5 Changes between 0.9.8za and 0.9.8zb [xx XXX xxxx]
7 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
8 for corner cases. (Certain input points at infinity could lead to
9 bogus results, with non-infinity inputs mapped to infinity too.)
12 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
14 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
15 handshake can force the use of weak keying material in OpenSSL
16 SSL/TLS clients and servers.
18 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
19 researching this issue. (CVE-2014-0224)
20 [KIKUCHI Masashi, Steve Henson]
22 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
23 OpenSSL DTLS client the code can be made to recurse eventually crashing
26 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
28 [Imre Rad, Steve Henson]
30 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
31 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
32 client or server. This is potentially exploitable to run arbitrary
33 code on a vulnerable client or server.
35 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
36 [Jüri Aedla, Steve Henson]
38 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
39 are subject to a denial of service attack.
41 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
42 this issue. (CVE-2014-3470)
43 [Felix Gröbert, Ivan Fratric, Steve Henson]
45 *) Fix for the attack described in the paper "Recovering OpenSSL
46 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
47 by Yuval Yarom and Naomi Benger. Details can be obtained from:
48 http://eprint.iacr.org/2014/140
50 Thanks to Yuval Yarom and Naomi Benger for discovering this
51 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
52 [Yuval Yarom and Naomi Benger]
54 Thanks to mancha for backporting the fix to the 0.9.8 branch.
56 *) Fix handling of warning-level alerts in SSL23 client mode so they
57 don't cause client-side termination (eg. on SNI unrecognized_name
58 warnings). Add client and server support for six additional alerts
59 per RFC 6066 and RFC 4279.
62 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
63 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
64 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
65 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
66 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
67 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
68 [Rob Stradling, Adam Langley]
70 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
72 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
74 This addresses the flaw in CBC record processing discovered by
75 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
76 at: http://www.isg.rhul.ac.uk/tls/
78 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
79 Security Group at Royal Holloway, University of London
80 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
81 Emilia Käsper for the initial patch.
83 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
85 *) Return an error when checking OCSP signatures when key is NULL.
86 This fixes a DoS attack. (CVE-2013-0166)
89 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
90 the right response is stapled. Also change SSL_get_certificate()
91 so it returns the certificate actually sent.
92 See http://rt.openssl.org/Ticket/Display.html?id=2836.
94 [Rob Stradling <rob.stradling@comodo.com>]
96 *) Fix possible deadlock when decoding public keys.
99 Changes between 0.9.8w and 0.9.8x [10 May 2012]
101 *) Sanity check record length before skipping explicit IV in DTLS
104 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
105 fuzzing as a service testing platform.
109 *) Initialise tkeylen properly when encrypting CMS messages.
110 Thanks to Solar Designer of Openwall for reporting this issue.
113 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
115 *) The fix for CVE-2012-2110 did not take into account that the
116 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
117 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
118 rejecting negative len parameter. (CVE-2012-2131)
119 [Tomas Hoger <thoger@redhat.com>]
121 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
123 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
124 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
125 in CRYPTO_realloc_clean.
127 Thanks to Tavis Ormandy, Google Security Team, for discovering this
128 issue and to Adam Langley <agl@chromium.org> for fixing it.
130 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
132 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
134 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
135 in CMS and PKCS7 code. When RSA decryption fails use a random key for
136 content decryption and always return the same error. Note: this attack
137 needs on average 2^20 messages so it only affects automated senders. The
138 old behaviour can be reenabled in the CMS code by setting the
139 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
140 an MMA defence is not necessary.
141 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
142 this issue. (CVE-2012-0884)
145 *) Fix CVE-2011-4619: make sure we really are receiving a
146 client hello before rejecting multiple SGC restarts. Thanks to
147 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
150 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
152 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
153 Thanks to Antonio Martin, Enterprise Secure Access Research and
154 Development, Cisco Systems, Inc. for discovering this bug and
155 preparing a fix. (CVE-2012-0050)
158 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
160 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
161 of the Vaudenay padding oracle attack on CBC mode encryption
162 which enables an efficient plaintext recovery attack against
163 the OpenSSL implementation of DTLS. Their attack exploits timing
164 differences arising during decryption processing. A research
165 paper describing this attack can be found at:
166 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
167 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
168 Security Group at Royal Holloway, University of London
169 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
170 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
171 for preparing the fix. (CVE-2011-4108)
172 [Robin Seggelmann, Michael Tuexen]
174 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
175 [Ben Laurie, Kasper <ekasper@google.com>]
177 *) Clear bytes used for block padding of SSL 3.0 records.
179 [Adam Langley (Google)]
181 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
182 Kadianakis <desnacked@gmail.com> for discovering this issue and
183 Adam Langley for preparing the fix. (CVE-2011-4619)
184 [Adam Langley (Google)]
186 *) Prevent malformed RFC3779 data triggering an assertion failure.
187 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
188 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
189 [Rob Austein <sra@hactrn.net>]
191 *) Fix ssl_ciph.c set-up race.
192 [Adam Langley (Google)]
194 *) Fix spurious failures in ecdsatest.c.
195 [Emilia Käsper (Google)]
197 *) Fix the BIO_f_buffer() implementation (which was mixing different
198 interpretations of the '..._len' fields).
199 [Adam Langley (Google)]
201 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
202 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
203 threads won't reuse the same blinding coefficients.
