5 Changes between 0.9.7 and 0.9.8 [xx XXX 2002]
7 *) Add ECDSA in new directory crypto/ecdsa/.
9 Add applications 'openssl ecdsaparam' and 'openssl ecdsa'
10 (these are variants of 'openssl dsaparam' and 'openssl dsa').
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>]
22 *) Add reference counting for EC_GROUP objects.
23 [Nils Larsch <nla@trustcenter.de>]
25 *) Include some named elliptic curves. These can be obtained from
28 EC_GROUP_new_by_name()
29 Also add a 'nid' field to EC_GROUP objects, which can be accessed
33 [Nils Larsch <nla@trustcenter.de, Bodo Moeller]
35 Changes between 0.9.6 and 0.9.7 [xx XXX 2002]
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.
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
46 +) Fix the 'app_verify_callback' interface so that the user-defined
47 argument is actually passed to the callback: In the
48 SSL_CTX_set_cert_verify_callback() prototype, the callback
49 declaration has been changed from
52 int (*cb)(X509_STORE_CTX *,void *);
53 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
54 i=s->ctx->app_verify_callback(&ctx)
56 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
58 To update applications using SSL_CTX_set_cert_verify_callback(),
59 a dummy argument can be added to their callback functions.
60 [D. K. Smetters <smetters@parc.xerox.com>]
62 +) Added the '4758cca' ENGINE to support IBM 4758 cards.
63 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
65 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
66 Yoram Zahavi <YoramZ@gilian.com>).
69 +) Add and OPENSSL_LOAD_CONF define which will cause
70 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
71 This allows older applications to transparently support certain
72 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
73 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
74 load the config file and OPENSSL_add_all_algorithms_conf() which will
75 always load it have also been added.
78 +) Config modules support in openssl utility.
80 Most commands now load modules from the config file,
81 though in a few (such as version) this isn't done
82 because it couldn't be used for anything.
84 In the case of ca and req the config file used is
85 the same as the utility itself: that is the -config
86 command line option can be used to specify an
90 +) Move default behaviour from OPENSSL_config(). If appname is NULL
91 use "openssl_conf" if filename is NULL use default openssl config file.
94 +) Add an argument to OPENSSL_config() to allow the use of an alternative
95 config section name. Add a new flag to tolerate a missing config file
96 and move code to CONF_modules_load_file().
99 *) Add information about CygWin 1.3 and on, and preserve proper
100 configuration for the versions before that.
101 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
103 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
104 check whether we deal with a copy of a session and do not delete from
105 the cache in this case. Problem reported by "Izhar Shoshani Levi"
106 <izhar@checkpoint.com>.
109 *) Do not store session data into the internal session cache, if it
110 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
111 flag is set). Proposed by Aslam <aslam@funk.com>.
114 +) Support for crypto accelerator cards from Accelerated Encryption
115 Processing, www.aep.ie. (Use engine 'aep')
116 The support was copied from 0.9.6c [engine] and adapted/corrected
117 to work with the new engine framework.
118 [AEP Inc. and Richard Levitte]
120 +) Support for SureWare crypto accelerator cards from Baltimore
121 Technologies. (Use engine 'sureware')
122 The support was copied from 0.9.6c [engine] and adapted
123 to work with the new engine framework.
126 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
130 +) Have the CHIL engine fork-safe (as defined by nCipher) and actually
131 make the newer ENGINE framework commands for the CHIL engine work.
132 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
134 +) Make it possible to produce shared libraries on ReliantUNIX.
135 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
137 *) Add the configuration target linux-s390x.
138 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
140 +) Add the configuration target debug-linux-ppro.
141 Make 'openssl rsa' use the general key loading routines
142 implemented in apps.c, and make those routines able to
143 handle the key format FORMAT_NETSCAPE and the variant
145 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
147 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
148 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
150 +) Add -keyform to rsautl, and document -engine.
151 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
153 +) Change BIO_new_file (crypto/bio/bss_file.c) to use new
154 BIO_R_NO_SUCH_FILE error code rather than the generic
155 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
160 ERR_peek_last_error_line
161 ERR_peek_last_error_line_data.
165 ERR_peek_error_line_data,
166 but report on the latest error recorded rather than the first one
167 still in the error queue.
168 [Ben Laurie, Bodo Moeller]
170 +) default_algorithms option in ENGINE config module. This allows things
172 default_algorithms = ALL
173 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
176 +) Prelminary ENGINE config module.
179 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
180 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
181 variable as an indication that a ClientHello message has been
182 received. As the flag value will be lost between multiple
183 invocations of ssl3_accept when using non-blocking I/O, the
184 function may not be aware that a handshake has actually taken
185 place, thus preventing a new session from being added to the
188 To avoid this problem, we now set s->new_session to 2 instead of
189 using a local variable.
190 [Lutz Jaenicke, Bodo Moeller]
192 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
193 if the SSL_R_LENGTH_MISMATCH error is detected.
194 [Geoff Thorpe, Bodo Moeller]
196 +) New experimental application configuration code.
199 *) New 'shared_ldflag' column in Configure platform table.
202 *) Fix EVP_CIPHER_mode macro.
203 ["Dan S. Camper" <dan@bti.net>]
205 +) Change the AES code to follow the same name structure as all other
206 symmetric ciphers, and behave the same way. Move everything to
207 the directory crypto/aes, thereby obsoleting crypto/rijndael.
208 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
210 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
211 type, we must throw them away by setting rr->length to 0.
212 [D P Chang <dpc@qualys.com>]
214 -) OpenSSL 0.9.6c released [21 dec 2001]
216 +) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
217 [Ben Laurie and Theo de Raadt]
219 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
220 <Dominikus.Scherkl@biodata.com>. (The previous implementation
221 worked incorrectly for those cases where range = 10..._2 and
222 3*range is two bits longer than range.)
225 *) Only add signing time to PKCS7 structures if it is not already
229 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
230 OBJ_ld_ce should be OBJ_id_ce.
231 Also some ip-pda OIDs in crypto/objects/objects.txt were
232 incorrect (cf. RFC 3039).
233 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
235 +) Add option to output public keys in req command.
236 [Massimiliano Pala madwolf@openca.org]
238 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
239 returns early because it has nothing to do.
240 [Andy Schneider <andy.schneider@bjss.co.uk>]
242 *) [In 0.9.6c-engine and 0.9.7 release:]
243 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
244 [Andy Schneider <andy.schneider@bjss.co.uk>]
246 -) [In 0.9.6c-engine release:]
247 Add support for Cryptographic Appliance's keyserver technology.
248 (Use engine 'keyclient')
249 [Cryptographic Appliances and Geoff Thorpe]
251 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
252 is called via tools/c89.sh because arguments have to be
253 rearranged (all '-L' options must appear before the first object
255 [Richard Shapiro <rshapiro@abinitio.com>]
257 +) Use wNAFs in EC_POINTs_mul() for improved efficiency
258 (up to about 10% better than before for P-192 and P-224).
261 -) [In 0.9.6c-engine release:]
262 Add support for Broadcom crypto accelerator cards, backported
264 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
266 -) [In 0.9.6c-engine release:]
267 Add support for SureWare crypto accelerator cards from
268 Baltimore Technologies. (Use engine 'sureware')
269 [Baltimore Technologies and Mark Cox]
271 -) [In 0.9.6c-engine release:]
272 Add support for crypto accelerator cards from Accelerated
273 Encryption Processing, www.aep.ie. (Use engine 'aep')
274 [AEP Inc. and Mark Cox]
276 *) Add a configuration entry for gcc on UnixWare.
277 [Gary Benson <gbenson@redhat.com>]
279 +) New functions/macros
281 SSL_CTX_set_msg_callback(ctx, cb)
282 SSL_CTX_set_msg_callback_arg(ctx, arg)
283 SSL_set_msg_callback(ssl, cb)
284 SSL_set_msg_callback_arg(ssl, arg)
286 to request calling a callback function
288 void cb(int write_p, int version, int content_type,
289 const void *buf, size_t len, SSL *ssl, void *arg)
291 whenever a protocol message has been completely received
292 (write_p == 0) or sent (write_p == 1). Here 'version' is the
293 protocol version according to which the SSL library interprets
294 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
295 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
296 the content type as defined in the SSL 3.0/TLS 1.0 protocol
297 specification (change_cipher_spec(20), alert(21), handshake(22)).
298 'buf' and 'len' point to the actual message, 'ssl' to the
299 SSL object, and 'arg' is the application-defined value set by
300 SSL[_CTX]_set_msg_callback_arg().
302 'openssl s_client' and 'openssl s_server' have new '-msg' options
303 to enable a callback that displays all protocol messages.
306 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
307 messages are stored in a single piece (fixed-length part and
308 variable-length part combined) and fix various bugs found on the way.
311 +) Change the shared library support so shared libraries are built as
312 soon as the corresponding static library is finished, and thereby get
313 openssl and the test programs linked against the shared library.
314 This still only happens when the keyword "shard" has been given to
315 the configuration scripts.
317 NOTE: shared library support is still an experimental thing, and
318 backward binary compatibility is still not guaranteed.
319 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
321 +) Add support for Subject Information Access extension.
322 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
324 +) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
325 additional bytes when new memory had to be allocated, not just
326 when reusing an existing buffer.
329 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
330 instead. BIO_gethostbyname() does not know what timeouts are
331 appropriate, so entries would stay in cache even when they have
333 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
335 +) New command line and configuration option 'utf8' for the req command.
336 This allows field values to be specified as UTF8 strings.
339 +) Add -multi and -mr options to "openssl speed" - giving multiple parallel
340 runs for the former and machine-readable output for the latter.
343 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
344 of the e-mail address in the DN (i.e., it will go into a certificate
345 extension only). The new configuration file option 'email_in_dn = no'
347 [Massimiliano Pala madwolf@openca.org]
349 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
350 faced with a pathologically small ClientHello fragment that does
351 not contain client_version: Instead of aborting with an error,
352 simply choose the highest available protocol version (i.e.,
353 TLS 1.0 unless it is disabled). In practice, ClientHello
354 messages are never sent like this, but this change gives us
355 strictly correct behaviour at least for TLS.
358 +) Change all functions with names starting with des_ to be starting
359 with DES_ instead. Add wrappers that are compatible with libdes,
360 but are named _ossl_old_des_*. Finally, add macros that map the
361 des_* symbols to the corresponding _ossl_old_des_*.
363 All this is done because there are increasing clashes with libdes
364 and other DES libraries that are currently used by other projects.
365 The old libdes interface (including crypt()) is provided if
366 <openssl/des_old.h> is included. For now, this automatically
367 happens in <openssl/des.h> unless OPENSSL_DISABLE_OLD_DES_SUPPORT is
368 defined. Note that crypt() is no longer declared in <openssl/des.h>.
370 NOTE: This is a major break of an old API into a new one. Software
371 authors are encouraged to switch to the DES_ style functions. Some
372 time in the future, des_old.h and the libdes compatibility functions
373 will be completely removed.
376 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
377 never resets s->method to s->ctx->method when called from within
378 one of the SSL handshake functions.
379 [Bodo Moeller; problem pointed out by Niko Baric]
381 +) Test for certificates which contain unsupported critical extensions.
382 If such a certificate is found during a verify operation it is
383 rejected by default: this behaviour can be overridden by either
384 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
385 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
386 X509_supported_extension() has also been added which returns 1 if a
387 particular extension is supported.
390 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
391 (sent using the client's version number) if client_version is
392 smaller than the protocol version in use. Also change
393 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
394 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
395 the client will at least see that alert.
398 +) Modify the behaviour of EVP cipher functions in similar way to digests
399 to retain compatibility with existing code.
402 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
403 compatibility with existing code. In particular the 'ctx' parameter does
404 not have to be to be initialized before the call to EVP_DigestInit() and
405 it is tidied up after a call to EVP_DigestFinal(). New function
406 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
407 EVP_MD_CTX_copy() changed to not require the destination to be
408 initialized valid and new function EVP_MD_CTX_copy_ex() added which
409 requires the destination to be valid.
411 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
412 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
415 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
416 so that complete 'Handshake' protocol structures are kept in memory
417 instead of overwriting 'msg_type' and 'length' with 'body' data.
420 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
424 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
425 [Massimo Santin via Richard Levitte]
427 +) Major restructuring to the underlying ENGINE code. This includes
428 reduction of linker bloat, separation of pure "ENGINE" manipulation
429 (initialisation, etc) from functionality dealing with implementations
430 of specific crypto iterfaces. This change also introduces integrated
431 support for symmetric ciphers and digest implementations - so ENGINEs
432 can now accelerate these by providing EVP_CIPHER and EVP_MD
433 implementations of their own. This is detailed in crypto/engine/README
434 as it couldn't be adequately described here. However, there are a few
435 API changes worth noting - some RSA, DSA, DH, and RAND functions that
436 were changed in the original introduction of ENGINE code have now
437 reverted back - the hooking from this code to ENGINE is now a good
438 deal more passive and at run-time, operations deal directly with
439 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
440 dereferencing through an ENGINE pointer any more. Also, the ENGINE
441 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
442 they were not being used by the framework as there is no concept of a
443 BIGNUM_METHOD and they could not be generalised to the new
444 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
445 ENGINE_cpy() has been removed as it cannot be consistently defined in
449 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
452 +) Change mkdef.pl to sort symbols that get the same entry number,
453 and make sure the automatically generated functions ERR_load_*
454 become part of libeay.num as well.
457 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
458 client receives HelloRequest while in a handshake.
459 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
461 +) New function SSL_renegotiate_pending(). This returns true once
462 renegotiation has been requested (either SSL_renegotiate() call
463 or HelloRequest/ClientHello receveived from the peer) and becomes
464 false once a handshake has been completed.
465 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
466 sends a HelloRequest, but does not ensure that a handshake takes
467 place. SSL_renegotiate_pending() is useful for checking if the
468 client has followed the request.)
471 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
472 By default, clients may request session resumption even during
473 renegotiation (if session ID contexts permit); with this option,
474 session resumption is possible only in the first handshake.
477 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
478 should end in 'break', not 'goto end' which circuments various
479 cleanups done in state SSL_ST_OK. But session related stuff
480 must be disabled for SSL_ST_OK in the case that we just sent a
483 Also avoid some overhead by not calling ssl_init_wbio_buffer()
484 before just sending a HelloRequest.
485 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
487 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
488 reveal whether illegal block cipher padding was found or a MAC
489 verification error occured. (Neither SSLerr() codes nor alerts
490 are directly visible to potential attackers, but the information
491 may leak via logfiles.)
493 Similar changes are not required for the SSL 2.0 implementation
494 because the number of padding bytes is sent in clear for SSL 2.0,
495 and the extra bytes are just ignored. However ssl/s2_pkt.c
496 failed to verify that the purported number of padding bytes is in
500 +) Add some demos for certificate and certificate request creation.
503 +) Make maximum certificate chain size accepted from the peer application
504 settable (SSL*_get/set_max_cert_list()), as proposed by
505 "Douglas E. Engert" <deengert@anl.gov>.
508 +) Add support for shared libraries for Unixware-7
509 (Boyd Lynn Gerber <gerberb@zenez.com>).
512 *) Add OpenUNIX-8 support including shared libraries
513 (Boyd Lynn Gerber <gerberb@zenez.com>).
516 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
517 'wristwatch attack' using huge encoding parameters (cf.
518 James H. Manger's CRYPTO 2001 paper). Note that the
519 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
520 encoding parameters and hence was not vulnerable.
523 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
524 be done prior to destruction. Use this to unload error strings from
525 ENGINEs that load their own error strings. NB: This adds two new API
526 functions to "get" and "set" this destroy handler in an ENGINE.
529 +) Alter all existing ENGINE implementations (except "openssl" and
530 "openbsd") to dynamically instantiate their own error strings. This
531 makes them more flexible to be built both as statically-linked ENGINEs
532 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
533 Also, add stub code to each that makes building them as self-contained
534 shared-libraries easier (see README.ENGINE).
