]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
disable '#ifdef DEBUG' sections
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
6
7 *) Add ECDSA in new directory crypto/ecdsa/.
8
9 Add applications 'openssl ecdsaparam' and 'openssl ecdsa'
10 (these are variants of 'openssl dsaparam' and 'openssl dsa').
11
12 ECDSA support is also included in various other files across the
13 library. Most notably,
14 - 'openssl req' now has a '-newkey ecdsa:file' option;
15 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
16 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
17 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
18 them suitable for ECDSA where domain parameters must be
19 extracted before the specific public key.
20 [Nils Larsch <nla@trustcenter.de>]
21
22 *) Add reference counting for EC_GROUP objects.
23 [Nils Larsch <nla@trustcenter.de>]
24
25 *) Include some named elliptic curves. These can be obtained from
26 the new functions
27 EC_GROUP_new_by_nid()
28 EC_GROUP_new_by_name()
29 Also add a 'nid' field to EC_GROUP objects, which can be accessed
30 via
31 EC_GROUP_set_nid()
32 EC_GROUP_get_nid()
33 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
34
35 Changes between 0.9.6 and 0.9.7 [xx XXX 2002]
36
37 OpenSSL 0.9.6a/0.9.6b/0.9.6c/0.9.6d (bugfix releases, 5 Apr 2001,
38 9 July 2001, 21 Dec 2001 and xx XXX 2002) and OpenSSL 0.9.7 were
39 developed in parallel, based on OpenSSL 0.9.6.
40
41 Change log entries are tagged as follows:
42 -) applies to 0.9.6a ... 0.9.6d only
43 *) applies to 0.9.6a ... 0.9.6d and 0.9.7
44 +) applies to 0.9.7 only
45
46 +) Added the '4758cca' ENGINE to support IBM 4758 cards.
47 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
48
49 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
50 Yoram Zahavi <YoramZ@gilian.com>).
51 [Lutz Jaenicke]
52
53 +) Add and OPENSSL_LOAD_CONF define which will cause
54 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
55 This allows older applications to transparently support certain
56 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
57 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
58 load the config file and OPENSSL_add_all_algorithms_conf() which will
59 always load it have also been added.
60 [Steve Henson]
61
62 +) Config modules support in openssl utility.
63
64 Most commands now load modules from the config file,
65 though in a few (such as version) this isn't done
66 because it couldn't be used for anything.
67
68 In the case of ca and req the config file used is
69 the same as the utility itself: that is the -config
70 command line option can be used to specify an
71 alternative file.
72 [Steve Henson]
73
74 +) Move default behaviour from OPENSSL_config(). If appname is NULL
75 use "openssl_conf" if filename is NULL use default openssl config file.
76 [Steve Henson]
77
78 +) Add an argument to OPENSSL_config() to allow the use of an alternative
79 config section name. Add a new flag to tolerate a missing config file
80 and move code to CONF_modules_load_file().
81 [Steve Henson]
82
83 *) Add information about CygWin 1.3 and on, and preserve proper
84 configuration for the versions before that.
85 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
86
87 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
88 check whether we deal with a copy of a session and do not delete from
89 the cache in this case. Problem reported by "Izhar Shoshani Levi"
90 <izhar@checkpoint.com>.
91 [Lutz Jaenicke]
92
93 *) Do not store session data into the internal session cache, if it
94 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
95 flag is set). Proposed by Aslam <aslam@funk.com>.
96 [Lutz Jaenicke]
97
98 +) Support for crypto accelerator cards from Accelerated Encryption
99 Processing, www.aep.ie. (Use engine 'aep')
100 The support was copied from 0.9.6c [engine] and adapted/corrected
101 to work with the new engine framework.
102 [AEP Inc. and Richard Levitte]
103
104 +) Support for SureWare crypto accelerator cards from Baltimore
105 Technologies. (Use engine 'sureware')
106 The support was copied from 0.9.6c [engine] and adapted
107 to work with the new engine framework.
108 [Richard Levitte]
109
110 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
111 value is 0.
112 [Richard Levitte]
113
114 +) Have the CHIL engine fork-safe (as defined by nCipher) and actually
115 make the newer ENGINE framework commands for the CHIL engine work.
116 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
117
118 +) Make it possible to produce shared libraries on ReliantUNIX.
119 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
120
121 *) Add the configuration target linux-s390x.
122 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
123
124 +) Add the configuration target debug-linux-ppro.
125 Make 'openssl rsa' use the general key loading routines
126 implemented in apps.c, and make those routines able to
127 handle the key format FORMAT_NETSCAPE and the variant
128 FORMAT_IISSGC.
129 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
130
131 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
132 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
133
134 +) Add -keyform to rsautl, and document -engine.
135 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
136
137 +) Change BIO_new_file (crypto/bio/bss_file.c) to use new
138 BIO_R_NO_SUCH_FILE error code rather than the generic
139 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
140 [Ben Laurie]
141
142 +) Add new functions
143 ERR_peek_last_error
144 ERR_peek_last_error_line
145 ERR_peek_last_error_line_data.
146 These are similar to
147 ERR_peek_error
148 ERR_peek_error_line
149 ERR_peek_error_line_data,
150 but report on the latest error recorded rather than the first one
151 still in the error queue.
152 [Ben Laurie, Bodo Moeller]
153
154 +) default_algorithms option in ENGINE config module. This allows things
155 like:
156 default_algorithms = ALL
157 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
158 [Steve Henson]
159
160 +) Prelminary ENGINE config module.
161 [Steve Henson]
162
163 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
164 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
165 variable as an indication that a ClientHello message has been
166 received. As the flag value will be lost between multiple
167 invocations of ssl3_accept when using non-blocking I/O, the
168 function may not be aware that a handshake has actually taken
169 place, thus preventing a new session from being added to the
170 session cache.
171
172 To avoid this problem, we now set s->new_session to 2 instead of
173 using a local variable.
174 [Lutz Jaenicke, Bodo Moeller]
175
176 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
177 if the SSL_R_LENGTH_MISMATCH error is detected.
178 [Geoff Thorpe, Bodo Moeller]
179
180 +) New experimental application configuration code.
181 [Steve Henson]
182
183 *) New 'shared_ldflag' column in Configure platform table.
184 [Richard Levitte]
185
186 *) Fix EVP_CIPHER_mode macro.
187 ["Dan S. Camper" <dan@bti.net>]
188
189 +) Change the AES code to follow the same name structure as all other
190 symmetric ciphers, and behave the same way. Move everything to
191 the directory crypto/aes, thereby obsoleting crypto/rijndael.
192 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
193
194 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
195 type, we must throw them away by setting rr->length to 0.
196 [D P Chang <dpc@qualys.com>]
197
198 -) OpenSSL 0.9.6c released [21 dec 2001]
199
200 +) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
201 [Ben Laurie and Theo de Raadt]
202
203 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
204 <Dominikus.Scherkl@biodata.com>. (The previous implementation
205 worked incorrectly for those cases where range = 10..._2 and
206 3*range is two bits longer than range.)
207 [Bodo Moeller]
208
209 *) Only add signing time to PKCS7 structures if it is not already
210 present.
211 [Steve Henson]
212
213 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
214 OBJ_ld_ce should be OBJ_id_ce.
215 Also some ip-pda OIDs in crypto/objects/objects.txt were
216 incorrect (cf. RFC 3039).
217 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
218
219 +) Add option to output public keys in req command.
220 [Massimiliano Pala madwolf@openca.org]
221
222 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
223 returns early because it has nothing to do.
224 [Andy Schneider <andy.schneider@bjss.co.uk>]
225
226 *) [In 0.9.6c-engine and 0.9.7 release:]
227 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
228 [Andy Schneider <andy.schneider@bjss.co.uk>]
229
230 -) [In 0.9.6c-engine release:]
231 Add support for Cryptographic Appliance's keyserver technology.
232 (Use engine 'keyclient')
233 [Cryptographic Appliances and Geoff Thorpe]
234
235 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
236 is called via tools/c89.sh because arguments have to be
237 rearranged (all '-L' options must appear before the first object
238 modules).
239 [Richard Shapiro <rshapiro@abinitio.com>]
240
241 +) Use wNAFs in EC_POINTs_mul() for improved efficiency
242 (up to about 10% better than before for P-192 and P-224).
243 [Bodo Moeller]
244
245 -) [In 0.9.6c-engine release:]
246 Add support for Broadcom crypto accelerator cards, backported
247 from 0.9.7.
248 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
249
250 -) [In 0.9.6c-engine release:]
251 Add support for SureWare crypto accelerator cards from
252 Baltimore Technologies. (Use engine 'sureware')
253 [Baltimore Technologies and Mark Cox]
254
255 -) [In 0.9.6c-engine release:]
256 Add support for crypto accelerator cards from Accelerated
257 Encryption Processing, www.aep.ie. (Use engine 'aep')
258 [AEP Inc. and Mark Cox]
259
260 *) Add a configuration entry for gcc on UnixWare.
261 [Gary Benson <gbenson@redhat.com>]
262
263 +) New functions/macros
264
265 SSL_CTX_set_msg_callback(ctx, cb)
266 SSL_CTX_set_msg_callback_arg(ctx, arg)
267 SSL_set_msg_callback(ssl, cb)
268 SSL_set_msg_callback_arg(ssl, arg)
269
270 to request calling a callback function
271
272 void cb(int write_p, int version, int content_type,
273 const void *buf, size_t len, SSL *ssl, void *arg)
274
275 whenever a protocol message has been completely received
276 (write_p == 0) or sent (write_p == 1). Here 'version' is the
277 protocol version according to which the SSL library interprets
278 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
279 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
280 the content type as defined in the SSL 3.0/TLS 1.0 protocol
281 specification (change_cipher_spec(20), alert(21), handshake(22)).
282 'buf' and 'len' point to the actual message, 'ssl' to the
283 SSL object, and 'arg' is the application-defined value set by
284 SSL[_CTX]_set_msg_callback_arg().
285
286 'openssl s_client' and 'openssl s_server' have new '-msg' options
287 to enable a callback that displays all protocol messages.
288 [Bodo Moeller]
289
290 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
291 messages are stored in a single piece (fixed-length part and
292 variable-length part combined) and fix various bugs found on the way.
293 [Bodo Moeller]
294
295 +) Change the shared library support so shared libraries are built as
296 soon as the corresponding static library is finished, and thereby get
297 openssl and the test programs linked against the shared library.
298 This still only happens when the keyword "shard" has been given to
299 the configuration scripts.
300
301 NOTE: shared library support is still an experimental thing, and
302 backward binary compatibility is still not guaranteed.
303 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
304
305 +) Add support for Subject Information Access extension.
306 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
307
308 +) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
309 additional bytes when new memory had to be allocated, not just
310 when reusing an existing buffer.
311 [Bodo Moeller]
312
313 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
314 instead. BIO_gethostbyname() does not know what timeouts are
315 appropriate, so entries would stay in cache even when they have
316 become invalid.
317 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
318
319 +) New command line and configuration option 'utf8' for the req command.
320 This allows field values to be specified as UTF8 strings.
321 [Steve Henson]
322
323 +) Add -multi and -mr options to "openssl speed" - giving multiple parallel
324 runs for the former and machine-readable output for the latter.
325 [Ben Laurie]
326
327 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
328 of the e-mail address in the DN (i.e., it will go into a certificate
329 extension only). The new configuration file option 'email_in_dn = no'
330 has the same effect.
331 [Massimiliano Pala madwolf@openca.org]
332
333 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
334 faced with a pathologically small ClientHello fragment that does
335 not contain client_version: Instead of aborting with an error,
336 simply choose the highest available protocol version (i.e.,
337 TLS 1.0 unless it is disabled). In practice, ClientHello
338 messages are never sent like this, but this change gives us
339 strictly correct behaviour at least for TLS.
340 [Bodo Moeller]
341
342 +) Change all functions with names starting with des_ to be starting
343 with DES_ instead. Add wrappers that are compatible with libdes,
344 but are named _ossl_old_des_*. Finally, add macros that map the
345 des_* symbols to the corresponding _ossl_old_des_*.
346
347 All this is done because there are increasing clashes with libdes
348 and other DES libraries that are currently used by other projects.
349 The old libdes interface (including crypt()) is provided if
350 <openssl/des_old.h> is included. For now, this automatically
351 happens in <openssl/des.h> unless OPENSSL_DISABLE_OLD_DES_SUPPORT is
352 defined. Note that crypt() is no longer declared in <openssl/des.h>.
353
354 NOTE: This is a major break of an old API into a new one. Software
355 authors are encouraged to switch to the DES_ style functions. Some
356 time in the future, des_old.h and the libdes compatibility functions
357 will be completely removed.
358 [Richard Levitte]
359
360 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
361 never resets s->method to s->ctx->method when called from within
362 one of the SSL handshake functions.
363 [Bodo Moeller; problem pointed out by Niko Baric]
364
365 +) Test for certificates which contain unsupported critical extensions.
366 If such a certificate is found during a verify operation it is
367 rejected by default: this behaviour can be overridden by either
368 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
369 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
370 X509_supported_extension() has also been added which returns 1 if a
371 particular extension is supported.
372 [Steve Henson]
373
374 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
375 (sent using the client's version number) if client_version is
376 smaller than the protocol version in use. Also change
377 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
378 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
379 the client will at least see that alert.
380 [Bodo Moeller]
381
382 +) Modify the behaviour of EVP cipher functions in similar way to digests
383 to retain compatibility with existing code.
384 [Steve Henson]
385
386 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
387 compatibility with existing code. In particular the 'ctx' parameter does
388 not have to be to be initialized before the call to EVP_DigestInit() and
389 it is tidied up after a call to EVP_DigestFinal(). New function
390 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
391 EVP_MD_CTX_copy() changed to not require the destination to be
392 initialized valid and new function EVP_MD_CTX_copy_ex() added which
393 requires the destination to be valid.
394
395 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
396 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
397 [Steve Henson]
398
399 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
400 so that complete 'Handshake' protocol structures are kept in memory
401 instead of overwriting 'msg_type' and 'length' with 'body' data.
402 [Bodo Moeller]
403
404 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
405 correctly.
406 [Bodo Moeller]
407
408 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
409 [Massimo Santin via Richard Levitte]
410
411 +) Major restructuring to the underlying ENGINE code. This includes
412 reduction of linker bloat, separation of pure "ENGINE" manipulation
413 (initialisation, etc) from functionality dealing with implementations
414 of specific crypto iterfaces. This change also introduces integrated
415 support for symmetric ciphers and digest implementations - so ENGINEs
416 can now accelerate these by providing EVP_CIPHER and EVP_MD
417 implementations of their own. This is detailed in crypto/engine/README
418 as it couldn't be adequately described here. However, there are a few
419 API changes worth noting - some RSA, DSA, DH, and RAND functions that
420 were changed in the original introduction of ENGINE code have now
421 reverted back - the hooking from this code to ENGINE is now a good
422 deal more passive and at run-time, operations deal directly with
423 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
424 dereferencing through an ENGINE pointer any more. Also, the ENGINE
425 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
426 they were not being used by the framework as there is no concept of a
427 BIGNUM_METHOD and they could not be generalised to the new
428 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
429 ENGINE_cpy() has been removed as it cannot be consistently defined in
430 the new code.
431 [Geoff Thorpe]
432
433 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
434 [Steve Henson]
435
436 +) Change mkdef.pl to sort symbols that get the same entry number,
437 and make sure the automatically generated functions ERR_load_*
438 become part of libeay.num as well.
439 [Richard Levitte]
440
441 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
442 client receives HelloRequest while in a handshake.
443 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
444
445 +) New function SSL_renegotiate_pending(). This returns true once
446 renegotiation has been requested (either SSL_renegotiate() call
447 or HelloRequest/ClientHello receveived from the peer) and becomes
448 false once a handshake has been completed.
449 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
450 sends a HelloRequest, but does not ensure that a handshake takes
451 place. SSL_renegotiate_pending() is useful for checking if the
452 client has followed the request.)
453 [Bodo Moeller]
454
455 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
456 By default, clients may request session resumption even during
457 renegotiation (if session ID contexts permit); with this option,
458 session resumption is possible only in the first handshake.
459 [Bodo Moeller]
460
461 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
462 should end in 'break', not 'goto end' which circuments various
463 cleanups done in state SSL_ST_OK. But session related stuff
464 must be disabled for SSL_ST_OK in the case that we just sent a
465 HelloRequest.
466
467 Also avoid some overhead by not calling ssl_init_wbio_buffer()
468 before just sending a HelloRequest.
469 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
470
471 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
472 reveal whether illegal block cipher padding was found or a MAC
473 verification error occured. (Neither SSLerr() codes nor alerts
474 are directly visible to potential attackers, but the information
475 may leak via logfiles.)
476
477 Similar changes are not required for the SSL 2.0 implementation
478 because the number of padding bytes is sent in clear for SSL 2.0,
479 and the extra bytes are just ignored. However ssl/s2_pkt.c
480 failed to verify that the purported number of padding bytes is in
481 the legal range.
482 [Bodo Moeller]
483
484 +) Add some demos for certificate and certificate request creation.
485 [Steve Henson]
486
487 +) Make maximum certificate chain size accepted from the peer application
488 settable (SSL*_get/set_max_cert_list()), as proposed by
489 "Douglas E. Engert" <deengert@anl.gov>.
490 [Lutz Jaenicke]
491
492 +) Add support for shared libraries for Unixware-7
493 (Boyd Lynn Gerber <gerberb@zenez.com>).
494 [Lutz Jaenicke]
495
496 *) Add OpenUNIX-8 support including shared libraries
497 (Boyd Lynn Gerber <gerberb@zenez.com>).
498 [Lutz Jaenicke]
499
500 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
501 'wristwatch attack' using huge encoding parameters (cf.
502 James H. Manger's CRYPTO 2001 paper). Note that the
503 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
504 encoding parameters and hence was not vulnerable.
505 [Bodo Moeller]
506
507 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
508 be done prior to destruction. Use this to unload error strings from
509 ENGINEs that load their own error strings. NB: This adds two new API
510 functions to "get" and "set" this destroy handler in an ENGINE.
511 [Geoff Thorpe]
512
513 +) Alter all existing ENGINE implementations (except "openssl" and
514 "openbsd") to dynamically instantiate their own error strings. This
515 makes them more flexible to be built both as statically-linked ENGINEs
516 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
517 Also, add stub code to each that makes building them as self-contained
518 shared-libraries easier (see README.ENGINE).
519 [Geoff Thorpe]
520
521 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
522 implementations into applications that are completely implemented in
523 self-contained shared-libraries. The "dynamic" ENGINE exposes control
524 commands that can be used to configure what shared-library to load and
525 to control aspects of the way it is handled. Also, made an update to
526 the README.ENGINE file that brings its information up-to-date and
527 provides some information and instructions on the "dynamic" ENGINE
528 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
529 [Geoff Thorpe]
530
531 *) BN_sqr() bug fix.
532 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
533
534 +) Make it possible to unload ranges of ERR strings with a new
535 "ERR_unload_strings" function.
536 [Geoff Thorpe]
537
538 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
539 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
540 followed by modular reduction.
541 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
542
543 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
544 equivalent based on BN_pseudo_rand() instead of BN_rand().
545 [Bodo Moeller]
546
547 +) Add a copy() function to EVP_MD.
548 [Ben Laurie]
549
550 +) Make EVP_MD routines take a context pointer instead of just the
551 md_data void pointer.
552 [Ben Laurie]
553
554 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
555 that the digest can only process a single chunk of data
556 (typically because it is provided by a piece of
557 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
558 is only going to provide a single chunk of data, and hence the
559 framework needn't accumulate the data for oneshot drivers.
560 [Ben Laurie]
561
562 +) As with "ERR", make it possible to replace the underlying "ex_data"
563 functions. This change also alters the storage and management of global
564 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
565 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
566 index counters. The API functions that use this state have been changed
567 to take a "class_index" rather than pointers to the class's local STACK
568 and counter, and there is now an API function to dynamically create new
569 classes. This centralisation allows us to (a) plug a lot of the
570 thread-safety problems that existed, and (b) makes it possible to clean
571 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
572 such data would previously have always leaked in application code and
573 workarounds were in place to make the memory debugging turn a blind eye
574 to it. Application code that doesn't use this new function will still
575 leak as before, but their memory debugging output will announce it now
576 rather than letting it slide.
577
578 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
579 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
580 has a return value to indicate success or failure.
581 [Geoff Thorpe]
582
583 +) Make it possible to replace the underlying "ERR" functions such that the
584 global state (2 LHASH tables and 2 locks) is only used by the "default"
585 implementation. This change also adds two functions to "get" and "set"
586 the implementation prior to it being automatically set the first time
587 any other ERR function takes place. Ie. an application can call "get",
588 pass the return value to a module it has just loaded, and that module
589 can call its own "set" function using that value. This means the
590 module's "ERR" operations will use (and modify) the error state in the
591 application and not in its own statically linked copy of OpenSSL code.
592 [Geoff Thorpe]
593
594 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
595 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
596 the operation, and provides a more encapsulated way for external code
597 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
598 to use these functions rather than manually incrementing the counts.
599
600 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
601 [Geoff Thorpe]
602
603 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
604 This function was broken, as the check for a new client hello message
605 to handle SGC did not allow these large messages.
606 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
607 [Lutz Jaenicke]
608
609 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
610 [Lutz Jaenicke]
611
612 +) Add EVP test program.
613 [Ben Laurie]
614
615 +) Add symmetric cipher support to ENGINE. Expect the API to change!
616 [Ben Laurie]
617
618 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
619 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
620 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
621 These allow a CRL to be built without having to access X509_CRL fields
622 directly. Modify 'ca' application to use new functions.
623 [Steve Henson]
624
625 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
626 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
627 [Lutz Jaenicke]
628
629 *) Rework the configuration and shared library support for Tru64 Unix.
630 The configuration part makes use of modern compiler features and
631 still retains old compiler behavior for those that run older versions
632 of the OS. The shared library support part includes a variant that
633 uses the RPATH feature, and is available through the special
634 configuration target "alpha-cc-rpath", which will never be selected
635 automatically.
636 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
637
638 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
639 with the same message size as in ssl3_get_certificate_request().
640 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
641 messages might inadvertently be reject as too long.
642 [Petr Lampa <lampa@fee.vutbr.cz>]
643
644 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
645 bug workarounds. Rollback attack detection is a security feature.
646 The problem will only arise on OpenSSL servers when TLSv1 is not
647 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
648 Software authors not wanting to support TLSv1 will have special reasons
649 for their choice and can explicitly enable this option.
650 [Bodo Moeller, Lutz Jaenicke]
651
652 +) Rationalise EVP so it can be extended: don't include a union of
653 cipher/digest structures, add init/cleanup functions. This also reduces
654 the number of header dependencies.
655 [Ben Laurie]
656
657 +) Make DES key schedule conform to the usual scheme, as well as
658 correcting its structure. This means that calls to DES functions
659 now have to pass a pointer to a des_key_schedule instead of a
660 plain des_key_schedule (which was actually always a pointer
661 anyway).
662 [Ben Laurie]
663
664 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
665 [Andy Polyakov]
666
667 *) Modified SSL library such that the verify_callback that has been set
668 specificly for an SSL object with SSL_set_verify() is actually being
669 used. Before the change, a verify_callback set with this function was
670 ignored and the verify_callback() set in the SSL_CTX at the time of
671 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
672 to allow the necessary settings.
673 [Lutz Jaenicke]
674
675 +) Initial reduction of linker bloat: the use of some functions, such as
676 PEM causes large amounts of unused functions to be linked in due to
677 poor organisation. For example pem_all.c contains every PEM function
678 which has a knock on effect of linking in large amounts of (unused)
679 ASN1 code. Grouping together similar functions and splitting unrelated
680 functions prevents this.
681 [Steve Henson]
682
683 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
684 explicitly to NULL, as at least on Solaris 8 this seems not always to be
685 done automatically (in contradiction to the requirements of the C
686 standard). This made problems when used from OpenSSH.
687 [Lutz Jaenicke]
688
689 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
690 dh->length and always used
691
692 BN_rand_range(priv_key, dh->p).
693
694 BN_rand_range() is not necessary for Diffie-Hellman, and this
695 specific range makes Diffie-Hellman unnecessarily inefficient if
696 dh->length (recommended exponent length) is much smaller than the
697 length of dh->p. We could use BN_rand_range() if the order of
698 the subgroup was stored in the DH structure, but we only have
699 dh->length.
700
701 So switch back to
702
703 BN_rand(priv_key, l, ...)
704
705 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
706 otherwise.
707 [Bodo Moeller]
708
709 *) In
710
711 RSA_eay_public_encrypt
712 RSA_eay_private_decrypt
713 RSA_eay_private_encrypt (signing)
714 RSA_eay_public_decrypt (signature verification)
715
716 (default implementations for RSA_public_encrypt,
717 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
718 always reject numbers >= n.
