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 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
17 check whether we deal with a copy of a session and do not delete from
18 the cache in this case. Problem reported by "Izhar Shoshani Levi"
19 <izhar@checkpoint.com>.
22 *) Do not store session data into the internal session cache, if it
23 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
24 flag is set). Proposed by Aslam <aslam@funk.com>.
27 +) Support for crypto accelerator cards from Accelerated Encryption
28 Processing, www.aep.ie. (Use engine 'aep')
29 The support was copied from 0.9.6c [engine] and adapted/corrected
30 to work with the new engine framework.
31 [AEP Inc. and Richard Levitte]
33 +) Support for SureWare crypto accelerator cards from Baltimore
34 Technologies. (Use engine 'sureware')
35 The support was copied from 0.9.6c [engine] and adapted
36 to work with the new engine framework.
39 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
43 +) Have the CHIL engine fork-safe (as defined by nCipher) and actually
44 make the newer ENGINE framework commands for the CHIL engine work.
45 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
47 +) Make it possible to produce shared libraries on ReliantUNIX.
48 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
50 *) Add the configuration target linux-s390x.
51 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
53 +) Add the configuration target debug-linux-ppro.
54 Make 'openssl rsa' use the general key loading routines
55 implemented in apps.c, and make those routines able to
56 handle the key format FORMAT_NETSCAPE and the variant
58 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
60 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
61 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
63 +) Add -keyform to rsautl, and document -engine.
64 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
66 +) Change BIO_new_file (crypto/bio/bss_file.c) to use new
67 BIO_R_NO_SUCH_FILE error code rather than the generic
68 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
73 ERR_peek_last_error_line
74 ERR_peek_last_error_line_data.
78 ERR_peek_error_line_data,
79 but report on the latest error recorded rather than the first one
80 still in the error queue.
81 [Ben Laurie, Bodo Moeller]
83 +) default_algorithms option in ENGINE config module. This allows things
85 default_algorithms = ALL
86 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
89 +) Prelminary ENGINE config module.
92 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
93 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
94 variable as an indication that a ClientHello message has been
95 received. As the flag value will be lost between multiple
96 invocations of ssl3_accept when using non-blocking I/O, the
97 function may not be aware that a handshake has actually taken
98 place, thus preventing a new session from being added to the
101 To avoid this problem, we now set s->new_session to 2 instead of
102 using a local variable.
103 [Lutz Jaenicke, Bodo Moeller]
105 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
106 if the SSL_R_LENGTH_MISMATCH error is detected.
107 [Geoff Thorpe, Bodo Moeller]
109 +) New experimental application configuration code.
112 *) New 'shared_ldflag' column in Configure platform table.
115 *) Fix EVP_CIPHER_mode macro.
116 ["Dan S. Camper" <dan@bti.net>]
118 +) Change the AES code to follow the same name structure as all other
119 symmetric ciphers, and behave the same way. Move everything to
120 the directory crypto/aes, thereby obsoleting crypto/rijndael.
121 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
123 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
124 type, we must throw them away by setting rr->length to 0.
125 [D P Chang <dpc@qualys.com>]
127 -) OpenSSL 0.9.6c released [21 dec 2001]
129 +) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
130 [Ben Laurie and Theo de Raadt]
132 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
133 <Dominikus.Scherkl@biodata.com>. (The previous implementation
134 worked incorrectly for those cases where range = 10..._2 and
135 3*range is two bits longer than range.)
138 *) Only add signing time to PKCS7 structures if it is not already
142 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
143 OBJ_ld_ce should be OBJ_id_ce.
144 Also some ip-pda OIDs in crypto/objects/objects.txt were
145 incorrect (cf. RFC 3039).
146 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
148 +) Add option to output public keys in req command.
149 [Massimiliano Pala madwolf@openca.org]
151 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
152 returns early because it has nothing to do.
153 [Andy Schneider <andy.schneider@bjss.co.uk>]
155 *) [In 0.9.6c-engine and 0.9.7 release:]
156 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
157 [Andy Schneider <andy.schneider@bjss.co.uk>]
159 -) [In 0.9.6c-engine release:]
160 Add support for Cryptographic Appliance's keyserver technology.
161 (Use engine 'keyclient')
162 [Cryptographic Appliances and Geoff Thorpe]
164 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
165 is called via tools/c89.sh because arguments have to be
166 rearranged (all '-L' options must appear before the first object
168 [Richard Shapiro <rshapiro@abinitio.com>]
170 +) Use wNAFs in EC_POINTs_mul() for improved efficiency
171 (up to about 10% better than before for P-192 and P-224).
174 -) [In 0.9.6c-engine release:]
175 Add support for Broadcom crypto accelerator cards, backported
177 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
179 -) [In 0.9.6c-engine release:]
180 Add support for SureWare crypto accelerator cards from
181 Baltimore Technologies. (Use engine 'sureware')
182 [Baltimore Technologies and Mark Cox]
184 -) [In 0.9.6c-engine release:]
185 Add support for crypto accelerator cards from Accelerated
186 Encryption Processing, www.aep.ie. (Use engine 'aep')
187 [AEP Inc. and Mark Cox]
189 *) Add a configuration entry for gcc on UnixWare.
190 [Gary Benson <gbenson@redhat.com>]
192 +) New functions/macros
194 SSL_CTX_set_msg_callback(ctx, cb)
195 SSL_CTX_set_msg_callback_arg(ctx, arg)
196 SSL_set_msg_callback(ssl, cb)
197 SSL_set_msg_callback_arg(ssl, arg)
199 to request calling a callback function
201 void cb(int write_p, int version, int content_type,
202 const void *buf, size_t len, SSL *ssl, void *arg)
204 whenever a protocol message has been completely received
205 (write_p == 0) or sent (write_p == 1). Here 'version' is the
206 protocol version according to which the SSL library interprets
207 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
208 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
209 the content type as defined in the SSL 3.0/TLS 1.0 protocol
210 specification (change_cipher_spec(20), alert(21), handshake(22)).
211 'buf' and 'len' point to the actual message, 'ssl' to the
212 SSL object, and 'arg' is the application-defined value set by
213 SSL[_CTX]_set_msg_callback_arg().
215 'openssl s_client' and 'openssl s_server' have new '-msg' options
216 to enable a callback that displays all protocol messages.
219 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
220 messages are stored in a single piece (fixed-length part and
221 variable-length part combined) and fix various bugs found on the way.
224 +) Change the shared library support so shared libraries are built as
225 soon as the corresponding static library is finished, and thereby get
226 openssl and the test programs linked against the shared library.
227 This still only happens when the keyword "shard" has been given to
228 the configuration scripts.
230 NOTE: shared library support is still an experimental thing, and
231 backward binary compatibility is still not guaranteed.
232 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
234 +) Add support for Subject Information Access extension.
235 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
237 +) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
238 additional bytes when new memory had to be allocated, not just
239 when reusing an existing buffer.
242 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
243 instead. BIO_gethostbyname() does not know what timeouts are
244 appropriate, so entries would stay in cache even when they have
246 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
248 +) New command line and configuration option 'utf8' for the req command.
249 This allows field values to be specified as UTF8 strings.
252 +) Add -multi and -mr options to "openssl speed" - giving multiple parallel
253 runs for the former and machine-readable output for the latter.
256 +) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
257 of the e-mail address in the DN (i.e., it will go into a certificate
258 extension only). The new configuration file option 'email_in_dn = no'
260 [Massimiliano Pala madwolf@openca.org]
262 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
263 faced with a pathologically small ClientHello fragment that does
264 not contain client_version: Instead of aborting with an error,
265 simply choose the highest available protocol version (i.e.,
266 TLS 1.0 unless it is disabled). In practice, ClientHello
267 messages are never sent like this, but this change gives us
268 strictly correct behaviour at least for TLS.
271 +) Change all functions with names starting with des_ to be starting
272 with DES_ instead. Add wrappers that are compatible with libdes,
273 but are named _ossl_old_des_*. Finally, add macros that map the
274 des_* symbols to the corresponding _ossl_old_des_*.
276 All this is done because there are increasing clashes with libdes
277 and other DES libraries that are currently used by other projects.
278 The old libdes interface (including crypt()) is provided if
279 <openssl/des_old.h> is included. For now, this automatically
280 happens in <openssl/des.h> unless OPENSSL_DISABLE_OLD_DES_SUPPORT is
281 defined. Note that crypt() is no longer declared in <openssl/des.h>.
283 NOTE: This is a major break of an old API into a new one. Software
284 authors are encouraged to switch to the DES_ style functions. Some
285 time in the future, des_old.h and the libdes compatibility functions
286 will be completely removed.
289 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
290 never resets s->method to s->ctx->method when called from within
291 one of the SSL handshake functions.
292 [Bodo Moeller; problem pointed out by Niko Baric]
294 +) Test for certificates which contain unsupported critical extensions.
295 If such a certificate is found during a verify operation it is
296 rejected by default: this behaviour can be overridden by either
297 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
298 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
299 X509_supported_extension() has also been added which returns 1 if a
300 particular extension is supported.
303 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
304 (sent using the client's version number) if client_version is
305 smaller than the protocol version in use. Also change
306 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
307 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
308 the client will at least see that alert.
311 +) Modify the behaviour of EVP cipher functions in similar way to digests
312 to retain compatibility with existing code.
315 +) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
316 compatibility with existing code. In particular the 'ctx' parameter does
317 not have to be to be initialized before the call to EVP_DigestInit() and
318 it is tidied up after a call to EVP_DigestFinal(). New function
319 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
320 EVP_MD_CTX_copy() changed to not require the destination to be
321 initialized valid and new function EVP_MD_CTX_copy_ex() added which
322 requires the destination to be valid.
324 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
325 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
328 +) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
329 so that complete 'Handshake' protocol structures are kept in memory
330 instead of overwriting 'msg_type' and 'length' with 'body' data.
333 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
337 +) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
338 [Massimo Santin via Richard Levitte]
340 +) Major restructuring to the underlying ENGINE code. This includes
341 reduction of linker bloat, separation of pure "ENGINE" manipulation
342 (initialisation, etc) from functionality dealing with implementations
343 of specific crypto iterfaces. This change also introduces integrated
344 support for symmetric ciphers and digest implementations - so ENGINEs
345 can now accelerate these by providing EVP_CIPHER and EVP_MD
346 implementations of their own. This is detailed in crypto/engine/README
347 as it couldn't be adequately described here. However, there are a few
348 API changes worth noting - some RSA, DSA, DH, and RAND functions that
349 were changed in the original introduction of ENGINE code have now
350 reverted back - the hooking from this code to ENGINE is now a good
351 deal more passive and at run-time, operations deal directly with
352 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
353 dereferencing through an ENGINE pointer any more. Also, the ENGINE
354 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
355 they were not being used by the framework as there is no concept of a
356 BIGNUM_METHOD and they could not be generalised to the new
357 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
358 ENGINE_cpy() has been removed as it cannot be consistently defined in
362 +) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
365 +) Change mkdef.pl to sort symbols that get the same entry number,
366 and make sure the automatically generated functions ERR_load_*
367 become part of libeay.num as well.
370 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
371 client receives HelloRequest while in a handshake.
372 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
374 +) New function SSL_renegotiate_pending(). This returns true once
375 renegotiation has been requested (either SSL_renegotiate() call
376 or HelloRequest/ClientHello receveived from the peer) and becomes
377 false once a handshake has been completed.
378 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
379 sends a HelloRequest, but does not ensure that a handshake takes
380 place. SSL_renegotiate_pending() is useful for checking if the
381 client has followed the request.)
384 +) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
385 By default, clients may request session resumption even during
386 renegotiation (if session ID contexts permit); with this option,
387 session resumption is possible only in the first handshake.
390 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
391 should end in 'break', not 'goto end' which circuments various
392 cleanups done in state SSL_ST_OK. But session related stuff
393 must be disabled for SSL_ST_OK in the case that we just sent a
396 Also avoid some overhead by not calling ssl_init_wbio_buffer()
397 before just sending a HelloRequest.
398 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
400 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
401 reveal whether illegal block cipher padding was found or a MAC
402 verification error occured. (Neither SSLerr() codes nor alerts
403 are directly visible to potential attackers, but the information
404 may leak via logfiles.)
406 Similar changes are not required for the SSL 2.0 implementation
407 because the number of padding bytes is sent in clear for SSL 2.0,
408 and the extra bytes are just ignored. However ssl/s2_pkt.c
409 failed to verify that the purported number of padding bytes is in
413 +) Add some demos for certificate and certificate request creation.
416 +) Make maximum certificate chain size accepted from the peer application
417 settable (SSL*_get/set_max_cert_list()), as proposed by
418 "Douglas E. Engert" <deengert@anl.gov>.
421 +) Add support for shared libraries for Unixware-7
422 (Boyd Lynn Gerber <gerberb@zenez.com>).
425 *) Add OpenUNIX-8 support including shared libraries
426 (Boyd Lynn Gerber <gerberb@zenez.com>).
429 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
430 'wristwatch attack' using huge encoding parameters (cf.
431 James H. Manger's CRYPTO 2001 paper). Note that the
432 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
433 encoding parameters and hence was not vulnerable.
436 +) Add a "destroy" handler to ENGINEs that allows structural cleanup to
437 be done prior to destruction. Use this to unload error strings from
438 ENGINEs that load their own error strings. NB: This adds two new API
439 functions to "get" and "set" this destroy handler in an ENGINE.
442 +) Alter all existing ENGINE implementations (except "openssl" and
443 "openbsd") to dynamically instantiate their own error strings. This
444 makes them more flexible to be built both as statically-linked ENGINEs
445 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
446 Also, add stub code to each that makes building them as self-contained
447 shared-libraries easier (see README.ENGINE).
450 +) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
451 implementations into applications that are completely implemented in
452 self-contained shared-libraries. The "dynamic" ENGINE exposes control
453 commands that can be used to configure what shared-library to load and
454 to control aspects of the way it is handled. Also, made an update to
455 the README.ENGINE file that brings its information up-to-date and
456 provides some information and instructions on the "dynamic" ENGINE
457 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
461 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
463 +) Make it possible to unload ranges of ERR strings with a new
464 "ERR_unload_strings" function.
467 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
468 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
469 followed by modular reduction.
470 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
472 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
473 equivalent based on BN_pseudo_rand() instead of BN_rand().
476 +) Add a copy() function to EVP_MD.
