie, Lutz Jaenicke]
6842
6843 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
6844 Yoram Zahavi <YoramZ@gilian.com>).
6845 [Lutz Jaenicke]
6846
6847 *) Add information about CygWin 1.3 and on, and preserve proper
6848 configuration for the versions before that.
6849 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6850
6851 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
6852 check whether we deal with a copy of a session and do not delete from
6853 the cache in this case. Problem reported by "Izhar Shoshani Levi"
6854 <izhar@checkpoint.com>.
6855 [Lutz Jaenicke]
6856
6857 *) Do not store session data into the internal session cache, if it
6858 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
6859 flag is set). Proposed by Aslam <aslam@funk.com>.
6860 [Lutz Jaenicke]
6861
6862 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
6863 value is 0.
6864 [Richard Levitte]
6865
6866 *) [In 0.9.6d-engine release:]
6867 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
6868 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
6869
6870 *) Add the configuration target linux-s390x.
6871 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
6872
6873 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
6874 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
6875 variable as an indication that a ClientHello message has been
6876 received. As the flag value will be lost between multiple
6877 invocations of ssl3_accept when using non-blocking I/O, the
6878 function may not be aware that a handshake has actually taken
6879 place, thus preventing a new session from being added to the
6880 session cache.
6881
6882 To avoid this problem, we now set s->new_session to 2 instead of
6883 using a local variable.
6884 [Lutz Jaenicke, Bodo Moeller]
6885
6886 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
6887 if the SSL_R_LENGTH_MISMATCH error is detected.
6888 [Geoff Thorpe, Bodo Moeller]
6889
6890 *) New 'shared_ldflag' column in Configure platform table.
6891 [Richard Levitte]
6892
6893 *) Fix EVP_CIPHER_mode macro.
6894 ["Dan S. Camper" <dan@bti.net>]
6895
6896 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
6897 type, we must throw them away by setting rr->length to 0.
6898 [D P Chang <dpc@qualys.com>]
6899
6900 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
6901
6902 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
6903 <Dominikus.Scherkl@biodata.com>. (The previous implementation
6904 worked incorrectly for those cases where range = 10..._2 and
6905 3*range is two bits longer than range.)
6906 [Bodo Moeller]
6907
6908 *) Only add signing time to PKCS7 structures if it is not already
6909 present.
6910 [Steve Henson]
6911
6912 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
6913 OBJ_ld_ce should be OBJ_id_ce.
6914 Also some ip-pda OIDs in crypto/objects/objects.txt were
6915 incorrect (cf. RFC 3039).
6916 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
6917
6918 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
6919 returns early because it has nothing to do.
6920 [Andy Schneider <andy.schneider@bjss.co.uk>]
6921
6922 *) [In 0.9.6c-engine release:]
6923 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
6924 [Andy Schneider <andy.schneider@bjss.co.uk>]
6925
6926 *) [In 0.9.6c-engine release:]
6927 Add support for Cryptographic Appliance's keyserver technology.
6928 (Use engine 'keyclient')
6929 [Cryptographic Appliances and Geoff Thorpe]
6930
6931 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
6932 is called via tools/c89.sh because arguments have to be
6933 rearranged (all '-L' options must appear before the first object
6934 modules).
6935 [Richard Shapiro <rshapiro@abinitio.com>]
6936
6937 *) [In 0.9.6c-engine release:]
6938 Add support for Broadcom crypto accelerator cards, backported
6939 from 0.9.7.
6940 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
6941
6942 *) [In 0.9.6c-engine release:]
6943 Add support for SureWare crypto accelerator cards from
6944 Baltimore Technologies. (Use engine 'sureware')
6945 [Baltimore Technologies and Mark Cox]
6946
6947 *) [In 0.9.6c-engine release:]
6948 Add support for crypto accelerator cards from Accelerated
6949 Encryption Processing, www.aep.ie. (Use engine 'aep')
6950 [AEP Inc. and Mark Cox]
6951
6952 *) Add a configuration entry for gcc on UnixWare.
6953 [Gary Benson <gbenson@redhat.com>]
6954
6955 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
6956 messages are stored in a single piece (fixed-length part and
6957 variable-length part combined) and fix various bugs found on the way.
6958 [Bodo Moeller]
6959
6960 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
6961 instead. BIO_gethostbyname() does not know what timeouts are
6962 appropriate, so entries would stay in cache even when they have
6963 become invalid.
6964 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
6965
6966 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
6967 faced with a pathologically small ClientHello fragment that does
6968 not contain client_version: Instead of aborting with an error,
6969 simply choose the highest available protocol version (i.e.,
6970 TLS 1.0 unless it is disabled). In practice, ClientHello
6971 messages are never sent like this, but this change gives us
6972 strictly correct behaviour at least for TLS.
6973 [Bodo Moeller]
6974
6975 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
6976 never resets s->method to s->ctx->method when called from within
6977 one of the SSL handshake functions.
6978 [Bodo Moeller; problem pointed out by Niko Baric]
6979
6980 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
6981 (sent using the client's version number) if client_version is
6982 smaller than the protocol version in use. Also change
6983 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
6984 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
6985 the client will at least see that alert.
6986 [Bodo Moeller]
6987
6988 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
6989 correctly.
6990 [Bodo Moeller]
6991
6992 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
6993 client receives HelloRequest while in a handshake.
6994 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
6995
6996 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
6997 should end in 'break', not 'goto end' which circumvents various
6998 cleanups done in state SSL_ST_OK. But session related stuff
6999 must be disabled for SSL_ST_OK in the case that we just sent a
7000 HelloRequest.
7001
7002 Also avoid some overhead by not calling ssl_init_wbio_buffer()
7003 before just sending a HelloRequest.
7004 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
7005
7006 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
7007 reveal whether illegal block cipher padding was found or a MAC
7008 verification error occured. (Neither SSLerr() codes nor alerts
7009 are directly visible to potential attackers, but the information
7010 may leak via logfiles.)
7011
7012 Similar changes are not required for the SSL 2.0 implementation
7013 because the number of padding bytes is sent in clear for SSL 2.0,
7014 and the extra bytes are just ignored. However ssl/s2_pkt.c
7015 failed to verify that the purported number of padding bytes is in
7016 the legal range.
7017 [Bodo Moeller]
7018
7019 *) Add OpenUNIX-8 support including shared libraries
7020 (Boyd Lynn Gerber <gerberb@zenez.com>).
7021 [Lutz Jaenicke]
7022
7023 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
7024 'wristwatch attack' using huge encoding parameters (cf.
7025 James H. Manger's CRYPTO 2001 paper). Note that the
7026 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
7027 encoding parameters and hence was not vulnerable.
7028 [Bodo Moeller]
7029
7030 *) BN_sqr() bug fix.
7031 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
7032
7033 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
7034 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
7035 followed by modular reduction.
7036 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
7037
7038 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
7039 equivalent based on BN_pseudo_rand() instead of BN_rand().
7040 [Bodo Moeller]
7041
7042 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
7043 This function was broken, as the check for a new client hello message
7044 to handle SGC did not allow these large messages.
7045 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
7046 [Lutz Jaenicke]
7047
7048 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
7049 [Lutz Jaenicke]
7050
7051 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
7052 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
7053 [Lutz Jaenicke]
7054
7055 *) Rework the configuration and shared library support for Tru64 Unix.
7056 The configuration part makes use of modern compiler features and
7057 still retains old compiler behavior for those that run older versions
7058 of the OS. The shared library support part includes a variant that
7059 uses the RPATH feature, and is available through the special
7060 configuration target "alpha-cc-rpath", which will never be selected
7061 automatically.
7062 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
7063
7064 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
7065 with the same message size as in ssl3_get_certificate_request().
7066 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
7067 messages might inadvertently be reject as too long.
7068 [Petr Lampa <lampa@fee.vutbr.cz>]
7069
7070 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
7071 [Andy Polyakov]
7072
7073 *) Modified SSL library such that the verify_callback that has been set
7074 specificly for an SSL object with SSL_set_verify() is actually being
7075 used. Before the change, a verify_callback set with this function was
7076 ignored and the verify_callback() set in the SSL_CTX at the time of
7077 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
7078 to allow the necessary settings.
7079 [Lutz Jaenicke]
7080
7081 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
7082 explicitly to NULL, as at least on Solaris 8 this seems not always to be
7083 done automatically (in contradiction to the requirements of the C
7084 standard). This made problems when used from OpenSSH.
7085 [Lutz Jaenicke]
7086
7087 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
7088 dh->length and always used
7089
7090 BN_rand_range(priv_key, dh->p).
7091
7092 BN_rand_range() is not necessary for Diffie-Hellman, and this
7093 specific range makes Diffie-Hellman unnecessarily inefficient if
7094 dh->length (recommended exponent length) is much smaller than the
7095 length of dh->p. We could use BN_rand_range() if the order of
7096 the subgroup was stored in the DH structure, but we only have
7097 dh->length.
7098
7099 So switch back to
7100
7101 BN_rand(priv_key, l, ...)
7102
7103 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
7104 otherwise.
7105 [Bodo Moeller]
7106
7107 *) In
7108
7109 RSA_eay_public_encrypt
7110 RSA_eay_private_decrypt
7111 RSA_eay_private_encrypt (signing)
7112 RSA_eay_public_decrypt (signature verification)
7113
7114 (default implementations for RSA_public_encrypt,
7115 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
7116 always reject numbers >= n.
7117 [Bodo Moeller]
7118
7119 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
7120 to synchronize access to 'locking_thread'. This is necessary on
7121 systems where access to 'locking_thread' (an 'unsigned long'
7122 variable) is not atomic.
7123 [Bodo Moeller]
7124
7125 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
7126 *before* setting the 'crypto_lock_rand' flag. The previous code had
7127 a race condition if 0 is a valid thread ID.
7128 [Travis Vitek <vitek@roguewave.com>]
7129
7130 *) Add support for shared libraries under Irix.
7131 [Albert Chin-A-Young <china@thewrittenword.com>]
7132
7133 *) Add configuration option to build on Linux on both big-endian and
7134 little-endian MIPS.
7135 [Ralf Baechle <ralf@uni-koblenz.de>]
7136
7137 *) Add the possibility to create shared libraries on HP-UX.
7138 [Richard Levitte]
7139
7140 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
7141
7142 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
7143 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
7144 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
7145 PRNG state recovery was possible based on the output of
7146 one PRNG request appropriately sized to gain knowledge on
7147 'md' followed by enough consecutive 1-byte PRNG requests
7148 to traverse all of 'state'.
7149
7150 1. When updating 'md_local' (the current thread's copy of 'md')
7151 during PRNG output generation, hash all of the previous
7152 'md_local' value, not just the half used for PRNG output.
7153
7154 2. Make the number of bytes from 'state' included into the hash
7155 independent from the number of PRNG bytes requested.
7156
7157 The first measure alone would be sufficient to avoid
7158 Markku-Juhani's attack. (Actually it had never occurred
7159 to me that the half of 'md_local' used for chaining was the
7160 half from which PRNG output bytes were taken -- I had always
7161 assumed that the secret half would be used.) The second
7162 measure makes sure that additional data from 'state' is never
7163 mixed into 'md_local' in small portions; this heuristically
7164 further strengthens the PRNG.
7165 [Bodo Moeller]
7166
7167 *) Fix crypto/bn/asm/mips3.s.
7168 [Andy Polyakov]
7169
7170 *) When only the key is given to "enc", the IV is undefined. Print out
7171 an error message in this case.
7172 [Lutz Jaenicke]
7173
7174 *) Handle special case when X509_NAME is empty in X509 printing routines.
7175 [Steve Henson]
7176
7177 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
7178 positive and less than q.
7179 [Bodo Moeller]
7180
7181 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
7182 used: it isn't thread safe and the add_lock_callback should handle
7183 that itself.
7184 [Paul Rose <Paul.Rose@bridge.com>]
7185
7186 *) Verify that incoming data obeys the block size in
7187 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
7188 [Bodo Moeller]
7189
7190 *) Fix OAEP check.
7191 [Ulf Möller, Bodo Möller]
7192
7193 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
7194 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
7195 when fixing the server behaviour for backwards-compatible 'client
7196 hello' messages. (Note that the attack is impractical against
7197 SSL 3.0 and TLS 1.0 anyway because length and version checking
7198 means that the probability of guessing a valid ciphertext is
7199 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
7200 paper.)
7201
7202 Before 0.9.5, the countermeasure (hide the error by generating a
7203 random 'decryption result') did not work properly because
7204 ERR_clear_error() was missing, meaning that SSL_get_error() would
7205 detect the supposedly ignored error.
7206
7207 Both problems are now fixed.
7208 [Bodo Moeller]
7209
7210 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
7211 (previously it was 1024).
7212 [Bodo Moeller]
7213
7214 *) Fix for compatibility mode trust settings: ignore trust settings
7215 unless some valid trust or reject settings are present.
7216 [Steve Henson]
7217
7218 *) Fix for blowfish EVP: its a variable length cipher.
7219 [Steve Henson]
7220
7221 *) Fix various bugs related to DSA S/MIME verification. Handle missing
7222 parameters in DSA public key structures and return an error in the
7223 DSA routines if parameters are absent.
7224 [Steve Henson]
7225
7226 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
7227 in the current directory if neither $RANDFILE nor $HOME was set.
7228 RAND_file_name() in 0.9.6a returned NULL in this case. This has
7229 caused some confusion to Windows users who haven't defined $HOME.
7230 Thus RAND_file_name() is changed again: e_os.h can define a
7231 DEFAULT_HOME, which will be used if $HOME is not set.
7232 For Windows, we use "C:"; on other platforms, we still require
7233 environment variables.
7234
7235 *) Move 'if (!initialized) RAND_poll()' into regions protected by
7236 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
7237 having multiple threads call RAND_poll() concurrently.
7238 [Bodo Moeller]
7239
7240 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
7241 combination of a flag and a thread ID variable.
7242 Otherwise while one thread is in ssleay_rand_bytes (which sets the
7243 flag), *other* threads can enter ssleay_add_bytes without obeying
7244 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
7245 that they do not hold after the first thread unsets add_do_not_lock).
7246 [Bodo Moeller]
7247
7248 *) Change bctest again: '-x' expressions are not available in all
7249 versions of 'test'.
7250 [Bodo Moeller]
7251
7252 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
7253
7254 *) Fix a couple of memory leaks in PKCS7_dataDecode()
7255 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
7256
7257 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
7258 the default extension for executables, if any. Also, make the perl
7259 scripts that use symlink() to test if it really exists and use "cp"
7260 if it doesn't. All this made OpenSSL compilable and installable in
7261 CygWin.
7262 [Richard Levitte]
7263
7264 *) Fix for asn1_GetSequence() for indefinite length constructed data.
7265 If SEQUENCE is length is indefinite just set c->slen to the total
7266 amount of data available.
7267 [Steve Henson, reported by shige@FreeBSD.org]
7268 [This change does not apply to 0.9.7.]
7269
7270 *) Change bctest to avoid here-documents inside command substitution
7271 (workaround for FreeBSD /bin/sh bug).
7272 For compatibility with Ultrix, avoid shell functions (introduced
7273 in the bctest version that searches along $PATH).
7274 [Bodo Moeller]
7275
7276 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
7277 with des_encrypt() defined on some operating systems, like Solaris
7278 and UnixWare.
7279 [Richard Levitte]
7280
7281 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
7282 On the Importance of Eliminating Errors in Cryptographic
7283 Computations, J. Cryptology 14 (2001) 2, 101-119,
7284 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
7285 [Ulf Moeller]
7286
7287 *) MIPS assembler BIGNUM division bug fix.
7288 [Andy Polyakov]
7289
7290 *) Disabled incorrect Alpha assembler code.
7291 [Richard Levitte]
7292
7293 *) Fix PKCS#7 decode routines so they correctly update the length
7294 after reading an EOC for the EXPLICIT tag.
7295 [Steve Henson]
7296 [This change does not apply to 0.9.7.]
7297
7298 *) Fix bug in PKCS#12 key generation routines. This was triggered
7299 if a 3DES key was generated with a 0 initial byte. Include
7300 PKCS12_BROKEN_KEYGEN compilation option to retain the old
7301 (but broken) behaviour.
7302 [Steve Henson]
7303
7304 *) Enhance bctest to search for a working bc along $PATH and print
7305 it when found.
7306 [Tim Rice <tim@multitalents.net> via Richard Levitte]
7307
7308 *) Fix memory leaks in err.c: free err_data string if necessary;
7309 don't write to the wrong index in ERR_set_error_data.
7310 [Bodo Moeller]
7311
7312 *) Implement ssl23_peek (analogous to ssl23_read), which previously
7313 did not exist.
7314 [Bodo Moeller]
7315
7316 *) Replace rdtsc with _emit statements for VC++ version 5.
7317 [Jeremy Cooper <jeremy@baymoo.org>]
7318
7319 *) Make it possible to reuse SSLv2 sessions.
7320 [Richard Levitte]
7321
7322 *) In copy_email() check for >= 0 as a return value for
7323 X509_NAME_get_index_by_NID() since 0 is a valid index.
7324 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
7325
7326 *) Avoid coredump with unsupported or invalid public keys by checking if
7327 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
7328 PKCS7_verify() fails with non detached data.
7329 [Steve Henson]
7330
7331 *) Don't use getenv in library functions when run as setuid/setgid.
7332 New function OPENSSL_issetugid().
7333 [Ulf Moeller]
7334
7335 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
7336 due to incorrect handling of multi-threading:
7337
7338 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
7339
7340 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
7341
7342 3. Count how many times MemCheck_off() has been called so that
7343 nested use can be treated correctly. This also avoids
7344 inband-signalling in the previous code (which relied on the
7345 assumption that thread ID 0 is impossible).
7346 [Bodo Moeller]
7347
7348 *) Add "-rand" option also to s_client and s_server.
7349 [Lutz Jaenicke]
7350
7351 *) Fix CPU detection on Irix 6.x.
7352 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
7353 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7354
7355 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
7356 was empty.
7357 [Steve Henson]
7358 [This change does not apply to 0.9.7.]
7359
7360 *) Use the cached encoding of an X509_NAME structure rather than
7361 copying it. This is apparently the reason for the libsafe "errors"
7362 but the code is actually correct.
7363 [Steve Henson]
7364
7365 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
7366 Bleichenbacher's DSA attack.
7367 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
7368 to be set and top=0 forces the highest bit to be set; top=-1 is new
7369 and leaves the highest bit random.
7370 [Ulf Moeller, Bodo Moeller]
7371
7372 *) In the NCONF_...-based implementations for CONF_... queries
7373 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
7374 a temporary CONF structure with the data component set to NULL
7375 (which gives segmentation faults in lh_retrieve).
7376 Instead, use NULL for the CONF pointer in CONF_get_string and
7377 CONF_get_number (which may use environment variables) and directly
7378 return NULL from CONF_get_section.
7379 [Bodo Moeller]
7380
7381 *) Fix potential buffer overrun for EBCDIC.