537 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
538 implementations into applications that are completely implemented in
539 self-contained shared-libraries. The "dynamic" ENGINE exposes control
540 commands that can be used to configure what shared-library to load and
541 to control aspects of the way it is handled. Also, made an update to
542 the README.ENGINE file that brings its information up-to-date and
543 provides some information and instructions on the "dynamic" ENGINE
544 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
548 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
550 +) Make it possible to unload ranges of ERR strings with a new
551 "ERR_unload_strings" function.
554 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
555 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
556 followed by modular reduction.
557 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
559 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
560 equivalent based on BN_pseudo_rand() instead of BN_rand().
563 +) Add a copy() function to EVP_MD.
566 +) Make EVP_MD routines take a context pointer instead of just the
567 md_data void pointer.
570 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
571 that the digest can only process a single chunk of data
572 (typically because it is provided by a piece of
573 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
574 is only going to provide a single chunk of data, and hence the
575 framework needn't accumulate the data for oneshot drivers.
578 +) As with "ERR", make it possible to replace the underlying "ex_data"
579 functions. This change also alters the storage and management of global
580 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
581 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
582 index counters. The API functions that use this state have been changed
583 to take a "class_index" rather than pointers to the class's local STACK
584 and counter, and there is now an API function to dynamically create new
585 classes. This centralisation allows us to (a) plug a lot of the
586 thread-safety problems that existed, and (b) makes it possible to clean
587 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
588 such data would previously have always leaked in application code and
589 workarounds were in place to make the memory debugging turn a blind eye
590 to it. Application code that doesn't use this new function will still
591 leak as before, but their memory debugging output will announce it now
592 rather than letting it slide.
594 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
595 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
596 has a return value to indicate success or failure.
599 +) Make it possible to replace the underlying "ERR" functions such that the
600 global state (2 LHASH tables and 2 locks) is only used by the "default"
601 implementation. This change also adds two functions to "get" and "set"
602 the implementation prior to it being automatically set the first time
603 any other ERR function takes place. Ie. an application can call "get",
604 pass the return value to a module it has just loaded, and that module
605 can call its own "set" function using that value. This means the
606 module's "ERR" operations will use (and modify) the error state in the
607 application and not in its own statically linked copy of OpenSSL code.
610 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
611 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
612 the operation, and provides a more encapsulated way for external code
613 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
614 to use these functions rather than manually incrementing the counts.
616 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
619 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
620 This function was broken, as the check for a new client hello message
621 to handle SGC did not allow these large messages.
622 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
625 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
628 +) Add EVP test program.
631 +) Add symmetric cipher support to ENGINE. Expect the API to change!
634 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
635 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
636 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
637 These allow a CRL to be built without having to access X509_CRL fields
638 directly. Modify 'ca' application to use new functions.
641 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
642 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
645 *) Rework the configuration and shared library support for Tru64 Unix.
646 The configuration part makes use of modern compiler features and
647 still retains old compiler behavior for those that run older versions
648 of the OS. The shared library support part includes a variant that
649 uses the RPATH feature, and is available through the special
650 configuration target "alpha-cc-rpath", which will never be selected
652 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
654 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
655 with the same message size as in ssl3_get_certificate_request().
656 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
657 messages might inadvertently be reject as too long.
658 [Petr Lampa <lampa@fee.vutbr.cz>]
660 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
661 bug workarounds. Rollback attack detection is a security feature.
662 The problem will only arise on OpenSSL servers when TLSv1 is not
663 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
664 Software authors not wanting to support TLSv1 will have special reasons
665 for their choice and can explicitly enable this option.
666 [Bodo Moeller, Lutz Jaenicke]
668 +) Rationalise EVP so it can be extended: don't include a union of
669 cipher/digest structures, add init/cleanup functions. This also reduces
670 the number of header dependencies.
673 +) Make DES key schedule conform to the usual scheme, as well as
674 correcting its structure. This means that calls to DES functions
675 now have to pass a pointer to a des_key_schedule instead of a
676 plain des_key_schedule (which was actually always a pointer
680 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
683 *) Modified SSL library such that the verify_callback that has been set
684 specificly for an SSL object with SSL_set_verify() is actually being
685 used. Before the change, a verify_callback set with this function was
686 ignored and the verify_callback() set in the SSL_CTX at the time of
687 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
688 to allow the necessary settings.
691 +) Initial reduction of linker bloat: the use of some functions, such as
692 PEM causes large amounts of unused functions to be linked in due to
693 poor organisation. For example pem_all.c contains every PEM function
694 which has a knock on effect of linking in large amounts of (unused)
695 ASN1 code. Grouping together similar functions and splitting unrelated
696 functions prevents this.
699 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
700 explicitly to NULL, as at least on Solaris 8 this seems not always to be
701 done automatically (in contradiction to the requirements of the C
702 standard). This made problems when used from OpenSSH.
705 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
706 dh->length and always used
708 BN_rand_range(priv_key, dh->p).
710 BN_rand_range() is not necessary for Diffie-Hellman, and this
711 specific range makes Diffie-Hellman unnecessarily inefficient if
712 dh->length (recommended exponent length) is much smaller than the
713 length of dh->p. We could use BN_rand_range() if the order of
714 the subgroup was stored in the DH structure, but we only have
719 BN_rand(priv_key, l, ...)
721 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
727 RSA_eay_public_encrypt
728 RSA_eay_private_decrypt
729 RSA_eay_private_encrypt (signing)
730 RSA_eay_public_decrypt (signature verification)
732 (default implementations for RSA_public_encrypt,
733 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
734 always reject numbers >= n.
737 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
738 to synchronize access to 'locking_thread'. This is necessary on
739 systems where access to 'locking_thread' (an 'unsigned long'
740 variable) is not atomic.
743 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
744 *before* setting the 'crypto_lock_rand' flag. The previous code had
745 a race condition if 0 is a valid thread ID.
746 [Travis Vitek <vitek@roguewave.com>]
748 +) Cleanup of EVP macros.
751 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
755 +) Add initial OCSP responder support to ocsp application. The
756 revocation information is handled using the text based index
757 use by the ca application. The responder can either handle
758 requests generated internally, supplied in files (for example
759 via a CGI script) or using an internal minimal server.
762 +) Add configuration choices to get zlib compression for TLS.
765 +) Changes to Kerberos SSL for RFC 2712 compliance:
766 1. Implemented real KerberosWrapper, instead of just using
767 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
768 2. Implemented optional authenticator field of KerberosWrapper.
770 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
771 and authenticator structs; see crypto/krb5/.
773 Generalized Kerberos calls to support multiple Kerberos libraries.
774 [Vern Staats <staatsvr@asc.hpc.mil>,
775 Jeffrey Altman <jaltman@columbia.edu>
778 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
779 already does with RSA. testdsa.h now has 'priv_key/pub_key'
780 values for each of the key sizes rather than having just
781 parameters (and 'speed' generating keys each time).
784 *) Add support for shared libraries under Irix.
785 [Albert Chin-A-Young <china@thewrittenword.com>]
787 *) Add configuration option to build on Linux on both big-endian and
789 [Ralf Baechle <ralf@uni-koblenz.de>]
791 *) Add the possibility to create shared libraries on HP-UX.
794 -) OpenSSL 0.9.6b released [9 July 2001]
796 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
797 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
798 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
799 PRNG state recovery was possible based on the output of
800 one PRNG request appropriately sized to gain knowledge on
801 'md' followed by enough consecutive 1-byte PRNG requests
802 to traverse all of 'state'.
804 1. When updating 'md_local' (the current thread's copy of 'md')
805 during PRNG output generation, hash all of the previous
806 'md_local' value, not just the half used for PRNG output.
808 2. Make the number of bytes from 'state' included into the hash
809 independent from the number of PRNG bytes requested.
811 The first measure alone would be sufficient to avoid
812 Markku-Juhani's attack. (Actually it had never occurred
813 to me that the half of 'md_local' used for chaining was the
814 half from which PRNG output bytes were taken -- I had always
815 assumed that the secret half would be used.) The second
816 measure makes sure that additional data from 'state' is never
817 mixed into 'md_local' in small portions; this heuristically
818 further strengthens the PRNG.
821 +) Speed up EVP routines.
824 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
825 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
826 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
827 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
829 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
830 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
831 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
834 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
836 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
839 *) Fix crypto/bn/asm/mips3.s.
842 *) When only the key is given to "enc", the IV is undefined. Print out
843 an error message in this case.
846 +) Added the OS2-EMX target.
847 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
849 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
850 to support NCONF routines in extension code. New function CONF_set_nconf()
851 to allow functions which take an NCONF to also handle the old LHASH
852 structure: this means that the old CONF compatible routines can be
853 retained (in particular wrt extensions) without having to duplicate the
854 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
857 *) Handle special case when X509_NAME is empty in X509 printing routines.
860 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
861 positive and less than q.
864 +) Enhance the general user interface with mechanisms for inner control
865 and with possibilities to have yes/no kind of prompts.
868 +) Change all calls to low level digest routines in the library and
869 applications to use EVP. Add missing calls to HMAC_cleanup() and
870 don't assume HMAC_CTX can be copied using memcpy().
871 [Verdon Walker <VWalker@novell.com>, Steve Henson]
873 +) Add the possibility to control engines through control names but with
874 arbitrary arguments instead of just a string.
875 Change the key loaders to take a UI_METHOD instead of a callback
876 function pointer. NOTE: this breaks binary compatibility with earlier
877 versions of OpenSSL [engine].
878 Adapt the nCipher code for these new conditions and add a card insertion
882 +) Enhance the general user interface with mechanisms to better support
883 dialog box interfaces, application-defined prompts, the possibility
884 to use defaults (for example default passwords from somewhere else)
885 and interrupts/cancellations.
888 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
889 used: it isn't thread safe and the add_lock_callback should handle
891 [Paul Rose <Paul.Rose@bridge.com>]
893 *) Verify that incoming data obeys the block size in
894 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
897 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
898 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
902 [Ulf Möller, Bodo Möller]
904 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
905 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
906 when fixing the server behaviour for backwards-compatible 'client
907 hello' messages. (Note that the attack is impractical against
908 SSL 3.0 and TLS 1.0 anyway because length and version checking
909 means that the probability of guessing a valid ciphertext is
910 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
913 Before 0.9.5, the countermeasure (hide the error by generating a
914 random 'decryption result') did not work properly because
915 ERR_clear_error() was missing, meaning that SSL_get_error() would
916 detect the supposedly ignored error.
918 Both problems are now fixed.
921 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
922 (previously it was 1024).
925 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
926 tidy up some unnecessarily weird code in 'sk_new()').
927 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
929 +) Change the key loading routines for ENGINEs to use the same kind
930 callback (pem_password_cb) as all other routines that need this
934 *) Fix for compatibility mode trust settings: ignore trust settings
935 unless some valid trust or reject settings are present.
938 *) Fix for blowfish EVP: its a variable length cipher.
941 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
942 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
943 than this minimum value is recommended.
946 +) New random seeder for OpenVMS, using the system process statistics
947 that are easily reachable.
950 +) Windows apparently can't transparently handle global
951 variables defined in DLLs. Initialisations such as:
953 const ASN1_ITEM *it = &ASN1_INTEGER_it;
955 wont compile. This is used by the any applications that need to
956 declare their own ASN1 modules. This was fixed by adding the option
957 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
958 needed for static libraries under Win32.
961 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
962 setting of purpose and trust fields. New X509_STORE trust and
963 purpose functions and tidy up setting in other SSL functions.
966 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
967 structure. These are inherited by X509_STORE_CTX when it is
968 initialised. This allows various defaults to be set in the
969 X509_STORE structure (such as flags for CRL checking and custom
970 purpose or trust settings) for functions which only use X509_STORE_CTX
971 internally such as S/MIME.
973 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
974 trust settings if they are not set in X509_STORE. This allows X509_STORE
975 purposes and trust (in S/MIME for example) to override any set by default.
977 Add command line options for CRL checking to smime, s_client and s_server
981 +) Initial CRL based revocation checking. If the CRL checking flag(s)
982 are set then the CRL is looked up in the X509_STORE structure and
983 its validity and signature checked, then if the certificate is found
984 in the CRL the verify fails with a revoked error.
986 Various new CRL related callbacks added to X509_STORE_CTX structure.
988 Command line options added to 'verify' application to support this.
990 This needs some additional work, such as being able to handle multiple
991 CRLs with different times, extension based lookup (rather than just
992 by subject name) and ultimately more complete V2 CRL extension
996 +) Add a general user interface API (crypto/ui/). This is designed
997 to replace things like des_read_password and friends (backward
998 compatibility functions using this new API are provided).
999 The purpose is to remove prompting functions from the DES code
1000 section as well as provide for prompting through dialog boxes in
1001 a window system and the like.
1004 *) Fix various bugs related to DSA S/MIME verification. Handle missing
1005 parameters in DSA public key structures and return an error in the
1006 DSA routines if parameters are absent.
1009 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
1010 in the current directory if neither $RANDFILE nor $HOME was set.
1011 RAND_file_name() in 0.9.6a returned NULL in this case. This has
1012 caused some confusion to Windows users who haven't defined $HOME.
1013 Thus RAND_file_name() is changed again: e_os.h can define a
1014 DEFAULT_HOME, which will be used if $HOME is not set.
1015 For Windows, we use "C:"; on other platforms, we still require
1016 environment variables.
1018 +) Add "ex_data" support to ENGINE so implementations can add state at a
1019 per-structure level rather than having to store it globally.
1022 +) Make it possible for ENGINE structures to be copied when retrieved by
1023 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1024 This causes the "original" ENGINE structure to act like a template,
1025 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1026 operational state can be localised to each ENGINE structure, despite the
1027 fact they all share the same "methods". New ENGINE structures returned in
1028 this case have no functional references and the return value is the single
1029 structural reference. This matches the single structural reference returned
1030 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1034 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1035 needs to match any other type at all we need to manually clear the
1039 +) Changes to the "openssl engine" utility to include;
1040 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1041 about an ENGINE's available control commands.
1042 - executing control commands from command line arguments using the
1043 '-pre' and '-post' switches. '-post' is only used if '-t' is
1044 specified and the ENGINE is successfully initialised. The syntax for
1045 the individual commands are colon-separated, for example;
1046 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1049 +) New dynamic control command support for ENGINEs. ENGINEs can now
1050 declare their own commands (numbers), names (strings), descriptions,
1051 and input types for run-time discovery by calling applications. A
1052 subset of these commands are implicitly classed as "executable"
1053 depending on their input type, and only these can be invoked through
1054 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1055 can be based on user input, config files, etc). The distinction is
1056 that "executable" commands cannot return anything other than a boolean
1057 result and can only support numeric or string input, whereas some
1058 discoverable commands may only be for direct use through
1059 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1060 pointers, or other custom uses. The "executable" commands are to
1061 support parameterisations of ENGINE behaviour that can be
1062 unambiguously defined by ENGINEs and used consistently across any
1063 OpenSSL-based application. Commands have been added to all the
1064 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1065 control over shared-library paths without source code alterations.
1068 +) Changed all ENGINE implementations to dynamically allocate their
1069 ENGINEs rather than declaring them statically. Apart from this being
1070 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1071 this also allows the implementations to compile without using the
1072 internal engine_int.h header.
1075 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1076 'const' value. Any code that should be able to modify a RAND_METHOD
1077 should already have non-const pointers to it (ie. they should only
1078 modify their own ones).
1081 +) Made a variety of little tweaks to the ENGINE code.
1082 - "atalla" and "ubsec" string definitions were moved from header files
1083 to C code. "nuron" string definitions were placed in variables
1084 rather than hard-coded - allowing parameterisation of these values
1085 later on via ctrl() commands.
1086 - Removed unused "#if 0"'d code.
1087 - Fixed engine list iteration code so it uses ENGINE_free() to release
1088 structural references.
1089 - Constified the RAND_METHOD element of ENGINE structures.
1090 - Constified various get/set functions as appropriate and added
1091 missing functions (including a catch-all ENGINE_cpy that duplicates
1092 all ENGINE values onto a new ENGINE except reference counts/state).
1093 - Removed NULL parameter checks in get/set functions. Setting a method
1094 or function to NULL is a way of cancelling out a previously set
1095 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1096 and doesn't justify the extra error symbols and code.