479 +) Make EVP_MD routines take a context pointer instead of just the
480 md_data void pointer.
483 +) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
484 that the digest can only process a single chunk of data
485 (typically because it is provided by a piece of
486 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
487 is only going to provide a single chunk of data, and hence the
488 framework needn't accumulate the data for oneshot drivers.
491 +) As with "ERR", make it possible to replace the underlying "ex_data"
492 functions. This change also alters the storage and management of global
493 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
494 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
495 index counters. The API functions that use this state have been changed
496 to take a "class_index" rather than pointers to the class's local STACK
497 and counter, and there is now an API function to dynamically create new
498 classes. This centralisation allows us to (a) plug a lot of the
499 thread-safety problems that existed, and (b) makes it possible to clean
500 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
501 such data would previously have always leaked in application code and
502 workarounds were in place to make the memory debugging turn a blind eye
503 to it. Application code that doesn't use this new function will still
504 leak as before, but their memory debugging output will announce it now
505 rather than letting it slide.
507 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
508 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
509 has a return value to indicate success or failure.
512 +) Make it possible to replace the underlying "ERR" functions such that the
513 global state (2 LHASH tables and 2 locks) is only used by the "default"
514 implementation. This change also adds two functions to "get" and "set"
515 the implementation prior to it being automatically set the first time
516 any other ERR function takes place. Ie. an application can call "get",
517 pass the return value to a module it has just loaded, and that module
518 can call its own "set" function using that value. This means the
519 module's "ERR" operations will use (and modify) the error state in the
520 application and not in its own statically linked copy of OpenSSL code.
523 +) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
524 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
525 the operation, and provides a more encapsulated way for external code
526 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
527 to use these functions rather than manually incrementing the counts.
529 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
532 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
533 This function was broken, as the check for a new client hello message
534 to handle SGC did not allow these large messages.
535 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
538 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
541 +) Add EVP test program.
544 +) Add symmetric cipher support to ENGINE. Expect the API to change!
547 +) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
548 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
549 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
550 These allow a CRL to be built without having to access X509_CRL fields
551 directly. Modify 'ca' application to use new functions.
554 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
555 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
558 *) Rework the configuration and shared library support for Tru64 Unix.
559 The configuration part makes use of modern compiler features and
560 still retains old compiler behavior for those that run older versions
561 of the OS. The shared library support part includes a variant that
562 uses the RPATH feature, and is available through the special
563 configuration target "alpha-cc-rpath", which will never be selected
565 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
567 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
568 with the same message size as in ssl3_get_certificate_request().
569 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
570 messages might inadvertently be reject as too long.
571 [Petr Lampa <lampa@fee.vutbr.cz>]
573 +) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
574 bug workarounds. Rollback attack detection is a security feature.
575 The problem will only arise on OpenSSL servers when TLSv1 is not
576 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
577 Software authors not wanting to support TLSv1 will have special reasons
578 for their choice and can explicitly enable this option.
579 [Bodo Moeller, Lutz Jaenicke]
581 +) Rationalise EVP so it can be extended: don't include a union of
582 cipher/digest structures, add init/cleanup functions. This also reduces
583 the number of header dependencies.
586 +) Make DES key schedule conform to the usual scheme, as well as
587 correcting its structure. This means that calls to DES functions
588 now have to pass a pointer to a des_key_schedule instead of a
589 plain des_key_schedule (which was actually always a pointer
593 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
596 *) Modified SSL library such that the verify_callback that has been set
597 specificly for an SSL object with SSL_set_verify() is actually being
598 used. Before the change, a verify_callback set with this function was
599 ignored and the verify_callback() set in the SSL_CTX at the time of
600 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
601 to allow the necessary settings.
604 +) Initial reduction of linker bloat: the use of some functions, such as
605 PEM causes large amounts of unused functions to be linked in due to
606 poor organisation. For example pem_all.c contains every PEM function
607 which has a knock on effect of linking in large amounts of (unused)
608 ASN1 code. Grouping together similar functions and splitting unrelated
609 functions prevents this.
612 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
613 explicitly to NULL, as at least on Solaris 8 this seems not always to be
614 done automatically (in contradiction to the requirements of the C
615 standard). This made problems when used from OpenSSH.
618 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
619 dh->length and always used
621 BN_rand_range(priv_key, dh->p).
623 BN_rand_range() is not necessary for Diffie-Hellman, and this
624 specific range makes Diffie-Hellman unnecessarily inefficient if
625 dh->length (recommended exponent length) is much smaller than the
626 length of dh->p. We could use BN_rand_range() if the order of
627 the subgroup was stored in the DH structure, but we only have
632 BN_rand(priv_key, l, ...)
634 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
640 RSA_eay_public_encrypt
641 RSA_eay_private_decrypt
642 RSA_eay_private_encrypt (signing)
643 RSA_eay_public_decrypt (signature verification)
645 (default implementations for RSA_public_encrypt,
646 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
647 always reject numbers >= n.
650 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
651 to synchronize access to 'locking_thread'. This is necessary on
652 systems where access to 'locking_thread' (an 'unsigned long'
653 variable) is not atomic.
656 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
657 *before* setting the 'crypto_lock_rand' flag. The previous code had
658 a race condition if 0 is a valid thread ID.
659 [Travis Vitek <vitek@roguewave.com>]
661 +) Cleanup of EVP macros.
664 +) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
668 +) Add initial OCSP responder support to ocsp application. The
669 revocation information is handled using the text based index
670 use by the ca application. The responder can either handle
671 requests generated internally, supplied in files (for example
672 via a CGI script) or using an internal minimal server.
675 +) Add configuration choices to get zlib compression for TLS.
678 +) Changes to Kerberos SSL for RFC 2712 compliance:
679 1. Implemented real KerberosWrapper, instead of just using
680 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
681 2. Implemented optional authenticator field of KerberosWrapper.
683 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
684 and authenticator structs; see crypto/krb5/.
686 Generalized Kerberos calls to support multiple Kerberos libraries.
687 [Vern Staats <staatsvr@asc.hpc.mil>,
688 Jeffrey Altman <jaltman@columbia.edu>
691 +) Cause 'openssl speed' to use fully hard-coded DSA keys as it
692 already does with RSA. testdsa.h now has 'priv_key/pub_key'
693 values for each of the key sizes rather than having just
694 parameters (and 'speed' generating keys each time).
697 *) Add support for shared libraries under Irix.
698 [Albert Chin-A-Young <china@thewrittenword.com>]
700 *) Add configuration option to build on Linux on both big-endian and
702 [Ralf Baechle <ralf@uni-koblenz.de>]
704 *) Add the possibility to create shared libraries on HP-UX.
707 -) OpenSSL 0.9.6b released [9 July 2001]
709 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
710 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
711 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
712 PRNG state recovery was possible based on the output of
713 one PRNG request appropriately sized to gain knowledge on
714 'md' followed by enough consecutive 1-byte PRNG requests
715 to traverse all of 'state'.
717 1. When updating 'md_local' (the current thread's copy of 'md')
718 during PRNG output generation, hash all of the previous
719 'md_local' value, not just the half used for PRNG output.
721 2. Make the number of bytes from 'state' included into the hash
722 independent from the number of PRNG bytes requested.
724 The first measure alone would be sufficient to avoid
725 Markku-Juhani's attack. (Actually it had never occurred
726 to me that the half of 'md_local' used for chaining was the
727 half from which PRNG output bytes were taken -- I had always
728 assumed that the secret half would be used.) The second
729 measure makes sure that additional data from 'state' is never
730 mixed into 'md_local' in small portions; this heuristically
731 further strengthens the PRNG.
734 +) Speed up EVP routines.
737 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
738 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
739 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
740 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
742 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
743 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
744 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
747 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
749 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
752 *) Fix crypto/bn/asm/mips3.s.
755 *) When only the key is given to "enc", the IV is undefined. Print out
756 an error message in this case.
759 +) Added the OS2-EMX target.
760 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
762 +) Rewrite apps to use NCONF routines instead of the old CONF. New functions
763 to support NCONF routines in extension code. New function CONF_set_nconf()
764 to allow functions which take an NCONF to also handle the old LHASH
765 structure: this means that the old CONF compatible routines can be
766 retained (in particular wrt extensions) without having to duplicate the
767 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
770 *) Handle special case when X509_NAME is empty in X509 printing routines.
773 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
774 positive and less than q.
777 +) Enhance the general user interface with mechanisms for inner control
778 and with possibilities to have yes/no kind of prompts.
781 +) Change all calls to low level digest routines in the library and
782 applications to use EVP. Add missing calls to HMAC_cleanup() and
783 don't assume HMAC_CTX can be copied using memcpy().
784 [Verdon Walker <VWalker@novell.com>, Steve Henson]
786 +) Add the possibility to control engines through control names but with
787 arbitrary arguments instead of just a string.
788 Change the key loaders to take a UI_METHOD instead of a callback
789 function pointer. NOTE: this breaks binary compatibility with earlier
790 versions of OpenSSL [engine].
791 Adapt the nCipher code for these new conditions and add a card insertion
795 +) Enhance the general user interface with mechanisms to better support
796 dialog box interfaces, application-defined prompts, the possibility
797 to use defaults (for example default passwords from somewhere else)
798 and interrupts/cancellations.
801 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
802 used: it isn't thread safe and the add_lock_callback should handle
804 [Paul Rose <Paul.Rose@bridge.com>]
806 *) Verify that incoming data obeys the block size in
807 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
810 +) Tidy up PKCS#12 attribute handling. Add support for the CSP name
811 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
815 [Ulf Möller, Bodo Möller]
817 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
818 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
819 when fixing the server behaviour for backwards-compatible 'client
820 hello' messages. (Note that the attack is impractical against
821 SSL 3.0 and TLS 1.0 anyway because length and version checking
822 means that the probability of guessing a valid ciphertext is
823 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
826 Before 0.9.5, the countermeasure (hide the error by generating a
827 random 'decryption result') did not work properly because
828 ERR_clear_error() was missing, meaning that SSL_get_error() would
829 detect the supposedly ignored error.
831 Both problems are now fixed.
834 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
835 (previously it was 1024).
838 +) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
839 tidy up some unnecessarily weird code in 'sk_new()').
840 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
842 +) Change the key loading routines for ENGINEs to use the same kind
843 callback (pem_password_cb) as all other routines that need this
847 *) Fix for compatibility mode trust settings: ignore trust settings
848 unless some valid trust or reject settings are present.
851 *) Fix for blowfish EVP: its a variable length cipher.
854 +) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
855 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
856 than this minimum value is recommended.
859 +) New random seeder for OpenVMS, using the system process statistics
860 that are easily reachable.
863 +) Windows apparently can't transparently handle global
864 variables defined in DLLs. Initialisations such as:
866 const ASN1_ITEM *it = &ASN1_INTEGER_it;
868 wont compile. This is used by the any applications that need to
869 declare their own ASN1 modules. This was fixed by adding the option
870 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
871 needed for static libraries under Win32.
874 +) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
875 setting of purpose and trust fields. New X509_STORE trust and
876 purpose functions and tidy up setting in other SSL functions.
879 +) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
880 structure. These are inherited by X509_STORE_CTX when it is
881 initialised. This allows various defaults to be set in the
882 X509_STORE structure (such as flags for CRL checking and custom
883 purpose or trust settings) for functions which only use X509_STORE_CTX
884 internally such as S/MIME.
886 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
887 trust settings if they are not set in X509_STORE. This allows X509_STORE
888 purposes and trust (in S/MIME for example) to override any set by default.
890 Add command line options for CRL checking to smime, s_client and s_server
894 +) Initial CRL based revocation checking. If the CRL checking flag(s)
895 are set then the CRL is looked up in the X509_STORE structure and
896 its validity and signature checked, then if the certificate is found
897 in the CRL the verify fails with a revoked error.
899 Various new CRL related callbacks added to X509_STORE_CTX structure.
901 Command line options added to 'verify' application to support this.
903 This needs some additional work, such as being able to handle multiple
904 CRLs with different times, extension based lookup (rather than just
905 by subject name) and ultimately more complete V2 CRL extension
909 +) Add a general user interface API (crypto/ui/). This is designed
910 to replace things like des_read_password and friends (backward
911 compatibility functions using this new API are provided).
912 The purpose is to remove prompting functions from the DES code
913 section as well as provide for prompting through dialog boxes in
914 a window system and the like.
917 *) Fix various bugs related to DSA S/MIME verification. Handle missing
918 parameters in DSA public key structures and return an error in the
919 DSA routines if parameters are absent.
922 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
923 in the current directory if neither $RANDFILE nor $HOME was set.
924 RAND_file_name() in 0.9.6a returned NULL in this case. This has
925 caused some confusion to Windows users who haven't defined $HOME.
926 Thus RAND_file_name() is changed again: e_os.h can define a
927 DEFAULT_HOME, which will be used if $HOME is not set.
928 For Windows, we use "C:"; on other platforms, we still require
929 environment variables.
931 +) Add "ex_data" support to ENGINE so implementations can add state at a
932 per-structure level rather than having to store it globally.
935 +) Make it possible for ENGINE structures to be copied when retrieved by
936 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
937 This causes the "original" ENGINE structure to act like a template,
938 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
939 operational state can be localised to each ENGINE structure, despite the
940 fact they all share the same "methods". New ENGINE structures returned in
941 this case have no functional references and the return value is the single
942 structural reference. This matches the single structural reference returned
943 by ENGINE_by_id() normally, when it is incremented on the pre-existing
947 +) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
948 needs to match any other type at all we need to manually clear the
952 +) Changes to the "openssl engine" utility to include;
953 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
954 about an ENGINE's available control commands.
955 - executing control commands from command line arguments using the
956 '-pre' and '-post' switches. '-post' is only used if '-t' is
957 specified and the ENGINE is successfully initialised. The syntax for
958 the individual commands are colon-separated, for example;
959 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
962 +) New dynamic control command support for ENGINEs. ENGINEs can now
963 declare their own commands (numbers), names (strings), descriptions,
964 and input types for run-time discovery by calling applications. A
965 subset of these commands are implicitly classed as "executable"
966 depending on their input type, and only these can be invoked through
967 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
968 can be based on user input, config files, etc). The distinction is
969 that "executable" commands cannot return anything other than a boolean
970 result and can only support numeric or string input, whereas some
971 discoverable commands may only be for direct use through
972 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
973 pointers, or other custom uses. The "executable" commands are to
974 support parameterisations of ENGINE behaviour that can be
975 unambiguously defined by ENGINEs and used consistently across any
976 OpenSSL-based application. Commands have been added to all the
977 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
978 control over shared-library paths without source code alterations.