7382 [Ulf Moeller]
7383
7384 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
7385 keyUsage if basicConstraints absent for a CA.
7386 [Steve Henson]
7387
7388 *) Make SMIME_write_PKCS7() write mail header values with a format that
7389 is more generally accepted (no spaces before the semicolon), since
7390 some programs can't parse those values properly otherwise. Also make
7391 sure BIO's that break lines after each write do not create invalid
7392 headers.
7393 [Richard Levitte]
7394
7395 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
7396 macros previously used would not encode an empty SEQUENCE OF
7397 and break the signature.
7398 [Steve Henson]
7399 [This change does not apply to 0.9.7.]
7400
7401 *) Zero the premaster secret after deriving the master secret in
7402 DH ciphersuites.
7403 [Steve Henson]
7404
7405 *) Add some EVP_add_digest_alias registrations (as found in
7406 OpenSSL_add_all_digests()) to SSL_library_init()
7407 aka OpenSSL_add_ssl_algorithms(). This provides improved
7408 compatibility with peers using X.509 certificates
7409 with unconventional AlgorithmIdentifier OIDs.
7410 [Bodo Moeller]
7411
7412 *) Fix for Irix with NO_ASM.
7413 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
7414
7415 *) ./config script fixes.
7416 [Ulf Moeller, Richard Levitte]
7417
7418 *) Fix 'openssl passwd -1'.
7419 [Bodo Moeller]
7420
7421 *) Change PKCS12_key_gen_asc() so it can cope with non null
7422 terminated strings whose length is passed in the passlen
7423 parameter, for example from PEM callbacks. This was done
7424 by adding an extra length parameter to asc2uni().
7425 [Steve Henson, reported by <oddissey@samsung.co.kr>]
7426
7427 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
7428 call failed, free the DSA structure.
7429 [Bodo Moeller]
7430
7431 *) Fix to uni2asc() to cope with zero length Unicode strings.
7432 These are present in some PKCS#12 files.
7433 [Steve Henson]
7434
7435 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
7436 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
7437 when writing a 32767 byte record.
7438 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
7439
7440 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
7441 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
7442
7443 (RSA objects have a reference count access to which is protected
7444 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
7445 so they are meant to be shared between threads.)
7446 [Bodo Moeller, Geoff Thorpe; original patch submitted by
7447 "Reddie, Steven" <Steven.Reddie@ca.com>]
7448
7449 *) Fix a deadlock in CRYPTO_mem_leaks().
7450 [Bodo Moeller]
7451
7452 *) Use better test patterns in bntest.
7453 [Ulf Möller]
7454
7455 *) rand_win.c fix for Borland C.
7456 [Ulf Möller]
7457
7458 *) BN_rshift bugfix for n == 0.
7459 [Bodo Moeller]
7460
7461 *) Add a 'bctest' script that checks for some known 'bc' bugs
7462 so that 'make test' does not abort just because 'bc' is broken.
7463 [Bodo Moeller]
7464
7465 *) Store verify_result within SSL_SESSION also for client side to
7466 avoid potential security hole. (Re-used sessions on the client side
7467 always resulted in verify_result==X509_V_OK, not using the original
7468 result of the server certificate verification.)
7469 [Lutz Jaenicke]
7470
7471 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
7472 SSL3_RT_APPLICATION_DATA, return 0.
7473 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
7474 [Bodo Moeller]
7475
7476 *) Fix SSL_peek:
7477 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
7478 releases, have been re-implemented by renaming the previous
7479 implementations of ssl2_read and ssl3_read to ssl2_read_internal
7480 and ssl3_read_internal, respectively, and adding 'peek' parameters
7481 to them. The new ssl[23]_{read,peek} functions are calls to
7482 ssl[23]_read_internal with the 'peek' flag set appropriately.
7483 A 'peek' parameter has also been added to ssl3_read_bytes, which
7484 does the actual work for ssl3_read_internal.
7485 [Bodo Moeller]
7486
7487 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
7488 the method-specific "init()" handler. Also clean up ex_data after
7489 calling the method-specific "finish()" handler. Previously, this was
7490 happening the other way round.
7491 [Geoff Thorpe]
7492
7493 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
7494 The previous value, 12, was not always sufficient for BN_mod_exp().
7495 [Bodo Moeller]
7496
7497 *) Make sure that shared libraries get the internal name engine with
7498 the full version number and not just 0. This should mark the
7499 shared libraries as not backward compatible. Of course, this should
7500 be changed again when we can guarantee backward binary compatibility.
7501 [Richard Levitte]
7502
7503 *) Fix typo in get_cert_by_subject() in by_dir.c
7504 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
7505
7506 *) Rework the system to generate shared libraries:
7507
7508 - Make note of the expected extension for the shared libraries and
7509 if there is a need for symbolic links from for example libcrypto.so.0
7510 to libcrypto.so.0.9.7. There is extended info in Configure for
7511 that.
7512
7513 - Make as few rebuilds of the shared libraries as possible.
7514
7515 - Still avoid linking the OpenSSL programs with the shared libraries.
7516
7517 - When installing, install the shared libraries separately from the
7518 static ones.
7519 [Richard Levitte]
7520
7521 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
7522
7523 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
7524 and not in SSL_clear because the latter is also used by the
7525 accept/connect functions; previously, the settings made by
7526 SSL_set_read_ahead would be lost during the handshake.
7527 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
7528
7529 *) Correct util/mkdef.pl to be selective about disabled algorithms.
7530 Previously, it would create entries for disableed algorithms no
7531 matter what.
7532 [Richard Levitte]
7533
7534 *) Added several new manual pages for SSL_* function.
7535 [Lutz Jaenicke]
7536
7537 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
7538
7539 *) In ssl23_get_client_hello, generate an error message when faced
7540 with an initial SSL 3.0/TLS record that is too small to contain the
7541 first two bytes of the ClientHello message, i.e. client_version.
7542 (Note that this is a pathologic case that probably has never happened
7543 in real life.) The previous approach was to use the version number
7544 from the record header as a substitute; but our protocol choice
7545 should not depend on that one because it is not authenticated
7546 by the Finished messages.
7547 [Bodo Moeller]
7548
7549 *) More robust randomness gathering functions for Windows.
7550 [Jeffrey Altman <jaltman@columbia.edu>]
7551
7552 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
7553 not set then we don't setup the error code for issuer check errors
7554 to avoid possibly overwriting other errors which the callback does
7555 handle. If an application does set the flag then we assume it knows
7556 what it is doing and can handle the new informational codes
7557 appropriately.
7558 [Steve Henson]
7559
7560 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
7561 a general "ANY" type, as such it should be able to decode anything
7562 including tagged types. However it didn't check the class so it would
7563 wrongly interpret tagged types in the same way as their universal
7564 counterpart and unknown types were just rejected. Changed so that the
7565 tagged and unknown types are handled in the same way as a SEQUENCE:
7566 that is the encoding is stored intact. There is also a new type
7567 "V_ASN1_OTHER" which is used when the class is not universal, in this
7568 case we have no idea what the actual type is so we just lump them all
7569 together.
7570 [Steve Henson]
7571
7572 *) On VMS, stdout may very well lead to a file that is written to
7573 in a record-oriented fashion. That means that every write() will
7574 write a separate record, which will be read separately by the
7575 programs trying to read from it. This can be very confusing.
7576
7577 The solution is to put a BIO filter in the way that will buffer
7578 text until a linefeed is reached, and then write everything a
7579 line at a time, so every record written will be an actual line,
7580 not chunks of lines and not (usually doesn't happen, but I've
7581 seen it once) several lines in one record. BIO_f_linebuffer() is
7582 the answer.
7583
7584 Currently, it's a VMS-only method, because that's where it has
7585 been tested well enough.
7586 [Richard Levitte]
7587
7588 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
7589 it can return incorrect results.
7590 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
7591 but it was in 0.9.6-beta[12].)
7592 [Bodo Moeller]
7593
7594 *) Disable the check for content being present when verifying detached
7595 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
7596 include zero length content when signing messages.
7597 [Steve Henson]
7598
7599 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
7600 BIO_ctrl (for BIO pairs).
7601 [Bodo Möller]
7602
7603 *) Add DSO method for VMS.
7604 [Richard Levitte]
7605
7606 *) Bug fix: Montgomery multiplication could produce results with the
7607 wrong sign.
7608 [Ulf Möller]
7609
7610 *) Add RPM specification openssl.spec and modify it to build three
7611 packages. The default package contains applications, application
7612 documentation and run-time libraries. The devel package contains
7613 include files, static libraries and function documentation. The
7614 doc package contains the contents of the doc directory. The original
7615 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
7616 [Richard Levitte]
7617
7618 *) Add a large number of documentation files for many SSL routines.
7619 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
7620
7621 *) Add a configuration entry for Sony News 4.
7622 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
7623
7624 *) Don't set the two most significant bits to one when generating a
7625 random number < q in the DSA library.
7626 [Ulf Möller]
7627
7628 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
7629 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
7630 the underlying transport is blocking) if a handshake took place.
7631 (The default behaviour is needed by applications such as s_client
7632 and s_server that use select() to determine when to use SSL_read;
7633 but for applications that know in advance when to expect data, it
7634 just makes things more complicated.)
7635 [Bodo Moeller]
7636
7637 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
7638 from EGD.
7639 [Ben Laurie]
7640
7641 *) Add a few more EBCDIC conditionals that make `req' and `x509'
7642 work better on such systems.
7643 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7644
7645 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
7646 Update PKCS12_parse() so it copies the friendlyName and the
7647 keyid to the certificates aux info.
7648 [Steve Henson]
7649
7650 *) Fix bug in PKCS7_verify() which caused an infinite loop
7651 if there was more than one signature.
7652 [Sven Uszpelkat <su@celocom.de>]
7653
7654 *) Major change in util/mkdef.pl to include extra information
7655 about each symbol, as well as presenting variables as well
7656 as functions. This change means that there's n more need
7657 to rebuild the .num files when some algorithms are excluded.
7658 [Richard Levitte]
7659
7660 *) Allow the verify time to be set by an application,
7661 rather than always using the current time.
7662 [Steve Henson]
7663
7664 *) Phase 2 verify code reorganisation. The certificate
7665 verify code now looks up an issuer certificate by a
7666 number of criteria: subject name, authority key id
7667 and key usage. It also verifies self signed certificates
7668 by the same criteria. The main comparison function is
7669 X509_check_issued() which performs these checks.
7670
7671 Lot of changes were necessary in order to support this
7672 without completely rewriting the lookup code.
7673
7674 Authority and subject key identifier are now cached.
7675
7676 The LHASH 'certs' is X509_STORE has now been replaced
7677 by a STACK_OF(X509_OBJECT). This is mainly because an
7678 LHASH can't store or retrieve multiple objects with
7679 the same hash value.
7680
7681 As a result various functions (which were all internal
7682 use only) have changed to handle the new X509_STORE
7683 structure. This will break anything that messed round
7684 with X509_STORE internally.
7685
7686 The functions X509_STORE_add_cert() now checks for an
7687 exact match, rather than just subject name.
7688
7689 The X509_STORE API doesn't directly support the retrieval
7690 of multiple certificates matching a given criteria, however
7691 this can be worked round by performing a lookup first
7692 (which will fill the cache with candidate certificates)
7693 and then examining the cache for matches. This is probably
7694 the best we can do without throwing out X509_LOOKUP
7695 entirely (maybe later...).
7696
7697 The X509_VERIFY_CTX structure has been enhanced considerably.
7698
7699 All certificate lookup operations now go via a get_issuer()
7700 callback. Although this currently uses an X509_STORE it
7701 can be replaced by custom lookups. This is a simple way
7702 to bypass the X509_STORE hackery necessary to make this
7703 work and makes it possible to use more efficient techniques
7704 in future. A very simple version which uses a simple
7705 STACK for its trusted certificate store is also provided
7706 using X509_STORE_CTX_trusted_stack().
7707
7708 The verify_cb() and verify() callbacks now have equivalents
7709 in the X509_STORE_CTX structure.
7710
7711 X509_STORE_CTX also has a 'flags' field which can be used
7712 to customise the verify behaviour.
7713 [Steve Henson]
7714
7715 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
7716 excludes S/MIME capabilities.
7717 [Steve Henson]
7718
7719 *) When a certificate request is read in keep a copy of the
7720 original encoding of the signed data and use it when outputing
7721 again. Signatures then use the original encoding rather than
7722 a decoded, encoded version which may cause problems if the
7723 request is improperly encoded.
7724 [Steve Henson]
7725
7726 *) For consistency with other BIO_puts implementations, call
7727 buffer_write(b, ...) directly in buffer_puts instead of calling
7728 BIO_write(b, ...).
7729
7730 In BIO_puts, increment b->num_write as in BIO_write.
7731 [Peter.Sylvester@EdelWeb.fr]
7732
7733 *) Fix BN_mul_word for the case where the word is 0. (We have to use
7734 BN_zero, we may not return a BIGNUM with an array consisting of
7735 words set to zero.)
7736 [Bodo Moeller]
7737
7738 *) Avoid calling abort() from within the library when problems are
7739 detected, except if preprocessor symbols have been defined
7740 (such as REF_CHECK, BN_DEBUG etc.).
7741 [Bodo Moeller]
7742
7743 *) New openssl application 'rsautl'. This utility can be
7744 used for low level RSA operations. DER public key
7745 BIO/fp routines also added.
7746 [Steve Henson]
7747
7748 *) New Configure entry and patches for compiling on QNX 4.
7749 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
7750
7751 *) A demo state-machine implementation was sponsored by
7752 Nuron (http://www.nuron.com/) and is now available in
7753 demos/state_machine.
7754 [Ben Laurie]
7755
7756 *) New options added to the 'dgst' utility for signature
7757 generation and verification.
7758 [Steve Henson]
7759
7760 *) Unrecognized PKCS#7 content types are now handled via a
7761 catch all ASN1_TYPE structure. This allows unsupported
7762 types to be stored as a "blob" and an application can
7763 encode and decode it manually.
7764 [Steve Henson]
7765
7766 *) Fix various signed/unsigned issues to make a_strex.c
7767 compile under VC++.
7768 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
7769
7770 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
7771 length if passed a buffer. ASN1_INTEGER_to_BN failed
7772 if passed a NULL BN and its argument was negative.
7773 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
7774
7775 *) Modification to PKCS#7 encoding routines to output definite
7776 length encoding. Since currently the whole structures are in
7777 memory there's not real point in using indefinite length
7778 constructed encoding. However if OpenSSL is compiled with
7779 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
7780 [Steve Henson]
7781
7782 *) Added BIO_vprintf() and BIO_vsnprintf().
7783 [Richard Levitte]
7784
7785 *) Added more prefixes to parse for in the the strings written
7786 through a logging bio, to cover all the levels that are available
7787 through syslog. The prefixes are now:
7788
7789 PANIC, EMERG, EMR => LOG_EMERG
7790 ALERT, ALR => LOG_ALERT
7791 CRIT, CRI => LOG_CRIT
7792 ERROR, ERR => LOG_ERR
7793 WARNING, WARN, WAR => LOG_WARNING
7794 NOTICE, NOTE, NOT => LOG_NOTICE
7795 INFO, INF => LOG_INFO
7796 DEBUG, DBG => LOG_DEBUG
7797
7798 and as before, if none of those prefixes are present at the
7799 beginning of the string, LOG_ERR is chosen.
7800
7801 On Win32, the LOG_* levels are mapped according to this:
7802
7803 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
7804 LOG_WARNING => EVENTLOG_WARNING_TYPE
7805 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
7806
7807 [Richard Levitte]
7808
7809 *) Made it possible to reconfigure with just the configuration
7810 argument "reconf" or "reconfigure". The command line arguments
7811 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
7812 and are retrieved from there when reconfiguring.
7813 [Richard Levitte]
7814
7815 *) MD4 implemented.
7816 [Assar Westerlund <assar@sics.se>, Richard Levitte]
7817
7818 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
7819 [Richard Levitte]
7820
7821 *) The obj_dat.pl script was messing up the sorting of object
7822 names. The reason was that it compared the quoted version
7823 of strings as a result "OCSP" > "OCSP Signing" because
7824 " > SPACE. Changed script to store unquoted versions of
7825 names and add quotes on output. It was also omitting some
7826 names from the lookup table if they were given a default
7827 value (that is if SN is missing it is given the same
7828 value as LN and vice versa), these are now added on the
7829 grounds that if an object has a name we should be able to
7830 look it up. Finally added warning output when duplicate
7831 short or long names are found.
7832 [Steve Henson]
7833
7834 *) Changes needed for Tandem NSK.
7835 [Scott Uroff <scott@xypro.com>]
7836
7837 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
7838 RSA_padding_check_SSLv23(), special padding was never detected
7839 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
7840 version rollback attacks was not effective.
7841
7842 In s23_clnt.c, don't use special rollback-attack detection padding
7843 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
7844 client; similarly, in s23_srvr.c, don't do the rollback check if
7845 SSL 2.0 is the only protocol enabled in the server.
7846 [Bodo Moeller]
7847
7848 *) Make it possible to get hexdumps of unprintable data with 'openssl
7849 asn1parse'. By implication, the functions ASN1_parse_dump() and
7850 BIO_dump_indent() are added.
7851 [Richard Levitte]
7852
7853 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
7854 these print out strings and name structures based on various
7855 flags including RFC2253 support and proper handling of
7856 multibyte characters. Added options to the 'x509' utility
7857 to allow the various flags to be set.
7858 [Steve Henson]
7859
7860 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
7861 Also change the functions X509_cmp_current_time() and
7862 X509_gmtime_adj() work with an ASN1_TIME structure,
7863 this will enable certificates using GeneralizedTime in validity
7864 dates to be checked.
7865 [Steve Henson]
7866
7867 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
7868 negative public key encodings) on by default,
7869 NO_NEG_PUBKEY_BUG can be set to disable it.
7870 [Steve Henson]
7871
7872 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
7873 content octets. An i2c_ASN1_OBJECT is unnecessary because
7874 the encoding can be trivially obtained from the structure.
7875 [Steve Henson]
7876
7877 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
7878 not read locks (CRYPTO_r_[un]lock).
7879 [Bodo Moeller]
7880
7881 *) A first attempt at creating official support for shared
7882 libraries through configuration. I've kept it so the
7883 default is static libraries only, and the OpenSSL programs
7884 are always statically linked for now, but there are
7885 preparations for dynamic linking in place.
7886 This has been tested on Linux and Tru64.
7887 [Richard Levitte]
7888
7889 *) Randomness polling function for Win9x, as described in:
7890 Peter Gutmann, Software Generation of Practically Strong
7891 Random Numbers.
7892 [Ulf Möller]
7893
7894 *) Fix so PRNG is seeded in req if using an already existing
7895 DSA key.
7896 [Steve Henson]
7897
7898 *) New options to smime application. -inform and -outform
7899 allow alternative formats for the S/MIME message including
7900 PEM and DER. The -content option allows the content to be
7901 specified separately. This should allow things like Netscape
7902 form signing output easier to verify.