1097 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1098 flags from engine_int.h to engine.h.
1099 - Changed prototypes for ENGINE handler functions (init(), finish(),
1100 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1103 *) Move 'if (!initialized) RAND_poll()' into regions protected by
1104 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
1105 having multiple threads call RAND_poll() concurrently.
1108 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
1109 combination of a flag and a thread ID variable.
1110 Otherwise while one thread is in ssleay_rand_bytes (which sets the
1111 flag), *other* threads can enter ssleay_add_bytes without obeying
1112 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
1113 that they do not hold after the first thread unsets add_do_not_lock).
1116 +) Implement binary inversion algorithm for BN_mod_inverse in addition
1117 to the algorithm using long division. The binary algorithm can be
1118 used only if the modulus is odd. On 32-bit systems, it is faster
1119 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1120 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1121 up to 450 bits. In 64-bit environments, the binary algorithm
1122 appears to be advantageous for much longer moduli; here we use it
1123 for moduli up to 2048 bits.
1126 *) Change bctest again: '-x' expressions are not available in all
1130 -) OpenSSL 0.9.6a released [5 Apr 2001]
1132 *) Fix a couple of memory leaks in PKCS7_dataDecode()
1133 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
1135 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
1136 the default extension for executables, if any. Also, make the perl
1137 scripts that use symlink() to test if it really exists and use "cp"
1138 if it doesn't. All this made OpenSSL compilable and installable in
1142 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1143 could not support the combine flag in choice fields.
1146 -) Fix for asn1_GetSequence() for indefinite length constructed data.
1147 If SEQUENCE is length is indefinite just set c->slen to the total
1148 amount of data available.
1149 [Steve Henson, reported by shige@FreeBSD.org]
1150 [This change does not apply to 0.9.7.]
1152 *) Change bctest to avoid here-documents inside command substitution
1153 (workaround for FreeBSD /bin/sh bug).
1154 For compatibility with Ultrix, avoid shell functions (introduced
1155 in the bctest version that searches along $PATH).
1158 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
1159 with des_encrypt() defined on some operating systems, like Solaris
1163 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
1164 On the Importance of Eliminating Errors in Cryptographic
1165 Computations, J. Cryptology 14 (2001) 2, 101-119,
1166 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
1169 *) MIPS assembler BIGNUM division bug fix.
1172 *) Disabled incorrect Alpha assembler code.
1175 -) Fix PKCS#7 decode routines so they correctly update the length
1176 after reading an EOC for the EXPLICIT tag.
1178 [This change does not apply to 0.9.7.]
1180 *) Fix bug in PKCS#12 key generation routines. This was triggered
1181 if a 3DES key was generated with a 0 initial byte. Include
1182 PKCS12_BROKEN_KEYGEN compilation option to retain the old
1183 (but broken) behaviour.
1186 *) Enhance bctest to search for a working bc along $PATH and print
1188 [Tim Rice <tim@multitalents.net> via Richard Levitte]
1190 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
1191 extensions from a certificate request to the certificate.
1194 +) Allow multiple 'certopt' and 'nameopt' options to be separated
1195 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1196 file: this allows the display of the certificate about to be
1197 signed to be customised, to allow certain fields to be included
1198 or excluded and extension details. The old system didn't display
1199 multicharacter strings properly, omitted fields not in the policy
1200 and couldn't display additional details such as extensions.
1203 *) Fix memory leaks in err.c: free err_data string if necessary;
1204 don't write to the wrong index in ERR_set_error_data.
1207 +) Function EC_POINTs_mul for multiple scalar multiplication
1208 of an arbitrary number of elliptic curve points
1209 \sum scalars[i]*points[i],
1210 optionally including the generator defined for the EC_GROUP:
1211 scalar*generator + \sum scalars[i]*points[i].
1213 EC_POINT_mul is a simple wrapper function for the typical case
1214 that the point list has just one item (besides the optional
1218 +) First EC_METHODs for curves over GF(p):
1220 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1221 operations and provides various method functions that can also
1222 operate with faster implementations of modular arithmetic.
1224 EC_GFp_mont_method() reuses most functions that are part of
1225 EC_GFp_simple_method, but uses Montgomery arithmetic.
1227 [Bodo Moeller; point addition and point doubling
1228 implementation directly derived from source code provided by
1229 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1231 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1232 crypto/ec/ec_lib.c):
1234 Curves are EC_GROUP objects (with an optional group generator)
1235 based on EC_METHODs that are built into the library.
1237 Points are EC_POINT objects based on EC_GROUP objects.
1239 Most of the framework would be able to handle curves over arbitrary
1240 finite fields, but as there are no obvious types for fields other
1241 than GF(p), some functions are limited to that for now.
1244 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1245 that the file contains a complete HTTP response.
1248 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1249 change the def and num file printf format specifier from "%-40sXXX"
1250 to "%-39s XXX". The latter will always guarantee a space after the
1251 field while the former will cause them to run together if the field
1252 is 40 of more characters long.
1255 +) Constify the cipher and digest 'method' functions and structures
1256 and modify related functions to take constant EVP_MD and EVP_CIPHER
1260 *) Implement ssl23_peek (analogous to ssl23_read), which previously
1264 *) Replace rdtsc with _emit statements for VC++ version 5.
1265 [Jeremy Cooper <jeremy@baymoo.org>]
1267 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1268 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1271 +) Modify EVP_Digest*() routines so they now return values. Although the
1272 internal software routines can never fail additional hardware versions
1276 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1278 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1279 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1282 ERR_R_NESTED_ASN1_ERROR
1284 ERR_R_MISSING_ASN1_EOS
1285 were 4 .. 9, conflicting with
1286 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1288 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1289 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1291 Add new error code 'ERR_R_INTERNAL_ERROR'.
1294 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1298 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1299 sets the subject name for a new request or supersedes the
1300 subject name in a given request. Formats that can be parsed are
1301 'CN=Some Name, OU=myOU, C=IT'
1303 'CN=Some Name/OU=myOU/C=IT'.
1305 Add options '-batch' and '-verbose' to 'openssl req'.
1306 [Massimiliano Pala <madwolf@hackmasters.net>]
1308 *) Make it possible to reuse SSLv2 sessions.
1311 +) Introduce the possibility to access global variables through
1312 functions on platform were that's the best way to handle exporting
1313 global variables in shared libraries. To enable this functionality,
1314 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1315 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1316 is normally done by Configure or something similar).
1318 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1319 in the source file (foo.c) like this:
1321 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1322 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1324 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1325 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1327 OPENSSL_DECLARE_GLOBAL(int,foo);
1328 #define foo OPENSSL_GLOBAL_REF(foo)
1329 OPENSSL_DECLARE_GLOBAL(double,bar);
1330 #define bar OPENSSL_GLOBAL_REF(bar)
1332 The #defines are very important, and therefore so is including the
1333 header file everywhere where the defined globals are used.
1335 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1336 of ASN.1 items, but that structure is a bit different.
1338 The largest change is in util/mkdef.pl which has been enhanced with
1339 better and easier to understand logic to choose which symbols should
1340 go into the Windows .def files as well as a number of fixes and code
1341 cleanup (among others, algorithm keywords are now sorted
1342 lexicographically to avoid constant rewrites).
1345 *) In copy_email() check for >= 0 as a return value for
1346 X509_NAME_get_index_by_NID() since 0 is a valid index.
1347 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1349 +) In BN_div() keep a copy of the sign of 'num' before writing the
1350 result to 'rm' because if rm==num the value will be overwritten
1351 and produce the wrong result if 'num' is negative: this caused
1352 problems with BN_mod() and BN_nnmod().
1355 +) Function OCSP_request_verify(). This checks the signature on an
1356 OCSP request and verifies the signer certificate. The signer
1357 certificate is just checked for a generic purpose and OCSP request
1361 +) Add OCSP_check_validity() function to check the validity of OCSP
1362 responses. OCSP responses are prepared in real time and may only
1363 be a few seconds old. Simply checking that the current time lies
1364 between thisUpdate and nextUpdate max reject otherwise valid responses
1365 caused by either OCSP responder or client clock inaccuracy. Instead
1366 we allow thisUpdate and nextUpdate to fall within a certain period of
1367 the current time. The age of the response can also optionally be
1368 checked. Two new options -validity_period and -status_age added to
1372 +) If signature or public key algorithm is unrecognized print out its
1373 OID rather that just UNKNOWN.
1376 *) Avoid coredump with unsupported or invalid public keys by checking if
1377 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1378 PKCS7_verify() fails with non detached data.
1381 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1382 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1383 ID to be generated from the issuer certificate alone which can then be
1384 passed to OCSP_id_issuer_cmp().
1387 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1388 ASN1 modules to export functions returning ASN1_ITEM pointers
1389 instead of the ASN1_ITEM structures themselves. This adds several
1390 new macros which allow the underlying ASN1 function/structure to
1391 be accessed transparently. As a result code should not use ASN1_ITEM
1392 references directly (such as &X509_it) but instead use the relevant
1393 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1394 use of the new ASN1 code on platforms where exporting structures
1395 is problematical (for example in shared libraries) but exporting
1396 functions returning pointers to structures is not.
1399 +) Add support for overriding the generation of SSL/TLS session IDs.
1400 These callbacks can be registered either in an SSL_CTX or per SSL.
1401 The purpose of this is to allow applications to control, if they wish,
1402 the arbitrary values chosen for use as session IDs, particularly as it
1403 can be useful for session caching in multiple-server environments. A
1404 command-line switch for testing this (and any client code that wishes
1405 to use such a feature) has been added to "s_server".
1406 [Geoff Thorpe, Lutz Jaenicke]
1408 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1409 of the form '#if defined(...) || defined(...) || ...' and
1410 '#if !defined(...) && !defined(...) && ...'. This also avoids
1411 the growing number of special cases it was previously handling.
1414 *) Don't use getenv in library functions when run as setuid/setgid.
1415 New function OPENSSL_issetugid().
1418 +) Make all configuration macros available for application by making
1419 sure they are available in opensslconf.h, by giving them names starting
1420 with "OPENSSL_" to avoid conflicts with other packages and by making
1421 sure e_os2.h will cover all platform-specific cases together with
1423 Additionally, it is now possible to define configuration/platform-
1424 specific names (called "system identities"). In the C code, these
1425 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1426 macro with the name beginning with "OPENSSL_SYS_", which is determined
1427 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1431 +) New option -set_serial to 'req' and 'x509' this allows the serial
1432 number to use to be specified on the command line. Previously self
1433 signed certificates were hard coded with serial number 0 and the
1434 CA options of 'x509' had to use a serial number in a file which was
1438 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1439 due to incorrect handling of multi-threading:
1441 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1443 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1445 3. Count how many times MemCheck_off() has been called so that
1446 nested use can be treated correctly. This also avoids
1447 inband-signalling in the previous code (which relied on the
1448 assumption that thread ID 0 is impossible).
1451 +) New options to 'ca' utility to support V2 CRL entry extensions.
1452 Currently CRL reason, invalidity date and hold instruction are
1453 supported. Add new CRL extensions to V3 code and some new objects.
1456 *) Add "-rand" option also to s_client and s_server.
1459 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1460 disable standard block padding (aka PKCS#5 padding) in the EVP
1461 API, which was previously mandatory. This means that the data is
1462 not padded in any way and so the total length much be a multiple
1463 of the block size, otherwise an error occurs.
1466 +) Initial (incomplete) OCSP SSL support.
1469 *) Fix CPU detection on Irix 6.x.
1470 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1471 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1473 +) New function OCSP_parse_url(). This splits up a URL into its host,
1474 port and path components: primarily to parse OCSP URLs. New -url
1475 option to ocsp utility.
1478 +) New nonce behavior. The return value of OCSP_check_nonce() now
1479 reflects the various checks performed. Applications can decide
1480 whether to tolerate certain situations such as an absent nonce
1481 in a response when one was present in a request: the ocsp application
1482 just prints out a warning. New function OCSP_add1_basic_nonce()
1483 this is to allow responders to include a nonce in a response even if
1484 the request is nonce-less.
1487 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1490 [This change does not apply to 0.9.7.]
1492 *) Use the cached encoding of an X509_NAME structure rather than
1493 copying it. This is apparently the reason for the libsafe "errors"
1494 but the code is actually correct.
1497 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1498 skipped when using openssl x509 multiple times on a single input file,
1499 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1502 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1503 set string type: to handle setting ASN1_TIME structures. Fix ca
1504 utility to correctly initialize revocation date of CRLs.
1507 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1508 the clients preferred ciphersuites and rather use its own preferences.
1509 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1510 Internet Explorer by ensuring unchanged hash method during stepup.
1511 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1514 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1515 to aes and add a new 'exist' option to print out symbols that don't
1519 +) Additional options to ocsp utility to allow flags to be set and
1520 additional certificates supplied.
1523 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1524 OCSP client a number of certificate to only verify the response
1528 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1529 Bleichenbacher's DSA attack.
1530 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1531 to be set and top=0 forces the highest bit to be set; top=-1 is new
1532 and leaves the highest bit random.
1533 [Ulf Moeller, Bodo Moeller]
1535 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1536 handle the new API. Currently only ECB, CBC modes supported. Add new
1537 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1538 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1539 [Ben Laurie, Steve Henson]
1541 *) In the NCONF_...-based implementations for CONF_... queries
1542 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1543 a temporary CONF structure with the data component set to NULL
1544 (which gives segmentation faults in lh_retrieve).
1545 Instead, use NULL for the CONF pointer in CONF_get_string and
1546 CONF_get_number (which may use environment variables) and directly
1547 return NULL from CONF_get_section.
1550 *) Fix potential buffer overrun for EBCDIC.
1553 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1554 request to response.
1557 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1558 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1559 extract information from a certificate request. OCSP_response_create()
1560 creates a response and optionally adds a basic response structure.
1561 OCSP_basic_add1_status() adds a complete single response to a basic
1562 response and returns the OCSP_SINGLERESP structure just added (to allow
1563 extensions to be included for example). OCSP_basic_add1_cert() adds a
1564 certificate to a basic response and OCSP_basic_sign() signs a basic
1565 response with various flags. New helper functions ASN1_TIME_check()
1566 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1567 (converts ASN1_TIME to GeneralizedTime).
1570 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1571 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1572 structure from a certificate. X509_pubkey_digest() digests the public_key
1573 contents: this is used in various key identifiers.
1576 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1577 keyUsage if basicConstraints absent for a CA.
1580 *) Make SMIME_write_PKCS7() write mail header values with a format that
1581 is more generally accepted (no spaces before the semicolon), since
1582 some programs can't parse those values properly otherwise. Also make
1583 sure BIO's that break lines after each write do not create invalid
1587 +) Make sk_sort() tolerate a NULL argument.
1588 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1590 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1591 passed by the function are trusted implicitly. If any of them signed the
1592 response then it is assumed to be valid and is not verified.
1595 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1596 macros previously used would not encode an empty SEQUENCE OF
1597 and break the signature.
1599 [This change does not apply to 0.9.7.]
1601 *) Zero the premaster secret after deriving the master secret in
1605 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1606 to data. This was previously part of the PKCS7 ASN1 code. This
1607 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1608 [Steve Henson, reported by Kenneth R. Robinette
1609 <support@securenetterm.com>]
1611 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1612 routines: without these tracing memory leaks is very painful.
1613 Fix leaks in PKCS12 and PKCS7 routines.
1616 *) Add some EVP_add_digest_alias registrations (as found in
1617 OpenSSL_add_all_digests()) to SSL_library_init()
1618 aka OpenSSL_add_ssl_algorithms(). This provides improved
1619 compatibility with peers using X.509 certificates
1620 with unconventional AlgorithmIdentifier OIDs.
1623 *) Fix for Irix with NO_ASM.
1624 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1626 *) ./config script fixes.
