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 *) Fix object definitions for Private and Enterprise: they were not
17 recognized in their shortname (=lowercase) representation. Extend
18 obj_dat.pl to issue an error when using undefined keywords instead
19 of silently ignoring the problem (Svenning Sorensen
20 <sss@sss.dnsalias.net>).
23 +) Make object definitions compliant to LDAP (RFC2256): SN is the short
24 form for "surname", serialNumber has no short form (Michael Bell
25 <michael.bell@rz.hu-berlin.de>).
28 *) Fix DH_generate_parameters() so that it works for 'non-standard'
29 generators, i.e. generators other than 2 and 5. (Previously, the
30 code did not properly initialise the 'add' and 'rem' values to
33 In the new general case, we do not insist that 'generator' is
34 actually a primitive root: This requirement is rather pointless;
35 a generator of the order-q subgroup is just as good, if not
39 *) Map new X509 verification errors to alerts. Discovered and submitted by
40 Tom Wu <tom@arcot.com>.
43 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
44 returning non-zero before the data has been completely received
45 when using non-blocking I/O.
46 [Bodo Moeller; problem pointed out by John Hughes]
48 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
49 [Ben Laurie, Lutz Jaenicke]
51 +) Add an "init" command to the ENGINE config module and auto initialize
52 ENGINEs. Without any "init" command the ENGINE will be initialized
53 after all ctrl commands have been executed on it. If init=1 the
54 ENGINE is initailized at that point (ctrls before that point are run
55 on the uninitialized ENGINE and after on the initialized one). If
56 init=0 then the ENGINE will not be iniatialized at all.
59 +) Fix the 'app_verify_callback' interface so that the user-defined
60 argument is actually passed to the callback: In the
61 SSL_CTX_set_cert_verify_callback() prototype, the callback
62 declaration has been changed from
65 int (*cb)(X509_STORE_CTX *,void *);
66 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
67 i=s->ctx->app_verify_callback(&ctx)
69 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
71 To update applications using SSL_CTX_set_cert_verify_callback(),
72 a dummy argument can be added to their callback functions.
73 [D. K. Smetters <smetters@parc.xerox.com>]
75 +) Added the '4758cca' ENGINE to support IBM 4758 cards.
76 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
78 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
79 Yoram Zahavi <YoramZ@gilian.com>).
82 +) Add and OPENSSL_LOAD_CONF define which will cause
83 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
84 This allows older applications to transparently support certain
85 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
86 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
87 load the config file and OPENSSL_add_all_algorithms_conf() which will
88 always load it have also been added.
91 +) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
92 Adjust NIDs and EVP layer.
93 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
95 +) Config modules support in openssl utility.
97 Most commands now load modules from the config file,
98 though in a few (such as version) this isn't done
99 because it couldn't be used for anything.
101 In the case of ca and req the config file used is
102 the same as the utility itself: that is the -config
103 command line option can be used to specify an
107 +) Move default behaviour from OPENSSL_config(). If appname is NULL
108 use "openssl_conf" if filename is NULL use default openssl config file.
111 +) Add an argument to OPENSSL_config() to allow the use of an alternative
112 config section name. Add a new flag to tolerate a missing config file
113 and move code to CONF_modules_load_file().
116 *) Add information about CygWin 1.3 and on, and preserve proper
117 configuration for the versions before that.
118 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
120 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
121 check whether we deal with a copy of a session and do not delete from
122 the cache in this case. Problem reported by "Izhar Shoshani Levi"
123 <izhar@checkpoint.com>.
126 *) Do not store session data into the internal session cache, if it
127 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
128 flag is set). Proposed by Aslam <aslam@funk.com>.
131 +) Support for crypto accelerator cards from Accelerated Encryption
132 Processing, www.aep.ie. (Use engine 'aep')
133 The support was copied from 0.9.6c [engine] and adapted/corrected
134 to work with the new engine framework.
135 [AEP Inc. and Richard Levitte]
137 +) Support for SureWare crypto accelerator cards from Baltimore
138 Technologies. (Use engine 'sureware')
139 The support was copied from 0.9.6c [engine] and adapted
140 to work with the new engine framework.
143 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
147 +) Have the CHIL engine fork-safe (as defined by nCipher) and actually
148 make the newer ENGINE framework commands for the CHIL engine work.
149 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
151 +) Make it possible to produce shared libraries on ReliantUNIX.
152 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
154 *) Add the configuration target linux-s390x.
155 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
157 +) Add the configuration target debug-linux-ppro.
158 Make 'openssl rsa' use the general key loading routines
159 implemented in apps.c, and make those routines able to
160 handle the key format FORMAT_NETSCAPE and the variant
162 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
164 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
165 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
167 +) Add -keyform to rsautl, and document -engine.
168 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
170 +) Change BIO_new_file (crypto/bio/bss_file.c) to use new
171 BIO_R_NO_SUCH_FILE error code rather than the generic
172 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
177 ERR_peek_last_error_line
178 ERR_peek_last_error_line_data.
182 ERR_peek_error_line_data,
183 but report on the latest error recorded rather than the first one
184 still in the error queue.
185 [Ben Laurie, Bodo Moeller]
187 +) default_algorithms option in ENGINE config module. This allows things
189 default_algorithms = ALL
190 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
193 +) Prelminary ENGINE config module.
196 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
197 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
198 variable as an indication that a ClientHello message has been
199 received. As the flag value will be lost between multiple
200 invocations of ssl3_accept when using non-blocking I/O, the
201 function may not be aware that a handshake has actually taken
202 place, thus preventing a new session from being added to the
205 To avoid this problem, we now set s->new_session to 2 instead of
206 using a local variable.
207 [Lutz Jaenicke, Bodo Moeller]
209 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
210 if the SSL_R_LENGTH_MISMATCH error is detected.
211 [Geoff Thorpe, Bodo Moeller]
213 +) New experimental application configuration code.
216 *) New 'shared_ldflag' column in Configure platform table.
219 *) Fix EVP_CIPHER_mode macro.
220 ["Dan S. Camper" <dan@bti.net>]
222 +) Change the AES code to follow the same name structure as all other
223 symmetric ciphers, and behave the same way. Move everything to
224 the directory crypto/aes, thereby obsoleting crypto/rijndael.
225 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
227 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
228 type, we must throw them away by setting rr->length to 0.
229 [D P Chang <dpc@qualys.com>]
231 -) OpenSSL 0.9.6c released [21 dec 2001]
233 +) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
234 [Ben Laurie and Theo de Raadt]
236 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
237 <Dominikus.Scherkl@biodata.com>. (The previous implementation
238 worked incorrectly for those cases where range = 10..._2 and
239 3*range is two bits longer than range.)
242 *) Only add signing time to PKCS7 structures if it is not already
246 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
247 OBJ_ld_ce should be OBJ_id_ce.
248 Also some ip-pda OIDs in crypto/objects/objects.txt were
249 incorrect (cf. RFC 3039).
250 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
252 +) Add option to output public keys in req command.
253 [Massimiliano Pala madwolf@openca.org]
255 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
256 returns early because it has nothing to do.
257 [Andy Schneider <andy.schneider@bjss.co.uk>]
259 *) [In 0.9.6c-engine and 0.9.7 release:]
260 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
261 [Andy Schneider <andy.schneider@bjss.co.uk>]
263 -) [In 0.9.6c-engine release:]
264 Add support for Cryptographic Appliance's keyserver technology.
265 (Use engine 'keyclient')
266 [Cryptographic Appliances and Geoff Thorpe]
268 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
269 is called via tools/c89.sh because arguments have to be
270 rearranged (all '-L' options must appear before the first object
272 [Richard Shapiro <rshapiro@abinitio.com>]
274 +) Use wNAFs in EC_POINTs_mul() for improved efficiency
275 (up to about 10% better than before for P-192 and P-224).
278 -) [In 0.9.6c-engine release:]
279 Add support for Broadcom crypto accelerator cards, backported
281 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
283 -) [In 0.9.6c-engine release:]
284 Add support for SureWare crypto accelerator cards from
285 Baltimore Technologies. (Use engine 'sureware')
286 [Baltimore Technologies and Mark Cox]
288 -) [In 0.9.6c-engine release:]
289 Add support for crypto accelerator cards from Accelerated
290 Encryption Processing, www.aep.ie. (Use engine 'aep')
291 [AEP Inc. and Mark Cox]
293 *) Add a configuration entry for gcc on UnixWare.
294 [Gary Benson <gbenson@redhat.com>]
296 +) New functions/macros
298 SSL_CTX_set_msg_callback(ctx, cb)
299 SSL_CTX_set_msg_callback_arg(ctx, arg)
300 SSL_set_msg_callback(ssl, cb)
301 SSL_set_msg_callback_arg(ssl, arg)
303 to request calling a callback function
305 void cb(int write_p, int version, int content_type,
306 const void *buf, size_t len, SSL *ssl, void *arg)
308 whenever a protocol message has been completely received
309 (write_p == 0) or sent (write_p == 1). Here 'version' is the
310 protocol version according to which the SSL library interprets
311 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
312 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
313 the content type as defined in the SSL 3.0/TLS 1.0 protocol
314 specification (change_cipher_spec(20), alert(21), handshake(22)).
315 'buf' and 'len' point to the actual message, 'ssl' to the
316 SSL object, and 'arg' is the application-defined value set by
317 SSL[_CTX]_set_msg_callback_arg().
319 'openssl s_client' and 'openssl s_server' have new '-msg' options
320 to enable a callback that displays all protocol messages.
323 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
324 messages are stored in a single piece (fixed-length part and
325 variable-length part combined) and fix various bugs found on the way.
328 +) Change the shared library support so shared libraries are built as
329 soon as the corresponding static library is finished, and thereby get
330 openssl and the test programs linked against the shared library.
331 This still only happens when the keyword "shard" has been given to
332 the configuration scripts.
334 NOTE: shared library support is still an experimental thing, and
335 backward binary compatibility is still not guaranteed.
336 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
338 +) Add support for Subject Information Access extension.
339 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
341 +) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
342 additional bytes when new memory had to be allocated, not just
343 when reusing an existing buffer.
346 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
347 instead. BIO_gethostbyname() does not know what timeouts are
348 appropriate, so entries would stay in cache even when they have
350 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
352 +) New command line and configuration option 'utf8' for the req command.
353 This allows field values to be specified as UTF8 strings.
356 +) Add -multi and -mr options to "openssl speed" - giving multiple parallel
357 runs for the former and machine-readable output for the latter.
360 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
361 of the e-mail address in the DN (i.e., it will go into a certificate
362 extension only). The new configuration file option 'email_in_dn = no'
364 [Massimiliano Pala madwolf@openca.org]
366 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
367 faced with a pathologically small ClientHello fragment that does
368 not contain client_version: Instead of aborting with an error,
369 simply choose the highest available protocol version (i.e.,
370 TLS 1.0 unless it is disabled). In practice, ClientHello
371 messages are never sent like this, but this change gives us
372 strictly correct behaviour at least for TLS.
375 +) Change all functions with names starting with des_ to be starting
376 with DES_ instead. Add wrappers that are compatible with libdes,
377 but are named _ossl_old_des_*. Finally, add macros that map the
378 des_* symbols to the corresponding _ossl_old_des_* if libdes
379 compatibility is desired. If OpenSSL 0.9.6c compatibility is
380 desired, the des_* symbols will be mapped to DES_*, with one
383 Since we provide two compatibility mappings, the user needs to
384 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
385 compatibility is desired. The default (i.e., when that macro
386 isn't defined) is OpenSSL 0.9.6c compatibility.
388 There are also macros that enable and disable the support of old
389 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
390 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
391 are defined, the default will apply: to support the old des routines.
393 In either case, one must include openssl/des.h to get the correct
394 definitions. Do not try to just include openssl/des_old.h, that
397 NOTE: This is a major break of an old API into a new one. Software
398 authors are encouraged to switch to the DES_ style functions. Some
399 time in the future, des_old.h and the libdes compatibility functions
400 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
401 default), and then completely removed.
404 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
405 never resets s->method to s->ctx->method when called from within
406 one of the SSL handshake functions.
407 [Bodo Moeller; problem pointed out by Niko Baric]
409 +) Test for certificates which contain unsupported critical extensions.
410 If such a certificate is found during a verify operation it is
411 rejected by default: this behaviour can be overridden by either
412 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
413 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
414 X509_supported_extension() has also been added which returns 1 if a
415 particular extension is supported.
418 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
419 (sent using the client's version number) if client_version is
420 smaller than the protocol version in use. Also change
421 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
422 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
423 the client will at least see that alert.
426 +) Modify the behaviour of EVP cipher functions in similar way to digests
427 to retain compatibility with existing code.
430 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
431 compatibility with existing code. In particular the 'ctx' parameter does
432 not have to be to be initialized before the call to EVP_DigestInit() and
433 it is tidied up after a call to EVP_DigestFinal(). New function
434 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
435 EVP_MD_CTX_copy() changed to not require the destination to be
436 initialized valid and new function EVP_MD_CTX_copy_ex() added which
437 requires the destination to be valid.
439 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
440 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
443 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
444 so that complete 'Handshake' protocol structures are kept in memory
445 instead of overwriting 'msg_type' and 'length' with 'body' data.
448 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
452 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
453 [Massimo Santin via Richard Levitte]
455 +) Major restructuring to the underlying ENGINE code. This includes
456 reduction of linker bloat, separation of pure "ENGINE" manipulation
457 (initialisation, etc) from functionality dealing with implementations
458 of specific crypto iterfaces. This change also introduces integrated
459 support for symmetric ciphers and digest implementations - so ENGINEs
460 can now accelerate these by providing EVP_CIPHER and EVP_MD
461 implementations of their own. This is detailed in crypto/engine/README
462 as it couldn't be adequately described here. However, there are a few
463 API changes worth noting - some RSA, DSA, DH, and RAND functions that
464 were changed in the original introduction of ENGINE code have now
465 reverted back - the hooking from this code to ENGINE is now a good
466 deal more passive and at run-time, operations deal directly with
467 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
468 dereferencing through an ENGINE pointer any more. Also, the ENGINE
469 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
470 they were not being used by the framework as there is no concept of a
471 BIGNUM_METHOD and they could not be generalised to the new
472 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
473 ENGINE_cpy() has been removed as it cannot be consistently defined in
477 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
480 +) Change mkdef.pl to sort symbols that get the same entry number,
481 and make sure the automatically generated functions ERR_load_*
482 become part of libeay.num as well.
485 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
486 client receives HelloRequest while in a handshake.
487 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
489 +) New function SSL_renegotiate_pending(). This returns true once
490 renegotiation has been requested (either SSL_renegotiate() call
491 or HelloRequest/ClientHello receveived from the peer) and becomes
492 false once a handshake has been completed.
493 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
494 sends a HelloRequest, but does not ensure that a handshake takes
495 place. SSL_renegotiate_pending() is useful for checking if the
496 client has followed the request.)
499 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
500 By default, clients may request session resumption even during
501 renegotiation (if session ID contexts permit); with this option,
502 session resumption is possible only in the first handshake.
505 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
506 should end in 'break', not 'goto end' which circuments various
507 cleanups done in state SSL_ST_OK. But session related stuff
508 must be disabled for SSL_ST_OK in the case that we just sent a
511 Also avoid some overhead by not calling ssl_init_wbio_buffer()
512 before just sending a HelloRequest.
513 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
515 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
516 reveal whether illegal block cipher padding was found or a MAC
517 verification error occured. (Neither SSLerr() codes nor alerts
518 are directly visible to potential attackers, but the information
519 may leak via logfiles.)
521 Similar changes are not required for the SSL 2.0 implementation
522 because the number of padding bytes is sent in clear for SSL 2.0,
523 and the extra bytes are just ignored. However ssl/s2_pkt.c
524 failed to verify that the purported number of padding bytes is in
528 +) Add some demos for certificate and certificate request creation.
531 +) Make maximum certificate chain size accepted from the peer application
532 settable (SSL*_get/set_max_cert_list()), as proposed by
533 "Douglas E. Engert" <deengert@anl.gov>.
536 +) Add support for shared libraries for Unixware-7
537 (Boyd Lynn Gerber <gerberb@zenez.com>).
540 *) Add OpenUNIX-8 support including shared libraries
541 (Boyd Lynn Gerber <gerberb@zenez.com>).
544 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
545 'wristwatch attack' using huge encoding parameters (cf.
546 James H. Manger's CRYPTO 2001 paper). Note that the
547 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
548 encoding parameters and hence was not vulnerable.
551 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
552 be done prior to destruction. Use this to unload error strings from
553 ENGINEs that load their own error strings. NB: This adds two new API
554 functions to "get" and "set" this destroy handler in an ENGINE.
557 +) Alter all existing ENGINE implementations (except "openssl" and
558 "openbsd") to dynamically instantiate their own error strings. This
559 makes them more flexible to be built both as statically-linked ENGINEs
560 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
561 Also, add stub code to each that makes building them as self-contained
562 shared-libraries easier (see README.ENGINE).
565 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
566 implementations into applications that are completely implemented in
567 self-contained shared-libraries. The "dynamic" ENGINE exposes control
568 commands that can be used to configure what shared-library to load and
569 to control aspects of the way it is handled. Also, made an update to
570 the README.ENGINE file that brings its information up-to-date and
571 provides some information and instructions on the "dynamic" ENGINE
572 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
576 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
578 +) Make it possible to unload ranges of ERR strings with a new
579 "ERR_unload_strings" function.
582 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
583 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
584 followed by modular reduction.
585 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
587 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
588 equivalent based on BN_pseudo_rand() instead of BN_rand().
591 +) Add a copy() function to EVP_MD.
594 +) Make EVP_MD routines take a context pointer instead of just the
595 md_data void pointer.
598 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
599 that the digest can only process a single chunk of data
600 (typically because it is provided by a piece of
601 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
602 is only going to provide a single chunk of data, and hence the
603 framework needn't accumulate the data for oneshot drivers.
606 +) As with "ERR", make it possible to replace the underlying "ex_data"
607 functions. This change also alters the storage and management of global
608 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
609 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
610 index counters. The API functions that use this state have been changed
611 to take a "class_index" rather than pointers to the class's local STACK
612 and counter, and there is now an API function to dynamically create new
613 classes. This centralisation allows us to (a) plug a lot of the
614 thread-safety problems that existed, and (b) makes it possible to clean
615 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
616 such data would previously have always leaked in application code and
617 workarounds were in place to make the memory debugging turn a blind eye
618 to it. Application code that doesn't use this new function will still
619 leak as before, but their memory debugging output will announce it now
620 rather than letting it slide.
622 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
623 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
624 has a return value to indicate success or failure.
627 +) Make it possible to replace the underlying "ERR" functions such that the
628 global state (2 LHASH tables and 2 locks) is only used by the "default"
629 implementation. This change also adds two functions to "get" and "set"
630 the implementation prior to it being automatically set the first time
631 any other ERR function takes place. Ie. an application can call "get",
632 pass the return value to a module it has just loaded, and that module
633 can call its own "set" function using that value. This means the
634 module's "ERR" operations will use (and modify) the error state in the
635 application and not in its own statically linked copy of OpenSSL code.
638 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
639 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
640 the operation, and provides a more encapsulated way for external code
641 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
642 to use these functions rather than manually incrementing the counts.
644 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
647 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
648 This function was broken, as the check for a new client hello message
649 to handle SGC did not allow these large messages.
650 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
653 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
656 +) Add EVP test program.
659 +) Add symmetric cipher support to ENGINE. Expect the API to change!
662 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
663 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
664 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
665 These allow a CRL to be built without having to access X509_CRL fields
666 directly. Modify 'ca' application to use new functions.
669 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
670 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
673 *) Rework the configuration and shared library support for Tru64 Unix.
674 The configuration part makes use of modern compiler features and
675 still retains old compiler behavior for those that run older versions
676 of the OS. The shared library support part includes a variant that
677 uses the RPATH feature, and is available through the special
678 configuration target "alpha-cc-rpath", which will never be selected
680 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
682 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
683 with the same message size as in ssl3_get_certificate_request().
684 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
685 messages might inadvertently be reject as too long.
686 [Petr Lampa <lampa@fee.vutbr.cz>]
688 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
689 bug workarounds. Rollback attack detection is a security feature.
690 The problem will only arise on OpenSSL servers when TLSv1 is not
691 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
692 Software authors not wanting to support TLSv1 will have special reasons
693 for their choice and can explicitly enable this option.
694 [Bodo Moeller, Lutz Jaenicke]
696 +) Rationalise EVP so it can be extended: don't include a union of
697 cipher/digest structures, add init/cleanup functions. This also reduces
698 the number of header dependencies.
703 EVP_MD_CTX_init(&md); /* new function call */
704 EVP_DigestInit(&md, EVP_sha1());
705 EVP_DigestUpdate(&md, in, len);
706 EVP_DigestFinal(&md, out, NULL);
707 EVP_MD_CTX_cleanup(&md); /* new function call */
711 +) Make DES key schedule conform to the usual scheme, as well as
712 correcting its structure. This means that calls to DES functions
713 now have to pass a pointer to a des_key_schedule instead of a
714 plain des_key_schedule (which was actually always a pointer
719 des_set_key_checked(..., &ks);
720 des_ncbc_encrypt(..., &ks, ...);
722 (Note that a later change renames 'des_...' into 'DES_...'.)