7903 [Steve Henson]
7904
7905 *) Fix the ASN1 encoding of tags using the 'long form'.
7906 [Steve Henson]
7907
7908 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
7909 STRING types. These convert content octets to and from the
7910 underlying type. The actual tag and length octets are
7911 already assumed to have been read in and checked. These
7912 are needed because all other string types have virtually
7913 identical handling apart from the tag. By having versions
7914 of the ASN1 functions that just operate on content octets
7915 IMPLICIT tagging can be handled properly. It also allows
7916 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
7917 and ASN1_INTEGER are identical apart from the tag.
7918 [Steve Henson]
7919
7920 *) Change the handling of OID objects as follows:
7921
7922 - New object identifiers are inserted in objects.txt, following
7923 the syntax given in objects.README.
7924 - objects.pl is used to process obj_mac.num and create a new
7925 obj_mac.h.
7926 - obj_dat.pl is used to create a new obj_dat.h, using the data in
7927 obj_mac.h.
7928
7929 This is currently kind of a hack, and the perl code in objects.pl
7930 isn't very elegant, but it works as I intended. The simplest way
7931 to check that it worked correctly is to look in obj_dat.h and
7932 check the array nid_objs and make sure the objects haven't moved
7933 around (this is important!). Additions are OK, as well as
7934 consistent name changes.
7935 [Richard Levitte]
7936
7937 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
7938 [Bodo Moeller]
7939
7940 *) Addition of the command line parameter '-rand file' to 'openssl req'.
7941 The given file adds to whatever has already been seeded into the
7942 random pool through the RANDFILE configuration file option or
7943 environment variable, or the default random state file.
7944 [Richard Levitte]
7945
7946 *) mkstack.pl now sorts each macro group into lexical order.
7947 Previously the output order depended on the order the files
7948 appeared in the directory, resulting in needless rewriting
7949 of safestack.h .
7950 [Steve Henson]
7951
7952 *) Patches to make OpenSSL compile under Win32 again. Mostly
7953 work arounds for the VC++ problem that it treats func() as
7954 func(void). Also stripped out the parts of mkdef.pl that
7955 added extra typesafe functions: these no longer exist.
7956 [Steve Henson]
7957
7958 *) Reorganisation of the stack code. The macros are now all
7959 collected in safestack.h . Each macro is defined in terms of
7960 a "stack macro" of the form SKM_<name>(type, a, b). The
7961 DEBUG_SAFESTACK is now handled in terms of function casts,
7962 this has the advantage of retaining type safety without the
7963 use of additional functions. If DEBUG_SAFESTACK is not defined
7964 then the non typesafe macros are used instead. Also modified the
7965 mkstack.pl script to handle the new form. Needs testing to see
7966 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
7967 the default if no major problems. Similar behaviour for ASN1_SET_OF
7968 and PKCS12_STACK_OF.
7969 [Steve Henson]
7970
7971 *) When some versions of IIS use the 'NET' form of private key the
7972 key derivation algorithm is different. Normally MD5(password) is
7973 used as a 128 bit RC4 key. In the modified case
7974 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
7975 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
7976 as the old Netscape_RSA functions except they have an additional
7977 'sgckey' parameter which uses the modified algorithm. Also added
7978 an -sgckey command line option to the rsa utility. Thanks to
7979 Adrian Peck <bertie@ncipher.com> for posting details of the modified
7980 algorithm to openssl-dev.
7981 [Steve Henson]
7982
7983 *) The evp_local.h macros were using 'c.##kname' which resulted in
7984 invalid expansion on some systems (SCO 5.0.5 for example).
7985 Corrected to 'c.kname'.
7986 [Phillip Porch <root@theporch.com>]
7987
7988 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
7989 a STACK of email addresses from a certificate or request, these look
7990 in the subject name and the subject alternative name extensions and
7991 omit any duplicate addresses.
7992 [Steve Henson]
7993
7994 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
7995 This makes DSA verification about 2 % faster.
7996 [Bodo Moeller]
7997
7998 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
7999 (meaning that now 2^5 values will be precomputed, which is only 4 KB
8000 plus overhead for 1024 bit moduli).
8001 This makes exponentiations about 0.5 % faster for 1024 bit
8002 exponents (as measured by "openssl speed rsa2048").
8003 [Bodo Moeller]
8004
8005 *) Rename memory handling macros to avoid conflicts with other
8006 software:
8007 Malloc => OPENSSL_malloc
8008 Malloc_locked => OPENSSL_malloc_locked
8009 Realloc => OPENSSL_realloc
8010 Free => OPENSSL_free
8011 [Richard Levitte]
8012
8013 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
8014 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
8015 [Bodo Moeller]
8016
8017 *) CygWin32 support.
8018 [John Jarvie <jjarvie@newsguy.com>]
8019
8020 *) The type-safe stack code has been rejigged. It is now only compiled
8021 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
8022 by default all type-specific stack functions are "#define"d back to
8023 standard stack functions. This results in more streamlined output
8024 but retains the type-safety checking possibilities of the original
8025 approach.
8026 [Geoff Thorpe]
8027
8028 *) The STACK code has been cleaned up, and certain type declarations
8029 that didn't make a lot of sense have been brought in line. This has
8030 also involved a cleanup of sorts in safestack.h to more correctly
8031 map type-safe stack functions onto their plain stack counterparts.
8032 This work has also resulted in a variety of "const"ifications of
8033 lots of the code, especially "_cmp" operations which should normally
8034 be prototyped with "const" parameters anyway.
8035 [Geoff Thorpe]
8036
8037 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
8038 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
8039 (The PRNG state consists of two parts, the large pool 'state' and 'md',
8040 where all of 'md' is used each time the PRNG is used, but 'state'
8041 is used only indexed by a cyclic counter. As entropy may not be
8042 well distributed from the beginning, 'md' is important as a
8043 chaining variable. However, the output function chains only half
8044 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
8045 all of 'md', and seeding with STATE_SIZE dummy bytes will result
8046 in all of 'state' being rewritten, with the new values depending
8047 on virtually all of 'md'. This overcomes the 80 bit limitation.)
8048 [Bodo Moeller]
8049
8050 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
8051 the handshake is continued after ssl_verify_cert_chain();
8052 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
8053 can lead to 'unexplainable' connection aborts later.
8054 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
8055
8056 *) Major EVP API cipher revision.
8057 Add hooks for extra EVP features. This allows various cipher
8058 parameters to be set in the EVP interface. Support added for variable
8059 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
8060 setting of RC2 and RC5 parameters.
8061
8062 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
8063 ciphers.
8064
8065 Remove lots of duplicated code from the EVP library. For example *every*
8066 cipher init() function handles the 'iv' in the same way according to the
8067 cipher mode. They also all do nothing if the 'key' parameter is NULL and
8068 for CFB and OFB modes they zero ctx->num.
8069
8070 New functionality allows removal of S/MIME code RC2 hack.
8071
8072 Most of the routines have the same form and so can be declared in terms
8073 of macros.
8074
8075 By shifting this to the top level EVP_CipherInit() it can be removed from
8076 all individual ciphers. If the cipher wants to handle IVs or keys
8077 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
8078 flags.
8079
8080 Change lots of functions like EVP_EncryptUpdate() to now return a
8081 value: although software versions of the algorithms cannot fail
8082 any installed hardware versions can.
8083 [Steve Henson]
8084
8085 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
8086 this option is set, tolerate broken clients that send the negotiated
8087 protocol version number instead of the requested protocol version
8088 number.
8089 [Bodo Moeller]
8090
8091 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
8092 i.e. non-zero for export ciphersuites, zero otherwise.
8093 Previous versions had this flag inverted, inconsistent with
8094 rsa_tmp_cb (..._TMP_RSA_CB).
8095 [Bodo Moeller; problem reported by Amit Chopra]
8096
8097 *) Add missing DSA library text string. Work around for some IIS
8098 key files with invalid SEQUENCE encoding.
8099 [Steve Henson]
8100
8101 *) Add a document (doc/standards.txt) that list all kinds of standards
8102 and so on that are implemented in OpenSSL.
8103 [Richard Levitte]
8104
8105 *) Enhance c_rehash script. Old version would mishandle certificates
8106 with the same subject name hash and wouldn't handle CRLs at all.
8107 Added -fingerprint option to crl utility, to support new c_rehash
8108 features.
8109 [Steve Henson]
8110
8111 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
8112 [Ulf Möller]
8113
8114 *) Fix for SSL server purpose checking. Server checking was
8115 rejecting certificates which had extended key usage present
8116 but no ssl client purpose.
8117 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
8118
8119 *) Make PKCS#12 code work with no password. The PKCS#12 spec
8120 is a little unclear about how a blank password is handled.
8121 Since the password in encoded as a BMPString with terminating
8122 double NULL a zero length password would end up as just the
8123 double NULL. However no password at all is different and is
8124 handled differently in the PKCS#12 key generation code. NS
8125 treats a blank password as zero length. MSIE treats it as no
8126 password on export: but it will try both on import. We now do
8127 the same: PKCS12_parse() tries zero length and no password if
8128 the password is set to "" or NULL (NULL is now a valid password:
8129 it wasn't before) as does the pkcs12 application.
8130 [Steve Henson]
8131
8132 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
8133 perror when PEM_read_bio_X509_REQ fails, the error message must
8134 be obtained from the error queue.
8135 [Bodo Moeller]
8136
8137 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
8138 it in ERR_remove_state if appropriate, and change ERR_get_state
8139 accordingly to avoid race conditions (this is necessary because
8140 thread_hash is no longer constant once set).
8141 [Bodo Moeller]
8142
8143 *) Bugfix for linux-elf makefile.one.
8144 [Ulf Möller]
8145
8146 *) RSA_get_default_method() will now cause a default
8147 RSA_METHOD to be chosen if one doesn't exist already.
8148 Previously this was only set during a call to RSA_new()
8149 or RSA_new_method(NULL) meaning it was possible for
8150 RSA_get_default_method() to return NULL.
8151 [Geoff Thorpe]
8152
8153 *) Added native name translation to the existing DSO code
8154 that will convert (if the flag to do so is set) filenames
8155 that are sufficiently small and have no path information
8156 into a canonical native form. Eg. "blah" converted to
8157 "libblah.so" or "blah.dll" etc.
8158 [Geoff Thorpe]
8159
8160 *) New function ERR_error_string_n(e, buf, len) which is like
8161 ERR_error_string(e, buf), but writes at most 'len' bytes
8162 including the 0 terminator. For ERR_error_string_n, 'buf'
8163 may not be NULL.
8164 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
8165
8166 *) CONF library reworked to become more general. A new CONF
8167 configuration file reader "class" is implemented as well as a
8168 new functions (NCONF_*, for "New CONF") to handle it. The now
8169 old CONF_* functions are still there, but are reimplemented to
8170 work in terms of the new functions. Also, a set of functions
8171 to handle the internal storage of the configuration data is
8172 provided to make it easier to write new configuration file
8173 reader "classes" (I can definitely see something reading a
8174 configuration file in XML format, for example), called _CONF_*,
8175 or "the configuration storage API"...
8176
8177 The new configuration file reading functions are:
8178
8179 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
8180 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
8181
8182 NCONF_default, NCONF_WIN32
8183
8184 NCONF_dump_fp, NCONF_dump_bio
8185
8186 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
8187 NCONF_new creates a new CONF object. This works in the same way
8188 as other interfaces in OpenSSL, like the BIO interface.
8189 NCONF_dump_* dump the internal storage of the configuration file,
8190 which is useful for debugging. All other functions take the same
8191 arguments as the old CONF_* functions wth the exception of the
8192 first that must be a `CONF *' instead of a `LHASH *'.
8193
8194 To make it easer to use the new classes with the old CONF_* functions,
8195 the function CONF_set_default_method is provided.
8196 [Richard Levitte]
8197
8198 *) Add '-tls1' option to 'openssl ciphers', which was already
8199 mentioned in the documentation but had not been implemented.
8200 (This option is not yet really useful because even the additional
8201 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
8202 [Bodo Moeller]
8203
8204 *) Initial DSO code added into libcrypto for letting OpenSSL (and
8205 OpenSSL-based applications) load shared libraries and bind to
8206 them in a portable way.
8207 [Geoff Thorpe, with contributions from Richard Levitte]
8208
8209 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
8210
8211 *) Make sure _lrotl and _lrotr are only used with MSVC.
8212
8213 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
8214 (the default implementation of RAND_status).
8215
8216 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
8217 to '-clrext' (= clear extensions), as intended and documented.
8218 [Bodo Moeller; inconsistency pointed out by Michael Attili
8219 <attili@amaxo.com>]
8220
8221 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
8222 was larger than the MD block size.
8223 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
8224
8225 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
8226 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
8227 using the passed key: if the passed key was a private key the result
8228 of X509_print(), for example, would be to print out all the private key
8229 components.
8230 [Steve Henson]
8231
8232 *) des_quad_cksum() byte order bug fix.
8233 [Ulf Möller, using the problem description in krb4-0.9.7, where
8234 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
8235
8236 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
8237 discouraged.
8238 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
8239
8240 *) For easily testing in shell scripts whether some command
8241 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
8242 returns with exit code 0 iff no command of the given name is available.
8243 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
8244 the output goes to stdout and nothing is printed to stderr.
8245 Additional arguments are always ignored.
8246
8247 Since for each cipher there is a command of the same name,
8248 the 'no-cipher' compilation switches can be tested this way.
8249
8250 ('openssl no-XXX' is not able to detect pseudo-commands such
8251 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
8252 [Bodo Moeller]
8253
8254 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
8255 [Bodo Moeller]
8256
8257 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
8258 is set; it will be thrown away anyway because each handshake creates
8259 its own key.
8260 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
8261 to parameters -- in previous versions (since OpenSSL 0.9.3) the
8262 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
8263 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
8264 [Bodo Moeller]
8265
8266 *) New s_client option -ign_eof: EOF at stdin is ignored, and
8267 'Q' and 'R' lose their special meanings (quit/renegotiate).
8268 This is part of what -quiet does; unlike -quiet, -ign_eof
8269 does not suppress any output.
8270 [Richard Levitte]
8271
8272 *) Add compatibility options to the purpose and trust code. The
8273 purpose X509_PURPOSE_ANY is "any purpose" which automatically
8274 accepts a certificate or CA, this was the previous behaviour,
8275 with all the associated security issues.
8276
8277 X509_TRUST_COMPAT is the old trust behaviour: only and
8278 automatically trust self signed roots in certificate store. A
8279 new trust setting X509_TRUST_DEFAULT is used to specify that
8280 a purpose has no associated trust setting and it should instead
8281 use the value in the default purpose.
8282 [Steve Henson]
8283
8284 *) Fix the PKCS#8 DSA private key code so it decodes keys again
8285 and fix a memory leak.
8286 [Steve Henson]
8287
8288 *) In util/mkerr.pl (which implements 'make errors'), preserve
8289 reason strings from the previous version of the .c file, as
8290 the default to have only downcase letters (and digits) in
8291 automatically generated reasons codes is not always appropriate.
8292 [Bodo Moeller]
8293
8294 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
8295 using strerror. Previously, ERR_reason_error_string() returned
8296 library names as reason strings for SYSerr; but SYSerr is a special
8297 case where small numbers are errno values, not library numbers.
8298 [Bodo Moeller]
8299
8300 *) Add '-dsaparam' option to 'openssl dhparam' application. This
8301 converts DSA parameters into DH parameters. (When creating parameters,
8302 DSA_generate_parameters is used.)
8303 [Bodo Moeller]
8304
8305 *) Include 'length' (recommended exponent length) in C code generated
8306 by 'openssl dhparam -C'.
8307 [Bodo Moeller]
8308
8309 *) The second argument to set_label in perlasm was already being used
8310 so couldn't be used as a "file scope" flag. Moved to third argument
8311 which was free.
8312 [Steve Henson]
8313
8314 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
8315 instead of RAND_bytes for encryption IVs and salts.
8316 [Bodo Moeller]
8317
8318 *) Include RAND_status() into RAND_METHOD instead of implementing
8319 it only for md_rand.c Otherwise replacing the PRNG by calling
8320 RAND_set_rand_method would be impossible.
8321 [Bodo Moeller]
8322
8323 *) Don't let DSA_generate_key() enter an infinite loop if the random
8324 number generation fails.
8325 [Bodo Moeller]
8326
8327 *) New 'rand' application for creating pseudo-random output.
8328 [Bodo Moeller]
8329
8330 *) Added configuration support for Linux/IA64
8331 [Rolf Haberrecker <rolf@suse.de>]
8332
8333 *) Assembler module support for Mingw32.
8334 [Ulf Möller]
8335
8336 *) Shared library support for HPUX (in shlib/).
8337 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
8338
8339 *) Shared library support for Solaris gcc.
8340 [Lutz Behnke <behnke@trustcenter.de>]
8341
8342 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
8343
8344 *) PKCS7_encrypt() was adding text MIME headers twice because they
8345 were added manually and by SMIME_crlf_copy().
8346 [Steve Henson]
8347
8348 *) In bntest.c don't call BN_rand with zero bits argument.
8349 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
8350
8351 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
8352 case was implemented. This caused BN_div_recp() to fail occasionally.
8353 [Ulf Möller]
8354
8355 *) Add an optional second argument to the set_label() in the perl
8356 assembly language builder. If this argument exists and is set
8357 to 1 it signals that the assembler should use a symbol whose
8358 scope is the entire file, not just the current function. This
8359 is needed with MASM which uses the format label:: for this scope.
8360 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
8361
8362 *) Change the ASN1 types so they are typedefs by default. Before
8363 almost all types were #define'd to ASN1_STRING which was causing
8364 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
8365 for example.
8366 [Steve Henson]
8367
8368 *) Change names of new functions to the new get1/get0 naming
8369 convention: After 'get1', the caller owns a reference count
8370 and has to call ..._free; 'get0' returns a pointer to some
8371 data structure without incrementing reference counters.
8372 (Some of the existing 'get' functions increment a reference
8373 counter, some don't.)
8374 Similarly, 'set1' and 'add1' functions increase reference
8375 counters or duplicate objects.
8376 [Steve Henson]
8377
8378 *) Allow for the possibility of temp RSA key generation failure:
8379 the code used to assume it always worked and crashed on failure.
8380 [Steve Henson]
8381
8382 *) Fix potential buffer overrun problem in BIO_printf().
8383 [Ulf Möller, using public domain code by Patrick Powell; problem
8384 pointed out by David Sacerdote <das33@cornell.edu>]
8385
8386 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
8387 RAND_egd() and RAND_status(). In the command line application,
8388 the EGD socket can be specified like a seed file using RANDFILE
8389 or -rand.
8390 [Ulf Möller]
8391
8392 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
8393 Some CAs (e.g. Verisign) distribute certificates in this form.
8394 [Steve Henson]
8395
8396 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
8397 list to exclude them. This means that no special compilation option
8398 is needed to use anonymous DH: it just needs to be included in the
8399 cipher list.