205 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
206 lock to call BN_BLINDING_invert_ex, and avoids one use of
207 BN_BLINDING_update for each BN_BLINDING structure (previously,
208 the last update always remained unused).
209 [Emilia Käsper (Google)]
211 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
212 for multi-threaded use of ECDH.
213 [Adam Langley (Google)]
215 *) Fix x509_name_ex_d2i memory leak on bad inputs.
218 *) Add protection against ECDSA timing attacks as mentioned in the paper
219 by Billy Bob Brumley and Nicola Tuveri, see:
221 http://eprint.iacr.org/2011/232.pdf
223 [Billy Bob Brumley and Nicola Tuveri]
225 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
227 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
228 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
230 *) Fix bug in string printing code: if *any* escaping is enabled we must
231 escape the escape character (backslash) or the resulting string is
235 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
237 *) Disable code workaround for ancient and obsolete Netscape browsers
238 and servers: an attacker can use it in a ciphersuite downgrade attack.
239 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
242 *) Fixed J-PAKE implementation error, originally discovered by
243 Sebastien Martini, further info and confirmation from Stefan
244 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
247 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
249 *) Fix extension code to avoid race conditions which can result in a buffer
250 overrun vulnerability: resumed sessions must not be modified as they can
251 be shared by multiple threads. CVE-2010-3864
254 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
257 *) Don't reencode certificate when calculating signature: cache and use
258 the original encoding instead. This makes signature verification of
259 some broken encodings work correctly.
262 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
263 is also one of the inputs.
264 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
266 *) Don't repeatedly append PBE algorithms to table if they already exist.
267 Sort table on each new add. This effectively makes the table read only
268 after all algorithms are added and subsequent calls to PKCS12_pbe_add
272 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
274 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
277 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
278 access or freeing data twice (CVE-2010-0742)
279 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
281 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
282 common in certificates and some applications which only call
283 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
287 Reduce copying into .apps and .test in makevms.com
288 Don't try to use blank CA certificate in CA.com
289 Allow use of C files from original directories in maketests.com
290 [Steven M. Schweda" <sms@antinode.info>]
292 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
294 *) When rejecting SSL/TLS records due to an incorrect version number, never
295 update s->server with a new major version number. As of
296 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
297 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
298 the previous behavior could result in a read attempt at NULL when
299 receiving specific incorrect SSL/TLS records once record payload
300 protection is active. (CVE-2010-0740)
301 [Bodo Moeller, Adam Langley <agl@chromium.org>]
303 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
304 could be crashed if the relevant tables were not present (e.g. chrooted).
305 [Tomas Hoger <thoger@redhat.com>]
307 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
309 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
310 [Martin Olsson, Neel Mehta]
312 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
313 accommodate for stack sorting, always a write lock!).
316 *) On some versions of WIN32 Heap32Next is very slow. This can cause
317 excessive delays in the RAND_poll(): over a minute. As a workaround
318 include a time check in the inner Heap32Next loop too.
321 *) The code that handled flushing of data in SSL/TLS originally used the
322 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
323 the problem outlined in PR#1949. The fix suggested there however can
324 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
325 of Apache). So instead simplify the code to flush unconditionally.
326 This should be fine since flushing with no data to flush is a no op.
329 *) Handle TLS versions 2.0 and later properly and correctly use the
330 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
331 off ancient servers have a habit of sticking around for a while...
334 *) Modify compression code so it frees up structures without using the
335 ex_data callbacks. This works around a problem where some applications
336 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
337 restarting) then use compression (e.g. SSL with compression) later.
338 This results in significant per-connection memory leaks and
339 has caused some security issues including CVE-2008-1678 and
343 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
344 change when encrypting or decrypting.
347 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
348 connect and renegotiate with servers which do not support RI.
349 Until RI is more widely deployed this option is enabled by default.
352 *) Add "missing" ssl ctrls to clear options and mode.
355 *) If client attempts to renegotiate and doesn't support RI respond with
356 a no_renegotiation alert as required by RFC5746. Some renegotiating
357 TLS clients will continue a connection gracefully when they receive
358 the alert. Unfortunately OpenSSL mishandled this alert and would hang
359 waiting for a server hello which it will never receive. Now we treat a
360 received no_renegotiation alert as a fatal error. This is because
361 applications requesting a renegotiation might well expect it to succeed
362 and would have no code in place to handle the server denying it so the
363 only safe thing to do is to terminate the connection.
366 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
367 peer supports secure renegotiation and 0 otherwise. Print out peer
368 renegotiation support in s_client/s_server.
371 *) Replace the highly broken and deprecated SPKAC certification method with
372 the updated NID creation version. This should correctly handle UTF8.
375 *) Implement RFC5746. Re-enable renegotiation but require the extension
376 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
377 turns out to be a bad idea. It has been replaced by
378 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
379 SSL_CTX_set_options(). This is really not recommended unless you
380 know what you are doing.
381 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
383 *) Fixes to stateless session resumption handling. Use initial_ctx when
384 issuing and attempting to decrypt tickets in case it has changed during
385 servername handling. Use a non-zero length session ID when attempting
386 stateless session resumption: this makes it possible to determine if
387 a resumption has occurred immediately after receiving server hello
388 (several places in OpenSSL subtly assume this) instead of later in
392 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
393 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
394 fixes for a few places where the return code is not checked
396 [Julia Lawall <julia@diku.dk>]
398 *) Add --strict-warnings option to Configure script to include devteam
399 warnings in other configurations.
