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