981 +) Changed all ENGINE implementations to dynamically allocate their
982 ENGINEs rather than declaring them statically. Apart from this being
983 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
984 this also allows the implementations to compile without using the
985 internal engine_int.h header.
988 +) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
989 'const' value. Any code that should be able to modify a RAND_METHOD
990 should already have non-const pointers to it (ie. they should only
991 modify their own ones).
994 +) Made a variety of little tweaks to the ENGINE code.
995 - "atalla" and "ubsec" string definitions were moved from header files
996 to C code. "nuron" string definitions were placed in variables
997 rather than hard-coded - allowing parameterisation of these values
998 later on via ctrl() commands.
999 - Removed unused "#if 0"'d code.
1000 - Fixed engine list iteration code so it uses ENGINE_free() to release
1001 structural references.
1002 - Constified the RAND_METHOD element of ENGINE structures.
1003 - Constified various get/set functions as appropriate and added
1004 missing functions (including a catch-all ENGINE_cpy that duplicates
1005 all ENGINE values onto a new ENGINE except reference counts/state).
1006 - Removed NULL parameter checks in get/set functions. Setting a method
1007 or function to NULL is a way of cancelling out a previously set
1008 value. Passing a NULL ENGINE parameter is just plain stupid anyway
1009 and doesn't justify the extra error symbols and code.
1010 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
1011 flags from engine_int.h to engine.h.
1012 - Changed prototypes for ENGINE handler functions (init(), finish(),
1013 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
1016 *) Move 'if (!initialized) RAND_poll()' into regions protected by
1017 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
1018 having multiple threads call RAND_poll() concurrently.
1021 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
1022 combination of a flag and a thread ID variable.
1023 Otherwise while one thread is in ssleay_rand_bytes (which sets the
1024 flag), *other* threads can enter ssleay_add_bytes without obeying
1025 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
1026 that they do not hold after the first thread unsets add_do_not_lock).
1029 +) Implement binary inversion algorithm for BN_mod_inverse in addition
1030 to the algorithm using long division. The binary algorithm can be
1031 used only if the modulus is odd. On 32-bit systems, it is faster
1032 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
1033 roughly 5-15% for 256-bit moduli), so we use it only for moduli
1034 up to 450 bits. In 64-bit environments, the binary algorithm
1035 appears to be advantageous for much longer moduli; here we use it
1036 for moduli up to 2048 bits.
1039 *) Change bctest again: '-x' expressions are not available in all
1043 -) OpenSSL 0.9.6a released [5 Apr 2001]
1045 *) Fix a couple of memory leaks in PKCS7_dataDecode()
1046 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
1048 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
1049 the default extension for executables, if any. Also, make the perl
1050 scripts that use symlink() to test if it really exists and use "cp"
1051 if it doesn't. All this made OpenSSL compilable and installable in
1055 +) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
1056 could not support the combine flag in choice fields.
1059 -) Fix for asn1_GetSequence() for indefinite length constructed data.
1060 If SEQUENCE is length is indefinite just set c->slen to the total
1061 amount of data available.
1062 [Steve Henson, reported by shige@FreeBSD.org]
1063 [This change does not apply to 0.9.7.]
1065 *) Change bctest to avoid here-documents inside command substitution
1066 (workaround for FreeBSD /bin/sh bug).
1067 For compatibility with Ultrix, avoid shell functions (introduced
1068 in the bctest version that searches along $PATH).
1071 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
1072 with des_encrypt() defined on some operating systems, like Solaris
1076 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
1077 On the Importance of Eliminating Errors in Cryptographic
1078 Computations, J. Cryptology 14 (2001) 2, 101-119,
1079 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
1082 *) MIPS assembler BIGNUM division bug fix.
1085 *) Disabled incorrect Alpha assembler code.
1088 -) Fix PKCS#7 decode routines so they correctly update the length
1089 after reading an EOC for the EXPLICIT tag.
1091 [This change does not apply to 0.9.7.]
1093 *) Fix bug in PKCS#12 key generation routines. This was triggered
1094 if a 3DES key was generated with a 0 initial byte. Include
1095 PKCS12_BROKEN_KEYGEN compilation option to retain the old
1096 (but broken) behaviour.
1099 *) Enhance bctest to search for a working bc along $PATH and print
1101 [Tim Rice <tim@multitalents.net> via Richard Levitte]
1103 +) Add a 'copy_extensions' option to the 'ca' utility. This copies
1104 extensions from a certificate request to the certificate.
1107 +) Allow multiple 'certopt' and 'nameopt' options to be separated
1108 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
1109 file: this allows the display of the certificate about to be
1110 signed to be customised, to allow certain fields to be included
1111 or excluded and extension details. The old system didn't display
1112 multicharacter strings properly, omitted fields not in the policy
1113 and couldn't display additional details such as extensions.
1116 *) Fix memory leaks in err.c: free err_data string if necessary;
1117 don't write to the wrong index in ERR_set_error_data.
1120 +) Function EC_POINTs_mul for multiple scalar multiplication
1121 of an arbitrary number of elliptic curve points
1122 \sum scalars[i]*points[i],
1123 optionally including the generator defined for the EC_GROUP:
1124 scalar*generator + \sum scalars[i]*points[i].
1126 EC_POINT_mul is a simple wrapper function for the typical case
1127 that the point list has just one item (besides the optional
1131 +) First EC_METHODs for curves over GF(p):
1133 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
1134 operations and provides various method functions that can also
1135 operate with faster implementations of modular arithmetic.
1137 EC_GFp_mont_method() reuses most functions that are part of
1138 EC_GFp_simple_method, but uses Montgomery arithmetic.
1140 [Bodo Moeller; point addition and point doubling
1141 implementation directly derived from source code provided by
1142 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
1144 +) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
1145 crypto/ec/ec_lib.c):
1147 Curves are EC_GROUP objects (with an optional group generator)
1148 based on EC_METHODs that are built into the library.
1150 Points are EC_POINT objects based on EC_GROUP objects.
1152 Most of the framework would be able to handle curves over arbitrary
1153 finite fields, but as there are no obvious types for fields other
1154 than GF(p), some functions are limited to that for now.
1157 +) Add the -HTTP option to s_server. It is similar to -WWW, but requires
1158 that the file contains a complete HTTP response.
1161 +) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
1162 change the def and num file printf format specifier from "%-40sXXX"
1163 to "%-39s XXX". The latter will always guarantee a space after the
1164 field while the former will cause them to run together if the field
1165 is 40 of more characters long.
1168 +) Constify the cipher and digest 'method' functions and structures
1169 and modify related functions to take constant EVP_MD and EVP_CIPHER
1173 *) Implement ssl23_peek (analogous to ssl23_read), which previously
1177 *) Replace rdtsc with _emit statements for VC++ version 5.
1178 [Jeremy Cooper <jeremy@baymoo.org>]
1180 +) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
1181 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
1184 +) Modify EVP_Digest*() routines so they now return values. Although the
1185 internal software routines can never fail additional hardware versions
1189 +) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
1191 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
1192 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
1195 ERR_R_NESTED_ASN1_ERROR
1197 ERR_R_MISSING_ASN1_EOS
1198 were 4 .. 9, conflicting with
1199 ERR_LIB_RSA (= ERR_R_RSA_LIB)
1201 ERR_LIB_PEM (= ERR_R_PEM_LIB).
1202 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
1204 Add new error code 'ERR_R_INTERNAL_ERROR'.
1207 +) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
1211 +) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
1212 sets the subject name for a new request or supersedes the
1213 subject name in a given request. Formats that can be parsed are
1214 'CN=Some Name, OU=myOU, C=IT'
1216 'CN=Some Name/OU=myOU/C=IT'.
1218 Add options '-batch' and '-verbose' to 'openssl req'.
1219 [Massimiliano Pala <madwolf@hackmasters.net>]
1221 *) Make it possible to reuse SSLv2 sessions.
1224 +) Introduce the possibility to access global variables through
1225 functions on platform were that's the best way to handle exporting
1226 global variables in shared libraries. To enable this functionality,
1227 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
1228 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
1229 is normally done by Configure or something similar).
1231 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
1232 in the source file (foo.c) like this:
1234 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
1235 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
1237 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
1238 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
1240 OPENSSL_DECLARE_GLOBAL(int,foo);
1241 #define foo OPENSSL_GLOBAL_REF(foo)
1242 OPENSSL_DECLARE_GLOBAL(double,bar);
1243 #define bar OPENSSL_GLOBAL_REF(bar)
1245 The #defines are very important, and therefore so is including the
1246 header file everywhere where the defined globals are used.
1248 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
1249 of ASN.1 items, but that structure is a bit different.
1251 The largest change is in util/mkdef.pl which has been enhanced with
1252 better and easier to understand logic to choose which symbols should
1253 go into the Windows .def files as well as a number of fixes and code
1254 cleanup (among others, algorithm keywords are now sorted
1255 lexicographically to avoid constant rewrites).
1258 *) In copy_email() check for >= 0 as a return value for
1259 X509_NAME_get_index_by_NID() since 0 is a valid index.
1260 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
1262 +) In BN_div() keep a copy of the sign of 'num' before writing the
1263 result to 'rm' because if rm==num the value will be overwritten
1264 and produce the wrong result if 'num' is negative: this caused
1265 problems with BN_mod() and BN_nnmod().
1268 +) Function OCSP_request_verify(). This checks the signature on an
1269 OCSP request and verifies the signer certificate. The signer
1270 certificate is just checked for a generic purpose and OCSP request
1274 +) Add OCSP_check_validity() function to check the validity of OCSP
1275 responses. OCSP responses are prepared in real time and may only
1276 be a few seconds old. Simply checking that the current time lies
1277 between thisUpdate and nextUpdate max reject otherwise valid responses
1278 caused by either OCSP responder or client clock inaccuracy. Instead
1279 we allow thisUpdate and nextUpdate to fall within a certain period of
1280 the current time. The age of the response can also optionally be
1281 checked. Two new options -validity_period and -status_age added to
1285 +) If signature or public key algorithm is unrecognized print out its
1286 OID rather that just UNKNOWN.
1289 *) Avoid coredump with unsupported or invalid public keys by checking if
1290 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
1291 PKCS7_verify() fails with non detached data.
1294 +) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
1295 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
1296 ID to be generated from the issuer certificate alone which can then be
1297 passed to OCSP_id_issuer_cmp().
1300 +) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
1301 ASN1 modules to export functions returning ASN1_ITEM pointers
1302 instead of the ASN1_ITEM structures themselves. This adds several
1303 new macros which allow the underlying ASN1 function/structure to
1304 be accessed transparently. As a result code should not use ASN1_ITEM
1305 references directly (such as &X509_it) but instead use the relevant
1306 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
1307 use of the new ASN1 code on platforms where exporting structures
1308 is problematical (for example in shared libraries) but exporting
1309 functions returning pointers to structures is not.
1312 +) Add support for overriding the generation of SSL/TLS session IDs.
1313 These callbacks can be registered either in an SSL_CTX or per SSL.
1314 The purpose of this is to allow applications to control, if they wish,
1315 the arbitrary values chosen for use as session IDs, particularly as it
1316 can be useful for session caching in multiple-server environments. A
1317 command-line switch for testing this (and any client code that wishes
1318 to use such a feature) has been added to "s_server".
1319 [Geoff Thorpe, Lutz Jaenicke]
1321 +) Modify mkdef.pl to recognise and parse preprocessor conditionals
1322 of the form '#if defined(...) || defined(...) || ...' and
1323 '#if !defined(...) && !defined(...) && ...'. This also avoids
1324 the growing number of special cases it was previously handling.
1327 *) Don't use getenv in library functions when run as setuid/setgid.
1328 New function OPENSSL_issetugid().
1331 +) Make all configuration macros available for application by making
1332 sure they are available in opensslconf.h, by giving them names starting
1333 with "OPENSSL_" to avoid conflicts with other packages and by making
1334 sure e_os2.h will cover all platform-specific cases together with
1336 Additionally, it is now possible to define configuration/platform-
1337 specific names (called "system identities"). In the C code, these
1338 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
1339 macro with the name beginning with "OPENSSL_SYS_", which is determined
1340 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
1344 +) New option -set_serial to 'req' and 'x509' this allows the serial
1345 number to use to be specified on the command line. Previously self
1346 signed certificates were hard coded with serial number 0 and the
1347 CA options of 'x509' had to use a serial number in a file which was
1351 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
1352 due to incorrect handling of multi-threading:
1354 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
1356 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
1358 3. Count how many times MemCheck_off() has been called so that
1359 nested use can be treated correctly. This also avoids
1360 inband-signalling in the previous code (which relied on the
1361 assumption that thread ID 0 is impossible).
1364 +) New options to 'ca' utility to support V2 CRL entry extensions.
1365 Currently CRL reason, invalidity date and hold instruction are
1366 supported. Add new CRL extensions to V3 code and some new objects.
1369 *) Add "-rand" option also to s_client and s_server.
1372 +) New function EVP_CIPHER_CTX_set_padding() this is used to
1373 disable standard block padding (aka PKCS#5 padding) in the EVP
1374 API, which was previously mandatory. This means that the data is
1375 not padded in any way and so the total length much be a multiple
1376 of the block size, otherwise an error occurs.
1379 +) Initial (incomplete) OCSP SSL support.
1382 *) Fix CPU detection on Irix 6.x.
1383 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
1384 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1386 +) New function OCSP_parse_url(). This splits up a URL into its host,
1387 port and path components: primarily to parse OCSP URLs. New -url
1388 option to ocsp utility.
1391 +) New nonce behavior. The return value of OCSP_check_nonce() now
1392 reflects the various checks performed. Applications can decide
1393 whether to tolerate certain situations such as an absent nonce
1394 in a response when one was present in a request: the ocsp application
1395 just prints out a warning. New function OCSP_add1_basic_nonce()
1396 this is to allow responders to include a nonce in a response even if
1397 the request is nonce-less.
1400 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
1403 [This change does not apply to 0.9.7.]
1405 *) Use the cached encoding of an X509_NAME structure rather than
1406 copying it. This is apparently the reason for the libsafe "errors"
1407 but the code is actually correct.