402 *) Add support for --libdir option and LIBDIR variable in makefiles. This
403 makes it possible to install openssl libraries in locations which
404 have names other than "lib", for example "/usr/lib64" which some
406 [Steve Henson, based on patch from Jeremy Utley]
408 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
409 X690 8.9.12 and can produce some misleading textual output of OIDs.
410 [Steve Henson, reported by Dan Kaminsky]
412 *) Delete MD2 from algorithm tables. This follows the recommendation in
413 several standards that it is not used in new applications due to
414 several cryptographic weaknesses. For binary compatibility reasons
415 the MD2 API is still compiled in by default.
418 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
422 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
423 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
425 [Guenter <lists@gknw.net>]
427 *) Fix the server certificate chain building code to use X509_verify_cert(),
428 it used to have an ad-hoc builder which was unable to cope with anything
429 other than a simple chain.
430 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
432 *) Don't check self signed certificate signatures in X509_verify_cert()
433 by default (a flag can override this): it just wastes time without
434 adding any security. As a useful side effect self signed root CAs
435 with non-FIPS digests are now usable in FIPS mode.
438 *) In dtls1_process_out_of_seq_message() the check if the current message
439 is already buffered was missing. For every new message was memory
440 allocated, allowing an attacker to perform an denial of service attack
441 with sending out of seq handshake messages until there is no memory
442 left. Additionally every future messege was buffered, even if the
443 sequence number made no sense and would be part of another handshake.
444 So only messages with sequence numbers less than 10 in advance will be
445 buffered. (CVE-2009-1378)
446 [Robin Seggelmann, discovered by Daniel Mentz]
448 *) Records are buffered if they arrive with a future epoch to be
449 processed after finishing the corresponding handshake. There is
450 currently no limitation to this buffer allowing an attacker to perform
451 a DOS attack with sending records with future epochs until there is no
452 memory left. This patch adds the pqueue_size() function to detemine
453 the size of a buffer and limits the record buffer to 100 entries.
455 [Robin Seggelmann, discovered by Daniel Mentz]
457 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
458 parent structure is freed. (CVE-2009-1379)
461 *) Handle non-blocking I/O properly in SSL_shutdown() call.
462 [Darryl Miles <darryl-mailinglists@netbauds.net>]
465 [Ilya O. <vrghost@gmail.com>]
467 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
469 *) Disable renegotiation completely - this fixes a severe security
470 problem (CVE-2009-3555) at the cost of breaking all
471 renegotiation. Renegotiation can be re-enabled by setting
472 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
473 run-time. This is really not recommended unless you know what
477 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
479 *) Don't set val to NULL when freeing up structures, it is freed up by
480 underlying code. If sizeof(void *) > sizeof(long) this can result in
481 zeroing past the valid field. (CVE-2009-0789)
482 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
484 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
485 checked correctly. This would allow some invalid signed attributes to
486 appear to verify correctly. (CVE-2009-0591)
487 [Ivan Nestlerode <inestlerode@us.ibm.com>]
489 *) Reject UniversalString and BMPString types with invalid lengths. This
490 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
491 a legal length. (CVE-2009-0590)
494 *) Set S/MIME signing as the default purpose rather than setting it
495 unconditionally. This allows applications to override it at the store
499 *) Permit restricted recursion of ASN1 strings. This is needed in practice
500 to handle some structures.
503 *) Improve efficiency of mem_gets: don't search whole buffer each time
505 [Jeremy Shapiro <jnshapir@us.ibm.com>]
507 *) New -hex option for openssl rand.
510 *) Print out UTF8String and NumericString when parsing ASN1.
513 *) Support NumericString type for name components.
516 *) Allow CC in the environment to override the automatically chosen
517 compiler. Note that nothing is done to ensure flags work with the
521 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
523 *) Properly check EVP_VerifyFinal() and similar return values
525 [Ben Laurie, Bodo Moeller, Google Security Team]
527 *) Enable TLS extensions by default.
530 *) Allow the CHIL engine to be loaded, whether the application is
531 multithreaded or not. (This does not release the developer from the
532 obligation to set up the dynamic locking callbacks.)
533 [Sander Temme <sander@temme.net>]
535 *) Use correct exit code if there is an error in dgst command.
536 [Steve Henson; problem pointed out by Roland Dirlewanger]
538 *) Tweak Configure so that you need to say "experimental-jpake" to enable
539 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
542 *) Add experimental JPAKE support, including demo authentication in
543 s_client and s_server.
546 *) Set the comparison function in v3_addr_canonize().
547 [Rob Austein <sra@hactrn.net>]
549 *) Add support for XMPP STARTTLS in s_client.
550 [Philip Paeps <philip@freebsd.org>]
552 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
553 to ensure that even with this option, only ciphersuites in the
554 server's preference list will be accepted. (Note that the option
555 applies only when resuming a session, so the earlier behavior was
556 just about the algorithm choice for symmetric cryptography.)
559 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
561 *) Fix NULL pointer dereference if a DTLS server received
562 ChangeCipherSpec as first record (CVE-2009-1386).
565 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
566 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
569 *) The fix in 0.9.8c that supposedly got rid of unsafe
570 double-checked locking was incomplete for RSA blinding,
571 addressing just one layer of what turns out to have been
572 doubly unsafe triple-checked locking.
574 So now fix this for real by retiring the MONT_HELPER macro
575 in crypto/rsa/rsa_eay.c.
