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