1627 [Ulf Moeller, Richard Levitte]
1629 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1630 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1631 effectively meant GeneralizedTime would never be used. Now it
1632 is initialised to -1 but X509_time_adj() now has to check the value
1633 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1634 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1635 [Steve Henson, reported by Kenneth R. Robinette
1636 <support@securenetterm.com>]
1638 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1639 result in a zero length in the ASN1_INTEGER structure which was
1640 not consistent with the structure when d2i_ASN1_INTEGER() was used
1641 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1642 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1643 where it did not print out a minus for negative ASN1_INTEGER.
1646 *) Fix 'openssl passwd -1'.
1649 +) Add summary printout to ocsp utility. The various functions which
1650 convert status values to strings have been renamed to:
1651 OCSP_response_status_str(), OCSP_cert_status_str() and
1652 OCSP_crl_reason_str() and are no longer static. New options
1653 to verify nonce values and to disable verification. OCSP response
1654 printout format cleaned up.
1657 +) Add additional OCSP certificate checks. These are those specified
1658 in RFC2560. This consists of two separate checks: the CA of the
1659 certificate being checked must either be the OCSP signer certificate
1660 or the issuer of the OCSP signer certificate. In the latter case the
1661 OCSP signer certificate must contain the OCSP signing extended key
1662 usage. This check is performed by attempting to match the OCSP
1663 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1664 in the OCSP_CERTID structures of the response.
1667 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1668 and related routines. This uses the standard OpenSSL certificate
1669 verify routines to perform initial checks (just CA validity) and
1670 to obtain the certificate chain. Then additional checks will be
1671 performed on the chain. Currently the root CA is checked to see
1672 if it is explicitly trusted for OCSP signing. This is used to set
1673 a root CA as a global signing root: that is any certificate that
1674 chains to that CA is an acceptable OCSP signing certificate.
1677 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1678 extensions from a separate configuration file.
1679 As when reading extensions from the main configuration file,
1680 the '-extensions ...' option may be used for specifying the
1682 [Massimiliano Pala <madwolf@comune.modena.it>]
1684 *) Change PKCS12_key_gen_asc() so it can cope with non null
1685 terminated strings whose length is passed in the passlen
1686 parameter, for example from PEM callbacks. This was done
1687 by adding an extra length parameter to asc2uni().
1688 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1690 +) New OCSP utility. Allows OCSP requests to be generated or
1691 read. The request can be sent to a responder and the output
1692 parsed, outputed or printed in text form. Not complete yet:
1693 still needs to check the OCSP response validity.
1696 +) New subcommands for 'openssl ca':
1697 'openssl ca -status <serial>' prints the status of the cert with
1698 the given serial number (according to the index file).
1699 'openssl ca -updatedb' updates the expiry status of certificates
1701 [Massimiliano Pala <madwolf@comune.modena.it>]
1703 +) New '-newreq-nodes' command option to CA.pl. This is like
1704 '-newreq', but calls 'openssl req' with the '-nodes' option
1705 so that the resulting key is not encrypted.
1706 [Damien Miller <djm@mindrot.org>]
1708 +) New configuration for the GNU Hurd.
1709 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1711 +) Initial code to implement OCSP basic response verify. This
1712 is currently incomplete. Currently just finds the signer's
1713 certificate and verifies the signature on the response.
1716 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1717 value of OPENSSLDIR. This is available via the new '-d' option
1718 to 'openssl version', and is also included in 'openssl version -a'.
1721 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1722 call failed, free the DSA structure.
1725 +) Allowing defining memory allocation callbacks that will be given
1726 file name and line number information in additional arguments
1727 (a const char* and an int). The basic functionality remains, as
1728 well as the original possibility to just replace malloc(),
1729 realloc() and free() by functions that do not know about these
1730 additional arguments. To register and find out the current
1731 settings for extended allocation functions, the following
1732 functions are provided:
1734 CRYPTO_set_mem_ex_functions
1735 CRYPTO_set_locked_mem_ex_functions
1736 CRYPTO_get_mem_ex_functions
1737 CRYPTO_get_locked_mem_ex_functions
1739 These work the same way as CRYPTO_set_mem_functions and friends.
1740 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1741 extended allocation function is enabled.
1742 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1743 a conventional allocation function is enabled.
1744 [Richard Levitte, Bodo Moeller]
1746 *) Fix to uni2asc() to cope with zero length Unicode strings.
1747 These are present in some PKCS#12 files.
1750 +) Finish off removing the remaining LHASH function pointer casts.
1751 There should no longer be any prototype-casting required when using
1752 the LHASH abstraction, and any casts that remain are "bugs". See
1753 the callback types and macros at the head of lhash.h for details
1754 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1757 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1758 If /dev/[u]random devices are not available or do not return enough
1759 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1761 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1762 /etc/entropy will be queried once each in this sequence, quering stops
1763 when enough entropy was collected without querying more sockets.
1766 +) Change the Unix RAND_poll() variant to be able to poll several
1767 random devices, as specified by DEVRANDOM, until a sufficient amount
1768 of data has been collected. We spend at most 10 ms on each file
1769 (select timeout) and read in non-blocking mode. DEVRANDOM now
1770 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1771 (previously it was just the string "/dev/urandom"), so on typical
1772 platforms the 10 ms delay will never occur.
1773 Also separate out the Unix variant to its own file, rand_unix.c.
1774 For VMS, there's a currently-empty rand_vms.c.
1777 +) Move OCSP client related routines to ocsp_cl.c. These
1778 provide utility functions which an application needing
1779 to issue a request to an OCSP responder and analyse the
1780 response will typically need: as opposed to those which an
1781 OCSP responder itself would need which will be added later.
1783 OCSP_request_sign() signs an OCSP request with an API similar
1784 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1785 response. OCSP_response_get1_basic() extracts basic response
1786 from response. OCSP_resp_find_status(): finds and extracts status
1787 information from an OCSP_CERTID structure (which will be created
1788 when the request structure is built). These are built from lower
1789 level functions which work on OCSP_SINGLERESP structures but
1790 wont normally be used unless the application wishes to examine
1791 extensions in the OCSP response for example.
1793 Replace nonce routines with a pair of functions.
1794 OCSP_request_add1_nonce() adds a nonce value and optionally
1795 generates a random value. OCSP_check_nonce() checks the
1796 validity of the nonce in an OCSP response.
1799 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1800 This doesn't copy the supplied OCSP_CERTID and avoids the
1801 need to free up the newly created id. Change return type
1802 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1803 This can then be used to add extensions to the request.
1804 Deleted OCSP_request_new(), since most of its functionality
1805 is now in OCSP_REQUEST_new() (and the case insensitive name
1806 clash) apart from the ability to set the request name which
1807 will be added elsewhere.
1810 +) Update OCSP API. Remove obsolete extensions argument from
1811 various functions. Extensions are now handled using the new
1812 OCSP extension code. New simple OCSP HTTP function which
1813 can be used to send requests and parse the response.
1816 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1817 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1818 uses the special reorder version of SET OF to sort the attributes
1819 and reorder them to match the encoded order. This resolves a long
1820 standing problem: a verify on a PKCS7 structure just after signing
1821 it used to fail because the attribute order did not match the
1822 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1823 it uses the received order. This is necessary to tolerate some broken
1824 software that does not order SET OF. This is handled by encoding
1825 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1826 to produce the required SET OF.
1829 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1830 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1831 files to get correct declarations of the ASN.1 item variables.
1834 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1835 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1836 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1837 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1838 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1839 ASN1_ITEM and no wrapper functions.
1842 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1843 replace the old function pointer based I/O routines. Change most of
1844 the *_d2i_bio() and *_d2i_fp() functions to use these.
1847 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1848 lines, recognice more "algorithms" that can be deselected, and make
1849 it complain about algorithm deselection that isn't recognised.
1852 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1853 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1854 to use new functions. Add NO_ASN1_OLD which can be set to remove
1855 some old style ASN1 functions: this can be used to determine if old
1856 code will still work when these eventually go away.
1859 +) New extension functions for OCSP structures, these follow the
1860 same conventions as certificates and CRLs.
1863 +) New function X509V3_add1_i2d(). This automatically encodes and
1864 adds an extension. Its behaviour can be customised with various
1865 flags to append, replace or delete. Various wrappers added for
1866 certifcates and CRLs.
1869 +) Fix to avoid calling the underlying ASN1 print routine when
1870 an extension cannot be parsed. Correct a typo in the
1871 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1874 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1875 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1876 when writing a 32767 byte record.
1877 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1879 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1880 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1882 (RSA objects have a reference count access to which is protected
1883 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1884 so they are meant to be shared between threads.)
1885 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1886 "Reddie, Steven" <Steven.Reddie@ca.com>]
1888 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1889 entries for variables.
1892 *) Fix a deadlock in CRYPTO_mem_leaks().
1895 +) Add functionality to apps/openssl.c for detecting locking
1896 problems: As the program is single-threaded, all we have
1897 to do is register a locking callback using an array for
1898 storing which locks are currently held by the program.
1901 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1902 SSL_get_ex_data_X509_STORE_idx(), which is used in
1903 ssl_verify_cert_chain() and thus can be called at any time
1904 during TLS/SSL handshakes so that thread-safety is essential.
1905 Unfortunately, the ex_data design is not at all suited
1906 for multi-threaded use, so it probably should be abolished.
1909 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1910 [Broadcom, tweaked and integrated by Geoff Thorpe]
1912 +) Move common extension printing code to new function
1913 X509V3_print_extensions(). Reorganise OCSP print routines and
1914 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1917 +) New function X509_signature_print() to remove duplication in some
1921 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1922 set (this was treated exactly the same as SET OF previously). This
1923 is used to reorder the STACK representing the structure to match the
1924 encoding. This will be used to get round a problem where a PKCS7
1925 structure which was signed could not be verified because the STACK
1926 order did not reflect the encoded order.
1929 +) Reimplement the OCSP ASN1 module using the new code.
1932 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1933 for its ASN1 operations. The old style function pointers still exist
1934 for now but they will eventually go away.
1937 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1938 completely replaces the old ASN1 functionality with a table driven
1939 encoder and decoder which interprets an ASN1_ITEM structure describing
1940 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1941 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1942 has also been converted to the new form.
1945 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1946 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1947 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1948 for negative moduli.
1951 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1952 of not touching the result's sign bit.
1955 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1959 +) Changed the LHASH code to use prototypes for callbacks, and created
1960 macros to declare and implement thin (optionally static) functions
1961 that provide type-safety and avoid function pointer casting for the
1962 type-specific callbacks.
1965 *) Use better test patterns in bntest.
1968 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1970 [Veers Staats <staatsvr@asc.hpc.mil>,
1971 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1973 *) rand_win.c fix for Borland C.
1976 *) BN_rshift bugfix for n == 0.
1979 +) Reformat the FAQ so the different questions and answers can be divided
1980 in sections depending on the subject.
1983 +) Have the zlib compression code load ZLIB.DLL dynamically under
1987 +) New function BN_mod_sqrt for computing square roots modulo a prime
1988 (using the probabilistic Tonelli-Shanks algorithm unless
1989 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1990 be handled deterministically).
1991 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1993 *) Add a 'bctest' script that checks for some known 'bc' bugs
1994 so that 'make test' does not abort just because 'bc' is broken.
1997 *) Store verify_result within SSL_SESSION also for client side to
1998 avoid potential security hole. (Re-used sessions on the client side
1999 always resulted in verify_result==X509_V_OK, not using the original
2000 result of the server certificate verification.)
2003 +) Make BN_mod_inverse faster by explicitly handling small quotients
2004 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2005 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2008 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2009 SSL3_RT_APPLICATION_DATA, return 0.
2010 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2014 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2015 releases, have been re-implemented by renaming the previous
2016 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2017 and ssl3_read_internal, respectively, and adding 'peek' parameters
2018 to them. The new ssl[23]_{read,peek} functions are calls to
2019 ssl[23]_read_internal with the 'peek' flag set appropriately.
2020 A 'peek' parameter has also been added to ssl3_read_bytes, which
2021 does the actual work for ssl3_read_internal.
2024 +) New function BN_kronecker.
2027 +) Fix BN_gcd so that it works on negative inputs; the result is
2028 positive unless both parameters are zero.
2029 Previously something reasonably close to an infinite loop was
2030 possible because numbers could be growing instead of shrinking
2031 in the implementation of Euclid's algorithm.
2034 +) Fix BN_is_word() and BN_is_one() macros to take into account the
2035 sign of the number in question.
2037 Fix BN_is_word(a,w) to work correctly for w == 0.
2039 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2040 because its test if the absolute value of 'a' equals 'w'.
2041 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2042 it exists mostly for use in the implementations of BN_is_zero(),
2043 BN_is_one(), and BN_is_word().
2046 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2047 the method-specific "init()" handler. Also clean up ex_data after
2048 calling the method-specific "finish()" handler. Previously, this was
2049 happening the other way round.
2052 +) New function BN_swap.
2055 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2056 the exponentiation functions are more likely to produce reasonable
2057 results on negative inputs.
2060 +) Change BN_mod_mul so that the result is always non-negative.
2061 Previously, it could be negative if one of the factors was negative;
2062 I don't think anyone really wanted that behaviour.
2065 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2066 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2067 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2068 and add new functions:
2077 BN_mod_lshift1_quick
2081 These functions always generate non-negative results.
2083 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2084 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2086 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2087 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2088 be reduced modulo m.
2089 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2091 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2092 was actually never needed) and in BN_mul(). The removal in BN_mul()
2093 required a small change in bn_mul_part_recursive() and the addition
2094 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2095 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2096 bn_sub_words() and bn_add_words() except they take arrays with
2100 +) In 'openssl passwd', verify passwords read from the terminal
2101 unless the '-salt' option is used (which usually means that
2102 verification would just waste user's time since the resulting
2103 hash is going to be compared with some given password hash)
2104 or the new '-noverify' option is used.
2106 This is an incompatible change, but it does not affect
2107 non-interactive use of 'openssl passwd' (passwords on the command
2108 line, '-stdin' option, '-in ...' option) and thus should not
2112 +) Remove all references to RSAref, since there's no more need for it.
2115 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2116 The previous value, 12, was not always sufficient for BN_mod_exp().
2119 +) Make DSO load along a path given through an environment variable
2120 (SHLIB_PATH) with shl_load().
2123 +) Constify the ENGINE code as a result of BIGNUM constification.
2124 Also constify the RSA code and most things related to it. In a
2125 few places, most notable in the depth of the ASN.1 code, ugly
2126 casts back to non-const were required (to be solved at a later
2130 +) Make it so the openssl application has all engines loaded by default.
2133 +) Constify the BIGNUM routines a little more.
2136 *) Make sure that shared libraries get the internal name engine with
2137 the full version number and not just 0. This should mark the
2138 shared libraries as not backward compatible. Of course, this should
2139 be changed again when we can guarantee backward binary compatibility.
2142 +) Add the following functions:
2144 ENGINE_load_cswift()
2146 ENGINE_load_atalla()
2148 ENGINE_load_builtin_engines()
2150 That way, an application can itself choose if external engines that
2151 are built-in in OpenSSL shall ever be used or not. The benefit is
2152 that applications won't have to be linked with libdl or other dso
2153 libraries unless it's really needed.
2155 Changed 'openssl engine' to load all engines on demand.
2156 Changed the engine header files to avoid the duplication of some
2157 declarations (they differed!).
2160 +) 'openssl engine' can now list capabilities.
2163 +) Better error reporting in 'openssl engine'.
2166 +) Never call load_dh_param(NULL) in s_server.
2169 +) Add engine application. It can currently list engines by name and
2170 identity, and test if they are actually available.
2173 +) Improve RPM specification file by forcing symbolic linking and making
2174 sure the installed documentation is also owned by root.root.
2175 [Damien Miller <djm@mindrot.org>]
2177 +) Give the OpenSSL applications more possibilities to make use of
2178 keys (public as well as private) handled by engines.
2181 +) Add OCSP code that comes from CertCo.
2184 +) Add VMS support for the Rijndael code.
2187 +) Added untested support for Nuron crypto accelerator.
2190 +) Add support for external cryptographic devices. This code was
2191 previously distributed separately as the "engine" branch.