577 [Bodo Moeller; problem pointed out by Marius Schilder]
579 *) Various precautionary measures:
581 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
583 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
584 (NB: This would require knowledge of the secret session ticket key
585 to exploit, in which case you'd be SOL either way.)
587 - Change bn_nist.c so that it will properly handle input BIGNUMs
588 outside the expected range.
590 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
593 [Neel Mehta, Bodo Moeller]
595 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
596 the load fails. Useful for distros.
597 [Ben Laurie and the FreeBSD team]
599 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
602 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
605 *) Expand ENGINE to support engine supplied SSL client certificate functions.
607 This work was sponsored by Logica.
610 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
611 keystores. Support for SSL/TLS client authentication too.
612 Not compiled unless enable-capieng specified to Configure.
614 This work was sponsored by Logica.
617 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
618 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
619 attribute creation routines such as certifcate requests and PKCS#12
623 Changes between 0.9.8g and 0.9.8h [28 May 2008]
625 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
626 handshake which could lead to a cilent crash as found using the
627 Codenomicon TLS test suite (CVE-2008-1672)
628 [Steve Henson, Mark Cox]
630 *) Fix double free in TLS server name extensions which could lead to
631 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
634 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
636 Clear the error queue to ensure that error entries left from
637 older function calls do not interfere with the correct operation.
638 [Lutz Jaenicke, Erik de Castro Lopo]
640 *) Remove root CA certificates of commercial CAs:
642 The OpenSSL project does not recommend any specific CA and does not
643 have any policy with respect to including or excluding any CA.
644 Therefore it does not make any sense to ship an arbitrary selection
645 of root CA certificates with the OpenSSL software.
648 *) RSA OAEP patches to fix two separate invalid memory reads.
649 The first one involves inputs when 'lzero' is greater than
650 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
651 before the beginning of from). The second one involves inputs where
652 the 'db' section contains nothing but zeroes (there is a one-byte
653 invalid read after the end of 'db').
654 [Ivan Nestlerode <inestlerode@us.ibm.com>]
656 *) Partial backport from 0.9.9-dev:
658 Introduce bn_mul_mont (dedicated Montgomery multiplication
659 procedure) as a candidate for BIGNUM assembler implementation.
660 While 0.9.9-dev uses assembler for various architectures, only
661 x86_64 is available by default here in the 0.9.8 branch, and
662 32-bit x86 is available through a compile-time setting.
664 To try the 32-bit x86 assembler implementation, use Configure
665 option "enable-montasm" (which exists only for this backport).
667 As "enable-montasm" for 32-bit x86 disclaims code stability
668 anyway, in this constellation we activate additional code
669 backported from 0.9.9-dev for further performance improvements,
670 namely BN_from_montgomery_word. (To enable this otherwise,
671 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
673 [Andy Polyakov (backport partially by Bodo Moeller)]
675 *) Add TLS session ticket callback. This allows an application to set
676 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
677 values. This is useful for key rollover for example where several key
678 sets may exist with different names.
681 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
682 This was broken until now in 0.9.8 releases, such that the only way
683 a registered ENGINE could be used (assuming it initialises
684 successfully on the host) was to explicitly set it as the default
685 for the relevant algorithms. This is in contradiction with 0.9.7
686 behaviour and the documentation. With this fix, when an ENGINE is
687 registered into a given algorithm's table of implementations, the
688 'uptodate' flag is reset so that auto-discovery will be used next
689 time a new context for that algorithm attempts to select an
691 [Ian Lister (tweaked by Geoff Thorpe)]
693 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
694 implemention in the following ways:
696 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
699 Lack of BER streaming support means one pass streaming processing is
700 only supported if data is detached: setting the streaming flag is
701 ignored for embedded content.
703 CMS support is disabled by default and must be explicitly enabled
704 with the enable-cms configuration option.
707 *) Update the GMP engine glue to do direct copies between BIGNUM and
708 mpz_t when openssl and GMP use the same limb size. Otherwise the
709 existing "conversion via a text string export" trick is still used.
710 [Paul Sheer <paulsheer@gmail.com>]
712 *) Zlib compression BIO. This is a filter BIO which compressed and
713 uncompresses any data passed through it.
716 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
717 RFC3394 compatible AES key wrapping.
720 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
721 sets string data without copying. X509_ALGOR_set0() and
722 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
723 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
724 from an X509_ATTRIBUTE structure optionally checking it occurs only
725 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
729 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
730 to get the expected BN_FLG_CONSTTIME behavior.
731 [Bodo Moeller (Google)]
735 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
736 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
737 - added some more tests to do_tests.pl
738 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
739 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
740 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
741 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
742 - various changes to netware.pl to enable gcc-cross builds on Win32
744 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
745 - various changes to fix missing prototype warnings
746 - fixed x86nasm.pl to create correct asm files for NASM COFF output
747 - added AES, WHIRLPOOL and CPUID assembler code to build files
748 - added missing AES assembler make rules to mk1mf.pl
749 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
750 [Guenter Knauf <eflash@gmx.net>]
752 *) Implement certificate status request TLS extension defined in RFC3546.
753 A client can set the appropriate parameters and receive the encoded
754 OCSP response via a callback. A server can query the supplied parameters
755 and set the encoded OCSP response in the callback. Add simplified examples
756 to s_client and s_server.