719 [Bodo Moeller]
720
721 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
722 to synchronize access to 'locking_thread'. This is necessary on
723 systems where access to 'locking_thread' (an 'unsigned long'
724 variable) is not atomic.
725 [Bodo Moeller]
726
727 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
728 *before* setting the 'crypto_lock_rand' flag. The previous code had
729 a race condition if 0 is a valid thread ID.
730 [Travis Vitek <vitek@roguewave.com>]
731
732 +) Cleanup of EVP macros.
733 [Ben Laurie]
734
735 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
736 correct _ecb suffix.
737 [Ben Laurie]
738
739 +) Add initial OCSP responder support to ocsp application. The
740 revocation information is handled using the text based index
741 use by the ca application. The responder can either handle
742 requests generated internally, supplied in files (for example
743 via a CGI script) or using an internal minimal server.
744 [Steve Henson]
745
746 +) Add configuration choices to get zlib compression for TLS.
747 [Richard Levitte]
748
749 +) Changes to Kerberos SSL for RFC 2712 compliance:
750 1. Implemented real KerberosWrapper, instead of just using
751 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
752 2. Implemented optional authenticator field of KerberosWrapper.
753
754 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
755 and authenticator structs; see crypto/krb5/.
756
757 Generalized Kerberos calls to support multiple Kerberos libraries.
758 [Vern Staats <staatsvr@asc.hpc.mil>,
759 Jeffrey Altman <jaltman@columbia.edu>
760 via Richard Levitte]
761
762 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
763 already does with RSA. testdsa.h now has 'priv_key/pub_key'
764 values for each of the key sizes rather than having just
765 parameters (and 'speed' generating keys each time).
766 [Geoff Thorpe]
767
768 *) Add support for shared libraries under Irix.
769 [Albert Chin-A-Young <china@thewrittenword.com>]
770
771 *) Add configuration option to build on Linux on both big-endian and
772 little-endian MIPS.
773 [Ralf Baechle <ralf@uni-koblenz.de>]
774
775 *) Add the possibility to create shared libraries on HP-UX.
776 [Richard Levitte]
777
778 -) OpenSSL 0.9.6b released [9 July 2001]
779
780 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
781 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
782 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
783 PRNG state recovery was possible based on the output of
784 one PRNG request appropriately sized to gain knowledge on
785 'md' followed by enough consecutive 1-byte PRNG requests
786 to traverse all of 'state'.
787
788 1. When updating 'md_local' (the current thread's copy of 'md')
789 during PRNG output generation, hash all of the previous
790 'md_local' value, not just the half used for PRNG output.
791
792 2. Make the number of bytes from 'state' included into the hash
793 independent from the number of PRNG bytes requested.
794
795 The first measure alone would be sufficient to avoid
796 Markku-Juhani's attack. (Actually it had never occurred
797 to me that the half of 'md_local' used for chaining was the
798 half from which PRNG output bytes were taken -- I had always
799 assumed that the secret half would be used.) The second
800 measure makes sure that additional data from 'state' is never
801 mixed into 'md_local' in small portions; this heuristically
802 further strengthens the PRNG.
803 [Bodo Moeller]
804
805 +) Speed up EVP routines.
806 Before:
807 encrypt
808 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
809 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
810 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
811 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
812 decrypt
813 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
814 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
815 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
816 After:
817 encrypt
818 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
819 decrypt
820 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
821 [Ben Laurie]
822
823 *) Fix crypto/bn/asm/mips3.s.
824 [Andy Polyakov]
825
826 *) When only the key is given to "enc", the IV is undefined. Print out
827 an error message in this case.
828 [Lutz Jaenicke]
829
830 +) Added the OS2-EMX target.
831 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
832
833 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
834 to support NCONF routines in extension code. New function CONF_set_nconf()
835 to allow functions which take an NCONF to also handle the old LHASH
836 structure: this means that the old CONF compatible routines can be
837 retained (in particular wrt extensions) without having to duplicate the
838 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
839 [Steve Henson]
840
841 *) Handle special case when X509_NAME is empty in X509 printing routines.
842 [Steve Henson]
843
844 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
845 positive and less than q.
846 [Bodo Moeller]
847
848 +) Enhance the general user interface with mechanisms for inner control
849 and with possibilities to have yes/no kind of prompts.
850 [Richard Levitte]
851
852 +) Change all calls to low level digest routines in the library and
853 applications to use EVP. Add missing calls to HMAC_cleanup() and
854 don't assume HMAC_CTX can be copied using memcpy().
855 [Verdon Walker <VWalker@novell.com>, Steve Henson]
856
857 +) Add the possibility to control engines through control names but with
858 arbitrary arguments instead of just a string.
859 Change the key loaders to take a UI_METHOD instead of a callback
860 function pointer. NOTE: this breaks binary compatibility with earlier
861 versions of OpenSSL [engine].
862 Adapt the nCipher code for these new conditions and add a card insertion
863 callback.
864 [Richard Levitte]
865
866 +) Enhance the general user interface with mechanisms to better support
867 dialog box interfaces, application-defined prompts, the possibility
868 to use defaults (for example default passwords from somewhere else)
869 and interrupts/cancellations.
870 [Richard Levitte]
871
872 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
873 used: it isn't thread safe and the add_lock_callback should handle
874 that itself.
875 [Paul Rose <Paul.Rose@bridge.com>]
876
877 *) Verify that incoming data obeys the block size in
878 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
879 [Bodo Moeller]
880
881 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
882 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
883 [Steve Henson]
884
885 *) Fix OAEP check.
886 [Ulf Möller, Bodo Möller]
887
888 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
889 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
890 when fixing the server behaviour for backwards-compatible 'client
891 hello' messages. (Note that the attack is impractical against
892 SSL 3.0 and TLS 1.0 anyway because length and version checking
893 means that the probability of guessing a valid ciphertext is
894 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
895 paper.)
896
897 Before 0.9.5, the countermeasure (hide the error by generating a
898 random 'decryption result') did not work properly because
899 ERR_clear_error() was missing, meaning that SSL_get_error() would
900 detect the supposedly ignored error.
901
902 Both problems are now fixed.
903 [Bodo Moeller]
904
905 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
906 (previously it was 1024).
907 [Bodo Moeller]
908
909 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
910 tidy up some unnecessarily weird code in 'sk_new()').
911 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
912
913 +) Change the key loading routines for ENGINEs to use the same kind
914 callback (pem_password_cb) as all other routines that need this
915 kind of callback.
916 [Richard Levitte]
917
918 *) Fix for compatibility mode trust settings: ignore trust settings
919 unless some valid trust or reject settings are present.
920 [Steve Henson]
921
922 *) Fix for blowfish EVP: its a variable length cipher.
923 [Steve Henson]
924
925 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
926 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
927 than this minimum value is recommended.
928 [Lutz Jaenicke]
929
930 +) New random seeder for OpenVMS, using the system process statistics
931 that are easily reachable.
932 [Richard Levitte]
933
934 +) Windows apparently can't transparently handle global
935 variables defined in DLLs. Initialisations such as:
936
937 const ASN1_ITEM *it = &ASN1_INTEGER_it;
938
939 wont compile. This is used by the any applications that need to
940 declare their own ASN1 modules. This was fixed by adding the option
941 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
942 needed for static libraries under Win32.
943 [Steve Henson]
944
945 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
946 setting of purpose and trust fields. New X509_STORE trust and
947 purpose functions and tidy up setting in other SSL functions.
948 [Steve Henson]
949
950 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
951 structure. These are inherited by X509_STORE_CTX when it is
952 initialised. This allows various defaults to be set in the
953 X509_STORE structure (such as flags for CRL checking and custom
954 purpose or trust settings) for functions which only use X509_STORE_CTX
955 internally such as S/MIME.
956
957 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
958 trust settings if they are not set in X509_STORE. This allows X509_STORE
959 purposes and trust (in S/MIME for example) to override any set by default.
960
961 Add command line options for CRL checking to smime, s_client and s_server
962 applications.
963 [Steve Henson]
964
965 +) Initial CRL based revocation checking. If the CRL checking flag(s)
966 are set then the CRL is looked up in the X509_STORE structure and
967 its validity and signature checked, then if the certificate is found
968 in the CRL the verify fails with a revoked error.
969
970 Various new CRL related callbacks added to X509_STORE_CTX structure.
971
972 Command line options added to 'verify' application to support this.
973
974 This needs some additional work, such as being able to handle multiple
975 CRLs with different times, extension based lookup (rather than just
976 by subject name) and ultimately more complete V2 CRL extension
977 handling.
978 [Steve Henson]
979
980 +) Add a general user interface API (crypto/ui/). This is designed
981 to replace things like des_read_password and friends (backward
982 compatibility functions using this new API are provided).
983 The purpose is to remove prompting functions from the DES code
984 section as well as provide for prompting through dialog boxes in
985 a window system and the like.
986 [Richard Levitte]
987
988 *) Fix various bugs related to DSA S/MIME verification. Handle missing
989 parameters in DSA public key structures and return an error in the
990 DSA routines if parameters are absent.
991 [Steve Henson]
992
993 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
994 in the current directory if neither $RANDFILE nor $HOME was set.
995 RAND_file_name() in 0.9.6a returned NULL in this case. This has
996 caused some confusion to Windows users who haven't defined $HOME.
997 Thus RAND_file_name() is changed again: e_os.h can define a
998 DEFAULT_HOME, which will be used if $HOME is not set.
999 For Windows, we use "C:"; on other platforms, we still require
1000 environment variables.
1001
1002 +) Add "ex_data" support to ENGINE so implementations can add state at a
1003 per-structure level rather than having to store it globally.
1004 [Geoff]
1005
1006 +) Make it possible for ENGINE structures to be copied when retrieved by
1007 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1008 This causes the "original" ENGINE structure to act like a template,
1009 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1010 operational state can be localised to each ENGINE structure, despite the
1011 fact they all share the same "methods". New ENGINE structures returned in
1012 this case have no functional references and the return value is the single
1013 structural reference. This matches the single structural reference returned
1014 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1015 ENGINE structure.
1016 [Geoff]
1017
1018 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1019 needs to match any other type at all we need to manually clear the
1020 tag cache.
1021 [Steve Henson]
1022
1023 +) Changes to the "openssl engine" utility to include;
1024 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1025 about an ENGINE's available control commands.
1026 - executing control commands from command line arguments using the
1027 '-pre' and '-post' switches. '-post' is only used if '-t' is
1028 specified and the ENGINE is successfully initialised. The syntax for
1029 the individual commands are colon-separated, for example;
1030 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1031 [Geoff]
1032
1033 +) New dynamic control command support for ENGINEs. ENGINEs can now
1034 declare their own commands (numbers), names (strings), descriptions,
1035 and input types for run-time discovery by calling applications. A
1036 subset of these commands are implicitly classed as "executable"
1037 depending on their input type, and only these can be invoked through
1038 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1039 can be based on user input, config files, etc). The distinction is
1040 that "executable" commands cannot return anything other than a boolean
1041 result and can only support numeric or string input, whereas some
1042 discoverable commands may only be for direct use through
1043 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1044 pointers, or other custom uses. The "executable" commands are to
1045 support parameterisations of ENGINE behaviour that can be
1046 unambiguously defined by ENGINEs and used consistently across any
1047 OpenSSL-based application. Commands have been added to all the
1048 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1049 control over shared-library paths without source code alterations.
1050 [Geoff]
1051
1052 +) Changed all ENGINE implementations to dynamically allocate their
1053 ENGINEs rather than declaring them statically. Apart from this being
1054 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1055 this also allows the implementations to compile without using the
1056 internal engine_int.h header.
1057 [Geoff]
1058
1059 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1060 'const' value. Any code that should be able to modify a RAND_METHOD
1061 should already have non-const pointers to it (ie. they should only
1062 modify their own ones).
1063 [Geoff]
1064
1065 +) Made a variety of little tweaks to the ENGINE code.
1066 - "atalla" and "ubsec" string definitions were moved from header files
1067 to C code. "nuron" string definitions were placed in variables
1068 rather than hard-coded - allowing parameterisation of these values
1069 later on via ctrl() commands.
1070 - Removed unused "#if 0"'d code.
1071 - Fixed engine list iteration code so it uses ENGINE_free() to release
1072 structural references.
1073 - Constified the RAND_METHOD element of ENGINE structures.
1074 - Constified various get/set functions as appropriate and added
1075 missing functions (including a catch-all ENGINE_cpy that duplicates
1076 all ENGINE values onto a new ENGINE except reference counts/state).
1077 - Removed NULL parameter checks in get/set functions. Setting a method
1078 or function to NULL is a way of cancelling out a previously set
1079 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1080 and doesn't justify the extra error symbols and code.
1081 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1082 flags from engine_int.h to engine.h.
1083 - Changed prototypes for ENGINE handler functions (init(), finish(),
1084 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1085 [Geoff]
1086
1087 *) Move 'if (!initialized) RAND_poll()' into regions protected by
1088 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
1089 having multiple threads call RAND_poll() concurrently.
1090 [Bodo Moeller]
1091
1092 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
1093 combination of a flag and a thread ID variable.
1094 Otherwise while one thread is in ssleay_rand_bytes (which sets the
1095 flag), *other* threads can enter ssleay_add_bytes without obeying
1096 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
1097 that they do not hold after the first thread unsets add_do_not_lock).
1098 [Bodo Moeller]
1099
1100 +) Implement binary inversion algorithm for BN_mod_inverse in addition
1101 to the algorithm using long division. The binary algorithm can be
1102 used only if the modulus is odd. On 32-bit systems, it is faster
1103 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1104 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1105 up to 450 bits. In 64-bit environments, the binary algorithm
1106 appears to be advantageous for much longer moduli; here we use it
1107 for moduli up to 2048 bits.
1108 [Bodo Moeller]
1109
1110 *) Change bctest again: '-x' expressions are not available in all
1111 versions of 'test'.
1112 [Bodo Moeller]
1113
1114 -) OpenSSL 0.9.6a released [5 Apr 2001]
1115
1116 *) Fix a couple of memory leaks in PKCS7_dataDecode()
1117 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
1118
1119 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
1120 the default extension for executables, if any. Also, make the perl
1121 scripts that use symlink() to test if it really exists and use "cp"
1122 if it doesn't. All this made OpenSSL compilable and installable in
1123 CygWin.
1124 [Richard Levitte]
1125
1126 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1127 could not support the combine flag in choice fields.
1128 [Steve Henson]
1129
1130 -) Fix for asn1_GetSequence() for indefinite length constructed data.
1131 If SEQUENCE is length is indefinite just set c->slen to the total
1132 amount of data available.
1133 [Steve Henson, reported by shige@FreeBSD.org]
1134 [This change does not apply to 0.9.7.]
1135
1136 *) Change bctest to avoid here-documents inside command substitution
1137 (workaround for FreeBSD /bin/sh bug).
1138 For compatibility with Ultrix, avoid shell functions (introduced
1139 in the bctest version that searches along $PATH).
1140 [Bodo Moeller]
1141
1142 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
1143 with des_encrypt() defined on some operating systems, like Solaris
1144 and UnixWare.
1145 [Richard Levitte]
1146
1147 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
1148 On the Importance of Eliminating Errors in Cryptographic
1149 Computations, J. Cryptology 14 (2001) 2, 101-119,
1150 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
1151 [Ulf Moeller]
1152
1153 *) MIPS assembler BIGNUM division bug fix.
1154 [Andy Polyakov]
1155
1156 *) Disabled incorrect Alpha assembler code.
1157 [Richard Levitte]
1158
1159 -) Fix PKCS#7 decode routines so they correctly update the length
1160 after reading an EOC for the EXPLICIT tag.
1161 [Steve Henson]
1162 [This change does not apply to 0.9.7.]
1163
1164 *) Fix bug in PKCS#12 key generation routines. This was triggered
1165 if a 3DES key was generated with a 0 initial byte. Include
1166 PKCS12_BROKEN_KEYGEN compilation option to retain the old
1167 (but broken) behaviour.
1168 [Steve Henson]
1169
1170 *) Enhance bctest to search for a working bc along $PATH and print
1171 it when found.
1172 [Tim Rice <tim@multitalents.net> via Richard Levitte]
1173
1174 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
1175 extensions from a certificate request to the certificate.
1176 [Steve Henson]
1177
1178 +) Allow multiple 'certopt' and 'nameopt' options to be separated
1179 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1180 file: this allows the display of the certificate about to be
1181 signed to be customised, to allow certain fields to be included
1182 or excluded and extension details. The old system didn't display
1183 multicharacter strings properly, omitted fields not in the policy
1184 and couldn't display additional details such as extensions.
1185 [Steve Henson]
1186
1187 *) Fix memory leaks in err.c: free err_data string if necessary;
1188 don't write to the wrong index in ERR_set_error_data.
1189 [Bodo Moeller]
1190
1191 +) Function EC_POINTs_mul for multiple scalar multiplication
1192 of an arbitrary number of elliptic curve points
1193 \sum scalars[i]*points[i],
1194 optionally including the generator defined for the EC_GROUP:
1195 scalar*generator + \sum scalars[i]*points[i].
1196
1197 EC_POINT_mul is a simple wrapper function for the typical case
1198 that the point list has just one item (besides the optional
1199 generator).
1200 [Bodo Moeller]
1201
1202 +) First EC_METHODs for curves over GF(p):
1203
1204 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1205 operations and provides various method functions that can also
1206 operate with faster implementations of modular arithmetic.
1207
1208 EC_GFp_mont_method() reuses most functions that are part of
1209 EC_GFp_simple_method, but uses Montgomery arithmetic.
1210
1211 [Bodo Moeller; point addition and point doubling
1212 implementation directly derived from source code provided by
1213 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1214
1215 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1216 crypto/ec/ec_lib.c):
1217
1218 Curves are EC_GROUP objects (with an optional group generator)
1219 based on EC_METHODs that are built into the library.
1220
1221 Points are EC_POINT objects based on EC_GROUP objects.
1222
1223 Most of the framework would be able to handle curves over arbitrary
1224 finite fields, but as there are no obvious types for fields other
1225 than GF(p), some functions are limited to that for now.
1226 [Bodo Moeller]
1227
1228 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1229 that the file contains a complete HTTP response.
1230 [Richard Levitte]
1231
1232 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1233 change the def and num file printf format specifier from "%-40sXXX"
1234 to "%-39s XXX". The latter will always guarantee a space after the
1235 field while the former will cause them to run together if the field
1236 is 40 of more characters long.
1237 [Steve Henson]
1238
1239 +) Constify the cipher and digest 'method' functions and structures
1240 and modify related functions to take constant EVP_MD and EVP_CIPHER
1241 pointers.
1242 [Steve Henson]
1243
1244 *) Implement ssl23_peek (analogous to ssl23_read), which previously
1245 did not exist.
1246 [Bodo Moeller]
1247
1248 *) Replace rdtsc with _emit statements for VC++ version 5.
1249 [Jeremy Cooper <jeremy@baymoo.org>]
1250
1251 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1252 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1253 [Bodo Moeller]
1254
1255 +) Modify EVP_Digest*() routines so they now return values. Although the
1256 internal software routines can never fail additional hardware versions
1257 might.
1258 [Steve Henson]
1259
1260 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1261
1262 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1263 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1264
1265 ASN1 error codes
1266 ERR_R_NESTED_ASN1_ERROR
1267 ...
1268 ERR_R_MISSING_ASN1_EOS
1269 were 4 .. 9, conflicting with
1270 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1271 ...
1272 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1273 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1274
1275 Add new error code 'ERR_R_INTERNAL_ERROR'.
1276 [Bodo Moeller]
1277
1278 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1279 suffices.
1280 [Bodo Moeller]
1281
1282 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1283 sets the subject name for a new request or supersedes the
1284 subject name in a given request. Formats that can be parsed are
1285 'CN=Some Name, OU=myOU, C=IT'
1286 and
1287 'CN=Some Name/OU=myOU/C=IT'.
1288
1289 Add options '-batch' and '-verbose' to 'openssl req'.
1290 [Massimiliano Pala <madwolf@hackmasters.net>]
1291
1292 *) Make it possible to reuse SSLv2 sessions.
1293 [Richard Levitte]
1294
1295 +) Introduce the possibility to access global variables through
1296 functions on platform were that's the best way to handle exporting
1297 global variables in shared libraries. To enable this functionality,
1298 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1299 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1300 is normally done by Configure or something similar).
1301
1302 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1303 in the source file (foo.c) like this:
1304
1305 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1306 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1307
1308 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1309 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1310
1311 OPENSSL_DECLARE_GLOBAL(int,foo);
1312 #define foo OPENSSL_GLOBAL_REF(foo)
1313 OPENSSL_DECLARE_GLOBAL(double,bar);
1314 #define bar OPENSSL_GLOBAL_REF(bar)
1315
1316 The #defines are very important, and therefore so is including the
1317 header file everywhere where the defined globals are used.
1318
1319 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1320 of ASN.1 items, but that structure is a bit different.
1321
1322 The largest change is in util/mkdef.pl which has been enhanced with
1323 better and easier to understand logic to choose which symbols should
1324 go into the Windows .def files as well as a number of fixes and code
1325 cleanup (among others, algorithm keywords are now sorted
1326 lexicographically to avoid constant rewrites).
1327 [Richard Levitte]
1328
1329 *) In copy_email() check for >= 0 as a return value for
1330 X509_NAME_get_index_by_NID() since 0 is a valid index.
1331 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1332
1333 +) In BN_div() keep a copy of the sign of 'num' before writing the
1334 result to 'rm' because if rm==num the value will be overwritten
1335 and produce the wrong result if 'num' is negative: this caused
1336 problems with BN_mod() and BN_nnmod().
1337 [Steve Henson]
1338
1339 +) Function OCSP_request_verify(). This checks the signature on an
1340 OCSP request and verifies the signer certificate. The signer
1341 certificate is just checked for a generic purpose and OCSP request
1342 trust settings.
1343 [Steve Henson]
1344
1345 +) Add OCSP_check_validity() function to check the validity of OCSP
1346 responses. OCSP responses are prepared in real time and may only
1347 be a few seconds old. Simply checking that the current time lies
1348 between thisUpdate and nextUpdate max reject otherwise valid responses
1349 caused by either OCSP responder or client clock inaccuracy. Instead
1350 we allow thisUpdate and nextUpdate to fall within a certain period of
1351 the current time. The age of the response can also optionally be
1352 checked. Two new options -validity_period and -status_age added to
1353 ocsp utility.
1354 [Steve Henson]
1355
1356 +) If signature or public key algorithm is unrecognized print out its
1357 OID rather that just UNKNOWN.
1358 [Steve Henson]
1359
1360 *) Avoid coredump with unsupported or invalid public keys by checking if
1361 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1362 PKCS7_verify() fails with non detached data.
1363 [Steve Henson]
1364
1365 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1366 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1367 ID to be generated from the issuer certificate alone which can then be
1368 passed to OCSP_id_issuer_cmp().
1369 [Steve Henson]
1370
1371 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1372 ASN1 modules to export functions returning ASN1_ITEM pointers
1373 instead of the ASN1_ITEM structures themselves. This adds several
1374 new macros which allow the underlying ASN1 function/structure to
1375 be accessed transparently. As a result code should not use ASN1_ITEM
1376 references directly (such as &X509_it) but instead use the relevant
1377 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1378 use of the new ASN1 code on platforms where exporting structures
1379 is problematical (for example in shared libraries) but exporting
1380 functions returning pointers to structures is not.
1381 [Steve Henson]
1382
1383 +) Add support for overriding the generation of SSL/TLS session IDs.
1384 These callbacks can be registered either in an SSL_CTX or per SSL.
1385 The purpose of this is to allow applications to control, if they wish,
1386 the arbitrary values chosen for use as session IDs, particularly as it
1387 can be useful for session caching in multiple-server environments. A
1388 command-line switch for testing this (and any client code that wishes
1389 to use such a feature) has been added to "s_server".
1390 [Geoff Thorpe, Lutz Jaenicke]
1391
1392 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1393 of the form '#if defined(...) || defined(...) || ...' and
1394 '#if !defined(...) && !defined(...) && ...'. This also avoids
1395 the growing number of special cases it was previously handling.
1396 [Richard Levitte]
1397
1398 *) Don't use getenv in library functions when run as setuid/setgid.
1399 New function OPENSSL_issetugid().
1400 [Ulf Moeller]
1401
1402 +) Make all configuration macros available for application by making
1403 sure they are available in opensslconf.h, by giving them names starting
1404 with "OPENSSL_" to avoid conflicts with other packages and by making
1405 sure e_os2.h will cover all platform-specific cases together with
1406 opensslconf.h.
1407 Additionally, it is now possible to define configuration/platform-
1408 specific names (called "system identities"). In the C code, these
1409 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1410 macro with the name beginning with "OPENSSL_SYS_", which is determined
1411 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1412 what is available.