1410 +) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
1411 skipped when using openssl x509 multiple times on a single input file,
1412 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
1415 +) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
1416 set string type: to handle setting ASN1_TIME structures. Fix ca
1417 utility to correctly initialize revocation date of CRLs.
1420 +) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
1421 the clients preferred ciphersuites and rather use its own preferences.
1422 Should help to work around M$ SGC (Server Gated Cryptography) bug in
1423 Internet Explorer by ensuring unchanged hash method during stepup.
1424 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
1427 +) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
1428 to aes and add a new 'exist' option to print out symbols that don't
1432 +) Additional options to ocsp utility to allow flags to be set and
1433 additional certificates supplied.
1436 +) Add the option -VAfile to 'openssl ocsp', so the user can give the
1437 OCSP client a number of certificate to only verify the response
1441 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
1442 Bleichenbacher's DSA attack.
1443 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
1444 to be set and top=0 forces the highest bit to be set; top=-1 is new
1445 and leaves the highest bit random.
1446 [Ulf Moeller, Bodo Moeller]
1448 +) Update Rijndael code to version 3.0 and change EVP AES ciphers to
1449 handle the new API. Currently only ECB, CBC modes supported. Add new
1450 AES OIDs. Add TLS AES ciphersuites as described in the "AES Ciphersuites
1451 for TLS" draft-ietf-tls-ciphersuite-03.txt.
1452 [Ben Laurie, Steve Henson]
1454 *) In the NCONF_...-based implementations for CONF_... queries
1455 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
1456 a temporary CONF structure with the data component set to NULL
1457 (which gives segmentation faults in lh_retrieve).
1458 Instead, use NULL for the CONF pointer in CONF_get_string and
1459 CONF_get_number (which may use environment variables) and directly
1460 return NULL from CONF_get_section.
1463 *) Fix potential buffer overrun for EBCDIC.
1466 +) New function OCSP_copy_nonce() to copy nonce value (if present) from
1467 request to response.
1470 +) Functions for OCSP responders. OCSP_request_onereq_count(),
1471 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
1472 extract information from a certificate request. OCSP_response_create()
1473 creates a response and optionally adds a basic response structure.
1474 OCSP_basic_add1_status() adds a complete single response to a basic
1475 response and returns the OCSP_SINGLERESP structure just added (to allow
1476 extensions to be included for example). OCSP_basic_add1_cert() adds a
1477 certificate to a basic response and OCSP_basic_sign() signs a basic
1478 response with various flags. New helper functions ASN1_TIME_check()
1479 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
1480 (converts ASN1_TIME to GeneralizedTime).
1483 +) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
1484 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
1485 structure from a certificate. X509_pubkey_digest() digests the public_key
1486 contents: this is used in various key identifiers.
1489 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
1490 keyUsage if basicConstraints absent for a CA.
1493 *) Make SMIME_write_PKCS7() write mail header values with a format that
1494 is more generally accepted (no spaces before the semicolon), since
1495 some programs can't parse those values properly otherwise. Also make
1496 sure BIO's that break lines after each write do not create invalid
1500 +) Make sk_sort() tolerate a NULL argument.
1501 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
1503 +) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
1504 passed by the function are trusted implicitly. If any of them signed the
1505 response then it is assumed to be valid and is not verified.
1508 -) Make the CRL encoding routines work with empty SEQUENCE OF. The
1509 macros previously used would not encode an empty SEQUENCE OF
1510 and break the signature.
1512 [This change does not apply to 0.9.7.]
1514 *) Zero the premaster secret after deriving the master secret in
1518 +) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
1519 to data. This was previously part of the PKCS7 ASN1 code. This
1520 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
1521 [Steve Henson, reported by Kenneth R. Robinette
1522 <support@securenetterm.com>]
1524 +) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
1525 routines: without these tracing memory leaks is very painful.
1526 Fix leaks in PKCS12 and PKCS7 routines.
1529 *) Add some EVP_add_digest_alias registrations (as found in
1530 OpenSSL_add_all_digests()) to SSL_library_init()
1531 aka OpenSSL_add_ssl_algorithms(). This provides improved
1532 compatibility with peers using X.509 certificates
1533 with unconventional AlgorithmIdentifier OIDs.
1536 *) Fix for Irix with NO_ASM.
1537 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
1539 *) ./config script fixes.
1540 [Ulf Moeller, Richard Levitte]
1542 +) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
1543 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
1544 effectively meant GeneralizedTime would never be used. Now it
1545 is initialised to -1 but X509_time_adj() now has to check the value
1546 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
1547 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
1548 [Steve Henson, reported by Kenneth R. Robinette
1549 <support@securenetterm.com>]
1551 +) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
1552 result in a zero length in the ASN1_INTEGER structure which was
1553 not consistent with the structure when d2i_ASN1_INTEGER() was used
1554 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
1555 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
1556 where it did not print out a minus for negative ASN1_INTEGER.
1559 *) Fix 'openssl passwd -1'.
1562 +) Add summary printout to ocsp utility. The various functions which
1563 convert status values to strings have been renamed to:
1564 OCSP_response_status_str(), OCSP_cert_status_str() and
1565 OCSP_crl_reason_str() and are no longer static. New options
1566 to verify nonce values and to disable verification. OCSP response
1567 printout format cleaned up.
1570 +) Add additional OCSP certificate checks. These are those specified
1571 in RFC2560. This consists of two separate checks: the CA of the
1572 certificate being checked must either be the OCSP signer certificate
1573 or the issuer of the OCSP signer certificate. In the latter case the
1574 OCSP signer certificate must contain the OCSP signing extended key
1575 usage. This check is performed by attempting to match the OCSP
1576 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
1577 in the OCSP_CERTID structures of the response.
1580 +) Initial OCSP certificate verification added to OCSP_basic_verify()
1581 and related routines. This uses the standard OpenSSL certificate
1582 verify routines to perform initial checks (just CA validity) and
1583 to obtain the certificate chain. Then additional checks will be
1584 performed on the chain. Currently the root CA is checked to see
1585 if it is explicitly trusted for OCSP signing. This is used to set
1586 a root CA as a global signing root: that is any certificate that
1587 chains to that CA is an acceptable OCSP signing certificate.
1590 +) New '-extfile ...' option to 'openssl ca' for reading X.509v3
1591 extensions from a separate configuration file.
1592 As when reading extensions from the main configuration file,
1593 the '-extensions ...' option may be used for specifying the
1595 [Massimiliano Pala <madwolf@comune.modena.it>]
1597 *) Change PKCS12_key_gen_asc() so it can cope with non null
1598 terminated strings whose length is passed in the passlen
1599 parameter, for example from PEM callbacks. This was done
1600 by adding an extra length parameter to asc2uni().
1601 [Steve Henson, reported by <oddissey@samsung.co.kr>]
1603 +) New OCSP utility. Allows OCSP requests to be generated or
1604 read. The request can be sent to a responder and the output
1605 parsed, outputed or printed in text form. Not complete yet:
1606 still needs to check the OCSP response validity.
1609 +) New subcommands for 'openssl ca':
1610 'openssl ca -status <serial>' prints the status of the cert with
1611 the given serial number (according to the index file).
1612 'openssl ca -updatedb' updates the expiry status of certificates
1614 [Massimiliano Pala <madwolf@comune.modena.it>]
1616 +) New '-newreq-nodes' command option to CA.pl. This is like
1617 '-newreq', but calls 'openssl req' with the '-nodes' option
1618 so that the resulting key is not encrypted.
1619 [Damien Miller <djm@mindrot.org>]
1621 +) New configuration for the GNU Hurd.
1622 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
1624 +) Initial code to implement OCSP basic response verify. This
1625 is currently incomplete. Currently just finds the signer's
1626 certificate and verifies the signature on the response.
1629 +) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
1630 value of OPENSSLDIR. This is available via the new '-d' option
1631 to 'openssl version', and is also included in 'openssl version -a'.
1634 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
1635 call failed, free the DSA structure.
1638 +) Allowing defining memory allocation callbacks that will be given
1639 file name and line number information in additional arguments
1640 (a const char* and an int). The basic functionality remains, as
1641 well as the original possibility to just replace malloc(),
1642 realloc() and free() by functions that do not know about these
1643 additional arguments. To register and find out the current
1644 settings for extended allocation functions, the following
1645 functions are provided:
1647 CRYPTO_set_mem_ex_functions
1648 CRYPTO_set_locked_mem_ex_functions
1649 CRYPTO_get_mem_ex_functions
1650 CRYPTO_get_locked_mem_ex_functions
1652 These work the same way as CRYPTO_set_mem_functions and friends.
1653 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
1654 extended allocation function is enabled.
1655 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
1656 a conventional allocation function is enabled.
1657 [Richard Levitte, Bodo Moeller]
1659 *) Fix to uni2asc() to cope with zero length Unicode strings.
1660 These are present in some PKCS#12 files.
1663 +) Finish off removing the remaining LHASH function pointer casts.
1664 There should no longer be any prototype-casting required when using
1665 the LHASH abstraction, and any casts that remain are "bugs". See
1666 the callback types and macros at the head of lhash.h for details
1667 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
1670 +) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
1671 If /dev/[u]random devices are not available or do not return enough
1672 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
1674 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
1675 /etc/entropy will be queried once each in this sequence, quering stops
1676 when enough entropy was collected without querying more sockets.
1679 +) Change the Unix RAND_poll() variant to be able to poll several
1680 random devices, as specified by DEVRANDOM, until a sufficient amount
1681 of data has been collected. We spend at most 10 ms on each file
1682 (select timeout) and read in non-blocking mode. DEVRANDOM now
1683 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
1684 (previously it was just the string "/dev/urandom"), so on typical
1685 platforms the 10 ms delay will never occur.
1686 Also separate out the Unix variant to its own file, rand_unix.c.
1687 For VMS, there's a currently-empty rand_vms.c.
1690 +) Move OCSP client related routines to ocsp_cl.c. These
1691 provide utility functions which an application needing
1692 to issue a request to an OCSP responder and analyse the
1693 response will typically need: as opposed to those which an
1694 OCSP responder itself would need which will be added later.
1696 OCSP_request_sign() signs an OCSP request with an API similar
1697 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
1698 response. OCSP_response_get1_basic() extracts basic response
1699 from response. OCSP_resp_find_status(): finds and extracts status
1700 information from an OCSP_CERTID structure (which will be created
1701 when the request structure is built). These are built from lower
1702 level functions which work on OCSP_SINGLERESP structures but
1703 wont normally be used unless the application wishes to examine
1704 extensions in the OCSP response for example.
1706 Replace nonce routines with a pair of functions.
1707 OCSP_request_add1_nonce() adds a nonce value and optionally
1708 generates a random value. OCSP_check_nonce() checks the
1709 validity of the nonce in an OCSP response.
1712 +) Change function OCSP_request_add() to OCSP_request_add0_id().
1713 This doesn't copy the supplied OCSP_CERTID and avoids the
1714 need to free up the newly created id. Change return type
1715 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
1716 This can then be used to add extensions to the request.
1717 Deleted OCSP_request_new(), since most of its functionality
1718 is now in OCSP_REQUEST_new() (and the case insensitive name
1719 clash) apart from the ability to set the request name which
1720 will be added elsewhere.
1723 +) Update OCSP API. Remove obsolete extensions argument from
1724 various functions. Extensions are now handled using the new
1725 OCSP extension code. New simple OCSP HTTP function which
1726 can be used to send requests and parse the response.
1729 +) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
1730 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
1731 uses the special reorder version of SET OF to sort the attributes
1732 and reorder them to match the encoded order. This resolves a long
1733 standing problem: a verify on a PKCS7 structure just after signing
1734 it used to fail because the attribute order did not match the
1735 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
1736 it uses the received order. This is necessary to tolerate some broken
1737 software that does not order SET OF. This is handled by encoding
1738 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
1739 to produce the required SET OF.
1742 +) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
1743 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
1744 files to get correct declarations of the ASN.1 item variables.
1747 +) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
1748 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
1749 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
1750 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
1751 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
1752 ASN1_ITEM and no wrapper functions.
1755 +) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
1756 replace the old function pointer based I/O routines. Change most of
1757 the *_d2i_bio() and *_d2i_fp() functions to use these.
1760 +) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
1761 lines, recognice more "algorithms" that can be deselected, and make
1762 it complain about algorithm deselection that isn't recognised.
1765 +) New ASN1 functions to handle dup, sign, verify, digest, pack and
1766 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
1767 to use new functions. Add NO_ASN1_OLD which can be set to remove
1768 some old style ASN1 functions: this can be used to determine if old
1769 code will still work when these eventually go away.
1772 +) New extension functions for OCSP structures, these follow the
1773 same conventions as certificates and CRLs.
1776 +) New function X509V3_add1_i2d(). This automatically encodes and
1777 adds an extension. Its behaviour can be customised with various
1778 flags to append, replace or delete. Various wrappers added for
1779 certifcates and CRLs.
1782 +) Fix to avoid calling the underlying ASN1 print routine when
1783 an extension cannot be parsed. Correct a typo in the
1784 OCSP_SERVICELOC extension. Tidy up print OCSP format.
1787 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
1788 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
1789 when writing a 32767 byte record.
1790 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
1792 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
1793 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
1795 (RSA objects have a reference count access to which is protected
1796 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
1797 so they are meant to be shared between threads.)
1798 [Bodo Moeller, Geoff Thorpe; original patch submitted by
1799 "Reddie, Steven" <Steven.Reddie@ca.com>]
1801 +) Make mkdef.pl parse some of the ASN1 macros and add apropriate
1802 entries for variables.
1805 *) Fix a deadlock in CRYPTO_mem_leaks().
1808 +) Add functionality to apps/openssl.c for detecting locking
1809 problems: As the program is single-threaded, all we have
1810 to do is register a locking callback using an array for
1811 storing which locks are currently held by the program.
1814 +) Use a lock around the call to CRYPTO_get_ex_new_index() in
1815 SSL_get_ex_data_X509_STORE_idx(), which is used in
1816 ssl_verify_cert_chain() and thus can be called at any time
1817 during TLS/SSL handshakes so that thread-safety is essential.
1818 Unfortunately, the ex_data design is not at all suited
1819 for multi-threaded use, so it probably should be abolished.
1822 +) Added Broadcom "ubsec" ENGINE to OpenSSL.
1823 [Broadcom, tweaked and integrated by Geoff Thorpe]
1825 +) Move common extension printing code to new function
1826 X509V3_print_extensions(). Reorganise OCSP print routines and
1827 implement some needed OCSP ASN1 functions. Add OCSP extensions.