8400 [Steve Henson]
8401
8402 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
8403 EVP_MD_type. The old functionality is available in a new macro called
8404 EVP_MD_md(). Change code that uses it and update docs.
8405 [Steve Henson]
8406
8407 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
8408 where the 'void *' argument is replaced by a function pointer argument.
8409 Previously 'void *' was abused to point to functions, which works on
8410 many platforms, but is not correct. As these functions are usually
8411 called by macros defined in OpenSSL header files, most source code
8412 should work without changes.
8413 [Richard Levitte]
8414
8415 *) <openssl/opensslconf.h> (which is created by Configure) now contains
8416 sections with information on -D... compiler switches used for
8417 compiling the library so that applications can see them. To enable
8418 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
8419 must be defined. E.g.,
8420 #define OPENSSL_ALGORITHM_DEFINES
8421 #include <openssl/opensslconf.h>
8422 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
8423 [Richard Levitte, Ulf and Bodo Möller]
8424
8425 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
8426 record layer.
8427 [Bodo Moeller]
8428
8429 *) Change the 'other' type in certificate aux info to a STACK_OF
8430 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
8431 the required ASN1 format: arbitrary types determined by an OID.
8432 [Steve Henson]
8433
8434 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
8435 argument to 'req'. This is not because the function is newer or
8436 better than others it just uses the work 'NEW' in the certificate
8437 request header lines. Some software needs this.
8438 [Steve Henson]
8439
8440 *) Reorganise password command line arguments: now passwords can be
8441 obtained from various sources. Delete the PEM_cb function and make
8442 it the default behaviour: i.e. if the callback is NULL and the
8443 usrdata argument is not NULL interpret it as a null terminated pass
8444 phrase. If usrdata and the callback are NULL then the pass phrase
8445 is prompted for as usual.
8446 [Steve Henson]
8447
8448 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
8449 the support is automatically enabled. The resulting binaries will
8450 autodetect the card and use it if present.
8451 [Ben Laurie and Compaq Inc.]
8452
8453 *) Work around for Netscape hang bug. This sends certificate request
8454 and server done in one record. Since this is perfectly legal in the
8455 SSL/TLS protocol it isn't a "bug" option and is on by default. See
8456 the bugs/SSLv3 entry for more info.
8457 [Steve Henson]
8458
8459 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
8460 [Andy Polyakov]
8461
8462 *) Add -rand argument to smime and pkcs12 applications and read/write
8463 of seed file.
8464 [Steve Henson]
8465
8466 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
8467 [Bodo Moeller]
8468
8469 *) Add command line password options to the remaining applications.
8470 [Steve Henson]
8471
8472 *) Bug fix for BN_div_recp() for numerators with an even number of
8473 bits.
8474 [Ulf Möller]
8475
8476 *) More tests in bntest.c, and changed test_bn output.
8477 [Ulf Möller]
8478
8479 *) ./config recognizes MacOS X now.
8480 [Andy Polyakov]
8481
8482 *) Bug fix for BN_div() when the first words of num and divsor are
8483 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
8484 [Ulf Möller]
8485
8486 *) Add support for various broken PKCS#8 formats, and command line
8487 options to produce them.
8488 [Steve Henson]
8489
8490 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
8491 get temporary BIGNUMs from a BN_CTX.
8492 [Ulf Möller]
8493
8494 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
8495 for p == 0.
8496 [Ulf Möller]
8497
8498 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
8499 include a #define from the old name to the new. The original intent
8500 was that statically linked binaries could for example just call
8501 SSLeay_add_all_ciphers() to just add ciphers to the table and not
8502 link with digests. This never worked because SSLeay_add_all_digests()
8503 and SSLeay_add_all_ciphers() were in the same source file so calling
8504 one would link with the other. They are now in separate source files.
8505 [Steve Henson]
8506
8507 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
8508 [Steve Henson]
8509
8510 *) Use a less unusual form of the Miller-Rabin primality test (it used
8511 a binary algorithm for exponentiation integrated into the Miller-Rabin
8512 loop, our standard modexp algorithms are faster).
8513 [Bodo Moeller]
8514
8515 *) Support for the EBCDIC character set completed.
8516 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
8517
8518 *) Source code cleanups: use const where appropriate, eliminate casts,
8519 use void * instead of char * in lhash.
8520 [Ulf Möller]
8521
8522 *) Bugfix: ssl3_send_server_key_exchange was not restartable
8523 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
8524 this the server could overwrite ephemeral keys that the client
8525 has already seen).
8526 [Bodo Moeller]
8527
8528 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
8529 using 50 iterations of the Rabin-Miller test.
8530
8531 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
8532 iterations of the Rabin-Miller test as required by the appendix
8533 to FIPS PUB 186[-1]) instead of DSA_is_prime.
8534 As BN_is_prime_fasttest includes trial division, DSA parameter
8535 generation becomes much faster.
8536
8537 This implies a change for the callback functions in DSA_is_prime
8538 and DSA_generate_parameters: The callback function is called once
8539 for each positive witness in the Rabin-Miller test, not just
8540 occasionally in the inner loop; and the parameters to the
8541 callback function now provide an iteration count for the outer
8542 loop rather than for the current invocation of the inner loop.
8543 DSA_generate_parameters additionally can call the callback
8544 function with an 'iteration count' of -1, meaning that a
8545 candidate has passed the trial division test (when q is generated
8546 from an application-provided seed, trial division is skipped).
8547 [Bodo Moeller]
8548
8549 *) New function BN_is_prime_fasttest that optionally does trial
8550 division before starting the Rabin-Miller test and has
8551 an additional BN_CTX * argument (whereas BN_is_prime always
8552 has to allocate at least one BN_CTX).
8553 'callback(1, -1, cb_arg)' is called when a number has passed the
8554 trial division stage.
8555 [Bodo Moeller]
8556
8557 *) Fix for bug in CRL encoding. The validity dates weren't being handled
8558 as ASN1_TIME.
8559 [Steve Henson]
8560
8561 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
8562 [Steve Henson]
8563
8564 *) New function BN_pseudo_rand().
8565 [Ulf Möller]
8566
8567 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
8568 bignum version of BN_from_montgomery() with the working code from
8569 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
8570 the comments.
8571 [Ulf Möller]
8572
8573 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
8574 made it impossible to use the same SSL_SESSION data structure in
8575 SSL2 clients in multiple threads.
8576 [Bodo Moeller]
8577
8578 *) The return value of RAND_load_file() no longer counts bytes obtained
8579 by stat(). RAND_load_file(..., -1) is new and uses the complete file
8580 to seed the PRNG (previously an explicit byte count was required).
8581 [Ulf Möller, Bodo Möller]
8582
8583 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
8584 used (char *) instead of (void *) and had casts all over the place.
8585 [Steve Henson]
8586
8587 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
8588 [Ulf Möller]
8589
8590 *) Retain source code compatibility for BN_prime_checks macro:
8591 BN_is_prime(..., BN_prime_checks, ...) now uses
8592 BN_prime_checks_for_size to determine the appropriate number of
8593 Rabin-Miller iterations.
8594 [Ulf Möller]
8595
8596 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
8597 DH_CHECK_P_NOT_SAFE_PRIME.
8598 (Check if this is true? OpenPGP calls them "strong".)
8599 [Ulf Möller]
8600
8601 *) Merge the functionality of "dh" and "gendh" programs into a new program
8602 "dhparam". The old programs are retained for now but will handle DH keys
8603 (instead of parameters) in future.
8604 [Steve Henson]
8605
8606 *) Make the ciphers, s_server and s_client programs check the return values
8607 when a new cipher list is set.
8608 [Steve Henson]
8609
8610 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
8611 ciphers. Before when the 56bit ciphers were enabled the sorting was
8612 wrong.
8613
8614 The syntax for the cipher sorting has been extended to support sorting by
8615 cipher-strength (using the strength_bits hard coded in the tables).
8616 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
8617
8618 Fix a bug in the cipher-command parser: when supplying a cipher command
8619 string with an "undefined" symbol (neither command nor alphanumeric
8620 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
8621 an error is flagged.
8622
8623 Due to the strength-sorting extension, the code of the
8624 ssl_create_cipher_list() function was completely rearranged. I hope that
8625 the readability was also increased :-)
8626 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
8627
8628 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
8629 for the first serial number and places 2 in the serial number file. This
8630 avoids problems when the root CA is created with serial number zero and
8631 the first user certificate has the same issuer name and serial number
8632 as the root CA.
8633 [Steve Henson]
8634
8635 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
8636 the new code. Add documentation for this stuff.
8637 [Steve Henson]
8638
8639 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
8640 X509_*() to X509at_*() on the grounds that they don't handle X509
8641 structures and behave in an analogous way to the X509v3 functions:
8642 they shouldn't be called directly but wrapper functions should be used
8643 instead.
8644
8645 So we also now have some wrapper functions that call the X509at functions
8646 when passed certificate requests. (TO DO: similar things can be done with
8647 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
8648 things. Some of these need some d2i or i2d and print functionality
8649 because they handle more complex structures.)
8650 [Steve Henson]
8651
8652 *) Add missing #ifndefs that caused missing symbols when building libssl
8653 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
8654 NO_RSA in ssl/s2*.c.
8655 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
8656
8657 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
8658 has a return value which indicates the quality of the random data
8659 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
8660 error queue. New function RAND_pseudo_bytes() generates output that is
8661 guaranteed to be unique but not unpredictable. RAND_add is like
8662 RAND_seed, but takes an extra argument for an entropy estimate
8663 (RAND_seed always assumes full entropy).
8664 [Ulf Möller]
8665
8666 *) Do more iterations of Rabin-Miller probable prime test (specifically,
8667 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
8668 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
8669 in crypto/bn/bn_prime.c for the complete table). This guarantees a
8670 false-positive rate of at most 2^-80 for random input.
8671 [Bodo Moeller]
8672
8673 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
8674 [Bodo Moeller]
8675
8676 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
8677 in the 0.9.5 release), this returns the chain
8678 from an X509_CTX structure with a dup of the stack and all
8679 the X509 reference counts upped: so the stack will exist
8680 after X509_CTX_cleanup() has been called. Modify pkcs12.c
8681 to use this.
8682
8683 Also make SSL_SESSION_print() print out the verify return
8684 code.
8685 [Steve Henson]
8686
8687 *) Add manpage for the pkcs12 command. Also change the default
8688 behaviour so MAC iteration counts are used unless the new
8689 -nomaciter option is used. This improves file security and
8690 only older versions of MSIE (4.0 for example) need it.
8691 [Steve Henson]
8692
8693 *) Honor the no-xxx Configure options when creating .DEF files.
8694 [Ulf Möller]
8695
8696 *) Add PKCS#10 attributes to field table: challengePassword,
8697 unstructuredName and unstructuredAddress. These are taken from
8698 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
8699 international characters are used.
8700
8701 More changes to X509_ATTRIBUTE code: allow the setting of types
8702 based on strings. Remove the 'loc' parameter when adding
8703 attributes because these will be a SET OF encoding which is sorted
8704 in ASN1 order.
8705 [Steve Henson]
8706
8707 *) Initial changes to the 'req' utility to allow request generation
8708 automation. This will allow an application to just generate a template
8709 file containing all the field values and have req construct the
8710 request.
8711
8712 Initial support for X509_ATTRIBUTE handling. Stacks of these are
8713 used all over the place including certificate requests and PKCS#7
8714 structures. They are currently handled manually where necessary with
8715 some primitive wrappers for PKCS#7. The new functions behave in a
8716 manner analogous to the X509 extension functions: they allow
8717 attributes to be looked up by NID and added.
8718
8719 Later something similar to the X509V3 code would be desirable to
8720 automatically handle the encoding, decoding and printing of the
8721 more complex types. The string types like challengePassword can
8722 be handled by the string table functions.
8723
8724 Also modified the multi byte string table handling. Now there is
8725 a 'global mask' which masks out certain types. The table itself
8726 can use the flag STABLE_NO_MASK to ignore the mask setting: this
8727 is useful when for example there is only one permissible type
8728 (as in countryName) and using the mask might result in no valid
8729 types at all.
8730 [Steve Henson]
8731
8732 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
8733 SSL_get_peer_finished to allow applications to obtain the latest
8734 Finished messages sent to the peer or expected from the peer,
8735 respectively. (SSL_get_peer_finished is usually the Finished message
8736 actually received from the peer, otherwise the protocol will be aborted.)
8737
8738 As the Finished message are message digests of the complete handshake
8739 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
8740 be used for external authentication procedures when the authentication
8741 provided by SSL/TLS is not desired or is not enough.
8742 [Bodo Moeller]
8743
8744 *) Enhanced support for Alpha Linux is added. Now ./config checks if
8745 the host supports BWX extension and if Compaq C is present on the
8746 $PATH. Just exploiting of the BWX extension results in 20-30%
8747 performance kick for some algorithms, e.g. DES and RC4 to mention
8748 a couple. Compaq C in turn generates ~20% faster code for MD5 and
8749 SHA1.
8750 [Andy Polyakov]
8751
8752 *) Add support for MS "fast SGC". This is arguably a violation of the
8753 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
8754 weak crypto and after checking the certificate is SGC a second one
8755 with strong crypto. MS SGC stops the first handshake after receiving
8756 the server certificate message and sends a second client hello. Since
8757 a server will typically do all the time consuming operations before
8758 expecting any further messages from the client (server key exchange
8759 is the most expensive) there is little difference between the two.
8760
8761 To get OpenSSL to support MS SGC we have to permit a second client
8762 hello message after we have sent server done. In addition we have to
8763 reset the MAC if we do get this second client hello.
8764 [Steve Henson]
8765
8766 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
8767 if a DER encoded private key is RSA or DSA traditional format. Changed
8768 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
8769 format DER encoded private key. Newer code should use PKCS#8 format which
8770 has the key type encoded in the ASN1 structure. Added DER private key
8771 support to pkcs8 application.
8772 [Steve Henson]
8773
8774 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
8775 ciphersuites has been selected (as required by the SSL 3/TLS 1
8776 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
8777 is set, we interpret this as a request to violate the specification
8778 (the worst that can happen is a handshake failure, and 'correct'
8779 behaviour would result in a handshake failure anyway).
8780 [Bodo Moeller]
8781
8782 *) In SSL_CTX_add_session, take into account that there might be multiple
8783 SSL_SESSION structures with the same session ID (e.g. when two threads
8784 concurrently obtain them from an external cache).
8785 The internal cache can handle only one SSL_SESSION with a given ID,
8786 so if there's a conflict, we now throw out the old one to achieve
8787 consistency.
8788 [Bodo Moeller]
8789
8790 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
8791 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
8792 some routines that use cipher OIDs: some ciphers do not have OIDs
8793 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
8794 example.
8795 [Steve Henson]
8796
8797 *) Simplify the trust setting structure and code. Now we just have
8798 two sequences of OIDs for trusted and rejected settings. These will
8799 typically have values the same as the extended key usage extension
8800 and any application specific purposes.
8801
8802 The trust checking code now has a default behaviour: it will just
8803 check for an object with the same NID as the passed id. Functions can
8804 be provided to override either the default behaviour or the behaviour
8805 for a given id. SSL client, server and email already have functions
8806 in place for compatibility: they check the NID and also return "trusted"
8807 if the certificate is self signed.
8808 [Steve Henson]
8809
8810 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
8811 traditional format into an EVP_PKEY structure.
8812 [Steve Henson]
8813
8814 *) Add a password callback function PEM_cb() which either prompts for
8815 a password if usr_data is NULL or otherwise assumes it is a null
8816 terminated password. Allow passwords to be passed on command line
8817 environment or config files in a few more utilities.
8818 [Steve Henson]
8819
8820 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
8821 keys. Add some short names for PKCS#8 PBE algorithms and allow them
8822 to be specified on the command line for the pkcs8 and pkcs12 utilities.
8823 Update documentation.
8824 [Steve Henson]
8825
8826 *) Support for ASN1 "NULL" type. This could be handled before by using
8827 ASN1_TYPE but there wasn't any function that would try to read a NULL
8828 and produce an error if it couldn't. For compatibility we also have
8829 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
8830 don't allocate anything because they don't need to.
8831 [Steve Henson]
8832
8833 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
8834 for details.
8835 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
8836
8837 *) Rebuild of the memory allocation routines used by OpenSSL code and
8838 possibly others as well. The purpose is to make an interface that
8839 provide hooks so anyone can build a separate set of allocation and
8840 deallocation routines to be used by OpenSSL, for example memory
8841 pool implementations, or something else, which was previously hard
8842 since Malloc(), Realloc() and Free() were defined as macros having
8843 the values malloc, realloc and free, respectively (except for Win32
8844 compilations). The same is provided for memory debugging code.
8845 OpenSSL already comes with functionality to find memory leaks, but
8846 this gives people a chance to debug other memory problems.
8847
8848 With these changes, a new set of functions and macros have appeared:
8849
8850 CRYPTO_set_mem_debug_functions() [F]
8851 CRYPTO_get_mem_debug_functions() [F]
8852 CRYPTO_dbg_set_options() [F]
8853 CRYPTO_dbg_get_options() [F]
8854 CRYPTO_malloc_debug_init() [M]
8855
8856 The memory debug functions are NULL by default, unless the library
8857 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
8858 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
8859 gives the standard debugging functions that come with OpenSSL) or
8860 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
8861 provided by the library user) must be used. When the standard
8862 debugging functions are used, CRYPTO_dbg_set_options can be used to
8863 request additional information:
8864 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
8865 the CRYPTO_MDEBUG_xxx macro when compiling the library.
8866
8867 Also, things like CRYPTO_set_mem_functions will always give the
8868 expected result (the new set of functions is used for allocation
8869 and deallocation) at all times, regardless of platform and compiler
8870 options.
8871
8872 To finish it up, some functions that were never use in any other
8873 way than through macros have a new API and new semantic:
8874
8875 CRYPTO_dbg_malloc()
8876 CRYPTO_dbg_realloc()
8877 CRYPTO_dbg_free()
8878
8879 All macros of value have retained their old syntax.
8880 [Richard Levitte and Bodo Moeller]
8881
8882 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
8883 ordering of SMIMECapabilities wasn't in "strength order" and there
8884 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
8885 algorithm.
8886 [Steve Henson]
8887
8888 *) Some ASN1 types with illegal zero length encoding (INTEGER,
8889 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
8890 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
8891
8892 *) Merge in my S/MIME library for OpenSSL. This provides a simple
8893 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
8894 functionality to handle multipart/signed properly) and a utility
8895 called 'smime' to call all this stuff. This is based on code I
8896 originally wrote for Celo who have kindly allowed it to be
8897 included in OpenSSL.
8898 [Steve Henson]
8899
8900 *) Add variants des_set_key_checked and des_set_key_unchecked of
8901 des_set_key (aka des_key_sched). Global variable des_check_key
8902 decides which of these is called by des_set_key; this way
8903 des_check_key behaves as it always did, but applications and
8904 the library itself, which was buggy for des_check_key == 1,
8905 have a cleaner way to pick the version they need.