1413 [Richard Levitte]
1414
1415 +) New option -set_serial to 'req' and 'x509' this allows the serial
1416 number to use to be specified on the command line. Previously self
1417 signed certificates were hard coded with serial number 0 and the
1418 CA options of 'x509' had to use a serial number in a file which was
1419 auto incremented.
1420 [Steve Henson]
1421
1422 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1423 due to incorrect handling of multi-threading:
1424
1425 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1426
1427 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1428
1429 3. Count how many times MemCheck_off() has been called so that
1430 nested use can be treated correctly. This also avoids
1431 inband-signalling in the previous code (which relied on the
1432 assumption that thread ID 0 is impossible).
1433 [Bodo Moeller]
1434
1435 +) New options to 'ca' utility to support V2 CRL entry extensions.
1436 Currently CRL reason, invalidity date and hold instruction are
1437 supported. Add new CRL extensions to V3 code and some new objects.
1438 [Steve Henson]
1439
1440 *) Add "-rand" option also to s_client and s_server.
1441 [Lutz Jaenicke]
1442
1443 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1444 disable standard block padding (aka PKCS#5 padding) in the EVP
1445 API, which was previously mandatory. This means that the data is
1446 not padded in any way and so the total length much be a multiple
1447 of the block size, otherwise an error occurs.
1448 [Steve Henson]
1449
1450 +) Initial (incomplete) OCSP SSL support.
1451 [Steve Henson]
1452
1453 *) Fix CPU detection on Irix 6.x.
1454 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1455 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1456
1457 +) New function OCSP_parse_url(). This splits up a URL into its host,
1458 port and path components: primarily to parse OCSP URLs. New -url
1459 option to ocsp utility.
1460 [Steve Henson]
1461
1462 +) New nonce behavior. The return value of OCSP_check_nonce() now
1463 reflects the various checks performed. Applications can decide
1464 whether to tolerate certain situations such as an absent nonce
1465 in a response when one was present in a request: the ocsp application
1466 just prints out a warning. New function OCSP_add1_basic_nonce()
1467 this is to allow responders to include a nonce in a response even if
1468 the request is nonce-less.
1469 [Steve Henson]
1470
1471 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1472 was empty.
1473 [Steve Henson]
1474 [This change does not apply to 0.9.7.]
1475
1476 *) Use the cached encoding of an X509_NAME structure rather than
1477 copying it. This is apparently the reason for the libsafe "errors"
1478 but the code is actually correct.
1479 [Steve Henson]
1480
1481 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1482 skipped when using openssl x509 multiple times on a single input file,
1483 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1484 [Bodo Moeller]
1485
1486 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1487 set string type: to handle setting ASN1_TIME structures. Fix ca
1488 utility to correctly initialize revocation date of CRLs.
1489 [Steve Henson]
1490
1491 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1492 the clients preferred ciphersuites and rather use its own preferences.
1493 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1494 Internet Explorer by ensuring unchanged hash method during stepup.
1495 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1496 [Lutz Jaenicke]
1497
1498 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1499 to aes and add a new 'exist' option to print out symbols that don't
1500 appear to exist.
1501 [Steve Henson]
1502
1503 +) Additional options to ocsp utility to allow flags to be set and
1504 additional certificates supplied.
1505 [Steve Henson]
1506
1507 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1508 OCSP client a number of certificate to only verify the response
1509 signature against.
1510 [Richard Levitte]
1511
1512 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1513 Bleichenbacher's DSA attack.
1514 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1515 to be set and top=0 forces the highest bit to be set; top=-1 is new
1516 and leaves the highest bit random.
1517 [Ulf Moeller, Bodo Moeller]
1518
1519 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1520 handle the new API. Currently only ECB, CBC modes supported. Add new
1521 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1522 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1523 [Ben Laurie, Steve Henson]
1524
1525 *) In the NCONF_...-based implementations for CONF_... queries
1526 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1527 a temporary CONF structure with the data component set to NULL
1528 (which gives segmentation faults in lh_retrieve).
1529 Instead, use NULL for the CONF pointer in CONF_get_string and
1530 CONF_get_number (which may use environment variables) and directly
1531 return NULL from CONF_get_section.
1532 [Bodo Moeller]
1533
1534 *) Fix potential buffer overrun for EBCDIC.
1535 [Ulf Moeller]
1536
1537 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1538 request to response.
1539 [Steve Henson]
1540
1541 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1542 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1543 extract information from a certificate request. OCSP_response_create()
1544 creates a response and optionally adds a basic response structure.
1545 OCSP_basic_add1_status() adds a complete single response to a basic
1546 response and returns the OCSP_SINGLERESP structure just added (to allow
1547 extensions to be included for example). OCSP_basic_add1_cert() adds a
1548 certificate to a basic response and OCSP_basic_sign() signs a basic
1549 response with various flags. New helper functions ASN1_TIME_check()
1550 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1551 (converts ASN1_TIME to GeneralizedTime).
1552 [Steve Henson]
1553
1554 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1555 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1556 structure from a certificate. X509_pubkey_digest() digests the public_key
1557 contents: this is used in various key identifiers.
1558 [Steve Henson]
1559
1560 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1561 keyUsage if basicConstraints absent for a CA.
1562 [Steve Henson]
1563
1564 *) Make SMIME_write_PKCS7() write mail header values with a format that
1565 is more generally accepted (no spaces before the semicolon), since
1566 some programs can't parse those values properly otherwise. Also make
1567 sure BIO's that break lines after each write do not create invalid
1568 headers.
1569 [Richard Levitte]
1570
1571 +) Make sk_sort() tolerate a NULL argument.
1572 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1573
1574 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1575 passed by the function are trusted implicitly. If any of them signed the
1576 response then it is assumed to be valid and is not verified.
1577 [Steve Henson]
1578
1579 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1580 macros previously used would not encode an empty SEQUENCE OF
1581 and break the signature.
1582 [Steve Henson]
1583 [This change does not apply to 0.9.7.]
1584
1585 *) Zero the premaster secret after deriving the master secret in
1586 DH ciphersuites.
1587 [Steve Henson]
1588
1589 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1590 to data. This was previously part of the PKCS7 ASN1 code. This
1591 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1592 [Steve Henson, reported by Kenneth R. Robinette
1593 <support@securenetterm.com>]
1594
1595 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1596 routines: without these tracing memory leaks is very painful.
1597 Fix leaks in PKCS12 and PKCS7 routines.
1598 [Steve Henson]
1599
1600 *) Add some EVP_add_digest_alias registrations (as found in
1601 OpenSSL_add_all_digests()) to SSL_library_init()
1602 aka OpenSSL_add_ssl_algorithms(). This provides improved
1603 compatibility with peers using X.509 certificates
1604 with unconventional AlgorithmIdentifier OIDs.
1605 [Bodo Moeller]
1606
1607 *) Fix for Irix with NO_ASM.
1608 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1609
1610 *) ./config script fixes.
1611 [Ulf Moeller, Richard Levitte]
1612
1613 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1614 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1615 effectively meant GeneralizedTime would never be used. Now it
1616 is initialised to -1 but X509_time_adj() now has to check the value
1617 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1618 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1619 [Steve Henson, reported by Kenneth R. Robinette
1620 <support@securenetterm.com>]
1621
1622 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1623 result in a zero length in the ASN1_INTEGER structure which was
1624 not consistent with the structure when d2i_ASN1_INTEGER() was used
1625 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1626 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1627 where it did not print out a minus for negative ASN1_INTEGER.
1628 [Steve Henson]
1629
1630 *) Fix 'openssl passwd -1'.
1631 [Bodo Moeller]
1632
1633 +) Add summary printout to ocsp utility. The various functions which
1634 convert status values to strings have been renamed to:
1635 OCSP_response_status_str(), OCSP_cert_status_str() and
1636 OCSP_crl_reason_str() and are no longer static. New options
1637 to verify nonce values and to disable verification. OCSP response
1638 printout format cleaned up.
1639 [Steve Henson]
1640
1641 +) Add additional OCSP certificate checks. These are those specified
1642 in RFC2560. This consists of two separate checks: the CA of the
1643 certificate being checked must either be the OCSP signer certificate
1644 or the issuer of the OCSP signer certificate. In the latter case the
1645 OCSP signer certificate must contain the OCSP signing extended key
1646 usage. This check is performed by attempting to match the OCSP
1647 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1648 in the OCSP_CERTID structures of the response.
1649 [Steve Henson]
1650
1651 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1652 and related routines. This uses the standard OpenSSL certificate
1653 verify routines to perform initial checks (just CA validity) and
1654 to obtain the certificate chain. Then additional checks will be
1655 performed on the chain. Currently the root CA is checked to see
1656 if it is explicitly trusted for OCSP signing. This is used to set
1657 a root CA as a global signing root: that is any certificate that
1658 chains to that CA is an acceptable OCSP signing certificate.
1659 [Steve Henson]
1660
1661 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1662 extensions from a separate configuration file.
1663 As when reading extensions from the main configuration file,
1664 the '-extensions ...' option may be used for specifying the
1665 section to use.
1666 [Massimiliano Pala <madwolf@comune.modena.it>]
1667
1668 *) Change PKCS12_key_gen_asc() so it can cope with non null
1669 terminated strings whose length is passed in the passlen
1670 parameter, for example from PEM callbacks. This was done
1671 by adding an extra length parameter to asc2uni().
1672 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1673
1674 +) New OCSP utility. Allows OCSP requests to be generated or
1675 read. The request can be sent to a responder and the output
1676 parsed, outputed or printed in text form. Not complete yet:
1677 still needs to check the OCSP response validity.
1678 [Steve Henson]
1679
1680 +) New subcommands for 'openssl ca':
1681 'openssl ca -status <serial>' prints the status of the cert with
1682 the given serial number (according to the index file).
1683 'openssl ca -updatedb' updates the expiry status of certificates
1684 in the index file.
1685 [Massimiliano Pala <madwolf@comune.modena.it>]
1686
1687 +) New '-newreq-nodes' command option to CA.pl. This is like
1688 '-newreq', but calls 'openssl req' with the '-nodes' option
1689 so that the resulting key is not encrypted.
1690 [Damien Miller <djm@mindrot.org>]
1691
1692 +) New configuration for the GNU Hurd.
1693 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1694
1695 +) Initial code to implement OCSP basic response verify. This
1696 is currently incomplete. Currently just finds the signer's
1697 certificate and verifies the signature on the response.
1698 [Steve Henson]
1699
1700 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1701 value of OPENSSLDIR. This is available via the new '-d' option
1702 to 'openssl version', and is also included in 'openssl version -a'.
1703 [Bodo Moeller]
1704
1705 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1706 call failed, free the DSA structure.
1707 [Bodo Moeller]
1708
1709 +) Allowing defining memory allocation callbacks that will be given
1710 file name and line number information in additional arguments
1711 (a const char* and an int). The basic functionality remains, as
1712 well as the original possibility to just replace malloc(),
1713 realloc() and free() by functions that do not know about these
1714 additional arguments. To register and find out the current
1715 settings for extended allocation functions, the following
1716 functions are provided:
1717
1718 CRYPTO_set_mem_ex_functions
1719 CRYPTO_set_locked_mem_ex_functions
1720 CRYPTO_get_mem_ex_functions
1721 CRYPTO_get_locked_mem_ex_functions
1722
1723 These work the same way as CRYPTO_set_mem_functions and friends.
1724 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1725 extended allocation function is enabled.
1726 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1727 a conventional allocation function is enabled.
1728 [Richard Levitte, Bodo Moeller]
1729
1730 *) Fix to uni2asc() to cope with zero length Unicode strings.
1731 These are present in some PKCS#12 files.
1732 [Steve Henson]
1733
1734 +) Finish off removing the remaining LHASH function pointer casts.
1735 There should no longer be any prototype-casting required when using
1736 the LHASH abstraction, and any casts that remain are "bugs". See
1737 the callback types and macros at the head of lhash.h for details
1738 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1739 [Geoff Thorpe]
1740
1741 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1742 If /dev/[u]random devices are not available or do not return enough
1743 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1744 be queried.
1745 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1746 /etc/entropy will be queried once each in this sequence, quering stops
1747 when enough entropy was collected without querying more sockets.
1748 [Lutz Jaenicke]
1749
1750 +) Change the Unix RAND_poll() variant to be able to poll several
1751 random devices, as specified by DEVRANDOM, until a sufficient amount
1752 of data has been collected. We spend at most 10 ms on each file
1753 (select timeout) and read in non-blocking mode. DEVRANDOM now
1754 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1755 (previously it was just the string "/dev/urandom"), so on typical
1756 platforms the 10 ms delay will never occur.
1757 Also separate out the Unix variant to its own file, rand_unix.c.
1758 For VMS, there's a currently-empty rand_vms.c.
1759 [Richard Levitte]
1760
1761 +) Move OCSP client related routines to ocsp_cl.c. These
1762 provide utility functions which an application needing
1763 to issue a request to an OCSP responder and analyse the
1764 response will typically need: as opposed to those which an
1765 OCSP responder itself would need which will be added later.
1766
1767 OCSP_request_sign() signs an OCSP request with an API similar
1768 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1769 response. OCSP_response_get1_basic() extracts basic response
1770 from response. OCSP_resp_find_status(): finds and extracts status
1771 information from an OCSP_CERTID structure (which will be created
1772 when the request structure is built). These are built from lower
1773 level functions which work on OCSP_SINGLERESP structures but
1774 wont normally be used unless the application wishes to examine
1775 extensions in the OCSP response for example.
1776
1777 Replace nonce routines with a pair of functions.
1778 OCSP_request_add1_nonce() adds a nonce value and optionally
1779 generates a random value. OCSP_check_nonce() checks the
1780 validity of the nonce in an OCSP response.
1781 [Steve Henson]
1782
1783 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1784 This doesn't copy the supplied OCSP_CERTID and avoids the
1785 need to free up the newly created id. Change return type
1786 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1787 This can then be used to add extensions to the request.
1788 Deleted OCSP_request_new(), since most of its functionality
1789 is now in OCSP_REQUEST_new() (and the case insensitive name
1790 clash) apart from the ability to set the request name which
1791 will be added elsewhere.
1792 [Steve Henson]
1793
1794 +) Update OCSP API. Remove obsolete extensions argument from
1795 various functions. Extensions are now handled using the new
1796 OCSP extension code. New simple OCSP HTTP function which
1797 can be used to send requests and parse the response.
1798 [Steve Henson]
1799
1800 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1801 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1802 uses the special reorder version of SET OF to sort the attributes
1803 and reorder them to match the encoded order. This resolves a long
1804 standing problem: a verify on a PKCS7 structure just after signing
1805 it used to fail because the attribute order did not match the
1806 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1807 it uses the received order. This is necessary to tolerate some broken
1808 software that does not order SET OF. This is handled by encoding
1809 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1810 to produce the required SET OF.
1811 [Steve Henson]
1812
1813 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1814 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1815 files to get correct declarations of the ASN.1 item variables.
1816 [Richard Levitte]
1817
1818 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1819 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1820 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1821 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1822 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1823 ASN1_ITEM and no wrapper functions.
1824 [Steve Henson]
1825
1826 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1827 replace the old function pointer based I/O routines. Change most of
1828 the *_d2i_bio() and *_d2i_fp() functions to use these.
1829 [Steve Henson]
1830
1831 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1832 lines, recognice more "algorithms" that can be deselected, and make
1833 it complain about algorithm deselection that isn't recognised.
1834 [Richard Levitte]
1835
1836 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1837 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1838 to use new functions. Add NO_ASN1_OLD which can be set to remove
1839 some old style ASN1 functions: this can be used to determine if old
1840 code will still work when these eventually go away.
1841 [Steve Henson]
1842
1843 +) New extension functions for OCSP structures, these follow the
1844 same conventions as certificates and CRLs.
1845 [Steve Henson]
1846
1847 +) New function X509V3_add1_i2d(). This automatically encodes and
1848 adds an extension. Its behaviour can be customised with various
1849 flags to append, replace or delete. Various wrappers added for
1850 certifcates and CRLs.
1851 [Steve Henson]
1852
1853 +) Fix to avoid calling the underlying ASN1 print routine when
1854 an extension cannot be parsed. Correct a typo in the
1855 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1856 [Steve Henson]
1857
1858 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1859 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1860 when writing a 32767 byte record.
1861 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1862
1863 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1864 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1865
1866 (RSA objects have a reference count access to which is protected
1867 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1868 so they are meant to be shared between threads.)
1869 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1870 "Reddie, Steven" <Steven.Reddie@ca.com>]
1871
1872 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1873 entries for variables.
1874 [Steve Henson]
1875
1876 *) Fix a deadlock in CRYPTO_mem_leaks().
1877 [Bodo Moeller]
1878
1879 +) Add functionality to apps/openssl.c for detecting locking
1880 problems: As the program is single-threaded, all we have
1881 to do is register a locking callback using an array for
1882 storing which locks are currently held by the program.
1883 [Bodo Moeller]
1884
1885 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1886 SSL_get_ex_data_X509_STORE_idx(), which is used in
1887 ssl_verify_cert_chain() and thus can be called at any time
1888 during TLS/SSL handshakes so that thread-safety is essential.
1889 Unfortunately, the ex_data design is not at all suited
1890 for multi-threaded use, so it probably should be abolished.
1891 [Bodo Moeller]
1892
1893 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1894 [Broadcom, tweaked and integrated by Geoff Thorpe]
1895
1896 +) Move common extension printing code to new function
1897 X509V3_print_extensions(). Reorganise OCSP print routines and
1898 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1899 [Steve Henson]
1900
1901 +) New function X509_signature_print() to remove duplication in some
1902 print routines.
1903 [Steve Henson]
1904
1905 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1906 set (this was treated exactly the same as SET OF previously). This
1907 is used to reorder the STACK representing the structure to match the
1908 encoding. This will be used to get round a problem where a PKCS7
1909 structure which was signed could not be verified because the STACK
1910 order did not reflect the encoded order.
1911 [Steve Henson]
1912
1913 +) Reimplement the OCSP ASN1 module using the new code.
1914 [Steve Henson]
1915
1916 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1917 for its ASN1 operations. The old style function pointers still exist
1918 for now but they will eventually go away.
1919 [Steve Henson]
1920
1921 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1922 completely replaces the old ASN1 functionality with a table driven
1923 encoder and decoder which interprets an ASN1_ITEM structure describing
1924 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1925 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1926 has also been converted to the new form.
1927 [Steve Henson]
1928
1929 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1930 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1931 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1932 for negative moduli.
1933 [Bodo Moeller]
1934
1935 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1936 of not touching the result's sign bit.
1937 [Bodo Moeller]
1938
1939 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1940 set.
1941 [Bodo Moeller]
1942
1943 +) Changed the LHASH code to use prototypes for callbacks, and created
1944 macros to declare and implement thin (optionally static) functions
1945 that provide type-safety and avoid function pointer casting for the
1946 type-specific callbacks.
1947 [Geoff Thorpe]
1948
1949 *) Use better test patterns in bntest.
1950 [Ulf Möller]
1951
1952 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1953 RFC 2712.
1954 [Veers Staats <staatsvr@asc.hpc.mil>,
1955 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1956
1957 *) rand_win.c fix for Borland C.
1958 [Ulf Möller]
1959
1960 *) BN_rshift bugfix for n == 0.
1961 [Bodo Moeller]
1962
1963 +) Reformat the FAQ so the different questions and answers can be divided
1964 in sections depending on the subject.
1965 [Richard Levitte]
1966
1967 +) Have the zlib compression code load ZLIB.DLL dynamically under
1968 Windows.
1969 [Richard Levitte]
1970
1971 +) New function BN_mod_sqrt for computing square roots modulo a prime
1972 (using the probabilistic Tonelli-Shanks algorithm unless
1973 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1974 be handled deterministically).
1975 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1976
1977 *) Add a 'bctest' script that checks for some known 'bc' bugs
1978 so that 'make test' does not abort just because 'bc' is broken.
1979 [Bodo Moeller]
1980
1981 *) Store verify_result within SSL_SESSION also for client side to
1982 avoid potential security hole. (Re-used sessions on the client side
1983 always resulted in verify_result==X509_V_OK, not using the original
1984 result of the server certificate verification.)
1985 [Lutz Jaenicke]
1986
1987 +) Make BN_mod_inverse faster by explicitly handling small quotients
1988 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1989 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1990 [Bodo Moeller]
1991
1992 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1993 SSL3_RT_APPLICATION_DATA, return 0.
1994 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1995 [Bodo Moeller]
1996
1997 *) Fix SSL_peek:
1998 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1999 releases, have been re-implemented by renaming the previous
2000 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2001 and ssl3_read_internal, respectively, and adding 'peek' parameters
2002 to them. The new ssl[23]_{read,peek} functions are calls to
2003 ssl[23]_read_internal with the 'peek' flag set appropriately.
2004 A 'peek' parameter has also been added to ssl3_read_bytes, which
2005 does the actual work for ssl3_read_internal.
2006 [Bodo Moeller]
2007
2008 +) New function BN_kronecker.
2009 [Bodo Moeller]
2010
2011 +) Fix BN_gcd so that it works on negative inputs; the result is
2012 positive unless both parameters are zero.
2013 Previously something reasonably close to an infinite loop was
2014 possible because numbers could be growing instead of shrinking
2015 in the implementation of Euclid's algorithm.
2016 [Bodo Moeller]
2017
2018 +) Fix BN_is_word() and BN_is_one() macros to take into account the
2019 sign of the number in question.
2020
2021 Fix BN_is_word(a,w) to work correctly for w == 0.
2022
2023 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2024 because its test if the absolute value of 'a' equals 'w'.
2025 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2026 it exists mostly for use in the implementations of BN_is_zero(),
2027 BN_is_one(), and BN_is_word().
2028 [Bodo Moeller]
2029
2030 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2031 the method-specific "init()" handler. Also clean up ex_data after
2032 calling the method-specific "finish()" handler. Previously, this was
2033 happening the other way round.
2034 [Geoff Thorpe]
2035
2036 +) New function BN_swap.
2037 [Bodo Moeller]
2038
2039 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2040 the exponentiation functions are more likely to produce reasonable
2041 results on negative inputs.
2042 [Bodo Moeller]
2043
2044 +) Change BN_mod_mul so that the result is always non-negative.
2045 Previously, it could be negative if one of the factors was negative;
2046 I don't think anyone really wanted that behaviour.
2047 [Bodo Moeller]
2048
2049 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2050 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2051 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2052 and add new functions:
2053
2054 BN_nnmod
2055 BN_mod_sqr
2056 BN_mod_add
2057 BN_mod_add_quick
2058 BN_mod_sub
2059 BN_mod_sub_quick
2060 BN_mod_lshift1
2061 BN_mod_lshift1_quick
2062 BN_mod_lshift
2063 BN_mod_lshift_quick
2064
2065 These functions always generate non-negative results.
2066
2067 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2068 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2069
2070 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2071 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2072 be reduced modulo m.
2073 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2074
2075 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2076 was actually never needed) and in BN_mul(). The removal in BN_mul()
2077 required a small change in bn_mul_part_recursive() and the addition
2078 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2079 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2080 bn_sub_words() and bn_add_words() except they take arrays with
2081 differing sizes.
2082 [Richard Levitte]
2083
2084 +) In 'openssl passwd', verify passwords read from the terminal
2085 unless the '-salt' option is used (which usually means that
2086 verification would just waste user's time since the resulting
2087 hash is going to be compared with some given password hash)
2088 or the new '-noverify' option is used.
2089
2090 This is an incompatible change, but it does not affect
2091 non-interactive use of 'openssl passwd' (passwords on the command
2092 line, '-stdin' option, '-in ...' option) and thus should not
2093 cause any problems.
2094 [Bodo Moeller]
2095
2096 +) Remove all references to RSAref, since there's no more need for it.
2097 [Richard Levitte]
2098
2099 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2100 The previous value, 12, was not always sufficient for BN_mod_exp().
2101 [Bodo Moeller]
2102
2103 +) Make DSO load along a path given through an environment variable
2104 (SHLIB_PATH) with shl_load().
2105 [Richard Levitte]
2106
2107 +) Constify the ENGINE code as a result of BIGNUM constification.
2108 Also constify the RSA code and most things related to it. In a
2109 few places, most notable in the depth of the ASN.1 code, ugly
2110 casts back to non-const were required (to be solved at a later
2111 time)
2112 [Richard Levitte]
2113
2114 +) Make it so the openssl application has all engines loaded by default.
2115 [Richard Levitte]
2116
2117 +) Constify the BIGNUM routines a little more.
2118 [Richard Levitte]
2119
2120 *) Make sure that shared libraries get the internal name engine with
2121 the full version number and not just 0. This should mark the
2122 shared libraries as not backward compatible. Of course, this should
2123 be changed again when we can guarantee backward binary compatibility.
2124 [Richard Levitte]
2125
2126 +) Add the following functions:
2127
2128 ENGINE_load_cswift()
2129 ENGINE_load_chil()
2130 ENGINE_load_atalla()
2131 ENGINE_load_nuron()
2132 ENGINE_load_builtin_engines()
2133
2134 That way, an application can itself choose if external engines that
2135 are built-in in OpenSSL shall ever be used or not. The benefit is
2136 that applications won't have to be linked with libdl or other dso
2137 libraries unless it's really needed.