1830 +) New function X509_signature_print() to remove duplication in some
1834 +) Add a special meaning when SET OF and SEQUENCE OF flags are both
1835 set (this was treated exactly the same as SET OF previously). This
1836 is used to reorder the STACK representing the structure to match the
1837 encoding. This will be used to get round a problem where a PKCS7
1838 structure which was signed could not be verified because the STACK
1839 order did not reflect the encoded order.
1842 +) Reimplement the OCSP ASN1 module using the new code.
1845 +) Update the X509V3 code to permit the use of an ASN1_ITEM structure
1846 for its ASN1 operations. The old style function pointers still exist
1847 for now but they will eventually go away.
1850 +) Merge in replacement ASN1 code from the ASN1 branch. This almost
1851 completely replaces the old ASN1 functionality with a table driven
1852 encoder and decoder which interprets an ASN1_ITEM structure describing
1853 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
1854 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
1855 has also been converted to the new form.
1858 +) Change BN_mod_exp_recp so that negative moduli are tolerated
1859 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
1860 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
1861 for negative moduli.
1864 +) Fix BN_uadd and BN_usub: Always return non-negative results instead
1865 of not touching the result's sign bit.
1868 +) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
1872 +) Changed the LHASH code to use prototypes for callbacks, and created
1873 macros to declare and implement thin (optionally static) functions
1874 that provide type-safety and avoid function pointer casting for the
1875 type-specific callbacks.
1878 *) Use better test patterns in bntest.
1881 +) Added Kerberos Cipher Suites to be used with TLS, as written in
1883 [Veers Staats <staatsvr@asc.hpc.mil>,
1884 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
1886 *) rand_win.c fix for Borland C.
1889 *) BN_rshift bugfix for n == 0.
1892 +) Reformat the FAQ so the different questions and answers can be divided
1893 in sections depending on the subject.
1896 +) Have the zlib compression code load ZLIB.DLL dynamically under
1900 +) New function BN_mod_sqrt for computing square roots modulo a prime
1901 (using the probabilistic Tonelli-Shanks algorithm unless
1902 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
1903 be handled deterministically).
1904 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
1906 *) Add a 'bctest' script that checks for some known 'bc' bugs
1907 so that 'make test' does not abort just because 'bc' is broken.
1910 *) Store verify_result within SSL_SESSION also for client side to
1911 avoid potential security hole. (Re-used sessions on the client side
1912 always resulted in verify_result==X509_V_OK, not using the original
1913 result of the server certificate verification.)
1916 +) Make BN_mod_inverse faster by explicitly handling small quotients
1917 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
1918 512 bits], about 30% for larger ones [1024 or 2048 bits].)
1921 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
1922 SSL3_RT_APPLICATION_DATA, return 0.
1923 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
1927 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
1928 releases, have been re-implemented by renaming the previous
1929 implementations of ssl2_read and ssl3_read to ssl2_read_internal
1930 and ssl3_read_internal, respectively, and adding 'peek' parameters
1931 to them. The new ssl[23]_{read,peek} functions are calls to
1932 ssl[23]_read_internal with the 'peek' flag set appropriately.
1933 A 'peek' parameter has also been added to ssl3_read_bytes, which
1934 does the actual work for ssl3_read_internal.
1937 +) New function BN_kronecker.
1940 +) Fix BN_gcd so that it works on negative inputs; the result is
1941 positive unless both parameters are zero.
1942 Previously something reasonably close to an infinite loop was
1943 possible because numbers could be growing instead of shrinking
1944 in the implementation of Euclid's algorithm.
1947 +) Fix BN_is_word() and BN_is_one() macros to take into account the
1948 sign of the number in question.
1950 Fix BN_is_word(a,w) to work correctly for w == 0.
1952 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
1953 because its test if the absolute value of 'a' equals 'w'.
1954 Note that BN_abs_is_word does *not* handle w == 0 reliably;
1955 it exists mostly for use in the implementations of BN_is_zero(),
1956 BN_is_one(), and BN_is_word().
1959 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
1960 the method-specific "init()" handler. Also clean up ex_data after
1961 calling the method-specific "finish()" handler. Previously, this was
1962 happening the other way round.
1965 +) New function BN_swap.
1968 +) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
1969 the exponentiation functions are more likely to produce reasonable
1970 results on negative inputs.
1973 +) Change BN_mod_mul so that the result is always non-negative.
1974 Previously, it could be negative if one of the factors was negative;
1975 I don't think anyone really wanted that behaviour.
1978 +) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
1979 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
1980 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
1981 and add new functions:
1990 BN_mod_lshift1_quick
1994 These functions always generate non-negative results.
1996 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
1997 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
1999 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
2000 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
2001 be reduced modulo m.
2002 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
2004 +) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2005 was actually never needed) and in BN_mul(). The removal in BN_mul()
2006 required a small change in bn_mul_part_recursive() and the addition
2007 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2008 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2009 bn_sub_words() and bn_add_words() except they take arrays with
2013 +) In 'openssl passwd', verify passwords read from the terminal
2014 unless the '-salt' option is used (which usually means that
2015 verification would just waste user's time since the resulting
2016 hash is going to be compared with some given password hash)
2017 or the new '-noverify' option is used.
2019 This is an incompatible change, but it does not affect
2020 non-interactive use of 'openssl passwd' (passwords on the command
2021 line, '-stdin' option, '-in ...' option) and thus should not
2025 +) Remove all references to RSAref, since there's no more need for it.
2028 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
2029 The previous value, 12, was not always sufficient for BN_mod_exp().
2032 +) Make DSO load along a path given through an environment variable
2033 (SHLIB_PATH) with shl_load().
2036 +) Constify the ENGINE code as a result of BIGNUM constification.
2037 Also constify the RSA code and most things related to it. In a
2038 few places, most notable in the depth of the ASN.1 code, ugly
2039 casts back to non-const were required (to be solved at a later
2043 +) Make it so the openssl application has all engines loaded by default.
2046 +) Constify the BIGNUM routines a little more.
2049 *) Make sure that shared libraries get the internal name engine with
2050 the full version number and not just 0. This should mark the
2051 shared libraries as not backward compatible. Of course, this should
2052 be changed again when we can guarantee backward binary compatibility.
2055 +) Add the following functions:
2057 ENGINE_load_cswift()
2059 ENGINE_load_atalla()
2061 ENGINE_load_builtin_engines()
2063 That way, an application can itself choose if external engines that
2064 are built-in in OpenSSL shall ever be used or not. The benefit is
2065 that applications won't have to be linked with libdl or other dso
2066 libraries unless it's really needed.
2068 Changed 'openssl engine' to load all engines on demand.
2069 Changed the engine header files to avoid the duplication of some
2070 declarations (they differed!).
2073 +) 'openssl engine' can now list capabilities.
2076 +) Better error reporting in 'openssl engine'.
2079 +) Never call load_dh_param(NULL) in s_server.
2082 +) Add engine application. It can currently list engines by name and
2083 identity, and test if they are actually available.
2086 +) Improve RPM specification file by forcing symbolic linking and making
2087 sure the installed documentation is also owned by root.root.
2088 [Damien Miller <djm@mindrot.org>]
2090 +) Give the OpenSSL applications more possibilities to make use of
2091 keys (public as well as private) handled by engines.
2094 +) Add OCSP code that comes from CertCo.
2097 +) Add VMS support for the Rijndael code.
2100 +) Added untested support for Nuron crypto accelerator.
2103 +) Add support for external cryptographic devices. This code was
2104 previously distributed separately as the "engine" branch.
2105 [Geoff Thorpe, Richard Levitte]
2107 +) Rework the filename-translation in the DSO code. It is now possible to
2108 have far greater control over how a "name" is turned into a filename
2109 depending on the operating environment and any oddities about the
2110 different shared library filenames on each system.
2113 +) Support threads on FreeBSD-elf in Configure.
2116 +) Fix for SHA1 assembly problem with MASM: it produces
2117 warnings about corrupt line number information when assembling
2118 with debugging information. This is caused by the overlapping
2120 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
2123 NCONF_get_number() has no error checking at all. As a replacement,
2124 NCONF_get_number_e() is defined (_e for "error checking") and is
2125 promoted strongly. The old NCONF_get_number is kept around for
2126 binary backward compatibility.
2127 Make it possible for methods to load from something other than a BIO,
2128 by providing a function pointer that is given a name instead of a BIO.
2129 For example, this could be used to load configuration data from an
2133 *) Fix typo in get_cert_by_subject() in by_dir.c
2134 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
2136 *) Rework the system to generate shared libraries:
2138 - Make note of the expected extension for the shared libraries and
2139 if there is a need for symbolic links from for example libcrypto.so.0
2140 to libcrypto.so.0.9.7. There is extended info in Configure for
2143 - Make as few rebuilds of the shared libraries as possible.
2145 - Still avoid linking the OpenSSL programs with the shared libraries.
2147 - When installing, install the shared libraries separately from the
2151 +) Fix for non blocking accept BIOs. Added new I/O special reason
2152 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
2153 with non blocking I/O was not possible because no retry code was
2154 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
2158 +) Added the beginnings of Rijndael support.
2161 +) Fix for bug in DirectoryString mask setting. Add support for
2162 X509_NAME_print_ex() in 'req' and X509_print_ex() function
2163 to allow certificate printing to more controllable, additional
2164 'certopt' option to 'x509' to allow new printing options to be
2168 +) Clean old EAY MD5 hack from e_os.h.
2171 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
2173 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
2174 and not in SSL_clear because the latter is also used by the
2175 accept/connect functions; previously, the settings made by
2176 SSL_set_read_ahead would be lost during the handshake.
2177 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
2179 *) Correct util/mkdef.pl to be selective about disabled algorithms.
2180 Previously, it would create entries for disableed algorithms no
2184 *) Added several new manual pages for SSL_* function.
2187 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
2189 *) In ssl23_get_client_hello, generate an error message when faced
2190 with an initial SSL 3.0/TLS record that is too small to contain the
2191 first two bytes of the ClientHello message, i.e. client_version.
2192 (Note that this is a pathologic case that probably has never happened
2193 in real life.) The previous approach was to use the version number
2194 from the record header as a substitute; but our protocol choice
2195 should not depend on that one because it is not authenticated
2196 by the Finished messages.
2199 *) More robust randomness gathering functions for Windows.
2200 [Jeffrey Altman <jaltman@columbia.edu>]
2202 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
2203 not set then we don't setup the error code for issuer check errors
2204 to avoid possibly overwriting other errors which the callback does
2205 handle. If an application does set the flag then we assume it knows
2206 what it is doing and can handle the new informational codes
2210 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
2211 a general "ANY" type, as such it should be able to decode anything
2212 including tagged types. However it didn't check the class so it would
2213 wrongly interpret tagged types in the same way as their universal
2214 counterpart and unknown types were just rejected. Changed so that the
2215 tagged and unknown types are handled in the same way as a SEQUENCE:
2216 that is the encoding is stored intact. There is also a new type
2217 "V_ASN1_OTHER" which is used when the class is not universal, in this
2218 case we have no idea what the actual type is so we just lump them all
2222 *) On VMS, stdout may very well lead to a file that is written to
2223 in a record-oriented fashion. That means that every write() will
2224 write a separate record, which will be read separately by the
2225 programs trying to read from it. This can be very confusing.
2227 The solution is to put a BIO filter in the way that will buffer
2228 text until a linefeed is reached, and then write everything a
2229 line at a time, so every record written will be an actual line,
2230 not chunks of lines and not (usually doesn't happen, but I've
2231 seen it once) several lines in one record. BIO_f_linebuffer() is
2234 Currently, it's a VMS-only method, because that's where it has
2235 been tested well enough.
2238 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
2239 it can return incorrect results.
2240 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
2241 but it was in 0.9.6-beta[12].)
2244 *) Disable the check for content being present when verifying detached
2245 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
2246 include zero length content when signing messages.
2249 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
2250 BIO_ctrl (for BIO pairs).
2253 *) Add DSO method for VMS.
2256 *) Bug fix: Montgomery multiplication could produce results with the
2260 *) Add RPM specification openssl.spec and modify it to build three
2261 packages. The default package contains applications, application
2262 documentation and run-time libraries. The devel package contains
2263 include files, static libraries and function documentation. The
2264 doc package contains the contents of the doc directory. The original
2265 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
2268 *) Add a large number of documentation files for many SSL routines.
2269 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
2271 *) Add a configuration entry for Sony News 4.
2272 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
2274 *) Don't set the two most significant bits to one when generating a
2275 random number < q in the DSA library.
2278 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
2279 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
2280 the underlying transport is blocking) if a handshake took place.
2281 (The default behaviour is needed by applications such as s_client
2282 and s_server that use select() to determine when to use SSL_read;
2283 but for applications that know in advance when to expect data, it
2284 just makes things more complicated.)
2287 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
2291 *) Add a few more EBCDIC conditionals that make `req' and `x509'
2292 work better on such systems.
2293 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
2295 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
2296 Update PKCS12_parse() so it copies the friendlyName and the
2297 keyid to the certificates aux info.
2300 *) Fix bug in PKCS7_verify() which caused an infinite loop
2301 if there was more than one signature.
2302 [Sven Uszpelkat <su@celocom.de>]
2304 *) Major change in util/mkdef.pl to include extra information
2305 about each symbol, as well as presentig variables as well
2306 as functions. This change means that there's n more need
2307 to rebuild the .num files when some algorithms are excluded.
2310 *) Allow the verify time to be set by an application,
2311 rather than always using the current time.
2314 *) Phase 2 verify code reorganisation. The certificate
2315 verify code now looks up an issuer certificate by a
2316 number of criteria: subject name, authority key id
2317 and key usage. It also verifies self signed certificates
2318 by the same criteria. The main comparison function is
2319 X509_check_issued() which performs these checks.
2321 Lot of changes were necessary in order to support this
2322 without completely rewriting the lookup code.
2324 Authority and subject key identifier are now cached.
2326 The LHASH 'certs' is X509_STORE has now been replaced
2327 by a STACK_OF(X509_OBJECT). This is mainly because an
2328 LHASH can't store or retrieve multiple objects with
2329 the same hash value.
2331 As a result various functions (which were all internal
2332 use only) have changed to handle the new X509_STORE
2333 structure. This will break anything that messed round
2334 with X509_STORE internally.