759 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
762 + Binary incompatibility of ssl_ctx_st structure
763 + DTLS interoperation with non-compliant servers
764 + Don't call get_session_cb() without proposed session
765 + Fix ia64 assembler code
766 [Andy Polyakov, Steve Henson]
768 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
770 *) DTLS Handshake overhaul. There were longstanding issues with
771 OpenSSL DTLS implementation, which were making it impossible for
772 RFC 4347 compliant client to communicate with OpenSSL server.
773 Unfortunately just fixing these incompatibilities would "cut off"
774 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
775 server keeps tolerating non RFC compliant syntax. The opposite is
776 not true, 0.9.8f client can not communicate with earlier server.
777 This update even addresses CVE-2007-4995.
780 *) Changes to avoid need for function casts in OpenSSL: some compilers
781 (gcc 4.2 and later) reject their use.
782 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
785 *) Add RFC4507 support to OpenSSL. This includes the corrections in
786 RFC4507bis. The encrypted ticket format is an encrypted encoded
787 SSL_SESSION structure, that way new session features are automatically
790 If a client application caches session in an SSL_SESSION structure
791 support is transparent because tickets are now stored in the encoded
794 The SSL_CTX structure automatically generates keys for ticket
795 protection in servers so again support should be possible
796 with no application modification.
798 If a client or server wishes to disable RFC4507 support then the option
799 SSL_OP_NO_TICKET can be set.
801 Add a TLS extension debugging callback to allow the contents of any client
802 or server extensions to be examined.
804 This work was sponsored by Google.
807 *) Add initial support for TLS extensions, specifically for the server_name
808 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
809 have new members for a host name. The SSL data structure has an
810 additional member SSL_CTX *initial_ctx so that new sessions can be
811 stored in that context to allow for session resumption, even after the
812 SSL has been switched to a new SSL_CTX in reaction to a client's
813 server_name extension.
815 New functions (subject to change):
818 SSL_get_servername_type()
821 New CTRL codes and macros (subject to change):
823 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
824 - SSL_CTX_set_tlsext_servername_callback()
825 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
826 - SSL_CTX_set_tlsext_servername_arg()
827 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
829 openssl s_client has a new '-servername ...' option.
831 openssl s_server has new options '-servername_host ...', '-cert2 ...',
832 '-key2 ...', '-servername_fatal' (subject to change). This allows
833 testing the HostName extension for a specific single host name ('-cert'
834 and '-key' remain fallbacks for handshakes without HostName
835 negotiation). If the unrecogninzed_name alert has to be sent, this by
836 default is a warning; it becomes fatal with the '-servername_fatal'
839 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
841 *) Add AES and SSE2 assembly language support to VC++ build.
844 *) Mitigate attack on final subtraction in Montgomery reduction.
847 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
848 (which previously caused an internal error).
851 *) Squeeze another 10% out of IGE mode when in != out.
854 *) AES IGE mode speedup.
855 [Dean Gaudet (Google)]
857 *) Add the Korean symmetric 128-bit cipher SEED (see
858 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
859 add SEED ciphersuites from RFC 4162:
861 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
862 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
863 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
864 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
866 To minimize changes between patchlevels in the OpenSSL 0.9.8
867 series, SEED remains excluded from compilation unless OpenSSL
868 is configured with 'enable-seed'.
871 *) Mitigate branch prediction attacks, which can be practical if a
872 single processor is shared, allowing a spy process to extract
873 information. For detailed background information, see
874 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
875 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
876 and Necessary Software Countermeasures"). The core of the change
877 are new versions BN_div_no_branch() and
878 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
879 respectively, which are slower, but avoid the security-relevant
880 conditional branches. These are automatically called by BN_div()
881 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
882 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
883 remove a conditional branch.
885 BN_FLG_CONSTTIME is the new name for the previous
886 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
887 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
888 in the exponent causes BN_mod_exp_mont() to use the alternative
889 implementation in BN_mod_exp_mont_consttime().) The old name
890 remains as a deprecated alias.
892 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
893 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
894 constant-time implementations for more than just exponentiation.
895 Here too the old name is kept as a deprecated alias.
897 BN_BLINDING_new() will now use BN_dup() for the modulus so that
898 the BN_BLINDING structure gets an independent copy of the
899 modulus. This means that the previous "BIGNUM *m" argument to
900 BN_BLINDING_new() and to BN_BLINDING_create_param() now
901 essentially becomes "const BIGNUM *m", although we can't actually
902 change this in the header file before 0.9.9. It allows
903 RSA_setup_blinding() to use BN_with_flags() on the modulus to
904 enable BN_FLG_CONSTTIME.
906 [Matthew D Wood (Intel Corp)]
908 *) In the SSL/TLS server implementation, be strict about session ID
909 context matching (which matters if an application uses a single
910 external cache for different purposes). Previously,
911 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
912 set. This did ensure strict client verification, but meant that,
913 with applications using a single external cache for quite
914 different requirements, clients could circumvent ciphersuite
915 restrictions for a given session ID context by starting a session
916 in a different context.
919 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
920 a ciphersuite string such as "DEFAULT:RSA" cannot enable
921 authentication-only ciphersuites.
924 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
925 not complete and could lead to a possible single byte overflow
926 (CVE-2007-5135) [Ben Laurie]
928 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
930 *) Since AES128 and AES256 (and similarly Camellia128 and
931 Camellia256) share a single mask bit in the logic of
932 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
933 kludge to work properly if AES128 is available and AES256 isn't
934 (or if Camellia128 is available and Camellia256 isn't).