2138
2139 Changed 'openssl engine' to load all engines on demand.
2140 Changed the engine header files to avoid the duplication of some
2141 declarations (they differed!).
2142 [Richard Levitte]
2143
2144 +) 'openssl engine' can now list capabilities.
2145 [Richard Levitte]
2146
2147 +) Better error reporting in 'openssl engine'.
2148 [Richard Levitte]
2149
2150 +) Never call load_dh_param(NULL) in s_server.
2151 [Bodo Moeller]
2152
2153 +) Add engine application. It can currently list engines by name and
2154 identity, and test if they are actually available.
2155 [Richard Levitte]
2156
2157 +) Improve RPM specification file by forcing symbolic linking and making
2158 sure the installed documentation is also owned by root.root.
2159 [Damien Miller <djm@mindrot.org>]
2160
2161 +) Give the OpenSSL applications more possibilities to make use of
2162 keys (public as well as private) handled by engines.
2163 [Richard Levitte]
2164
2165 +) Add OCSP code that comes from CertCo.
2166 [Richard Levitte]
2167
2168 +) Add VMS support for the Rijndael code.
2169 [Richard Levitte]
2170
2171 +) Added untested support for Nuron crypto accelerator.
2172 [Ben Laurie]
2173
2174 +) Add support for external cryptographic devices. This code was
2175 previously distributed separately as the "engine" branch.
2176 [Geoff Thorpe, Richard Levitte]
2177
2178 +) Rework the filename-translation in the DSO code. It is now possible to
2179 have far greater control over how a "name" is turned into a filename
2180 depending on the operating environment and any oddities about the
2181 different shared library filenames on each system.
2182 [Geoff Thorpe]
2183
2184 +) Support threads on FreeBSD-elf in Configure.
2185 [Richard Levitte]
2186
2187 +) Fix for SHA1 assembly problem with MASM: it produces
2188 warnings about corrupt line number information when assembling
2189 with debugging information. This is caused by the overlapping
2190 of two sections.
2191 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2192
2193 +) NCONF changes.
2194 NCONF_get_number() has no error checking at all. As a replacement,
2195 NCONF_get_number_e() is defined (_e for "error checking") and is
2196 promoted strongly. The old NCONF_get_number is kept around for
2197 binary backward compatibility.
2198 Make it possible for methods to load from something other than a BIO,
2199 by providing a function pointer that is given a name instead of a BIO.
2200 For example, this could be used to load configuration data from an
2201 LDAP server.
2202 [Richard Levitte]
2203
2204 *) Fix typo in get_cert_by_subject() in by_dir.c
2205 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2206
2207 *) Rework the system to generate shared libraries:
2208
2209 - Make note of the expected extension for the shared libraries and
2210 if there is a need for symbolic links from for example libcrypto.so.0
2211 to libcrypto.so.0.9.7. There is extended info in Configure for
2212 that.
2213
2214 - Make as few rebuilds of the shared libraries as possible.
2215
2216 - Still avoid linking the OpenSSL programs with the shared libraries.
2217
2218 - When installing, install the shared libraries separately from the
2219 static ones.
2220 [Richard Levitte]
2221
2222 +) Fix for non blocking accept BIOs. Added new I/O special reason
2223 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2224 with non blocking I/O was not possible because no retry code was
2225 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2226 this case.
2227 [Steve Henson]
2228
2229 +) Added the beginnings of Rijndael support.
2230 [Ben Laurie]
2231
2232 +) Fix for bug in DirectoryString mask setting. Add support for
2233 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2234 to allow certificate printing to more controllable, additional
2235 'certopt' option to 'x509' to allow new printing options to be
2236 set.
2237 [Steve Henson]
2238
2239 +) Clean old EAY MD5 hack from e_os.h.
2240 [Richard Levitte]
2241
2242 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2243
2244 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2245 and not in SSL_clear because the latter is also used by the
2246 accept/connect functions; previously, the settings made by
2247 SSL_set_read_ahead would be lost during the handshake.
2248 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2249
2250 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2251 Previously, it would create entries for disableed algorithms no
2252 matter what.
2253 [Richard Levitte]
2254
2255 *) Added several new manual pages for SSL_* function.
2256 [Lutz Jaenicke]
2257
2258 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2259
2260 *) In ssl23_get_client_hello, generate an error message when faced
2261 with an initial SSL 3.0/TLS record that is too small to contain the
2262 first two bytes of the ClientHello message, i.e. client_version.
2263 (Note that this is a pathologic case that probably has never happened
2264 in real life.) The previous approach was to use the version number
2265 from the record header as a substitute; but our protocol choice
2266 should not depend on that one because it is not authenticated
2267 by the Finished messages.
2268 [Bodo Moeller]
2269
2270 *) More robust randomness gathering functions for Windows.
2271 [Jeffrey Altman <jaltman@columbia.edu>]
2272
2273 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2274 not set then we don't setup the error code for issuer check errors
2275 to avoid possibly overwriting other errors which the callback does
2276 handle. If an application does set the flag then we assume it knows
2277 what it is doing and can handle the new informational codes
2278 appropriately.
2279 [Steve Henson]
2280
2281 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2282 a general "ANY" type, as such it should be able to decode anything
2283 including tagged types. However it didn't check the class so it would
2284 wrongly interpret tagged types in the same way as their universal
2285 counterpart and unknown types were just rejected. Changed so that the
2286 tagged and unknown types are handled in the same way as a SEQUENCE:
2287 that is the encoding is stored intact. There is also a new type
2288 "V_ASN1_OTHER" which is used when the class is not universal, in this
2289 case we have no idea what the actual type is so we just lump them all
2290 together.
2291 [Steve Henson]
2292
2293 *) On VMS, stdout may very well lead to a file that is written to
2294 in a record-oriented fashion. That means that every write() will
2295 write a separate record, which will be read separately by the
2296 programs trying to read from it. This can be very confusing.
2297
2298 The solution is to put a BIO filter in the way that will buffer
2299 text until a linefeed is reached, and then write everything a
2300 line at a time, so every record written will be an actual line,
2301 not chunks of lines and not (usually doesn't happen, but I've
2302 seen it once) several lines in one record. BIO_f_linebuffer() is
2303 the answer.
2304
2305 Currently, it's a VMS-only method, because that's where it has
2306 been tested well enough.
2307 [Richard Levitte]
2308
2309 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2310 it can return incorrect results.
2311 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2312 but it was in 0.9.6-beta[12].)
2313 [Bodo Moeller]
2314
2315 *) Disable the check for content being present when verifying detached
2316 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2317 include zero length content when signing messages.
2318 [Steve Henson]
2319
2320 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2321 BIO_ctrl (for BIO pairs).
2322 [Bodo Möller]
2323
2324 *) Add DSO method for VMS.
2325 [Richard Levitte]
2326
2327 *) Bug fix: Montgomery multiplication could produce results with the
2328 wrong sign.
2329 [Ulf Möller]
2330
2331 *) Add RPM specification openssl.spec and modify it to build three
2332 packages. The default package contains applications, application
2333 documentation and run-time libraries. The devel package contains
2334 include files, static libraries and function documentation. The
2335 doc package contains the contents of the doc directory. The original
2336 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2337 [Richard Levitte]
2338
2339 *) Add a large number of documentation files for many SSL routines.
2340 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2341
2342 *) Add a configuration entry for Sony News 4.
2343 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2344
2345 *) Don't set the two most significant bits to one when generating a
2346 random number < q in the DSA library.
2347 [Ulf Möller]
2348
2349 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2350 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2351 the underlying transport is blocking) if a handshake took place.
2352 (The default behaviour is needed by applications such as s_client
2353 and s_server that use select() to determine when to use SSL_read;
2354 but for applications that know in advance when to expect data, it
2355 just makes things more complicated.)
2356 [Bodo Moeller]
2357
2358 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2359 from EGD.
2360 [Ben Laurie]
2361
2362 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2363 work better on such systems.
2364 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2365
2366 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2367 Update PKCS12_parse() so it copies the friendlyName and the
2368 keyid to the certificates aux info.
2369 [Steve Henson]
2370
2371 *) Fix bug in PKCS7_verify() which caused an infinite loop
2372 if there was more than one signature.
2373 [Sven Uszpelkat <su@celocom.de>]
2374
2375 *) Major change in util/mkdef.pl to include extra information
2376 about each symbol, as well as presentig variables as well
2377 as functions. This change means that there's n more need
2378 to rebuild the .num files when some algorithms are excluded.
2379 [Richard Levitte]
2380
2381 *) Allow the verify time to be set by an application,
2382 rather than always using the current time.
2383 [Steve Henson]
2384
2385 *) Phase 2 verify code reorganisation. The certificate
2386 verify code now looks up an issuer certificate by a
2387 number of criteria: subject name, authority key id
2388 and key usage. It also verifies self signed certificates
2389 by the same criteria. The main comparison function is
2390 X509_check_issued() which performs these checks.
2391
2392 Lot of changes were necessary in order to support this
2393 without completely rewriting the lookup code.
2394
2395 Authority and subject key identifier are now cached.
2396
2397 The LHASH 'certs' is X509_STORE has now been replaced
2398 by a STACK_OF(X509_OBJECT). This is mainly because an
2399 LHASH can't store or retrieve multiple objects with
2400 the same hash value.
2401
2402 As a result various functions (which were all internal
2403 use only) have changed to handle the new X509_STORE
2404 structure. This will break anything that messed round
2405 with X509_STORE internally.
2406
2407 The functions X509_STORE_add_cert() now checks for an
2408 exact match, rather than just subject name.
2409
2410 The X509_STORE API doesn't directly support the retrieval
2411 of multiple certificates matching a given criteria, however
2412 this can be worked round by performing a lookup first
2413 (which will fill the cache with candidate certificates)
2414 and then examining the cache for matches. This is probably
2415 the best we can do without throwing out X509_LOOKUP
2416 entirely (maybe later...).
2417
2418 The X509_VERIFY_CTX structure has been enhanced considerably.
2419
2420 All certificate lookup operations now go via a get_issuer()
2421 callback. Although this currently uses an X509_STORE it
2422 can be replaced by custom lookups. This is a simple way
2423 to bypass the X509_STORE hackery necessary to make this
2424 work and makes it possible to use more efficient techniques
2425 in future. A very simple version which uses a simple
2426 STACK for its trusted certificate store is also provided
2427 using X509_STORE_CTX_trusted_stack().
2428
2429 The verify_cb() and verify() callbacks now have equivalents
2430 in the X509_STORE_CTX structure.
2431
2432 X509_STORE_CTX also has a 'flags' field which can be used
2433 to customise the verify behaviour.
2434 [Steve Henson]
2435
2436 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2437 excludes S/MIME capabilities.
2438 [Steve Henson]
2439
2440 *) When a certificate request is read in keep a copy of the
2441 original encoding of the signed data and use it when outputing
2442 again. Signatures then use the original encoding rather than
2443 a decoded, encoded version which may cause problems if the
2444 request is improperly encoded.
2445 [Steve Henson]
2446
2447 *) For consistency with other BIO_puts implementations, call
2448 buffer_write(b, ...) directly in buffer_puts instead of calling
2449 BIO_write(b, ...).
2450
2451 In BIO_puts, increment b->num_write as in BIO_write.
2452 [Peter.Sylvester@EdelWeb.fr]
2453
2454 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2455 BN_zero, we may not return a BIGNUM with an array consisting of
2456 words set to zero.)
2457 [Bodo Moeller]
2458
2459 *) Avoid calling abort() from within the library when problems are
2460 detected, except if preprocessor symbols have been defined
2461 (such as REF_CHECK, BN_DEBUG etc.).
2462 [Bodo Moeller]
2463
2464 *) New openssl application 'rsautl'. This utility can be
2465 used for low level RSA operations. DER public key
2466 BIO/fp routines also added.
2467 [Steve Henson]
2468
2469 *) New Configure entry and patches for compiling on QNX 4.
2470 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2471
2472 *) A demo state-machine implementation was sponsored by
2473 Nuron (http://www.nuron.com/) and is now available in
2474 demos/state_machine.
2475 [Ben Laurie]
2476
2477 *) New options added to the 'dgst' utility for signature
2478 generation and verification.
2479 [Steve Henson]
2480
2481 *) Unrecognized PKCS#7 content types are now handled via a
2482 catch all ASN1_TYPE structure. This allows unsupported
2483 types to be stored as a "blob" and an application can
2484 encode and decode it manually.
2485 [Steve Henson]
2486
2487 *) Fix various signed/unsigned issues to make a_strex.c
2488 compile under VC++.
2489 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2490
2491 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2492 length if passed a buffer. ASN1_INTEGER_to_BN failed
2493 if passed a NULL BN and its argument was negative.
2494 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2495
2496 *) Modification to PKCS#7 encoding routines to output definite
2497 length encoding. Since currently the whole structures are in
2498 memory there's not real point in using indefinite length
2499 constructed encoding. However if OpenSSL is compiled with
2500 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2501 [Steve Henson]
2502
2503 *) Added BIO_vprintf() and BIO_vsnprintf().
2504 [Richard Levitte]
2505
2506 *) Added more prefixes to parse for in the the strings written
2507 through a logging bio, to cover all the levels that are available
2508 through syslog. The prefixes are now:
2509
2510 PANIC, EMERG, EMR => LOG_EMERG
2511 ALERT, ALR => LOG_ALERT
2512 CRIT, CRI => LOG_CRIT
2513 ERROR, ERR => LOG_ERR
2514 WARNING, WARN, WAR => LOG_WARNING
2515 NOTICE, NOTE, NOT => LOG_NOTICE
2516 INFO, INF => LOG_INFO
2517 DEBUG, DBG => LOG_DEBUG
2518
2519 and as before, if none of those prefixes are present at the
2520 beginning of the string, LOG_ERR is chosen.
2521
2522 On Win32, the LOG_* levels are mapped according to this:
2523
2524 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2525 LOG_WARNING => EVENTLOG_WARNING_TYPE
2526 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2527
2528 [Richard Levitte]
2529
2530 *) Made it possible to reconfigure with just the configuration
2531 argument "reconf" or "reconfigure". The command line arguments
2532 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2533 and are retrieved from there when reconfiguring.
2534 [Richard Levitte]
2535
2536 *) MD4 implemented.
2537 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2538
2539 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2540 [Richard Levitte]
2541
2542 *) The obj_dat.pl script was messing up the sorting of object
2543 names. The reason was that it compared the quoted version
2544 of strings as a result "OCSP" > "OCSP Signing" because
2545 " > SPACE. Changed script to store unquoted versions of
2546 names and add quotes on output. It was also omitting some
2547 names from the lookup table if they were given a default
2548 value (that is if SN is missing it is given the same
2549 value as LN and vice versa), these are now added on the
2550 grounds that if an object has a name we should be able to
2551 look it up. Finally added warning output when duplicate
2552 short or long names are found.
2553 [Steve Henson]
2554
2555 *) Changes needed for Tandem NSK.
2556 [Scott Uroff <scott@xypro.com>]
2557
2558 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2559 RSA_padding_check_SSLv23(), special padding was never detected
2560 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2561 version rollback attacks was not effective.
2562
2563 In s23_clnt.c, don't use special rollback-attack detection padding
2564 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2565 client; similarly, in s23_srvr.c, don't do the rollback check if
2566 SSL 2.0 is the only protocol enabled in the server.
2567 [Bodo Moeller]
2568
2569 *) Make it possible to get hexdumps of unprintable data with 'openssl
2570 asn1parse'. By implication, the functions ASN1_parse_dump() and
2571 BIO_dump_indent() are added.
2572 [Richard Levitte]
2573
2574 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2575 these print out strings and name structures based on various
2576 flags including RFC2253 support and proper handling of
2577 multibyte characters. Added options to the 'x509' utility
2578 to allow the various flags to be set.
2579 [Steve Henson]
2580
2581 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2582 Also change the functions X509_cmp_current_time() and
2583 X509_gmtime_adj() work with an ASN1_TIME structure,
2584 this will enable certificates using GeneralizedTime in validity
2585 dates to be checked.
2586 [Steve Henson]
2587
2588 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2589 negative public key encodings) on by default,
2590 NO_NEG_PUBKEY_BUG can be set to disable it.
2591 [Steve Henson]
2592
2593 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2594 content octets. An i2c_ASN1_OBJECT is unnecessary because
2595 the encoding can be trivially obtained from the structure.
2596 [Steve Henson]
2597
2598 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2599 not read locks (CRYPTO_r_[un]lock).
2600 [Bodo Moeller]
2601
2602 *) A first attempt at creating official support for shared
2603 libraries through configuration. I've kept it so the
2604 default is static libraries only, and the OpenSSL programs
2605 are always statically linked for now, but there are
2606 preparations for dynamic linking in place.
2607 This has been tested on Linux and Tru64.
2608 [Richard Levitte]
2609
2610 *) Randomness polling function for Win9x, as described in:
2611 Peter Gutmann, Software Generation of Practically Strong
2612 Random Numbers.
2613 [Ulf Möller]
2614
2615 *) Fix so PRNG is seeded in req if using an already existing
2616 DSA key.
2617 [Steve Henson]
2618
2619 *) New options to smime application. -inform and -outform
2620 allow alternative formats for the S/MIME message including
2621 PEM and DER. The -content option allows the content to be
2622 specified separately. This should allow things like Netscape
2623 form signing output easier to verify.
2624 [Steve Henson]
2625
2626 *) Fix the ASN1 encoding of tags using the 'long form'.
2627 [Steve Henson]
2628
2629 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2630 STRING types. These convert content octets to and from the
2631 underlying type. The actual tag and length octets are
2632 already assumed to have been read in and checked. These
2633 are needed because all other string types have virtually
2634 identical handling apart from the tag. By having versions
2635 of the ASN1 functions that just operate on content octets
2636 IMPLICIT tagging can be handled properly. It also allows
2637 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2638 and ASN1_INTEGER are identical apart from the tag.
2639 [Steve Henson]
2640
2641 *) Change the handling of OID objects as follows:
2642
2643 - New object identifiers are inserted in objects.txt, following
2644 the syntax given in objects.README.
2645 - objects.pl is used to process obj_mac.num and create a new
2646 obj_mac.h.
2647 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2648 obj_mac.h.
2649
2650 This is currently kind of a hack, and the perl code in objects.pl
2651 isn't very elegant, but it works as I intended. The simplest way
2652 to check that it worked correctly is to look in obj_dat.h and
2653 check the array nid_objs and make sure the objects haven't moved
2654 around (this is important!). Additions are OK, as well as
2655 consistent name changes.
2656 [Richard Levitte]
2657
2658 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2659 [Bodo Moeller]
2660
2661 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2662 The given file adds to whatever has already been seeded into the
2663 random pool through the RANDFILE configuration file option or
2664 environment variable, or the default random state file.
2665 [Richard Levitte]
2666
2667 *) mkstack.pl now sorts each macro group into lexical order.
2668 Previously the output order depended on the order the files
2669 appeared in the directory, resulting in needless rewriting
2670 of safestack.h .
2671 [Steve Henson]
2672
2673 *) Patches to make OpenSSL compile under Win32 again. Mostly
2674 work arounds for the VC++ problem that it treats func() as
2675 func(void). Also stripped out the parts of mkdef.pl that
2676 added extra typesafe functions: these no longer exist.
2677 [Steve Henson]
2678
2679 *) Reorganisation of the stack code. The macros are now all
2680 collected in safestack.h . Each macro is defined in terms of
2681 a "stack macro" of the form SKM_<name>(type, a, b). The
2682 DEBUG_SAFESTACK is now handled in terms of function casts,
2683 this has the advantage of retaining type safety without the
2684 use of additional functions. If DEBUG_SAFESTACK is not defined
2685 then the non typesafe macros are used instead. Also modified the
2686 mkstack.pl script to handle the new form. Needs testing to see
2687 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2688 the default if no major problems. Similar behaviour for ASN1_SET_OF
2689 and PKCS12_STACK_OF.
2690 [Steve Henson]
2691
2692 *) When some versions of IIS use the 'NET' form of private key the
2693 key derivation algorithm is different. Normally MD5(password) is
2694 used as a 128 bit RC4 key. In the modified case
2695 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2696 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2697 as the old Netscape_RSA functions except they have an additional
2698 'sgckey' parameter which uses the modified algorithm. Also added
2699 an -sgckey command line option to the rsa utility. Thanks to
2700 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2701 algorithm to openssl-dev.
2702 [Steve Henson]
2703
2704 *) The evp_local.h macros were using 'c.##kname' which resulted in
2705 invalid expansion on some systems (SCO 5.0.5 for example).
2706 Corrected to 'c.kname'.
2707 [Phillip Porch <root@theporch.com>]
2708
2709 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2710 a STACK of email addresses from a certificate or request, these look
2711 in the subject name and the subject alternative name extensions and
2712 omit any duplicate addresses.
2713 [Steve Henson]
2714
2715 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2716 This makes DSA verification about 2 % faster.
2717 [Bodo Moeller]
2718
2719 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2720 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2721 plus overhead for 1024 bit moduli).
2722 This makes exponentiations about 0.5 % faster for 1024 bit
2723 exponents (as measured by "openssl speed rsa2048").
2724 [Bodo Moeller]
2725
2726 *) Rename memory handling macros to avoid conflicts with other
2727 software:
2728 Malloc => OPENSSL_malloc
2729 Malloc_locked => OPENSSL_malloc_locked
2730 Realloc => OPENSSL_realloc
2731 Free => OPENSSL_free
2732 [Richard Levitte]
2733
2734 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2735 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2736 [Bodo Moeller]
2737
2738 *) CygWin32 support.
2739 [John Jarvie <jjarvie@newsguy.com>]
2740
2741 *) The type-safe stack code has been rejigged. It is now only compiled
2742 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2743 by default all type-specific stack functions are "#define"d back to
2744 standard stack functions. This results in more streamlined output
2745 but retains the type-safety checking possibilities of the original
2746 approach.
2747 [Geoff Thorpe]
2748
2749 *) The STACK code has been cleaned up, and certain type declarations
2750 that didn't make a lot of sense have been brought in line. This has
2751 also involved a cleanup of sorts in safestack.h to more correctly
2752 map type-safe stack functions onto their plain stack counterparts.
2753 This work has also resulted in a variety of "const"ifications of
2754 lots of the code, especially "_cmp" operations which should normally
2755 be prototyped with "const" parameters anyway.
2756 [Geoff Thorpe]
2757
2758 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2759 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2760 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2761 where all of 'md' is used each time the PRNG is used, but 'state'
2762 is used only indexed by a cyclic counter. As entropy may not be
2763 well distributed from the beginning, 'md' is important as a
2764 chaining variable. However, the output function chains only half
2765 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2766 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2767 in all of 'state' being rewritten, with the new values depending
2768 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2769 [Bodo Moeller]
2770
2771 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2772 the handshake is continued after ssl_verify_cert_chain();
2773 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2774 can lead to 'unexplainable' connection aborts later.
2775 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2776
2777 *) Major EVP API cipher revision.
2778 Add hooks for extra EVP features. This allows various cipher
2779 parameters to be set in the EVP interface. Support added for variable
2780 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2781 setting of RC2 and RC5 parameters.
2782
2783 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2784 ciphers.
2785
2786 Remove lots of duplicated code from the EVP library. For example *every*
2787 cipher init() function handles the 'iv' in the same way according to the
2788 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2789 for CFB and OFB modes they zero ctx->num.
2790
2791 New functionality allows removal of S/MIME code RC2 hack.
2792
2793 Most of the routines have the same form and so can be declared in terms
2794 of macros.
2795
2796 By shifting this to the top level EVP_CipherInit() it can be removed from
2797 all individual ciphers. If the cipher wants to handle IVs or keys
2798 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2799 flags.
2800
2801 Change lots of functions like EVP_EncryptUpdate() to now return a
2802 value: although software versions of the algorithms cannot fail
2803 any installed hardware versions can.
2804 [Steve Henson]
2805
2806 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2807 this option is set, tolerate broken clients that send the negotiated
2808 protocol version number instead of the requested protocol version
2809 number.
2810 [Bodo Moeller]
2811
2812 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2813 i.e. non-zero for export ciphersuites, zero otherwise.
2814 Previous versions had this flag inverted, inconsistent with
2815 rsa_tmp_cb (..._TMP_RSA_CB).
2816 [Bodo Moeller; problem reported by Amit Chopra]
2817
2818 *) Add missing DSA library text string. Work around for some IIS
2819 key files with invalid SEQUENCE encoding.
2820 [Steve Henson]
2821
2822 *) Add a document (doc/standards.txt) that list all kinds of standards
2823 and so on that are implemented in OpenSSL.
2824 [Richard Levitte]
2825
2826 *) Enhance c_rehash script. Old version would mishandle certificates
2827 with the same subject name hash and wouldn't handle CRLs at all.