2336 The functions X509_STORE_add_cert() now checks for an
2337 exact match, rather than just subject name.
2339 The X509_STORE API doesn't directly support the retrieval
2340 of multiple certificates matching a given criteria, however
2341 this can be worked round by performing a lookup first
2342 (which will fill the cache with candidate certificates)
2343 and then examining the cache for matches. This is probably
2344 the best we can do without throwing out X509_LOOKUP
2345 entirely (maybe later...).
2347 The X509_VERIFY_CTX structure has been enhanced considerably.
2349 All certificate lookup operations now go via a get_issuer()
2350 callback. Although this currently uses an X509_STORE it
2351 can be replaced by custom lookups. This is a simple way
2352 to bypass the X509_STORE hackery necessary to make this
2353 work and makes it possible to use more efficient techniques
2354 in future. A very simple version which uses a simple
2355 STACK for its trusted certificate store is also provided
2356 using X509_STORE_CTX_trusted_stack().
2358 The verify_cb() and verify() callbacks now have equivalents
2359 in the X509_STORE_CTX structure.
2361 X509_STORE_CTX also has a 'flags' field which can be used
2362 to customise the verify behaviour.
2365 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
2366 excludes S/MIME capabilities.
2369 *) When a certificate request is read in keep a copy of the
2370 original encoding of the signed data and use it when outputing
2371 again. Signatures then use the original encoding rather than
2372 a decoded, encoded version which may cause problems if the
2373 request is improperly encoded.
2376 *) For consistency with other BIO_puts implementations, call
2377 buffer_write(b, ...) directly in buffer_puts instead of calling
2380 In BIO_puts, increment b->num_write as in BIO_write.
2381 [Peter.Sylvester@EdelWeb.fr]
2383 *) Fix BN_mul_word for the case where the word is 0. (We have to use
2384 BN_zero, we may not return a BIGNUM with an array consisting of
2388 *) Avoid calling abort() from within the library when problems are
2389 detected, except if preprocessor symbols have been defined
2390 (such as REF_CHECK, BN_DEBUG etc.).
2393 *) New openssl application 'rsautl'. This utility can be
2394 used for low level RSA operations. DER public key
2395 BIO/fp routines also added.
2398 *) New Configure entry and patches for compiling on QNX 4.
2399 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
2401 *) A demo state-machine implementation was sponsored by
2402 Nuron (http://www.nuron.com/) and is now available in
2403 demos/state_machine.
2406 *) New options added to the 'dgst' utility for signature
2407 generation and verification.
2410 *) Unrecognized PKCS#7 content types are now handled via a
2411 catch all ASN1_TYPE structure. This allows unsupported
2412 types to be stored as a "blob" and an application can
2413 encode and decode it manually.
2416 *) Fix various signed/unsigned issues to make a_strex.c
2418 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
2420 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
2421 length if passed a buffer. ASN1_INTEGER_to_BN failed
2422 if passed a NULL BN and its argument was negative.
2423 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
2425 *) Modification to PKCS#7 encoding routines to output definite
2426 length encoding. Since currently the whole structures are in
2427 memory there's not real point in using indefinite length
2428 constructed encoding. However if OpenSSL is compiled with
2429 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
2432 *) Added BIO_vprintf() and BIO_vsnprintf().
2435 *) Added more prefixes to parse for in the the strings written
2436 through a logging bio, to cover all the levels that are available
2437 through syslog. The prefixes are now:
2439 PANIC, EMERG, EMR => LOG_EMERG
2440 ALERT, ALR => LOG_ALERT
2441 CRIT, CRI => LOG_CRIT
2442 ERROR, ERR => LOG_ERR
2443 WARNING, WARN, WAR => LOG_WARNING
2444 NOTICE, NOTE, NOT => LOG_NOTICE
2445 INFO, INF => LOG_INFO
2446 DEBUG, DBG => LOG_DEBUG
2448 and as before, if none of those prefixes are present at the
2449 beginning of the string, LOG_ERR is chosen.
2451 On Win32, the LOG_* levels are mapped according to this:
2453 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
2454 LOG_WARNING => EVENTLOG_WARNING_TYPE
2455 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
2459 *) Made it possible to reconfigure with just the configuration
2460 argument "reconf" or "reconfigure". The command line arguments
2461 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
2462 and are retrieved from there when reconfiguring.
2466 [Assar Westerlund <assar@sics.se>, Richard Levitte]
2468 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
2471 *) The obj_dat.pl script was messing up the sorting of object
2472 names. The reason was that it compared the quoted version
2473 of strings as a result "OCSP" > "OCSP Signing" because
2474 " > SPACE. Changed script to store unquoted versions of
2475 names and add quotes on output. It was also omitting some
2476 names from the lookup table if they were given a default
2477 value (that is if SN is missing it is given the same
2478 value as LN and vice versa), these are now added on the
2479 grounds that if an object has a name we should be able to
2480 look it up. Finally added warning output when duplicate
2481 short or long names are found.
2484 *) Changes needed for Tandem NSK.
2485 [Scott Uroff <scott@xypro.com>]
2487 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
2488 RSA_padding_check_SSLv23(), special padding was never detected
2489 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
2490 version rollback attacks was not effective.
2492 In s23_clnt.c, don't use special rollback-attack detection padding
2493 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
2494 client; similarly, in s23_srvr.c, don't do the rollback check if
2495 SSL 2.0 is the only protocol enabled in the server.
2498 *) Make it possible to get hexdumps of unprintable data with 'openssl
2499 asn1parse'. By implication, the functions ASN1_parse_dump() and
2500 BIO_dump_indent() are added.
2503 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
2504 these print out strings and name structures based on various
2505 flags including RFC2253 support and proper handling of
2506 multibyte characters. Added options to the 'x509' utility
2507 to allow the various flags to be set.
2510 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
2511 Also change the functions X509_cmp_current_time() and
2512 X509_gmtime_adj() work with an ASN1_TIME structure,
2513 this will enable certificates using GeneralizedTime in validity
2514 dates to be checked.
2517 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
2518 negative public key encodings) on by default,
2519 NO_NEG_PUBKEY_BUG can be set to disable it.
2522 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
2523 content octets. An i2c_ASN1_OBJECT is unnecessary because
2524 the encoding can be trivially obtained from the structure.
2527 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
2528 not read locks (CRYPTO_r_[un]lock).
2531 *) A first attempt at creating official support for shared
2532 libraries through configuration. I've kept it so the
2533 default is static libraries only, and the OpenSSL programs
2534 are always statically linked for now, but there are
2535 preparations for dynamic linking in place.
2536 This has been tested on Linux and Tru64.
2539 *) Randomness polling function for Win9x, as described in:
2540 Peter Gutmann, Software Generation of Practically Strong
2544 *) Fix so PRNG is seeded in req if using an already existing
2548 *) New options to smime application. -inform and -outform
2549 allow alternative formats for the S/MIME message including
2550 PEM and DER. The -content option allows the content to be
2551 specified separately. This should allow things like Netscape
2552 form signing output easier to verify.
2555 *) Fix the ASN1 encoding of tags using the 'long form'.
2558 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
2559 STRING types. These convert content octets to and from the
2560 underlying type. The actual tag and length octets are
2561 already assumed to have been read in and checked. These
2562 are needed because all other string types have virtually
2563 identical handling apart from the tag. By having versions
2564 of the ASN1 functions that just operate on content octets
2565 IMPLICIT tagging can be handled properly. It also allows
2566 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
2567 and ASN1_INTEGER are identical apart from the tag.
2570 *) Change the handling of OID objects as follows:
2572 - New object identifiers are inserted in objects.txt, following
2573 the syntax given in objects.README.
2574 - objects.pl is used to process obj_mac.num and create a new
2576 - obj_dat.pl is used to create a new obj_dat.h, using the data in
2579 This is currently kind of a hack, and the perl code in objects.pl
2580 isn't very elegant, but it works as I intended. The simplest way
2581 to check that it worked correctly is to look in obj_dat.h and
2582 check the array nid_objs and make sure the objects haven't moved
2583 around (this is important!). Additions are OK, as well as
2584 consistent name changes.
2587 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
2590 *) Addition of the command line parameter '-rand file' to 'openssl req'.
2591 The given file adds to whatever has already been seeded into the
2592 random pool through the RANDFILE configuration file option or
2593 environment variable, or the default random state file.
2596 *) mkstack.pl now sorts each macro group into lexical order.
2597 Previously the output order depended on the order the files
2598 appeared in the directory, resulting in needless rewriting
2602 *) Patches to make OpenSSL compile under Win32 again. Mostly
2603 work arounds for the VC++ problem that it treats func() as
2604 func(void). Also stripped out the parts of mkdef.pl that
2605 added extra typesafe functions: these no longer exist.
2608 *) Reorganisation of the stack code. The macros are now all
2609 collected in safestack.h . Each macro is defined in terms of
2610 a "stack macro" of the form SKM_<name>(type, a, b). The
2611 DEBUG_SAFESTACK is now handled in terms of function casts,
2612 this has the advantage of retaining type safety without the
2613 use of additional functions. If DEBUG_SAFESTACK is not defined
2614 then the non typesafe macros are used instead. Also modified the
2615 mkstack.pl script to handle the new form. Needs testing to see
2616 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
2617 the default if no major problems. Similar behaviour for ASN1_SET_OF
2618 and PKCS12_STACK_OF.
2621 *) When some versions of IIS use the 'NET' form of private key the
2622 key derivation algorithm is different. Normally MD5(password) is
2623 used as a 128 bit RC4 key. In the modified case
2624 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
2625 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
2626 as the old Netscape_RSA functions except they have an additional
2627 'sgckey' parameter which uses the modified algorithm. Also added
2628 an -sgckey command line option to the rsa utility. Thanks to
2629 Adrian Peck <bertie@ncipher.com> for posting details of the modified
2630 algorithm to openssl-dev.
2633 *) The evp_local.h macros were using 'c.##kname' which resulted in
2634 invalid expansion on some systems (SCO 5.0.5 for example).
2635 Corrected to 'c.kname'.
2636 [Phillip Porch <root@theporch.com>]
2638 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
2639 a STACK of email addresses from a certificate or request, these look
2640 in the subject name and the subject alternative name extensions and
2641 omit any duplicate addresses.
2644 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
2645 This makes DSA verification about 2 % faster.
2648 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
2649 (meaning that now 2^5 values will be precomputed, which is only 4 KB
2650 plus overhead for 1024 bit moduli).
2651 This makes exponentiations about 0.5 % faster for 1024 bit
2652 exponents (as measured by "openssl speed rsa2048").
2655 *) Rename memory handling macros to avoid conflicts with other
2657 Malloc => OPENSSL_malloc
2658 Malloc_locked => OPENSSL_malloc_locked
2659 Realloc => OPENSSL_realloc
2660 Free => OPENSSL_free
2663 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
2664 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
2667 *) CygWin32 support.
2668 [John Jarvie <jjarvie@newsguy.com>]
2670 *) The type-safe stack code has been rejigged. It is now only compiled
2671 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
2672 by default all type-specific stack functions are "#define"d back to
2673 standard stack functions. This results in more streamlined output
2674 but retains the type-safety checking possibilities of the original
2678 *) The STACK code has been cleaned up, and certain type declarations
2679 that didn't make a lot of sense have been brought in line. This has
2680 also involved a cleanup of sorts in safestack.h to more correctly
2681 map type-safe stack functions onto their plain stack counterparts.
2682 This work has also resulted in a variety of "const"ifications of
2683 lots of the code, especially "_cmp" operations which should normally
2684 be prototyped with "const" parameters anyway.
2687 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
2688 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
2689 (The PRNG state consists of two parts, the large pool 'state' and 'md',
2690 where all of 'md' is used each time the PRNG is used, but 'state'
2691 is used only indexed by a cyclic counter. As entropy may not be
2692 well distributed from the beginning, 'md' is important as a
2693 chaining variable. However, the output function chains only half
2694 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
2695 all of 'md', and seeding with STATE_SIZE dummy bytes will result
2696 in all of 'state' being rewritten, with the new values depending
2697 on virtually all of 'md'. This overcomes the 80 bit limitation.)
2700 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
2701 the handshake is continued after ssl_verify_cert_chain();
2702 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
2703 can lead to 'unexplainable' connection aborts later.
2704 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
2706 *) Major EVP API cipher revision.
2707 Add hooks for extra EVP features. This allows various cipher
2708 parameters to be set in the EVP interface. Support added for variable
2709 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
2710 setting of RC2 and RC5 parameters.
2712 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
2715 Remove lots of duplicated code from the EVP library. For example *every*
2716 cipher init() function handles the 'iv' in the same way according to the
2717 cipher mode. They also all do nothing if the 'key' parameter is NULL and
2718 for CFB and OFB modes they zero ctx->num.
2720 New functionality allows removal of S/MIME code RC2 hack.
2722 Most of the routines have the same form and so can be declared in terms
2725 By shifting this to the top level EVP_CipherInit() it can be removed from
2726 all individual ciphers. If the cipher wants to handle IVs or keys
2727 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
2730 Change lots of functions like EVP_EncryptUpdate() to now return a
2731 value: although software versions of the algorithms cannot fail
2732 any installed hardware versions can.
2735 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
2736 this option is set, tolerate broken clients that send the negotiated
2737 protocol version number instead of the requested protocol version
2741 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
2742 i.e. non-zero for export ciphersuites, zero otherwise.
2743 Previous versions had this flag inverted, inconsistent with
2744 rsa_tmp_cb (..._TMP_RSA_CB).
2745 [Bodo Moeller; problem reported by Amit Chopra]
2747 *) Add missing DSA library text string. Work around for some IIS
2748 key files with invalid SEQUENCE encoding.
2751 *) Add a document (doc/standards.txt) that list all kinds of standards
2752 and so on that are implemented in OpenSSL.
2755 *) Enhance c_rehash script. Old version would mishandle certificates
2756 with the same subject name hash and wouldn't handle CRLs at all.
2757 Added -fingerprint option to crl utility, to support new c_rehash
2761 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
2764 *) Fix for SSL server purpose checking. Server checking was
2765 rejecting certificates which had extended key usage present
2766 but no ssl client purpose.
2767 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
2769 *) Make PKCS#12 code work with no password. The PKCS#12 spec
2770 is a little unclear about how a blank password is handled.