8906 [Bodo Moeller]
8907
8908 *) New function PKCS12_newpass() which changes the password of a
8909 PKCS12 structure.
8910 [Steve Henson]
8911
8912 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
8913 dynamic mix. In both cases the ids can be used as an index into the
8914 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
8915 functions so they accept a list of the field values and the
8916 application doesn't need to directly manipulate the X509_TRUST
8917 structure.
8918 [Steve Henson]
8919
8920 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
8921 need initialising.
8922 [Steve Henson]
8923
8924 *) Modify the way the V3 extension code looks up extensions. This now
8925 works in a similar way to the object code: we have some "standard"
8926 extensions in a static table which is searched with OBJ_bsearch()
8927 and the application can add dynamic ones if needed. The file
8928 crypto/x509v3/ext_dat.h now has the info: this file needs to be
8929 updated whenever a new extension is added to the core code and kept
8930 in ext_nid order. There is a simple program 'tabtest.c' which checks
8931 this. New extensions are not added too often so this file can readily
8932 be maintained manually.
8933
8934 There are two big advantages in doing things this way. The extensions
8935 can be looked up immediately and no longer need to be "added" using
8936 X509V3_add_standard_extensions(): this function now does nothing.
8937 [Side note: I get *lots* of email saying the extension code doesn't
8938 work because people forget to call this function]
8939 Also no dynamic allocation is done unless new extensions are added:
8940 so if we don't add custom extensions there is no need to call
8941 X509V3_EXT_cleanup().
8942 [Steve Henson]
8943
8944 *) Modify enc utility's salting as follows: make salting the default. Add a
8945 magic header, so unsalted files fail gracefully instead of just decrypting
8946 to garbage. This is because not salting is a big security hole, so people
8947 should be discouraged from doing it.
8948 [Ben Laurie]
8949
8950 *) Fixes and enhancements to the 'x509' utility. It allowed a message
8951 digest to be passed on the command line but it only used this
8952 parameter when signing a certificate. Modified so all relevant
8953 operations are affected by the digest parameter including the
8954 -fingerprint and -x509toreq options. Also -x509toreq choked if a
8955 DSA key was used because it didn't fix the digest.
8956 [Steve Henson]
8957
8958 *) Initial certificate chain verify code. Currently tests the untrusted
8959 certificates for consistency with the verify purpose (which is set
8960 when the X509_STORE_CTX structure is set up) and checks the pathlength.
8961
8962 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
8963 this is because it will reject chains with invalid extensions whereas
8964 every previous version of OpenSSL and SSLeay made no checks at all.
8965
8966 Trust code: checks the root CA for the relevant trust settings. Trust
8967 settings have an initial value consistent with the verify purpose: e.g.
8968 if the verify purpose is for SSL client use it expects the CA to be
8969 trusted for SSL client use. However the default value can be changed to
8970 permit custom trust settings: one example of this would be to only trust
8971 certificates from a specific "secure" set of CAs.
8972
8973 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
8974 which should be used for version portability: especially since the
8975 verify structure is likely to change more often now.
8976
8977 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
8978 to set them. If not set then assume SSL clients will verify SSL servers
8979 and vice versa.
8980
8981 Two new options to the verify program: -untrusted allows a set of
8982 untrusted certificates to be passed in and -purpose which sets the
8983 intended purpose of the certificate. If a purpose is set then the
8984 new chain verify code is used to check extension consistency.
8985 [Steve Henson]
8986
8987 *) Support for the authority information access extension.
8988 [Steve Henson]
8989
8990 *) Modify RSA and DSA PEM read routines to transparently handle
8991 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
8992 public keys in a format compatible with certificate
8993 SubjectPublicKeyInfo structures. Unfortunately there were already
8994 functions called *_PublicKey_* which used various odd formats so
8995 these are retained for compatibility: however the DSA variants were
8996 never in a public release so they have been deleted. Changed dsa/rsa
8997 utilities to handle the new format: note no releases ever handled public
8998 keys so we should be OK.
8999
9000 The primary motivation for this change is to avoid the same fiasco
9001 that dogs private keys: there are several incompatible private key
9002 formats some of which are standard and some OpenSSL specific and
9003 require various evil hacks to allow partial transparent handling and
9004 even then it doesn't work with DER formats. Given the option anything
9005 other than PKCS#8 should be dumped: but the other formats have to
9006 stay in the name of compatibility.
9007
9008 With public keys and the benefit of hindsight one standard format
9009 is used which works with EVP_PKEY, RSA or DSA structures: though
9010 it clearly returns an error if you try to read the wrong kind of key.
9011
9012 Added a -pubkey option to the 'x509' utility to output the public key.
9013 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
9014 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
9015 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
9016 that do the same as the EVP_PKEY_assign_*() except they up the
9017 reference count of the added key (they don't "swallow" the
9018 supplied key).
9019 [Steve Henson]
9020
9021 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
9022 CRLs would fail if the file contained no certificates or no CRLs:
9023 added a new function to read in both types and return the number
9024 read: this means that if none are read it will be an error. The
9025 DER versions of the certificate and CRL reader would always fail
9026 because it isn't possible to mix certificates and CRLs in DER format
9027 without choking one or the other routine. Changed this to just read
9028 a certificate: this is the best we can do. Also modified the code
9029 in apps/verify.c to take notice of return codes: it was previously
9030 attempting to read in certificates from NULL pointers and ignoring
9031 any errors: this is one reason why the cert and CRL reader seemed
9032 to work. It doesn't check return codes from the default certificate
9033 routines: these may well fail if the certificates aren't installed.
9034 [Steve Henson]
9035
9036 *) Code to support otherName option in GeneralName.
9037 [Steve Henson]
9038
9039 *) First update to verify code. Change the verify utility
9040 so it warns if it is passed a self signed certificate:
9041 for consistency with the normal behaviour. X509_verify
9042 has been modified to it will now verify a self signed
9043 certificate if *exactly* the same certificate appears
9044 in the store: it was previously impossible to trust a
9045 single self signed certificate. This means that:
9046 openssl verify ss.pem
9047 now gives a warning about a self signed certificate but
9048 openssl verify -CAfile ss.pem ss.pem
9049 is OK.
9050 [Steve Henson]
9051
9052 *) For servers, store verify_result in SSL_SESSION data structure
9053 (and add it to external session representation).
9054 This is needed when client certificate verifications fails,
9055 but an application-provided verification callback (set by
9056 SSL_CTX_set_cert_verify_callback) allows accepting the session
9057 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
9058 but returns 1): When the session is reused, we have to set
9059 ssl->verify_result to the appropriate error code to avoid
9060 security holes.
9061 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
9062
9063 *) Fix a bug in the new PKCS#7 code: it didn't consider the
9064 case in PKCS7_dataInit() where the signed PKCS7 structure
9065 didn't contain any existing data because it was being created.
9066 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
9067
9068 *) Add a salt to the key derivation routines in enc.c. This
9069 forms the first 8 bytes of the encrypted file. Also add a
9070 -S option to allow a salt to be input on the command line.
9071 [Steve Henson]
9072
9073 *) New function X509_cmp(). Oddly enough there wasn't a function
9074 to compare two certificates. We do this by working out the SHA1
9075 hash and comparing that. X509_cmp() will be needed by the trust
9076 code.
9077 [Steve Henson]
9078
9079 *) SSL_get1_session() is like SSL_get_session(), but increments
9080 the reference count in the SSL_SESSION returned.
9081 [Geoff Thorpe <geoff@eu.c2.net>]
9082
9083 *) Fix for 'req': it was adding a null to request attributes.
9084 Also change the X509_LOOKUP and X509_INFO code to handle
9085 certificate auxiliary information.
9086 [Steve Henson]
9087
9088 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
9089 the 'enc' command.
9090 [Steve Henson]
9091
9092 *) Add the possibility to add extra information to the memory leak
9093 detecting output, to form tracebacks, showing from where each
9094 allocation was originated: CRYPTO_push_info("constant string") adds
9095 the string plus current file name and line number to a per-thread
9096 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
9097 is like calling CYRPTO_pop_info() until the stack is empty.
9098 Also updated memory leak detection code to be multi-thread-safe.
9099 [Richard Levitte]
9100
9101 *) Add options -text and -noout to pkcs7 utility and delete the
9102 encryption options which never did anything. Update docs.
9103 [Steve Henson]
9104
9105 *) Add options to some of the utilities to allow the pass phrase
9106 to be included on either the command line (not recommended on
9107 OSes like Unix) or read from the environment. Update the
9108 manpages and fix a few bugs.
9109 [Steve Henson]
9110
9111 *) Add a few manpages for some of the openssl commands.
9112 [Steve Henson]
9113
9114 *) Fix the -revoke option in ca. It was freeing up memory twice,
9115 leaking and not finding already revoked certificates.
9116 [Steve Henson]
9117
9118 *) Extensive changes to support certificate auxiliary information.
9119 This involves the use of X509_CERT_AUX structure and X509_AUX
9120 functions. An X509_AUX function such as PEM_read_X509_AUX()
9121 can still read in a certificate file in the usual way but it
9122 will also read in any additional "auxiliary information". By
9123 doing things this way a fair degree of compatibility can be
9124 retained: existing certificates can have this information added
9125 using the new 'x509' options.
9126
9127 Current auxiliary information includes an "alias" and some trust
9128 settings. The trust settings will ultimately be used in enhanced
9129 certificate chain verification routines: currently a certificate
9130 can only be trusted if it is self signed and then it is trusted
9131 for all purposes.
9132 [Steve Henson]
9133
9134 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
9135 The problem was that one of the replacement routines had not been working
9136 since SSLeay releases. For now the offending routine has been replaced
9137 with non-optimised assembler. Even so, this now gives around 95%
9138 performance improvement for 1024 bit RSA signs.
9139 [Mark Cox]
9140
9141 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
9142 handling. Most clients have the effective key size in bits equal to
9143 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
9144 A few however don't do this and instead use the size of the decrypted key
9145 to determine the RC2 key length and the AlgorithmIdentifier to determine
9146 the effective key length. In this case the effective key length can still
9147 be 40 bits but the key length can be 168 bits for example. This is fixed
9148 by manually forcing an RC2 key into the EVP_PKEY structure because the
9149 EVP code can't currently handle unusual RC2 key sizes: it always assumes
9150 the key length and effective key length are equal.
9151 [Steve Henson]
9152
9153 *) Add a bunch of functions that should simplify the creation of
9154 X509_NAME structures. Now you should be able to do:
9155 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
9156 and have it automatically work out the correct field type and fill in
9157 the structures. The more adventurous can try:
9158 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
9159 and it will (hopefully) work out the correct multibyte encoding.
9160 [Steve Henson]
9161
9162 *) Change the 'req' utility to use the new field handling and multibyte
9163 copy routines. Before the DN field creation was handled in an ad hoc
9164 way in req, ca, and x509 which was rather broken and didn't support
9165 BMPStrings or UTF8Strings. Since some software doesn't implement
9166 BMPStrings or UTF8Strings yet, they can be enabled using the config file
9167 using the dirstring_type option. See the new comment in the default
9168 openssl.cnf for more info.
9169 [Steve Henson]
9170
9171 *) Make crypto/rand/md_rand.c more robust:
9172 - Assure unique random numbers after fork().
9173 - Make sure that concurrent threads access the global counter and
9174 md serializably so that we never lose entropy in them
9175 or use exactly the same state in multiple threads.
9176 Access to the large state is not always serializable because
9177 the additional locking could be a performance killer, and
9178 md should be large enough anyway.
9179 [Bodo Moeller]
9180
9181 *) New file apps/app_rand.c with commonly needed functionality
9182 for handling the random seed file.
9183
9184 Use the random seed file in some applications that previously did not:
9185 ca,
9186 dsaparam -genkey (which also ignored its '-rand' option),
9187 s_client,
9188 s_server,
9189 x509 (when signing).
9190 Except on systems with /dev/urandom, it is crucial to have a random
9191 seed file at least for key creation, DSA signing, and for DH exchanges;
9192 for RSA signatures we could do without one.
9193
9194 gendh and gendsa (unlike genrsa) used to read only the first byte
9195 of each file listed in the '-rand' option. The function as previously
9196 found in genrsa is now in app_rand.c and is used by all programs
9197 that support '-rand'.
9198 [Bodo Moeller]
9199
9200 *) In RAND_write_file, use mode 0600 for creating files;
9201 don't just chmod when it may be too late.
9202 [Bodo Moeller]
9203
9204 *) Report an error from X509_STORE_load_locations
9205 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
9206 [Bill Perry]
9207
9208 *) New function ASN1_mbstring_copy() this copies a string in either
9209 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
9210 into an ASN1_STRING type. A mask of permissible types is passed
9211 and it chooses the "minimal" type to use or an error if not type
9212 is suitable.
9213 [Steve Henson]
9214
9215 *) Add function equivalents to the various macros in asn1.h. The old
9216 macros are retained with an M_ prefix. Code inside the library can
9217 use the M_ macros. External code (including the openssl utility)
9218 should *NOT* in order to be "shared library friendly".
9219 [Steve Henson]
9220
9221 *) Add various functions that can check a certificate's extensions
9222 to see if it usable for various purposes such as SSL client,
9223 server or S/MIME and CAs of these types. This is currently
9224 VERY EXPERIMENTAL but will ultimately be used for certificate chain
9225 verification. Also added a -purpose flag to x509 utility to
9226 print out all the purposes.
9227 [Steve Henson]
9228
9229 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
9230 functions.
9231 [Steve Henson]
9232
9233 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
9234 for, obtain and decode and extension and obtain its critical flag.
9235 This allows all the necessary extension code to be handled in a
9236 single function call.
9237 [Steve Henson]
9238
9239 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
9240 platforms. See crypto/rc4/rc4_enc.c for further details.
9241 [Andy Polyakov]
9242
9243 *) New -noout option to asn1parse. This causes no output to be produced
9244 its main use is when combined with -strparse and -out to extract data
9245 from a file (which may not be in ASN.1 format).
9246 [Steve Henson]
9247
9248 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
9249 when producing the local key id.
9250 [Richard Levitte <levitte@stacken.kth.se>]
9251
9252 *) New option -dhparam in s_server. This allows a DH parameter file to be
9253 stated explicitly. If it is not stated then it tries the first server
9254 certificate file. The previous behaviour hard coded the filename
9255 "server.pem".
9256 [Steve Henson]
9257
9258 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
9259 a public key to be input or output. For example:
9260 openssl rsa -in key.pem -pubout -out pubkey.pem
9261 Also added necessary DSA public key functions to handle this.
9262 [Steve Henson]
9263
9264 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
9265 in the message. This was handled by allowing
9266 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
9267 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
9268
9269 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
9270 to the end of the strings whereas this didn't. This would cause problems
9271 if strings read with d2i_ASN1_bytes() were later modified.
9272 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
9273
9274 *) Fix for base64 decode bug. When a base64 bio reads only one line of
9275 data and it contains EOF it will end up returning an error. This is
9276 caused by input 46 bytes long. The cause is due to the way base64
9277 BIOs find the start of base64 encoded data. They do this by trying a
9278 trial decode on each line until they find one that works. When they
9279 do a flag is set and it starts again knowing it can pass all the
9280 data directly through the decoder. Unfortunately it doesn't reset
9281 the context it uses. This means that if EOF is reached an attempt
9282 is made to pass two EOFs through the context and this causes the
9283 resulting error. This can also cause other problems as well. As is
9284 usual with these problems it takes *ages* to find and the fix is
9285 trivial: move one line.
9286 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
9287
9288 *) Ugly workaround to get s_client and s_server working under Windows. The
9289 old code wouldn't work because it needed to select() on sockets and the
9290 tty (for keypresses and to see if data could be written). Win32 only
9291 supports select() on sockets so we select() with a 1s timeout on the
9292 sockets and then see if any characters are waiting to be read, if none
9293 are present then we retry, we also assume we can always write data to
9294 the tty. This isn't nice because the code then blocks until we've
9295 received a complete line of data and it is effectively polling the
9296 keyboard at 1s intervals: however it's quite a bit better than not
9297 working at all :-) A dedicated Windows application might handle this
9298 with an event loop for example.
9299 [Steve Henson]
9300
9301 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
9302 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
9303 will be called when RSA_sign() and RSA_verify() are used. This is useful
9304 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
9305 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
9306 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
9307 This necessitated the support of an extra signature type NID_md5_sha1
9308 for SSL signatures and modifications to the SSL library to use it instead
9309 of calling RSA_public_decrypt() and RSA_private_encrypt().
9310 [Steve Henson]
9311
9312 *) Add new -verify -CAfile and -CApath options to the crl program, these
9313 will lookup a CRL issuers certificate and verify the signature in a
9314 similar way to the verify program. Tidy up the crl program so it
9315 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
9316 less strict. It will now permit CRL extensions even if it is not
9317 a V2 CRL: this will allow it to tolerate some broken CRLs.
9318 [Steve Henson]
9319
9320 *) Initialize all non-automatic variables each time one of the openssl
9321 sub-programs is started (this is necessary as they may be started
9322 multiple times from the "OpenSSL>" prompt).
9323 [Lennart Bang, Bodo Moeller]
9324
9325 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
9326 removing all other RSA functionality (this is what NO_RSA does). This
9327 is so (for example) those in the US can disable those operations covered
9328 by the RSA patent while allowing storage and parsing of RSA keys and RSA
9329 key generation.
9330 [Steve Henson]
9331
9332 *) Non-copying interface to BIO pairs.
9333 (still largely untested)
9334 [Bodo Moeller]
9335
9336 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
9337 ASCII string. This was handled independently in various places before.
9338 [Steve Henson]
9339
9340 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
9341 UTF8 strings a character at a time.
9342 [Steve Henson]
9343
9344 *) Use client_version from client hello to select the protocol
9345 (s23_srvr.c) and for RSA client key exchange verification
9346 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
9347 [Bodo Moeller]
9348
9349 *) Add various utility functions to handle SPKACs, these were previously
9350 handled by poking round in the structure internals. Added new function
9351 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
9352 print, verify and generate SPKACs. Based on an original idea from
9353 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
9354 [Steve Henson]
9355
9356 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
9357 [Andy Polyakov]
9358
9359 *) Allow the config file extension section to be overwritten on the
9360 command line. Based on an original idea from Massimiliano Pala
9361 <madwolf@comune.modena.it>. The new option is called -extensions
9362 and can be applied to ca, req and x509. Also -reqexts to override
9363 the request extensions in req and -crlexts to override the crl extensions
9364 in ca.
9365 [Steve Henson]
9366
9367 *) Add new feature to the SPKAC handling in ca. Now you can include
9368 the same field multiple times by preceding it by "XXXX." for example:
9369 1.OU="Unit name 1"
9370 2.OU="Unit name 2"
9371 this is the same syntax as used in the req config file.
9372 [Steve Henson]
9373
9374 *) Allow certificate extensions to be added to certificate requests. These
9375 are specified in a 'req_extensions' option of the req section of the
9376 config file. They can be printed out with the -text option to req but
9377 are otherwise ignored at present.