2828 Added -fingerprint option to crl utility, to support new c_rehash
2829 features.
2830 [Steve Henson]
2831
2832 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2833 [Ulf Möller]
2834
2835 *) Fix for SSL server purpose checking. Server checking was
2836 rejecting certificates which had extended key usage present
2837 but no ssl client purpose.
2838 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2839
2840 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2841 is a little unclear about how a blank password is handled.
2842 Since the password in encoded as a BMPString with terminating
2843 double NULL a zero length password would end up as just the
2844 double NULL. However no password at all is different and is
2845 handled differently in the PKCS#12 key generation code. NS
2846 treats a blank password as zero length. MSIE treats it as no
2847 password on export: but it will try both on import. We now do
2848 the same: PKCS12_parse() tries zero length and no password if
2849 the password is set to "" or NULL (NULL is now a valid password:
2850 it wasn't before) as does the pkcs12 application.
2851 [Steve Henson]
2852
2853 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2854 perror when PEM_read_bio_X509_REQ fails, the error message must
2855 be obtained from the error queue.
2856 [Bodo Moeller]
2857
2858 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2859 it in ERR_remove_state if appropriate, and change ERR_get_state
2860 accordingly to avoid race conditions (this is necessary because
2861 thread_hash is no longer constant once set).
2862 [Bodo Moeller]
2863
2864 *) Bugfix for linux-elf makefile.one.
2865 [Ulf Möller]
2866
2867 *) RSA_get_default_method() will now cause a default
2868 RSA_METHOD to be chosen if one doesn't exist already.
2869 Previously this was only set during a call to RSA_new()
2870 or RSA_new_method(NULL) meaning it was possible for
2871 RSA_get_default_method() to return NULL.
2872 [Geoff Thorpe]
2873
2874 *) Added native name translation to the existing DSO code
2875 that will convert (if the flag to do so is set) filenames
2876 that are sufficiently small and have no path information
2877 into a canonical native form. Eg. "blah" converted to
2878 "libblah.so" or "blah.dll" etc.
2879 [Geoff Thorpe]
2880
2881 *) New function ERR_error_string_n(e, buf, len) which is like
2882 ERR_error_string(e, buf), but writes at most 'len' bytes
2883 including the 0 terminator. For ERR_error_string_n, 'buf'
2884 may not be NULL.
2885 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2886
2887 *) CONF library reworked to become more general. A new CONF
2888 configuration file reader "class" is implemented as well as a
2889 new functions (NCONF_*, for "New CONF") to handle it. The now
2890 old CONF_* functions are still there, but are reimplemented to
2891 work in terms of the new functions. Also, a set of functions
2892 to handle the internal storage of the configuration data is
2893 provided to make it easier to write new configuration file
2894 reader "classes" (I can definitely see something reading a
2895 configuration file in XML format, for example), called _CONF_*,
2896 or "the configuration storage API"...
2897
2898 The new configuration file reading functions are:
2899
2900 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2901 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2902
2903 NCONF_default, NCONF_WIN32
2904
2905 NCONF_dump_fp, NCONF_dump_bio
2906
2907 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2908 NCONF_new creates a new CONF object. This works in the same way
2909 as other interfaces in OpenSSL, like the BIO interface.
2910 NCONF_dump_* dump the internal storage of the configuration file,
2911 which is useful for debugging. All other functions take the same
2912 arguments as the old CONF_* functions wth the exception of the
2913 first that must be a `CONF *' instead of a `LHASH *'.
2914
2915 To make it easer to use the new classes with the old CONF_* functions,
2916 the function CONF_set_default_method is provided.
2917 [Richard Levitte]
2918
2919 *) Add '-tls1' option to 'openssl ciphers', which was already
2920 mentioned in the documentation but had not been implemented.
2921 (This option is not yet really useful because even the additional
2922 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2923 [Bodo Moeller]
2924
2925 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2926 OpenSSL-based applications) load shared libraries and bind to
2927 them in a portable way.
2928 [Geoff Thorpe, with contributions from Richard Levitte]
2929
2930 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2931
2932 *) Make sure _lrotl and _lrotr are only used with MSVC.
2933
2934 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2935 (the default implementation of RAND_status).
2936
2937 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2938 to '-clrext' (= clear extensions), as intended and documented.
2939 [Bodo Moeller; inconsistency pointed out by Michael Attili
2940 <attili@amaxo.com>]
2941
2942 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2943 was larger than the MD block size.
2944 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2945
2946 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2947 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2948 using the passed key: if the passed key was a private key the result
2949 of X509_print(), for example, would be to print out all the private key
2950 components.
2951 [Steve Henson]
2952
2953 *) des_quad_cksum() byte order bug fix.
2954 [Ulf Möller, using the problem description in krb4-0.9.7, where
2955 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2956
2957 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2958 discouraged.
2959 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2960
2961 *) For easily testing in shell scripts whether some command
2962 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2963 returns with exit code 0 iff no command of the given name is available.
2964 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2965 the output goes to stdout and nothing is printed to stderr.
2966 Additional arguments are always ignored.
2967
2968 Since for each cipher there is a command of the same name,
2969 the 'no-cipher' compilation switches can be tested this way.
2970
2971 ('openssl no-XXX' is not able to detect pseudo-commands such
2972 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2973 [Bodo Moeller]
2974
2975 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2976 [Bodo Moeller]
2977
2978 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2979 is set; it will be thrown away anyway because each handshake creates
2980 its own key.
2981 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2982 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2983 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2984 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2985 [Bodo Moeller]
2986
2987 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2988 'Q' and 'R' lose their special meanings (quit/renegotiate).
2989 This is part of what -quiet does; unlike -quiet, -ign_eof
2990 does not suppress any output.
2991 [Richard Levitte]
2992
2993 *) Add compatibility options to the purpose and trust code. The
2994 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2995 accepts a certificate or CA, this was the previous behaviour,
2996 with all the associated security issues.
2997
2998 X509_TRUST_COMPAT is the old trust behaviour: only and
2999 automatically trust self signed roots in certificate store. A
3000 new trust setting X509_TRUST_DEFAULT is used to specify that
3001 a purpose has no associated trust setting and it should instead
3002 use the value in the default purpose.
3003 [Steve Henson]
3004
3005 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3006 and fix a memory leak.
3007 [Steve Henson]
3008
3009 *) In util/mkerr.pl (which implements 'make errors'), preserve
3010 reason strings from the previous version of the .c file, as
3011 the default to have only downcase letters (and digits) in
3012 automatically generated reasons codes is not always appropriate.
3013 [Bodo Moeller]
3014
3015 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3016 using strerror. Previously, ERR_reason_error_string() returned
3017 library names as reason strings for SYSerr; but SYSerr is a special
3018 case where small numbers are errno values, not library numbers.
3019 [Bodo Moeller]
3020
3021 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3022 converts DSA parameters into DH parameters. (When creating parameters,
3023 DSA_generate_parameters is used.)
3024 [Bodo Moeller]
3025
3026 *) Include 'length' (recommended exponent length) in C code generated
3027 by 'openssl dhparam -C'.
3028 [Bodo Moeller]
3029
3030 *) The second argument to set_label in perlasm was already being used
3031 so couldn't be used as a "file scope" flag. Moved to third argument
3032 which was free.
3033 [Steve Henson]
3034
3035 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3036 instead of RAND_bytes for encryption IVs and salts.
3037 [Bodo Moeller]
3038
3039 *) Include RAND_status() into RAND_METHOD instead of implementing
3040 it only for md_rand.c Otherwise replacing the PRNG by calling
3041 RAND_set_rand_method would be impossible.
3042 [Bodo Moeller]
3043
3044 *) Don't let DSA_generate_key() enter an infinite loop if the random
3045 number generation fails.
3046 [Bodo Moeller]
3047
3048 *) New 'rand' application for creating pseudo-random output.
3049 [Bodo Moeller]
3050
3051 *) Added configuration support for Linux/IA64
3052 [Rolf Haberrecker <rolf@suse.de>]
3053
3054 *) Assembler module support for Mingw32.
3055 [Ulf Möller]
3056
3057 *) Shared library support for HPUX (in shlib/).
3058 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3059
3060 *) Shared library support for Solaris gcc.
3061 [Lutz Behnke <behnke@trustcenter.de>]
3062
3063 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3064
3065 *) PKCS7_encrypt() was adding text MIME headers twice because they
3066 were added manually and by SMIME_crlf_copy().
3067 [Steve Henson]
3068
3069 *) In bntest.c don't call BN_rand with zero bits argument.
3070 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3071
3072 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3073 case was implemented. This caused BN_div_recp() to fail occasionally.
3074 [Ulf Möller]
3075
3076 *) Add an optional second argument to the set_label() in the perl
3077 assembly language builder. If this argument exists and is set
3078 to 1 it signals that the assembler should use a symbol whose
3079 scope is the entire file, not just the current function. This
3080 is needed with MASM which uses the format label:: for this scope.
3081 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3082
3083 *) Change the ASN1 types so they are typedefs by default. Before
3084 almost all types were #define'd to ASN1_STRING which was causing
3085 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3086 for example.
3087 [Steve Henson]
3088
3089 *) Change names of new functions to the new get1/get0 naming
3090 convention: After 'get1', the caller owns a reference count
3091 and has to call ..._free; 'get0' returns a pointer to some
3092 data structure without incrementing reference counters.
3093 (Some of the existing 'get' functions increment a reference
3094 counter, some don't.)
3095 Similarly, 'set1' and 'add1' functions increase reference
3096 counters or duplicate objects.
3097 [Steve Henson]
3098
3099 *) Allow for the possibility of temp RSA key generation failure:
3100 the code used to assume it always worked and crashed on failure.
3101 [Steve Henson]
3102
3103 *) Fix potential buffer overrun problem in BIO_printf().
3104 [Ulf Möller, using public domain code by Patrick Powell; problem
3105 pointed out by David Sacerdote <das33@cornell.edu>]
3106
3107 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3108 RAND_egd() and RAND_status(). In the command line application,
3109 the EGD socket can be specified like a seed file using RANDFILE
3110 or -rand.
3111 [Ulf Möller]
3112
3113 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3114 Some CAs (e.g. Verisign) distribute certificates in this form.
3115 [Steve Henson]
3116
3117 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3118 list to exclude them. This means that no special compilation option
3119 is needed to use anonymous DH: it just needs to be included in the
3120 cipher list.
3121 [Steve Henson]
3122
3123 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3124 EVP_MD_type. The old functionality is available in a new macro called
3125 EVP_MD_md(). Change code that uses it and update docs.
3126 [Steve Henson]
3127
3128 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3129 where the 'void *' argument is replaced by a function pointer argument.
3130 Previously 'void *' was abused to point to functions, which works on
3131 many platforms, but is not correct. As these functions are usually
3132 called by macros defined in OpenSSL header files, most source code
3133 should work without changes.
3134 [Richard Levitte]
3135
3136 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3137 sections with information on -D... compiler switches used for
3138 compiling the library so that applications can see them. To enable
3139 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3140 must be defined. E.g.,
3141 #define OPENSSL_ALGORITHM_DEFINES
3142 #include <openssl/opensslconf.h>
3143 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3144 [Richard Levitte, Ulf and Bodo Möller]
3145
3146 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3147 record layer.
3148 [Bodo Moeller]
3149
3150 *) Change the 'other' type in certificate aux info to a STACK_OF
3151 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3152 the required ASN1 format: arbitrary types determined by an OID.
3153 [Steve Henson]
3154
3155 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3156 argument to 'req'. This is not because the function is newer or
3157 better than others it just uses the work 'NEW' in the certificate
3158 request header lines. Some software needs this.
3159 [Steve Henson]
3160
3161 *) Reorganise password command line arguments: now passwords can be
3162 obtained from various sources. Delete the PEM_cb function and make
3163 it the default behaviour: i.e. if the callback is NULL and the
3164 usrdata argument is not NULL interpret it as a null terminated pass
3165 phrase. If usrdata and the callback are NULL then the pass phrase
3166 is prompted for as usual.
3167 [Steve Henson]
3168
3169 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3170 the support is automatically enabled. The resulting binaries will
3171 autodetect the card and use it if present.
3172 [Ben Laurie and Compaq Inc.]
3173
3174 *) Work around for Netscape hang bug. This sends certificate request
3175 and server done in one record. Since this is perfectly legal in the
3176 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3177 the bugs/SSLv3 entry for more info.
3178 [Steve Henson]
3179
3180 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3181 [Andy Polyakov]
3182
3183 *) Add -rand argument to smime and pkcs12 applications and read/write
3184 of seed file.
3185 [Steve Henson]
3186
3187 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3188 [Bodo Moeller]
3189
3190 *) Add command line password options to the remaining applications.
3191 [Steve Henson]
3192
3193 *) Bug fix for BN_div_recp() for numerators with an even number of
3194 bits.
3195 [Ulf Möller]
3196
3197 *) More tests in bntest.c, and changed test_bn output.
3198 [Ulf Möller]
3199
3200 *) ./config recognizes MacOS X now.
3201 [Andy Polyakov]
3202
3203 *) Bug fix for BN_div() when the first words of num and divsor are
3204 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3205 [Ulf Möller]
3206
3207 *) Add support for various broken PKCS#8 formats, and command line
3208 options to produce them.
3209 [Steve Henson]
3210
3211 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3212 get temporary BIGNUMs from a BN_CTX.
3213 [Ulf Möller]
3214
3215 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3216 for p == 0.
3217 [Ulf Möller]
3218
3219 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3220 include a #define from the old name to the new. The original intent
3221 was that statically linked binaries could for example just call
3222 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3223 link with digests. This never worked becayse SSLeay_add_all_digests()
3224 and SSLeay_add_all_ciphers() were in the same source file so calling
3225 one would link with the other. They are now in separate source files.
3226 [Steve Henson]
3227
3228 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3229 [Steve Henson]
3230
3231 *) Use a less unusual form of the Miller-Rabin primality test (it used
3232 a binary algorithm for exponentiation integrated into the Miller-Rabin
3233 loop, our standard modexp algorithms are faster).
3234 [Bodo Moeller]
3235
3236 *) Support for the EBCDIC character set completed.
3237 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3238
3239 *) Source code cleanups: use const where appropriate, eliminate casts,
3240 use void * instead of char * in lhash.
3241 [Ulf Möller]
3242
3243 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3244 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3245 this the server could overwrite ephemeral keys that the client
3246 has already seen).
3247 [Bodo Moeller]
3248
3249 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3250 using 50 iterations of the Rabin-Miller test.
3251
3252 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3253 iterations of the Rabin-Miller test as required by the appendix
3254 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3255 As BN_is_prime_fasttest includes trial division, DSA parameter
3256 generation becomes much faster.
3257
3258 This implies a change for the callback functions in DSA_is_prime
3259 and DSA_generate_parameters: The callback function is called once
3260 for each positive witness in the Rabin-Miller test, not just
3261 occasionally in the inner loop; and the parameters to the
3262 callback function now provide an iteration count for the outer
3263 loop rather than for the current invocation of the inner loop.
3264 DSA_generate_parameters additionally can call the callback
3265 function with an 'iteration count' of -1, meaning that a
3266 candidate has passed the trial division test (when q is generated
3267 from an application-provided seed, trial division is skipped).
3268 [Bodo Moeller]
3269
3270 *) New function BN_is_prime_fasttest that optionally does trial
3271 division before starting the Rabin-Miller test and has
3272 an additional BN_CTX * argument (whereas BN_is_prime always
3273 has to allocate at least one BN_CTX).
3274 'callback(1, -1, cb_arg)' is called when a number has passed the
3275 trial division stage.
3276 [Bodo Moeller]
3277
3278 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3279 as ASN1_TIME.
3280 [Steve Henson]
3281
3282 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3283 [Steve Henson]
3284
3285 *) New function BN_pseudo_rand().
3286 [Ulf Möller]
3287
3288 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3289 bignum version of BN_from_montgomery() with the working code from
3290 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3291 the comments.
3292 [Ulf Möller]
3293
3294 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3295 made it impossible to use the same SSL_SESSION data structure in
3296 SSL2 clients in multiple threads.
3297 [Bodo Moeller]
3298
3299 *) The return value of RAND_load_file() no longer counts bytes obtained
3300 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3301 to seed the PRNG (previously an explicit byte count was required).
3302 [Ulf Möller, Bodo Möller]
3303
3304 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3305 used (char *) instead of (void *) and had casts all over the place.
3306 [Steve Henson]
3307
3308 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3309 [Ulf Möller]
3310
3311 *) Retain source code compatibility for BN_prime_checks macro:
3312 BN_is_prime(..., BN_prime_checks, ...) now uses
3313 BN_prime_checks_for_size to determine the appropriate number of
3314 Rabin-Miller iterations.
3315 [Ulf Möller]
3316
3317 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3318 DH_CHECK_P_NOT_SAFE_PRIME.
3319 (Check if this is true? OpenPGP calls them "strong".)
3320 [Ulf Möller]
3321
3322 *) Merge the functionality of "dh" and "gendh" programs into a new program
3323 "dhparam". The old programs are retained for now but will handle DH keys
3324 (instead of parameters) in future.
3325 [Steve Henson]
3326
3327 *) Make the ciphers, s_server and s_client programs check the return values
3328 when a new cipher list is set.
3329 [Steve Henson]
3330
3331 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3332 ciphers. Before when the 56bit ciphers were enabled the sorting was
3333 wrong.
3334
3335 The syntax for the cipher sorting has been extended to support sorting by
3336 cipher-strength (using the strength_bits hard coded in the tables).
3337 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3338
3339 Fix a bug in the cipher-command parser: when supplying a cipher command
3340 string with an "undefined" symbol (neither command nor alphanumeric
3341 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3342 an error is flagged.
3343
3344 Due to the strength-sorting extension, the code of the
3345 ssl_create_cipher_list() function was completely rearranged. I hope that
3346 the readability was also increased :-)
3347 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3348
3349 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3350 for the first serial number and places 2 in the serial number file. This
3351 avoids problems when the root CA is created with serial number zero and
3352 the first user certificate has the same issuer name and serial number
3353 as the root CA.
3354 [Steve Henson]
3355
3356 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3357 the new code. Add documentation for this stuff.
3358 [Steve Henson]
3359
3360 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3361 X509_*() to X509at_*() on the grounds that they don't handle X509
3362 structures and behave in an analagous way to the X509v3 functions:
3363 they shouldn't be called directly but wrapper functions should be used
3364 instead.
3365
3366 So we also now have some wrapper functions that call the X509at functions
3367 when passed certificate requests. (TO DO: similar things can be done with
3368 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3369 things. Some of these need some d2i or i2d and print functionality
3370 because they handle more complex structures.)
3371 [Steve Henson]
3372
3373 *) Add missing #ifndefs that caused missing symbols when building libssl
3374 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3375 NO_RSA in ssl/s2*.c.
3376 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3377
3378 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3379 has a return value which indicates the quality of the random data
3380 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3381 error queue. New function RAND_pseudo_bytes() generates output that is
3382 guaranteed to be unique but not unpredictable. RAND_add is like
3383 RAND_seed, but takes an extra argument for an entropy estimate
3384 (RAND_seed always assumes full entropy).
3385 [Ulf Möller]
3386
3387 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3388 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3389 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3390 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3391 false-positive rate of at most 2^-80 for random input.
3392 [Bodo Moeller]
3393
3394 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3395 [Bodo Moeller]
3396
3397 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3398 in the 0.9.5 release), this returns the chain
3399 from an X509_CTX structure with a dup of the stack and all
3400 the X509 reference counts upped: so the stack will exist
3401 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3402 to use this.
3403
3404 Also make SSL_SESSION_print() print out the verify return
3405 code.
3406 [Steve Henson]
3407
3408 *) Add manpage for the pkcs12 command. Also change the default
3409 behaviour so MAC iteration counts are used unless the new
3410 -nomaciter option is used. This improves file security and
3411 only older versions of MSIE (4.0 for example) need it.
3412 [Steve Henson]
3413
3414 *) Honor the no-xxx Configure options when creating .DEF files.
3415 [Ulf Möller]
3416
3417 *) Add PKCS#10 attributes to field table: challengePassword,
3418 unstructuredName and unstructuredAddress. These are taken from
3419 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3420 international characters are used.
3421
3422 More changes to X509_ATTRIBUTE code: allow the setting of types
3423 based on strings. Remove the 'loc' parameter when adding
3424 attributes because these will be a SET OF encoding which is sorted
3425 in ASN1 order.
3426 [Steve Henson]
3427
3428 *) Initial changes to the 'req' utility to allow request generation
3429 automation. This will allow an application to just generate a template
3430 file containing all the field values and have req construct the
3431 request.
3432
3433 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3434 used all over the place including certificate requests and PKCS#7
3435 structures. They are currently handled manually where necessary with
3436 some primitive wrappers for PKCS#7. The new functions behave in a
3437 manner analogous to the X509 extension functions: they allow
3438 attributes to be looked up by NID and added.
3439
3440 Later something similar to the X509V3 code would be desirable to
3441 automatically handle the encoding, decoding and printing of the
3442 more complex types. The string types like challengePassword can
3443 be handled by the string table functions.
3444
3445 Also modified the multi byte string table handling. Now there is
3446 a 'global mask' which masks out certain types. The table itself
3447 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3448 is useful when for example there is only one permissible type
3449 (as in countryName) and using the mask might result in no valid
3450 types at all.
3451 [Steve Henson]
3452
3453 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3454 SSL_get_peer_finished to allow applications to obtain the latest
3455 Finished messages sent to the peer or expected from the peer,
3456 respectively. (SSL_get_peer_finished is usually the Finished message
3457 actually received from the peer, otherwise the protocol will be aborted.)
3458
3459 As the Finished message are message digests of the complete handshake
3460 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3461 be used for external authentication procedures when the authentication
3462 provided by SSL/TLS is not desired or is not enough.
3463 [Bodo Moeller]
3464
3465 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3466 the host supports BWX extension and if Compaq C is present on the
3467 $PATH. Just exploiting of the BWX extension results in 20-30%
3468 performance kick for some algorithms, e.g. DES and RC4 to mention
3469 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3470 SHA1.
3471 [Andy Polyakov]
3472
3473 *) Add support for MS "fast SGC". This is arguably a violation of the
3474 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3475 weak crypto and after checking the certificate is SGC a second one
3476 with strong crypto. MS SGC stops the first handshake after receiving
3477 the server certificate message and sends a second client hello. Since
3478 a server will typically do all the time consuming operations before
3479 expecting any further messages from the client (server key exchange
3480 is the most expensive) there is little difference between the two.
3481
3482 To get OpenSSL to support MS SGC we have to permit a second client
3483 hello message after we have sent server done. In addition we have to
3484 reset the MAC if we do get this second client hello.
3485 [Steve Henson]
3486
3487 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3488 if a DER encoded private key is RSA or DSA traditional format. Changed
3489 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3490 format DER encoded private key. Newer code should use PKCS#8 format which
3491 has the key type encoded in the ASN1 structure. Added DER private key
3492 support to pkcs8 application.
3493 [Steve Henson]
3494
3495 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3496 ciphersuites has been selected (as required by the SSL 3/TLS 1
3497 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3498 is set, we interpret this as a request to violate the specification
3499 (the worst that can happen is a handshake failure, and 'correct'
3500 behaviour would result in a handshake failure anyway).
3501 [Bodo Moeller]
3502
3503 *) In SSL_CTX_add_session, take into account that there might be multiple
3504 SSL_SESSION structures with the same session ID (e.g. when two threads
3505 concurrently obtain them from an external cache).
3506 The internal cache can handle only one SSL_SESSION with a given ID,
3507 so if there's a conflict, we now throw out the old one to achieve
3508 consistency.
3509 [Bodo Moeller]
3510
3511 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3512 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3513 some routines that use cipher OIDs: some ciphers do not have OIDs
3514 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3515 example.
3516 [Steve Henson]
3517
3518 *) Simplify the trust setting structure and code. Now we just have
3519 two sequences of OIDs for trusted and rejected settings. These will
3520 typically have values the same as the extended key usage extension
3521 and any application specific purposes.
3522
3523 The trust checking code now has a default behaviour: it will just
3524 check for an object with the same NID as the passed id. Functions can
3525 be provided to override either the default behaviour or the behaviour
3526 for a given id. SSL client, server and email already have functions
3527 in place for compatibility: they check the NID and also return "trusted"
3528 if the certificate is self signed.
3529 [Steve Henson]
3530
3531 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3532 traditional format into an EVP_PKEY structure.
3533 [Steve Henson]
3534
3535 *) Add a password callback function PEM_cb() which either prompts for
3536 a password if usr_data is NULL or otherwise assumes it is a null
3537 terminated password. Allow passwords to be passed on command line
3538 environment or config files in a few more utilities.
3539 [Steve Henson]
3540
3541 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3542 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3543 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3544 Update documentation.