937 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
938 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
939 When a point or a seed is encoded in a BIT STRING, we need to
940 prevent the removal of trailing zero bits to get the proper DER
941 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
942 of a NamedBitList, for which trailing 0 bits need to be removed.)
945 *) Have SSL/TLS server implementation tolerate "mismatched" record
946 protocol version while receiving ClientHello even if the
947 ClientHello is fragmented. (The server can't insist on the
948 particular protocol version it has chosen before the ServerHello
949 message has informed the client about his choice.)
952 *) Add RFC 3779 support.
953 [Rob Austein for ARIN, Ben Laurie]
955 *) Load error codes if they are not already present instead of using a
956 static variable. This allows them to be cleanly unloaded and reloaded.
957 Improve header file function name parsing.
960 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
961 or CAPABILITY handshake as required by RFCs.
964 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
966 *) Introduce limits to prevent malicious keys being able to
967 cause a denial of service. (CVE-2006-2940)
968 [Steve Henson, Bodo Moeller]
970 *) Fix ASN.1 parsing of certain invalid structures that can result
971 in a denial of service. (CVE-2006-2937) [Steve Henson]
973 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
974 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
976 *) Fix SSL client code which could crash if connecting to a
977 malicious SSLv2 server. (CVE-2006-4343)
978 [Tavis Ormandy and Will Drewry, Google Security Team]
980 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
981 match only those. Before that, "AES256-SHA" would be interpreted
982 as a pattern and match "AES128-SHA" too (since AES128-SHA got
983 the same strength classification in 0.9.7h) as we currently only
984 have a single AES bit in the ciphersuite description bitmap.
985 That change, however, also applied to ciphersuite strings such as
986 "RC4-MD5" that intentionally matched multiple ciphersuites --
987 namely, SSL 2.0 ciphersuites in addition to the more common ones
988 from SSL 3.0/TLS 1.0.
990 So we change the selection algorithm again: Naming an explicit
991 ciphersuite selects this one ciphersuite, and any other similar
992 ciphersuite (same bitmap) from *other* protocol versions.
993 Thus, "RC4-MD5" again will properly select both the SSL 2.0
994 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
996 Since SSL 2.0 does not have any ciphersuites for which the
997 128/256 bit distinction would be relevant, this works for now.
998 The proper fix will be to use different bits for AES128 and
999 AES256, which would have avoided the problems from the beginning;
1000 however, bits are scarce, so we can only do this in a new release
1001 (not just a patchlevel) when we can change the SSL_CIPHER
1002 definition to split the single 'unsigned long mask' bitmap into
1003 multiple values to extend the available space.
1007 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1009 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1010 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1012 *) Add AES IGE and biIGE modes.
1015 *) Change the Unix randomness entropy gathering to use poll() when
1016 possible instead of select(), since the latter has some
1017 undesirable limitations.
1018 [Darryl Miles via Richard Levitte and Bodo Moeller]
1020 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1021 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1022 cannot be implicitly activated as part of, e.g., the "AES" alias.
1023 However, please upgrade to OpenSSL 0.9.9[-dev] for
1024 non-experimental use of the ECC ciphersuites to get TLS extension
1025 support, which is required for curve and point format negotiation
1026 to avoid potential handshake problems.
1029 *) Disable rogue ciphersuites:
1031 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1032 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1033 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1035 The latter two were purportedly from
1036 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1039 Also deactivate the remaining ciphersuites from
1040 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1041 unofficial, and the ID has long expired.
1044 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1045 dual-core machines) and other potential thread-safety issues.
1048 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1049 versions), which is now available for royalty-free use
1050 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1051 Also, add Camellia TLS ciphersuites from RFC 4132.
1053 To minimize changes between patchlevels in the OpenSSL 0.9.8
1054 series, Camellia remains excluded from compilation unless OpenSSL
1055 is configured with 'enable-camellia'.
1058 *) Disable the padding bug check when compression is in use. The padding
1059 bug check assumes the first packet is of even length, this is not
1060 necessarily true if compresssion is enabled and can result in false
1061 positives causing handshake failure. The actual bug test is ancient
1062 code so it is hoped that implementations will either have fixed it by
1063 now or any which still have the bug do not support compression.
1066 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1068 *) When applying a cipher rule check to see if string match is an explicit
1069 cipher suite and only match that one cipher suite if it is.
1072 *) Link in manifests for VC++ if needed.
1073 [Austin Ziegler <halostatue@gmail.com>]
1075 *) Update support for ECC-based TLS ciphersuites according to
1076 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1077 TLS extensions, which are supported starting with the 0.9.9
1078 branch, not in the OpenSSL 0.9.8 branch).
1081 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1082 opaque EVP_CIPHER_CTX handling.
1085 *) Fixes and enhancements to zlib compression code. We now only use
1086 "zlib1.dll" and use the default __cdecl calling convention on Win32
1087 to conform with the standards mentioned here:
1088 http://www.zlib.net/DLL_FAQ.txt
1089 Static zlib linking now works on Windows and the new --with-zlib-include
1090 --with-zlib-lib options to Configure can be used to supply the location
1091 of the headers and library. Gracefully handle case where zlib library
1095 *) Several fixes and enhancements to the OID generation code. The old code
1096 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1097 handle numbers larger than ULONG_MAX, truncated printing and had a
1098 non standard OBJ_obj2txt() behaviour.
1101 *) Add support for building of engines under engine/ as shared libraries
1102 under VC++ build system.