2771 Since the password in encoded as a BMPString with terminating
2772 double NULL a zero length password would end up as just the
2773 double NULL. However no password at all is different and is
2774 handled differently in the PKCS#12 key generation code. NS
2775 treats a blank password as zero length. MSIE treats it as no
2776 password on export: but it will try both on import. We now do
2777 the same: PKCS12_parse() tries zero length and no password if
2778 the password is set to "" or NULL (NULL is now a valid password:
2779 it wasn't before) as does the pkcs12 application.
2782 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
2783 perror when PEM_read_bio_X509_REQ fails, the error message must
2784 be obtained from the error queue.
2787 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
2788 it in ERR_remove_state if appropriate, and change ERR_get_state
2789 accordingly to avoid race conditions (this is necessary because
2790 thread_hash is no longer constant once set).
2793 *) Bugfix for linux-elf makefile.one.
2796 *) RSA_get_default_method() will now cause a default
2797 RSA_METHOD to be chosen if one doesn't exist already.
2798 Previously this was only set during a call to RSA_new()
2799 or RSA_new_method(NULL) meaning it was possible for
2800 RSA_get_default_method() to return NULL.
2803 *) Added native name translation to the existing DSO code
2804 that will convert (if the flag to do so is set) filenames
2805 that are sufficiently small and have no path information
2806 into a canonical native form. Eg. "blah" converted to
2807 "libblah.so" or "blah.dll" etc.
2810 *) New function ERR_error_string_n(e, buf, len) which is like
2811 ERR_error_string(e, buf), but writes at most 'len' bytes
2812 including the 0 terminator. For ERR_error_string_n, 'buf'
2814 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
2816 *) CONF library reworked to become more general. A new CONF
2817 configuration file reader "class" is implemented as well as a
2818 new functions (NCONF_*, for "New CONF") to handle it. The now
2819 old CONF_* functions are still there, but are reimplemented to
2820 work in terms of the new functions. Also, a set of functions
2821 to handle the internal storage of the configuration data is
2822 provided to make it easier to write new configuration file
2823 reader "classes" (I can definitely see something reading a
2824 configuration file in XML format, for example), called _CONF_*,
2825 or "the configuration storage API"...
2827 The new configuration file reading functions are:
2829 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
2830 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
2832 NCONF_default, NCONF_WIN32
2834 NCONF_dump_fp, NCONF_dump_bio
2836 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
2837 NCONF_new creates a new CONF object. This works in the same way
2838 as other interfaces in OpenSSL, like the BIO interface.
2839 NCONF_dump_* dump the internal storage of the configuration file,
2840 which is useful for debugging. All other functions take the same
2841 arguments as the old CONF_* functions wth the exception of the
2842 first that must be a `CONF *' instead of a `LHASH *'.
2844 To make it easer to use the new classes with the old CONF_* functions,
2845 the function CONF_set_default_method is provided.
2848 *) Add '-tls1' option to 'openssl ciphers', which was already
2849 mentioned in the documentation but had not been implemented.
2850 (This option is not yet really useful because even the additional
2851 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
2854 *) Initial DSO code added into libcrypto for letting OpenSSL (and
2855 OpenSSL-based applications) load shared libraries and bind to
2856 them in a portable way.
2857 [Geoff Thorpe, with contributions from Richard Levitte]
2859 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
2861 *) Make sure _lrotl and _lrotr are only used with MSVC.
2863 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
2864 (the default implementation of RAND_status).
2866 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
2867 to '-clrext' (= clear extensions), as intended and documented.
2868 [Bodo Moeller; inconsistency pointed out by Michael Attili
2871 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
2872 was larger than the MD block size.
2873 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
2875 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
2876 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
2877 using the passed key: if the passed key was a private key the result
2878 of X509_print(), for example, would be to print out all the private key
2882 *) des_quad_cksum() byte order bug fix.
2883 [Ulf Möller, using the problem description in krb4-0.9.7, where
2884 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
2886 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
2888 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
2890 *) For easily testing in shell scripts whether some command
2891 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
2892 returns with exit code 0 iff no command of the given name is available.
2893 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
2894 the output goes to stdout and nothing is printed to stderr.
2895 Additional arguments are always ignored.
2897 Since for each cipher there is a command of the same name,
2898 the 'no-cipher' compilation switches can be tested this way.
2900 ('openssl no-XXX' is not able to detect pseudo-commands such
2901 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
2904 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
2907 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
2908 is set; it will be thrown away anyway because each handshake creates
2910 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
2911 to parameters -- in previous versions (since OpenSSL 0.9.3) the
2912 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
2913 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
2916 *) New s_client option -ign_eof: EOF at stdin is ignored, and
2917 'Q' and 'R' lose their special meanings (quit/renegotiate).
2918 This is part of what -quiet does; unlike -quiet, -ign_eof
2919 does not suppress any output.
2922 *) Add compatibility options to the purpose and trust code. The
2923 purpose X509_PURPOSE_ANY is "any purpose" which automatically
2924 accepts a certificate or CA, this was the previous behaviour,
2925 with all the associated security issues.
2927 X509_TRUST_COMPAT is the old trust behaviour: only and
2928 automatically trust self signed roots in certificate store. A
2929 new trust setting X509_TRUST_DEFAULT is used to specify that
2930 a purpose has no associated trust setting and it should instead
2931 use the value in the default purpose.
2934 *) Fix the PKCS#8 DSA private key code so it decodes keys again
2935 and fix a memory leak.
2938 *) In util/mkerr.pl (which implements 'make errors'), preserve
2939 reason strings from the previous version of the .c file, as
2940 the default to have only downcase letters (and digits) in
2941 automatically generated reasons codes is not always appropriate.
2944 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
2945 using strerror. Previously, ERR_reason_error_string() returned
2946 library names as reason strings for SYSerr; but SYSerr is a special
2947 case where small numbers are errno values, not library numbers.
2950 *) Add '-dsaparam' option to 'openssl dhparam' application. This
2951 converts DSA parameters into DH parameters. (When creating parameters,
2952 DSA_generate_parameters is used.)
2955 *) Include 'length' (recommended exponent length) in C code generated
2956 by 'openssl dhparam -C'.
2959 *) The second argument to set_label in perlasm was already being used
2960 so couldn't be used as a "file scope" flag. Moved to third argument
2964 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
2965 instead of RAND_bytes for encryption IVs and salts.
2968 *) Include RAND_status() into RAND_METHOD instead of implementing
2969 it only for md_rand.c Otherwise replacing the PRNG by calling
2970 RAND_set_rand_method would be impossible.
2973 *) Don't let DSA_generate_key() enter an infinite loop if the random
2974 number generation fails.
2977 *) New 'rand' application for creating pseudo-random output.
2980 *) Added configuration support for Linux/IA64
2981 [Rolf Haberrecker <rolf@suse.de>]
2983 *) Assembler module support for Mingw32.
2986 *) Shared library support for HPUX (in shlib/).
2987 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
2989 *) Shared library support for Solaris gcc.
2990 [Lutz Behnke <behnke@trustcenter.de>]
2992 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
2994 *) PKCS7_encrypt() was adding text MIME headers twice because they
2995 were added manually and by SMIME_crlf_copy().
2998 *) In bntest.c don't call BN_rand with zero bits argument.
2999 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
3001 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
3002 case was implemented. This caused BN_div_recp() to fail occasionally.
3005 *) Add an optional second argument to the set_label() in the perl
3006 assembly language builder. If this argument exists and is set
3007 to 1 it signals that the assembler should use a symbol whose
3008 scope is the entire file, not just the current function. This
3009 is needed with MASM which uses the format label:: for this scope.
3010 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
3012 *) Change the ASN1 types so they are typedefs by default. Before
3013 almost all types were #define'd to ASN1_STRING which was causing
3014 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
3018 *) Change names of new functions to the new get1/get0 naming
3019 convention: After 'get1', the caller owns a reference count
3020 and has to call ..._free; 'get0' returns a pointer to some
3021 data structure without incrementing reference counters.
3022 (Some of the existing 'get' functions increment a reference
3023 counter, some don't.)
3024 Similarly, 'set1' and 'add1' functions increase reference
3025 counters or duplicate objects.
3028 *) Allow for the possibility of temp RSA key generation failure:
3029 the code used to assume it always worked and crashed on failure.
3032 *) Fix potential buffer overrun problem in BIO_printf().
3033 [Ulf Möller, using public domain code by Patrick Powell; problem
3034 pointed out by David Sacerdote <das33@cornell.edu>]
3036 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
3037 RAND_egd() and RAND_status(). In the command line application,
3038 the EGD socket can be specified like a seed file using RANDFILE
3042 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
3043 Some CAs (e.g. Verisign) distribute certificates in this form.
3046 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
3047 list to exclude them. This means that no special compilation option
3048 is needed to use anonymous DH: it just needs to be included in the
3052 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
3053 EVP_MD_type. The old functionality is available in a new macro called
3054 EVP_MD_md(). Change code that uses it and update docs.
3057 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
3058 where the 'void *' argument is replaced by a function pointer argument.
3059 Previously 'void *' was abused to point to functions, which works on
3060 many platforms, but is not correct. As these functions are usually
3061 called by macros defined in OpenSSL header files, most source code
3062 should work without changes.
3065 *) <openssl/opensslconf.h> (which is created by Configure) now contains
3066 sections with information on -D... compiler switches used for
3067 compiling the library so that applications can see them. To enable
3068 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
3069 must be defined. E.g.,
3070 #define OPENSSL_ALGORITHM_DEFINES
3071 #include <openssl/opensslconf.h>
3072 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
3073 [Richard Levitte, Ulf and Bodo Möller]
3075 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
3079 *) Change the 'other' type in certificate aux info to a STACK_OF
3080 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
3081 the required ASN1 format: arbitrary types determined by an OID.
3084 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
3085 argument to 'req'. This is not because the function is newer or
3086 better than others it just uses the work 'NEW' in the certificate
3087 request header lines. Some software needs this.
3090 *) Reorganise password command line arguments: now passwords can be
3091 obtained from various sources. Delete the PEM_cb function and make
3092 it the default behaviour: i.e. if the callback is NULL and the
3093 usrdata argument is not NULL interpret it as a null terminated pass
3094 phrase. If usrdata and the callback are NULL then the pass phrase
3095 is prompted for as usual.
3098 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
3099 the support is automatically enabled. The resulting binaries will
3100 autodetect the card and use it if present.
3101 [Ben Laurie and Compaq Inc.]
3103 *) Work around for Netscape hang bug. This sends certificate request
3104 and server done in one record. Since this is perfectly legal in the
3105 SSL/TLS protocol it isn't a "bug" option and is on by default. See
3106 the bugs/SSLv3 entry for more info.
3109 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
3112 *) Add -rand argument to smime and pkcs12 applications and read/write
3116 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
3119 *) Add command line password options to the remaining applications.
3122 *) Bug fix for BN_div_recp() for numerators with an even number of
3126 *) More tests in bntest.c, and changed test_bn output.
3129 *) ./config recognizes MacOS X now.
3132 *) Bug fix for BN_div() when the first words of num and divsor are
3133 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
3136 *) Add support for various broken PKCS#8 formats, and command line
3137 options to produce them.
3140 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
3141 get temporary BIGNUMs from a BN_CTX.
3144 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
3148 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
3149 include a #define from the old name to the new. The original intent
3150 was that statically linked binaries could for example just call
3151 SSLeay_add_all_ciphers() to just add ciphers to the table and not
3152 link with digests. This never worked becayse SSLeay_add_all_digests()
3153 and SSLeay_add_all_ciphers() were in the same source file so calling
3154 one would link with the other. They are now in separate source files.
3157 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
3160 *) Use a less unusual form of the Miller-Rabin primality test (it used
3161 a binary algorithm for exponentiation integrated into the Miller-Rabin
3162 loop, our standard modexp algorithms are faster).
3165 *) Support for the EBCDIC character set completed.
3166 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
3168 *) Source code cleanups: use const where appropriate, eliminate casts,
3169 use void * instead of char * in lhash.
3172 *) Bugfix: ssl3_send_server_key_exchange was not restartable
3173 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
3174 this the server could overwrite ephemeral keys that the client
3178 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
3179 using 50 iterations of the Rabin-Miller test.
3181 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
3182 iterations of the Rabin-Miller test as required by the appendix
3183 to FIPS PUB 186[-1]) instead of DSA_is_prime.
3184 As BN_is_prime_fasttest includes trial division, DSA parameter
3185 generation becomes much faster.
3187 This implies a change for the callback functions in DSA_is_prime
3188 and DSA_generate_parameters: The callback function is called once
3189 for each positive witness in the Rabin-Miller test, not just
3190 occasionally in the inner loop; and the parameters to the
3191 callback function now provide an iteration count for the outer
3192 loop rather than for the current invocation of the inner loop.
3193 DSA_generate_parameters additionally can call the callback
3194 function with an 'iteration count' of -1, meaning that a
3195 candidate has passed the trial division test (when q is generated
3196 from an application-provided seed, trial division is skipped).
3199 *) New function BN_is_prime_fasttest that optionally does trial
3200 division before starting the Rabin-Miller test and has
3201 an additional BN_CTX * argument (whereas BN_is_prime always
3202 has to allocate at least one BN_CTX).
3203 'callback(1, -1, cb_arg)' is called when a number has passed the
3204 trial division stage.
3207 *) Fix for bug in CRL encoding. The validity dates weren't being handled
3211 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
3214 *) New function BN_pseudo_rand().
3217 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
3218 bignum version of BN_from_montgomery() with the working code from
3219 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
3223 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
3224 made it impossible to use the same SSL_SESSION data structure in
3225 SSL2 clients in multiple threads.
3228 *) The return value of RAND_load_file() no longer counts bytes obtained
3229 by stat(). RAND_load_file(..., -1) is new and uses the complete file
3230 to seed the PRNG (previously an explicit byte count was required).
3231 [Ulf Möller, Bodo Möller]
3233 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
3234 used (char *) instead of (void *) and had casts all over the place.
3237 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
3240 *) Retain source code compatibility for BN_prime_checks macro:
3241 BN_is_prime(..., BN_prime_checks, ...) now uses
3242 BN_prime_checks_for_size to determine the appropriate number of
3243 Rabin-Miller iterations.
3246 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
3247 DH_CHECK_P_NOT_SAFE_PRIME.
3248 (Check if this is true? OpenPGP calls them "strong".)