3545 [Steve Henson]
3546
3547 *) Support for ASN1 "NULL" type. This could be handled before by using
3548 ASN1_TYPE but there wasn't any function that would try to read a NULL
3549 and produce an error if it couldn't. For compatibility we also have
3550 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3551 don't allocate anything because they don't need to.
3552 [Steve Henson]
3553
3554 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3555 for details.
3556 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3557
3558 *) Rebuild of the memory allocation routines used by OpenSSL code and
3559 possibly others as well. The purpose is to make an interface that
3560 provide hooks so anyone can build a separate set of allocation and
3561 deallocation routines to be used by OpenSSL, for example memory
3562 pool implementations, or something else, which was previously hard
3563 since Malloc(), Realloc() and Free() were defined as macros having
3564 the values malloc, realloc and free, respectively (except for Win32
3565 compilations). The same is provided for memory debugging code.
3566 OpenSSL already comes with functionality to find memory leaks, but
3567 this gives people a chance to debug other memory problems.
3568
3569 With these changes, a new set of functions and macros have appeared:
3570
3571 CRYPTO_set_mem_debug_functions() [F]
3572 CRYPTO_get_mem_debug_functions() [F]
3573 CRYPTO_dbg_set_options() [F]
3574 CRYPTO_dbg_get_options() [F]
3575 CRYPTO_malloc_debug_init() [M]
3576
3577 The memory debug functions are NULL by default, unless the library
3578 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3579 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3580 gives the standard debugging functions that come with OpenSSL) or
3581 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3582 provided by the library user) must be used. When the standard
3583 debugging functions are used, CRYPTO_dbg_set_options can be used to
3584 request additional information:
3585 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3586 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3587
3588 Also, things like CRYPTO_set_mem_functions will always give the
3589 expected result (the new set of functions is used for allocation
3590 and deallocation) at all times, regardless of platform and compiler
3591 options.
3592
3593 To finish it up, some functions that were never use in any other
3594 way than through macros have a new API and new semantic:
3595
3596 CRYPTO_dbg_malloc()
3597 CRYPTO_dbg_realloc()
3598 CRYPTO_dbg_free()
3599
3600 All macros of value have retained their old syntax.
3601 [Richard Levitte and Bodo Moeller]
3602
3603 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3604 ordering of SMIMECapabilities wasn't in "strength order" and there
3605 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3606 algorithm.
3607 [Steve Henson]
3608
3609 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3610 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3611 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3612
3613 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3614 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3615 functionality to handle multipart/signed properly) and a utility
3616 called 'smime' to call all this stuff. This is based on code I
3617 originally wrote for Celo who have kindly allowed it to be
3618 included in OpenSSL.
3619 [Steve Henson]
3620
3621 *) Add variants des_set_key_checked and des_set_key_unchecked of
3622 des_set_key (aka des_key_sched). Global variable des_check_key
3623 decides which of these is called by des_set_key; this way
3624 des_check_key behaves as it always did, but applications and
3625 the library itself, which was buggy for des_check_key == 1,
3626 have a cleaner way to pick the version they need.
3627 [Bodo Moeller]
3628
3629 *) New function PKCS12_newpass() which changes the password of a
3630 PKCS12 structure.
3631 [Steve Henson]
3632
3633 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3634 dynamic mix. In both cases the ids can be used as an index into the
3635 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3636 functions so they accept a list of the field values and the
3637 application doesn't need to directly manipulate the X509_TRUST
3638 structure.
3639 [Steve Henson]
3640
3641 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3642 need initialising.
3643 [Steve Henson]
3644
3645 *) Modify the way the V3 extension code looks up extensions. This now
3646 works in a similar way to the object code: we have some "standard"
3647 extensions in a static table which is searched with OBJ_bsearch()
3648 and the application can add dynamic ones if needed. The file
3649 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3650 updated whenever a new extension is added to the core code and kept
3651 in ext_nid order. There is a simple program 'tabtest.c' which checks
3652 this. New extensions are not added too often so this file can readily
3653 be maintained manually.
3654
3655 There are two big advantages in doing things this way. The extensions
3656 can be looked up immediately and no longer need to be "added" using
3657 X509V3_add_standard_extensions(): this function now does nothing.
3658 [Side note: I get *lots* of email saying the extension code doesn't
3659 work because people forget to call this function]
3660 Also no dynamic allocation is done unless new extensions are added:
3661 so if we don't add custom extensions there is no need to call
3662 X509V3_EXT_cleanup().
3663 [Steve Henson]
3664
3665 *) Modify enc utility's salting as follows: make salting the default. Add a
3666 magic header, so unsalted files fail gracefully instead of just decrypting
3667 to garbage. This is because not salting is a big security hole, so people
3668 should be discouraged from doing it.
3669 [Ben Laurie]
3670
3671 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3672 digest to be passed on the command line but it only used this
3673 parameter when signing a certificate. Modified so all relevant
3674 operations are affected by the digest parameter including the
3675 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3676 DSA key was used because it didn't fix the digest.
3677 [Steve Henson]
3678
3679 *) Initial certificate chain verify code. Currently tests the untrusted
3680 certificates for consistency with the verify purpose (which is set
3681 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3682
3683 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3684 this is because it will reject chains with invalid extensions whereas
3685 every previous version of OpenSSL and SSLeay made no checks at all.
3686
3687 Trust code: checks the root CA for the relevant trust settings. Trust
3688 settings have an initial value consistent with the verify purpose: e.g.
3689 if the verify purpose is for SSL client use it expects the CA to be
3690 trusted for SSL client use. However the default value can be changed to
3691 permit custom trust settings: one example of this would be to only trust
3692 certificates from a specific "secure" set of CAs.
3693
3694 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3695 which should be used for version portability: especially since the
3696 verify structure is likely to change more often now.
3697
3698 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3699 to set them. If not set then assume SSL clients will verify SSL servers
3700 and vice versa.
3701
3702 Two new options to the verify program: -untrusted allows a set of
3703 untrusted certificates to be passed in and -purpose which sets the
3704 intended purpose of the certificate. If a purpose is set then the
3705 new chain verify code is used to check extension consistency.
3706 [Steve Henson]
3707
3708 *) Support for the authority information access extension.
3709 [Steve Henson]
3710
3711 *) Modify RSA and DSA PEM read routines to transparently handle
3712 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3713 public keys in a format compatible with certificate
3714 SubjectPublicKeyInfo structures. Unfortunately there were already
3715 functions called *_PublicKey_* which used various odd formats so
3716 these are retained for compatibility: however the DSA variants were
3717 never in a public release so they have been deleted. Changed dsa/rsa
3718 utilities to handle the new format: note no releases ever handled public
3719 keys so we should be OK.
3720
3721 The primary motivation for this change is to avoid the same fiasco
3722 that dogs private keys: there are several incompatible private key
3723 formats some of which are standard and some OpenSSL specific and
3724 require various evil hacks to allow partial transparent handling and
3725 even then it doesn't work with DER formats. Given the option anything
3726 other than PKCS#8 should be dumped: but the other formats have to
3727 stay in the name of compatibility.
3728
3729 With public keys and the benefit of hindsight one standard format
3730 is used which works with EVP_PKEY, RSA or DSA structures: though
3731 it clearly returns an error if you try to read the wrong kind of key.
3732
3733 Added a -pubkey option to the 'x509' utility to output the public key.
3734 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3735 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3736 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3737 that do the same as the EVP_PKEY_assign_*() except they up the
3738 reference count of the added key (they don't "swallow" the
3739 supplied key).
3740 [Steve Henson]
3741
3742 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3743 CRLs would fail if the file contained no certificates or no CRLs:
3744 added a new function to read in both types and return the number
3745 read: this means that if none are read it will be an error. The
3746 DER versions of the certificate and CRL reader would always fail
3747 because it isn't possible to mix certificates and CRLs in DER format
3748 without choking one or the other routine. Changed this to just read
3749 a certificate: this is the best we can do. Also modified the code
3750 in apps/verify.c to take notice of return codes: it was previously
3751 attempting to read in certificates from NULL pointers and ignoring
3752 any errors: this is one reason why the cert and CRL reader seemed
3753 to work. It doesn't check return codes from the default certificate
3754 routines: these may well fail if the certificates aren't installed.
3755 [Steve Henson]
3756
3757 *) Code to support otherName option in GeneralName.
3758 [Steve Henson]
3759
3760 *) First update to verify code. Change the verify utility
3761 so it warns if it is passed a self signed certificate:
3762 for consistency with the normal behaviour. X509_verify
3763 has been modified to it will now verify a self signed
3764 certificate if *exactly* the same certificate appears
3765 in the store: it was previously impossible to trust a
3766 single self signed certificate. This means that:
3767 openssl verify ss.pem
3768 now gives a warning about a self signed certificate but
3769 openssl verify -CAfile ss.pem ss.pem
3770 is OK.
3771 [Steve Henson]
3772
3773 *) For servers, store verify_result in SSL_SESSION data structure
3774 (and add it to external session representation).
3775 This is needed when client certificate verifications fails,
3776 but an application-provided verification callback (set by
3777 SSL_CTX_set_cert_verify_callback) allows accepting the session
3778 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3779 but returns 1): When the session is reused, we have to set
3780 ssl->verify_result to the appropriate error code to avoid
3781 security holes.
3782 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3783
3784 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3785 case in PKCS7_dataInit() where the signed PKCS7 structure
3786 didn't contain any existing data because it was being created.
3787 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3788
3789 *) Add a salt to the key derivation routines in enc.c. This
3790 forms the first 8 bytes of the encrypted file. Also add a
3791 -S option to allow a salt to be input on the command line.
3792 [Steve Henson]
3793
3794 *) New function X509_cmp(). Oddly enough there wasn't a function
3795 to compare two certificates. We do this by working out the SHA1
3796 hash and comparing that. X509_cmp() will be needed by the trust
3797 code.
3798 [Steve Henson]
3799
3800 *) SSL_get1_session() is like SSL_get_session(), but increments
3801 the reference count in the SSL_SESSION returned.
3802 [Geoff Thorpe <geoff@eu.c2.net>]
3803
3804 *) Fix for 'req': it was adding a null to request attributes.
3805 Also change the X509_LOOKUP and X509_INFO code to handle
3806 certificate auxiliary information.
3807 [Steve Henson]
3808
3809 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3810 the 'enc' command.
3811 [Steve Henson]
3812
3813 *) Add the possibility to add extra information to the memory leak
3814 detecting output, to form tracebacks, showing from where each
3815 allocation was originated: CRYPTO_push_info("constant string") adds
3816 the string plus current file name and line number to a per-thread
3817 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3818 is like calling CYRPTO_pop_info() until the stack is empty.
3819 Also updated memory leak detection code to be multi-thread-safe.
3820 [Richard Levitte]
3821
3822 *) Add options -text and -noout to pkcs7 utility and delete the
3823 encryption options which never did anything. Update docs.
3824 [Steve Henson]
3825
3826 *) Add options to some of the utilities to allow the pass phrase
3827 to be included on either the command line (not recommended on
3828 OSes like Unix) or read from the environment. Update the
3829 manpages and fix a few bugs.
3830 [Steve Henson]
3831
3832 *) Add a few manpages for some of the openssl commands.
3833 [Steve Henson]
3834
3835 *) Fix the -revoke option in ca. It was freeing up memory twice,
3836 leaking and not finding already revoked certificates.
3837 [Steve Henson]
3838
3839 *) Extensive changes to support certificate auxiliary information.
3840 This involves the use of X509_CERT_AUX structure and X509_AUX
3841 functions. An X509_AUX function such as PEM_read_X509_AUX()
3842 can still read in a certificate file in the usual way but it
3843 will also read in any additional "auxiliary information". By
3844 doing things this way a fair degree of compatibility can be
3845 retained: existing certificates can have this information added
3846 using the new 'x509' options.
3847
3848 Current auxiliary information includes an "alias" and some trust
3849 settings. The trust settings will ultimately be used in enhanced
3850 certificate chain verification routines: currently a certificate
3851 can only be trusted if it is self signed and then it is trusted
3852 for all purposes.
3853 [Steve Henson]
3854
3855 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3856 The problem was that one of the replacement routines had not been working
3857 since SSLeay releases. For now the offending routine has been replaced
3858 with non-optimised assembler. Even so, this now gives around 95%
3859 performance improvement for 1024 bit RSA signs.
3860 [Mark Cox]
3861
3862 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3863 handling. Most clients have the effective key size in bits equal to
3864 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3865 A few however don't do this and instead use the size of the decrypted key
3866 to determine the RC2 key length and the AlgorithmIdentifier to determine
3867 the effective key length. In this case the effective key length can still
3868 be 40 bits but the key length can be 168 bits for example. This is fixed
3869 by manually forcing an RC2 key into the EVP_PKEY structure because the
3870 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3871 the key length and effective key length are equal.
3872 [Steve Henson]
3873
3874 *) Add a bunch of functions that should simplify the creation of
3875 X509_NAME structures. Now you should be able to do:
3876 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3877 and have it automatically work out the correct field type and fill in
3878 the structures. The more adventurous can try:
3879 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3880 and it will (hopefully) work out the correct multibyte encoding.
3881 [Steve Henson]
3882
3883 *) Change the 'req' utility to use the new field handling and multibyte
3884 copy routines. Before the DN field creation was handled in an ad hoc
3885 way in req, ca, and x509 which was rather broken and didn't support
3886 BMPStrings or UTF8Strings. Since some software doesn't implement
3887 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3888 using the dirstring_type option. See the new comment in the default
3889 openssl.cnf for more info.
3890 [Steve Henson]
3891
3892 *) Make crypto/rand/md_rand.c more robust:
3893 - Assure unique random numbers after fork().
3894 - Make sure that concurrent threads access the global counter and
3895 md serializably so that we never lose entropy in them
3896 or use exactly the same state in multiple threads.
3897 Access to the large state is not always serializable because
3898 the additional locking could be a performance killer, and
3899 md should be large enough anyway.
3900 [Bodo Moeller]
3901
3902 *) New file apps/app_rand.c with commonly needed functionality
3903 for handling the random seed file.
3904
3905 Use the random seed file in some applications that previously did not:
3906 ca,
3907 dsaparam -genkey (which also ignored its '-rand' option),
3908 s_client,
3909 s_server,
3910 x509 (when signing).
3911 Except on systems with /dev/urandom, it is crucial to have a random
3912 seed file at least for key creation, DSA signing, and for DH exchanges;
3913 for RSA signatures we could do without one.
3914
3915 gendh and gendsa (unlike genrsa) used to read only the first byte
3916 of each file listed in the '-rand' option. The function as previously
3917 found in genrsa is now in app_rand.c and is used by all programs
3918 that support '-rand'.
3919 [Bodo Moeller]
3920
3921 *) In RAND_write_file, use mode 0600 for creating files;
3922 don't just chmod when it may be too late.
3923 [Bodo Moeller]
3924
3925 *) Report an error from X509_STORE_load_locations
3926 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3927 [Bill Perry]
3928
3929 *) New function ASN1_mbstring_copy() this copies a string in either
3930 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3931 into an ASN1_STRING type. A mask of permissible types is passed
3932 and it chooses the "minimal" type to use or an error if not type
3933 is suitable.
3934 [Steve Henson]
3935
3936 *) Add function equivalents to the various macros in asn1.h. The old
3937 macros are retained with an M_ prefix. Code inside the library can
3938 use the M_ macros. External code (including the openssl utility)
3939 should *NOT* in order to be "shared library friendly".
3940 [Steve Henson]
3941
3942 *) Add various functions that can check a certificate's extensions
3943 to see if it usable for various purposes such as SSL client,
3944 server or S/MIME and CAs of these types. This is currently
3945 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3946 verification. Also added a -purpose flag to x509 utility to
3947 print out all the purposes.
3948 [Steve Henson]
3949
3950 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3951 functions.
3952 [Steve Henson]
3953
3954 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3955 for, obtain and decode and extension and obtain its critical flag.
3956 This allows all the necessary extension code to be handled in a
3957 single function call.
3958 [Steve Henson]
3959
3960 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3961 platforms. See crypto/rc4/rc4_enc.c for further details.
3962 [Andy Polyakov]
3963
3964 *) New -noout option to asn1parse. This causes no output to be produced
3965 its main use is when combined with -strparse and -out to extract data
3966 from a file (which may not be in ASN.1 format).
3967 [Steve Henson]
3968
3969 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3970 when producing the local key id.
3971 [Richard Levitte <levitte@stacken.kth.se>]
3972
3973 *) New option -dhparam in s_server. This allows a DH parameter file to be
3974 stated explicitly. If it is not stated then it tries the first server
3975 certificate file. The previous behaviour hard coded the filename
3976 "server.pem".
3977 [Steve Henson]
3978
3979 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3980 a public key to be input or output. For example:
3981 openssl rsa -in key.pem -pubout -out pubkey.pem
3982 Also added necessary DSA public key functions to handle this.
3983 [Steve Henson]
3984
3985 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
3986 in the message. This was handled by allowing
3987 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
3988 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
3989
3990 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
3991 to the end of the strings whereas this didn't. This would cause problems
3992 if strings read with d2i_ASN1_bytes() were later modified.
3993 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
3994
3995 *) Fix for base64 decode bug. When a base64 bio reads only one line of
3996 data and it contains EOF it will end up returning an error. This is
3997 caused by input 46 bytes long. The cause is due to the way base64
3998 BIOs find the start of base64 encoded data. They do this by trying a
3999 trial decode on each line until they find one that works. When they
4000 do a flag is set and it starts again knowing it can pass all the
4001 data directly through the decoder. Unfortunately it doesn't reset
4002 the context it uses. This means that if EOF is reached an attempt
4003 is made to pass two EOFs through the context and this causes the
4004 resulting error. This can also cause other problems as well. As is
4005 usual with these problems it takes *ages* to find and the fix is
4006 trivial: move one line.
4007 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4008
4009 *) Ugly workaround to get s_client and s_server working under Windows. The
4010 old code wouldn't work because it needed to select() on sockets and the
4011 tty (for keypresses and to see if data could be written). Win32 only
4012 supports select() on sockets so we select() with a 1s timeout on the
4013 sockets and then see if any characters are waiting to be read, if none
4014 are present then we retry, we also assume we can always write data to
4015 the tty. This isn't nice because the code then blocks until we've
4016 received a complete line of data and it is effectively polling the
4017 keyboard at 1s intervals: however it's quite a bit better than not
4018 working at all :-) A dedicated Windows application might handle this
4019 with an event loop for example.
4020 [Steve Henson]
4021
4022 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4023 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4024 will be called when RSA_sign() and RSA_verify() are used. This is useful
4025 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4026 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4027 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4028 This necessitated the support of an extra signature type NID_md5_sha1
4029 for SSL signatures and modifications to the SSL library to use it instead
4030 of calling RSA_public_decrypt() and RSA_private_encrypt().
4031 [Steve Henson]
4032
4033 *) Add new -verify -CAfile and -CApath options to the crl program, these
4034 will lookup a CRL issuers certificate and verify the signature in a
4035 similar way to the verify program. Tidy up the crl program so it
4036 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4037 less strict. It will now permit CRL extensions even if it is not
4038 a V2 CRL: this will allow it to tolerate some broken CRLs.
4039 [Steve Henson]
4040
4041 *) Initialize all non-automatic variables each time one of the openssl
4042 sub-programs is started (this is necessary as they may be started
4043 multiple times from the "OpenSSL>" prompt).
4044 [Lennart Bang, Bodo Moeller]
4045
4046 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4047 removing all other RSA functionality (this is what NO_RSA does). This
4048 is so (for example) those in the US can disable those operations covered
4049 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4050 key generation.
4051 [Steve Henson]
4052
4053 *) Non-copying interface to BIO pairs.
4054 (still largely untested)
4055 [Bodo Moeller]
4056
4057 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4058 ASCII string. This was handled independently in various places before.
4059 [Steve Henson]
4060
4061 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4062 UTF8 strings a character at a time.
4063 [Steve Henson]
4064
4065 *) Use client_version from client hello to select the protocol
4066 (s23_srvr.c) and for RSA client key exchange verification
4067 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4068 [Bodo Moeller]
4069
4070 *) Add various utility functions to handle SPKACs, these were previously
4071 handled by poking round in the structure internals. Added new function
4072 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4073 print, verify and generate SPKACs. Based on an original idea from
4074 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4075 [Steve Henson]
4076
4077 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4078 [Andy Polyakov]
4079
4080 *) Allow the config file extension section to be overwritten on the
4081 command line. Based on an original idea from Massimiliano Pala
4082 <madwolf@comune.modena.it>. The new option is called -extensions
4083 and can be applied to ca, req and x509. Also -reqexts to override
4084 the request extensions in req and -crlexts to override the crl extensions
4085 in ca.
4086 [Steve Henson]
4087
4088 *) Add new feature to the SPKAC handling in ca. Now you can include
4089 the same field multiple times by preceding it by "XXXX." for example:
4090 1.OU="Unit name 1"
4091 2.OU="Unit name 2"
4092 this is the same syntax as used in the req config file.
4093 [Steve Henson]
4094
4095 *) Allow certificate extensions to be added to certificate requests. These
4096 are specified in a 'req_extensions' option of the req section of the
4097 config file. They can be printed out with the -text option to req but
4098 are otherwise ignored at present.
4099 [Steve Henson]
4100
4101 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4102 data read consists of only the final block it would not decrypted because
4103 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4104 A misplaced 'break' also meant the decrypted final block might not be
4105 copied until the next read.
4106 [Steve Henson]
4107
4108 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4109 a few extra parameters to the DH structure: these will be useful if
4110 for example we want the value of 'q' or implement X9.42 DH.
4111 [Steve Henson]
4112
4113 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4114 provides hooks that allow the default DSA functions or functions on a
4115 "per key" basis to be replaced. This allows hardware acceleration and
4116 hardware key storage to be handled without major modification to the
4117 library. Also added low level modexp hooks and CRYPTO_EX structure and
4118 associated functions.
4119 [Steve Henson]
4120
4121 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4122 as "read only": it can't be written to and the buffer it points to will
4123 not be freed. Reading from a read only BIO is much more efficient than
4124 a normal memory BIO. This was added because there are several times when
4125 an area of memory needs to be read from a BIO. The previous method was
4126 to create a memory BIO and write the data to it, this results in two
4127 copies of the data and an O(n^2) reading algorithm. There is a new
4128 function BIO_new_mem_buf() which creates a read only memory BIO from
4129 an area of memory. Also modified the PKCS#7 routines to use read only
4130 memory BIOs.
4131 [Steve Henson]
4132
4133 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4134 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4135 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4136 but a retry condition occured while trying to read the rest.
4137 [Bodo Moeller]
4138
4139 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4140 NID_pkcs7_encrypted by default: this was wrong since this should almost
4141 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4142 the encrypted data type: this is a more sensible place to put it and it
4143 allows the PKCS#12 code to be tidied up that duplicated this
4144 functionality.
4145 [Steve Henson]
4146
4147 *) Changed obj_dat.pl script so it takes its input and output files on
4148 the command line. This should avoid shell escape redirection problems
4149 under Win32.
4150 [Steve Henson]
4151
4152 *) Initial support for certificate extension requests, these are included
4153 in things like Xenroll certificate requests. Included functions to allow
4154 extensions to be obtained and added.
4155 [Steve Henson]
4156
4157 *) -crlf option to s_client and s_server for sending newlines as
4158 CRLF (as required by many protocols).
4159 [Bodo Moeller]
4160
4161 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4162
4163 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4164 [Ralf S. Engelschall]
4165
4166 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4167 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4168
4169 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4170 program.
4171 [Steve Henson]
4172
4173 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4174 DH parameters/keys (q is lost during that conversion, but the resulting
4175 DH parameters contain its length).
4176
4177 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4178 much faster than DH_generate_parameters (which creates parameters
4179 where p = 2*q + 1), and also the smaller q makes DH computations
4180 much more efficient (160-bit exponentiation instead of 1024-bit
4181 exponentiation); so this provides a convenient way to support DHE
4182 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4183 utter importance to use
4184 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4185 or
4186 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4187 when such DH parameters are used, because otherwise small subgroup
4188 attacks may become possible!
4189 [Bodo Moeller]
4190
4191 *) Avoid memory leak in i2d_DHparams.
4192 [Bodo Moeller]
4193
4194 *) Allow the -k option to be used more than once in the enc program:
4195 this allows the same encrypted message to be read by multiple recipients.
4196 [Steve Henson]
4197
4198 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4199 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4200 it will always use the numerical form of the OID, even if it has a short
4201 or long name.
4202 [Steve Henson]
4203
4204 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4205 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4206 otherwise bn_mod_exp was called. In the case of hardware keys for example
4207 no private key components need be present and it might store extra data
4208 in the RSA structure, which cannot be accessed from bn_mod_exp.
4209 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4210 private key operations.
4211 [Steve Henson]
4212
4213 *) Added support for SPARC Linux.