2192 [Geoff Thorpe, Richard Levitte]
2194 +) Rework the filename-translation in the DSO code. It is now possible to
2195 have far greater control over how a "name" is turned into a filename
2196 depending on the operating environment and any oddities about the
2197 different shared library filenames on each system.
2200 +) Support threads on FreeBSD-elf in Configure.
2203 +) Fix for SHA1 assembly problem with MASM: it produces
2204 warnings about corrupt line number information when assembling
2205 with debugging information. This is caused by the overlapping
2207 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2210 NCONF_get_number() has no error checking at all. As a replacement,
2211 NCONF_get_number_e() is defined (_e for "error checking") and is
2212 promoted strongly. The old NCONF_get_number is kept around for
2213 binary backward compatibility.
2214 Make it possible for methods to load from something other than a BIO,
2215 by providing a function pointer that is given a name instead of a BIO.
2216 For example, this could be used to load configuration data from an
2220 *) Fix typo in get_cert_by_subject() in by_dir.c
2221 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2223 *) Rework the system to generate shared libraries:
2225 - Make note of the expected extension for the shared libraries and
2226 if there is a need for symbolic links from for example libcrypto.so.0
2227 to libcrypto.so.0.9.7. There is extended info in Configure for
2230 - Make as few rebuilds of the shared libraries as possible.
2232 - Still avoid linking the OpenSSL programs with the shared libraries.
2234 - When installing, install the shared libraries separately from the
2238 +) Fix for non blocking accept BIOs. Added new I/O special reason
2239 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2240 with non blocking I/O was not possible because no retry code was
2241 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2245 +) Added the beginnings of Rijndael support.
2248 +) Fix for bug in DirectoryString mask setting. Add support for
2249 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2250 to allow certificate printing to more controllable, additional
2251 'certopt' option to 'x509' to allow new printing options to be
2255 +) Clean old EAY MD5 hack from e_os.h.
2258 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2260 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2261 and not in SSL_clear because the latter is also used by the
2262 accept/connect functions; previously, the settings made by
2263 SSL_set_read_ahead would be lost during the handshake.
2264 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2266 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2267 Previously, it would create entries for disableed algorithms no
2271 *) Added several new manual pages for SSL_* function.
2274 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2276 *) In ssl23_get_client_hello, generate an error message when faced
2277 with an initial SSL 3.0/TLS record that is too small to contain the
2278 first two bytes of the ClientHello message, i.e. client_version.
2279 (Note that this is a pathologic case that probably has never happened
2280 in real life.) The previous approach was to use the version number
2281 from the record header as a substitute; but our protocol choice
2282 should not depend on that one because it is not authenticated
2283 by the Finished messages.
2286 *) More robust randomness gathering functions for Windows.
2287 [Jeffrey Altman <jaltman@columbia.edu>]
2289 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2290 not set then we don't setup the error code for issuer check errors
2291 to avoid possibly overwriting other errors which the callback does
2292 handle. If an application does set the flag then we assume it knows
2293 what it is doing and can handle the new informational codes
2297 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2298 a general "ANY" type, as such it should be able to decode anything
2299 including tagged types. However it didn't check the class so it would
2300 wrongly interpret tagged types in the same way as their universal
2301 counterpart and unknown types were just rejected. Changed so that the
2302 tagged and unknown types are handled in the same way as a SEQUENCE:
2303 that is the encoding is stored intact. There is also a new type
2304 "V_ASN1_OTHER" which is used when the class is not universal, in this
2305 case we have no idea what the actual type is so we just lump them all
2309 *) On VMS, stdout may very well lead to a file that is written to
2310 in a record-oriented fashion. That means that every write() will
2311 write a separate record, which will be read separately by the
2312 programs trying to read from it. This can be very confusing.
2314 The solution is to put a BIO filter in the way that will buffer
2315 text until a linefeed is reached, and then write everything a
2316 line at a time, so every record written will be an actual line,
2317 not chunks of lines and not (usually doesn't happen, but I've
2318 seen it once) several lines in one record. BIO_f_linebuffer() is
2321 Currently, it's a VMS-only method, because that's where it has
2322 been tested well enough.
2325 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2326 it can return incorrect results.
2327 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2328 but it was in 0.9.6-beta[12].)
2331 *) Disable the check for content being present when verifying detached
2332 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2333 include zero length content when signing messages.
2336 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2337 BIO_ctrl (for BIO pairs).
2340 *) Add DSO method for VMS.
2343 *) Bug fix: Montgomery multiplication could produce results with the
2347 *) Add RPM specification openssl.spec and modify it to build three
2348 packages. The default package contains applications, application
2349 documentation and run-time libraries. The devel package contains
2350 include files, static libraries and function documentation. The
2351 doc package contains the contents of the doc directory. The original
2352 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2355 *) Add a large number of documentation files for many SSL routines.
2356 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2358 *) Add a configuration entry for Sony News 4.
2359 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2361 *) Don't set the two most significant bits to one when generating a
2362 random number < q in the DSA library.
2365 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2366 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2367 the underlying transport is blocking) if a handshake took place.
2368 (The default behaviour is needed by applications such as s_client
2369 and s_server that use select() to determine when to use SSL_read;
2370 but for applications that know in advance when to expect data, it
2371 just makes things more complicated.)
2374 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2378 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2379 work better on such systems.
2380 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2382 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2383 Update PKCS12_parse() so it copies the friendlyName and the
2384 keyid to the certificates aux info.
2387 *) Fix bug in PKCS7_verify() which caused an infinite loop
2388 if there was more than one signature.
2389 [Sven Uszpelkat <su@celocom.de>]
2391 *) Major change in util/mkdef.pl to include extra information
2392 about each symbol, as well as presentig variables as well
2393 as functions. This change means that there's n more need
2394 to rebuild the .num files when some algorithms are excluded.
2397 *) Allow the verify time to be set by an application,
2398 rather than always using the current time.
2401 *) Phase 2 verify code reorganisation. The certificate
2402 verify code now looks up an issuer certificate by a
2403 number of criteria: subject name, authority key id
2404 and key usage. It also verifies self signed certificates
2405 by the same criteria. The main comparison function is
2406 X509_check_issued() which performs these checks.
2408 Lot of changes were necessary in order to support this
2409 without completely rewriting the lookup code.
2411 Authority and subject key identifier are now cached.
2413 The LHASH 'certs' is X509_STORE has now been replaced
2414 by a STACK_OF(X509_OBJECT). This is mainly because an
2415 LHASH can't store or retrieve multiple objects with
2416 the same hash value.
2418 As a result various functions (which were all internal
2419 use only) have changed to handle the new X509_STORE
2420 structure. This will break anything that messed round
2421 with X509_STORE internally.
2423 The functions X509_STORE_add_cert() now checks for an
2424 exact match, rather than just subject name.
2426 The X509_STORE API doesn't directly support the retrieval
2427 of multiple certificates matching a given criteria, however
2428 this can be worked round by performing a lookup first
2429 (which will fill the cache with candidate certificates)
2430 and then examining the cache for matches. This is probably
2431 the best we can do without throwing out X509_LOOKUP
2432 entirely (maybe later...).
2434 The X509_VERIFY_CTX structure has been enhanced considerably.
2436 All certificate lookup operations now go via a get_issuer()
2437 callback. Although this currently uses an X509_STORE it
2438 can be replaced by custom lookups. This is a simple way
2439 to bypass the X509_STORE hackery necessary to make this
2440 work and makes it possible to use more efficient techniques
2441 in future. A very simple version which uses a simple
2442 STACK for its trusted certificate store is also provided
2443 using X509_STORE_CTX_trusted_stack().
2445 The verify_cb() and verify() callbacks now have equivalents
2446 in the X509_STORE_CTX structure.
2448 X509_STORE_CTX also has a 'flags' field which can be used
2449 to customise the verify behaviour.
2452 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2453 excludes S/MIME capabilities.
2456 *) When a certificate request is read in keep a copy of the
2457 original encoding of the signed data and use it when outputing
2458 again. Signatures then use the original encoding rather than
2459 a decoded, encoded version which may cause problems if the
2460 request is improperly encoded.
2463 *) For consistency with other BIO_puts implementations, call
2464 buffer_write(b, ...) directly in buffer_puts instead of calling
2467 In BIO_puts, increment b->num_write as in BIO_write.
2468 [Peter.Sylvester@EdelWeb.fr]
2470 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2471 BN_zero, we may not return a BIGNUM with an array consisting of
2475 *) Avoid calling abort() from within the library when problems are
2476 detected, except if preprocessor symbols have been defined
2477 (such as REF_CHECK, BN_DEBUG etc.).
2480 *) New openssl application 'rsautl'. This utility can be
2481 used for low level RSA operations. DER public key
2482 BIO/fp routines also added.
2485 *) New Configure entry and patches for compiling on QNX 4.
2486 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2488 *) A demo state-machine implementation was sponsored by
2489 Nuron (http://www.nuron.com/) and is now available in
2490 demos/state_machine.
2493 *) New options added to the 'dgst' utility for signature
2494 generation and verification.
2497 *) Unrecognized PKCS#7 content types are now handled via a
2498 catch all ASN1_TYPE structure. This allows unsupported
2499 types to be stored as a "blob" and an application can
2500 encode and decode it manually.
2503 *) Fix various signed/unsigned issues to make a_strex.c
2505 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2507 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2508 length if passed a buffer. ASN1_INTEGER_to_BN failed
2509 if passed a NULL BN and its argument was negative.
2510 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2512 *) Modification to PKCS#7 encoding routines to output definite
2513 length encoding. Since currently the whole structures are in
2514 memory there's not real point in using indefinite length
2515 constructed encoding. However if OpenSSL is compiled with
2516 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2519 *) Added BIO_vprintf() and BIO_vsnprintf().
2522 *) Added more prefixes to parse for in the the strings written
2523 through a logging bio, to cover all the levels that are available
2524 through syslog. The prefixes are now:
2526 PANIC, EMERG, EMR => LOG_EMERG
2527 ALERT, ALR => LOG_ALERT
2528 CRIT, CRI => LOG_CRIT
2529 ERROR, ERR => LOG_ERR
2530 WARNING, WARN, WAR => LOG_WARNING
2531 NOTICE, NOTE, NOT => LOG_NOTICE
2532 INFO, INF => LOG_INFO
2533 DEBUG, DBG => LOG_DEBUG
2535 and as before, if none of those prefixes are present at the
2536 beginning of the string, LOG_ERR is chosen.
2538 On Win32, the LOG_* levels are mapped according to this:
2540 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2541 LOG_WARNING => EVENTLOG_WARNING_TYPE
2542 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2546 *) Made it possible to reconfigure with just the configuration
2547 argument "reconf" or "reconfigure". The command line arguments
2548 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2549 and are retrieved from there when reconfiguring.
2553 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2555 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2558 *) The obj_dat.pl script was messing up the sorting of object
2559 names. The reason was that it compared the quoted version
2560 of strings as a result "OCSP" > "OCSP Signing" because
2561 " > SPACE. Changed script to store unquoted versions of
2562 names and add quotes on output. It was also omitting some
2563 names from the lookup table if they were given a default
2564 value (that is if SN is missing it is given the same
2565 value as LN and vice versa), these are now added on the
2566 grounds that if an object has a name we should be able to
2567 look it up. Finally added warning output when duplicate
2568 short or long names are found.
2571 *) Changes needed for Tandem NSK.
2572 [Scott Uroff <scott@xypro.com>]
2574 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2575 RSA_padding_check_SSLv23(), special padding was never detected
2576 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2577 version rollback attacks was not effective.
2579 In s23_clnt.c, don't use special rollback-attack detection padding
2580 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2581 client; similarly, in s23_srvr.c, don't do the rollback check if
2582 SSL 2.0 is the only protocol enabled in the server.
2585 *) Make it possible to get hexdumps of unprintable data with 'openssl
2586 asn1parse'. By implication, the functions ASN1_parse_dump() and
2587 BIO_dump_indent() are added.
2590 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2591 these print out strings and name structures based on various
2592 flags including RFC2253 support and proper handling of
2593 multibyte characters. Added options to the 'x509' utility
2594 to allow the various flags to be set.
2597 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2598 Also change the functions X509_cmp_current_time() and
2599 X509_gmtime_adj() work with an ASN1_TIME structure,
2600 this will enable certificates using GeneralizedTime in validity
2601 dates to be checked.
2604 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2605 negative public key encodings) on by default,
2606 NO_NEG_PUBKEY_BUG can be set to disable it.
2609 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2610 content octets. An i2c_ASN1_OBJECT is unnecessary because
2611 the encoding can be trivially obtained from the structure.
2614 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2615 not read locks (CRYPTO_r_[un]lock).
2618 *) A first attempt at creating official support for shared
2619 libraries through configuration. I've kept it so the
2620 default is static libraries only, and the OpenSSL programs
2621 are always statically linked for now, but there are
2622 preparations for dynamic linking in place.
2623 This has been tested on Linux and Tru64.
2626 *) Randomness polling function for Win9x, as described in:
2627 Peter Gutmann, Software Generation of Practically Strong
2631 *) Fix so PRNG is seeded in req if using an already existing
2635 *) New options to smime application. -inform and -outform
2636 allow alternative formats for the S/MIME message including
2637 PEM and DER. The -content option allows the content to be
2638 specified separately. This should allow things like Netscape
2639 form signing output easier to verify.
2642 *) Fix the ASN1 encoding of tags using the 'long form'.
2645 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2646 STRING types. These convert content octets to and from the
2647 underlying type. The actual tag and length octets are
2648 already assumed to have been read in and checked. These
2649 are needed because all other string types have virtually
2650 identical handling apart from the tag. By having versions
2651 of the ASN1 functions that just operate on content octets
2652 IMPLICIT tagging can be handled properly. It also allows
2653 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2654 and ASN1_INTEGER are identical apart from the tag.
2657 *) Change the handling of OID objects as follows:
2659 - New object identifiers are inserted in objects.txt, following
2660 the syntax given in objects.README.
2661 - objects.pl is used to process obj_mac.num and create a new
2663 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2666 This is currently kind of a hack, and the perl code in objects.pl
2667 isn't very elegant, but it works as I intended. The simplest way
2668 to check that it worked correctly is to look in obj_dat.h and
2669 check the array nid_objs and make sure the objects haven't moved
2670 around (this is important!). Additions are OK, as well as
2671 consistent name changes.
2674 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2677 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2678 The given file adds to whatever has already been seeded into the
2679 random pool through the RANDFILE configuration file option or
2680 environment variable, or the default random state file.
2683 *) mkstack.pl now sorts each macro group into lexical order.
2684 Previously the output order depended on the order the files
2685 appeared in the directory, resulting in needless rewriting
2689 *) Patches to make OpenSSL compile under Win32 again. Mostly
2690 work arounds for the VC++ problem that it treats func() as
2691 func(void). Also stripped out the parts of mkdef.pl that
2692 added extra typesafe functions: these no longer exist.
2695 *) Reorganisation of the stack code. The macros are now all
2696 collected in safestack.h . Each macro is defined in terms of
2697 a "stack macro" of the form SKM_<name>(type, a, b). The
2698 DEBUG_SAFESTACK is now handled in terms of function casts,
2699 this has the advantage of retaining type safety without the
2700 use of additional functions. If DEBUG_SAFESTACK is not defined
2701 then the non typesafe macros are used instead. Also modified the
2702 mkstack.pl script to handle the new form. Needs testing to see
2703 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2704 the default if no major problems. Similar behaviour for ASN1_SET_OF
2705 and PKCS12_STACK_OF.
2708 *) When some versions of IIS use the 'NET' form of private key the
2709 key derivation algorithm is different. Normally MD5(password) is
2710 used as a 128 bit RC4 key. In the modified case
2711 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2712 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2713 as the old Netscape_RSA functions except they have an additional
2714 'sgckey' parameter which uses the modified algorithm. Also added
2715 an -sgckey command line option to the rsa utility. Thanks to
2716 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2717 algorithm to openssl-dev.
2720 *) The evp_local.h macros were using 'c.##kname' which resulted in
2721 invalid expansion on some systems (SCO 5.0.5 for example).