9378 [Steve Henson]
9379
9380 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
9381 data read consists of only the final block it would not decrypted because
9382 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
9383 A misplaced 'break' also meant the decrypted final block might not be
9384 copied until the next read.
9385 [Steve Henson]
9386
9387 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
9388 a few extra parameters to the DH structure: these will be useful if
9389 for example we want the value of 'q' or implement X9.42 DH.
9390 [Steve Henson]
9391
9392 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
9393 provides hooks that allow the default DSA functions or functions on a
9394 "per key" basis to be replaced. This allows hardware acceleration and
9395 hardware key storage to be handled without major modification to the
9396 library. Also added low level modexp hooks and CRYPTO_EX structure and
9397 associated functions.
9398 [Steve Henson]
9399
9400 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
9401 as "read only": it can't be written to and the buffer it points to will
9402 not be freed. Reading from a read only BIO is much more efficient than
9403 a normal memory BIO. This was added because there are several times when
9404 an area of memory needs to be read from a BIO. The previous method was
9405 to create a memory BIO and write the data to it, this results in two
9406 copies of the data and an O(n^2) reading algorithm. There is a new
9407 function BIO_new_mem_buf() which creates a read only memory BIO from
9408 an area of memory. Also modified the PKCS#7 routines to use read only
9409 memory BIOs.
9410 [Steve Henson]
9411
9412 *) Bugfix: ssl23_get_client_hello did not work properly when called in
9413 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
9414 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
9415 but a retry condition occured while trying to read the rest.
9416 [Bodo Moeller]
9417
9418 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
9419 NID_pkcs7_encrypted by default: this was wrong since this should almost
9420 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
9421 the encrypted data type: this is a more sensible place to put it and it
9422 allows the PKCS#12 code to be tidied up that duplicated this
9423 functionality.
9424 [Steve Henson]
9425
9426 *) Changed obj_dat.pl script so it takes its input and output files on
9427 the command line. This should avoid shell escape redirection problems
9428 under Win32.
9429 [Steve Henson]
9430
9431 *) Initial support for certificate extension requests, these are included
9432 in things like Xenroll certificate requests. Included functions to allow
9433 extensions to be obtained and added.
9434 [Steve Henson]
9435
9436 *) -crlf option to s_client and s_server for sending newlines as
9437 CRLF (as required by many protocols).
9438 [Bodo Moeller]
9439
9440 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
9441
9442 *) Install libRSAglue.a when OpenSSL is built with RSAref.
9443 [Ralf S. Engelschall]
9444
9445 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
9446 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
9447
9448 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
9449 program.
9450 [Steve Henson]
9451
9452 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
9453 DH parameters/keys (q is lost during that conversion, but the resulting
9454 DH parameters contain its length).
9455
9456 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
9457 much faster than DH_generate_parameters (which creates parameters
9458 where p = 2*q + 1), and also the smaller q makes DH computations
9459 much more efficient (160-bit exponentiation instead of 1024-bit
9460 exponentiation); so this provides a convenient way to support DHE
9461 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
9462 utter importance to use
9463 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9464 or
9465 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
9466 when such DH parameters are used, because otherwise small subgroup
9467 attacks may become possible!
9468 [Bodo Moeller]
9469
9470 *) Avoid memory leak in i2d_DHparams.
9471 [Bodo Moeller]
9472
9473 *) Allow the -k option to be used more than once in the enc program:
9474 this allows the same encrypted message to be read by multiple recipients.
9475 [Steve Henson]
9476
9477 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
9478 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
9479 it will always use the numerical form of the OID, even if it has a short
9480 or long name.
9481 [Steve Henson]
9482
9483 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
9484 method only got called if p,q,dmp1,dmq1,iqmp components were present,
9485 otherwise bn_mod_exp was called. In the case of hardware keys for example
9486 no private key components need be present and it might store extra data
9487 in the RSA structure, which cannot be accessed from bn_mod_exp.
9488 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
9489 private key operations.
9490 [Steve Henson]
9491
9492 *) Added support for SPARC Linux.
9493 [Andy Polyakov]
9494
9495 *) pem_password_cb function type incompatibly changed from
9496 typedef int pem_password_cb(char *buf, int size, int rwflag);
9497 to
9498 ....(char *buf, int size, int rwflag, void *userdata);
9499 so that applications can pass data to their callbacks:
9500 The PEM[_ASN1]_{read,write}... functions and macros now take an
9501 additional void * argument, which is just handed through whenever
9502 the password callback is called.
9503 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
9504
9505 New function SSL_CTX_set_default_passwd_cb_userdata.
9506
9507 Compatibility note: As many C implementations push function arguments
9508 onto the stack in reverse order, the new library version is likely to
9509 interoperate with programs that have been compiled with the old
9510 pem_password_cb definition (PEM_whatever takes some data that
9511 happens to be on the stack as its last argument, and the callback
9512 just ignores this garbage); but there is no guarantee whatsoever that
9513 this will work.
9514
9515 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
9516 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
9517 problems not only on Windows, but also on some Unix platforms.
9518 To avoid problematic command lines, these definitions are now in an
9519 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
9520 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
9521 [Bodo Moeller]
9522
9523 *) MIPS III/IV assembler module is reimplemented.
9524 [Andy Polyakov]
9525
9526 *) More DES library cleanups: remove references to srand/rand and
9527 delete an unused file.
9528 [Ulf Möller]
9529
9530 *) Add support for the the free Netwide assembler (NASM) under Win32,
9531 since not many people have MASM (ml) and it can be hard to obtain.
9532 This is currently experimental but it seems to work OK and pass all
9533 the tests. Check out INSTALL.W32 for info.
9534 [Steve Henson]
9535
9536 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
9537 without temporary keys kept an extra copy of the server key,
9538 and connections with temporary keys did not free everything in case
9539 of an error.
9540 [Bodo Moeller]
9541
9542 *) New function RSA_check_key and new openssl rsa option -check
9543 for verifying the consistency of RSA keys.
9544 [Ulf Moeller, Bodo Moeller]
9545
9546 *) Various changes to make Win32 compile work:
9547 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
9548 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
9549 comparison" warnings.
9550 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
9551 [Steve Henson]
9552
9553 *) Add a debugging option to PKCS#5 v2 key generation function: when
9554 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
9555 derived keys are printed to stderr.
9556 [Steve Henson]
9557
9558 *) Copy the flags in ASN1_STRING_dup().
9559 [Roman E. Pavlov <pre@mo.msk.ru>]
9560
9561 *) The x509 application mishandled signing requests containing DSA
9562 keys when the signing key was also DSA and the parameters didn't match.
9563
9564 It was supposed to omit the parameters when they matched the signing key:
9565 the verifying software was then supposed to automatically use the CA's
9566 parameters if they were absent from the end user certificate.
9567
9568 Omitting parameters is no longer recommended. The test was also
9569 the wrong way round! This was probably due to unusual behaviour in
9570 EVP_cmp_parameters() which returns 1 if the parameters match.
9571 This meant that parameters were omitted when they *didn't* match and
9572 the certificate was useless. Certificates signed with 'ca' didn't have
9573 this bug.
9574 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
9575
9576 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
9577 The interface is as follows:
9578 Applications can use
9579 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
9580 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
9581 "off" is now the default.
9582 The library internally uses
9583 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
9584 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
9585 to disable memory-checking temporarily.
9586
9587 Some inconsistent states that previously were possible (and were
9588 even the default) are now avoided.
9589
9590 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
9591 with each memory chunk allocated; this is occasionally more helpful
9592 than just having a counter.
9593
9594 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
9595
9596 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
9597 extensions.
9598 [Bodo Moeller]
9599
9600 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
9601 which largely parallels "options", but is for changing API behaviour,
9602 whereas "options" are about protocol behaviour.
9603 Initial "mode" flags are:
9604
9605 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
9606 a single record has been written.
9607 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
9608 retries use the same buffer location.
9609 (But all of the contents must be
9610 copied!)
9611 [Bodo Moeller]
9612
9613 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
9614 worked.
9615
9616 *) Fix problems with no-hmac etc.
9617 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
9618
9619 *) New functions RSA_get_default_method(), RSA_set_method() and
9620 RSA_get_method(). These allows replacement of RSA_METHODs without having
9621 to mess around with the internals of an RSA structure.
9622 [Steve Henson]
9623
9624 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
9625 Also really enable memory leak checks in openssl.c and in some
9626 test programs.
9627 [Chad C. Mulligan, Bodo Moeller]
9628
9629 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
9630 up the length of negative integers. This has now been simplified to just
9631 store the length when it is first determined and use it later, rather
9632 than trying to keep track of where data is copied and updating it to
9633 point to the end.
9634 [Steve Henson, reported by Brien Wheeler
9635 <bwheeler@authentica-security.com>]
9636
9637 *) Add a new function PKCS7_signatureVerify. This allows the verification
9638 of a PKCS#7 signature but with the signing certificate passed to the
9639 function itself. This contrasts with PKCS7_dataVerify which assumes the
9640 certificate is present in the PKCS#7 structure. This isn't always the
9641 case: certificates can be omitted from a PKCS#7 structure and be
9642 distributed by "out of band" means (such as a certificate database).
9643 [Steve Henson]
9644
9645 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
9646 function prototypes in pem.h, also change util/mkdef.pl to add the
9647 necessary function names.
9648 [Steve Henson]
9649
9650 *) mk1mf.pl (used by Windows builds) did not properly read the
9651 options set by Configure in the top level Makefile, and Configure
9652 was not even able to write more than one option correctly.
9653 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
9654 [Bodo Moeller]
9655
9656 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
9657 file to be loaded from a BIO or FILE pointer. The BIO version will
9658 for example allow memory BIOs to contain config info.
9659 [Steve Henson]
9660
9661 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
9662 Whoever hopes to achieve shared-library compatibility across versions
9663 must use this, not the compile-time macro.
9664 (Exercise 0.9.4: Which is the minimum library version required by
9665 such programs?)
9666 Note: All this applies only to multi-threaded programs, others don't
9667 need locks.
9668 [Bodo Moeller]
9669
9670 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
9671 through a BIO pair triggered the default case, i.e.
9672 SSLerr(...,SSL_R_UNKNOWN_STATE).
9673 [Bodo Moeller]
9674
9675 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
9676 can use the SSL library even if none of the specific BIOs is
9677 appropriate.
9678 [Bodo Moeller]
9679
9680 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
9681 for the encoded length.
9682 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
9683
9684 *) Add initial documentation of the X509V3 functions.
9685 [Steve Henson]
9686
9687 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
9688 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
9689 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
9690 secure PKCS#8 private key format with a high iteration count.
9691 [Steve Henson]
9692
9693 *) Fix determination of Perl interpreter: A perl or perl5
9694 _directory_ in $PATH was also accepted as the interpreter.
9695 [Ralf S. Engelschall]
9696
9697 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
9698 wrong with it but it was very old and did things like calling
9699 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
9700 unusual formatting.
9701 [Steve Henson]
9702
9703 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
9704 to use the new extension code.
9705 [Steve Henson]
9706
9707 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
9708 with macros. This should make it easier to change their form, add extra
9709 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
9710 constant.
9711 [Steve Henson]
9712
9713 *) Add to configuration table a new entry that can specify an alternative
9714 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
9715 according to Mark Crispin <MRC@Panda.COM>.
9716 [Bodo Moeller]
9717
9718 #if 0
9719 *) DES CBC did not update the IV. Weird.
9720 [Ben Laurie]
9721 #else
9722 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
9723 Changing the behaviour of the former might break existing programs --
9724 where IV updating is needed, des_ncbc_encrypt can be used.
9725 #endif
9726
9727 *) When bntest is run from "make test" it drives bc to check its
9728 calculations, as well as internally checking them. If an internal check
9729 fails, it needs to cause bc to give a non-zero result or make test carries
9730 on without noticing the failure. Fixed.
9731 [Ben Laurie]
9732
9733 *) DES library cleanups.
9734 [Ulf Möller]
9735
9736 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
9737 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
9738 ciphers. NOTE: although the key derivation function has been verified
9739 against some published test vectors it has not been extensively tested
9740 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
9741 of v2.0.
9742 [Steve Henson]
9743
9744 *) Instead of "mkdir -p", which is not fully portable, use new
9745 Perl script "util/mkdir-p.pl".
9746 [Bodo Moeller]
9747
9748 *) Rewrite the way password based encryption (PBE) is handled. It used to
9749 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
9750 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
9751 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
9752 the 'parameter' field of the AlgorithmIdentifier is passed to the
9753 underlying key generation function so it must do its own ASN1 parsing.
9754 This has also changed the EVP_PBE_CipherInit() function which now has a
9755 'parameter' argument instead of literal salt and iteration count values
9756 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
9757 [Steve Henson]
9758
9759 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
9760 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
9761 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
9762 KEY" because this clashed with PKCS#8 unencrypted string. Since this
9763 value was just used as a "magic string" and not used directly its
9764 value doesn't matter.
9765 [Steve Henson]
9766
9767 *) Introduce some semblance of const correctness to BN. Shame C doesn't
9768 support mutable.
9769 [Ben Laurie]
9770
9771 *) "linux-sparc64" configuration (ultrapenguin).
9772 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
9773 "linux-sparc" configuration.
9774 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
9775
9776 *) config now generates no-xxx options for missing ciphers.
9777 [Ulf Möller]
9778
9779 *) Support the EBCDIC character set (work in progress).
9780 File ebcdic.c not yet included because it has a different license.
9781 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9782
9783 *) Support BS2000/OSD-POSIX.
9784 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9785
9786 *) Make callbacks for key generation use void * instead of char *.
9787 [Ben Laurie]
9788
9789 *) Make S/MIME samples compile (not yet tested).
9790 [Ben Laurie]
9791
9792 *) Additional typesafe stacks.
9793 [Ben Laurie]
9794
9795 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
9796 [Bodo Moeller]
9797
9798
9799 Changes between 0.9.3 and 0.9.3a [29 May 1999]
9800
9801 *) New configuration variant "sco5-gcc".
9802
9803 *) Updated some demos.
9804 [Sean O Riordain, Wade Scholine]
9805
9806 *) Add missing BIO_free at exit of pkcs12 application.
9807 [Wu Zhigang]
9808
9809 *) Fix memory leak in conf.c.
9810 [Steve Henson]
9811
9812 *) Updates for Win32 to assembler version of MD5.
9813 [Steve Henson]
9814
9815 *) Set #! path to perl in apps/der_chop to where we found it
9816 instead of using a fixed path.
9817 [Bodo Moeller]
9818
9819 *) SHA library changes for irix64-mips4-cc.
9820 [Andy Polyakov]
9821
9822 *) Improvements for VMS support.
9823 [Richard Levitte]
9824
9825
9826 Changes between 0.9.2b and 0.9.3 [24 May 1999]
9827
9828 *) Bignum library bug fix. IRIX 6 passes "make test" now!
9829 This also avoids the problems with SC4.2 and unpatched SC5.
9830 [Andy Polyakov <appro@fy.chalmers.se>]
9831
9832 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
9833 These are required because of the typesafe stack would otherwise break
9834 existing code. If old code used a structure member which used to be STACK
9835 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
9836 sk_num or sk_value it would produce an error because the num, data members
9837 are not present in STACK_OF. Now it just produces a warning. sk_set
9838 replaces the old method of assigning a value to sk_value
9839 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
9840 that does this will no longer work (and should use sk_set instead) but
9841 this could be regarded as a "questionable" behaviour anyway.
9842 [Steve Henson]
9843
9844 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
9845 correctly handle encrypted S/MIME data.
9846 [Steve Henson]
9847
9848 *) Change type of various DES function arguments from des_cblock
9849 (which means, in function argument declarations, pointer to char)
9850 to des_cblock * (meaning pointer to array with 8 char elements),
9851 which allows the compiler to do more typechecking; it was like
9852 that back in SSLeay, but with lots of ugly casts.
9853
9854 Introduce new type const_des_cblock.
9855 [Bodo Moeller]
9856
9857 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
9858 problems: find RecipientInfo structure that matches recipient certificate
9859 and initialise the ASN1 structures properly based on passed cipher.
9860 [Steve Henson]
9861
9862 *) Belatedly make the BN tests actually check the results.
9863 [Ben Laurie]
9864
9865 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
9866 to and from BNs: it was completely broken. New compilation option
9867 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
9868 key elements as negative integers.
9869 [Steve Henson]
9870
9871 *) Reorganize and speed up MD5.
9872 [Andy Polyakov <appro@fy.chalmers.se>]
9873
9874 *) VMS support.
9875 [Richard Levitte <richard@levitte.org>]
9876
9877 *) New option -out to asn1parse to allow the parsed structure to be
9878 output to a file. This is most useful when combined with the -strparse
9879 option to examine the output of things like OCTET STRINGS.
9880 [Steve Henson]
9881
9882 *) Make SSL library a little more fool-proof by not requiring any longer
9883 that SSL_set_{accept,connect}_state be called before
9884 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
9885 in many applications because usually everything *appeared* to work as
9886 intended anyway -- now it really works as intended).
9887 [Bodo Moeller]
9888
9889 *) Move openssl.cnf out of lib/.
9890 [Ulf Möller]
9891
9892 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
9893 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
9894 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
9895 [Ralf S. Engelschall]
9896
9897 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
9898 handle PKCS#7 enveloped data properly.
9899 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
9900
9901 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
9902 copying pointers. The cert_st handling is changed by this in
9903 various ways (and thus what used to be known as ctx->default_cert
9904 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
9905 any longer when s->cert does not give us what we need).
9906 ssl_cert_instantiate becomes obsolete by this change.
9907 As soon as we've got the new code right (possibly it already is?),
9908 we have solved a couple of bugs of the earlier code where s->cert
9909 was used as if it could not have been shared with other SSL structures.
9910
9911 Note that using the SSL API in certain dirty ways now will result
9912 in different behaviour than observed with earlier library versions:
9913 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
9914 does not influence s as it used to.
9915
9916 In order to clean up things more thoroughly, inside SSL_SESSION
9917 we don't use CERT any longer, but a new structure SESS_CERT
9918 that holds per-session data (if available); currently, this is
9919 the peer's certificate chain and, for clients, the server's certificate
9920 and temporary key. CERT holds only those values that can have
9921 meaningful defaults in an SSL_CTX.
9922 [Bodo Moeller]
9923
9924 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
9925 from the internal representation. Various PKCS#7 fixes: remove some
9926 evil casts and set the enc_dig_alg field properly based on the signing
9927 key type.
9928 [Steve Henson]
9929
9930 *) Allow PKCS#12 password to be set from the command line or the
9931 environment. Let 'ca' get its config file name from the environment
9932 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
9933 and 'x509').
9934 [Steve Henson]
9935
9936 *) Allow certificate policies extension to use an IA5STRING for the
9937 organization field. This is contrary to the PKIX definition but
9938 VeriSign uses it and IE5 only recognises this form. Document 'x509'
9939 extension option.