4214 [Andy Polyakov]
4215
4216 *) pem_password_cb function type incompatibly changed from
4217 typedef int pem_password_cb(char *buf, int size, int rwflag);
4218 to
4219 ....(char *buf, int size, int rwflag, void *userdata);
4220 so that applications can pass data to their callbacks:
4221 The PEM[_ASN1]_{read,write}... functions and macros now take an
4222 additional void * argument, which is just handed through whenever
4223 the password callback is called.
4224 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4225
4226 New function SSL_CTX_set_default_passwd_cb_userdata.
4227
4228 Compatibility note: As many C implementations push function arguments
4229 onto the stack in reverse order, the new library version is likely to
4230 interoperate with programs that have been compiled with the old
4231 pem_password_cb definition (PEM_whatever takes some data that
4232 happens to be on the stack as its last argument, and the callback
4233 just ignores this garbage); but there is no guarantee whatsoever that
4234 this will work.
4235
4236 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4237 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4238 problems not only on Windows, but also on some Unix platforms.
4239 To avoid problematic command lines, these definitions are now in an
4240 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4241 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4242 [Bodo Moeller]
4243
4244 *) MIPS III/IV assembler module is reimplemented.
4245 [Andy Polyakov]
4246
4247 *) More DES library cleanups: remove references to srand/rand and
4248 delete an unused file.
4249 [Ulf Möller]
4250
4251 *) Add support for the the free Netwide assembler (NASM) under Win32,
4252 since not many people have MASM (ml) and it can be hard to obtain.
4253 This is currently experimental but it seems to work OK and pass all
4254 the tests. Check out INSTALL.W32 for info.
4255 [Steve Henson]
4256
4257 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4258 without temporary keys kept an extra copy of the server key,
4259 and connections with temporary keys did not free everything in case
4260 of an error.
4261 [Bodo Moeller]
4262
4263 *) New function RSA_check_key and new openssl rsa option -check
4264 for verifying the consistency of RSA keys.
4265 [Ulf Moeller, Bodo Moeller]
4266
4267 *) Various changes to make Win32 compile work:
4268 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4269 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4270 comparison" warnings.
4271 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4272 [Steve Henson]
4273
4274 *) Add a debugging option to PKCS#5 v2 key generation function: when
4275 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4276 derived keys are printed to stderr.
4277 [Steve Henson]
4278
4279 *) Copy the flags in ASN1_STRING_dup().
4280 [Roman E. Pavlov <pre@mo.msk.ru>]
4281
4282 *) The x509 application mishandled signing requests containing DSA
4283 keys when the signing key was also DSA and the parameters didn't match.
4284
4285 It was supposed to omit the parameters when they matched the signing key:
4286 the verifying software was then supposed to automatically use the CA's
4287 parameters if they were absent from the end user certificate.
4288
4289 Omitting parameters is no longer recommended. The test was also
4290 the wrong way round! This was probably due to unusual behaviour in
4291 EVP_cmp_parameters() which returns 1 if the parameters match.
4292 This meant that parameters were omitted when they *didn't* match and
4293 the certificate was useless. Certificates signed with 'ca' didn't have
4294 this bug.
4295 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4296
4297 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4298 The interface is as follows:
4299 Applications can use
4300 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4301 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4302 "off" is now the default.
4303 The library internally uses
4304 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4305 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4306 to disable memory-checking temporarily.
4307
4308 Some inconsistent states that previously were possible (and were
4309 even the default) are now avoided.
4310
4311 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4312 with each memory chunk allocated; this is occasionally more helpful
4313 than just having a counter.
4314
4315 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4316
4317 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4318 extensions.
4319 [Bodo Moeller]
4320
4321 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4322 which largely parallels "options", but is for changing API behaviour,
4323 whereas "options" are about protocol behaviour.
4324 Initial "mode" flags are:
4325
4326 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4327 a single record has been written.
4328 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4329 retries use the same buffer location.
4330 (But all of the contents must be
4331 copied!)
4332 [Bodo Moeller]
4333
4334 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4335 worked.
4336
4337 *) Fix problems with no-hmac etc.
4338 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4339
4340 *) New functions RSA_get_default_method(), RSA_set_method() and
4341 RSA_get_method(). These allows replacement of RSA_METHODs without having
4342 to mess around with the internals of an RSA structure.
4343 [Steve Henson]
4344
4345 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4346 Also really enable memory leak checks in openssl.c and in some
4347 test programs.
4348 [Chad C. Mulligan, Bodo Moeller]
4349
4350 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4351 up the length of negative integers. This has now been simplified to just
4352 store the length when it is first determined and use it later, rather
4353 than trying to keep track of where data is copied and updating it to
4354 point to the end.
4355 [Steve Henson, reported by Brien Wheeler
4356 <bwheeler@authentica-security.com>]
4357
4358 *) Add a new function PKCS7_signatureVerify. This allows the verification
4359 of a PKCS#7 signature but with the signing certificate passed to the
4360 function itself. This contrasts with PKCS7_dataVerify which assumes the
4361 certificate is present in the PKCS#7 structure. This isn't always the
4362 case: certificates can be omitted from a PKCS#7 structure and be
4363 distributed by "out of band" means (such as a certificate database).
4364 [Steve Henson]
4365
4366 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4367 function prototypes in pem.h, also change util/mkdef.pl to add the
4368 necessary function names.
4369 [Steve Henson]
4370
4371 *) mk1mf.pl (used by Windows builds) did not properly read the
4372 options set by Configure in the top level Makefile, and Configure
4373 was not even able to write more than one option correctly.
4374 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4375 [Bodo Moeller]
4376
4377 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4378 file to be loaded from a BIO or FILE pointer. The BIO version will
4379 for example allow memory BIOs to contain config info.
4380 [Steve Henson]
4381
4382 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4383 Whoever hopes to achieve shared-library compatibility across versions
4384 must use this, not the compile-time macro.
4385 (Exercise 0.9.4: Which is the minimum library version required by
4386 such programs?)
4387 Note: All this applies only to multi-threaded programs, others don't
4388 need locks.
4389 [Bodo Moeller]
4390
4391 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4392 through a BIO pair triggered the default case, i.e.
4393 SSLerr(...,SSL_R_UNKNOWN_STATE).
4394 [Bodo Moeller]
4395
4396 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4397 can use the SSL library even if none of the specific BIOs is
4398 appropriate.
4399 [Bodo Moeller]
4400
4401 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4402 for the encoded length.
4403 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4404
4405 *) Add initial documentation of the X509V3 functions.
4406 [Steve Henson]
4407
4408 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4409 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4410 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4411 secure PKCS#8 private key format with a high iteration count.
4412 [Steve Henson]
4413
4414 *) Fix determination of Perl interpreter: A perl or perl5
4415 _directory_ in $PATH was also accepted as the interpreter.
4416 [Ralf S. Engelschall]
4417
4418 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4419 wrong with it but it was very old and did things like calling
4420 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4421 unusual formatting.
4422 [Steve Henson]
4423
4424 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4425 to use the new extension code.
4426 [Steve Henson]
4427
4428 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4429 with macros. This should make it easier to change their form, add extra
4430 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4431 constant.
4432 [Steve Henson]
4433
4434 *) Add to configuration table a new entry that can specify an alternative
4435 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4436 according to Mark Crispin <MRC@Panda.COM>.
4437 [Bodo Moeller]
4438
4439 #if 0
4440 *) DES CBC did not update the IV. Weird.
4441 [Ben Laurie]
4442 #else
4443 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4444 Changing the behaviour of the former might break existing programs --
4445 where IV updating is needed, des_ncbc_encrypt can be used.
4446 #endif
4447
4448 *) When bntest is run from "make test" it drives bc to check its
4449 calculations, as well as internally checking them. If an internal check
4450 fails, it needs to cause bc to give a non-zero result or make test carries
4451 on without noticing the failure. Fixed.
4452 [Ben Laurie]
4453
4454 *) DES library cleanups.
4455 [Ulf Möller]
4456
4457 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4458 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4459 ciphers. NOTE: although the key derivation function has been verified
4460 against some published test vectors it has not been extensively tested
4461 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4462 of v2.0.
4463 [Steve Henson]
4464
4465 *) Instead of "mkdir -p", which is not fully portable, use new
4466 Perl script "util/mkdir-p.pl".
4467 [Bodo Moeller]
4468
4469 *) Rewrite the way password based encryption (PBE) is handled. It used to
4470 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4471 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4472 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4473 the 'parameter' field of the AlgorithmIdentifier is passed to the
4474 underlying key generation function so it must do its own ASN1 parsing.
4475 This has also changed the EVP_PBE_CipherInit() function which now has a
4476 'parameter' argument instead of literal salt and iteration count values
4477 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4478 [Steve Henson]
4479
4480 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4481 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4482 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4483 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4484 value was just used as a "magic string" and not used directly its
4485 value doesn't matter.
4486 [Steve Henson]
4487
4488 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4489 support mutable.
4490 [Ben Laurie]
4491
4492 *) "linux-sparc64" configuration (ultrapenguin).
4493 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4494 "linux-sparc" configuration.
4495 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4496
4497 *) config now generates no-xxx options for missing ciphers.
4498 [Ulf Möller]
4499
4500 *) Support the EBCDIC character set (work in progress).
4501 File ebcdic.c not yet included because it has a different license.
4502 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4503
4504 *) Support BS2000/OSD-POSIX.
4505 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4506
4507 *) Make callbacks for key generation use void * instead of char *.
4508 [Ben Laurie]
4509
4510 *) Make S/MIME samples compile (not yet tested).
4511 [Ben Laurie]
4512
4513 *) Additional typesafe stacks.
4514 [Ben Laurie]
4515
4516 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4517 [Bodo Moeller]
4518
4519
4520 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4521
4522 *) New configuration variant "sco5-gcc".
4523
4524 *) Updated some demos.
4525 [Sean O Riordain, Wade Scholine]
4526
4527 *) Add missing BIO_free at exit of pkcs12 application.
4528 [Wu Zhigang]
4529
4530 *) Fix memory leak in conf.c.
4531 [Steve Henson]
4532
4533 *) Updates for Win32 to assembler version of MD5.
4534 [Steve Henson]
4535
4536 *) Set #! path to perl in apps/der_chop to where we found it
4537 instead of using a fixed path.
4538 [Bodo Moeller]
4539
4540 *) SHA library changes for irix64-mips4-cc.
4541 [Andy Polyakov]
4542
4543 *) Improvements for VMS support.
4544 [Richard Levitte]
4545
4546
4547 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4548
4549 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4550 This also avoids the problems with SC4.2 and unpatched SC5.
4551 [Andy Polyakov <appro@fy.chalmers.se>]
4552
4553 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4554 These are required because of the typesafe stack would otherwise break
4555 existing code. If old code used a structure member which used to be STACK
4556 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4557 sk_num or sk_value it would produce an error because the num, data members
4558 are not present in STACK_OF. Now it just produces a warning. sk_set
4559 replaces the old method of assigning a value to sk_value
4560 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4561 that does this will no longer work (and should use sk_set instead) but
4562 this could be regarded as a "questionable" behaviour anyway.
4563 [Steve Henson]
4564
4565 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4566 correctly handle encrypted S/MIME data.
4567 [Steve Henson]
4568
4569 *) Change type of various DES function arguments from des_cblock
4570 (which means, in function argument declarations, pointer to char)
4571 to des_cblock * (meaning pointer to array with 8 char elements),
4572 which allows the compiler to do more typechecking; it was like
4573 that back in SSLeay, but with lots of ugly casts.
4574
4575 Introduce new type const_des_cblock.
4576 [Bodo Moeller]
4577
4578 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4579 problems: find RecipientInfo structure that matches recipient certificate
4580 and initialise the ASN1 structures properly based on passed cipher.
4581 [Steve Henson]
4582
4583 *) Belatedly make the BN tests actually check the results.
4584 [Ben Laurie]
4585
4586 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4587 to and from BNs: it was completely broken. New compilation option
4588 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4589 key elements as negative integers.
4590 [Steve Henson]
4591
4592 *) Reorganize and speed up MD5.
4593 [Andy Polyakov <appro@fy.chalmers.se>]
4594
4595 *) VMS support.
4596 [Richard Levitte <richard@levitte.org>]
4597
4598 *) New option -out to asn1parse to allow the parsed structure to be
4599 output to a file. This is most useful when combined with the -strparse
4600 option to examine the output of things like OCTET STRINGS.
4601 [Steve Henson]
4602
4603 *) Make SSL library a little more fool-proof by not requiring any longer
4604 that SSL_set_{accept,connect}_state be called before
4605 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4606 in many applications because usually everything *appeared* to work as
4607 intended anyway -- now it really works as intended).
4608 [Bodo Moeller]
4609
4610 *) Move openssl.cnf out of lib/.
4611 [Ulf Möller]
4612
4613 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4614 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4615 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4616 [Ralf S. Engelschall]
4617
4618 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4619 handle PKCS#7 enveloped data properly.
4620 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4621
4622 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4623 copying pointers. The cert_st handling is changed by this in
4624 various ways (and thus what used to be known as ctx->default_cert
4625 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4626 any longer when s->cert does not give us what we need).
4627 ssl_cert_instantiate becomes obsolete by this change.
4628 As soon as we've got the new code right (possibly it already is?),
4629 we have solved a couple of bugs of the earlier code where s->cert
4630 was used as if it could not have been shared with other SSL structures.
4631
4632 Note that using the SSL API in certain dirty ways now will result
4633 in different behaviour than observed with earlier library versions:
4634 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4635 does not influence s as it used to.
4636
4637 In order to clean up things more thoroughly, inside SSL_SESSION
4638 we don't use CERT any longer, but a new structure SESS_CERT
4639 that holds per-session data (if available); currently, this is
4640 the peer's certificate chain and, for clients, the server's certificate
4641 and temporary key. CERT holds only those values that can have
4642 meaningful defaults in an SSL_CTX.
4643 [Bodo Moeller]
4644
4645 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4646 from the internal representation. Various PKCS#7 fixes: remove some
4647 evil casts and set the enc_dig_alg field properly based on the signing
4648 key type.
4649 [Steve Henson]
4650
4651 *) Allow PKCS#12 password to be set from the command line or the
4652 environment. Let 'ca' get its config file name from the environment
4653 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4654 and 'x509').
4655 [Steve Henson]
4656
4657 *) Allow certificate policies extension to use an IA5STRING for the
4658 organization field. This is contrary to the PKIX definition but
4659 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4660 extension option.
4661 [Steve Henson]
4662
4663 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4664 without disallowing inline assembler and the like for non-pedantic builds.
4665 [Ben Laurie]
4666
4667 *) Support Borland C++ builder.
4668 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4669
4670 *) Support Mingw32.
4671 [Ulf Möller]
4672
4673 *) SHA-1 cleanups and performance enhancements.
4674 [Andy Polyakov <appro@fy.chalmers.se>]
4675
4676 *) Sparc v8plus assembler for the bignum library.
4677 [Andy Polyakov <appro@fy.chalmers.se>]
4678
4679 *) Accept any -xxx and +xxx compiler options in Configure.
4680 [Ulf Möller]
4681
4682 *) Update HPUX configuration.
4683 [Anonymous]
4684
4685 *) Add missing sk_<type>_unshift() function to safestack.h
4686 [Ralf S. Engelschall]
4687
4688 *) New function SSL_CTX_use_certificate_chain_file that sets the
4689 "extra_cert"s in addition to the certificate. (This makes sense
4690 only for "PEM" format files, as chains as a whole are not
4691 DER-encoded.)
4692 [Bodo Moeller]
4693
4694 *) Support verify_depth from the SSL API.
4695 x509_vfy.c had what can be considered an off-by-one-error:
4696 Its depth (which was not part of the external interface)
4697 was actually counting the number of certificates in a chain;
4698 now it really counts the depth.
4699 [Bodo Moeller]
4700
4701 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4702 instead of X509err, which often resulted in confusing error
4703 messages since the error codes are not globally unique
4704 (e.g. an alleged error in ssl3_accept when a certificate
4705 didn't match the private key).
4706
4707 *) New function SSL_CTX_set_session_id_context that allows to set a default
4708 value (so that you don't need SSL_set_session_id_context for each
4709 connection using the SSL_CTX).
4710 [Bodo Moeller]
4711
4712 *) OAEP decoding bug fix.
4713 [Ulf Möller]
4714
4715 *) Support INSTALL_PREFIX for package builders, as proposed by
4716 David Harris.
4717 [Bodo Moeller]
4718
4719 *) New Configure options "threads" and "no-threads". For systems
4720 where the proper compiler options are known (currently Solaris
4721 and Linux), "threads" is the default.
4722 [Bodo Moeller]
4723
4724 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4725 [Bodo Moeller]
4726
4727 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4728 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4729 such as /usr/local/bin.
4730 [Bodo Moeller]
4731
4732 *) "make linux-shared" to build shared libraries.
4733 [Niels Poppe <niels@netbox.org>]
4734
4735 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4736 [Ulf Möller]
4737
4738 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4739 extension adding in x509 utility.
4740 [Steve Henson]
4741
4742 *) Remove NOPROTO sections and error code comments.
4743 [Ulf Möller]
4744
4745 *) Partial rewrite of the DEF file generator to now parse the ANSI
4746 prototypes.
4747 [Steve Henson]
4748
4749 *) New Configure options --prefix=DIR and --openssldir=DIR.
4750 [Ulf Möller]
4751
4752 *) Complete rewrite of the error code script(s). It is all now handled
4753 by one script at the top level which handles error code gathering,
4754 header rewriting and C source file generation. It should be much better
4755 than the old method: it now uses a modified version of Ulf's parser to
4756 read the ANSI prototypes in all header files (thus the old K&R definitions
4757 aren't needed for error creation any more) and do a better job of
4758 translating function codes into names. The old 'ASN1 error code imbedded
4759 in a comment' is no longer necessary and it doesn't use .err files which
4760 have now been deleted. Also the error code call doesn't have to appear all
4761 on one line (which resulted in some large lines...).
4762 [Steve Henson]
4763
4764 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4765 [Bodo Moeller]
4766
4767 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4768 0 (which usually indicates a closed connection), but continue reading.
4769 [Bodo Moeller]
4770
4771 *) Fix some race conditions.
4772 [Bodo Moeller]
4773
4774 *) Add support for CRL distribution points extension. Add Certificate
4775 Policies and CRL distribution points documentation.
4776 [Steve Henson]
4777
4778 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4779 [Ulf Möller]
4780
4781 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4782 8 of keying material. Merlin has also confirmed interop with this fix
4783 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4784 [Merlin Hughes <merlin@baltimore.ie>]
4785
4786 *) Fix lots of warnings.
4787 [Richard Levitte <levitte@stacken.kth.se>]
4788
4789 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4790 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4791 [Richard Levitte <levitte@stacken.kth.se>]
4792
4793 *) Fix problems with sizeof(long) == 8.
4794 [Andy Polyakov <appro@fy.chalmers.se>]
4795
4796 *) Change functions to ANSI C.
4797 [Ulf Möller]
4798
4799 *) Fix typos in error codes.
4800 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4801
4802 *) Remove defunct assembler files from Configure.
4803 [Ulf Möller]
4804
4805 *) SPARC v8 assembler BIGNUM implementation.
4806 [Andy Polyakov <appro@fy.chalmers.se>]
4807
4808 *) Support for Certificate Policies extension: both print and set.
4809 Various additions to support the r2i method this uses.
4810 [Steve Henson]
4811
4812 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4813 return a const string when you are expecting an allocated buffer.
4814 [Ben Laurie]
4815
4816 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4817 types DirectoryString and DisplayText.
4818 [Steve Henson]
4819
4820 *) Add code to allow r2i extensions to access the configuration database,
4821 add an LHASH database driver and add several ctx helper functions.
4822 [Steve Henson]
4823
4824 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4825 fail when they extended the size of a BIGNUM.
4826 [Steve Henson]
4827
4828 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4829 support typesafe stack.
4830 [Steve Henson]
4831
4832 *) Fix typo in SSL_[gs]et_options().
4833 [Nils Frostberg <nils@medcom.se>]
4834
4835 *) Delete various functions and files that belonged to the (now obsolete)
4836 old X509V3 handling code.
4837 [Steve Henson]
4838
4839 *) New Configure option "rsaref".
4840 [Ulf Möller]
4841
4842 *) Don't auto-generate pem.h.
4843 [Bodo Moeller]
4844
4845 *) Introduce type-safe ASN.1 SETs.
4846 [Ben Laurie]
4847
4848 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4849 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4850
4851 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4852 that links with OpenSSL (well at least cause lots of warnings), but fear
4853 not: the conversion is trivial, and it eliminates loads of evil casts. A
4854 few STACKed things have been converted already. Feel free to convert more.
4855 In the fullness of time, I'll do away with the STACK type altogether.
4856 [Ben Laurie]
4857
4858 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4859 specified in <certfile> by updating the entry in the index.txt file.
4860 This way one no longer has to edit the index.txt file manually for
4861 revoking a certificate. The -revoke option does the gory details now.
4862 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4863
4864 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4865 `-text' option at all and this way the `-noout -text' combination was
4866 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4867 [Ralf S. Engelschall]
4868
4869 *) Make sure a corresponding plain text error message exists for the
4870 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4871 verify callback function determined that a certificate was revoked.
4872 [Ralf S. Engelschall]
4873
4874 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4875 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4876 all available cipers including rc5, which was forgotten until now.
4877 In order to let the testing shell script know which algorithms
4878 are available, a new (up to now undocumented) command
4879 "openssl list-cipher-commands" is used.
4880 [Bodo Moeller]
4881
4882 *) Bugfix: s_client occasionally would sleep in select() when
4883 it should have checked SSL_pending() first.
4884 [Bodo Moeller]
4885
4886 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4887 the raw DSA values prior to ASN.1 encoding.
4888 [Ulf Möller]
4889
4890 *) Tweaks to Configure
4891 [Niels Poppe <niels@netbox.org>]
4892
4893 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4894 yet...
4895 [Steve Henson]
4896
4897 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4898 [Ulf Möller]
4899
4900 *) New config option to avoid instructions that are illegal on the 80386.
4901 The default code is faster, but requires at least a 486.
4902 [Ulf Möller]
4903
4904 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4905 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4906 same as SSL2_VERSION anyway.
4907 [Bodo Moeller]
4908
4909 *) New "-showcerts" option for s_client.
4910 [Bodo Moeller]
4911
4912 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4913 application. Various cleanups and fixes.
4914 [Steve Henson]
4915
4916 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4917 modify error routines to work internally. Add error codes and PBE init
4918 to library startup routines.
4919 [Steve Henson]
4920
4921 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4922 packing functions to asn1 and evp. Changed function names and error
4923 codes along the way.
4924 [Steve Henson]
4925
4926 *) PKCS12 integration: and so it begins... First of several patches to
4927 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4928 objects to objects.h
4929 [Steve Henson]
4930
4931 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4932 and display support for Thawte strong extranet extension.
4933 [Steve Henson]
4934
4935 *) Add LinuxPPC support.
4936 [Jeff Dubrule <igor@pobox.org>]
4937
4938 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
4939 bn_div_words in alpha.s.
4940 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
4941
4942 *) Make sure the RSA OAEP test is skipped under -DRSAref because
4943 OAEP isn't supported when OpenSSL is built with RSAref.
4944 [Ulf Moeller <ulf@fitug.de>]
4945
4946 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
4947 so they no longer are missing under -DNOPROTO.
4948 [Soren S. Jorvang <soren@t.dk>]
4949
4950
4951 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
4952
4953 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
4954 doesn't work when the session is reused. Coming soon!
4955 [Ben Laurie]
4956
4957 *) Fix a security hole, that allows sessions to be reused in the wrong
4958 context thus bypassing client cert protection! All software that uses
4959 client certs and session caches in multiple contexts NEEDS PATCHING to
4960 allow session reuse! A fuller solution is in the works.
4961 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
4962
4963 *) Some more source tree cleanups (removed obsolete files
4964 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
4965 permission on "config" script to be executable) and a fix for the INSTALL
4966 document.
4967 [Ulf Moeller <ulf@fitug.de>]
4968
4969 *) Remove some legacy and erroneous uses of malloc, free instead of
4970 Malloc, Free.
4971 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
4972
4973 *) Make rsa_oaep_test return non-zero on error.
4974 [Ulf Moeller <ulf@fitug.de>]
4975
4976 *) Add support for native Solaris shared libraries. Configure
4977 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
4978 if someone would make that last step automatic.
4979 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
4980
4981 *) ctx_size was not built with the right compiler during "make links". Fixed.
4982 [Ben Laurie]
4983
4984 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
4985 except NULL ciphers". This means the default cipher list will no longer
4986 enable NULL ciphers. They need to be specifically enabled e.g. with
4987 the string "DEFAULT:eNULL".