725 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
728 *) Modified SSL library such that the verify_callback that has been set
729 specificly for an SSL object with SSL_set_verify() is actually being
730 used. Before the change, a verify_callback set with this function was
731 ignored and the verify_callback() set in the SSL_CTX at the time of
732 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
733 to allow the necessary settings.
736 +) Initial reduction of linker bloat: the use of some functions, such as
737 PEM causes large amounts of unused functions to be linked in due to
738 poor organisation. For example pem_all.c contains every PEM function
739 which has a knock on effect of linking in large amounts of (unused)
740 ASN1 code. Grouping together similar functions and splitting unrelated
741 functions prevents this.
744 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
745 explicitly to NULL, as at least on Solaris 8 this seems not always to be
746 done automatically (in contradiction to the requirements of the C
747 standard). This made problems when used from OpenSSH.
750 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
751 dh->length and always used
753 BN_rand_range(priv_key, dh->p).
755 BN_rand_range() is not necessary for Diffie-Hellman, and this
756 specific range makes Diffie-Hellman unnecessarily inefficient if
757 dh->length (recommended exponent length) is much smaller than the
758 length of dh->p. We could use BN_rand_range() if the order of
759 the subgroup was stored in the DH structure, but we only have
764 BN_rand(priv_key, l, ...)
766 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
772 RSA_eay_public_encrypt
773 RSA_eay_private_decrypt
774 RSA_eay_private_encrypt (signing)
775 RSA_eay_public_decrypt (signature verification)
777 (default implementations for RSA_public_encrypt,
778 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
779 always reject numbers >= n.
782 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
783 to synchronize access to 'locking_thread'. This is necessary on
784 systems where access to 'locking_thread' (an 'unsigned long'
785 variable) is not atomic.
788 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
789 *before* setting the 'crypto_lock_rand' flag. The previous code had
790 a race condition if 0 is a valid thread ID.
791 [Travis Vitek <vitek@roguewave.com>]
793 +) Cleanup of EVP macros.
796 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
800 +) Add initial OCSP responder support to ocsp application. The
801 revocation information is handled using the text based index
802 use by the ca application. The responder can either handle
803 requests generated internally, supplied in files (for example
804 via a CGI script) or using an internal minimal server.
807 +) Add configuration choices to get zlib compression for TLS.
810 +) Changes to Kerberos SSL for RFC 2712 compliance:
811 1. Implemented real KerberosWrapper, instead of just using
812 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
813 2. Implemented optional authenticator field of KerberosWrapper.
815 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
816 and authenticator structs; see crypto/krb5/.
818 Generalized Kerberos calls to support multiple Kerberos libraries.
819 [Vern Staats <staatsvr@asc.hpc.mil>,
820 Jeffrey Altman <jaltman@columbia.edu>
823 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
824 already does with RSA. testdsa.h now has 'priv_key/pub_key'
825 values for each of the key sizes rather than having just
826 parameters (and 'speed' generating keys each time).
829 *) Add support for shared libraries under Irix.
830 [Albert Chin-A-Young <china@thewrittenword.com>]
832 *) Add configuration option to build on Linux on both big-endian and
834 [Ralf Baechle <ralf@uni-koblenz.de>]
836 *) Add the possibility to create shared libraries on HP-UX.
839 -) OpenSSL 0.9.6b released [9 July 2001]
841 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
842 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
843 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
844 PRNG state recovery was possible based on the output of
845 one PRNG request appropriately sized to gain knowledge on
846 'md' followed by enough consecutive 1-byte PRNG requests
847 to traverse all of 'state'.
849 1. When updating 'md_local' (the current thread's copy of 'md')
850 during PRNG output generation, hash all of the previous
851 'md_local' value, not just the half used for PRNG output.
853 2. Make the number of bytes from 'state' included into the hash
854 independent from the number of PRNG bytes requested.
856 The first measure alone would be sufficient to avoid
857 Markku-Juhani's attack. (Actually it had never occurred
858 to me that the half of 'md_local' used for chaining was the
859 half from which PRNG output bytes were taken -- I had always
860 assumed that the secret half would be used.) The second
861 measure makes sure that additional data from 'state' is never
862 mixed into 'md_local' in small portions; this heuristically
863 further strengthens the PRNG.
866 +) Speed up EVP routines.
869 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
870 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
871 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
872 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
874 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
875 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
876 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
879 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
881 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
884 *) Fix crypto/bn/asm/mips3.s.
887 *) When only the key is given to "enc", the IV is undefined. Print out
888 an error message in this case.
891 +) Added the OS2-EMX target.
892 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
894 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
895 to support NCONF routines in extension code. New function CONF_set_nconf()
896 to allow functions which take an NCONF to also handle the old LHASH
897 structure: this means that the old CONF compatible routines can be
898 retained (in particular wrt extensions) without having to duplicate the
899 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
902 *) Handle special case when X509_NAME is empty in X509 printing routines.
905 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
906 positive and less than q.
909 +) Enhance the general user interface with mechanisms for inner control
910 and with possibilities to have yes/no kind of prompts.
913 +) Change all calls to low level digest routines in the library and
914 applications to use EVP. Add missing calls to HMAC_cleanup() and
915 don't assume HMAC_CTX can be copied using memcpy().
916 [Verdon Walker <VWalker@novell.com>, Steve Henson]
918 +) Add the possibility to control engines through control names but with
919 arbitrary arguments instead of just a string.
920 Change the key loaders to take a UI_METHOD instead of a callback
921 function pointer. NOTE: this breaks binary compatibility with earlier
922 versions of OpenSSL [engine].
923 Adapt the nCipher code for these new conditions and add a card insertion
927 +) Enhance the general user interface with mechanisms to better support
928 dialog box interfaces, application-defined prompts, the possibility
929 to use defaults (for example default passwords from somewhere else)
930 and interrupts/cancellations.
933 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
934 used: it isn't thread safe and the add_lock_callback should handle
936 [Paul Rose <Paul.Rose@bridge.com>]
938 *) Verify that incoming data obeys the block size in
939 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
942 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
943 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
947 [Ulf Möller, Bodo Möller]
949 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
950 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
951 when fixing the server behaviour for backwards-compatible 'client
952 hello' messages. (Note that the attack is impractical against
953 SSL 3.0 and TLS 1.0 anyway because length and version checking
954 means that the probability of guessing a valid ciphertext is
955 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
958 Before 0.9.5, the countermeasure (hide the error by generating a
959 random 'decryption result') did not work properly because
960 ERR_clear_error() was missing, meaning that SSL_get_error() would
961 detect the supposedly ignored error.
963 Both problems are now fixed.
966 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
967 (previously it was 1024).
970 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
971 tidy up some unnecessarily weird code in 'sk_new()').
972 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
974 +) Change the key loading routines for ENGINEs to use the same kind
975 callback (pem_password_cb) as all other routines that need this
979 *) Fix for compatibility mode trust settings: ignore trust settings
980 unless some valid trust or reject settings are present.
983 *) Fix for blowfish EVP: its a variable length cipher.
986 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
987 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
988 than this minimum value is recommended.
991 +) New random seeder for OpenVMS, using the system process statistics
992 that are easily reachable.
995 +) Windows apparently can't transparently handle global
996 variables defined in DLLs. Initialisations such as:
998 const ASN1_ITEM *it = &ASN1_INTEGER_it;
1000 wont compile. This is used by the any applications that need to
1001 declare their own ASN1 modules. This was fixed by adding the option
1002 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
1003 needed for static libraries under Win32.
1006 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
1007 setting of purpose and trust fields. New X509_STORE trust and
1008 purpose functions and tidy up setting in other SSL functions.
1011 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
1012 structure. These are inherited by X509_STORE_CTX when it is
1013 initialised. This allows various defaults to be set in the
1014 X509_STORE structure (such as flags for CRL checking and custom
1015 purpose or trust settings) for functions which only use X509_STORE_CTX
1016 internally such as S/MIME.
1018 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
1019 trust settings if they are not set in X509_STORE. This allows X509_STORE
1020 purposes and trust (in S/MIME for example) to override any set by default.
1022 Add command line options for CRL checking to smime, s_client and s_server
1026 +) Initial CRL based revocation checking. If the CRL checking flag(s)
1027 are set then the CRL is looked up in the X509_STORE structure and
1028 its validity and signature checked, then if the certificate is found
1029 in the CRL the verify fails with a revoked error.
1031 Various new CRL related callbacks added to X509_STORE_CTX structure.
1033 Command line options added to 'verify' application to support this.
1035 This needs some additional work, such as being able to handle multiple
1036 CRLs with different times, extension based lookup (rather than just
1037 by subject name) and ultimately more complete V2 CRL extension
1041 +) Add a general user interface API (crypto/ui/). This is designed
1042 to replace things like des_read_password and friends (backward
1043 compatibility functions using this new API are provided).
1044 The purpose is to remove prompting functions from the DES code
1045 section as well as provide for prompting through dialog boxes in
1046 a window system and the like.
1049 *) Fix various bugs related to DSA S/MIME verification. Handle missing
1050 parameters in DSA public key structures and return an error in the
1051 DSA routines if parameters are absent.
1054 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
1055 in the current directory if neither $RANDFILE nor $HOME was set.
1056 RAND_file_name() in 0.9.6a returned NULL in this case. This has
1057 caused some confusion to Windows users who haven't defined $HOME.
1058 Thus RAND_file_name() is changed again: e_os.h can define a
1059 DEFAULT_HOME, which will be used if $HOME is not set.
1060 For Windows, we use "C:"; on other platforms, we still require
1061 environment variables.
1063 +) Add "ex_data" support to ENGINE so implementations can add state at a
1064 per-structure level rather than having to store it globally.
1067 +) Make it possible for ENGINE structures to be copied when retrieved by
1068 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
1069 This causes the "original" ENGINE structure to act like a template,
1070 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
1071 operational state can be localised to each ENGINE structure, despite the
1072 fact they all share the same "methods". New ENGINE structures returned in
1073 this case have no functional references and the return value is the single
1074 structural reference. This matches the single structural reference returned
1075 by ENGINE_by_id() normally, when it is incremented on the pre-existing
1079 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
1080 needs to match any other type at all we need to manually clear the
1084 +) Changes to the "openssl engine" utility to include;
1085 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
1086 about an ENGINE's available control commands.
1087 - executing control commands from command line arguments using the
1088 '-pre' and '-post' switches. '-post' is only used if '-t' is
1089 specified and the ENGINE is successfully initialised. The syntax for
1090 the individual commands are colon-separated, for example;
1091 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
1094 +) New dynamic control command support for ENGINEs. ENGINEs can now
1095 declare their own commands (numbers), names (strings), descriptions,
1096 and input types for run-time discovery by calling applications. A
1097 subset of these commands are implicitly classed as "executable"
1098 depending on their input type, and only these can be invoked through
1099 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
1100 can be based on user input, config files, etc). The distinction is
1101 that "executable" commands cannot return anything other than a boolean
1102 result and can only support numeric or string input, whereas some
1103 discoverable commands may only be for direct use through
1104 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
1105 pointers, or other custom uses. The "executable" commands are to
1106 support parameterisations of ENGINE behaviour that can be
1107 unambiguously defined by ENGINEs and used consistently across any
1108 OpenSSL-based application. Commands have been added to all the
1109 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
1110 control over shared-library paths without source code alterations.
1113 +) Changed all ENGINE implementations to dynamically allocate their
1114 ENGINEs rather than declaring them statically. Apart from this being
1115 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
1116 this also allows the implementations to compile without using the
1117 internal engine_int.h header.
1120 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
1121 'const' value. Any code that should be able to modify a RAND_METHOD
1122 should already have non-const pointers to it (ie. they should only
1123 modify their own ones).
1126 +) Made a variety of little tweaks to the ENGINE code.
1127 - "atalla" and "ubsec" string definitions were moved from header files
1128 to C code. "nuron" string definitions were placed in variables
1129 rather than hard-coded - allowing parameterisation of these values
1130 later on via ctrl() commands.
1131 - Removed unused "#if 0"'d code.
1132 - Fixed engine list iteration code so it uses ENGINE_free() to release
1133 structural references.
1134 - Constified the RAND_METHOD element of ENGINE structures.
1135 - Constified various get/set functions as appropriate and added
1136 missing functions (including a catch-all ENGINE_cpy that duplicates
1137 all ENGINE values onto a new ENGINE except reference counts/state).
1138 - Removed NULL parameter checks in get/set functions. Setting a method
1139 or function to NULL is a way of cancelling out a previously set
1140 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1141 and doesn't justify the extra error symbols and code.
1142 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1143 flags from engine_int.h to engine.h.
1144 - Changed prototypes for ENGINE handler functions (init(), finish(),
1145 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1148 *) Move 'if (!initialized) RAND_poll()' into regions protected by
1149 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
1150 having multiple threads call RAND_poll() concurrently.
1153 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
1154 combination of a flag and a thread ID variable.
1155 Otherwise while one thread is in ssleay_rand_bytes (which sets the
1156 flag), *other* threads can enter ssleay_add_bytes without obeying
1157 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
1158 that they do not hold after the first thread unsets add_do_not_lock).
1161 +) Implement binary inversion algorithm for BN_mod_inverse in addition
1162 to the algorithm using long division. The binary algorithm can be
1163 used only if the modulus is odd. On 32-bit systems, it is faster
1164 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1165 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1166 up to 450 bits. In 64-bit environments, the binary algorithm
1167 appears to be advantageous for much longer moduli; here we use it
1168 for moduli up to 2048 bits.
1171 *) Change bctest again: '-x' expressions are not available in all
1175 -) OpenSSL 0.9.6a released [5 Apr 2001]
1177 *) Fix a couple of memory leaks in PKCS7_dataDecode()
1178 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
1180 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
1181 the default extension for executables, if any. Also, make the perl
1182 scripts that use symlink() to test if it really exists and use "cp"
1183 if it doesn't. All this made OpenSSL compilable and installable in
1187 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1188 could not support the combine flag in choice fields.
1191 -) Fix for asn1_GetSequence() for indefinite length constructed data.
1192 If SEQUENCE is length is indefinite just set c->slen to the total
1193 amount of data available.
1194 [Steve Henson, reported by shige@FreeBSD.org]
1195 [This change does not apply to 0.9.7.]
1197 *) Change bctest to avoid here-documents inside command substitution
1198 (workaround for FreeBSD /bin/sh bug).
1199 For compatibility with Ultrix, avoid shell functions (introduced
1200 in the bctest version that searches along $PATH).
1203 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
1204 with des_encrypt() defined on some operating systems, like Solaris
1208 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
1209 On the Importance of Eliminating Errors in Cryptographic
1210 Computations, J. Cryptology 14 (2001) 2, 101-119,
1211 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
1214 *) MIPS assembler BIGNUM division bug fix.
1217 *) Disabled incorrect Alpha assembler code.
1220 -) Fix PKCS#7 decode routines so they correctly update the length
1221 after reading an EOC for the EXPLICIT tag.
1223 [This change does not apply to 0.9.7.]
1225 *) Fix bug in PKCS#12 key generation routines. This was triggered
1226 if a 3DES key was generated with a 0 initial byte. Include
1227 PKCS12_BROKEN_KEYGEN compilation option to retain the old
1228 (but broken) behaviour.
1231 *) Enhance bctest to search for a working bc along $PATH and print
1233 [Tim Rice <tim@multitalents.net> via Richard Levitte]
1235 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
1236 extensions from a certificate request to the certificate.
1239 +) Allow multiple 'certopt' and 'nameopt' options to be separated
1240 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1241 file: this allows the display of the certificate about to be
1242 signed to be customised, to allow certain fields to be included
1243 or excluded and extension details. The old system didn't display
1244 multicharacter strings properly, omitted fields not in the policy
1245 and couldn't display additional details such as extensions.
1248 *) Fix memory leaks in err.c: free err_data string if necessary;
1249 don't write to the wrong index in ERR_set_error_data.
1252 +) Function EC_POINTs_mul for multiple scalar multiplication
1253 of an arbitrary number of elliptic curve points
1254 \sum scalars[i]*points[i],
1255 optionally including the generator defined for the EC_GROUP:
1256 scalar*generator + \sum scalars[i]*points[i].
1258 EC_POINT_mul is a simple wrapper function for the typical case
1259 that the point list has just one item (besides the optional
1263 +) First EC_METHODs for curves over GF(p):
1265 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1266 operations and provides various method functions that can also
1267 operate with faster implementations of modular arithmetic.
1269 EC_GFp_mont_method() reuses most functions that are part of
1270 EC_GFp_simple_method, but uses Montgomery arithmetic.
1272 [Bodo Moeller; point addition and point doubling
1273 implementation directly derived from source code provided by
1274 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1276 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1277 crypto/ec/ec_lib.c):
1279 Curves are EC_GROUP objects (with an optional group generator)
1280 based on EC_METHODs that are built into the library.
1282 Points are EC_POINT objects based on EC_GROUP objects.
1284 Most of the framework would be able to handle curves over arbitrary
1285 finite fields, but as there are no obvious types for fields other
1286 than GF(p), some functions are limited to that for now.
1289 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1290 that the file contains a complete HTTP response.
1293 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1294 change the def and num file printf format specifier from "%-40sXXX"
1295 to "%-39s XXX". The latter will always guarantee a space after the
1296 field while the former will cause them to run together if the field
1297 is 40 of more characters long.
1300 +) Constify the cipher and digest 'method' functions and structures
1301 and modify related functions to take constant EVP_MD and EVP_CIPHER
1305 *) Implement ssl23_peek (analogous to ssl23_read), which previously
1309 *) Replace rdtsc with _emit statements for VC++ version 5.
1310 [Jeremy Cooper <jeremy@baymoo.org>]
1312 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1313 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1316 +) Modify EVP_Digest*() routines so they now return values. Although the
1317 internal software routines can never fail additional hardware versions
1321 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1323 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1324 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1327 ERR_R_NESTED_ASN1_ERROR
1329 ERR_R_MISSING_ASN1_EOS
1330 were 4 .. 9, conflicting with
1331 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1333 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1334 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1336 Add new error code 'ERR_R_INTERNAL_ERROR'.
1339 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1343 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1344 sets the subject name for a new request or supersedes the
1345 subject name in a given request. Formats that can be parsed are
1346 'CN=Some Name, OU=myOU, C=IT'
1348 'CN=Some Name/OU=myOU/C=IT'.
1350 Add options '-batch' and '-verbose' to 'openssl req'.
1351 [Massimiliano Pala <madwolf@hackmasters.net>]
1353 *) Make it possible to reuse SSLv2 sessions.
1356 +) Introduce the possibility to access global variables through
1357 functions on platform were that's the best way to handle exporting
1358 global variables in shared libraries. To enable this functionality,
1359 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1360 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1361 is normally done by Configure or something similar).
1363 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1364 in the source file (foo.c) like this:
1366 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1367 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1369 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1370 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1372 OPENSSL_DECLARE_GLOBAL(int,foo);
1373 #define foo OPENSSL_GLOBAL_REF(foo)
1374 OPENSSL_DECLARE_GLOBAL(double,bar);
1375 #define bar OPENSSL_GLOBAL_REF(bar)
1377 The #defines are very important, and therefore so is including the
1378 header file everywhere where the defined globals are used.
1380 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1381 of ASN.1 items, but that structure is a bit different.
1383 The largest change is in util/mkdef.pl which has been enhanced with
1384 better and easier to understand logic to choose which symbols should
1385 go into the Windows .def files as well as a number of fixes and code
1386 cleanup (among others, algorithm keywords are now sorted
1387 lexicographically to avoid constant rewrites).
1390 *) In copy_email() check for >= 0 as a return value for
1391 X509_NAME_get_index_by_NID() since 0 is a valid index.
1392 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1394 +) In BN_div() keep a copy of the sign of 'num' before writing the
1395 result to 'rm' because if rm==num the value will be overwritten
1396 and produce the wrong result if 'num' is negative: this caused
1397 problems with BN_mod() and BN_nnmod().
1400 +) Function OCSP_request_verify(). This checks the signature on an
1401 OCSP request and verifies the signer certificate. The signer
1402 certificate is just checked for a generic purpose and OCSP request
1406 +) Add OCSP_check_validity() function to check the validity of OCSP
1407 responses. OCSP responses are prepared in real time and may only
1408 be a few seconds old. Simply checking that the current time lies
1409 between thisUpdate and nextUpdate max reject otherwise valid responses
1410 caused by either OCSP responder or client clock inaccuracy. Instead
1411 we allow thisUpdate and nextUpdate to fall within a certain period of
1412 the current time. The age of the response can also optionally be
1413 checked. Two new options -validity_period and -status_age added to
1417 +) If signature or public key algorithm is unrecognized print out its
1418 OID rather that just UNKNOWN.
1421 *) Avoid coredump with unsupported or invalid public keys by checking if
1422 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1423 PKCS7_verify() fails with non detached data.
1426 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1427 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1428 ID to be generated from the issuer certificate alone which can then be
1429 passed to OCSP_id_issuer_cmp().
1432 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1433 ASN1 modules to export functions returning ASN1_ITEM pointers
1434 instead of the ASN1_ITEM structures themselves. This adds several
1435 new macros which allow the underlying ASN1 function/structure to
1436 be accessed transparently. As a result code should not use ASN1_ITEM
1437 references directly (such as &X509_it) but instead use the relevant
1438 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1439 use of the new ASN1 code on platforms where exporting structures
1440 is problematical (for example in shared libraries) but exporting
1441 functions returning pointers to structures is not.