2722 Corrected to 'c.kname'.
2723 [Phillip Porch <root@theporch.com>]
2725 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2726 a STACK of email addresses from a certificate or request, these look
2727 in the subject name and the subject alternative name extensions and
2728 omit any duplicate addresses.
2731 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2732 This makes DSA verification about 2 % faster.
2735 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2736 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2737 plus overhead for 1024 bit moduli).
2738 This makes exponentiations about 0.5 % faster for 1024 bit
2739 exponents (as measured by "openssl speed rsa2048").
2742 *) Rename memory handling macros to avoid conflicts with other
2744 Malloc => OPENSSL_malloc
2745 Malloc_locked => OPENSSL_malloc_locked
2746 Realloc => OPENSSL_realloc
2747 Free => OPENSSL_free
2750 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2751 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2754 *) CygWin32 support.
2755 [John Jarvie <jjarvie@newsguy.com>]
2757 *) The type-safe stack code has been rejigged. It is now only compiled
2758 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2759 by default all type-specific stack functions are "#define"d back to
2760 standard stack functions. This results in more streamlined output
2761 but retains the type-safety checking possibilities of the original
2765 *) The STACK code has been cleaned up, and certain type declarations
2766 that didn't make a lot of sense have been brought in line. This has
2767 also involved a cleanup of sorts in safestack.h to more correctly
2768 map type-safe stack functions onto their plain stack counterparts.
2769 This work has also resulted in a variety of "const"ifications of
2770 lots of the code, especially "_cmp" operations which should normally
2771 be prototyped with "const" parameters anyway.
2774 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2775 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2776 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2777 where all of 'md' is used each time the PRNG is used, but 'state'
2778 is used only indexed by a cyclic counter. As entropy may not be
2779 well distributed from the beginning, 'md' is important as a
2780 chaining variable. However, the output function chains only half
2781 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2782 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2783 in all of 'state' being rewritten, with the new values depending
2784 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2787 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2788 the handshake is continued after ssl_verify_cert_chain();
2789 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2790 can lead to 'unexplainable' connection aborts later.
2791 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2793 *) Major EVP API cipher revision.
2794 Add hooks for extra EVP features. This allows various cipher
2795 parameters to be set in the EVP interface. Support added for variable
2796 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2797 setting of RC2 and RC5 parameters.
2799 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2802 Remove lots of duplicated code from the EVP library. For example *every*
2803 cipher init() function handles the 'iv' in the same way according to the
2804 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2805 for CFB and OFB modes they zero ctx->num.
2807 New functionality allows removal of S/MIME code RC2 hack.
2809 Most of the routines have the same form and so can be declared in terms
2812 By shifting this to the top level EVP_CipherInit() it can be removed from
2813 all individual ciphers. If the cipher wants to handle IVs or keys
2814 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2817 Change lots of functions like EVP_EncryptUpdate() to now return a
2818 value: although software versions of the algorithms cannot fail
2819 any installed hardware versions can.
2822 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2823 this option is set, tolerate broken clients that send the negotiated
2824 protocol version number instead of the requested protocol version
2828 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2829 i.e. non-zero for export ciphersuites, zero otherwise.
2830 Previous versions had this flag inverted, inconsistent with
2831 rsa_tmp_cb (..._TMP_RSA_CB).
2832 [Bodo Moeller; problem reported by Amit Chopra]
2834 *) Add missing DSA library text string. Work around for some IIS
2835 key files with invalid SEQUENCE encoding.
2838 *) Add a document (doc/standards.txt) that list all kinds of standards
2839 and so on that are implemented in OpenSSL.
2842 *) Enhance c_rehash script. Old version would mishandle certificates
2843 with the same subject name hash and wouldn't handle CRLs at all.
2844 Added -fingerprint option to crl utility, to support new c_rehash
2848 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2851 *) Fix for SSL server purpose checking. Server checking was
2852 rejecting certificates which had extended key usage present
2853 but no ssl client purpose.
2854 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2856 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2857 is a little unclear about how a blank password is handled.
2858 Since the password in encoded as a BMPString with terminating
2859 double NULL a zero length password would end up as just the
2860 double NULL. However no password at all is different and is
2861 handled differently in the PKCS#12 key generation code. NS
2862 treats a blank password as zero length. MSIE treats it as no
2863 password on export: but it will try both on import. We now do
2864 the same: PKCS12_parse() tries zero length and no password if
2865 the password is set to "" or NULL (NULL is now a valid password:
2866 it wasn't before) as does the pkcs12 application.
2869 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2870 perror when PEM_read_bio_X509_REQ fails, the error message must
2871 be obtained from the error queue.
2874 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2875 it in ERR_remove_state if appropriate, and change ERR_get_state
2876 accordingly to avoid race conditions (this is necessary because
2877 thread_hash is no longer constant once set).
2880 *) Bugfix for linux-elf makefile.one.
2883 *) RSA_get_default_method() will now cause a default
2884 RSA_METHOD to be chosen if one doesn't exist already.
2885 Previously this was only set during a call to RSA_new()
2886 or RSA_new_method(NULL) meaning it was possible for
2887 RSA_get_default_method() to return NULL.
2890 *) Added native name translation to the existing DSO code
2891 that will convert (if the flag to do so is set) filenames
2892 that are sufficiently small and have no path information
2893 into a canonical native form. Eg. "blah" converted to
2894 "libblah.so" or "blah.dll" etc.
2897 *) New function ERR_error_string_n(e, buf, len) which is like
2898 ERR_error_string(e, buf), but writes at most 'len' bytes
2899 including the 0 terminator. For ERR_error_string_n, 'buf'
2901 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2903 *) CONF library reworked to become more general. A new CONF
2904 configuration file reader "class" is implemented as well as a
2905 new functions (NCONF_*, for "New CONF") to handle it. The now
2906 old CONF_* functions are still there, but are reimplemented to
2907 work in terms of the new functions. Also, a set of functions
2908 to handle the internal storage of the configuration data is
2909 provided to make it easier to write new configuration file
2910 reader "classes" (I can definitely see something reading a
2911 configuration file in XML format, for example), called _CONF_*,
2912 or "the configuration storage API"...
2914 The new configuration file reading functions are:
2916 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2917 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2919 NCONF_default, NCONF_WIN32
2921 NCONF_dump_fp, NCONF_dump_bio
2923 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2924 NCONF_new creates a new CONF object. This works in the same way
2925 as other interfaces in OpenSSL, like the BIO interface.
2926 NCONF_dump_* dump the internal storage of the configuration file,
2927 which is useful for debugging. All other functions take the same
2928 arguments as the old CONF_* functions wth the exception of the
2929 first that must be a `CONF *' instead of a `LHASH *'.
2931 To make it easer to use the new classes with the old CONF_* functions,
2932 the function CONF_set_default_method is provided.
2935 *) Add '-tls1' option to 'openssl ciphers', which was already
2936 mentioned in the documentation but had not been implemented.
2937 (This option is not yet really useful because even the additional
2938 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2941 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2942 OpenSSL-based applications) load shared libraries and bind to
2943 them in a portable way.
2944 [Geoff Thorpe, with contributions from Richard Levitte]
2946 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2948 *) Make sure _lrotl and _lrotr are only used with MSVC.
2950 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2951 (the default implementation of RAND_status).
2953 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2954 to '-clrext' (= clear extensions), as intended and documented.
2955 [Bodo Moeller; inconsistency pointed out by Michael Attili
2958 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2959 was larger than the MD block size.
2960 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2962 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2963 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2964 using the passed key: if the passed key was a private key the result
2965 of X509_print(), for example, would be to print out all the private key
2969 *) des_quad_cksum() byte order bug fix.
2970 [Ulf Möller, using the problem description in krb4-0.9.7, where
2971 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2973 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2975 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2977 *) For easily testing in shell scripts whether some command
2978 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2979 returns with exit code 0 iff no command of the given name is available.
2980 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2981 the output goes to stdout and nothing is printed to stderr.
2982 Additional arguments are always ignored.
2984 Since for each cipher there is a command of the same name,
2985 the 'no-cipher' compilation switches can be tested this way.
2987 ('openssl no-XXX' is not able to detect pseudo-commands such
2988 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2991 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2994 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2995 is set; it will be thrown away anyway because each handshake creates
2997 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2998 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2999 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3000 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3003 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3004 'Q' and 'R' lose their special meanings (quit/renegotiate).
3005 This is part of what -quiet does; unlike -quiet, -ign_eof
3006 does not suppress any output.
3009 *) Add compatibility options to the purpose and trust code. The
3010 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3011 accepts a certificate or CA, this was the previous behaviour,
3012 with all the associated security issues.
3014 X509_TRUST_COMPAT is the old trust behaviour: only and
3015 automatically trust self signed roots in certificate store. A
3016 new trust setting X509_TRUST_DEFAULT is used to specify that
3017 a purpose has no associated trust setting and it should instead
3018 use the value in the default purpose.
3021 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3022 and fix a memory leak.
3025 *) In util/mkerr.pl (which implements 'make errors'), preserve
3026 reason strings from the previous version of the .c file, as
3027 the default to have only downcase letters (and digits) in
3028 automatically generated reasons codes is not always appropriate.
3031 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3032 using strerror. Previously, ERR_reason_error_string() returned
3033 library names as reason strings for SYSerr; but SYSerr is a special
3034 case where small numbers are errno values, not library numbers.
3037 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3038 converts DSA parameters into DH parameters. (When creating parameters,
3039 DSA_generate_parameters is used.)
3042 *) Include 'length' (recommended exponent length) in C code generated
3043 by 'openssl dhparam -C'.
3046 *) The second argument to set_label in perlasm was already being used
3047 so couldn't be used as a "file scope" flag. Moved to third argument
3051 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3052 instead of RAND_bytes for encryption IVs and salts.
3055 *) Include RAND_status() into RAND_METHOD instead of implementing
3056 it only for md_rand.c Otherwise replacing the PRNG by calling
3057 RAND_set_rand_method would be impossible.
3060 *) Don't let DSA_generate_key() enter an infinite loop if the random
3061 number generation fails.
3064 *) New 'rand' application for creating pseudo-random output.
3067 *) Added configuration support for Linux/IA64
3068 [Rolf Haberrecker <rolf@suse.de>]
3070 *) Assembler module support for Mingw32.
3073 *) Shared library support for HPUX (in shlib/).
3074 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3076 *) Shared library support for Solaris gcc.
3077 [Lutz Behnke <behnke@trustcenter.de>]
3079 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3081 *) PKCS7_encrypt() was adding text MIME headers twice because they
3082 were added manually and by SMIME_crlf_copy().
3085 *) In bntest.c don't call BN_rand with zero bits argument.
3086 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3088 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3089 case was implemented. This caused BN_div_recp() to fail occasionally.
3092 *) Add an optional second argument to the set_label() in the perl
3093 assembly language builder. If this argument exists and is set
3094 to 1 it signals that the assembler should use a symbol whose
3095 scope is the entire file, not just the current function. This
3096 is needed with MASM which uses the format label:: for this scope.
3097 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3099 *) Change the ASN1 types so they are typedefs by default. Before
3100 almost all types were #define'd to ASN1_STRING which was causing
3101 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3105 *) Change names of new functions to the new get1/get0 naming
3106 convention: After 'get1', the caller owns a reference count
3107 and has to call ..._free; 'get0' returns a pointer to some
3108 data structure without incrementing reference counters.
3109 (Some of the existing 'get' functions increment a reference
3110 counter, some don't.)
3111 Similarly, 'set1' and 'add1' functions increase reference
3112 counters or duplicate objects.
3115 *) Allow for the possibility of temp RSA key generation failure:
3116 the code used to assume it always worked and crashed on failure.
3119 *) Fix potential buffer overrun problem in BIO_printf().
3120 [Ulf Möller, using public domain code by Patrick Powell; problem
3121 pointed out by David Sacerdote <das33@cornell.edu>]
3123 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3124 RAND_egd() and RAND_status(). In the command line application,
3125 the EGD socket can be specified like a seed file using RANDFILE
3129 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3130 Some CAs (e.g. Verisign) distribute certificates in this form.
3133 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3134 list to exclude them. This means that no special compilation option
3135 is needed to use anonymous DH: it just needs to be included in the
3139 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3140 EVP_MD_type. The old functionality is available in a new macro called
3141 EVP_MD_md(). Change code that uses it and update docs.
3144 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3145 where the 'void *' argument is replaced by a function pointer argument.
3146 Previously 'void *' was abused to point to functions, which works on
3147 many platforms, but is not correct. As these functions are usually
3148 called by macros defined in OpenSSL header files, most source code
3149 should work without changes.
3152 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3153 sections with information on -D... compiler switches used for
3154 compiling the library so that applications can see them. To enable
3155 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3156 must be defined. E.g.,
3157 #define OPENSSL_ALGORITHM_DEFINES
3158 #include <openssl/opensslconf.h>
3159 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3160 [Richard Levitte, Ulf and Bodo Möller]
3162 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3166 *) Change the 'other' type in certificate aux info to a STACK_OF
3167 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3168 the required ASN1 format: arbitrary types determined by an OID.
3171 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3172 argument to 'req'. This is not because the function is newer or
3173 better than others it just uses the work 'NEW' in the certificate
3174 request header lines. Some software needs this.
3177 *) Reorganise password command line arguments: now passwords can be
3178 obtained from various sources. Delete the PEM_cb function and make
3179 it the default behaviour: i.e. if the callback is NULL and the
3180 usrdata argument is not NULL interpret it as a null terminated pass
3181 phrase. If usrdata and the callback are NULL then the pass phrase
3182 is prompted for as usual.
3185 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3186 the support is automatically enabled. The resulting binaries will
3187 autodetect the card and use it if present.
3188 [Ben Laurie and Compaq Inc.]
3190 *) Work around for Netscape hang bug. This sends certificate request
3191 and server done in one record. Since this is perfectly legal in the
3192 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3193 the bugs/SSLv3 entry for more info.
3196 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3199 *) Add -rand argument to smime and pkcs12 applications and read/write
3203 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3206 *) Add command line password options to the remaining applications.
3209 *) Bug fix for BN_div_recp() for numerators with an even number of
3213 *) More tests in bntest.c, and changed test_bn output.
3216 *) ./config recognizes MacOS X now.
3219 *) Bug fix for BN_div() when the first words of num and divsor are
3220 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3223 *) Add support for various broken PKCS#8 formats, and command line
3224 options to produce them.
3227 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3228 get temporary BIGNUMs from a BN_CTX.
3231 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3235 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3236 include a #define from the old name to the new. The original intent
3237 was that statically linked binaries could for example just call
3238 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3239 link with digests. This never worked becayse SSLeay_add_all_digests()
3240 and SSLeay_add_all_ciphers() were in the same source file so calling
3241 one would link with the other. They are now in separate source files.
3244 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3247 *) Use a less unusual form of the Miller-Rabin primality test (it used
3248 a binary algorithm for exponentiation integrated into the Miller-Rabin
3249 loop, our standard modexp algorithms are faster).
3252 *) Support for the EBCDIC character set completed.
3253 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3255 *) Source code cleanups: use const where appropriate, eliminate casts,
3256 use void * instead of char * in lhash.
3259 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3260 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3261 this the server could overwrite ephemeral keys that the client
3265 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3266 using 50 iterations of the Rabin-Miller test.
3268 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3269 iterations of the Rabin-Miller test as required by the appendix
3270 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3271 As BN_is_prime_fasttest includes trial division, DSA parameter
3272 generation becomes much faster.
3274 This implies a change for the callback functions in DSA_is_prime
3275 and DSA_generate_parameters: The callback function is called once
3276 for each positive witness in the Rabin-Miller test, not just
3277 occasionally in the inner loop; and the parameters to the
3278 callback function now provide an iteration count for the outer
3279 loop rather than for the current invocation of the inner loop.
3280 DSA_generate_parameters additionally can call the callback
3281 function with an 'iteration count' of -1, meaning that a
3282 candidate has passed the trial division test (when q is generated
3283 from an application-provided seed, trial division is skipped).