4988 [Steve Henson]
4989
4990 *) Fix to RSA private encryption routines: if p < q then it would
4991 occasionally produce an invalid result. This will only happen with
4992 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
4993 [Steve Henson]
4994
4995 *) Be less restrictive and allow also `perl util/perlpath.pl
4996 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
4997 because this way one can also use an interpreter named `perl5' (which is
4998 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
4999 installed as `perl').
5000 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5001
5002 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5003 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5004
5005 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5006 advapi32.lib to Win32 build and change the pem test comparision
5007 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5008 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5009 and crypto/des/ede_cbcm_enc.c.
5010 [Steve Henson]
5011
5012 *) DES quad checksum was broken on big-endian architectures. Fixed.
5013 [Ben Laurie]
5014
5015 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5016 Win32 test batch file so it (might) work again. The Win32 test batch file
5017 is horrible: I feel ill....
5018 [Steve Henson]
5019
5020 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5021 in e_os.h. Audit of header files to check ANSI and non ANSI
5022 sections: 10 functions were absent from non ANSI section and not exported
5023 from Windows DLLs. Fixed up libeay.num for new functions.
5024 [Steve Henson]
5025
5026 *) Make `openssl version' output lines consistent.
5027 [Ralf S. Engelschall]
5028
5029 *) Fix Win32 symbol export lists for BIO functions: Added
5030 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5031 to ms/libeay{16,32}.def.
5032 [Ralf S. Engelschall]
5033
5034 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5035 fine under Unix and passes some trivial tests I've now added. But the
5036 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5037 added to make sure no one expects that this stuff really works in the
5038 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5039 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5040 openssl_bio.xs.
5041 [Ralf S. Engelschall]
5042
5043 *) Fix the generation of two part addresses in perl.
5044 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5045
5046 *) Add config entry for Linux on MIPS.
5047 [John Tobey <jtobey@channel1.com>]
5048
5049 *) Make links whenever Configure is run, unless we are on Windoze.
5050 [Ben Laurie]
5051
5052 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5053 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5054 in CRLs.
5055 [Steve Henson]
5056
5057 *) Add a useful kludge to allow package maintainers to specify compiler and
5058 other platforms details on the command line without having to patch the
5059 Configure script everytime: One now can use ``perl Configure
5060 <id>:<details>'', i.e. platform ids are allowed to have details appended
5061 to them (seperated by colons). This is treated as there would be a static
5062 pre-configured entry in Configure's %table under key <id> with value
5063 <details> and ``perl Configure <id>'' is called. So, when you want to
5064 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5065 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5066 now, which overrides the FreeBSD-elf entry on-the-fly.
5067 [Ralf S. Engelschall]
5068
5069 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5070 [Ben Laurie]
5071
5072 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5073 on the `perl Configure ...' command line. This way one can compile
5074 OpenSSL libraries with Position Independent Code (PIC) which is needed
5075 for linking it into DSOs.
5076 [Ralf S. Engelschall]
5077
5078 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5079 Fixed.
5080 [Ben Laurie]
5081
5082 *) Cleaned up the LICENSE document: The official contact for any license
5083 questions now is the OpenSSL core team under openssl-core@openssl.org.
5084 And add a paragraph about the dual-license situation to make sure people
5085 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5086 to the OpenSSL toolkit.
5087 [Ralf S. Engelschall]
5088
5089 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5090 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5091 Additonally cleaned up the `make links' target: Remove unnecessary
5092 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5093 to speed processing and no longer clutter the display with confusing
5094 stuff. Instead only the actually done links are displayed.
5095 [Ralf S. Engelschall]
5096
5097 *) Permit null encryption ciphersuites, used for authentication only. It used
5098 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5099 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5100 encryption.
5101 [Ben Laurie]
5102
5103 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5104 signed attributes when verifying signatures (this would break them),
5105 the detached data encoding was wrong and public keys obtained using
5106 X509_get_pubkey() weren't freed.
5107 [Steve Henson]
5108
5109 *) Add text documentation for the BUFFER functions. Also added a work around
5110 to a Win95 console bug. This was triggered by the password read stuff: the
5111 last character typed gets carried over to the next fread(). If you were
5112 generating a new cert request using 'req' for example then the last
5113 character of the passphrase would be CR which would then enter the first
5114 field as blank.
5115 [Steve Henson]
5116
5117 *) Added the new `Includes OpenSSL Cryptography Software' button as
5118 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5119 button and can be used by applications based on OpenSSL to show the
5120 relationship to the OpenSSL project.
5121 [Ralf S. Engelschall]
5122
5123 *) Remove confusing variables in function signatures in files
5124 ssl/ssl_lib.c and ssl/ssl.h.
5125 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5126
5127 *) Don't install bss_file.c under PREFIX/include/
5128 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5129
5130 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5131 functions that return function pointers and has support for NT specific
5132 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5133 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5134 unsigned to signed types: this was killing the Win32 compile.
5135 [Steve Henson]
5136
5137 *) Add new certificate file to stack functions,
5138 SSL_add_dir_cert_subjects_to_stack() and
5139 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5140 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5141 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5142 This means that Apache-SSL and similar packages don't have to mess around
5143 to add as many CAs as they want to the preferred list.
5144 [Ben Laurie]
5145
5146 *) Experiment with doxygen documentation. Currently only partially applied to
5147 ssl/ssl_lib.c.
5148 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5149 openssl.doxy as the configuration file.
5150 [Ben Laurie]
5151
5152 *) Get rid of remaining C++-style comments which strict C compilers hate.
5153 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5154
5155 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5156 compiled in by default: it has problems with large keys.
5157 [Steve Henson]
5158
5159 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5160 DH private keys and/or callback functions which directly correspond to
5161 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5162 is needed for applications which have to configure certificates on a
5163 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5164 (e.g. s_server).
5165 For the RSA certificate situation is makes no difference, but
5166 for the DSA certificate situation this fixes the "no shared cipher"
5167 problem where the OpenSSL cipher selection procedure failed because the
5168 temporary keys were not overtaken from the context and the API provided
5169 no way to reconfigure them.
5170 The new functions now let applications reconfigure the stuff and they
5171 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5172 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5173 non-public-API function ssl_cert_instantiate() is used as a helper
5174 function and also to reduce code redundancy inside ssl_rsa.c.
5175 [Ralf S. Engelschall]
5176
5177 *) Move s_server -dcert and -dkey options out of the undocumented feature
5178 area because they are useful for the DSA situation and should be
5179 recognized by the users.
5180 [Ralf S. Engelschall]
5181
5182 *) Fix the cipher decision scheme for export ciphers: the export bits are
5183 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5184 SSL_EXP_MASK. So, the original variable has to be used instead of the
5185 already masked variable.
5186 [Richard Levitte <levitte@stacken.kth.se>]
5187
5188 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5189 [Richard Levitte <levitte@stacken.kth.se>]
5190
5191 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5192 from `int' to `unsigned int' because it's a length and initialized by
5193 EVP_DigestFinal() which expects an `unsigned int *'.
5194 [Richard Levitte <levitte@stacken.kth.se>]
5195
5196 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5197 script. Instead use the usual Shell->Perl transition trick.
5198 [Ralf S. Engelschall]
5199
5200 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5201 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5202 -noout -modulus' as it's already the case for `openssl rsa -noout
5203 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5204 currently the public key is printed (a decision which was already done by
5205 `openssl dsa -modulus' in the past) which serves a similar purpose.
5206 Additionally the NO_RSA no longer completely removes the whole -modulus
5207 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5208 now, too.
5209 [Ralf S. Engelschall]
5210
5211 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5212 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5213 [Arne Ansper <arne@ats.cyber.ee>]
5214
5215 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5216 to be added. Now both 'req' and 'ca' can use new objects defined in the
5217 config file.
5218 [Steve Henson]
5219
5220 *) Add cool BIO that does syslog (or event log on NT).
5221 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5222
5223 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5224 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5225 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5226 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5227 [Ben Laurie]
5228
5229 *) Add preliminary config info for new extension code.
5230 [Steve Henson]
5231
5232 *) Make RSA_NO_PADDING really use no padding.
5233 [Ulf Moeller <ulf@fitug.de>]
5234
5235 *) Generate errors when private/public key check is done.
5236 [Ben Laurie]
5237
5238 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5239 for some CRL extensions and new objects added.
5240 [Steve Henson]
5241
5242 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5243 key usage extension and fuller support for authority key id.
5244 [Steve Henson]
5245
5246 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5247 padding method for RSA, which is recommended for new applications in PKCS
5248 #1 v2.0 (RFC 2437, October 1998).
5249 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5250 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5251 against Bleichbacher's attack on RSA.
5252 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5253 Ben Laurie]
5254
5255 *) Updates to the new SSL compression code
5256 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5257
5258 *) Fix so that the version number in the master secret, when passed
5259 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5260 (because the server will not accept higher), that the version number
5261 is 0x03,0x01, not 0x03,0x00
5262 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5263
5264 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5265 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5266 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5267 [Steve Henson]
5268
5269 *) Support for RAW extensions where an arbitrary extension can be
5270 created by including its DER encoding. See apps/openssl.cnf for
5271 an example.
5272 [Steve Henson]
5273
5274 *) Make sure latest Perl versions don't interpret some generated C array
5275 code as Perl array code in the crypto/err/err_genc.pl script.
5276 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5277
5278 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5279 not many people have the assembler. Various Win32 compilation fixes and
5280 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5281 build instructions.
5282 [Steve Henson]
5283
5284 *) Modify configure script 'Configure' to automatically create crypto/date.h
5285 file under Win32 and also build pem.h from pem.org. New script
5286 util/mkfiles.pl to create the MINFO file on environments that can't do a
5287 'make files': perl util/mkfiles.pl >MINFO should work.
5288 [Steve Henson]
5289
5290 *) Major rework of DES function declarations, in the pursuit of correctness
5291 and purity. As a result, many evil casts evaporated, and some weirdness,
5292 too. You may find this causes warnings in your code. Zapping your evil
5293 casts will probably fix them. Mostly.
5294 [Ben Laurie]
5295
5296 *) Fix for a typo in asn1.h. Bug fix to object creation script
5297 obj_dat.pl. It considered a zero in an object definition to mean
5298 "end of object": none of the objects in objects.h have any zeros
5299 so it wasn't spotted.
5300 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5301
5302 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5303 Masking (CBCM). In the absence of test vectors, the best I have been able
5304 to do is check that the decrypt undoes the encrypt, so far. Send me test
5305 vectors if you have them.
5306 [Ben Laurie]
5307
5308 *) Correct calculation of key length for export ciphers (too much space was
5309 allocated for null ciphers). This has not been tested!
5310 [Ben Laurie]
5311
5312 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5313 message is now correct (it understands "crypto" and "ssl" on its
5314 command line). There is also now an "update" option. This will update
5315 the util/ssleay.num and util/libeay.num files with any new functions.
5316 If you do a:
5317 perl util/mkdef.pl crypto ssl update
5318 it will update them.
5319 [Steve Henson]
5320
5321 *) Overhauled the Perl interface (perl/*):
5322 - ported BN stuff to OpenSSL's different BN library
5323 - made the perl/ source tree CVS-aware
5324 - renamed the package from SSLeay to OpenSSL (the files still contain
5325 their history because I've copied them in the repository)
5326 - removed obsolete files (the test scripts will be replaced
5327 by better Test::Harness variants in the future)
5328 [Ralf S. Engelschall]
5329
5330 *) First cut for a very conservative source tree cleanup:
5331 1. merge various obsolete readme texts into doc/ssleay.txt
5332 where we collect the old documents and readme texts.
5333 2. remove the first part of files where I'm already sure that we no
5334 longer need them because of three reasons: either they are just temporary
5335 files which were left by Eric or they are preserved original files where
5336 I've verified that the diff is also available in the CVS via "cvs diff
5337 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5338 the crypto/md/ stuff).
5339 [Ralf S. Engelschall]
5340
5341 *) More extension code. Incomplete support for subject and issuer alt
5342 name, issuer and authority key id. Change the i2v function parameters
5343 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5344 what that's for :-) Fix to ASN1 macro which messed up
5345 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5346 [Steve Henson]
5347
5348 *) Preliminary support for ENUMERATED type. This is largely copied from the
5349 INTEGER code.
5350 [Steve Henson]
5351
5352 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5353 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5354
5355 *) Make sure `make rehash' target really finds the `openssl' program.
5356 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5357
5358 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5359 like to hear about it if this slows down other processors.
5360 [Ben Laurie]
5361
5362 *) Add CygWin32 platform information to Configure script.
5363 [Alan Batie <batie@aahz.jf.intel.com>]
5364
5365 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5366 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5367
5368 *) New program nseq to manipulate netscape certificate sequences
5369 [Steve Henson]
5370
5371 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5372 few typos.
5373 [Steve Henson]
5374
5375 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5376 but the BN code had some problems that would cause failures when
5377 doing certificate verification and some other functions.
5378 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5379
5380 *) Add ASN1 and PEM code to support netscape certificate sequences.
5381 [Steve Henson]
5382
5383 *) Add ASN1 and PEM code to support netscape certificate sequences.
5384 [Steve Henson]
5385
5386 *) Add several PKIX and private extended key usage OIDs.
5387 [Steve Henson]
5388
5389 *) Modify the 'ca' program to handle the new extension code. Modify
5390 openssl.cnf for new extension format, add comments.
5391 [Steve Henson]
5392
5393 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5394 and add a sample to openssl.cnf so req -x509 now adds appropriate
5395 CA extensions.
5396 [Steve Henson]
5397
5398 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5399 error code, add initial support to X509_print() and x509 application.
5400 [Steve Henson]
5401
5402 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5403 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5404 stuff is currently isolated and isn't even compiled yet.
5405 [Steve Henson]
5406
5407 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5408 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5409 Removed the versions check from X509 routines when loading extensions:
5410 this allows certain broken certificates that don't set the version
5411 properly to be processed.
5412 [Steve Henson]
5413
5414 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5415 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5416 can still be regenerated with "make depend".
5417 [Ben Laurie]
5418
5419 *) Spelling mistake in C version of CAST-128.
5420 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5421
5422 *) Changes to the error generation code. The perl script err-code.pl
5423 now reads in the old error codes and retains the old numbers, only
5424 adding new ones if necessary. It also only changes the .err files if new
5425 codes are added. The makefiles have been modified to only insert errors
5426 when needed (to avoid needlessly modifying header files). This is done
5427 by only inserting errors if the .err file is newer than the auto generated
5428 C file. To rebuild all the error codes from scratch (the old behaviour)
5429 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5430 or delete all the .err files.
5431 [Steve Henson]
5432
5433 *) CAST-128 was incorrectly implemented for short keys. The C version has
5434 been fixed, but is untested. The assembler versions are also fixed, but
5435 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5436 to regenerate it if needed.
5437 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5438 Hagino <itojun@kame.net>]
5439
5440 *) File was opened incorrectly in randfile.c.
5441 [Ulf Möller <ulf@fitug.de>]
5442
5443 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5444 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5445 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5446 al: it's just almost always a UTCTime. Note this patch adds new error
5447 codes so do a "make errors" if there are problems.
5448 [Steve Henson]
5449
5450 *) Correct Linux 1 recognition in config.
5451 [Ulf Möller <ulf@fitug.de>]
5452
5453 *) Remove pointless MD5 hash when using DSA keys in ca.
5454 [Anonymous <nobody@replay.com>]
5455
5456 *) Generate an error if given an empty string as a cert directory. Also
5457 generate an error if handed NULL (previously returned 0 to indicate an
5458 error, but didn't set one).
5459 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5460
5461 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5462 [Ben Laurie]
5463
5464 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5465 parameters. This was causing a warning which killed off the Win32 compile.
5466 [Steve Henson]
5467
5468 *) Remove C++ style comments from crypto/bn/bn_local.h.
5469 [Neil Costigan <neil.costigan@celocom.com>]
5470
5471 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5472 based on a text string, looking up short and long names and finally
5473 "dot" format. The "dot" format stuff didn't work. Added new function
5474 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5475 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5476 OID is not part of the table.
5477 [Steve Henson]
5478
5479 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5480 X509_LOOKUP_by_alias().
5481 [Ben Laurie]
5482
5483 *) Sort openssl functions by name.
5484 [Ben Laurie]
5485
5486 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5487 encryption from sample DSA keys (in case anyone is interested the password
5488 was "1234").
5489 [Steve Henson]
5490
5491 *) Make _all_ *_free functions accept a NULL pointer.
5492 [Frans Heymans <fheymans@isaserver.be>]
5493
5494 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5495 NULL pointers.
5496 [Anonymous <nobody@replay.com>]
5497
5498 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5499 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5500
5501 *) Don't blow it for numeric -newkey arguments to apps/req.
5502 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5503
5504 *) Temp key "for export" tests were wrong in s3_srvr.c.
5505 [Anonymous <nobody@replay.com>]
5506
5507 *) Add prototype for temp key callback functions
5508 SSL_CTX_set_tmp_{rsa,dh}_callback().
5509 [Ben Laurie]
5510
5511 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5512 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5513 [Steve Henson]
5514
5515 *) X509_name_add_entry() freed the wrong thing after an error.
5516 [Arne Ansper <arne@ats.cyber.ee>]
5517
5518 *) rsa_eay.c would attempt to free a NULL context.
5519 [Arne Ansper <arne@ats.cyber.ee>]
5520
5521 *) BIO_s_socket() had a broken should_retry() on Windoze.
5522 [Arne Ansper <arne@ats.cyber.ee>]
5523
5524 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5525 [Arne Ansper <arne@ats.cyber.ee>]
5526
5527 *) Make sure the already existing X509_STORE->depth variable is initialized
5528 in X509_STORE_new(), but document the fact that this variable is still
5529 unused in the certificate verification process.
5530 [Ralf S. Engelschall]
5531
5532 *) Fix the various library and apps files to free up pkeys obtained from
5533 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5534 [Steve Henson]
5535
5536 *) Fix reference counting in X509_PUBKEY_get(). This makes
5537 demos/maurice/example2.c work, amongst others, probably.
5538 [Steve Henson and Ben Laurie]
5539
5540 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5541 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5542 are no longer created. This way we have a single and consistent command
5543 line interface `openssl <command>', similar to `cvs <command>'.
5544 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5545
5546 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5547 BIT STRING wrapper always have zero unused bits.
5548 [Steve Henson]
5549
5550 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5551 [Steve Henson]
5552
5553 *) Make the top-level INSTALL documentation easier to understand.
5554 [Paul Sutton]
5555
5556 *) Makefiles updated to exit if an error occurs in a sub-directory
5557 make (including if user presses ^C) [Paul Sutton]
5558
5559 *) Make Montgomery context stuff explicit in RSA data structure.
5560 [Ben Laurie]
5561
5562 *) Fix build order of pem and err to allow for generated pem.h.
5563 [Ben Laurie]
5564
5565 *) Fix renumbering bug in X509_NAME_delete_entry().
5566 [Ben Laurie]
5567
5568 *) Enhanced the err-ins.pl script so it makes the error library number
5569 global and can add a library name. This is needed for external ASN1 and
5570 other error libraries.
5571 [Steve Henson]
5572
5573 *) Fixed sk_insert which never worked properly.
5574 [Steve Henson]
5575
5576 *) Fix ASN1 macros so they can handle indefinite length construted
5577 EXPLICIT tags. Some non standard certificates use these: they can now
5578 be read in.
5579 [Steve Henson]
5580
5581 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5582 into a single doc/ssleay.txt bundle. This way the information is still
5583 preserved but no longer messes up this directory. Now it's new room for
5584 the new set of documenation files.
5585 [Ralf S. Engelschall]
5586
5587 *) SETs were incorrectly DER encoded. This was a major pain, because they
5588 shared code with SEQUENCEs, which aren't coded the same. This means that
5589 almost everything to do with SETs or SEQUENCEs has either changed name or
5590 number of arguments.
5591 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5592
5593 *) Fix test data to work with the above.
5594 [Ben Laurie]
5595
5596 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5597 was already fixed by Eric for 0.9.1 it seems.
5598 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5599
5600 *) Autodetect FreeBSD3.
5601 [Ben Laurie]
5602
5603 *) Fix various bugs in Configure. This affects the following platforms:
5604 nextstep
5605 ncr-scde
5606 unixware-2.0
5607 unixware-2.0-pentium
5608 sco5-cc.
5609 [Ben Laurie]
5610
5611 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5612 before they are needed.
5613 [Ben Laurie]
5614
5615 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5616 [Ben Laurie]
5617
5618
5619 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5620
5621 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5622 changed SSLeay to OpenSSL in version strings.
5623 [Ralf S. Engelschall]
5624
5625 *) Some fixups to the top-level documents.
5626 [Paul Sutton]
5627
5628 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5629 because the symlink to include/ was missing.
5630 [Ralf S. Engelschall]
5631
5632 *) Incorporated the popular no-RSA/DSA-only patches
5633 which allow to compile a RSA-free SSLeay.
5634 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5635
5636 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5637 when "ssleay" is still not found.
5638 [Ralf S. Engelschall]
5639
5640 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5641 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5642
5643 *) Updated the README file.
5644 [Ralf S. Engelschall]
5645
5646 *) Added various .cvsignore files in the CVS repository subdirs
5647 to make a "cvs update" really silent.
5648 [Ralf S. Engelschall]
5649
5650 *) Recompiled the error-definition header files and added
5651 missing symbols to the Win32 linker tables.
5652 [Ralf S. Engelschall]
5653
5654 *) Cleaned up the top-level documents;
5655 o new files: CHANGES and LICENSE
5656 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5657 o merged COPYRIGHT into LICENSE
5658 o removed obsolete TODO file
5659 o renamed MICROSOFT to INSTALL.W32
5660 [Ralf S. Engelschall]
5661
5662 *) Removed dummy files from the 0.9.1b source tree:
5663 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5664 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5665 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5666 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5667 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5668 [Ralf S. Engelschall]
5669
5670 *) Added various platform portability fixes.
5671 [Mark J. Cox]
5672
5673 *) The Genesis of the OpenSSL rpject:
5674 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5675 Young and Tim J. Hudson created while they were working for C2Net until
5676 summer 1998.
5677 [The OpenSSL Project]
5678
5679
5680 Changes between 0.9.0b and 0.9.1b [not released]
5681
5682 *) Updated a few CA certificates under certs/
5683 [Eric A. Young]
5684
5685 *) Changed some BIGNUM api stuff.
5686 [Eric A. Young]
5687
5688 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5689 DGUX x86, Linux Alpha, etc.
5690 [Eric A. Young]
5691
5692 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5693 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5694 available).
5695 [Eric A. Young]
5696
5697 *) Add -strparse option to asn1pars program which parses nested
5698 binary structures
5699 [Dr Stephen Henson <shenson@bigfoot.com>]
5700
5701 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5702 [Eric A. Young]
5703
5704 *) DSA fix for "ca" program.
5705 [Eric A. Young]
5706
5707 *) Added "-genkey" option to "dsaparam" program.
5708 [Eric A. Young]
5709
5710 *) Added RIPE MD160 (rmd160) message digest.
5711 [Eric A. Young]
5712
5713 *) Added -a (all) option to "ssleay version" command.
5714 [Eric A. Young]
5715
5716 *) Added PLATFORM define which is the id given to Configure.
5717 [Eric A. Young]
5718
5719 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5720 [Eric A. Young]
5721
5722 *) Extended the ASN.1 parser routines.
5723 [Eric A. Young]
5724
5725 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5726 [Eric A. Young]
5727
5728 *) Added a BN_CTX to the BN library.
5729 [Eric A. Young]
5730
5731 *) Fixed the weak key values in DES library
5732 [Eric A. Young]
5733
5734 *) Changed API in EVP library for cipher aliases.
5735 [Eric A. Young]
5736
5737 *) Added support for RC2/64bit cipher.
5738 [Eric A. Young]
5739
5740 *) Converted the lhash library to the crypto/mem.c functions.
5741 [Eric A. Young]
5742
5743 *) Added more recognized ASN.1 object ids.
5744 [Eric A. Young]
5745
5746 *) Added more RSA padding checks for SSL/TLS.
5747 [Eric A. Young]
5748
5749 *) Added BIO proxy/filter functionality.
5750 [Eric A. Young]
5751
5752 *) Added extra_certs to SSL_CTX which can be used
5753 send extra CA certificates to the client in the CA cert chain sending
5754 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5755 [Eric A. Young]
5756
5757 *) Now Fortezza is denied in the authentication phase because
5758 this is key exchange mechanism is not supported by SSLeay at all.
5759 [Eric A. Young]
5760
5761 *) Additional PKCS1 checks.
5762 [Eric A. Young]
5763
5764 *) Support the string "TLSv1" for all TLS v1 ciphers.
5765 [Eric A. Young]
5766
5767 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5768 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5769 [Eric A. Young]
5770
5771 *) Fixed a few memory leaks.
5772 [Eric A. Young]
5773
5774 *) Fixed various code and comment typos.
5775 [Eric A. Young]
5776
5777 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5778 bytes sent in the client random.
5779 [Edward Bishop <ebishop@spyglass.com>]
5780