1105 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1106 Hopefully, we will not see any false combination of paths any more.
1109 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1111 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1112 (part of SSL_OP_ALL). This option used to disable the
1113 countermeasure against man-in-the-middle protocol-version
1114 rollback in the SSL 2.0 server implementation, which is a bad
1115 idea. (CVE-2005-2969)
1117 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1118 for Information Security, National Institute of Advanced Industrial
1119 Science and Technology [AIST], Japan)]
1121 *) Add two function to clear and return the verify parameter flags.
1124 *) Keep cipherlists sorted in the source instead of sorting them at
1125 runtime, thus removing the need for a lock.
1128 *) Avoid some small subgroup attacks in Diffie-Hellman.
1129 [Nick Mathewson and Ben Laurie]
1131 *) Add functions for well-known primes.
1134 *) Extended Windows CE support.
1135 [Satoshi Nakamura and Andy Polyakov]
1137 *) Initialize SSL_METHOD structures at compile time instead of during
1138 runtime, thus removing the need for a lock.
1141 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1142 attempting to decrypt each encrypted key in turn. Add support to
1146 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1148 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1151 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1154 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1155 key into the same file any more.
1158 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1161 *) Add -utf8 command line and config file option to 'ca'.
1162 [Stefan <stf@udoma.org]
1164 *) Removed the macro des_crypt(), as it seems to conflict with some
1165 libraries. Use DES_crypt().
1168 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1169 involves renaming the source and generated shared-libs for
1170 both. The engines will accept the corrected or legacy ids
1171 ('ncipher' and '4758_cca' respectively) when binding. NB,
1172 this only applies when building 'shared'.
1173 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1175 *) Add attribute functions to EVP_PKEY structure. Modify
1176 PKCS12_create() to recognize a CSP name attribute and
1177 use it. Make -CSP option work again in pkcs12 utility.
1180 *) Add new functionality to the bn blinding code:
1181 - automatic re-creation of the BN_BLINDING parameters after
1182 a fixed number of uses (currently 32)
1183 - add new function for parameter creation
1184 - introduce flags to control the update behaviour of the
1185 BN_BLINDING parameters
1186 - hide BN_BLINDING structure
1187 Add a second BN_BLINDING slot to the RSA structure to improve
1188 performance when a single RSA object is shared among several
1192 *) Add support for DTLS.
1193 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1195 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1196 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1199 *) Remove buggy and incompletet DH cert support from
1200 ssl/ssl_rsa.c and ssl/s3_both.c
1203 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1204 the apps/openssl applications.
1207 *) Compile clean with "-Wall -Wmissing-prototypes
1208 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1209 DEBUG_SAFESTACK must also be set.
1212 *) Change ./Configure so that certain algorithms can be disabled by default.
1213 The new counterpiece to "no-xxx" is "enable-xxx".
1215 The patented RC5 and MDC2 algorithms will now be disabled unless
1216 "enable-rc5" and "enable-mdc2", respectively, are specified.
1218 (IDEA remains enabled despite being patented. This is because IDEA
1219 is frequently required for interoperability, and there is no license
1220 fee for non-commercial use. As before, "no-idea" can be used to
1221 avoid this algorithm.)
1225 *) Add processing of proxy certificates (see RFC 3820). This work was
1226 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1227 EGEE (Enabling Grids for E-science in Europe).
1230 *) RC4 performance overhaul on modern architectures/implementations, such
1231 as Intel P4, IA-64 and AMD64.
1234 *) New utility extract-section.pl. This can be used specify an alternative
1235 section number in a pod file instead of having to treat each file as
1236 a separate case in Makefile. This can be done by adding two lines to the
1239 =for comment openssl_section:XXX
1241 The blank line is mandatory.
1245 *) New arguments -certform, -keyform and -pass for s_client and s_server
1246 to allow alternative format key and certificate files and passphrase
1250 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1251 update associated structures and add various utility functions.
1253 Add new policy related verify parameters, include policy checking in
1254 standard verify code. Enhance 'smime' application with extra parameters
1255 to support policy checking and print out.
1258 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1259 Nehemiah processors. These extensions support AES encryption in hardware
1260 as well as RNG (though RNG support is currently disabled).
1261 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1263 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1266 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1267 [Andy Polyakov and a number of other people]
1269 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1270 implementation contributed by IBM.
1271 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1273 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1274 exponent rather than 'unsigned long'. There is a corresponding change to
1275 the new 'rsa_keygen' element of the RSA_METHOD structure.
1276 [Jelte Jansen, Geoff Thorpe]
1278 *) Functionality for creating the initial serial number file is now
1279 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1281 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1282 number file to 1, which is bound to cause problems. To avoid
1283 the problems while respecting compatibility between different 0.9.7
1284 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1285 CA.pl for serial number initialization. With the new release 0.9.8,
1286 we can fix the problem directly in the 'ca' utility.)
1289 *) Reduced header interdepencies by declaring more opaque objects in
1290 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1291 give fewer recursive includes, which could break lazy source code - so
1292 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1293 developers should define this symbol when building and using openssl to
1294 ensure they track the recommended behaviour, interfaces, [etc], but
1295 backwards-compatible behaviour prevails when this isn't defined.
1298 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1301 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1302 This will generate a random key of the appropriate length based on the
1303 cipher context. The EVP_CIPHER can provide its own random key generation
1304 routine to support keys of a specific form. This is used in the des and
1305 3des routines to generate a key of the correct parity. Update S/MIME
1306 code to use new functions and hence generate correct parity DES keys.