3286 *) New function BN_is_prime_fasttest that optionally does trial
3287 division before starting the Rabin-Miller test and has
3288 an additional BN_CTX * argument (whereas BN_is_prime always
3289 has to allocate at least one BN_CTX).
3290 'callback(1, -1, cb_arg)' is called when a number has passed the
3291 trial division stage.
3294 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3298 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3301 *) New function BN_pseudo_rand().
3304 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3305 bignum version of BN_from_montgomery() with the working code from
3306 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3310 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3311 made it impossible to use the same SSL_SESSION data structure in
3312 SSL2 clients in multiple threads.
3315 *) The return value of RAND_load_file() no longer counts bytes obtained
3316 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3317 to seed the PRNG (previously an explicit byte count was required).
3318 [Ulf Möller, Bodo Möller]
3320 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3321 used (char *) instead of (void *) and had casts all over the place.
3324 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3327 *) Retain source code compatibility for BN_prime_checks macro:
3328 BN_is_prime(..., BN_prime_checks, ...) now uses
3329 BN_prime_checks_for_size to determine the appropriate number of
3330 Rabin-Miller iterations.
3333 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3334 DH_CHECK_P_NOT_SAFE_PRIME.
3335 (Check if this is true? OpenPGP calls them "strong".)
3338 *) Merge the functionality of "dh" and "gendh" programs into a new program
3339 "dhparam". The old programs are retained for now but will handle DH keys
3340 (instead of parameters) in future.
3343 *) Make the ciphers, s_server and s_client programs check the return values
3344 when a new cipher list is set.
3347 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3348 ciphers. Before when the 56bit ciphers were enabled the sorting was
3351 The syntax for the cipher sorting has been extended to support sorting by
3352 cipher-strength (using the strength_bits hard coded in the tables).
3353 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3355 Fix a bug in the cipher-command parser: when supplying a cipher command
3356 string with an "undefined" symbol (neither command nor alphanumeric
3357 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3358 an error is flagged.
3360 Due to the strength-sorting extension, the code of the
3361 ssl_create_cipher_list() function was completely rearranged. I hope that
3362 the readability was also increased :-)
3363 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3365 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3366 for the first serial number and places 2 in the serial number file. This
3367 avoids problems when the root CA is created with serial number zero and
3368 the first user certificate has the same issuer name and serial number
3372 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3373 the new code. Add documentation for this stuff.
3376 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3377 X509_*() to X509at_*() on the grounds that they don't handle X509
3378 structures and behave in an analagous way to the X509v3 functions:
3379 they shouldn't be called directly but wrapper functions should be used
3382 So we also now have some wrapper functions that call the X509at functions
3383 when passed certificate requests. (TO DO: similar things can be done with
3384 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3385 things. Some of these need some d2i or i2d and print functionality
3386 because they handle more complex structures.)
3389 *) Add missing #ifndefs that caused missing symbols when building libssl
3390 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3391 NO_RSA in ssl/s2*.c.
3392 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3394 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3395 has a return value which indicates the quality of the random data
3396 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3397 error queue. New function RAND_pseudo_bytes() generates output that is
3398 guaranteed to be unique but not unpredictable. RAND_add is like
3399 RAND_seed, but takes an extra argument for an entropy estimate
3400 (RAND_seed always assumes full entropy).
3403 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3404 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3405 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3406 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3407 false-positive rate of at most 2^-80 for random input.
3410 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3413 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3414 in the 0.9.5 release), this returns the chain
3415 from an X509_CTX structure with a dup of the stack and all
3416 the X509 reference counts upped: so the stack will exist
3417 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3420 Also make SSL_SESSION_print() print out the verify return
3424 *) Add manpage for the pkcs12 command. Also change the default
3425 behaviour so MAC iteration counts are used unless the new
3426 -nomaciter option is used. This improves file security and
3427 only older versions of MSIE (4.0 for example) need it.
3430 *) Honor the no-xxx Configure options when creating .DEF files.
3433 *) Add PKCS#10 attributes to field table: challengePassword,
3434 unstructuredName and unstructuredAddress. These are taken from
3435 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3436 international characters are used.
3438 More changes to X509_ATTRIBUTE code: allow the setting of types
3439 based on strings. Remove the 'loc' parameter when adding
3440 attributes because these will be a SET OF encoding which is sorted
3444 *) Initial changes to the 'req' utility to allow request generation
3445 automation. This will allow an application to just generate a template
3446 file containing all the field values and have req construct the
3449 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3450 used all over the place including certificate requests and PKCS#7
3451 structures. They are currently handled manually where necessary with
3452 some primitive wrappers for PKCS#7. The new functions behave in a
3453 manner analogous to the X509 extension functions: they allow
3454 attributes to be looked up by NID and added.
3456 Later something similar to the X509V3 code would be desirable to
3457 automatically handle the encoding, decoding and printing of the
3458 more complex types. The string types like challengePassword can
3459 be handled by the string table functions.
3461 Also modified the multi byte string table handling. Now there is
3462 a 'global mask' which masks out certain types. The table itself
3463 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3464 is useful when for example there is only one permissible type
3465 (as in countryName) and using the mask might result in no valid
3469 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3470 SSL_get_peer_finished to allow applications to obtain the latest
3471 Finished messages sent to the peer or expected from the peer,
3472 respectively. (SSL_get_peer_finished is usually the Finished message
3473 actually received from the peer, otherwise the protocol will be aborted.)
3475 As the Finished message are message digests of the complete handshake
3476 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3477 be used for external authentication procedures when the authentication
3478 provided by SSL/TLS is not desired or is not enough.
3481 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3482 the host supports BWX extension and if Compaq C is present on the
3483 $PATH. Just exploiting of the BWX extension results in 20-30%
3484 performance kick for some algorithms, e.g. DES and RC4 to mention
3485 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3489 *) Add support for MS "fast SGC". This is arguably a violation of the
3490 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3491 weak crypto and after checking the certificate is SGC a second one
3492 with strong crypto. MS SGC stops the first handshake after receiving
3493 the server certificate message and sends a second client hello. Since
3494 a server will typically do all the time consuming operations before
3495 expecting any further messages from the client (server key exchange
3496 is the most expensive) there is little difference between the two.
3498 To get OpenSSL to support MS SGC we have to permit a second client
3499 hello message after we have sent server done. In addition we have to
3500 reset the MAC if we do get this second client hello.
3503 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3504 if a DER encoded private key is RSA or DSA traditional format. Changed
3505 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3506 format DER encoded private key. Newer code should use PKCS#8 format which
3507 has the key type encoded in the ASN1 structure. Added DER private key
3508 support to pkcs8 application.
3511 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3512 ciphersuites has been selected (as required by the SSL 3/TLS 1
3513 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3514 is set, we interpret this as a request to violate the specification
3515 (the worst that can happen is a handshake failure, and 'correct'
3516 behaviour would result in a handshake failure anyway).
3519 *) In SSL_CTX_add_session, take into account that there might be multiple
3520 SSL_SESSION structures with the same session ID (e.g. when two threads
3521 concurrently obtain them from an external cache).
3522 The internal cache can handle only one SSL_SESSION with a given ID,
3523 so if there's a conflict, we now throw out the old one to achieve
3527 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3528 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3529 some routines that use cipher OIDs: some ciphers do not have OIDs
3530 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3534 *) Simplify the trust setting structure and code. Now we just have
3535 two sequences of OIDs for trusted and rejected settings. These will
3536 typically have values the same as the extended key usage extension
3537 and any application specific purposes.
3539 The trust checking code now has a default behaviour: it will just
3540 check for an object with the same NID as the passed id. Functions can
3541 be provided to override either the default behaviour or the behaviour
3542 for a given id. SSL client, server and email already have functions
3543 in place for compatibility: they check the NID and also return "trusted"
3544 if the certificate is self signed.
3547 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3548 traditional format into an EVP_PKEY structure.
3551 *) Add a password callback function PEM_cb() which either prompts for
3552 a password if usr_data is NULL or otherwise assumes it is a null
3553 terminated password. Allow passwords to be passed on command line
3554 environment or config files in a few more utilities.
3557 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3558 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3559 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3560 Update documentation.
3563 *) Support for ASN1 "NULL" type. This could be handled before by using
3564 ASN1_TYPE but there wasn't any function that would try to read a NULL
3565 and produce an error if it couldn't. For compatibility we also have
3566 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3567 don't allocate anything because they don't need to.
3570 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3572 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3574 *) Rebuild of the memory allocation routines used by OpenSSL code and
3575 possibly others as well. The purpose is to make an interface that
3576 provide hooks so anyone can build a separate set of allocation and
3577 deallocation routines to be used by OpenSSL, for example memory
3578 pool implementations, or something else, which was previously hard
3579 since Malloc(), Realloc() and Free() were defined as macros having
3580 the values malloc, realloc and free, respectively (except for Win32
3581 compilations). The same is provided for memory debugging code.
3582 OpenSSL already comes with functionality to find memory leaks, but
3583 this gives people a chance to debug other memory problems.
3585 With these changes, a new set of functions and macros have appeared:
3587 CRYPTO_set_mem_debug_functions() [F]
3588 CRYPTO_get_mem_debug_functions() [F]
3589 CRYPTO_dbg_set_options() [F]
3590 CRYPTO_dbg_get_options() [F]
3591 CRYPTO_malloc_debug_init() [M]
3593 The memory debug functions are NULL by default, unless the library
3594 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3595 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3596 gives the standard debugging functions that come with OpenSSL) or
3597 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3598 provided by the library user) must be used. When the standard
3599 debugging functions are used, CRYPTO_dbg_set_options can be used to
3600 request additional information:
3601 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3602 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3604 Also, things like CRYPTO_set_mem_functions will always give the
3605 expected result (the new set of functions is used for allocation
3606 and deallocation) at all times, regardless of platform and compiler
3609 To finish it up, some functions that were never use in any other
3610 way than through macros have a new API and new semantic:
3613 CRYPTO_dbg_realloc()
3616 All macros of value have retained their old syntax.
3617 [Richard Levitte and Bodo Moeller]
3619 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3620 ordering of SMIMECapabilities wasn't in "strength order" and there
3621 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3625 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3626 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3627 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3629 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3630 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3631 functionality to handle multipart/signed properly) and a utility
3632 called 'smime' to call all this stuff. This is based on code I
3633 originally wrote for Celo who have kindly allowed it to be
3634 included in OpenSSL.
3637 *) Add variants des_set_key_checked and des_set_key_unchecked of
3638 des_set_key (aka des_key_sched). Global variable des_check_key
3639 decides which of these is called by des_set_key; this way
3640 des_check_key behaves as it always did, but applications and
3641 the library itself, which was buggy for des_check_key == 1,
3642 have a cleaner way to pick the version they need.
3645 *) New function PKCS12_newpass() which changes the password of a
3649 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3650 dynamic mix. In both cases the ids can be used as an index into the
3651 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3652 functions so they accept a list of the field values and the
3653 application doesn't need to directly manipulate the X509_TRUST
3657 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3661 *) Modify the way the V3 extension code looks up extensions. This now
3662 works in a similar way to the object code: we have some "standard"
3663 extensions in a static table which is searched with OBJ_bsearch()
3664 and the application can add dynamic ones if needed. The file
3665 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3666 updated whenever a new extension is added to the core code and kept
3667 in ext_nid order. There is a simple program 'tabtest.c' which checks
3668 this. New extensions are not added too often so this file can readily
3669 be maintained manually.
3671 There are two big advantages in doing things this way. The extensions
3672 can be looked up immediately and no longer need to be "added" using
3673 X509V3_add_standard_extensions(): this function now does nothing.
3674 [Side note: I get *lots* of email saying the extension code doesn't
3675 work because people forget to call this function]
3676 Also no dynamic allocation is done unless new extensions are added:
3677 so if we don't add custom extensions there is no need to call
3678 X509V3_EXT_cleanup().
3681 *) Modify enc utility's salting as follows: make salting the default. Add a
3682 magic header, so unsalted files fail gracefully instead of just decrypting
3683 to garbage. This is because not salting is a big security hole, so people
3684 should be discouraged from doing it.
3687 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3688 digest to be passed on the command line but it only used this
3689 parameter when signing a certificate. Modified so all relevant
3690 operations are affected by the digest parameter including the
3691 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3692 DSA key was used because it didn't fix the digest.
3695 *) Initial certificate chain verify code. Currently tests the untrusted
3696 certificates for consistency with the verify purpose (which is set
3697 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3699 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3700 this is because it will reject chains with invalid extensions whereas
3701 every previous version of OpenSSL and SSLeay made no checks at all.
3703 Trust code: checks the root CA for the relevant trust settings. Trust
3704 settings have an initial value consistent with the verify purpose: e.g.
3705 if the verify purpose is for SSL client use it expects the CA to be
3706 trusted for SSL client use. However the default value can be changed to
3707 permit custom trust settings: one example of this would be to only trust
3708 certificates from a specific "secure" set of CAs.
3710 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3711 which should be used for version portability: especially since the
3712 verify structure is likely to change more often now.
3714 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3715 to set them. If not set then assume SSL clients will verify SSL servers
3718 Two new options to the verify program: -untrusted allows a set of
3719 untrusted certificates to be passed in and -purpose which sets the
3720 intended purpose of the certificate. If a purpose is set then the
3721 new chain verify code is used to check extension consistency.
3724 *) Support for the authority information access extension.
3727 *) Modify RSA and DSA PEM read routines to transparently handle
3728 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3729 public keys in a format compatible with certificate
3730 SubjectPublicKeyInfo structures. Unfortunately there were already
3731 functions called *_PublicKey_* which used various odd formats so
3732 these are retained for compatibility: however the DSA variants were
3733 never in a public release so they have been deleted. Changed dsa/rsa
3734 utilities to handle the new format: note no releases ever handled public
3735 keys so we should be OK.
3737 The primary motivation for this change is to avoid the same fiasco
3738 that dogs private keys: there are several incompatible private key
3739 formats some of which are standard and some OpenSSL specific and
3740 require various evil hacks to allow partial transparent handling and
3741 even then it doesn't work with DER formats. Given the option anything
3742 other than PKCS#8 should be dumped: but the other formats have to
3743 stay in the name of compatibility.
3745 With public keys and the benefit of hindsight one standard format
3746 is used which works with EVP_PKEY, RSA or DSA structures: though
3747 it clearly returns an error if you try to read the wrong kind of key.
3749 Added a -pubkey option to the 'x509' utility to output the public key.
3750 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3751 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3752 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3753 that do the same as the EVP_PKEY_assign_*() except they up the
3754 reference count of the added key (they don't "swallow" the
3758 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3759 CRLs would fail if the file contained no certificates or no CRLs:
3760 added a new function to read in both types and return the number
3761 read: this means that if none are read it will be an error. The
3762 DER versions of the certificate and CRL reader would always fail
3763 because it isn't possible to mix certificates and CRLs in DER format
3764 without choking one or the other routine. Changed this to just read
3765 a certificate: this is the best we can do. Also modified the code
3766 in apps/verify.c to take notice of return codes: it was previously
3767 attempting to read in certificates from NULL pointers and ignoring
3768 any errors: this is one reason why the cert and CRL reader seemed
3769 to work. It doesn't check return codes from the default certificate
3770 routines: these may well fail if the certificates aren't installed.
3773 *) Code to support otherName option in GeneralName.
3776 *) First update to verify code. Change the verify utility
3777 so it warns if it is passed a self signed certificate:
3778 for consistency with the normal behaviour. X509_verify
3779 has been modified to it will now verify a self signed
3780 certificate if *exactly* the same certificate appears
3781 in the store: it was previously impossible to trust a
3782 single self signed certificate. This means that:
3783 openssl verify ss.pem
3784 now gives a warning about a self signed certificate but
3785 openssl verify -CAfile ss.pem ss.pem
3789 *) For servers, store verify_result in SSL_SESSION data structure
3790 (and add it to external session representation).
3791 This is needed when client certificate verifications fails,
3792 but an application-provided verification callback (set by
3793 SSL_CTX_set_cert_verify_callback) allows accepting the session
3794 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3795 but returns 1): When the session is reused, we have to set
3796 ssl->verify_result to the appropriate error code to avoid
3798 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3800 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3801 case in PKCS7_dataInit() where the signed PKCS7 structure
3802 didn't contain any existing data because it was being created.