9940 [Steve Henson]
9941
9942 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
9943 without disallowing inline assembler and the like for non-pedantic builds.
9944 [Ben Laurie]
9945
9946 *) Support Borland C++ builder.
9947 [Janez Jere <jj@void.si>, modified by Ulf Möller]
9948
9949 *) Support Mingw32.
9950 [Ulf Möller]
9951
9952 *) SHA-1 cleanups and performance enhancements.
9953 [Andy Polyakov <appro@fy.chalmers.se>]
9954
9955 *) Sparc v8plus assembler for the bignum library.
9956 [Andy Polyakov <appro@fy.chalmers.se>]
9957
9958 *) Accept any -xxx and +xxx compiler options in Configure.
9959 [Ulf Möller]
9960
9961 *) Update HPUX configuration.
9962 [Anonymous]
9963
9964 *) Add missing sk_<type>_unshift() function to safestack.h
9965 [Ralf S. Engelschall]
9966
9967 *) New function SSL_CTX_use_certificate_chain_file that sets the
9968 "extra_cert"s in addition to the certificate. (This makes sense
9969 only for "PEM" format files, as chains as a whole are not
9970 DER-encoded.)
9971 [Bodo Moeller]
9972
9973 *) Support verify_depth from the SSL API.
9974 x509_vfy.c had what can be considered an off-by-one-error:
9975 Its depth (which was not part of the external interface)
9976 was actually counting the number of certificates in a chain;
9977 now it really counts the depth.
9978 [Bodo Moeller]
9979
9980 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
9981 instead of X509err, which often resulted in confusing error
9982 messages since the error codes are not globally unique
9983 (e.g. an alleged error in ssl3_accept when a certificate
9984 didn't match the private key).
9985
9986 *) New function SSL_CTX_set_session_id_context that allows to set a default
9987 value (so that you don't need SSL_set_session_id_context for each
9988 connection using the SSL_CTX).
9989 [Bodo Moeller]
9990
9991 *) OAEP decoding bug fix.
9992 [Ulf Möller]
9993
9994 *) Support INSTALL_PREFIX for package builders, as proposed by
9995 David Harris.
9996 [Bodo Moeller]
9997
9998 *) New Configure options "threads" and "no-threads". For systems
9999 where the proper compiler options are known (currently Solaris
10000 and Linux), "threads" is the default.
10001 [Bodo Moeller]
10002
10003 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
10004 [Bodo Moeller]
10005
10006 *) Install various scripts to $(OPENSSLDIR)/misc, not to
10007 $(INSTALLTOP)/bin -- they shouldn't clutter directories
10008 such as /usr/local/bin.
10009 [Bodo Moeller]
10010
10011 *) "make linux-shared" to build shared libraries.
10012 [Niels Poppe <niels@netbox.org>]
10013
10014 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
10015 [Ulf Möller]
10016
10017 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
10018 extension adding in x509 utility.
10019 [Steve Henson]
10020
10021 *) Remove NOPROTO sections and error code comments.
10022 [Ulf Möller]
10023
10024 *) Partial rewrite of the DEF file generator to now parse the ANSI
10025 prototypes.
10026 [Steve Henson]
10027
10028 *) New Configure options --prefix=DIR and --openssldir=DIR.
10029 [Ulf Möller]
10030
10031 *) Complete rewrite of the error code script(s). It is all now handled
10032 by one script at the top level which handles error code gathering,
10033 header rewriting and C source file generation. It should be much better
10034 than the old method: it now uses a modified version of Ulf's parser to
10035 read the ANSI prototypes in all header files (thus the old K&R definitions
10036 aren't needed for error creation any more) and do a better job of
10037 translating function codes into names. The old 'ASN1 error code imbedded
10038 in a comment' is no longer necessary and it doesn't use .err files which
10039 have now been deleted. Also the error code call doesn't have to appear all
10040 on one line (which resulted in some large lines...).
10041 [Steve Henson]
10042
10043 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
10044 [Bodo Moeller]
10045
10046 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
10047 0 (which usually indicates a closed connection), but continue reading.
10048 [Bodo Moeller]
10049
10050 *) Fix some race conditions.
10051 [Bodo Moeller]
10052
10053 *) Add support for CRL distribution points extension. Add Certificate
10054 Policies and CRL distribution points documentation.
10055 [Steve Henson]
10056
10057 *) Move the autogenerated header file parts to crypto/opensslconf.h.
10058 [Ulf Möller]
10059
10060 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
10061 8 of keying material. Merlin has also confirmed interop with this fix
10062 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
10063 [Merlin Hughes <merlin@baltimore.ie>]
10064
10065 *) Fix lots of warnings.
10066 [Richard Levitte <levitte@stacken.kth.se>]
10067
10068 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
10069 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
10070 [Richard Levitte <levitte@stacken.kth.se>]
10071
10072 *) Fix problems with sizeof(long) == 8.
10073 [Andy Polyakov <appro@fy.chalmers.se>]
10074
10075 *) Change functions to ANSI C.
10076 [Ulf Möller]
10077
10078 *) Fix typos in error codes.
10079 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
10080
10081 *) Remove defunct assembler files from Configure.
10082 [Ulf Möller]
10083
10084 *) SPARC v8 assembler BIGNUM implementation.
10085 [Andy Polyakov <appro@fy.chalmers.se>]
10086
10087 *) Support for Certificate Policies extension: both print and set.
10088 Various additions to support the r2i method this uses.
10089 [Steve Henson]
10090
10091 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
10092 return a const string when you are expecting an allocated buffer.
10093 [Ben Laurie]
10094
10095 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
10096 types DirectoryString and DisplayText.
10097 [Steve Henson]
10098
10099 *) Add code to allow r2i extensions to access the configuration database,
10100 add an LHASH database driver and add several ctx helper functions.
10101 [Steve Henson]
10102
10103 *) Fix an evil bug in bn_expand2() which caused various BN functions to
10104 fail when they extended the size of a BIGNUM.
10105 [Steve Henson]
10106
10107 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
10108 support typesafe stack.
10109 [Steve Henson]
10110
10111 *) Fix typo in SSL_[gs]et_options().
10112 [Nils Frostberg <nils@medcom.se>]
10113
10114 *) Delete various functions and files that belonged to the (now obsolete)
10115 old X509V3 handling code.
10116 [Steve Henson]
10117
10118 *) New Configure option "rsaref".
10119 [Ulf Möller]
10120
10121 *) Don't auto-generate pem.h.
10122 [Bodo Moeller]
10123
10124 *) Introduce type-safe ASN.1 SETs.
10125 [Ben Laurie]
10126
10127 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
10128 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
10129
10130 *) Introduce type-safe STACKs. This will almost certainly break lots of code
10131 that links with OpenSSL (well at least cause lots of warnings), but fear
10132 not: the conversion is trivial, and it eliminates loads of evil casts. A
10133 few STACKed things have been converted already. Feel free to convert more.
10134 In the fullness of time, I'll do away with the STACK type altogether.
10135 [Ben Laurie]
10136
10137 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
10138 specified in <certfile> by updating the entry in the index.txt file.
10139 This way one no longer has to edit the index.txt file manually for
10140 revoking a certificate. The -revoke option does the gory details now.
10141 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
10142
10143 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
10144 `-text' option at all and this way the `-noout -text' combination was
10145 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
10146 [Ralf S. Engelschall]
10147
10148 *) Make sure a corresponding plain text error message exists for the
10149 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
10150 verify callback function determined that a certificate was revoked.
10151 [Ralf S. Engelschall]
10152
10153 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
10154 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
10155 all available cipers including rc5, which was forgotten until now.
10156 In order to let the testing shell script know which algorithms
10157 are available, a new (up to now undocumented) command
10158 "openssl list-cipher-commands" is used.
10159 [Bodo Moeller]
10160
10161 *) Bugfix: s_client occasionally would sleep in select() when
10162 it should have checked SSL_pending() first.
10163 [Bodo Moeller]
10164
10165 *) New functions DSA_do_sign and DSA_do_verify to provide access to
10166 the raw DSA values prior to ASN.1 encoding.
10167 [Ulf Möller]
10168
10169 *) Tweaks to Configure
10170 [Niels Poppe <niels@netbox.org>]
10171
10172 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
10173 yet...
10174 [Steve Henson]
10175
10176 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
10177 [Ulf Möller]
10178
10179 *) New config option to avoid instructions that are illegal on the 80386.
10180 The default code is faster, but requires at least a 486.
10181 [Ulf Möller]
10182
10183 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
10184 SSL2_SERVER_VERSION (not used at all) macros, which are now the
10185 same as SSL2_VERSION anyway.
10186 [Bodo Moeller]
10187
10188 *) New "-showcerts" option for s_client.
10189 [Bodo Moeller]
10190
10191 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
10192 application. Various cleanups and fixes.
10193 [Steve Henson]
10194
10195 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
10196 modify error routines to work internally. Add error codes and PBE init
10197 to library startup routines.
10198 [Steve Henson]
10199
10200 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
10201 packing functions to asn1 and evp. Changed function names and error
10202 codes along the way.
10203 [Steve Henson]
10204
10205 *) PKCS12 integration: and so it begins... First of several patches to
10206 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
10207 objects to objects.h
10208 [Steve Henson]
10209
10210 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
10211 and display support for Thawte strong extranet extension.
10212 [Steve Henson]
10213
10214 *) Add LinuxPPC support.
10215 [Jeff Dubrule <igor@pobox.org>]
10216
10217 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
10218 bn_div_words in alpha.s.
10219 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
10220
10221 *) Make sure the RSA OAEP test is skipped under -DRSAref because
10222 OAEP isn't supported when OpenSSL is built with RSAref.
10223 [Ulf Moeller <ulf@fitug.de>]
10224
10225 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
10226 so they no longer are missing under -DNOPROTO.
10227 [Soren S. Jorvang <soren@t.dk>]
10228
10229
10230 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
10231
10232 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
10233 doesn't work when the session is reused. Coming soon!
10234 [Ben Laurie]
10235
10236 *) Fix a security hole, that allows sessions to be reused in the wrong
10237 context thus bypassing client cert protection! All software that uses
10238 client certs and session caches in multiple contexts NEEDS PATCHING to
10239 allow session reuse! A fuller solution is in the works.
10240 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
10241
10242 *) Some more source tree cleanups (removed obsolete files
10243 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
10244 permission on "config" script to be executable) and a fix for the INSTALL
10245 document.
10246 [Ulf Moeller <ulf@fitug.de>]
10247
10248 *) Remove some legacy and erroneous uses of malloc, free instead of
10249 Malloc, Free.
10250 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
10251
10252 *) Make rsa_oaep_test return non-zero on error.
10253 [Ulf Moeller <ulf@fitug.de>]
10254
10255 *) Add support for native Solaris shared libraries. Configure
10256 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
10257 if someone would make that last step automatic.
10258 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
10259
10260 *) ctx_size was not built with the right compiler during "make links". Fixed.
10261 [Ben Laurie]
10262
10263 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
10264 except NULL ciphers". This means the default cipher list will no longer
10265 enable NULL ciphers. They need to be specifically enabled e.g. with
10266 the string "DEFAULT:eNULL".
10267 [Steve Henson]
10268
10269 *) Fix to RSA private encryption routines: if p < q then it would
10270 occasionally produce an invalid result. This will only happen with
10271 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
10272 [Steve Henson]
10273
10274 *) Be less restrictive and allow also `perl util/perlpath.pl
10275 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
10276 because this way one can also use an interpreter named `perl5' (which is
10277 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
10278 installed as `perl').
10279 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10280
10281 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
10282 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10283
10284 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
10285 advapi32.lib to Win32 build and change the pem test comparison
10286 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
10287 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
10288 and crypto/des/ede_cbcm_enc.c.
10289 [Steve Henson]
10290
10291 *) DES quad checksum was broken on big-endian architectures. Fixed.
10292 [Ben Laurie]
10293
10294 *) Comment out two functions in bio.h that aren't implemented. Fix up the
10295 Win32 test batch file so it (might) work again. The Win32 test batch file
10296 is horrible: I feel ill....
10297 [Steve Henson]
10298
10299 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
10300 in e_os.h. Audit of header files to check ANSI and non ANSI
10301 sections: 10 functions were absent from non ANSI section and not exported
10302 from Windows DLLs. Fixed up libeay.num for new functions.
10303 [Steve Henson]
10304
10305 *) Make `openssl version' output lines consistent.
10306 [Ralf S. Engelschall]
10307
10308 *) Fix Win32 symbol export lists for BIO functions: Added
10309 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
10310 to ms/libeay{16,32}.def.
10311 [Ralf S. Engelschall]
10312
10313 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
10314 fine under Unix and passes some trivial tests I've now added. But the
10315 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
10316 added to make sure no one expects that this stuff really works in the
10317 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
10318 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
10319 openssl_bio.xs.
10320 [Ralf S. Engelschall]
10321
10322 *) Fix the generation of two part addresses in perl.
10323 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
10324
10325 *) Add config entry for Linux on MIPS.
10326 [John Tobey <jtobey@channel1.com>]
10327
10328 *) Make links whenever Configure is run, unless we are on Windoze.
10329 [Ben Laurie]
10330
10331 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
10332 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
10333 in CRLs.
10334 [Steve Henson]
10335
10336 *) Add a useful kludge to allow package maintainers to specify compiler and
10337 other platforms details on the command line without having to patch the
10338 Configure script everytime: One now can use ``perl Configure
10339 <id>:<details>'', i.e. platform ids are allowed to have details appended
10340 to them (separated by colons). This is treated as there would be a static
10341 pre-configured entry in Configure's %table under key <id> with value
10342 <details> and ``perl Configure <id>'' is called. So, when you want to
10343 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
10344 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
10345 now, which overrides the FreeBSD-elf entry on-the-fly.
10346 [Ralf S. Engelschall]
10347
10348 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
10349 [Ben Laurie]
10350
10351 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
10352 on the `perl Configure ...' command line. This way one can compile
10353 OpenSSL libraries with Position Independent Code (PIC) which is needed
10354 for linking it into DSOs.
10355 [Ralf S. Engelschall]
10356
10357 *) Remarkably, export ciphers were totally broken and no-one had noticed!
10358 Fixed.
10359 [Ben Laurie]
10360
10361 *) Cleaned up the LICENSE document: The official contact for any license
10362 questions now is the OpenSSL core team under openssl-core@openssl.org.
10363 And add a paragraph about the dual-license situation to make sure people
10364 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
10365 to the OpenSSL toolkit.
10366 [Ralf S. Engelschall]
10367
10368 *) General source tree makefile cleanups: Made `making xxx in yyy...'
10369 display consistent in the source tree and replaced `/bin/rm' by `rm'.
10370 Additionally cleaned up the `make links' target: Remove unnecessary
10371 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
10372 to speed processing and no longer clutter the display with confusing
10373 stuff. Instead only the actually done links are displayed.
10374 [Ralf S. Engelschall]
10375
10376 *) Permit null encryption ciphersuites, used for authentication only. It used
10377 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
10378 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
10379 encryption.
10380 [Ben Laurie]
10381
10382 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
10383 signed attributes when verifying signatures (this would break them),
10384 the detached data encoding was wrong and public keys obtained using
10385 X509_get_pubkey() weren't freed.
10386 [Steve Henson]
10387
10388 *) Add text documentation for the BUFFER functions. Also added a work around
10389 to a Win95 console bug. This was triggered by the password read stuff: the
10390 last character typed gets carried over to the next fread(). If you were
10391 generating a new cert request using 'req' for example then the last
10392 character of the passphrase would be CR which would then enter the first
10393 field as blank.
10394 [Steve Henson]
10395
10396 *) Added the new `Includes OpenSSL Cryptography Software' button as
10397 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
10398 button and can be used by applications based on OpenSSL to show the
10399 relationship to the OpenSSL project.
10400 [Ralf S. Engelschall]
10401
10402 *) Remove confusing variables in function signatures in files
10403 ssl/ssl_lib.c and ssl/ssl.h.
10404 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10405
10406 *) Don't install bss_file.c under PREFIX/include/
10407 [Lennart Bong <lob@kulthea.stacken.kth.se>]
10408
10409 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
10410 functions that return function pointers and has support for NT specific
10411 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
10412 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
10413 unsigned to signed types: this was killing the Win32 compile.
10414 [Steve Henson]
10415
10416 *) Add new certificate file to stack functions,
10417 SSL_add_dir_cert_subjects_to_stack() and
10418 SSL_add_file_cert_subjects_to_stack(). These largely supplant
10419 SSL_load_client_CA_file(), and can be used to add multiple certs easily
10420 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
10421 This means that Apache-SSL and similar packages don't have to mess around
10422 to add as many CAs as they want to the preferred list.
10423 [Ben Laurie]
10424
10425 *) Experiment with doxygen documentation. Currently only partially applied to
10426 ssl/ssl_lib.c.
10427 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
10428 openssl.doxy as the configuration file.
10429 [Ben Laurie]
10430
10431 *) Get rid of remaining C++-style comments which strict C compilers hate.
10432 [Ralf S. Engelschall, pointed out by Carlos Amengual]
10433
10434 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
10435 compiled in by default: it has problems with large keys.
10436 [Steve Henson]
10437
10438 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
10439 DH private keys and/or callback functions which directly correspond to
10440 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
10441 is needed for applications which have to configure certificates on a
10442 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
10443 (e.g. s_server).
10444 For the RSA certificate situation is makes no difference, but
10445 for the DSA certificate situation this fixes the "no shared cipher"
10446 problem where the OpenSSL cipher selection procedure failed because the
10447 temporary keys were not overtaken from the context and the API provided
10448 no way to reconfigure them.
10449 The new functions now let applications reconfigure the stuff and they
10450 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
10451 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
10452 non-public-API function ssl_cert_instantiate() is used as a helper
10453 function and also to reduce code redundancy inside ssl_rsa.c.
10454 [Ralf S. Engelschall]
10455
10456 *) Move s_server -dcert and -dkey options out of the undocumented feature
10457 area because they are useful for the DSA situation and should be
10458 recognized by the users.
10459 [Ralf S. Engelschall]
10460
10461 *) Fix the cipher decision scheme for export ciphers: the export bits are
10462 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
10463 SSL_EXP_MASK. So, the original variable has to be used instead of the
10464 already masked variable.
10465 [Richard Levitte <levitte@stacken.kth.se>]
10466
10467 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
10468 [Richard Levitte <levitte@stacken.kth.se>]
10469
10470 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
10471 from `int' to `unsigned int' because it's a length and initialized by
10472 EVP_DigestFinal() which expects an `unsigned int *'.
10473 [Richard Levitte <levitte@stacken.kth.se>]
10474
10475 *) Don't hard-code path to Perl interpreter on shebang line of Configure
10476 script. Instead use the usual Shell->Perl transition trick.
10477 [Ralf S. Engelschall]
10478
10479 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
10480 (in addition to RSA certificates) to match the behaviour of `openssl dsa
10481 -noout -modulus' as it's already the case for `openssl rsa -noout
10482 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
10483 currently the public key is printed (a decision which was already done by
10484 `openssl dsa -modulus' in the past) which serves a similar purpose.