1444 +) Add support for overriding the generation of SSL/TLS session IDs.
1445 These callbacks can be registered either in an SSL_CTX or per SSL.
1446 The purpose of this is to allow applications to control, if they wish,
1447 the arbitrary values chosen for use as session IDs, particularly as it
1448 can be useful for session caching in multiple-server environments. A
1449 command-line switch for testing this (and any client code that wishes
1450 to use such a feature) has been added to "s_server".
1451 [Geoff Thorpe, Lutz Jaenicke]
1453 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1454 of the form '#if defined(...) || defined(...) || ...' and
1455 '#if !defined(...) && !defined(...) && ...'. This also avoids
1456 the growing number of special cases it was previously handling.
1459 *) Don't use getenv in library functions when run as setuid/setgid.
1460 New function OPENSSL_issetugid().
1463 +) Make all configuration macros available for application by making
1464 sure they are available in opensslconf.h, by giving them names starting
1465 with "OPENSSL_" to avoid conflicts with other packages and by making
1466 sure e_os2.h will cover all platform-specific cases together with
1468 Additionally, it is now possible to define configuration/platform-
1469 specific names (called "system identities"). In the C code, these
1470 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1471 macro with the name beginning with "OPENSSL_SYS_", which is determined
1472 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1476 +) New option -set_serial to 'req' and 'x509' this allows the serial
1477 number to use to be specified on the command line. Previously self
1478 signed certificates were hard coded with serial number 0 and the
1479 CA options of 'x509' had to use a serial number in a file which was
1483 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1484 due to incorrect handling of multi-threading:
1486 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1488 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1490 3. Count how many times MemCheck_off() has been called so that
1491 nested use can be treated correctly. This also avoids
1492 inband-signalling in the previous code (which relied on the
1493 assumption that thread ID 0 is impossible).
1496 +) New options to 'ca' utility to support V2 CRL entry extensions.
1497 Currently CRL reason, invalidity date and hold instruction are
1498 supported. Add new CRL extensions to V3 code and some new objects.
1501 *) Add "-rand" option also to s_client and s_server.
1504 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1505 disable standard block padding (aka PKCS#5 padding) in the EVP
1506 API, which was previously mandatory. This means that the data is
1507 not padded in any way and so the total length much be a multiple
1508 of the block size, otherwise an error occurs.
1511 +) Initial (incomplete) OCSP SSL support.
1514 *) Fix CPU detection on Irix 6.x.
1515 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1516 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1518 +) New function OCSP_parse_url(). This splits up a URL into its host,
1519 port and path components: primarily to parse OCSP URLs. New -url
1520 option to ocsp utility.
1523 +) New nonce behavior. The return value of OCSP_check_nonce() now
1524 reflects the various checks performed. Applications can decide
1525 whether to tolerate certain situations such as an absent nonce
1526 in a response when one was present in a request: the ocsp application
1527 just prints out a warning. New function OCSP_add1_basic_nonce()
1528 this is to allow responders to include a nonce in a response even if
1529 the request is nonce-less.
1532 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1535 [This change does not apply to 0.9.7.]
1537 *) Use the cached encoding of an X509_NAME structure rather than
1538 copying it. This is apparently the reason for the libsafe "errors"
1539 but the code is actually correct.
1542 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1543 skipped when using openssl x509 multiple times on a single input file,
1544 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1547 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1548 set string type: to handle setting ASN1_TIME structures. Fix ca
1549 utility to correctly initialize revocation date of CRLs.
1552 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1553 the clients preferred ciphersuites and rather use its own preferences.
1554 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1555 Internet Explorer by ensuring unchanged hash method during stepup.
1556 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1559 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1560 to aes and add a new 'exist' option to print out symbols that don't
1564 +) Additional options to ocsp utility to allow flags to be set and
1565 additional certificates supplied.
1568 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1569 OCSP client a number of certificate to only verify the response
1573 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1574 Bleichenbacher's DSA attack.
1575 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1576 to be set and top=0 forces the highest bit to be set; top=-1 is new
1577 and leaves the highest bit random.
1578 [Ulf Moeller, Bodo Moeller]
1580 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1581 handle the new API. Currently only ECB, CBC modes supported. Add new
1582 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1583 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1584 [Ben Laurie, Steve Henson]
1586 *) In the NCONF_...-based implementations for CONF_... queries
1587 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1588 a temporary CONF structure with the data component set to NULL
1589 (which gives segmentation faults in lh_retrieve).
1590 Instead, use NULL for the CONF pointer in CONF_get_string and
1591 CONF_get_number (which may use environment variables) and directly
1592 return NULL from CONF_get_section.
1595 *) Fix potential buffer overrun for EBCDIC.
1598 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1599 request to response.
1602 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1603 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1604 extract information from a certificate request. OCSP_response_create()
1605 creates a response and optionally adds a basic response structure.
1606 OCSP_basic_add1_status() adds a complete single response to a basic
1607 response and returns the OCSP_SINGLERESP structure just added (to allow
1608 extensions to be included for example). OCSP_basic_add1_cert() adds a
1609 certificate to a basic response and OCSP_basic_sign() signs a basic
1610 response with various flags. New helper functions ASN1_TIME_check()
1611 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1612 (converts ASN1_TIME to GeneralizedTime).
1615 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1616 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1617 structure from a certificate. X509_pubkey_digest() digests the public_key
1618 contents: this is used in various key identifiers.
1621 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1622 keyUsage if basicConstraints absent for a CA.
1625 *) Make SMIME_write_PKCS7() write mail header values with a format that
1626 is more generally accepted (no spaces before the semicolon), since
1627 some programs can't parse those values properly otherwise. Also make
1628 sure BIO's that break lines after each write do not create invalid
1632 +) Make sk_sort() tolerate a NULL argument.
1633 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1635 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1636 passed by the function are trusted implicitly. If any of them signed the
1637 response then it is assumed to be valid and is not verified.
1640 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1641 macros previously used would not encode an empty SEQUENCE OF
1642 and break the signature.
1644 [This change does not apply to 0.9.7.]
1646 *) Zero the premaster secret after deriving the master secret in
1650 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1651 to data. This was previously part of the PKCS7 ASN1 code. This
1652 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1653 [Steve Henson, reported by Kenneth R. Robinette
1654 <support@securenetterm.com>]
1656 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1657 routines: without these tracing memory leaks is very painful.
1658 Fix leaks in PKCS12 and PKCS7 routines.
1661 *) Add some EVP_add_digest_alias registrations (as found in
1662 OpenSSL_add_all_digests()) to SSL_library_init()
1663 aka OpenSSL_add_ssl_algorithms(). This provides improved
1664 compatibility with peers using X.509 certificates
1665 with unconventional AlgorithmIdentifier OIDs.
1668 *) Fix for Irix with NO_ASM.
1669 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1671 *) ./config script fixes.
1672 [Ulf Moeller, Richard Levitte]
1674 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1675 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1676 effectively meant GeneralizedTime would never be used. Now it
1677 is initialised to -1 but X509_time_adj() now has to check the value
1678 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1679 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1680 [Steve Henson, reported by Kenneth R. Robinette
1681 <support@securenetterm.com>]
1683 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1684 result in a zero length in the ASN1_INTEGER structure which was
1685 not consistent with the structure when d2i_ASN1_INTEGER() was used
1686 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1687 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1688 where it did not print out a minus for negative ASN1_INTEGER.
1691 *) Fix 'openssl passwd -1'.
1694 +) Add summary printout to ocsp utility. The various functions which
1695 convert status values to strings have been renamed to:
1696 OCSP_response_status_str(), OCSP_cert_status_str() and
1697 OCSP_crl_reason_str() and are no longer static. New options
1698 to verify nonce values and to disable verification. OCSP response
1699 printout format cleaned up.
1702 +) Add additional OCSP certificate checks. These are those specified
1703 in RFC2560. This consists of two separate checks: the CA of the
1704 certificate being checked must either be the OCSP signer certificate
1705 or the issuer of the OCSP signer certificate. In the latter case the
1706 OCSP signer certificate must contain the OCSP signing extended key
1707 usage. This check is performed by attempting to match the OCSP
1708 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1709 in the OCSP_CERTID structures of the response.
1712 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1713 and related routines. This uses the standard OpenSSL certificate
1714 verify routines to perform initial checks (just CA validity) and
1715 to obtain the certificate chain. Then additional checks will be
1716 performed on the chain. Currently the root CA is checked to see
1717 if it is explicitly trusted for OCSP signing. This is used to set
1718 a root CA as a global signing root: that is any certificate that
1719 chains to that CA is an acceptable OCSP signing certificate.
1722 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1723 extensions from a separate configuration file.
1724 As when reading extensions from the main configuration file,
1725 the '-extensions ...' option may be used for specifying the
1727 [Massimiliano Pala <madwolf@comune.modena.it>]
1729 *) Change PKCS12_key_gen_asc() so it can cope with non null
1730 terminated strings whose length is passed in the passlen
1731 parameter, for example from PEM callbacks. This was done
1732 by adding an extra length parameter to asc2uni().
1733 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1735 +) New OCSP utility. Allows OCSP requests to be generated or
1736 read. The request can be sent to a responder and the output
1737 parsed, outputed or printed in text form. Not complete yet:
1738 still needs to check the OCSP response validity.
1741 +) New subcommands for 'openssl ca':
1742 'openssl ca -status <serial>' prints the status of the cert with
1743 the given serial number (according to the index file).
1744 'openssl ca -updatedb' updates the expiry status of certificates
1746 [Massimiliano Pala <madwolf@comune.modena.it>]
1748 +) New '-newreq-nodes' command option to CA.pl. This is like
1749 '-newreq', but calls 'openssl req' with the '-nodes' option
1750 so that the resulting key is not encrypted.
1751 [Damien Miller <djm@mindrot.org>]
1753 +) New configuration for the GNU Hurd.
1754 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1756 +) Initial code to implement OCSP basic response verify. This
1757 is currently incomplete. Currently just finds the signer's
1758 certificate and verifies the signature on the response.
1761 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1762 value of OPENSSLDIR. This is available via the new '-d' option
1763 to 'openssl version', and is also included in 'openssl version -a'.
1766 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1767 call failed, free the DSA structure.
1770 +) Allowing defining memory allocation callbacks that will be given
1771 file name and line number information in additional arguments
1772 (a const char* and an int). The basic functionality remains, as
1773 well as the original possibility to just replace malloc(),
1774 realloc() and free() by functions that do not know about these
1775 additional arguments. To register and find out the current
1776 settings for extended allocation functions, the following
1777 functions are provided:
1779 CRYPTO_set_mem_ex_functions
1780 CRYPTO_set_locked_mem_ex_functions
1781 CRYPTO_get_mem_ex_functions
1782 CRYPTO_get_locked_mem_ex_functions
1784 These work the same way as CRYPTO_set_mem_functions and friends.
1785 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1786 extended allocation function is enabled.
1787 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1788 a conventional allocation function is enabled.
1789 [Richard Levitte, Bodo Moeller]
1791 *) Fix to uni2asc() to cope with zero length Unicode strings.
1792 These are present in some PKCS#12 files.
1795 +) Finish off removing the remaining LHASH function pointer casts.
1796 There should no longer be any prototype-casting required when using
1797 the LHASH abstraction, and any casts that remain are "bugs". See
1798 the callback types and macros at the head of lhash.h for details
1799 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1802 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1803 If /dev/[u]random devices are not available or do not return enough
1804 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1806 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1807 /etc/entropy will be queried once each in this sequence, quering stops
1808 when enough entropy was collected without querying more sockets.
1811 +) Change the Unix RAND_poll() variant to be able to poll several
1812 random devices, as specified by DEVRANDOM, until a sufficient amount
1813 of data has been collected. We spend at most 10 ms on each file
1814 (select timeout) and read in non-blocking mode. DEVRANDOM now
1815 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1816 (previously it was just the string "/dev/urandom"), so on typical
1817 platforms the 10 ms delay will never occur.
1818 Also separate out the Unix variant to its own file, rand_unix.c.
1819 For VMS, there's a currently-empty rand_vms.c.
1822 +) Move OCSP client related routines to ocsp_cl.c. These
1823 provide utility functions which an application needing
1824 to issue a request to an OCSP responder and analyse the
1825 response will typically need: as opposed to those which an
1826 OCSP responder itself would need which will be added later.
1828 OCSP_request_sign() signs an OCSP request with an API similar
1829 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1830 response. OCSP_response_get1_basic() extracts basic response
1831 from response. OCSP_resp_find_status(): finds and extracts status
1832 information from an OCSP_CERTID structure (which will be created
1833 when the request structure is built). These are built from lower
1834 level functions which work on OCSP_SINGLERESP structures but
1835 wont normally be used unless the application wishes to examine
1836 extensions in the OCSP response for example.
1838 Replace nonce routines with a pair of functions.
1839 OCSP_request_add1_nonce() adds a nonce value and optionally
1840 generates a random value. OCSP_check_nonce() checks the
1841 validity of the nonce in an OCSP response.
1844 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1845 This doesn't copy the supplied OCSP_CERTID and avoids the
1846 need to free up the newly created id. Change return type
1847 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1848 This can then be used to add extensions to the request.
1849 Deleted OCSP_request_new(), since most of its functionality
1850 is now in OCSP_REQUEST_new() (and the case insensitive name
1851 clash) apart from the ability to set the request name which
1852 will be added elsewhere.
1855 +) Update OCSP API. Remove obsolete extensions argument from
1856 various functions. Extensions are now handled using the new
1857 OCSP extension code. New simple OCSP HTTP function which
1858 can be used to send requests and parse the response.
1861 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1862 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1863 uses the special reorder version of SET OF to sort the attributes
1864 and reorder them to match the encoded order. This resolves a long
1865 standing problem: a verify on a PKCS7 structure just after signing
1866 it used to fail because the attribute order did not match the
1867 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1868 it uses the received order. This is necessary to tolerate some broken
1869 software that does not order SET OF. This is handled by encoding
1870 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1871 to produce the required SET OF.
1874 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1875 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1876 files to get correct declarations of the ASN.1 item variables.
1879 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1880 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1881 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1882 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1883 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1884 ASN1_ITEM and no wrapper functions.
1887 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1888 replace the old function pointer based I/O routines. Change most of
1889 the *_d2i_bio() and *_d2i_fp() functions to use these.
1892 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1893 lines, recognice more "algorithms" that can be deselected, and make
1894 it complain about algorithm deselection that isn't recognised.
1897 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1898 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1899 to use new functions. Add NO_ASN1_OLD which can be set to remove
1900 some old style ASN1 functions: this can be used to determine if old
1901 code will still work when these eventually go away.
1904 +) New extension functions for OCSP structures, these follow the
1905 same conventions as certificates and CRLs.
1908 +) New function X509V3_add1_i2d(). This automatically encodes and
1909 adds an extension. Its behaviour can be customised with various
1910 flags to append, replace or delete. Various wrappers added for
1911 certifcates and CRLs.
1914 +) Fix to avoid calling the underlying ASN1 print routine when
1915 an extension cannot be parsed. Correct a typo in the
1916 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1919 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1920 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1921 when writing a 32767 byte record.
1922 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1924 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1925 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1927 (RSA objects have a reference count access to which is protected
1928 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1929 so they are meant to be shared between threads.)
1930 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1931 "Reddie, Steven" <Steven.Reddie@ca.com>]
1933 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1934 entries for variables.
1937 *) Fix a deadlock in CRYPTO_mem_leaks().
1940 +) Add functionality to apps/openssl.c for detecting locking
1941 problems: As the program is single-threaded, all we have
1942 to do is register a locking callback using an array for
1943 storing which locks are currently held by the program.
1946 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1947 SSL_get_ex_data_X509_STORE_idx(), which is used in
1948 ssl_verify_cert_chain() and thus can be called at any time
1949 during TLS/SSL handshakes so that thread-safety is essential.
1950 Unfortunately, the ex_data design is not at all suited
1951 for multi-threaded use, so it probably should be abolished.
1954 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1955 [Broadcom, tweaked and integrated by Geoff Thorpe]
1957 +) Move common extension printing code to new function
1958 X509V3_print_extensions(). Reorganise OCSP print routines and
1959 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1962 +) New function X509_signature_print() to remove duplication in some
1966 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1967 set (this was treated exactly the same as SET OF previously). This
1968 is used to reorder the STACK representing the structure to match the
1969 encoding. This will be used to get round a problem where a PKCS7
1970 structure which was signed could not be verified because the STACK
1971 order did not reflect the encoded order.
1974 +) Reimplement the OCSP ASN1 module using the new code.
1977 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1978 for its ASN1 operations. The old style function pointers still exist
1979 for now but they will eventually go away.
1982 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1983 completely replaces the old ASN1 functionality with a table driven
1984 encoder and decoder which interprets an ASN1_ITEM structure describing
1985 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1986 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1987 has also been converted to the new form.
1990 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1991 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1992 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1993 for negative moduli.
1996 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1997 of not touching the result's sign bit.
2000 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
2004 +) Changed the LHASH code to use prototypes for callbacks, and created
2005 macros to declare and implement thin (optionally static) functions
2006 that provide type-safety and avoid function pointer casting for the
2007 type-specific callbacks.
2010 *) Use better test patterns in bntest.
2013 +) Added Kerberos Cipher Suites to be used with TLS, as written in
2015 [Veers Staats <staatsvr@asc.hpc.mil>,
2016 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
2018 *) rand_win.c fix for Borland C.
2021 *) BN_rshift bugfix for n == 0.
2024 +) Reformat the FAQ so the different questions and answers can be divided
2025 in sections depending on the subject.
2028 +) Have the zlib compression code load ZLIB.DLL dynamically under
2032 +) New function BN_mod_sqrt for computing square roots modulo a prime
2033 (using the probabilistic Tonelli-Shanks algorithm unless
2034 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
2035 be handled deterministically).
2036 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2038 *) Add a 'bctest' script that checks for some known 'bc' bugs
2039 so that 'make test' does not abort just because 'bc' is broken.
2042 *) Store verify_result within SSL_SESSION also for client side to
2043 avoid potential security hole. (Re-used sessions on the client side
2044 always resulted in verify_result==X509_V_OK, not using the original
2045 result of the server certificate verification.)
2048 +) Make BN_mod_inverse faster by explicitly handling small quotients
2049 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
2050 512 bits], about 30% for larger ones [1024 or 2048 bits].)
2053 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
2054 SSL3_RT_APPLICATION_DATA, return 0.
2055 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
2059 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
2060 releases, have been re-implemented by renaming the previous
2061 implementations of ssl2_read and ssl3_read to ssl2_read_internal
2062 and ssl3_read_internal, respectively, and adding 'peek' parameters
2063 to them. The new ssl[23]_{read,peek} functions are calls to
2064 ssl[23]_read_internal with the 'peek' flag set appropriately.
2065 A 'peek' parameter has also been added to ssl3_read_bytes, which
2066 does the actual work for ssl3_read_internal.
2069 +) New function BN_kronecker.
2072 +) Fix BN_gcd so that it works on negative inputs; the result is
2073 positive unless both parameters are zero.
2074 Previously something reasonably close to an infinite loop was
2075 possible because numbers could be growing instead of shrinking
2076 in the implementation of Euclid's algorithm.
2079 +) Fix BN_is_word() and BN_is_one() macros to take into account the
2080 sign of the number in question.
2082 Fix BN_is_word(a,w) to work correctly for w == 0.
2084 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
2085 because its test if the absolute value of 'a' equals 'w'.
2086 Note that BN_abs_is_word does *not* handle w == 0 reliably;
2087 it exists mostly for use in the implementations of BN_is_zero(),
2088 BN_is_one(), and BN_is_word().
2091 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
2092 the method-specific "init()" handler. Also clean up ex_data after
2093 calling the method-specific "finish()" handler. Previously, this was
2094 happening the other way round.
2097 +) New function BN_swap.
2100 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
2101 the exponentiation functions are more likely to produce reasonable
2102 results on negative inputs.
2105 +) Change BN_mod_mul so that the result is always non-negative.
2106 Previously, it could be negative if one of the factors was negative;
2107 I don't think anyone really wanted that behaviour.
2110 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
2111 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
2112 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
2113 and add new functions:
2122 BN_mod_lshift1_quick
2126 These functions always generate non-negative results.
2128 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
2129 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
2131 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2132 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2133 be reduced modulo m.
2134 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2136 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2137 was actually never needed) and in BN_mul(). The removal in BN_mul()
2138 required a small change in bn_mul_part_recursive() and the addition
2139 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2140 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2141 bn_sub_words() and bn_add_words() except they take arrays with
2145 +) In 'openssl passwd', verify passwords read from the terminal
2146 unless the '-salt' option is used (which usually means that
2147 verification would just waste user's time since the resulting
2148 hash is going to be compared with some given password hash)
2149 or the new '-noverify' option is used.
2151 This is an incompatible change, but it does not affect
2152 non-interactive use of 'openssl passwd' (passwords on the command
2153 line, '-stdin' option, '-in ...' option) and thus should not
2157 +) Remove all references to RSAref, since there's no more need for it.
2160 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2161 The previous value, 12, was not always sufficient for BN_mod_exp().
2164 +) Make DSO load along a path given through an environment variable
2165 (SHLIB_PATH) with shl_load().
2168 +) Constify the ENGINE code as a result of BIGNUM constification.
