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