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