2169 Also constify the RSA code and most things related to it. In a
2170 few places, most notable in the depth of the ASN.1 code, ugly
2171 casts back to non-const were required (to be solved at a later
2175 +) Make it so the openssl application has all engines loaded by default.
2178 +) Constify the BIGNUM routines a little more.
2181 *) Make sure that shared libraries get the internal name engine with
2182 the full version number and not just 0. This should mark the
2183 shared libraries as not backward compatible. Of course, this should
2184 be changed again when we can guarantee backward binary compatibility.
2187 +) Add the following functions:
2189 ENGINE_load_cswift()
2191 ENGINE_load_atalla()
2193 ENGINE_load_builtin_engines()
2195 That way, an application can itself choose if external engines that
2196 are built-in in OpenSSL shall ever be used or not. The benefit is
2197 that applications won't have to be linked with libdl or other dso
2198 libraries unless it's really needed.
2200 Changed 'openssl engine' to load all engines on demand.
2201 Changed the engine header files to avoid the duplication of some
2202 declarations (they differed!).
2205 +) 'openssl engine' can now list capabilities.
2208 +) Better error reporting in 'openssl engine'.
2211 +) Never call load_dh_param(NULL) in s_server.
2214 +) Add engine application. It can currently list engines by name and
2215 identity, and test if they are actually available.
2218 +) Improve RPM specification file by forcing symbolic linking and making
2219 sure the installed documentation is also owned by root.root.
2220 [Damien Miller <djm@mindrot.org>]
2222 +) Give the OpenSSL applications more possibilities to make use of
2223 keys (public as well as private) handled by engines.
2226 +) Add OCSP code that comes from CertCo.
2229 +) Add VMS support for the Rijndael code.
2232 +) Added untested support for Nuron crypto accelerator.
2235 +) Add support for external cryptographic devices. This code was
2236 previously distributed separately as the "engine" branch.
2237 [Geoff Thorpe, Richard Levitte]
2239 +) Rework the filename-translation in the DSO code. It is now possible to
2240 have far greater control over how a "name" is turned into a filename
2241 depending on the operating environment and any oddities about the
2242 different shared library filenames on each system.
2245 +) Support threads on FreeBSD-elf in Configure.
2248 +) Fix for SHA1 assembly problem with MASM: it produces
2249 warnings about corrupt line number information when assembling
2250 with debugging information. This is caused by the overlapping
2252 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2255 NCONF_get_number() has no error checking at all. As a replacement,
2256 NCONF_get_number_e() is defined (_e for "error checking") and is
2257 promoted strongly. The old NCONF_get_number is kept around for
2258 binary backward compatibility.
2259 Make it possible for methods to load from something other than a BIO,
2260 by providing a function pointer that is given a name instead of a BIO.
2261 For example, this could be used to load configuration data from an
2265 *) Fix typo in get_cert_by_subject() in by_dir.c
2266 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2268 *) Rework the system to generate shared libraries:
2270 - Make note of the expected extension for the shared libraries and
2271 if there is a need for symbolic links from for example libcrypto.so.0
2272 to libcrypto.so.0.9.7. There is extended info in Configure for
2275 - Make as few rebuilds of the shared libraries as possible.
2277 - Still avoid linking the OpenSSL programs with the shared libraries.
2279 - When installing, install the shared libraries separately from the
2283 +) Fix for non blocking accept BIOs. Added new I/O special reason
2284 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2285 with non blocking I/O was not possible because no retry code was
2286 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2290 +) Added the beginnings of Rijndael support.
2293 +) Fix for bug in DirectoryString mask setting. Add support for
2294 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2295 to allow certificate printing to more controllable, additional
2296 'certopt' option to 'x509' to allow new printing options to be
2300 +) Clean old EAY MD5 hack from e_os.h.
2303 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2305 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2306 and not in SSL_clear because the latter is also used by the
2307 accept/connect functions; previously, the settings made by
2308 SSL_set_read_ahead would be lost during the handshake.
2309 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2311 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2312 Previously, it would create entries for disableed algorithms no
2316 *) Added several new manual pages for SSL_* function.
2319 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2321 *) In ssl23_get_client_hello, generate an error message when faced
2322 with an initial SSL 3.0/TLS record that is too small to contain the
2323 first two bytes of the ClientHello message, i.e. client_version.
2324 (Note that this is a pathologic case that probably has never happened
2325 in real life.) The previous approach was to use the version number
2326 from the record header as a substitute; but our protocol choice
2327 should not depend on that one because it is not authenticated
2328 by the Finished messages.
2331 *) More robust randomness gathering functions for Windows.
2332 [Jeffrey Altman <jaltman@columbia.edu>]
2334 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2335 not set then we don't setup the error code for issuer check errors
2336 to avoid possibly overwriting other errors which the callback does
2337 handle. If an application does set the flag then we assume it knows
2338 what it is doing and can handle the new informational codes
2342 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2343 a general "ANY" type, as such it should be able to decode anything
2344 including tagged types. However it didn't check the class so it would
2345 wrongly interpret tagged types in the same way as their universal
2346 counterpart and unknown types were just rejected. Changed so that the
2347 tagged and unknown types are handled in the same way as a SEQUENCE:
2348 that is the encoding is stored intact. There is also a new type
2349 "V_ASN1_OTHER" which is used when the class is not universal, in this
2350 case we have no idea what the actual type is so we just lump them all
2354 *) On VMS, stdout may very well lead to a file that is written to
2355 in a record-oriented fashion. That means that every write() will
2356 write a separate record, which will be read separately by the
2357 programs trying to read from it. This can be very confusing.
2359 The solution is to put a BIO filter in the way that will buffer
2360 text until a linefeed is reached, and then write everything a
2361 line at a time, so every record written will be an actual line,
2362 not chunks of lines and not (usually doesn't happen, but I've
2363 seen it once) several lines in one record. BIO_f_linebuffer() is
2366 Currently, it's a VMS-only method, because that's where it has
2367 been tested well enough.
2370 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2371 it can return incorrect results.
2372 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2373 but it was in 0.9.6-beta[12].)
2376 *) Disable the check for content being present when verifying detached
2377 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2378 include zero length content when signing messages.
2381 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2382 BIO_ctrl (for BIO pairs).
2385 *) Add DSO method for VMS.
2388 *) Bug fix: Montgomery multiplication could produce results with the
2392 *) Add RPM specification openssl.spec and modify it to build three
2393 packages. The default package contains applications, application
2394 documentation and run-time libraries. The devel package contains
2395 include files, static libraries and function documentation. The
2396 doc package contains the contents of the doc directory. The original
2397 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2400 *) Add a large number of documentation files for many SSL routines.
2401 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2403 *) Add a configuration entry for Sony News 4.
2404 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2406 *) Don't set the two most significant bits to one when generating a
2407 random number < q in the DSA library.
2410 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2411 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2412 the underlying transport is blocking) if a handshake took place.
2413 (The default behaviour is needed by applications such as s_client
2414 and s_server that use select() to determine when to use SSL_read;
2415 but for applications that know in advance when to expect data, it
2416 just makes things more complicated.)
2419 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2423 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2424 work better on such systems.
2425 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2427 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2428 Update PKCS12_parse() so it copies the friendlyName and the
2429 keyid to the certificates aux info.
2432 *) Fix bug in PKCS7_verify() which caused an infinite loop
2433 if there was more than one signature.
2434 [Sven Uszpelkat <su@celocom.de>]
2436 *) Major change in util/mkdef.pl to include extra information
2437 about each symbol, as well as presentig variables as well
2438 as functions. This change means that there's n more need
2439 to rebuild the .num files when some algorithms are excluded.
2442 *) Allow the verify time to be set by an application,
2443 rather than always using the current time.
2446 *) Phase 2 verify code reorganisation. The certificate
2447 verify code now looks up an issuer certificate by a
2448 number of criteria: subject name, authority key id
2449 and key usage. It also verifies self signed certificates
2450 by the same criteria. The main comparison function is
2451 X509_check_issued() which performs these checks.
2453 Lot of changes were necessary in order to support this
2454 without completely rewriting the lookup code.
2456 Authority and subject key identifier are now cached.
2458 The LHASH 'certs' is X509_STORE has now been replaced
2459 by a STACK_OF(X509_OBJECT). This is mainly because an
2460 LHASH can't store or retrieve multiple objects with
2461 the same hash value.
2463 As a result various functions (which were all internal
2464 use only) have changed to handle the new X509_STORE
2465 structure. This will break anything that messed round
2466 with X509_STORE internally.
2468 The functions X509_STORE_add_cert() now checks for an
2469 exact match, rather than just subject name.
2471 The X509_STORE API doesn't directly support the retrieval
2472 of multiple certificates matching a given criteria, however
2473 this can be worked round by performing a lookup first
2474 (which will fill the cache with candidate certificates)
2475 and then examining the cache for matches. This is probably
2476 the best we can do without throwing out X509_LOOKUP
2477 entirely (maybe later...).
2479 The X509_VERIFY_CTX structure has been enhanced considerably.
2481 All certificate lookup operations now go via a get_issuer()
2482 callback. Although this currently uses an X509_STORE it
2483 can be replaced by custom lookups. This is a simple way
2484 to bypass the X509_STORE hackery necessary to make this
2485 work and makes it possible to use more efficient techniques
2486 in future. A very simple version which uses a simple
2487 STACK for its trusted certificate store is also provided
2488 using X509_STORE_CTX_trusted_stack().
2490 The verify_cb() and verify() callbacks now have equivalents
2491 in the X509_STORE_CTX structure.
2493 X509_STORE_CTX also has a 'flags' field which can be used
2494 to customise the verify behaviour.
2497 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2498 excludes S/MIME capabilities.
2501 *) When a certificate request is read in keep a copy of the
2502 original encoding of the signed data and use it when outputing
2503 again. Signatures then use the original encoding rather than
2504 a decoded, encoded version which may cause problems if the
2505 request is improperly encoded.
2508 *) For consistency with other BIO_puts implementations, call
2509 buffer_write(b, ...) directly in buffer_puts instead of calling
2512 In BIO_puts, increment b->num_write as in BIO_write.
2513 [Peter.Sylvester@EdelWeb.fr]
2515 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2516 BN_zero, we may not return a BIGNUM with an array consisting of
2520 *) Avoid calling abort() from within the library when problems are
2521 detected, except if preprocessor symbols have been defined
2522 (such as REF_CHECK, BN_DEBUG etc.).
2525 *) New openssl application 'rsautl'. This utility can be
2526 used for low level RSA operations. DER public key
2527 BIO/fp routines also added.
2530 *) New Configure entry and patches for compiling on QNX 4.
2531 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2533 *) A demo state-machine implementation was sponsored by
2534 Nuron (http://www.nuron.com/) and is now available in
2535 demos/state_machine.
2538 *) New options added to the 'dgst' utility for signature
2539 generation and verification.
2542 *) Unrecognized PKCS#7 content types are now handled via a
2543 catch all ASN1_TYPE structure. This allows unsupported
2544 types to be stored as a "blob" and an application can
2545 encode and decode it manually.
2548 *) Fix various signed/unsigned issues to make a_strex.c
2550 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2552 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2553 length if passed a buffer. ASN1_INTEGER_to_BN failed
2554 if passed a NULL BN and its argument was negative.
2555 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2557 *) Modification to PKCS#7 encoding routines to output definite
2558 length encoding. Since currently the whole structures are in
2559 memory there's not real point in using indefinite length
2560 constructed encoding. However if OpenSSL is compiled with
2561 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2564 *) Added BIO_vprintf() and BIO_vsnprintf().
2567 *) Added more prefixes to parse for in the the strings written
2568 through a logging bio, to cover all the levels that are available
2569 through syslog. The prefixes are now:
2571 PANIC, EMERG, EMR => LOG_EMERG
2572 ALERT, ALR => LOG_ALERT
2573 CRIT, CRI => LOG_CRIT
2574 ERROR, ERR => LOG_ERR
2575 WARNING, WARN, WAR => LOG_WARNING
2576 NOTICE, NOTE, NOT => LOG_NOTICE
2577 INFO, INF => LOG_INFO
2578 DEBUG, DBG => LOG_DEBUG
2580 and as before, if none of those prefixes are present at the
2581 beginning of the string, LOG_ERR is chosen.
2583 On Win32, the LOG_* levels are mapped according to this:
2585 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2586 LOG_WARNING => EVENTLOG_WARNING_TYPE
2587 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2591 *) Made it possible to reconfigure with just the configuration
2592 argument "reconf" or "reconfigure". The command line arguments
2593 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2594 and are retrieved from there when reconfiguring.
2598 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2600 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2603 *) The obj_dat.pl script was messing up the sorting of object
2604 names. The reason was that it compared the quoted version
2605 of strings as a result "OCSP" > "OCSP Signing" because
2606 " > SPACE. Changed script to store unquoted versions of
2607 names and add quotes on output. It was also omitting some
2608 names from the lookup table if they were given a default
2609 value (that is if SN is missing it is given the same
2610 value as LN and vice versa), these are now added on the
2611 grounds that if an object has a name we should be able to
2612 look it up. Finally added warning output when duplicate
2613 short or long names are found.
2616 *) Changes needed for Tandem NSK.
2617 [Scott Uroff <scott@xypro.com>]
2619 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2620 RSA_padding_check_SSLv23(), special padding was never detected
2621 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2622 version rollback attacks was not effective.
2624 In s23_clnt.c, don't use special rollback-attack detection padding
2625 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2626 client; similarly, in s23_srvr.c, don't do the rollback check if
2627 SSL 2.0 is the only protocol enabled in the server.
2630 *) Make it possible to get hexdumps of unprintable data with 'openssl
2631 asn1parse'. By implication, the functions ASN1_parse_dump() and
2632 BIO_dump_indent() are added.
2635 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2636 these print out strings and name structures based on various
2637 flags including RFC2253 support and proper handling of
2638 multibyte characters. Added options to the 'x509' utility
2639 to allow the various flags to be set.
2642 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2643 Also change the functions X509_cmp_current_time() and
2644 X509_gmtime_adj() work with an ASN1_TIME structure,
2645 this will enable certificates using GeneralizedTime in validity
2646 dates to be checked.
2649 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2650 negative public key encodings) on by default,
2651 NO_NEG_PUBKEY_BUG can be set to disable it.
2654 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2655 content octets. An i2c_ASN1_OBJECT is unnecessary because
2656 the encoding can be trivially obtained from the structure.
2659 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2660 not read locks (CRYPTO_r_[un]lock).
2663 *) A first attempt at creating official support for shared
2664 libraries through configuration. I've kept it so the
2665 default is static libraries only, and the OpenSSL programs
2666 are always statically linked for now, but there are
2667 preparations for dynamic linking in place.
2668 This has been tested on Linux and Tru64.
2671 *) Randomness polling function for Win9x, as described in:
2672 Peter Gutmann, Software Generation of Practically Strong
2676 *) Fix so PRNG is seeded in req if using an already existing
2680 *) New options to smime application. -inform and -outform
2681 allow alternative formats for the S/MIME message including
2682 PEM and DER. The -content option allows the content to be
2683 specified separately. This should allow things like Netscape
2684 form signing output easier to verify.
2687 *) Fix the ASN1 encoding of tags using the 'long form'.
2690 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2691 STRING types. These convert content octets to and from the
2692 underlying type. The actual tag and length octets are
2693 already assumed to have been read in and checked. These
2694 are needed because all other string types have virtually
2695 identical handling apart from the tag. By having versions
2696 of the ASN1 functions that just operate on content octets
2697 IMPLICIT tagging can be handled properly. It also allows
2698 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2699 and ASN1_INTEGER are identical apart from the tag.
2702 *) Change the handling of OID objects as follows:
2704 - New object identifiers are inserted in objects.txt, following
2705 the syntax given in objects.README.
2706 - objects.pl is used to process obj_mac.num and create a new
2708 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2711 This is currently kind of a hack, and the perl code in objects.pl
2712 isn't very elegant, but it works as I intended. The simplest way
2713 to check that it worked correctly is to look in obj_dat.h and
2714 check the array nid_objs and make sure the objects haven't moved
2715 around (this is important!). Additions are OK, as well as
2716 consistent name changes.
2719 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2722 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2723 The given file adds to whatever has already been seeded into the
2724 random pool through the RANDFILE configuration file option or
2725 environment variable, or the default random state file.
2728 *) mkstack.pl now sorts each macro group into lexical order.
2729 Previously the output order depended on the order the files
2730 appeared in the directory, resulting in needless rewriting
2734 *) Patches to make OpenSSL compile under Win32 again. Mostly
2735 work arounds for the VC++ problem that it treats func() as
2736 func(void). Also stripped out the parts of mkdef.pl that
2737 added extra typesafe functions: these no longer exist.
2740 *) Reorganisation of the stack code. The macros are now all
2741 collected in safestack.h . Each macro is defined in terms of
2742 a "stack macro" of the form SKM_<name>(type, a, b). The
2743 DEBUG_SAFESTACK is now handled in terms of function casts,
2744 this has the advantage of retaining type safety without the
2745 use of additional functions. If DEBUG_SAFESTACK is not defined
2746 then the non typesafe macros are used instead. Also modified the
2747 mkstack.pl script to handle the new form. Needs testing to see
2748 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2749 the default if no major problems. Similar behaviour for ASN1_SET_OF
2750 and PKCS12_STACK_OF.
2753 *) When some versions of IIS use the 'NET' form of private key the
2754 key derivation algorithm is different. Normally MD5(password) is
2755 used as a 128 bit RC4 key. In the modified case
2756 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2757 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2758 as the old Netscape_RSA functions except they have an additional
2759 'sgckey' parameter which uses the modified algorithm. Also added
2760 an -sgckey command line option to the rsa utility. Thanks to
2761 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2762 algorithm to openssl-dev.
2765 *) The evp_local.h macros were using 'c.##kname' which resulted in
2766 invalid expansion on some systems (SCO 5.0.5 for example).
2767 Corrected to 'c.kname'.
2768 [Phillip Porch <root@theporch.com>]
2770 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2771 a STACK of email addresses from a certificate or request, these look
2772 in the subject name and the subject alternative name extensions and
2773 omit any duplicate addresses.
2776 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2777 This makes DSA verification about 2 % faster.
2780 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2781 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2782 plus overhead for 1024 bit moduli).
2783 This makes exponentiations about 0.5 % faster for 1024 bit
2784 exponents (as measured by "openssl speed rsa2048").
2787 *) Rename memory handling macros to avoid conflicts with other
2789 Malloc => OPENSSL_malloc
2790 Malloc_locked => OPENSSL_malloc_locked
2791 Realloc => OPENSSL_realloc
2792 Free => OPENSSL_free
2795 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2796 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2799 *) CygWin32 support.
2800 [John Jarvie <jjarvie@newsguy.com>]
2802 *) The type-safe stack code has been rejigged. It is now only compiled
2803 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2804 by default all type-specific stack functions are "#define"d back to
2805 standard stack functions. This results in more streamlined output
2806 but retains the type-safety checking possibilities of the original
2810 *) The STACK code has been cleaned up, and certain type declarations
2811 that didn't make a lot of sense have been brought in line. This has
2812 also involved a cleanup of sorts in safestack.h to more correctly
2813 map type-safe stack functions onto their plain stack counterparts.
2814 This work has also resulted in a variety of "const"ifications of
2815 lots of the code, especially "_cmp" operations which should normally
2816 be prototyped with "const" parameters anyway.
2819 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2820 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2821 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2822 where all of 'md' is used each time the PRNG is used, but 'state'
2823 is used only indexed by a cyclic counter. As entropy may not be
2824 well distributed from the beginning, 'md' is important as a
2825 chaining variable. However, the output function chains only half
2826 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2827 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2828 in all of 'state' being rewritten, with the new values depending
2829 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2832 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2833 the handshake is continued after ssl_verify_cert_chain();
2834 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2835 can lead to 'unexplainable' connection aborts later.
2836 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2838 *) Major EVP API cipher revision.
2839 Add hooks for extra EVP features. This allows various cipher
2840 parameters to be set in the EVP interface. Support added for variable
2841 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2842 setting of RC2 and RC5 parameters.
2844 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2847 Remove lots of duplicated code from the EVP library. For example *every*
2848 cipher init() function handles the 'iv' in the same way according to the
2849 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2850 for CFB and OFB modes they zero ctx->num.
2852 New functionality allows removal of S/MIME code RC2 hack.
2854 Most of the routines have the same form and so can be declared in terms
2857 By shifting this to the top level EVP_CipherInit() it can be removed from
2858 all individual ciphers. If the cipher wants to handle IVs or keys
2859 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2862 Change lots of functions like EVP_EncryptUpdate() to now return a
2863 value: although software versions of the algorithms cannot fail
2864 any installed hardware versions can.
2867 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2868 this option is set, tolerate broken clients that send the negotiated
2869 protocol version number instead of the requested protocol version
2873 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2874 i.e. non-zero for export ciphersuites, zero otherwise.
2875 Previous versions had this flag inverted, inconsistent with
2876 rsa_tmp_cb (..._TMP_RSA_CB).
2877 [Bodo Moeller; problem reported by Amit Chopra]
2879 *) Add missing DSA library text string. Work around for some IIS
2880 key files with invalid SEQUENCE encoding.
2883 *) Add a document (doc/standards.txt) that list all kinds of standards
2884 and so on that are implemented in OpenSSL.