3251 *) Merge the functionality of "dh" and "gendh" programs into a new program
3252 "dhparam". The old programs are retained for now but will handle DH keys
3253 (instead of parameters) in future.
3256 *) Make the ciphers, s_server and s_client programs check the return values
3257 when a new cipher list is set.
3260 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
3261 ciphers. Before when the 56bit ciphers were enabled the sorting was
3264 The syntax for the cipher sorting has been extended to support sorting by
3265 cipher-strength (using the strength_bits hard coded in the tables).
3266 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
3268 Fix a bug in the cipher-command parser: when supplying a cipher command
3269 string with an "undefined" symbol (neither command nor alphanumeric
3270 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
3271 an error is flagged.
3273 Due to the strength-sorting extension, the code of the
3274 ssl_create_cipher_list() function was completely rearranged. I hope that
3275 the readability was also increased :-)
3276 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
3278 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
3279 for the first serial number and places 2 in the serial number file. This
3280 avoids problems when the root CA is created with serial number zero and
3281 the first user certificate has the same issuer name and serial number
3285 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
3286 the new code. Add documentation for this stuff.
3289 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
3290 X509_*() to X509at_*() on the grounds that they don't handle X509
3291 structures and behave in an analagous way to the X509v3 functions:
3292 they shouldn't be called directly but wrapper functions should be used
3295 So we also now have some wrapper functions that call the X509at functions
3296 when passed certificate requests. (TO DO: similar things can be done with
3297 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
3298 things. Some of these need some d2i or i2d and print functionality
3299 because they handle more complex structures.)
3302 *) Add missing #ifndefs that caused missing symbols when building libssl
3303 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
3304 NO_RSA in ssl/s2*.c.
3305 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
3307 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
3308 has a return value which indicates the quality of the random data
3309 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
3310 error queue. New function RAND_pseudo_bytes() generates output that is
3311 guaranteed to be unique but not unpredictable. RAND_add is like
3312 RAND_seed, but takes an extra argument for an entropy estimate
3313 (RAND_seed always assumes full entropy).
3316 *) Do more iterations of Rabin-Miller probable prime test (specifically,
3317 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
3318 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
3319 in crypto/bn/bn_prime.c for the complete table). This guarantees a
3320 false-positive rate of at most 2^-80 for random input.
3323 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
3326 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
3327 in the 0.9.5 release), this returns the chain
3328 from an X509_CTX structure with a dup of the stack and all
3329 the X509 reference counts upped: so the stack will exist
3330 after X509_CTX_cleanup() has been called. Modify pkcs12.c
3333 Also make SSL_SESSION_print() print out the verify return
3337 *) Add manpage for the pkcs12 command. Also change the default
3338 behaviour so MAC iteration counts are used unless the new
3339 -nomaciter option is used. This improves file security and
3340 only older versions of MSIE (4.0 for example) need it.
3343 *) Honor the no-xxx Configure options when creating .DEF files.
3346 *) Add PKCS#10 attributes to field table: challengePassword,
3347 unstructuredName and unstructuredAddress. These are taken from
3348 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
3349 international characters are used.
3351 More changes to X509_ATTRIBUTE code: allow the setting of types
3352 based on strings. Remove the 'loc' parameter when adding
3353 attributes because these will be a SET OF encoding which is sorted
3357 *) Initial changes to the 'req' utility to allow request generation
3358 automation. This will allow an application to just generate a template
3359 file containing all the field values and have req construct the
3362 Initial support for X509_ATTRIBUTE handling. Stacks of these are
3363 used all over the place including certificate requests and PKCS#7
3364 structures. They are currently handled manually where necessary with
3365 some primitive wrappers for PKCS#7. The new functions behave in a
3366 manner analogous to the X509 extension functions: they allow
3367 attributes to be looked up by NID and added.
3369 Later something similar to the X509V3 code would be desirable to
3370 automatically handle the encoding, decoding and printing of the
3371 more complex types. The string types like challengePassword can
3372 be handled by the string table functions.
3374 Also modified the multi byte string table handling. Now there is
3375 a 'global mask' which masks out certain types. The table itself
3376 can use the flag STABLE_NO_MASK to ignore the mask setting: this
3377 is useful when for example there is only one permissible type
3378 (as in countryName) and using the mask might result in no valid
3382 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
3383 SSL_get_peer_finished to allow applications to obtain the latest
3384 Finished messages sent to the peer or expected from the peer,
3385 respectively. (SSL_get_peer_finished is usually the Finished message
3386 actually received from the peer, otherwise the protocol will be aborted.)
3388 As the Finished message are message digests of the complete handshake
3389 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
3390 be used for external authentication procedures when the authentication
3391 provided by SSL/TLS is not desired or is not enough.
3394 *) Enhanced support for Alpha Linux is added. Now ./config checks if
3395 the host supports BWX extension and if Compaq C is present on the
3396 $PATH. Just exploiting of the BWX extension results in 20-30%
3397 performance kick for some algorithms, e.g. DES and RC4 to mention
3398 a couple. Compaq C in turn generates ~20% faster code for MD5 and
3402 *) Add support for MS "fast SGC". This is arguably a violation of the
3403 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
3404 weak crypto and after checking the certificate is SGC a second one
3405 with strong crypto. MS SGC stops the first handshake after receiving
3406 the server certificate message and sends a second client hello. Since
3407 a server will typically do all the time consuming operations before
3408 expecting any further messages from the client (server key exchange
3409 is the most expensive) there is little difference between the two.
3411 To get OpenSSL to support MS SGC we have to permit a second client
3412 hello message after we have sent server done. In addition we have to
3413 reset the MAC if we do get this second client hello.
3416 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
3417 if a DER encoded private key is RSA or DSA traditional format. Changed
3418 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
3419 format DER encoded private key. Newer code should use PKCS#8 format which
3420 has the key type encoded in the ASN1 structure. Added DER private key
3421 support to pkcs8 application.
3424 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
3425 ciphersuites has been selected (as required by the SSL 3/TLS 1
3426 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
3427 is set, we interpret this as a request to violate the specification
3428 (the worst that can happen is a handshake failure, and 'correct'
3429 behaviour would result in a handshake failure anyway).
3432 *) In SSL_CTX_add_session, take into account that there might be multiple
3433 SSL_SESSION structures with the same session ID (e.g. when two threads
3434 concurrently obtain them from an external cache).
3435 The internal cache can handle only one SSL_SESSION with a given ID,
3436 so if there's a conflict, we now throw out the old one to achieve
3440 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
3441 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
3442 some routines that use cipher OIDs: some ciphers do not have OIDs
3443 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
3447 *) Simplify the trust setting structure and code. Now we just have
3448 two sequences of OIDs for trusted and rejected settings. These will
3449 typically have values the same as the extended key usage extension
3450 and any application specific purposes.
3452 The trust checking code now has a default behaviour: it will just
3453 check for an object with the same NID as the passed id. Functions can
3454 be provided to override either the default behaviour or the behaviour
3455 for a given id. SSL client, server and email already have functions
3456 in place for compatibility: they check the NID and also return "trusted"
3457 if the certificate is self signed.
3460 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
3461 traditional format into an EVP_PKEY structure.
3464 *) Add a password callback function PEM_cb() which either prompts for
3465 a password if usr_data is NULL or otherwise assumes it is a null
3466 terminated password. Allow passwords to be passed on command line
3467 environment or config files in a few more utilities.
3470 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
3471 keys. Add some short names for PKCS#8 PBE algorithms and allow them
3472 to be specified on the command line for the pkcs8 and pkcs12 utilities.
3473 Update documentation.
3476 *) Support for ASN1 "NULL" type. This could be handled before by using
3477 ASN1_TYPE but there wasn't any function that would try to read a NULL
3478 and produce an error if it couldn't. For compatibility we also have
3479 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
3480 don't allocate anything because they don't need to.
3483 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
3485 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
3487 *) Rebuild of the memory allocation routines used by OpenSSL code and
3488 possibly others as well. The purpose is to make an interface that
3489 provide hooks so anyone can build a separate set of allocation and
3490 deallocation routines to be used by OpenSSL, for example memory
3491 pool implementations, or something else, which was previously hard
3492 since Malloc(), Realloc() and Free() were defined as macros having
3493 the values malloc, realloc and free, respectively (except for Win32
3494 compilations). The same is provided for memory debugging code.
3495 OpenSSL already comes with functionality to find memory leaks, but
3496 this gives people a chance to debug other memory problems.
3498 With these changes, a new set of functions and macros have appeared:
3500 CRYPTO_set_mem_debug_functions() [F]
3501 CRYPTO_get_mem_debug_functions() [F]
3502 CRYPTO_dbg_set_options() [F]
3503 CRYPTO_dbg_get_options() [F]
3504 CRYPTO_malloc_debug_init() [M]
3506 The memory debug functions are NULL by default, unless the library
3507 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
3508 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
3509 gives the standard debugging functions that come with OpenSSL) or
3510 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
3511 provided by the library user) must be used. When the standard
3512 debugging functions are used, CRYPTO_dbg_set_options can be used to
3513 request additional information:
3514 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
3515 the CRYPTO_MDEBUG_xxx macro when compiling the library.
3517 Also, things like CRYPTO_set_mem_functions will always give the
3518 expected result (the new set of functions is used for allocation
3519 and deallocation) at all times, regardless of platform and compiler
3522 To finish it up, some functions that were never use in any other
3523 way than through macros have a new API and new semantic:
3526 CRYPTO_dbg_realloc()
3529 All macros of value have retained their old syntax.
3530 [Richard Levitte and Bodo Moeller]
3532 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
3533 ordering of SMIMECapabilities wasn't in "strength order" and there
3534 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
3538 *) Some ASN1 types with illegal zero length encoding (INTEGER,
3539 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
3540 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
3542 *) Merge in my S/MIME library for OpenSSL. This provides a simple
3543 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
3544 functionality to handle multipart/signed properly) and a utility
3545 called 'smime' to call all this stuff. This is based on code I
3546 originally wrote for Celo who have kindly allowed it to be
3547 included in OpenSSL.
3550 *) Add variants des_set_key_checked and des_set_key_unchecked of
3551 des_set_key (aka des_key_sched). Global variable des_check_key
3552 decides which of these is called by des_set_key; this way
3553 des_check_key behaves as it always did, but applications and
3554 the library itself, which was buggy for des_check_key == 1,
3555 have a cleaner way to pick the version they need.
3558 *) New function PKCS12_newpass() which changes the password of a
3562 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
3563 dynamic mix. In both cases the ids can be used as an index into the
3564 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
3565 functions so they accept a list of the field values and the
3566 application doesn't need to directly manipulate the X509_TRUST
3570 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
3574 *) Modify the way the V3 extension code looks up extensions. This now
3575 works in a similar way to the object code: we have some "standard"
3576 extensions in a static table which is searched with OBJ_bsearch()
3577 and the application can add dynamic ones if needed. The file
3578 crypto/x509v3/ext_dat.h now has the info: this file needs to be
3579 updated whenever a new extension is added to the core code and kept
3580 in ext_nid order. There is a simple program 'tabtest.c' which checks
3581 this. New extensions are not added too often so this file can readily
3582 be maintained manually.
3584 There are two big advantages in doing things this way. The extensions
3585 can be looked up immediately and no longer need to be "added" using
3586 X509V3_add_standard_extensions(): this function now does nothing.
3587 [Side note: I get *lots* of email saying the extension code doesn't
3588 work because people forget to call this function]
3589 Also no dynamic allocation is done unless new extensions are added:
3590 so if we don't add custom extensions there is no need to call
3591 X509V3_EXT_cleanup().
3594 *) Modify enc utility's salting as follows: make salting the default. Add a
3595 magic header, so unsalted files fail gracefully instead of just decrypting
3596 to garbage. This is because not salting is a big security hole, so people
3597 should be discouraged from doing it.
3600 *) Fixes and enhancements to the 'x509' utility. It allowed a message
3601 digest to be passed on the command line but it only used this
3602 parameter when signing a certificate. Modified so all relevant
3603 operations are affected by the digest parameter including the
3604 -fingerprint and -x509toreq options. Also -x509toreq choked if a
3605 DSA key was used because it didn't fix the digest.
3608 *) Initial certificate chain verify code. Currently tests the untrusted
3609 certificates for consistency with the verify purpose (which is set
3610 when the X509_STORE_CTX structure is set up) and checks the pathlength.
3612 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
3613 this is because it will reject chains with invalid extensions whereas
3614 every previous version of OpenSSL and SSLeay made no checks at all.
3616 Trust code: checks the root CA for the relevant trust settings. Trust
3617 settings have an initial value consistent with the verify purpose: e.g.
3618 if the verify purpose is for SSL client use it expects the CA to be
3619 trusted for SSL client use. However the default value can be changed to
3620 permit custom trust settings: one example of this would be to only trust
3621 certificates from a specific "secure" set of CAs.
3623 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
3624 which should be used for version portability: especially since the
3625 verify structure is likely to change more often now.
3627 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
3628 to set them. If not set then assume SSL clients will verify SSL servers
3631 Two new options to the verify program: -untrusted allows a set of
3632 untrusted certificates to be passed in and -purpose which sets the
3633 intended purpose of the certificate. If a purpose is set then the
3634 new chain verify code is used to check extension consistency.
3637 *) Support for the authority information access extension.
3640 *) Modify RSA and DSA PEM read routines to transparently handle
3641 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
3642 public keys in a format compatible with certificate
3643 SubjectPublicKeyInfo structures. Unfortunately there were already
3644 functions called *_PublicKey_* which used various odd formats so
3645 these are retained for compatibility: however the DSA variants were
3646 never in a public release so they have been deleted. Changed dsa/rsa
3647 utilities to handle the new format: note no releases ever handled public
3648 keys so we should be OK.
3650 The primary motivation for this change is to avoid the same fiasco
3651 that dogs private keys: there are several incompatible private key
3652 formats some of which are standard and some OpenSSL specific and
3653 require various evil hacks to allow partial transparent handling and
3654 even then it doesn't work with DER formats. Given the option anything
3655 other than PKCS#8 should be dumped: but the other formats have to
3656 stay in the name of compatibility.
3658 With public keys and the benefit of hindsight one standard format
3659 is used which works with EVP_PKEY, RSA or DSA structures: though
3660 it clearly returns an error if you try to read the wrong kind of key.
3662 Added a -pubkey option to the 'x509' utility to output the public key.
3663 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
3664 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
projects / openssl.git / blob