3803 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3805 *) Add a salt to the key derivation routines in enc.c. This
3806 forms the first 8 bytes of the encrypted file. Also add a
3807 -S option to allow a salt to be input on the command line.
3810 *) New function X509_cmp(). Oddly enough there wasn't a function
3811 to compare two certificates. We do this by working out the SHA1
3812 hash and comparing that. X509_cmp() will be needed by the trust
3816 *) SSL_get1_session() is like SSL_get_session(), but increments
3817 the reference count in the SSL_SESSION returned.
3818 [Geoff Thorpe <geoff@eu.c2.net>]
3820 *) Fix for 'req': it was adding a null to request attributes.
3821 Also change the X509_LOOKUP and X509_INFO code to handle
3822 certificate auxiliary information.
3825 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3829 *) Add the possibility to add extra information to the memory leak
3830 detecting output, to form tracebacks, showing from where each
3831 allocation was originated: CRYPTO_push_info("constant string") adds
3832 the string plus current file name and line number to a per-thread
3833 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3834 is like calling CYRPTO_pop_info() until the stack is empty.
3835 Also updated memory leak detection code to be multi-thread-safe.
3838 *) Add options -text and -noout to pkcs7 utility and delete the
3839 encryption options which never did anything. Update docs.
3842 *) Add options to some of the utilities to allow the pass phrase
3843 to be included on either the command line (not recommended on
3844 OSes like Unix) or read from the environment. Update the
3845 manpages and fix a few bugs.
3848 *) Add a few manpages for some of the openssl commands.
3851 *) Fix the -revoke option in ca. It was freeing up memory twice,
3852 leaking and not finding already revoked certificates.
3855 *) Extensive changes to support certificate auxiliary information.
3856 This involves the use of X509_CERT_AUX structure and X509_AUX
3857 functions. An X509_AUX function such as PEM_read_X509_AUX()
3858 can still read in a certificate file in the usual way but it
3859 will also read in any additional "auxiliary information". By
3860 doing things this way a fair degree of compatibility can be
3861 retained: existing certificates can have this information added
3862 using the new 'x509' options.
3864 Current auxiliary information includes an "alias" and some trust
3865 settings. The trust settings will ultimately be used in enhanced
3866 certificate chain verification routines: currently a certificate
3867 can only be trusted if it is self signed and then it is trusted
3871 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3872 The problem was that one of the replacement routines had not been working
3873 since SSLeay releases. For now the offending routine has been replaced
3874 with non-optimised assembler. Even so, this now gives around 95%
3875 performance improvement for 1024 bit RSA signs.
3878 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3879 handling. Most clients have the effective key size in bits equal to
3880 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3881 A few however don't do this and instead use the size of the decrypted key
3882 to determine the RC2 key length and the AlgorithmIdentifier to determine
3883 the effective key length. In this case the effective key length can still
3884 be 40 bits but the key length can be 168 bits for example. This is fixed
3885 by manually forcing an RC2 key into the EVP_PKEY structure because the
3886 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3887 the key length and effective key length are equal.
3890 *) Add a bunch of functions that should simplify the creation of
3891 X509_NAME structures. Now you should be able to do:
3892 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3893 and have it automatically work out the correct field type and fill in
3894 the structures. The more adventurous can try:
3895 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3896 and it will (hopefully) work out the correct multibyte encoding.
3899 *) Change the 'req' utility to use the new field handling and multibyte
3900 copy routines. Before the DN field creation was handled in an ad hoc
3901 way in req, ca, and x509 which was rather broken and didn't support
3902 BMPStrings or UTF8Strings. Since some software doesn't implement
3903 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3904 using the dirstring_type option. See the new comment in the default
3905 openssl.cnf for more info.
3908 *) Make crypto/rand/md_rand.c more robust:
3909 - Assure unique random numbers after fork().
3910 - Make sure that concurrent threads access the global counter and
3911 md serializably so that we never lose entropy in them
3912 or use exactly the same state in multiple threads.
3913 Access to the large state is not always serializable because
3914 the additional locking could be a performance killer, and
3915 md should be large enough anyway.
3918 *) New file apps/app_rand.c with commonly needed functionality
3919 for handling the random seed file.
3921 Use the random seed file in some applications that previously did not:
3923 dsaparam -genkey (which also ignored its '-rand' option),
3926 x509 (when signing).
3927 Except on systems with /dev/urandom, it is crucial to have a random
3928 seed file at least for key creation, DSA signing, and for DH exchanges;
3929 for RSA signatures we could do without one.
3931 gendh and gendsa (unlike genrsa) used to read only the first byte
3932 of each file listed in the '-rand' option. The function as previously
3933 found in genrsa is now in app_rand.c and is used by all programs
3934 that support '-rand'.
3937 *) In RAND_write_file, use mode 0600 for creating files;
3938 don't just chmod when it may be too late.
3941 *) Report an error from X509_STORE_load_locations
3942 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3945 *) New function ASN1_mbstring_copy() this copies a string in either
3946 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3947 into an ASN1_STRING type. A mask of permissible types is passed
3948 and it chooses the "minimal" type to use or an error if not type
3952 *) Add function equivalents to the various macros in asn1.h. The old
3953 macros are retained with an M_ prefix. Code inside the library can
3954 use the M_ macros. External code (including the openssl utility)
3955 should *NOT* in order to be "shared library friendly".
3958 *) Add various functions that can check a certificate's extensions
3959 to see if it usable for various purposes such as SSL client,
3960 server or S/MIME and CAs of these types. This is currently
3961 VERY EXPERIMENTAL but will ultimately be used for certificate chain
3962 verification. Also added a -purpose flag to x509 utility to
3963 print out all the purposes.
3966 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
3970 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
3971 for, obtain and decode and extension and obtain its critical flag.
3972 This allows all the necessary extension code to be handled in a
3973 single function call.
3976 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
3977 platforms. See crypto/rc4/rc4_enc.c for further details.
3980 *) New -noout option to asn1parse. This causes no output to be produced
3981 its main use is when combined with -strparse and -out to extract data
3982 from a file (which may not be in ASN.1 format).
3985 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
3986 when producing the local key id.
3987 [Richard Levitte <levitte@stacken.kth.se>]
3989 *) New option -dhparam in s_server. This allows a DH parameter file to be
3990 stated explicitly. If it is not stated then it tries the first server
3991 certificate file. The previous behaviour hard coded the filename
3995 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
3996 a public key to be input or output. For example:
3997 openssl rsa -in key.pem -pubout -out pubkey.pem
3998 Also added necessary DSA public key functions to handle this.
4001 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4002 in the message. This was handled by allowing
4003 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4004 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4006 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4007 to the end of the strings whereas this didn't. This would cause problems
4008 if strings read with d2i_ASN1_bytes() were later modified.
4009 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4011 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4012 data and it contains EOF it will end up returning an error. This is
4013 caused by input 46 bytes long. The cause is due to the way base64
4014 BIOs find the start of base64 encoded data. They do this by trying a
4015 trial decode on each line until they find one that works. When they
4016 do a flag is set and it starts again knowing it can pass all the
4017 data directly through the decoder. Unfortunately it doesn't reset
4018 the context it uses. This means that if EOF is reached an attempt
4019 is made to pass two EOFs through the context and this causes the
4020 resulting error. This can also cause other problems as well. As is
4021 usual with these problems it takes *ages* to find and the fix is
4022 trivial: move one line.
4023 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4025 *) Ugly workaround to get s_client and s_server working under Windows. The
4026 old code wouldn't work because it needed to select() on sockets and the
4027 tty (for keypresses and to see if data could be written). Win32 only
4028 supports select() on sockets so we select() with a 1s timeout on the
4029 sockets and then see if any characters are waiting to be read, if none
4030 are present then we retry, we also assume we can always write data to
4031 the tty. This isn't nice because the code then blocks until we've
4032 received a complete line of data and it is effectively polling the
4033 keyboard at 1s intervals: however it's quite a bit better than not
4034 working at all :-) A dedicated Windows application might handle this
4035 with an event loop for example.
4038 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4039 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4040 will be called when RSA_sign() and RSA_verify() are used. This is useful
4041 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4042 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4043 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4044 This necessitated the support of an extra signature type NID_md5_sha1
4045 for SSL signatures and modifications to the SSL library to use it instead
4046 of calling RSA_public_decrypt() and RSA_private_encrypt().
4049 *) Add new -verify -CAfile and -CApath options to the crl program, these
4050 will lookup a CRL issuers certificate and verify the signature in a
4051 similar way to the verify program. Tidy up the crl program so it
4052 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4053 less strict. It will now permit CRL extensions even if it is not
4054 a V2 CRL: this will allow it to tolerate some broken CRLs.
4057 *) Initialize all non-automatic variables each time one of the openssl
4058 sub-programs is started (this is necessary as they may be started
4059 multiple times from the "OpenSSL>" prompt).
4060 [Lennart Bang, Bodo Moeller]
4062 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4063 removing all other RSA functionality (this is what NO_RSA does). This
4064 is so (for example) those in the US can disable those operations covered
4065 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4069 *) Non-copying interface to BIO pairs.
4070 (still largely untested)
4073 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4074 ASCII string. This was handled independently in various places before.
4077 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4078 UTF8 strings a character at a time.
4081 *) Use client_version from client hello to select the protocol
4082 (s23_srvr.c) and for RSA client key exchange verification
4083 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4086 *) Add various utility functions to handle SPKACs, these were previously
4087 handled by poking round in the structure internals. Added new function
4088 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4089 print, verify and generate SPKACs. Based on an original idea from
4090 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4093 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4096 *) Allow the config file extension section to be overwritten on the
4097 command line. Based on an original idea from Massimiliano Pala
4098 <madwolf@comune.modena.it>. The new option is called -extensions
4099 and can be applied to ca, req and x509. Also -reqexts to override
4100 the request extensions in req and -crlexts to override the crl extensions
4104 *) Add new feature to the SPKAC handling in ca. Now you can include
4105 the same field multiple times by preceding it by "XXXX." for example:
4108 this is the same syntax as used in the req config file.
4111 *) Allow certificate extensions to be added to certificate requests. These
4112 are specified in a 'req_extensions' option of the req section of the
4113 config file. They can be printed out with the -text option to req but
4114 are otherwise ignored at present.
4117 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4118 data read consists of only the final block it would not decrypted because
4119 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4120 A misplaced 'break' also meant the decrypted final block might not be
4121 copied until the next read.
4124 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4125 a few extra parameters to the DH structure: these will be useful if
4126 for example we want the value of 'q' or implement X9.42 DH.
4129 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4130 provides hooks that allow the default DSA functions or functions on a
4131 "per key" basis to be replaced. This allows hardware acceleration and
4132 hardware key storage to be handled without major modification to the
4133 library. Also added low level modexp hooks and CRYPTO_EX structure and
4134 associated functions.
4137 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4138 as "read only": it can't be written to and the buffer it points to will
4139 not be freed. Reading from a read only BIO is much more efficient than
4140 a normal memory BIO. This was added because there are several times when
4141 an area of memory needs to be read from a BIO. The previous method was
4142 to create a memory BIO and write the data to it, this results in two
4143 copies of the data and an O(n^2) reading algorithm. There is a new
4144 function BIO_new_mem_buf() which creates a read only memory BIO from
4145 an area of memory. Also modified the PKCS#7 routines to use read only
4149 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4150 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4151 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4152 but a retry condition occured while trying to read the rest.
4155 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4156 NID_pkcs7_encrypted by default: this was wrong since this should almost
4157 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4158 the encrypted data type: this is a more sensible place to put it and it
4159 allows the PKCS#12 code to be tidied up that duplicated this
4163 *) Changed obj_dat.pl script so it takes its input and output files on
4164 the command line. This should avoid shell escape redirection problems
4168 *) Initial support for certificate extension requests, these are included
4169 in things like Xenroll certificate requests. Included functions to allow
4170 extensions to be obtained and added.
4173 *) -crlf option to s_client and s_server for sending newlines as
4174 CRLF (as required by many protocols).
4177 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4179 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4180 [Ralf S. Engelschall]
4182 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4183 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4185 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4189 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4190 DH parameters/keys (q is lost during that conversion, but the resulting
4191 DH parameters contain its length).
4193 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4194 much faster than DH_generate_parameters (which creates parameters
4195 where p = 2*q + 1), and also the smaller q makes DH computations
4196 much more efficient (160-bit exponentiation instead of 1024-bit
4197 exponentiation); so this provides a convenient way to support DHE
4198 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4199 utter importance to use
4200 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4202 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4203 when such DH parameters are used, because otherwise small subgroup
4204 attacks may become possible!
4207 *) Avoid memory leak in i2d_DHparams.
4210 *) Allow the -k option to be used more than once in the enc program:
4211 this allows the same encrypted message to be read by multiple recipients.
4214 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4215 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4216 it will always use the numerical form of the OID, even if it has a short
4220 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4221 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4222 otherwise bn_mod_exp was called. In the case of hardware keys for example
4223 no private key components need be present and it might store extra data
4224 in the RSA structure, which cannot be accessed from bn_mod_exp.
4225 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4226 private key operations.
4229 *) Added support for SPARC Linux.
4232 *) pem_password_cb function type incompatibly changed from
4233 typedef int pem_password_cb(char *buf, int size, int rwflag);
4235 ....(char *buf, int size, int rwflag, void *userdata);
4236 so that applications can pass data to their callbacks:
4237 The PEM[_ASN1]_{read,write}... functions and macros now take an
4238 additional void * argument, which is just handed through whenever
4239 the password callback is called.
4240 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4242 New function SSL_CTX_set_default_passwd_cb_userdata.
4244 Compatibility note: As many C implementations push function arguments
4245 onto the stack in reverse order, the new library version is likely to
4246 interoperate with programs that have been compiled with the old
4247 pem_password_cb definition (PEM_whatever takes some data that
4248 happens to be on the stack as its last argument, and the callback
4249 just ignores this garbage); but there is no guarantee whatsoever that
4252 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4253 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4254 problems not only on Windows, but also on some Unix platforms.
4255 To avoid problematic command lines, these definitions are now in an
4256 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4257 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4260 *) MIPS III/IV assembler module is reimplemented.
4263 *) More DES library cleanups: remove references to srand/rand and
4264 delete an unused file.
4267 *) Add support for the the free Netwide assembler (NASM) under Win32,
4268 since not many people have MASM (ml) and it can be hard to obtain.
4269 This is currently experimental but it seems to work OK and pass all
4270 the tests. Check out INSTALL.W32 for info.
4273 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4274 without temporary keys kept an extra copy of the server key,
4275 and connections with temporary keys did not free everything in case
4279 *) New function RSA_check_key and new openssl rsa option -check
4280 for verifying the consistency of RSA keys.
4281 [Ulf Moeller, Bodo Moeller]
4283 *) Various changes to make Win32 compile work:
4284 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4285 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4286 comparison" warnings.
4287 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4290 *) Add a debugging option to PKCS#5 v2 key generation function: when
4291 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4292 derived keys are printed to stderr.
4295 *) Copy the flags in ASN1_STRING_dup().
4296 [Roman E. Pavlov <pre@mo.msk.ru>]
4298 *) The x509 application mishandled signing requests containing DSA
4299 keys when the signing key was also DSA and the parameters didn't match.
4301 It was supposed to omit the parameters when they matched the signing key:
4302 the verifying software was then supposed to automatically use the CA's
4303 parameters if they were absent from the end user certificate.
4305 Omitting parameters is no longer recommended. The test was also
4306 the wrong way round! This was probably due to unusual behaviour in
4307 EVP_cmp_parameters() which returns 1 if the parameters match.
4308 This meant that parameters were omitted when they *didn't* match and
4309 the certificate was useless. Certificates signed with 'ca' didn't have
4311 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4313 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4314 The interface is as follows:
4315 Applications can use
4316 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4317 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4318 "off" is now the default.
4319 The library internally uses
4320 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4321 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka
projects / openssl.git / blob