2887 *) Enhance c_rehash script. Old version would mishandle certificates
2888 with the same subject name hash and wouldn't handle CRLs at all.
2889 Added -fingerprint option to crl utility, to support new c_rehash
2893 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2896 *) Fix for SSL server purpose checking. Server checking was
2897 rejecting certificates which had extended key usage present
2898 but no ssl client purpose.
2899 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2901 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2902 is a little unclear about how a blank password is handled.
2903 Since the password in encoded as a BMPString with terminating
2904 double NULL a zero length password would end up as just the
2905 double NULL. However no password at all is different and is
2906 handled differently in the PKCS#12 key generation code. NS
2907 treats a blank password as zero length. MSIE treats it as no
2908 password on export: but it will try both on import. We now do
2909 the same: PKCS12_parse() tries zero length and no password if
2910 the password is set to "" or NULL (NULL is now a valid password:
2911 it wasn't before) as does the pkcs12 application.
2914 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2915 perror when PEM_read_bio_X509_REQ fails, the error message must
2916 be obtained from the error queue.
2919 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2920 it in ERR_remove_state if appropriate, and change ERR_get_state
2921 accordingly to avoid race conditions (this is necessary because
2922 thread_hash is no longer constant once set).
2925 *) Bugfix for linux-elf makefile.one.
2928 *) RSA_get_default_method() will now cause a default
2929 RSA_METHOD to be chosen if one doesn't exist already.
2930 Previously this was only set during a call to RSA_new()
2931 or RSA_new_method(NULL) meaning it was possible for
2932 RSA_get_default_method() to return NULL.
2935 *) Added native name translation to the existing DSO code
2936 that will convert (if the flag to do so is set) filenames
2937 that are sufficiently small and have no path information
2938 into a canonical native form. Eg. "blah" converted to
2939 "libblah.so" or "blah.dll" etc.
2942 *) New function ERR_error_string_n(e, buf, len) which is like
2943 ERR_error_string(e, buf), but writes at most 'len' bytes
2944 including the 0 terminator. For ERR_error_string_n, 'buf'
2946 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2948 *) CONF library reworked to become more general. A new CONF
2949 configuration file reader "class" is implemented as well as a
2950 new functions (NCONF_*, for "New CONF") to handle it. The now
2951 old CONF_* functions are still there, but are reimplemented to
2952 work in terms of the new functions. Also, a set of functions
2953 to handle the internal storage of the configuration data is
2954 provided to make it easier to write new configuration file
2955 reader "classes" (I can definitely see something reading a
2956 configuration file in XML format, for example), called _CONF_*,
2957 or "the configuration storage API"...
2959 The new configuration file reading functions are:
2961 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2962 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2964 NCONF_default, NCONF_WIN32
2966 NCONF_dump_fp, NCONF_dump_bio
2968 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2969 NCONF_new creates a new CONF object. This works in the same way
2970 as other interfaces in OpenSSL, like the BIO interface.
2971 NCONF_dump_* dump the internal storage of the configuration file,
2972 which is useful for debugging. All other functions take the same
2973 arguments as the old CONF_* functions wth the exception of the
2974 first that must be a `CONF *' instead of a `LHASH *'.
2976 To make it easer to use the new classes with the old CONF_* functions,
2977 the function CONF_set_default_method is provided.
2980 *) Add '-tls1' option to 'openssl ciphers', which was already
2981 mentioned in the documentation but had not been implemented.
2982 (This option is not yet really useful because even the additional
2983 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2986 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2987 OpenSSL-based applications) load shared libraries and bind to
2988 them in a portable way.
2989 [Geoff Thorpe, with contributions from Richard Levitte]
2991 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2993 *) Make sure _lrotl and _lrotr are only used with MSVC.
2995 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2996 (the default implementation of RAND_status).
2998 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2999 to '-clrext' (= clear extensions), as intended and documented.
3000 [Bodo Moeller; inconsistency pointed out by Michael Attili
3003 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
3004 was larger than the MD block size.
3005 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
3007 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
3008 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
3009 using the passed key: if the passed key was a private key the result
3010 of X509_print(), for example, would be to print out all the private key
3014 *) des_quad_cksum() byte order bug fix.
3015 [Ulf Möller, using the problem description in krb4-0.9.7, where
3016 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
3018 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
3020 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
3022 *) For easily testing in shell scripts whether some command
3023 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
3024 returns with exit code 0 iff no command of the given name is available.
3025 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
3026 the output goes to stdout and nothing is printed to stderr.
3027 Additional arguments are always ignored.
3029 Since for each cipher there is a command of the same name,
3030 the 'no-cipher' compilation switches can be tested this way.
3032 ('openssl no-XXX' is not able to detect pseudo-commands such
3033 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
3036 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
3039 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
3040 is set; it will be thrown away anyway because each handshake creates
3042 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
3043 to parameters -- in previous versions (since OpenSSL 0.9.3) the
3044 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
3045 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
3048 *) New s_client option -ign_eof: EOF at stdin is ignored, and
3049 'Q' and 'R' lose their special meanings (quit/renegotiate).
3050 This is part of what -quiet does; unlike -quiet, -ign_eof
3051 does not suppress any output.
3054 *) Add compatibility options to the purpose and trust code. The
3055 purpose X509_PURPOSE_ANY is "any purpose" which automatically
3056 accepts a certificate or CA, this was the previous behaviour,
3057 with all the associated security issues.
3059 X509_TRUST_COMPAT is the old trust behaviour: only and
3060 automatically trust self signed roots in certificate store. A
3061 new trust setting X509_TRUST_DEFAULT is used to specify that
3062 a purpose has no associated trust setting and it should instead
3063 use the value in the default purpose.
3066 *) Fix the PKCS#8 DSA private key code so it decodes keys again
3067 and fix a memory leak.
3070 *) In util/mkerr.pl (which implements 'make errors'), preserve
3071 reason strings from the previous version of the .c file, as
3072 the default to have only downcase letters (and digits) in
3073 automatically generated reasons codes is not always appropriate.
3076 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
3077 using strerror. Previously, ERR_reason_error_string() returned
3078 library names as reason strings for SYSerr; but SYSerr is a special
3079 case where small numbers are errno values, not library numbers.
3082 *) Add '-dsaparam' option to 'openssl dhparam' application. This
3083 converts DSA parameters into DH parameters. (When creating parameters,
3084 DSA_generate_parameters is used.)
3087 *) Include 'length' (recommended exponent length) in C code generated
3088 by 'openssl dhparam -C'.
3091 *) The second argument to set_label in perlasm was already being used
3092 so couldn't be used as a "file scope" flag. Moved to third argument
3096 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
3097 instead of RAND_bytes for encryption IVs and salts.
3100 *) Include RAND_status() into RAND_METHOD instead of implementing
3101 it only for md_rand.c Otherwise replacing the PRNG by calling
3102 RAND_set_rand_method would be impossible.
3105 *) Don't let DSA_generate_key() enter an infinite loop if the random
3106 number generation fails.
3109 *) New 'rand' application for creating pseudo-random output.
3112 *) Added configuration support for Linux/IA64
3113 [Rolf Haberrecker <rolf@suse.de>]
3115 *) Assembler module support for Mingw32.
3118 *) Shared library support for HPUX (in shlib/).
3119 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
3121 *) Shared library support for Solaris gcc.
3122 [Lutz Behnke <behnke@trustcenter.de>]
3124 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
3126 *) PKCS7_encrypt() was adding text MIME headers twice because they
3127 were added manually and by SMIME_crlf_copy().
3130 *) In bntest.c don't call BN_rand with zero bits argument.
3131 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3133 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3134 case was implemented. This caused BN_div_recp() to fail occasionally.
3137 *) Add an optional second argument to the set_label() in the perl
3138 assembly language builder. If this argument exists and is set
3139 to 1 it signals that the assembler should use a symbol whose
3140 scope is the entire file, not just the current function. This
3141 is needed with MASM which uses the format label:: for this scope.
3142 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3144 *) Change the ASN1 types so they are typedefs by default. Before
3145 almost all types were #define'd to ASN1_STRING which was causing
3146 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3150 *) Change names of new functions to the new get1/get0 naming
3151 convention: After 'get1', the caller owns a reference count
3152 and has to call ..._free; 'get0' returns a pointer to some
3153 data structure without incrementing reference counters.
3154 (Some of the existing 'get' functions increment a reference
3155 counter, some don't.)
3156 Similarly, 'set1' and 'add1' functions increase reference
3157 counters or duplicate objects.
3160 *) Allow for the possibility of temp RSA key generation failure:
3161 the code used to assume it always worked and crashed on failure.
3164 *) Fix potential buffer overrun problem in BIO_printf().
3165 [Ulf Möller, using public domain code by Patrick Powell; problem
3166 pointed out by David Sacerdote <das33@cornell.edu>]
3168 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3169 RAND_egd() and RAND_status(). In the command line application,
3170 the EGD socket can be specified like a seed file using RANDFILE
3174 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3175 Some CAs (e.g. Verisign) distribute certificates in this form.
3178 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3179 list to exclude them. This means that no special compilation option
3180 is needed to use anonymous DH: it just needs to be included in the
3184 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3185 EVP_MD_type. The old functionality is available in a new macro called
3186 EVP_MD_md(). Change code that uses it and update docs.
3189 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3190 where the 'void *' argument is replaced by a function pointer argument.
3191 Previously 'void *' was abused to point to functions, which works on
3192 many platforms, but is not correct. As these functions are usually
3193 called by macros defined in OpenSSL header files, most source code
3194 should work without changes.
3197 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3198 sections with information on -D... compiler switches used for
3199 compiling the library so that applications can see them. To enable
3200 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3201 must be defined. E.g.,
3202 #define OPENSSL_ALGORITHM_DEFINES
3203 #include <openssl/opensslconf.h>
3204 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3205 [Richard Levitte, Ulf and Bodo Möller]
3207 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3211 *) Change the 'other' type in certificate aux info to a STACK_OF
3212 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3213 the required ASN1 format: arbitrary types determined by an OID.
3216 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3217 argument to 'req'. This is not because the function is newer or
3218 better than others it just uses the work 'NEW' in the certificate
3219 request header lines. Some software needs this.
3222 *) Reorganise password command line arguments: now passwords can be
3223 obtained from various sources. Delete the PEM_cb function and make
3224 it the default behaviour: i.e. if the callback is NULL and the
3225 usrdata argument is not NULL interpret it as a null terminated pass
3226 phrase. If usrdata and the callback are NULL then the pass phrase
3227 is prompted for as usual.
3230 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3231 the support is automatically enabled. The resulting binaries will
3232 autodetect the card and use it if present.
3233 [Ben Laurie and Compaq Inc.]
3235 *) Work around for Netscape hang bug. This sends certificate request
3236 and server done in one record. Since this is perfectly legal in the
3237 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3238 the bugs/SSLv3 entry for more info.
3241 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3244 *) Add -rand argument to smime and pkcs12 applications and read/write
3248 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3251 *) Add command line password options to the remaining applications.
3254 *) Bug fix for BN_div_recp() for numerators with an even number of
3258 *) More tests in bntest.c, and changed test_bn output.
3261 *) ./config recognizes MacOS X now.
3264 *) Bug fix for BN_div() when the first words of num and divsor are
3265 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3268 *) Add support for various broken PKCS#8 formats, and command line
3269 options to produce them.
3272 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3273 get temporary BIGNUMs from a BN_CTX.
3276 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3280 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3281 include a #define from the old name to the new. The original intent
3282 was that statically linked binaries could for example just call
3283 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3284 link with digests. This never worked becayse SSLeay_add_all_digests()
3285 and SSLeay_add_all_ciphers() were in the same source file so calling
3286 one would link with the other. They are now in separate source files.
3289 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3292 *) Use a less unusual form of the Miller-Rabin primality test (it used
3293 a binary algorithm for exponentiation integrated into the Miller-Rabin
3294 loop, our standard modexp algorithms are faster).
3297 *) Support for the EBCDIC character set completed.
3298 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3300 *) Source code cleanups: use const where appropriate, eliminate casts,
3301 use void * instead of char * in lhash.
3304 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3305 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3306 this the server could overwrite ephemeral keys that the client
3310 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3311 using 50 iterations of the Rabin-Miller test.
3313 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3314 iterations of the Rabin-Miller test as required by the appendix
3315 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3316 As BN_is_prime_fasttest includes trial division, DSA parameter
3317 generation becomes much faster.
3319 This implies a change for the callback functions in DSA_is_prime
3320 and DSA_generate_parameters: The callback function is called once
3321 for each positive witness in the Rabin-Miller test, not just
3322 occasionally in the inner loop; and the parameters to the
3323 callback function now provide an iteration count for the outer
3324 loop rather than for the current invocation of the inner loop.
3325 DSA_generate_parameters additionally can call the callback
3326 function with an 'iteration count' of -1, meaning that a
3327 candidate has passed the trial division test (when q is generated
3328 from an application-provided seed, trial division is skipped).
3331 *) New function BN_is_prime_fasttest that optionally does trial
3332 division before starting the Rabin-Miller test and has
3333 an additional BN_CTX * argument (whereas BN_is_prime always
3334 has to allocate at least one BN_CTX).
3335 'callback(1, -1, cb_arg)' is called when a number has passed the
3336 trial division stage.
3339 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3343 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3346 *) New function BN_pseudo_rand().
3349 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3350 bignum version of BN_from_montgomery() with the working code from
3351 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3355 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3356 made it impossible to use the same SSL_SESSION data structure in
3357 SSL2 clients in multiple threads.
3360 *) The return value of RAND_load_file() no longer counts bytes obtained
3361 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3362 to seed the PRNG (previously an explicit byte count was required).
3363 [Ulf Möller, Bodo Möller]
3365 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3366 used (char *) instead of (void *) and had casts all over the place.
3369 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3372 *) Retain source code compatibility for BN_prime_checks macro:
3373 BN_is_prime(..., BN_prime_checks, ...) now uses
3374 BN_prime_checks_for_size to determine the appropriate number of
3375 Rabin-Miller iterations.
3378 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3379 DH_CHECK_P_NOT_SAFE_PRIME.
3380 (Check if this is true? OpenPGP calls them "strong".)
3383 *) Merge the functionality of "dh" and "gendh" programs into a new program
3384 "dhparam". The old programs are retained for now but will handle DH keys
3385 (instead of parameters) in future.
3388 *) Make the ciphers, s_server and s_client programs check the return values
3389 when a new cipher list is set.
3392 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3393 ciphers. Before when the 56bit ciphers were enabled the sorting was
3396 The syntax for the cipher sorting has been extended to support sorting by
3397 cipher-strength (using the strength_bits hard coded in the tables).
3398 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3400 Fix a bug in the cipher-command parser: when supplying a cipher command
3401 string with an "undefined" symbol (neither command nor alphanumeric
3402 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3403 an error is flagged.
3405 Due to the strength-sorting extension, the code of the
3406 ssl_create_cipher_list() function was completely rearranged. I hope that
3407 the readability was also increased :-)
3408 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3410 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3411 for the first serial number and places 2 in the serial number file. This
3412 avoids problems when the root CA is created with serial number zero and
3413 the first user certificate has the same issuer name and serial number
3417 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3418 the new code. Add documentation for this stuff.
3421 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3422 X509_*() to X509at_*() on the grounds that they don't handle X509
3423 structures and behave in an analagous way to the X509v3 functions:
3424 they shouldn't be called directly but wrapper functions should be used
3427 So we also now have some wrapper functions that call the X509at functions
3428 when passed certificate requests. (TO DO: similar things can be done with
3429 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3430 things. Some of these need some d2i or i2d and print functionality
3431 because they handle more complex structures.)
3434 *) Add missing #ifndefs that caused missing symbols when building libssl
3435 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3436 NO_RSA in ssl/s2*.c.
3437 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3439 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3440 has a return value which indicates the quality of the random data
3441 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3442 error queue. New function RAND_pseudo_bytes() generates output that is
3443 guaranteed to be unique but not unpredictable. RAND_add is like
3444 RAND_seed, but takes an extra argument for an entropy estimate
3445 (RAND_seed always assumes full entropy).
3448 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3449 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3450 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3451 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3452 false-positive rate of at most 2^-80 for random input.
3455 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3458 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3459 in the 0.9.5 release), this returns the chain
3460 from an X509_CTX structure with a dup of the stack and all
3461 the X509 reference counts upped: so the stack will exist
3462 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3465 Also make SSL_SESSION_print() print out the verify return
3469 *) Add manpage for the pkcs12 command. Also change the default
3470 behaviour so MAC iteration counts are used unless the new
3471 -nomaciter option is used. This improves file security and
3472 only older versions of MSIE (4.0 for example) need it.
3475 *) Honor the no-xxx Configure options when creating .DEF files.
3478 *) Add PKCS#10 attributes to field table: challengePassword,
3479 unstructuredName and unstructuredAddress. These are taken from
3480 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3481 international characters are used.
3483 More changes to X509_ATTRIBUTE code: allow the setting of types
3484 based on strings. Remove the 'loc' parameter when adding
3485 attributes because these will be a SET OF encoding which is sorted
3489 *) Initial changes to the 'req' utility to allow request generation
3490 automation. This will allow an application to just generate a template
3491 file containing all the field values and have req construct the
3494 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3495 used all over the place including certificate requests and PKCS#7
3496 structures. They are currently handled manually where necessary with
3497 some primitive wrappers for PKCS#7. The new functions behave in a
3498 manner analogous to the X509 extension functions: they allow
3499 attributes to be looked up by NID and added.
3501 Later something similar to the X509V3 code would be desirable to
3502 automatically handle the encoding, decoding and printing of the
3503 more complex types. The string types like challengePassword can
3504 be handled by the string table functions.
3506 Also modified the multi byte string table handling. Now there is
3507 a 'global mask' which masks out certain types. The table itself
3508 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3509 is useful when for example there is only one permissible type
3510 (as in countryName) and using the mask might result in no valid
3514 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3515 SSL_get_peer_finished to allow applications to obtain the latest
3516 Finished messages sent to the peer or expected from the peer,
3517 respectively. (SSL_get_peer_finished is usually the Finished message
3518 actually received from the peer, otherwise the protocol will be aborted.)
3520 As the Finished message are message digests of the complete handshake
3521 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3522 be used for external authentication procedures when the authentication
3523 provided by SSL/TLS is not desired or is not enough.
3526 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3527 the host supports BWX extension and if Compaq C is present on the
3528 $PATH. Just exploiting of the BWX extension results in 20-30%
3529 performance kick for some algorithms, e.g. DES and RC4 to mention
3530 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3534 *) Add support for MS "fast SGC". This is arguably a violation of the
3535 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3536 weak crypto and after checking the certificate is SGC a second one
3537 with strong crypto. MS SGC stops the first handshake after receiving
3538 the server certificate message and sends a second client hello. Since
3539 a server will typically do all the time consuming operations before
3540 expecting any further messages from the client (server key exchange
3541 is the most expensive) there is little difference between the two.
3543 To get OpenSSL to support MS SGC we have to permit a second client
3544 hello message after we have sent server done. In addition we have to
3545 reset the MAC if we do get this second client hello.
3548 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3549 if a DER encoded private key is RSA or DSA traditional format. Changed
3550 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3551 format DER encoded private key. Newer code should use PKCS#8 format which
3552 has the key type encoded in the ASN1 structure. Added DER private key
3553 support to pkcs8 application.
3556 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3557 ciphersuites has been selected (as required by the SSL 3/TLS 1
3558 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3559 is set, we interpret this as a request to violate the specification
3560 (the worst that can happen is a handshake failure, and 'correct'
3561 behaviour would result in a handshake failure anyway).
3564 *) In SSL_CTX_add_session, take into account that there might be multiple
3565 SSL_SESSION structures with the same session ID (e.g. when two threads
3566 concurrently obtain them from an external cache).
3567 The internal cache can handle only one SSL_SESSION with a given ID,
3568 so if there's a conflict, we now throw out the old one to achieve
3572 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3573 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3574 some routines that use cipher OIDs: some ciphers do not have OIDs
3575 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3579 *) Simplify the trust setting structure and code. Now we just have
3580 two sequences of OIDs for trusted and rejected settings. These will
3581 typically have values the same as the extended key usage extension
3582 and any application specific purposes.
3584 The trust checking code now has a default behaviour: it will just
3585 check for an object with the same NID as the passed id. Functions can
3586 be provided to override either the default behaviour or the behaviour
3587 for a given id. SSL client, server and email already have functions
3588 in place for compatibility: they check the NID and also return "trusted"
3589 if the certificate is self signed.
3592 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3593 traditional format into an EVP_PKEY structure.
3596 *) Add a password callback function PEM_cb() which either prompts for
3597 a password if usr_data is NULL or otherwise assumes it is a null
3598 terminated password. Allow passwords to be passed on command line
3599 environment or config files in a few more utilities.
3602 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3603 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3604 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3605 Update documentation.
3608 *) Support for ASN1 "NULL" type. This could be handled before by using
3609 ASN1_TYPE but there wasn't any function that would try to read a NULL
3610 and produce an error if it couldn't. For compatibility we also have
3611 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3612 don't allocate anything because they don't need to.
3615 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3617 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3619 *) Rebuild of the memory allocation routines used by OpenSSL code and
3620 possibly others as well. The purpose is to make an interface that
3621 provide hooks so anyone can build a separate set of allocation and
3622 deallocation routines to be used by OpenSSL, for example memory
3623 pool implementations, or something else, which was previously hard
3624 since Malloc(), Realloc() and Free() were defined as macros having
3625 the values malloc, realloc and free, respectively (except for Win32
3626 compilations). The same is provided for memory debugging code.