10485 Additionally the NO_RSA no longer completely removes the whole -modulus
10486 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
10487 now, too.
10488 [Ralf S. Engelschall]
10489
10490 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
10491 BIO. See the source (crypto/evp/bio_ok.c) for more info.
10492 [Arne Ansper <arne@ats.cyber.ee>]
10493
10494 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
10495 to be added. Now both 'req' and 'ca' can use new objects defined in the
10496 config file.
10497 [Steve Henson]
10498
10499 *) Add cool BIO that does syslog (or event log on NT).
10500 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
10501
10502 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
10503 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
10504 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
10505 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
10506 [Ben Laurie]
10507
10508 *) Add preliminary config info for new extension code.
10509 [Steve Henson]
10510
10511 *) Make RSA_NO_PADDING really use no padding.
10512 [Ulf Moeller <ulf@fitug.de>]
10513
10514 *) Generate errors when private/public key check is done.
10515 [Ben Laurie]
10516
10517 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
10518 for some CRL extensions and new objects added.
10519 [Steve Henson]
10520
10521 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
10522 key usage extension and fuller support for authority key id.
10523 [Steve Henson]
10524
10525 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
10526 padding method for RSA, which is recommended for new applications in PKCS
10527 #1 v2.0 (RFC 2437, October 1998).
10528 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
10529 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
10530 against Bleichbacher's attack on RSA.
10531 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
10532 Ben Laurie]
10533
10534 *) Updates to the new SSL compression code
10535 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10536
10537 *) Fix so that the version number in the master secret, when passed
10538 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
10539 (because the server will not accept higher), that the version number
10540 is 0x03,0x01, not 0x03,0x00
10541 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10542
10543 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
10544 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
10545 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
10546 [Steve Henson]
10547
10548 *) Support for RAW extensions where an arbitrary extension can be
10549 created by including its DER encoding. See apps/openssl.cnf for
10550 an example.
10551 [Steve Henson]
10552
10553 *) Make sure latest Perl versions don't interpret some generated C array
10554 code as Perl array code in the crypto/err/err_genc.pl script.
10555 [Lars Weber <3weber@informatik.uni-hamburg.de>]
10556
10557 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
10558 not many people have the assembler. Various Win32 compilation fixes and
10559 update to the INSTALL.W32 file with (hopefully) more accurate Win32
10560 build instructions.
10561 [Steve Henson]
10562
10563 *) Modify configure script 'Configure' to automatically create crypto/date.h
10564 file under Win32 and also build pem.h from pem.org. New script
10565 util/mkfiles.pl to create the MINFO file on environments that can't do a
10566 'make files': perl util/mkfiles.pl >MINFO should work.
10567 [Steve Henson]
10568
10569 *) Major rework of DES function declarations, in the pursuit of correctness
10570 and purity. As a result, many evil casts evaporated, and some weirdness,
10571 too. You may find this causes warnings in your code. Zapping your evil
10572 casts will probably fix them. Mostly.
10573 [Ben Laurie]
10574
10575 *) Fix for a typo in asn1.h. Bug fix to object creation script
10576 obj_dat.pl. It considered a zero in an object definition to mean
10577 "end of object": none of the objects in objects.h have any zeros
10578 so it wasn't spotted.
10579 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
10580
10581 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
10582 Masking (CBCM). In the absence of test vectors, the best I have been able
10583 to do is check that the decrypt undoes the encrypt, so far. Send me test
10584 vectors if you have them.
10585 [Ben Laurie]
10586
10587 *) Correct calculation of key length for export ciphers (too much space was
10588 allocated for null ciphers). This has not been tested!
10589 [Ben Laurie]
10590
10591 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
10592 message is now correct (it understands "crypto" and "ssl" on its
10593 command line). There is also now an "update" option. This will update
10594 the util/ssleay.num and util/libeay.num files with any new functions.
10595 If you do a:
10596 perl util/mkdef.pl crypto ssl update
10597 it will update them.
10598 [Steve Henson]
10599
10600 *) Overhauled the Perl interface (perl/*):
10601 - ported BN stuff to OpenSSL's different BN library
10602 - made the perl/ source tree CVS-aware
10603 - renamed the package from SSLeay to OpenSSL (the files still contain
10604 their history because I've copied them in the repository)
10605 - removed obsolete files (the test scripts will be replaced
10606 by better Test::Harness variants in the future)
10607 [Ralf S. Engelschall]
10608
10609 *) First cut for a very conservative source tree cleanup:
10610 1. merge various obsolete readme texts into doc/ssleay.txt
10611 where we collect the old documents and readme texts.
10612 2. remove the first part of files where I'm already sure that we no
10613 longer need them because of three reasons: either they are just temporary
10614 files which were left by Eric or they are preserved original files where
10615 I've verified that the diff is also available in the CVS via "cvs diff
10616 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
10617 the crypto/md/ stuff).
10618 [Ralf S. Engelschall]
10619
10620 *) More extension code. Incomplete support for subject and issuer alt
10621 name, issuer and authority key id. Change the i2v function parameters
10622 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
10623 what that's for :-) Fix to ASN1 macro which messed up
10624 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
10625 [Steve Henson]
10626
10627 *) Preliminary support for ENUMERATED type. This is largely copied from the
10628 INTEGER code.
10629 [Steve Henson]
10630
10631 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
10632 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10633
10634 *) Make sure `make rehash' target really finds the `openssl' program.
10635 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
10636
10637 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
10638 like to hear about it if this slows down other processors.
10639 [Ben Laurie]
10640
10641 *) Add CygWin32 platform information to Configure script.
10642 [Alan Batie <batie@aahz.jf.intel.com>]
10643
10644 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
10645 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
10646
10647 *) New program nseq to manipulate netscape certificate sequences
10648 [Steve Henson]
10649
10650 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
10651 few typos.
10652 [Steve Henson]
10653
10654 *) Fixes to BN code. Previously the default was to define BN_RECURSION
10655 but the BN code had some problems that would cause failures when
10656 doing certificate verification and some other functions.
10657 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
10658
10659 *) Add ASN1 and PEM code to support netscape certificate sequences.
10660 [Steve Henson]
10661
10662 *) Add ASN1 and PEM code to support netscape certificate sequences.
10663 [Steve Henson]
10664
10665 *) Add several PKIX and private extended key usage OIDs.
10666 [Steve Henson]
10667
10668 *) Modify the 'ca' program to handle the new extension code. Modify
10669 openssl.cnf for new extension format, add comments.
10670 [Steve Henson]
10671
10672 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
10673 and add a sample to openssl.cnf so req -x509 now adds appropriate
10674 CA extensions.
10675 [Steve Henson]
10676
10677 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
10678 error code, add initial support to X509_print() and x509 application.
10679 [Steve Henson]
10680
10681 *) Takes a deep breath and start adding X509 V3 extension support code. Add
10682 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
10683 stuff is currently isolated and isn't even compiled yet.
10684 [Steve Henson]
10685
10686 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
10687 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
10688 Removed the versions check from X509 routines when loading extensions:
10689 this allows certain broken certificates that don't set the version
10690 properly to be processed.
10691 [Steve Henson]
10692
10693 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
10694 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
10695 can still be regenerated with "make depend".
10696 [Ben Laurie]
10697
10698 *) Spelling mistake in C version of CAST-128.
10699 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
10700
10701 *) Changes to the error generation code. The perl script err-code.pl
10702 now reads in the old error codes and retains the old numbers, only
10703 adding new ones if necessary. It also only changes the .err files if new
10704 codes are added. The makefiles have been modified to only insert errors
10705 when needed (to avoid needlessly modifying header files). This is done
10706 by only inserting errors if the .err file is newer than the auto generated
10707 C file. To rebuild all the error codes from scratch (the old behaviour)
10708 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
10709 or delete all the .err files.
10710 [Steve Henson]
10711
10712 *) CAST-128 was incorrectly implemented for short keys. The C version has
10713 been fixed, but is untested. The assembler versions are also fixed, but
10714 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
10715 to regenerate it if needed.
10716 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
10717 Hagino <itojun@kame.net>]
10718
10719 *) File was opened incorrectly in randfile.c.
10720 [Ulf Möller <ulf@fitug.de>]
10721
10722 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
10723 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
10724 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
10725 al: it's just almost always a UTCTime. Note this patch adds new error
10726 codes so do a "make errors" if there are problems.
10727 [Steve Henson]
10728
10729 *) Correct Linux 1 recognition in config.
10730 [Ulf Möller <ulf@fitug.de>]
10731
10732 *) Remove pointless MD5 hash when using DSA keys in ca.
10733 [Anonymous <nobody@replay.com>]
10734
10735 *) Generate an error if given an empty string as a cert directory. Also
10736 generate an error if handed NULL (previously returned 0 to indicate an
10737 error, but didn't set one).
10738 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
10739
10740 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
10741 [Ben Laurie]
10742
10743 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
10744 parameters. This was causing a warning which killed off the Win32 compile.
10745 [Steve Henson]
10746
10747 *) Remove C++ style comments from crypto/bn/bn_local.h.
10748 [Neil Costigan <neil.costigan@celocom.com>]
10749
10750 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
10751 based on a text string, looking up short and long names and finally
10752 "dot" format. The "dot" format stuff didn't work. Added new function
10753 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
10754 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
10755 OID is not part of the table.
10756 [Steve Henson]
10757
10758 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
10759 X509_LOOKUP_by_alias().
10760 [Ben Laurie]
10761
10762 *) Sort openssl functions by name.
10763 [Ben Laurie]
10764
10765 *) Get the gendsa program working (hopefully) and add it to app list. Remove
10766 encryption from sample DSA keys (in case anyone is interested the password
10767 was "1234").
10768 [Steve Henson]
10769
10770 *) Make _all_ *_free functions accept a NULL pointer.
10771 [Frans Heymans <fheymans@isaserver.be>]
10772
10773 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
10774 NULL pointers.
10775 [Anonymous <nobody@replay.com>]
10776
10777 *) s_server should send the CAfile as acceptable CAs, not its own cert.
10778 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10779
10780 *) Don't blow it for numeric -newkey arguments to apps/req.
10781 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
10782
10783 *) Temp key "for export" tests were wrong in s3_srvr.c.
10784 [Anonymous <nobody@replay.com>]
10785
10786 *) Add prototype for temp key callback functions
10787 SSL_CTX_set_tmp_{rsa,dh}_callback().
10788 [Ben Laurie]
10789
10790 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
10791 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
10792 [Steve Henson]
10793
10794 *) X509_name_add_entry() freed the wrong thing after an error.
10795 [Arne Ansper <arne@ats.cyber.ee>]
10796
10797 *) rsa_eay.c would attempt to free a NULL context.
10798 [Arne Ansper <arne@ats.cyber.ee>]
10799
10800 *) BIO_s_socket() had a broken should_retry() on Windoze.
10801 [Arne Ansper <arne@ats.cyber.ee>]
10802
10803 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
10804 [Arne Ansper <arne@ats.cyber.ee>]
10805
10806 *) Make sure the already existing X509_STORE->depth variable is initialized
10807 in X509_STORE_new(), but document the fact that this variable is still
10808 unused in the certificate verification process.
10809 [Ralf S. Engelschall]
10810
10811 *) Fix the various library and apps files to free up pkeys obtained from
10812 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
10813 [Steve Henson]
10814
10815 *) Fix reference counting in X509_PUBKEY_get(). This makes
10816 demos/maurice/example2.c work, amongst others, probably.
10817 [Steve Henson and Ben Laurie]
10818
10819 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
10820 `openssl' and second, the shortcut symlinks for the `openssl <command>'
10821 are no longer created. This way we have a single and consistent command
10822 line interface `openssl <command>', similar to `cvs <command>'.
10823 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
10824
10825 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
10826 BIT STRING wrapper always have zero unused bits.
10827 [Steve Henson]
10828
10829 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
10830 [Steve Henson]
10831
10832 *) Make the top-level INSTALL documentation easier to understand.
10833 [Paul Sutton]
10834
10835 *) Makefiles updated to exit if an error occurs in a sub-directory
10836 make (including if user presses ^C) [Paul Sutton]
10837
10838 *) Make Montgomery context stuff explicit in RSA data structure.
10839 [Ben Laurie]
10840
10841 *) Fix build order of pem and err to allow for generated pem.h.
10842 [Ben Laurie]
10843
10844 *) Fix renumbering bug in X509_NAME_delete_entry().
10845 [Ben Laurie]
10846
10847 *) Enhanced the err-ins.pl script so it makes the error library number
10848 global and can add a library name. This is needed for external ASN1 and
10849 other error libraries.
10850 [Steve Henson]
10851
10852 *) Fixed sk_insert which never worked properly.
10853 [Steve Henson]
10854
10855 *) Fix ASN1 macros so they can handle indefinite length constructed
10856 EXPLICIT tags. Some non standard certificates use these: they can now
10857 be read in.
10858 [Steve Henson]
10859
10860 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
10861 into a single doc/ssleay.txt bundle. This way the information is still
10862 preserved but no longer messes up this directory. Now it's new room for
10863 the new set of documentation files.
10864 [Ralf S. Engelschall]
10865
10866 *) SETs were incorrectly DER encoded. This was a major pain, because they
10867 shared code with SEQUENCEs, which aren't coded the same. This means that
10868 almost everything to do with SETs or SEQUENCEs has either changed name or
10869 number of arguments.
10870 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
10871
10872 *) Fix test data to work with the above.
10873 [Ben Laurie]
10874
10875 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
10876 was already fixed by Eric for 0.9.1 it seems.
10877 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
10878
10879 *) Autodetect FreeBSD3.
10880 [Ben Laurie]
10881
10882 *) Fix various bugs in Configure. This affects the following platforms:
10883 nextstep
10884 ncr-scde
10885 unixware-2.0
10886 unixware-2.0-pentium
10887 sco5-cc.
10888 [Ben Laurie]
10889
10890 *) Eliminate generated files from CVS. Reorder tests to regenerate files
10891 before they are needed.
10892 [Ben Laurie]
10893
10894 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
10895 [Ben Laurie]
10896
10897
10898 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
10899
10900 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
10901 changed SSLeay to OpenSSL in version strings.
10902 [Ralf S. Engelschall]
10903
10904 *) Some fixups to the top-level documents.
10905 [Paul Sutton]
10906
10907 *) Fixed the nasty bug where rsaref.h was not found under compile-time
10908 because the symlink to include/ was missing.
10909 [Ralf S. Engelschall]
10910
10911 *) Incorporated the popular no-RSA/DSA-only patches
10912 which allow to compile a RSA-free SSLeay.
10913 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
10914
10915 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
10916 when "ssleay" is still not found.
10917 [Ralf S. Engelschall]
10918
10919 *) Added more platforms to Configure: Cray T3E, HPUX 11,
10920 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
10921
10922 *) Updated the README file.
10923 [Ralf S. Engelschall]
10924
10925 *) Added various .cvsignore files in the CVS repository subdirs
10926 to make a "cvs update" really silent.
10927 [Ralf S. Engelschall]
10928
10929 *) Recompiled the error-definition header files and added
10930 missing symbols to the Win32 linker tables.
10931 [Ralf S. Engelschall]
10932
10933 *) Cleaned up the top-level documents;
10934 o new files: CHANGES and LICENSE
10935 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
10936 o merged COPYRIGHT into LICENSE
10937 o removed obsolete TODO file
10938 o renamed MICROSOFT to INSTALL.W32
10939 [Ralf S. Engelschall]
10940
10941 *) Removed dummy files from the 0.9.1b source tree:
10942 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
10943 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
10944 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
10945 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
10946 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
10947 [Ralf S. Engelschall]
10948
10949 *) Added various platform portability fixes.
10950 [Mark J. Cox]
10951
10952 *) The Genesis of the OpenSSL rpject:
10953 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
10954 Young and Tim J. Hudson created while they were working for C2Net until
10955 summer 1998.
10956 [The OpenSSL Project]
10957
10958
10959 Changes between 0.9.0b and 0.9.1b [not released]
10960
10961 *) Updated a few CA certificates under certs/
10962 [Eric A. Young]
10963
10964 *) Changed some BIGNUM api stuff.
10965 [Eric A. Young]
10966
10967 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
10968 DGUX x86, Linux Alpha, etc.
10969 [Eric A. Young]
10970
10971 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
10972 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
10973 available).
10974 [Eric A. Young]
10975
10976 *) Add -strparse option to asn1pars program which parses nested
10977 binary structures
10978 [Dr Stephen Henson <shenson@bigfoot.com>]
10979
10980 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
10981 [Eric A. Young]
10982
10983 *) DSA fix for "ca" program.
10984 [Eric A. Young]
10985
10986 *) Added "-genkey" option to "dsaparam" program.
10987 [Eric A. Young]
10988
10989 *) Added RIPE MD160 (rmd160) message digest.
10990 [Eric A. Young]
10991
10992 *) Added -a (all) option to "ssleay version" command.
10993 [Eric A. Young]
10994
10995 *) Added PLATFORM define which is the id given to Configure.
10996 [Eric A. Young]
10997
10998 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
10999 [Eric A. Young]
11000
11001 *) Extended the ASN.1 parser routines.
11002 [Eric A. Young]
11003
11004 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
11005 [Eric A. Young]
11006
11007 *) Added a BN_CTX to the BN library.
11008 [Eric A. Young]
11009
11010 *) Fixed the weak key values in DES library
11011 [Eric A. Young]
11012
11013 *) Changed API in EVP library for cipher aliases.
11014 [Eric A. Young]
11015
11016 *) Added support for RC2/64bit cipher.
11017 [Eric A. Young]
11018
11019 *) Converted the lhash library to the crypto/mem.c functions.
11020 [Eric A. Young]
11021
11022 *) Added more recognized ASN.1 object ids.
11023 [Eric A. Young]
11024
11025 *) Added more RSA padding checks for SSL/TLS.
11026 [Eric A. Young]
11027
11028 *) Added BIO proxy/filter functionality.
11029 [Eric A. Young]
11030
11031 *) Added extra_certs to SSL_CTX which can be used
11032 send extra CA certificates to the client in the CA cert chain sending
11033 process. It can be configured with SSL_CTX_add_extra_chain_cert().
11034 [Eric A. Young]
11035
11036 *) Now Fortezza is denied in the authentication phase because
11037 this is key exchange mechanism is not supported by SSLeay at all.
11038 [Eric A. Young]
11039
11040 *) Additional PKCS1 checks.
11041 [Eric A. Young]
11042
11043 *) Support the string "TLSv1" for all TLS v1 ciphers.
11044 [Eric A. Young]
11045
11046 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
11047 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
11048 [Eric A. Young]
11049
11050 *) Fixed a few memory leaks.
11051 [Eric A. Young]
11052
11053 *) Fixed various code and comment typos.
11054 [Eric A. Young]
11055
11056 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
11057 bytes sent in the client random.
11058 [Edward Bishop <ebishop@spyglass.com>]
11059