3627 OpenSSL already comes with functionality to find memory leaks, but
3628 this gives people a chance to debug other memory problems.
3630 With these changes, a new set of functions and macros have appeared:
3632 CRYPTO_set_mem_debug_functions() [F]
3633 CRYPTO_get_mem_debug_functions() [F]
3634 CRYPTO_dbg_set_options() [F]
3635 CRYPTO_dbg_get_options() [F]
3636 CRYPTO_malloc_debug_init() [M]
3638 The memory debug functions are NULL by default, unless the library
3639 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3640 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3641 gives the standard debugging functions that come with OpenSSL) or
3642 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3643 provided by the library user) must be used. When the standard
3644 debugging functions are used, CRYPTO_dbg_set_options can be used to
3645 request additional information:
3646 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3647 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3649 Also, things like CRYPTO_set_mem_functions will always give the
3650 expected result (the new set of functions is used for allocation
3651 and deallocation) at all times, regardless of platform and compiler
3654 To finish it up, some functions that were never use in any other
3655 way than through macros have a new API and new semantic:
3658 CRYPTO_dbg_realloc()
3661 All macros of value have retained their old syntax.
3662 [Richard Levitte and Bodo Moeller]
3664 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3665 ordering of SMIMECapabilities wasn't in "strength order" and there
3666 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3670 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3671 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3672 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3674 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3675 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3676 functionality to handle multipart/signed properly) and a utility
3677 called 'smime' to call all this stuff. This is based on code I
3678 originally wrote for Celo who have kindly allowed it to be
3679 included in OpenSSL.
3682 *) Add variants des_set_key_checked and des_set_key_unchecked of
3683 des_set_key (aka des_key_sched). Global variable des_check_key
3684 decides which of these is called by des_set_key; this way
3685 des_check_key behaves as it always did, but applications and
3686 the library itself, which was buggy for des_check_key == 1,
3687 have a cleaner way to pick the version they need.
3690 *) New function PKCS12_newpass() which changes the password of a
3694 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3695 dynamic mix. In both cases the ids can be used as an index into the
3696 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3697 functions so they accept a list of the field values and the
3698 application doesn't need to directly manipulate the X509_TRUST
3702 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3706 *) Modify the way the V3 extension code looks up extensions. This now
3707 works in a similar way to the object code: we have some "standard"
3708 extensions in a static table which is searched with OBJ_bsearch()
3709 and the application can add dynamic ones if needed. The file
3710 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3711 updated whenever a new extension is added to the core code and kept
3712 in ext_nid order. There is a simple program 'tabtest.c' which checks
3713 this. New extensions are not added too often so this file can readily
3714 be maintained manually.
3716 There are two big advantages in doing things this way. The extensions
3717 can be looked up immediately and no longer need to be "added" using
3718 X509V3_add_standard_extensions(): this function now does nothing.
3719 [Side note: I get *lots* of email saying the extension code doesn't
3720 work because people forget to call this function]
3721 Also no dynamic allocation is done unless new extensions are added:
3722 so if we don't add custom extensions there is no need to call
3723 X509V3_EXT_cleanup().
3726 *) Modify enc utility's salting as follows: make salting the default. Add a
3727 magic header, so unsalted files fail gracefully instead of just decrypting
3728 to garbage. This is because not salting is a big security hole, so people
3729 should be discouraged from doing it.
3732 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3733 digest to be passed on the command line but it only used this
3734 parameter when signing a certificate. Modified so all relevant
3735 operations are affected by the digest parameter including the
3736 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3737 DSA key was used because it didn't fix the digest.
3740 *) Initial certificate chain verify code. Currently tests the untrusted
3741 certificates for consistency with the verify purpose (which is set
3742 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3744 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3745 this is because it will reject chains with invalid extensions whereas
3746 every previous version of OpenSSL and SSLeay made no checks at all.
3748 Trust code: checks the root CA for the relevant trust settings. Trust
3749 settings have an initial value consistent with the verify purpose: e.g.
3750 if the verify purpose is for SSL client use it expects the CA to be
3751 trusted for SSL client use. However the default value can be changed to
3752 permit custom trust settings: one example of this would be to only trust
3753 certificates from a specific "secure" set of CAs.
3755 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3756 which should be used for version portability: especially since the
3757 verify structure is likely to change more often now.
3759 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3760 to set them. If not set then assume SSL clients will verify SSL servers
3763 Two new options to the verify program: -untrusted allows a set of
3764 untrusted certificates to be passed in and -purpose which sets the
3765 intended purpose of the certificate. If a purpose is set then the
3766 new chain verify code is used to check extension consistency.
3769 *) Support for the authority information access extension.
3772 *) Modify RSA and DSA PEM read routines to transparently handle
3773 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3774 public keys in a format compatible with certificate
3775 SubjectPublicKeyInfo structures. Unfortunately there were already
3776 functions called *_PublicKey_* which used various odd formats so
3777 these are retained for compatibility: however the DSA variants were
3778 never in a public release so they have been deleted. Changed dsa/rsa
3779 utilities to handle the new format: note no releases ever handled public
3780 keys so we should be OK.
3782 The primary motivation for this change is to avoid the same fiasco
3783 that dogs private keys: there are several incompatible private key
3784 formats some of which are standard and some OpenSSL specific and
3785 require various evil hacks to allow partial transparent handling and
3786 even then it doesn't work with DER formats. Given the option anything
3787 other than PKCS#8 should be dumped: but the other formats have to
3788 stay in the name of compatibility.
3790 With public keys and the benefit of hindsight one standard format
3791 is used which works with EVP_PKEY, RSA or DSA structures: though
3792 it clearly returns an error if you try to read the wrong kind of key.
3794 Added a -pubkey option to the 'x509' utility to output the public key.
3795 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3796 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
3797 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
3798 that do the same as the EVP_PKEY_assign_*() except they up the
3799 reference count of the added key (they don't "swallow" the
3803 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
3804 CRLs would fail if the file contained no certificates or no CRLs:
3805 added a new function to read in both types and return the number
3806 read: this means that if none are read it will be an error. The
3807 DER versions of the certificate and CRL reader would always fail
3808 because it isn't possible to mix certificates and CRLs in DER format
3809 without choking one or the other routine. Changed this to just read
3810 a certificate: this is the best we can do. Also modified the code
3811 in apps/verify.c to take notice of return codes: it was previously
3812 attempting to read in certificates from NULL pointers and ignoring
3813 any errors: this is one reason why the cert and CRL reader seemed
3814 to work. It doesn't check return codes from the default certificate
3815 routines: these may well fail if the certificates aren't installed.
3818 *) Code to support otherName option in GeneralName.
3821 *) First update to verify code. Change the verify utility
3822 so it warns if it is passed a self signed certificate:
3823 for consistency with the normal behaviour. X509_verify
3824 has been modified to it will now verify a self signed
3825 certificate if *exactly* the same certificate appears
3826 in the store: it was previously impossible to trust a
3827 single self signed certificate. This means that:
3828 openssl verify ss.pem
3829 now gives a warning about a self signed certificate but
3830 openssl verify -CAfile ss.pem ss.pem
3834 *) For servers, store verify_result in SSL_SESSION data structure
3835 (and add it to external session representation).
3836 This is needed when client certificate verifications fails,
3837 but an application-provided verification callback (set by
3838 SSL_CTX_set_cert_verify_callback) allows accepting the session
3839 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
3840 but returns 1): When the session is reused, we have to set
3841 ssl->verify_result to the appropriate error code to avoid
3843 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
3845 *) Fix a bug in the new PKCS#7 code: it didn't consider the
3846 case in PKCS7_dataInit() where the signed PKCS7 structure
3847 didn't contain any existing data because it was being created.
3848 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
3850 *) Add a salt to the key derivation routines in enc.c. This
3851 forms the first 8 bytes of the encrypted file. Also add a
3852 -S option to allow a salt to be input on the command line.
3855 *) New function X509_cmp(). Oddly enough there wasn't a function
3856 to compare two certificates. We do this by working out the SHA1
3857 hash and comparing that. X509_cmp() will be needed by the trust
3861 *) SSL_get1_session() is like SSL_get_session(), but increments
3862 the reference count in the SSL_SESSION returned.
3863 [Geoff Thorpe <geoff@eu.c2.net>]
3865 *) Fix for 'req': it was adding a null to request attributes.
3866 Also change the X509_LOOKUP and X509_INFO code to handle
3867 certificate auxiliary information.
3870 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
3874 *) Add the possibility to add extra information to the memory leak
3875 detecting output, to form tracebacks, showing from where each
3876 allocation was originated: CRYPTO_push_info("constant string") adds
3877 the string plus current file name and line number to a per-thread
3878 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
3879 is like calling CYRPTO_pop_info() until the stack is empty.
3880 Also updated memory leak detection code to be multi-thread-safe.
3883 *) Add options -text and -noout to pkcs7 utility and delete the
3884 encryption options which never did anything. Update docs.
3887 *) Add options to some of the utilities to allow the pass phrase
3888 to be included on either the command line (not recommended on
3889 OSes like Unix) or read from the environment. Update the
3890 manpages and fix a few bugs.
3893 *) Add a few manpages for some of the openssl commands.
3896 *) Fix the -revoke option in ca. It was freeing up memory twice,
3897 leaking and not finding already revoked certificates.
3900 *) Extensive changes to support certificate auxiliary information.
3901 This involves the use of X509_CERT_AUX structure and X509_AUX
3902 functions. An X509_AUX function such as PEM_read_X509_AUX()
3903 can still read in a certificate file in the usual way but it
3904 will also read in any additional "auxiliary information". By
3905 doing things this way a fair degree of compatibility can be
3906 retained: existing certificates can have this information added
3907 using the new 'x509' options.
3909 Current auxiliary information includes an "alias" and some trust
3910 settings. The trust settings will ultimately be used in enhanced
3911 certificate chain verification routines: currently a certificate
3912 can only be trusted if it is self signed and then it is trusted
3916 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
3917 The problem was that one of the replacement routines had not been working
3918 since SSLeay releases. For now the offending routine has been replaced
3919 with non-optimised assembler. Even so, this now gives around 95%
3920 performance improvement for 1024 bit RSA signs.
3923 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
3924 handling. Most clients have the effective key size in bits equal to
3925 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
3926 A few however don't do this and instead use the size of the decrypted key
3927 to determine the RC2 key length and the AlgorithmIdentifier to determine
3928 the effective key length. In this case the effective key length can still
3929 be 40 bits but the key length can be 168 bits for example. This is fixed
3930 by manually forcing an RC2 key into the EVP_PKEY structure because the
3931 EVP code can't currently handle unusual RC2 key sizes: it always assumes
3932 the key length and effective key length are equal.
3935 *) Add a bunch of functions that should simplify the creation of
3936 X509_NAME structures. Now you should be able to do:
3937 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
3938 and have it automatically work out the correct field type and fill in
3939 the structures. The more adventurous can try:
3940 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
3941 and it will (hopefully) work out the correct multibyte encoding.
3944 *) Change the 'req' utility to use the new field handling and multibyte
3945 copy routines. Before the DN field creation was handled in an ad hoc
3946 way in req, ca, and x509 which was rather broken and didn't support
3947 BMPStrings or UTF8Strings. Since some software doesn't implement
3948 BMPStrings or UTF8Strings yet, they can be enabled using the config file
3949 using the dirstring_type option. See the new comment in the default
3950 openssl.cnf for more info.
3953 *) Make crypto/rand/md_rand.c more robust:
3954 - Assure unique random numbers after fork().
3955 - Make sure that concurrent threads access the global counter and
3956 md serializably so that we never lose entropy in them
3957 or use exactly the same state in multiple threads.
3958 Access to the large state is not always serializable because
3959 the additional locking could be a performance killer, and
3960 md should be large enough anyway.
3963 *) New file apps/app_rand.c with commonly needed functionality
3964 for handling the random seed file.
3966 Use the random seed file in some applications that previously did not:
3968 dsaparam -genkey (which also ignored its '-rand' option),
3971 x509 (when signing).
3972 Except on systems with /dev/urandom, it is crucial to have a random
3973 seed file at least for key creation, DSA signing, and for DH exchanges;
3974 for RSA signatures we could do without one.
3976 gendh and gendsa (unlike genrsa) used to read only the first byte
3977 of each file listed in the '-rand' option. The function as previously
3978 found in genrsa is now in app_rand.c and is used by all programs
3979 that support '-rand'.
3982 *) In RAND_write_file, use mode 0600 for creating files;
3983 don't just chmod when it may be too late.
3986 *) Report an error from X509_STORE_load_locations
3987 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
3990 *) New function ASN1_mbstring_copy() this copies a string in either
3991 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
3992 into an ASN1_STRING type. A mask of permissible types is passed
3993 and it chooses the "minimal" type to use or an error if not type
3997 *) Add function equivalents to the various macros in asn1.h. The old
3998 macros are retained with an M_ prefix. Code inside the library can
3999 use the M_ macros. External code (including the openssl utility)
4000 should *NOT* in order to be "shared library friendly".
4003 *) Add various functions that can check a certificate's extensions
4004 to see if it usable for various purposes such as SSL client,
4005 server or S/MIME and CAs of these types. This is currently
4006 VERY EXPERIMENTAL but will ultimately be used for certificate chain
4007 verification. Also added a -purpose flag to x509 utility to
4008 print out all the purposes.
4011 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
4015 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
4016 for, obtain and decode and extension and obtain its critical flag.
4017 This allows all the necessary extension code to be handled in a
4018 single function call.
4021 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
4022 platforms. See crypto/rc4/rc4_enc.c for further details.
4025 *) New -noout option to asn1parse. This causes no output to be produced
4026 its main use is when combined with -strparse and -out to extract data
4027 from a file (which may not be in ASN.1 format).
4030 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
4031 when producing the local key id.
4032 [Richard Levitte <levitte@stacken.kth.se>]
4034 *) New option -dhparam in s_server. This allows a DH parameter file to be
4035 stated explicitly. If it is not stated then it tries the first server
4036 certificate file. The previous behaviour hard coded the filename
4040 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
4041 a public key to be input or output. For example:
4042 openssl rsa -in key.pem -pubout -out pubkey.pem
4043 Also added necessary DSA public key functions to handle this.
4046 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
4047 in the message. This was handled by allowing
4048 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
4049 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
4051 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
4052 to the end of the strings whereas this didn't. This would cause problems
4053 if strings read with d2i_ASN1_bytes() were later modified.
4054 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
4056 *) Fix for base64 decode bug. When a base64 bio reads only one line of
4057 data and it contains EOF it will end up returning an error. This is
4058 caused by input 46 bytes long. The cause is due to the way base64
4059 BIOs find the start of base64 encoded data. They do this by trying a
4060 trial decode on each line until they find one that works. When they
4061 do a flag is set and it starts again knowing it can pass all the
4062 data directly through the decoder. Unfortunately it doesn't reset
4063 the context it uses. This means that if EOF is reached an attempt
4064 is made to pass two EOFs through the context and this causes the
4065 resulting error. This can also cause other problems as well. As is
4066 usual with these problems it takes *ages* to find and the fix is
4067 trivial: move one line.
4068 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
4070 *) Ugly workaround to get s_client and s_server working under Windows. The
4071 old code wouldn't work because it needed to select() on sockets and the
4072 tty (for keypresses and to see if data could be written). Win32 only
4073 supports select() on sockets so we select() with a 1s timeout on the
4074 sockets and then see if any characters are waiting to be read, if none
4075 are present then we retry, we also assume we can always write data to
4076 the tty. This isn't nice because the code then blocks until we've
4077 received a complete line of data and it is effectively polling the
4078 keyboard at 1s intervals: however it's quite a bit better than not
4079 working at all :-) A dedicated Windows application might handle this
4080 with an event loop for example.
4083 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
4084 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
4085 will be called when RSA_sign() and RSA_verify() are used. This is useful
4086 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
4087 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
4088 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
4089 This necessitated the support of an extra signature type NID_md5_sha1
4090 for SSL signatures and modifications to the SSL library to use it instead
4091 of calling RSA_public_decrypt() and RSA_private_encrypt().
4094 *) Add new -verify -CAfile and -CApath options to the crl program, these
4095 will lookup a CRL issuers certificate and verify the signature in a
4096 similar way to the verify program. Tidy up the crl program so it
4097 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
4098 less strict. It will now permit CRL extensions even if it is not
4099 a V2 CRL: this will allow it to tolerate some broken CRLs.
4102 *) Initialize all non-automatic variables each time one of the openssl
4103 sub-programs is started (this is necessary as they may be started
4104 multiple times from the "OpenSSL>" prompt).
4105 [Lennart Bang, Bodo Moeller]
4107 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
4108 removing all other RSA functionality (this is what NO_RSA does). This
4109 is so (for example) those in the US can disable those operations covered
4110 by the RSA patent while allowing storage and parsing of RSA keys and RSA
4114 *) Non-copying interface to BIO pairs.
4115 (still largely untested)
4118 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
4119 ASCII string. This was handled independently in various places before.
4122 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
4123 UTF8 strings a character at a time.
4126 *) Use client_version from client hello to select the protocol
4127 (s23_srvr.c) and for RSA client key exchange verification
4128 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
4131 *) Add various utility functions to handle SPKACs, these were previously
4132 handled by poking round in the structure internals. Added new function
4133 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
4134 print, verify and generate SPKACs. Based on an original idea from
4135 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
4138 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
4141 *) Allow the config file extension section to be overwritten on the
4142 command line. Based on an original idea from Massimiliano Pala
4143 <madwolf@comune.modena.it>. The new option is called -extensions
4144 and can be applied to ca, req and x509. Also -reqexts to override
4145 the request extensions in req and -crlexts to override the crl extensions
4149 *) Add new feature to the SPKAC handling in ca. Now you can include
4150 the same field multiple times by preceding it by "XXXX." for example:
4153 this is the same syntax as used in the req config file.
4156 *) Allow certificate extensions to be added to certificate requests. These
4157 are specified in a 'req_extensions' option of the req section of the
4158 config file. They can be printed out with the -text option to req but
4159 are otherwise ignored at present.
4162 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
4163 data read consists of only the final block it would not decrypted because
4164 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
4165 A misplaced 'break' also meant the decrypted final block might not be
4166 copied until the next read.
4169 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
4170 a few extra parameters to the DH structure: these will be useful if
4171 for example we want the value of 'q' or implement X9.42 DH.
4174 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
4175 provides hooks that allow the default DSA functions or functions on a
4176 "per key" basis to be replaced. This allows hardware acceleration and
4177 hardware key storage to be handled without major modification to the
4178 library. Also added low level modexp hooks and CRYPTO_EX structure and
4179 associated functions.
4182 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
4183 as "read only": it can't be written to and the buffer it points to will
4184 not be freed. Reading from a read only BIO is much more efficient than
4185 a normal memory BIO. This was added because there are several times when
4186 an area of memory needs to be read from a BIO. The previous method was
4187 to create a memory BIO and write the data to it, this results in two
4188 copies of the data and an O(n^2) reading algorithm. There is a new
4189 function BIO_new_mem_buf() which creates a read only memory BIO from
4190 an area of memory. Also modified the PKCS#7 routines to use read only
4194 *) Bugfix: ssl23_get_client_hello did not work properly when called in
4195 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
4196 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
4197 but a retry condition occured while trying to read the rest.
4200 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
4201 NID_pkcs7_encrypted by default: this was wrong since this should almost
4202 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
4203 the encrypted data type: this is a more sensible place to put it and it
4204 allows the PKCS#12 code to be tidied up that duplicated this
4208 *) Changed obj_dat.pl script so it takes its input and output files on
4209 the command line. This should avoid shell escape redirection problems
4213 *) Initial support for certificate extension requests, these are included
4214 in things like Xenroll certificate requests. Included functions to allow
4215 extensions to be obtained and added.
4218 *) -crlf option to s_client and s_server for sending newlines as
4219 CRLF (as required by many protocols).
4222 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
4224 *) Install libRSAglue.a when OpenSSL is built with RSAref.
4225 [Ralf S. Engelschall]
4227 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
4228 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
4230 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
4234 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
4235 DH parameters/keys (q is lost during that conversion, but the resulting
4236 DH parameters contain its length).
4238 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
4239 much faster than DH_generate_parameters (which creates parameters
4240 where p = 2*q + 1), and also the smaller q makes DH computations
4241 much more efficient (160-bit exponentiation instead of 1024-bit
4242 exponentiation); so this provides a convenient way to support DHE
4243 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
4244 utter importance to use
4245 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4247 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
4248 when such DH parameters are used, because otherwise small subgroup
4249 attacks may become possible!
4252 *) Avoid memory leak in i2d_DHparams.
4255 *) Allow the -k option to be used more than once in the enc program:
4256 this allows the same encrypted message to be read by multiple recipients.
4259 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
4260 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
4261 it will always use the numerical form of the OID, even if it has a short
4265 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
4266 method only got called if p,q,dmp1,dmq1,iqmp components were present,
4267 otherwise bn_mod_exp was called. In the case of hardware keys for example
4268 no private key components need be present and it might store extra data
4269 in the RSA structure, which cannot be accessed from bn_mod_exp.
4270 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
4271 private key operations.
4274 *) Added support for SPARC Linux.
4277 *) pem_password_cb function type incompatibly changed from
4278 typedef int pem_password_cb(char *buf, int size, int rwflag);
4280 ....(char *buf, int size, int rwflag, void *userdata);
4281 so that applications can pass data to their callbacks:
4282 The PEM[_ASN1]_{read,write}... functions and macros now take an
4283 additional void * argument, which is just handed through whenever
4284 the password callback is called.
4285 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
4287 New function SSL_CTX_set_default_passwd_cb_userdata.
4289 Compatibility note: As many C implementations push function arguments
4290 onto the stack in reverse order, the new library version is likely to
4291 interoperate with programs that have been compiled with the old
4292 pem_password_cb definition (PEM_whatever takes some data that
4293 happens to be on the stack as its last argument, and the callback
4294 just ignores this garbage); but there is no guarantee whatsoever that
4297 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
4298 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
4299 problems not only on Windows, but also on some Unix platforms.
4300 To avoid problematic command lines, these definitions are now in an
4301 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
4302 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
4305 *) MIPS III/IV assembler module is reimplemented.
4308 *) More DES library cleanups: remove references to srand/rand and
4309 delete an unused file.
4312 *) Add support for the the free Netwide assembler (NASM) under Win32,
4313 since not many people have MASM (ml) and it can be hard to obtain.
4314 This is currently experimental but it seems to work OK and pass all
4315 the tests. Check out INSTALL.W32 for info.
4318 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
4319 without temporary keys kept an extra copy of the server key,
4320 and connections with temporary keys did not free everything in case
4324 *) New function RSA_check_key and new openssl rsa option -check
4325 for verifying the consistency of RSA keys.
4326 [Ulf Moeller, Bodo Moeller]
4328 *) Various changes to make Win32 compile work:
4329 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
4330 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
4331 comparison" warnings.
4332 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
4335 *) Add a debugging option to PKCS#5 v2 key generation function: when
4336 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
4337 derived keys are printed to stderr.
4340 *) Copy the flags in ASN1_STRING_dup().
4341 [Roman E. Pavlov <pre@mo.msk.ru>]
4343 *) The x509 application mishandled signing requests containing DSA
4344 keys when the signing key was also DSA and the parameters didn't match.
4346 It was supposed to omit the parameters when they matched the signing key:
4347 the verifying software was then supposed to automatically use the CA's
4348 parameters if they were absent from the end user certificate.
4350 Omitting parameters is no longer recommended. The test was also
4351 the wrong way round! This was probably due to unusual behaviour in
4352 EVP_cmp_parameters() which returns 1 if the parameters match.
4353 This meant that parameters were omitted when they *didn't* match and
4354 the certificate was useless. Certificates signed with 'ca' didn't have
4356 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
4358 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
4359 The interface is as follows:
4360 Applications can use
4361 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
4362 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
4363 "off" is now the default.
4364 The library internally uses
4365 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
4366 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
4367 to disable memory-checking temporarily.
4369 Some inconsistent states that previously were possible (and were
4370 even the default) are now avoided.
4372 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
4373 with each memory chunk allocated; this is occasionally more helpful
4374 than just having a counter.
4376 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
4378 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
4382 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
4383 which largely parallels "options", but is for changing API behaviour,
4384 whereas "options" are about protocol behaviour.
4385 Initial "mode" flags are:
4387 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
4388 a single record has been written.
4389 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
4390 retries use the same buffer location.
4391 (But all of the contents must be
4395 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
4398 *) Fix problems with no-hmac etc.
4399 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
4401 *) New functions RSA_get_default_method(), RSA_set_method() and
4402 RSA_get_method(). These allows replacement of RSA_METHODs without having
4403 to mess around with the internals of an RSA structure.
4406 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
4407 Also really enable memory leak checks in openssl.c and in some
4409 [Chad C. Mulligan, Bodo Moeller]
4411 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
4412 up the length of negative integers. This has now been simplified to just
4413 store the length when it is first determined and use it later, rather
4414 than trying to keep track of where data is copied and updating it to
4416 [Steve Henson, reported by Brien Wheeler
4417 <bwheeler@authentica-security.com>]
4419 *) Add a new function PKCS7_signatureVerify. This allows the verification
4420 of a PKCS#7 signature but with the signing certificate passed to the
4421 function itself. This contrasts with PKCS7_dataVerify which assumes the
4422 certificate is present in the PKCS#7 structure. This isn't always the
4423 case: certificates can be omitted from a PKCS#7 structure and be
4424 distributed by "out of band" means (such as a certificate database).
4427 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
4428 function prototypes in pem.h, also change util/mkdef.pl to add the
4429 necessary function names.
4432 *) mk1mf.pl (used by Windows builds) did not properly read the
4433 options set by Configure in the top level Makefile, and Configure
4434 was not even able to write more than one option correctly.
4435 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
4438 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
4439 file to be loaded from a BIO or FILE pointer. The BIO version will
4440 for example allow memory BIOs to contain config info.
4443 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
4444 Whoever hopes to achieve shared-library compatibility across versions
4445 must use this, not the compile-time macro.
4446 (Exercise 0.9.4: Which is the minimum library version required by
4448 Note: All this applies only to multi-threaded programs, others don't
4452 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
4453 through a BIO pair triggered the default case, i.e.
4454 SSLerr(...,SSL_R_UNKNOWN_STATE).
4457 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
4458 can use the SSL library even if none of the specific BIOs is
4462 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
4463 for the encoded length.
4464 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
4466 *) Add initial documentation of the X509V3 functions.
4469 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
4470 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
4471 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
4472 secure PKCS#8 private key format with a high iteration count.
4475 *) Fix determination of Perl interpreter: A perl or perl5
4476 _directory_ in $PATH was also accepted as the interpreter.
4477 [Ralf S. Engelschall]
4479 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
4480 wrong with it but it was very old and did things like calling
4481 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
4485 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
4486 to use the new extension code.
4489 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
4490 with macros. This should make it easier to change their form, add extra
4491 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
4495 *) Add to configuration table a new entry that can specify an alternative
4496 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
4497 according to Mark Crispin <MRC@Panda.COM>.
4501 *) DES CBC did not update the IV. Weird.
4504 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
4505 Changing the behaviour of the former might break existing programs --
4506 where IV updating is needed, des_ncbc_encrypt can be used.
4509 *) When bntest is run from "make test" it drives bc to check its
4510 calculations, as well as internally checking them. If an internal check
4511 fails, it needs to cause bc to give a non-zero result or make test carries
4512 on without noticing the failure. Fixed.
4515 *) DES library cleanups.
4518 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
4519 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
4520 ciphers. NOTE: although the key derivation function has been verified
4521 against some published test vectors it has not been extensively tested
4522 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
4526 *) Instead of "mkdir -p", which is not fully portable, use new
4527 Perl script "util/mkdir-p.pl".
4530 *) Rewrite the way password based encryption (PBE) is handled. It used to
4531 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
4532 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
4533 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
4534 the 'parameter' field of the AlgorithmIdentifier is passed to the
4535 underlying key generation function so it must do its own ASN1 parsing.
4536 This has also changed the EVP_PBE_CipherInit() function which now has a
4537 'parameter' argument instead of literal salt and iteration count values
4538 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
4541 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
4542 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
4543 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
4544 KEY" because this clashed with PKCS#8 unencrypted string. Since this
4545 value was just used as a "magic string" and not used directly its
4546 value doesn't matter.
4549 *) Introduce some semblance of const correctness to BN. Shame C doesn't
4553 *) "linux-sparc64" configuration (ultrapenguin).
4554 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
4555 "linux-sparc" configuration.
4556 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
4558 *) config now generates no-xxx options for missing ciphers.
4561 *) Support the EBCDIC character set (work in progress).
4562 File ebcdic.c not yet included because it has a different license.
4563 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4565 *) Support BS2000/OSD-POSIX.
4566 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
4568 *) Make callbacks for key generation use void * instead of char *.
4571 *) Make S/MIME samples compile (not yet tested).
4574 *) Additional typesafe stacks.
4577 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
4581 Changes between 0.9.3 and 0.9.3a [29 May 1999]
4583 *) New configuration variant "sco5-gcc".
4585 *) Updated some demos.
4586 [Sean O Riordain, Wade Scholine]
4588 *) Add missing BIO_free at exit of pkcs12 application.
4591 *) Fix memory leak in conf.c.
4594 *) Updates for Win32 to assembler version of MD5.
4597 *) Set #! path to perl in apps/der_chop to where we found it
4598 instead of using a fixed path.
4601 *) SHA library changes for irix64-mips4-cc.
4604 *) Improvements for VMS support.
4608 Changes between 0.9.2b and 0.9.3 [24 May 1999]
4610 *) Bignum library bug fix. IRIX 6 passes "make test" now!
4611 This also avoids the problems with SC4.2 and unpatched SC5.
4612 [Andy Polyakov <appro@fy.chalmers.se>]
4614 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
4615 These are required because of the typesafe stack would otherwise break
4616 existing code. If old code used a structure member which used to be STACK
4617 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
4618 sk_num or sk_value it would produce an error because the num, data members
4619 are not present in STACK_OF. Now it just produces a warning. sk_set
4620 replaces the old method of assigning a value to sk_value
4621 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
4622 that does this will no longer work (and should use sk_set instead) but
4623 this could be regarded as a "questionable" behaviour anyway.
4626 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
4627 correctly handle encrypted S/MIME data.
4630 *) Change type of various DES function arguments from des_cblock
4631 (which means, in function argument declarations, pointer to char)
4632 to des_cblock * (meaning pointer to array with 8 char elements),
4633 which allows the compiler to do more typechecking; it was like
4634 that back in SSLeay, but with lots of ugly casts.
4636 Introduce new type const_des_cblock.
4639 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
4640 problems: find RecipientInfo structure that matches recipient certificate
4641 and initialise the ASN1 structures properly based on passed cipher.
4644 *) Belatedly make the BN tests actually check the results.
4647 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
4648 to and from BNs: it was completely broken. New compilation option
4649 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
4650 key elements as negative integers.
4653 *) Reorganize and speed up MD5.
4654 [Andy Polyakov <appro@fy.chalmers.se>]
4657 [Richard Levitte <richard@levitte.org>]
4659 *) New option -out to asn1parse to allow the parsed structure to be
4660 output to a file. This is most useful when combined with the -strparse
4661 option to examine the output of things like OCTET STRINGS.
4664 *) Make SSL library a little more fool-proof by not requiring any longer
4665 that SSL_set_{accept,connect}_state be called before
4666 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
4667 in many applications because usually everything *appeared* to work as
4668 intended anyway -- now it really works as intended).
4671 *) Move openssl.cnf out of lib/.
4674 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
4675 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
4676 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
4677 [Ralf S. Engelschall]
4679 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
4680 handle PKCS#7 enveloped data properly.
4681 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
4683 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
4684 copying pointers. The cert_st handling is changed by this in
4685 various ways (and thus what used to be known as ctx->default_cert
4686 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
4687 any longer when s->cert does not give us what we need).
4688 ssl_cert_instantiate becomes obsolete by this change.
4689 As soon as we've got the new code right (possibly it already is?),
4690 we have solved a couple of bugs of the earlier code where s->cert
4691 was used as if it could not have been shared with other SSL structures.
4693 Note that using the SSL API in certain dirty ways now will result
4694 in different behaviour than observed with earlier library versions:
4695 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
4696 does not influence s as it used to.
4698 In order to clean up things more thoroughly, inside SSL_SESSION
4699 we don't use CERT any longer, but a new structure SESS_CERT
4700 that holds per-session data (if available); currently, this is
4701 the peer's certificate chain and, for clients, the server's certificate
4702 and temporary key. CERT holds only those values that can have
4703 meaningful defaults in an SSL_CTX.
4706 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
4707 from the internal representation. Various PKCS#7 fixes: remove some
4708 evil casts and set the enc_dig_alg field properly based on the signing
4712 *) Allow PKCS#12 password to be set from the command line or the
4713 environment. Let 'ca' get its config file name from the environment
4714 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
4718 *) Allow certificate policies extension to use an IA5STRING for the
4719 organization field. This is contrary to the PKIX definition but
4720 VeriSign uses it and IE5 only recognises this form. Document 'x509'
4724 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
4725 without disallowing inline assembler and the like for non-pedantic builds.
4728 *) Support Borland C++ builder.
4729 [Janez Jere <jj@void.si>, modified by Ulf Möller]
4734 *) SHA-1 cleanups and performance enhancements.
4735 [Andy Polyakov <appro@fy.chalmers.se>]
4737 *) Sparc v8plus assembler for the bignum library.
4738 [Andy Polyakov <appro@fy.chalmers.se>]
4740 *) Accept any -xxx and +xxx compiler options in Configure.
4743 *) Update HPUX configuration.
4746 *) Add missing sk_<type>_unshift() function to safestack.h
4747 [Ralf S. Engelschall]
4749 *) New function SSL_CTX_use_certificate_chain_file that sets the
4750 "extra_cert"s in addition to the certificate. (This makes sense
4751 only for "PEM" format files, as chains as a whole are not
4755 *) Support verify_depth from the SSL API.
4756 x509_vfy.c had what can be considered an off-by-one-error:
4757 Its depth (which was not part of the external interface)
4758 was actually counting the number of certificates in a chain;
4759 now it really counts the depth.
4762 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
4763 instead of X509err, which often resulted in confusing error
4764 messages since the error codes are not globally unique
4765 (e.g. an alleged error in ssl3_accept when a certificate
4766 didn't match the private key).
4768 *) New function SSL_CTX_set_session_id_context that allows to set a default
4769 value (so that you don't need SSL_set_session_id_context for each
4770 connection using the SSL_CTX).
4773 *) OAEP decoding bug fix.
4776 *) Support INSTALL_PREFIX for package builders, as proposed by
4780 *) New Configure options "threads" and "no-threads". For systems
4781 where the proper compiler options are known (currently Solaris
4782 and Linux), "threads" is the default.
4785 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
4788 *) Install various scripts to $(OPENSSLDIR)/misc, not to
4789 $(INSTALLTOP)/bin -- they shouldn't clutter directories
4790 such as /usr/local/bin.
4793 *) "make linux-shared" to build shared libraries.
4794 [Niels Poppe <niels@netbox.org>]
4796 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
4799 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
4800 extension adding in x509 utility.
4803 *) Remove NOPROTO sections and error code comments.
4806 *) Partial rewrite of the DEF file generator to now parse the ANSI
4810 *) New Configure options --prefix=DIR and --openssldir=DIR.
4813 *) Complete rewrite of the error code script(s). It is all now handled
4814 by one script at the top level which handles error code gathering,
4815 header rewriting and C source file generation. It should be much better
4816 than the old method: it now uses a modified version of Ulf's parser to
4817 read the ANSI prototypes in all header files (thus the old K&R definitions
4818 aren't needed for error creation any more) and do a better job of
4819 translating function codes into names. The old 'ASN1 error code imbedded
4820 in a comment' is no longer necessary and it doesn't use .err files which
4821 have now been deleted. Also the error code call doesn't have to appear all
4822 on one line (which resulted in some large lines...).
4825 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
4828 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
4829 0 (which usually indicates a closed connection), but continue reading.
4832 *) Fix some race conditions.
4835 *) Add support for CRL distribution points extension. Add Certificate
4836 Policies and CRL distribution points documentation.
4839 *) Move the autogenerated header file parts to crypto/opensslconf.h.
4842 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
4843 8 of keying material. Merlin has also confirmed interop with this fix
4844 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
4845 [Merlin Hughes <merlin@baltimore.ie>]
4847 *) Fix lots of warnings.
4848 [Richard Levitte <levitte@stacken.kth.se>]
4850 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
4851 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
4852 [Richard Levitte <levitte@stacken.kth.se>]
4854 *) Fix problems with sizeof(long) == 8.
4855 [Andy Polyakov <appro@fy.chalmers.se>]
4857 *) Change functions to ANSI C.
4860 *) Fix typos in error codes.
4861 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
4863 *) Remove defunct assembler files from Configure.
4866 *) SPARC v8 assembler BIGNUM implementation.
4867 [Andy Polyakov <appro@fy.chalmers.se>]
4869 *) Support for Certificate Policies extension: both print and set.
4870 Various additions to support the r2i method this uses.
4873 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
4874 return a const string when you are expecting an allocated buffer.
4877 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
4878 types DirectoryString and DisplayText.
4881 *) Add code to allow r2i extensions to access the configuration database,
4882 add an LHASH database driver and add several ctx helper functions.
4885 *) Fix an evil bug in bn_expand2() which caused various BN functions to
4886 fail when they extended the size of a BIGNUM.
4889 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
4890 support typesafe stack.
4893 *) Fix typo in SSL_[gs]et_options().
4894 [Nils Frostberg <nils@medcom.se>]
4896 *) Delete various functions and files that belonged to the (now obsolete)
4897 old X509V3 handling code.
4900 *) New Configure option "rsaref".
4903 *) Don't auto-generate pem.h.
4906 *) Introduce type-safe ASN.1 SETs.
4909 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
4910 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
4912 *) Introduce type-safe STACKs. This will almost certainly break lots of code
4913 that links with OpenSSL (well at least cause lots of warnings), but fear
4914 not: the conversion is trivial, and it eliminates loads of evil casts. A
4915 few STACKed things have been converted already. Feel free to convert more.
4916 In the fullness of time, I'll do away with the STACK type altogether.
4919 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
4920 specified in <certfile> by updating the entry in the index.txt file.
4921 This way one no longer has to edit the index.txt file manually for
4922 revoking a certificate. The -revoke option does the gory details now.
4923 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
4925 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
4926 `-text' option at all and this way the `-noout -text' combination was
4927 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
4928 [Ralf S. Engelschall]
4930 *) Make sure a corresponding plain text error message exists for the
4931 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
4932 verify callback function determined that a certificate was revoked.
4933 [Ralf S. Engelschall]
4935 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
4936 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
4937 all available cipers including rc5, which was forgotten until now.
4938 In order to let the testing shell script know which algorithms
4939 are available, a new (up to now undocumented) command
4940 "openssl list-cipher-commands" is used.
4943 *) Bugfix: s_client occasionally would sleep in select() when
4944 it should have checked SSL_pending() first.
4947 *) New functions DSA_do_sign and DSA_do_verify to provide access to
4948 the raw DSA values prior to ASN.1 encoding.
4951 *) Tweaks to Configure
4952 [Niels Poppe <niels@netbox.org>]
4954 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
4958 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
4961 *) New config option to avoid instructions that are illegal on the 80386.
4962 The default code is faster, but requires at least a 486.
4965 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
4966 SSL2_SERVER_VERSION (not used at all) macros, which are now the
4967 same as SSL2_VERSION anyway.
4970 *) New "-showcerts" option for s_client.
4973 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
4974 application. Various cleanups and fixes.
4977 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
4978 modify error routines to work internally. Add error codes and PBE init
4979 to library startup routines.
4982 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
4983 packing functions to asn1 and evp. Changed function names and error
4984 codes along the way.
4987 *) PKCS12 integration: and so it begins... First of several patches to
4988 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
4989 objects to objects.h
4992 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
4993 and display support for Thawte strong extranet extension.
4996 *) Add LinuxPPC support.
4997 [Jeff Dubrule <igor@pobox.org>]
4999 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
5000 bn_div_words in alpha.s.
5001 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
5003 *) Make sure the RSA OAEP test is skipped under -DRSAref because
5004 OAEP isn't supported when OpenSSL is built with RSAref.
5005 [Ulf Moeller <ulf@fitug.de>]
5007 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
5008 so they no longer are missing under -DNOPROTO.
5009 [Soren S. Jorvang <soren@t.dk>]
5012 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
5014 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
5015 doesn't work when the session is reused. Coming soon!
5018 *) Fix a security hole, that allows sessions to be reused in the wrong
5019 context thus bypassing client cert protection! All software that uses
5020 client certs and session caches in multiple contexts NEEDS PATCHING to
5021 allow session reuse! A fuller solution is in the works.
5022 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
5024 *) Some more source tree cleanups (removed obsolete files
5025 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
5026 permission on "config" script to be executable) and a fix for the INSTALL
5028 [Ulf Moeller <ulf@fitug.de>]
5030 *) Remove some legacy and erroneous uses of malloc, free instead of
5032 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
5034 *) Make rsa_oaep_test return non-zero on error.
5035 [Ulf Moeller <ulf@fitug.de>]
5037 *) Add support for native Solaris shared libraries. Configure
5038 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
5039 if someone would make that last step automatic.
5040 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
5042 *) ctx_size was not built with the right compiler during "make links". Fixed.
5045 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
5046 except NULL ciphers". This means the default cipher list will no longer
5047 enable NULL ciphers. They need to be specifically enabled e.g. with
5048 the string "DEFAULT:eNULL".
5051 *) Fix to RSA private encryption routines: if p < q then it would
5052 occasionally produce an invalid result. This will only happen with
5053 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
5056 *) Be less restrictive and allow also `perl util/perlpath.pl
5057 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
5058 because this way one can also use an interpreter named `perl5' (which is
5059 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
5060 installed as `perl').
5061 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5063 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
5064 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5066 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
5067 advapi32.lib to Win32 build and change the pem test comparision
5068 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
5069 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
5070 and crypto/des/ede_cbcm_enc.c.
5073 *) DES quad checksum was broken on big-endian architectures. Fixed.
5076 *) Comment out two functions in bio.h that aren't implemented. Fix up the
5077 Win32 test batch file so it (might) work again. The Win32 test batch file
5078 is horrible: I feel ill....
5081 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
5082 in e_os.h. Audit of header files to check ANSI and non ANSI
5083 sections: 10 functions were absent from non ANSI section and not exported
5084 from Windows DLLs. Fixed up libeay.num for new functions.
5087 *) Make `openssl version' output lines consistent.
5088 [Ralf S. Engelschall]
5090 *) Fix Win32 symbol export lists for BIO functions: Added
5091 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
5092 to ms/libeay{16,32}.def.
5093 [Ralf S. Engelschall]
5095 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
5096 fine under Unix and passes some trivial tests I've now added. But the
5097 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
5098 added to make sure no one expects that this stuff really works in the
5099 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
5100 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
5102 [Ralf S. Engelschall]
5104 *) Fix the generation of two part addresses in perl.
5105 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
5107 *) Add config entry for Linux on MIPS.
5108 [John Tobey <jtobey@channel1.com>]
5110 *) Make links whenever Configure is run, unless we are on Windoze.
5113 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
5114 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
5118 *) Add a useful kludge to allow package maintainers to specify compiler and
5119 other platforms details on the command line without having to patch the
5120 Configure script everytime: One now can use ``perl Configure
5121 <id>:<details>'', i.e. platform ids are allowed to have details appended
5122 to them (seperated by colons). This is treated as there would be a static
5123 pre-configured entry in Configure's %table under key <id> with value
5124 <details> and ``perl Configure <id>'' is called. So, when you want to
5125 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
5126 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
5127 now, which overrides the FreeBSD-elf entry on-the-fly.
5128 [Ralf S. Engelschall]
5130 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
5133 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
5134 on the `perl Configure ...' command line. This way one can compile
5135 OpenSSL libraries with Position Independent Code (PIC) which is needed
5136 for linking it into DSOs.
5137 [Ralf S. Engelschall]
5139 *) Remarkably, export ciphers were totally broken and no-one had noticed!
5143 *) Cleaned up the LICENSE document: The official contact for any license
5144 questions now is the OpenSSL core team under openssl-core@openssl.org.
5145 And add a paragraph about the dual-license situation to make sure people
5146 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
5147 to the OpenSSL toolkit.
5148 [Ralf S. Engelschall]
5150 *) General source tree makefile cleanups: Made `making xxx in yyy...'
5151 display consistent in the source tree and replaced `/bin/rm' by `rm'.
5152 Additonally cleaned up the `make links' target: Remove unnecessary
5153 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
5154 to speed processing and no longer clutter the display with confusing
5155 stuff. Instead only the actually done links are displayed.
5156 [Ralf S. Engelschall]
5158 *) Permit null encryption ciphersuites, used for authentication only. It used
5159 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
5160 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
5164 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
5165 signed attributes when verifying signatures (this would break them),
5166 the detached data encoding was wrong and public keys obtained using
5167 X509_get_pubkey() weren't freed.
5170 *) Add text documentation for the BUFFER functions. Also added a work around
5171 to a Win95 console bug. This was triggered by the password read stuff: the
5172 last character typed gets carried over to the next fread(). If you were
5173 generating a new cert request using 'req' for example then the last
5174 character of the passphrase would be CR which would then enter the first
5178 *) Added the new `Includes OpenSSL Cryptography Software' button as
5179 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
5180 button and can be used by applications based on OpenSSL to show the
5181 relationship to the OpenSSL project.
5182 [Ralf S. Engelschall]
5184 *) Remove confusing variables in function signatures in files
5185 ssl/ssl_lib.c and ssl/ssl.h.
5186 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5188 *) Don't install bss_file.c under PREFIX/include/
5189 [Lennart Bong <lob@kulthea.stacken.kth.se>]
5191 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
5192 functions that return function pointers and has support for NT specific
5193 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
5194 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
5195 unsigned to signed types: this was killing the Win32 compile.
5198 *) Add new certificate file to stack functions,
5199 SSL_add_dir_cert_subjects_to_stack() and
5200 SSL_add_file_cert_subjects_to_stack(). These largely supplant
5201 SSL_load_client_CA_file(), and can be used to add multiple certs easily
5202 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
5203 This means that Apache-SSL and similar packages don't have to mess around
5204 to add as many CAs as they want to the preferred list.
5207 *) Experiment with doxygen documentation. Currently only partially applied to
5209 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
5210 openssl.doxy as the configuration file.
5213 *) Get rid of remaining C++-style comments which strict C compilers hate.
5214 [Ralf S. Engelschall, pointed out by Carlos Amengual]
5216 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
5217 compiled in by default: it has problems with large keys.
5220 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
5221 DH private keys and/or callback functions which directly correspond to
5222 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
5223 is needed for applications which have to configure certificates on a
5224 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
5226 For the RSA certificate situation is makes no difference, but
5227 for the DSA certificate situation this fixes the "no shared cipher"
5228 problem where the OpenSSL cipher selection procedure failed because the
5229 temporary keys were not overtaken from the context and the API provided
5230 no way to reconfigure them.
5231 The new functions now let applications reconfigure the stuff and they
5232 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
5233 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
5234 non-public-API function ssl_cert_instantiate() is used as a helper
5235 function and also to reduce code redundancy inside ssl_rsa.c.
5236 [Ralf S. Engelschall]
5238 *) Move s_server -dcert and -dkey options out of the undocumented feature
5239 area because they are useful for the DSA situation and should be
5240 recognized by the users.
5241 [Ralf S. Engelschall]
5243 *) Fix the cipher decision scheme for export ciphers: the export bits are
5244 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
5245 SSL_EXP_MASK. So, the original variable has to be used instead of the
5246 already masked variable.
5247 [Richard Levitte <levitte@stacken.kth.se>]
5249 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
5250 [Richard Levitte <levitte@stacken.kth.se>]
5252 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
5253 from `int' to `unsigned int' because it's a length and initialized by
5254 EVP_DigestFinal() which expects an `unsigned int *'.
5255 [Richard Levitte <levitte@stacken.kth.se>]
5257 *) Don't hard-code path to Perl interpreter on shebang line of Configure
5258 script. Instead use the usual Shell->Perl transition trick.
5259 [Ralf S. Engelschall]
5261 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
5262 (in addition to RSA certificates) to match the behaviour of `openssl dsa
5263 -noout -modulus' as it's already the case for `openssl rsa -noout
5264 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
5265 currently the public key is printed (a decision which was already done by
5266 `openssl dsa -modulus' in the past) which serves a similar purpose.
5267 Additionally the NO_RSA no longer completely removes the whole -modulus
5268 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
5270 [Ralf S. Engelschall]
5272 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
5273 BIO. See the source (crypto/evp/bio_ok.c) for more info.
5274 [Arne Ansper <arne@ats.cyber.ee>]
5276 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
5277 to be added. Now both 'req' and 'ca' can use new objects defined in the
5281 *) Add cool BIO that does syslog (or event log on NT).
5282 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
5284 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
5285 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
5286 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
5287 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
5290 *) Add preliminary config info for new extension code.
5293 *) Make RSA_NO_PADDING really use no padding.
5294 [Ulf Moeller <ulf@fitug.de>]
5296 *) Generate errors when private/public key check is done.
5299 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
5300 for some CRL extensions and new objects added.
5303 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
5304 key usage extension and fuller support for authority key id.
5307 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
5308 padding method for RSA, which is recommended for new applications in PKCS
5309 #1 v2.0 (RFC 2437, October 1998).
5310 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
5311 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
5312 against Bleichbacher's attack on RSA.
5313 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
5316 *) Updates to the new SSL compression code
5317 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5319 *) Fix so that the version number in the master secret, when passed
5320 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
5321 (because the server will not accept higher), that the version number
5322 is 0x03,0x01, not 0x03,0x00
5323 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5325 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
5326 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
5327 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
5330 *) Support for RAW extensions where an arbitrary extension can be
5331 created by including its DER encoding. See apps/openssl.cnf for
5335 *) Make sure latest Perl versions don't interpret some generated C array
5336 code as Perl array code in the crypto/err/err_genc.pl script.
5337 [Lars Weber <3weber@informatik.uni-hamburg.de>]
5339 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
5340 not many people have the assembler. Various Win32 compilation fixes and
5341 update to the INSTALL.W32 file with (hopefully) more accurate Win32
5345 *) Modify configure script 'Configure' to automatically create crypto/date.h
5346 file under Win32 and also build pem.h from pem.org. New script
5347 util/mkfiles.pl to create the MINFO file on environments that can't do a
5348 'make files': perl util/mkfiles.pl >MINFO should work.
5351 *) Major rework of DES function declarations, in the pursuit of correctness
5352 and purity. As a result, many evil casts evaporated, and some weirdness,
5353 too. You may find this causes warnings in your code. Zapping your evil
5354 casts will probably fix them. Mostly.
5357 *) Fix for a typo in asn1.h. Bug fix to object creation script
5358 obj_dat.pl. It considered a zero in an object definition to mean
5359 "end of object": none of the objects in objects.h have any zeros
5360 so it wasn't spotted.
5361 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
5363 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
5364 Masking (CBCM). In the absence of test vectors, the best I have been able
5365 to do is check that the decrypt undoes the encrypt, so far. Send me test
5366 vectors if you have them.
5369 *) Correct calculation of key length for export ciphers (too much space was
5370 allocated for null ciphers). This has not been tested!
5373 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
5374 message is now correct (it understands "crypto" and "ssl" on its
5375 command line). There is also now an "update" option. This will update
5376 the util/ssleay.num and util/libeay.num files with any new functions.
5378 perl util/mkdef.pl crypto ssl update
5379 it will update them.
5382 *) Overhauled the Perl interface (perl/*):
5383 - ported BN stuff to OpenSSL's different BN library
5384 - made the perl/ source tree CVS-aware
5385 - renamed the package from SSLeay to OpenSSL (the files still contain
5386 their history because I've copied them in the repository)
5387 - removed obsolete files (the test scripts will be replaced
5388 by better Test::Harness variants in the future)
5389 [Ralf S. Engelschall]
5391 *) First cut for a very conservative source tree cleanup:
5392 1. merge various obsolete readme texts into doc/ssleay.txt
5393 where we collect the old documents and readme texts.
5394 2. remove the first part of files where I'm already sure that we no
5395 longer need them because of three reasons: either they are just temporary
5396 files which were left by Eric or they are preserved original files where
5397 I've verified that the diff is also available in the CVS via "cvs diff
5398 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
5399 the crypto/md/ stuff).
5400 [Ralf S. Engelschall]
5402 *) More extension code. Incomplete support for subject and issuer alt
5403 name, issuer and authority key id. Change the i2v function parameters
5404 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
5405 what that's for :-) Fix to ASN1 macro which messed up
5406 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
5409 *) Preliminary support for ENUMERATED type. This is largely copied from the
5413 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
5414 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5416 *) Make sure `make rehash' target really finds the `openssl' program.
5417 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
5419 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
5420 like to hear about it if this slows down other processors.
5423 *) Add CygWin32 platform information to Configure script.
5424 [Alan Batie <batie@aahz.jf.intel.com>]
5426 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
5427 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
5429 *) New program nseq to manipulate netscape certificate sequences
5432 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
5436 *) Fixes to BN code. Previously the default was to define BN_RECURSION
5437 but the BN code had some problems that would cause failures when
5438 doing certificate verification and some other functions.
5439 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
5441 *) Add ASN1 and PEM code to support netscape certificate sequences.
5444 *) Add ASN1 and PEM code to support netscape certificate sequences.
5447 *) Add several PKIX and private extended key usage OIDs.
5450 *) Modify the 'ca' program to handle the new extension code. Modify
5451 openssl.cnf for new extension format, add comments.
5454 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
5455 and add a sample to openssl.cnf so req -x509 now adds appropriate
5459 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
5460 error code, add initial support to X509_print() and x509 application.
5463 *) Takes a deep breath and start addding X509 V3 extension support code. Add
5464 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
5465 stuff is currently isolated and isn't even compiled yet.
5468 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
5469 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
5470 Removed the versions check from X509 routines when loading extensions:
5471 this allows certain broken certificates that don't set the version
5472 properly to be processed.
5475 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
5476 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
5477 can still be regenerated with "make depend".
5480 *) Spelling mistake in C version of CAST-128.
5481 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
5483 *) Changes to the error generation code. The perl script err-code.pl
5484 now reads in the old error codes and retains the old numbers, only
5485 adding new ones if necessary. It also only changes the .err files if new
5486 codes are added. The makefiles have been modified to only insert errors
5487 when needed (to avoid needlessly modifying header files). This is done
5488 by only inserting errors if the .err file is newer than the auto generated
5489 C file. To rebuild all the error codes from scratch (the old behaviour)
5490 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
5491 or delete all the .err files.
5494 *) CAST-128 was incorrectly implemented for short keys. The C version has
5495 been fixed, but is untested. The assembler versions are also fixed, but
5496 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
5497 to regenerate it if needed.
5498 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
5499 Hagino <itojun@kame.net>]
5501 *) File was opened incorrectly in randfile.c.
5502 [Ulf Möller <ulf@fitug.de>]
5504 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
5505 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
5506 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
5507 al: it's just almost always a UTCTime. Note this patch adds new error
5508 codes so do a "make errors" if there are problems.
5511 *) Correct Linux 1 recognition in config.
5512 [Ulf Möller <ulf@fitug.de>]
5514 *) Remove pointless MD5 hash when using DSA keys in ca.
5515 [Anonymous <nobody@replay.com>]
5517 *) Generate an error if given an empty string as a cert directory. Also
5518 generate an error if handed NULL (previously returned 0 to indicate an
5519 error, but didn't set one).
5520 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
5522 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
5525 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
5526 parameters. This was causing a warning which killed off the Win32 compile.
5529 *) Remove C++ style comments from crypto/bn/bn_local.h.
5530 [Neil Costigan <neil.costigan@celocom.com>]
5532 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
5533 based on a text string, looking up short and long names and finally
5534 "dot" format. The "dot" format stuff didn't work. Added new function
5535 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
5536 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
5537 OID is not part of the table.
5540 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
5541 X509_LOOKUP_by_alias().
5544 *) Sort openssl functions by name.
5547 *) Get the gendsa program working (hopefully) and add it to app list. Remove
5548 encryption from sample DSA keys (in case anyone is interested the password
5552 *) Make _all_ *_free functions accept a NULL pointer.
5553 [Frans Heymans <fheymans@isaserver.be>]
5555 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
5557 [Anonymous <nobody@replay.com>]
5559 *) s_server should send the CAfile as acceptable CAs, not its own cert.
5560 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5562 *) Don't blow it for numeric -newkey arguments to apps/req.
5563 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
5565 *) Temp key "for export" tests were wrong in s3_srvr.c.
5566 [Anonymous <nobody@replay.com>]
5568 *) Add prototype for temp key callback functions
5569 SSL_CTX_set_tmp_{rsa,dh}_callback().
5572 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
5573 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
5576 *) X509_name_add_entry() freed the wrong thing after an error.
5577 [Arne Ansper <arne@ats.cyber.ee>]
5579 *) rsa_eay.c would attempt to free a NULL context.
5580 [Arne Ansper <arne@ats.cyber.ee>]
5582 *) BIO_s_socket() had a broken should_retry() on Windoze.
5583 [Arne Ansper <arne@ats.cyber.ee>]
5585 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
5586 [Arne Ansper <arne@ats.cyber.ee>]
5588 *) Make sure the already existing X509_STORE->depth variable is initialized
5589 in X509_STORE_new(), but document the fact that this variable is still
5590 unused in the certificate verification process.
5591 [Ralf S. Engelschall]
5593 *) Fix the various library and apps files to free up pkeys obtained from
5594 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
5597 *) Fix reference counting in X509_PUBKEY_get(). This makes
5598 demos/maurice/example2.c work, amongst others, probably.
5599 [Steve Henson and Ben Laurie]
5601 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
5602 `openssl' and second, the shortcut symlinks for the `openssl <command>'
5603 are no longer created. This way we have a single and consistent command
5604 line interface `openssl <command>', similar to `cvs <command>'.
5605 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
5607 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
5608 BIT STRING wrapper always have zero unused bits.
5611 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
5614 *) Make the top-level INSTALL documentation easier to understand.
5617 *) Makefiles updated to exit if an error occurs in a sub-directory
5618 make (including if user presses ^C) [Paul Sutton]
5620 *) Make Montgomery context stuff explicit in RSA data structure.
5623 *) Fix build order of pem and err to allow for generated pem.h.
5626 *) Fix renumbering bug in X509_NAME_delete_entry().
5629 *) Enhanced the err-ins.pl script so it makes the error library number
5630 global and can add a library name. This is needed for external ASN1 and
5631 other error libraries.
5634 *) Fixed sk_insert which never worked properly.
5637 *) Fix ASN1 macros so they can handle indefinite length construted
5638 EXPLICIT tags. Some non standard certificates use these: they can now
5642 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
5643 into a single doc/ssleay.txt bundle. This way the information is still
5644 preserved but no longer messes up this directory. Now it's new room for
5645 the new set of documenation files.
5646 [Ralf S. Engelschall]
5648 *) SETs were incorrectly DER encoded. This was a major pain, because they
5649 shared code with SEQUENCEs, which aren't coded the same. This means that
5650 almost everything to do with SETs or SEQUENCEs has either changed name or
5651 number of arguments.
5652 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
5654 *) Fix test data to work with the above.
5657 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
5658 was already fixed by Eric for 0.9.1 it seems.
5659 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
5661 *) Autodetect FreeBSD3.
5664 *) Fix various bugs in Configure. This affects the following platforms:
5668 unixware-2.0-pentium
5672 *) Eliminate generated files from CVS. Reorder tests to regenerate files
5673 before they are needed.
5676 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
5680 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
5682 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
5683 changed SSLeay to OpenSSL in version strings.
5684 [Ralf S. Engelschall]
5686 *) Some fixups to the top-level documents.
5689 *) Fixed the nasty bug where rsaref.h was not found under compile-time
5690 because the symlink to include/ was missing.
5691 [Ralf S. Engelschall]
5693 *) Incorporated the popular no-RSA/DSA-only patches
5694 which allow to compile a RSA-free SSLeay.
5695 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
5697 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
5698 when "ssleay" is still not found.
5699 [Ralf S. Engelschall]
5701 *) Added more platforms to Configure: Cray T3E, HPUX 11,
5702 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
5704 *) Updated the README file.
5705 [Ralf S. Engelschall]
5707 *) Added various .cvsignore files in the CVS repository subdirs
5708 to make a "cvs update" really silent.
5709 [Ralf S. Engelschall]
5711 *) Recompiled the error-definition header files and added
5712 missing symbols to the Win32 linker tables.
5713 [Ralf S. Engelschall]
5715 *) Cleaned up the top-level documents;
5716 o new files: CHANGES and LICENSE
5717 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
5718 o merged COPYRIGHT into LICENSE
5719 o removed obsolete TODO file
5720 o renamed MICROSOFT to INSTALL.W32
5721 [Ralf S. Engelschall]
5723 *) Removed dummy files from the 0.9.1b source tree:
5724 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
5725 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
5726 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
5727 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
5728 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
5729 [Ralf S. Engelschall]
5731 *) Added various platform portability fixes.
5734 *) The Genesis of the OpenSSL rpject:
5735 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
5736 Young and Tim J. Hudson created while they were working for C2Net until
5738 [The OpenSSL Project]
5741 Changes between 0.9.0b and 0.9.1b [not released]
5743 *) Updated a few CA certificates under certs/
5746 *) Changed some BIGNUM api stuff.
5749 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
5750 DGUX x86, Linux Alpha, etc.
5753 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
5754 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
5758 *) Add -strparse option to asn1pars program which parses nested
5760 [Dr Stephen Henson <shenson@bigfoot.com>]
5762 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
5765 *) DSA fix for "ca" program.
5768 *) Added "-genkey" option to "dsaparam" program.
5771 *) Added RIPE MD160 (rmd160) message digest.
5774 *) Added -a (all) option to "ssleay version" command.
5777 *) Added PLATFORM define which is the id given to Configure.
5780 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
5783 *) Extended the ASN.1 parser routines.
5786 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
5789 *) Added a BN_CTX to the BN library.
5792 *) Fixed the weak key values in DES library
5795 *) Changed API in EVP library for cipher aliases.
5798 *) Added support for RC2/64bit cipher.
5801 *) Converted the lhash library to the crypto/mem.c functions.
5804 *) Added more recognized ASN.1 object ids.
5807 *) Added more RSA padding checks for SSL/TLS.
5810 *) Added BIO proxy/filter functionality.
5813 *) Added extra_certs to SSL_CTX which can be used
5814 send extra CA certificates to the client in the CA cert chain sending
5815 process. It can be configured with SSL_CTX_add_extra_chain_cert().
5818 *) Now Fortezza is denied in the authentication phase because
5819 this is key exchange mechanism is not supported by SSLeay at all.
5822 *) Additional PKCS1 checks.
5825 *) Support the string "TLSv1" for all TLS v1 ciphers.
5828 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
5829 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
5832 *) Fixed a few memory leaks.
5835 *) Fixed various code and comment typos.
5838 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
5839 bytes sent in the client random.
5840 [Edward Bishop <ebishop@spyglass.com>]