1307 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1308 valid (weak or incorrect parity).
1311 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1312 as looking them up. This is useful when the verified structure may contain
1313 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1314 present unless the new PKCS7_NO_CRL flag is asserted.
1317 *) Extend ASN1 oid configuration module. It now additionally accepts the
1320 shortName = some long name, 1.2.3.4
1323 *) Reimplemented the BN_CTX implementation. There is now no more static
1324 limitation on the number of variables it can handle nor the depth of the
1325 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1326 information can now expand as required, and rather than having a single
1327 static array of bignums, BN_CTX now uses a linked-list of such arrays
1328 allowing it to expand on demand whilst maintaining the usefulness of
1329 BN_CTX's "bundling".
1332 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1333 to allow all RSA operations to function using a single BN_CTX.
1336 *) Preliminary support for certificate policy evaluation and checking. This
1337 is initially intended to pass the tests outlined in "Conformance Testing
1338 of Relying Party Client Certificate Path Processing Logic" v1.07.
1341 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1342 remained unused and not that useful. A variety of other little bignum
1343 tweaks and fixes have also been made continuing on from the audit (see
1347 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1348 associated ASN1, EVP and SSL functions and old ASN1 macros.
1351 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1352 and this should never fail. So the return value from the use of
1353 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1354 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1357 *) BN_CTX_get() should return zero-valued bignums, providing the same
1358 initialised value as BN_new().
1359 [Geoff Thorpe, suggested by Ulf Möller]
1361 *) Support for inhibitAnyPolicy certificate extension.
1364 *) An audit of the BIGNUM code is underway, for which debugging code is
1365 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1366 is considered valid when processing BIGNUMs, and causes execution to
1367 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1368 further steps are taken to deliberately pollute unused data in BIGNUM
1369 structures to try and expose faulty code further on. For now, openssl will
1370 (in its default mode of operation) continue to tolerate the inconsistent
1371 forms that it has tolerated in the past, but authors and packagers should
1372 consider trying openssl and their own applications when compiled with
1373 these debugging symbols defined. It will help highlight potential bugs in
1374 their own code, and will improve the test coverage for OpenSSL itself. At
1375 some point, these tighter rules will become openssl's default to improve
1376 maintainability, though the assert()s and other overheads will remain only
1377 in debugging configurations. See bn.h for more details.
1378 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1380 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1381 that can only be obtained through BN_CTX_new() (which implicitly
1382 initialises it). The presence of this function only made it possible
1383 to overwrite an existing structure (and cause memory leaks).
1386 *) Because of the callback-based approach for implementing LHASH as a
1387 template type, lh_insert() adds opaque objects to hash-tables and
1388 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1389 to clean up those corresponding objects before destroying the hash table
1390 (and losing the object pointers). So some over-zealous constifications in
1391 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1392 objects as "const" and the lh_doall[_arg] callback wrappers are not
1393 prototyped to have "const" restrictions on the object pointers they are
1394 given (and so aren't required to cast them away any more).
1397 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1398 (speed) prefers to use its own implementation. The two implementations
1399 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1400 its object type properly exposed (MS_TM) instead of casting to/from "char
1401 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1402 aren't necessarily the greatest nomenclatures - but this is what was used
1403 internally to the implementation so I've used that for now.
1406 *) Ensure that deprecated functions do not get compiled when
1407 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1408 the self-tests were still using deprecated key-generation functions so
1409 these have been updated also.
1412 *) Reorganise PKCS#7 code to separate the digest location functionality
1413 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1414 New function PKCS7_set_digest() to set the digest type for PKCS#7
1415 digestedData type. Add additional code to correctly generate the
1416 digestedData type and add support for this type in PKCS7 initialization
1420 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1421 structure of type "other".
1424 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1425 sure the loop does correctly stop and breaking ("division by zero")
1426 modulus operations are not performed. The (pre-generated) prime
1427 table crypto/bn/bn_prime.h was already correct, but it could not be
1428 re-generated on some platforms because of the "division by zero"
1429 situation in the script.
1430 [Ralf S. Engelschall]
1432 *) Update support for ECC-based TLS ciphersuites according to
1433 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1434 SHA-1 now is only used for "small" curves (where the
1435 representation of a field element takes up to 24 bytes); for
1436 larger curves, the field element resulting from ECDH is directly
1437 used as premaster secret.
1438 [Douglas Stebila (Sun Microsystems Laboratories)]
1440 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1441 curve secp160r1 to the tests.
1442 [Douglas Stebila (Sun Microsystems Laboratories)]
1444 *) Add the possibility to load symbols globally with DSO.
1445 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1447 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1448 control of the error stack.
1451 *) Add support for STORE in ENGINE.
1454 *) Add the STORE type. The intention is to provide a common interface
1455 to certificate and key stores, be they simple file-based stores, or
1456 HSM-type store, or LDAP stores, or...
1457 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1460 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1461 pass a list of arguments to any function as well as provide a way
1462 for a function to pass data back to the caller.
1465 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1466 works like BUF_strdup() but can be used to duplicate a portion of
1467 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1471 *) Add the function sk_find_ex() which works like sk_find(), but will
1472 return an index to an element even if an exact match couldn't be
1473 found. The index is guaranteed to point at the element where the
1474 searched-for key would be inserted to preserve sorting order.
1477 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1478 takes an extra flags argument for optional functionality. Currently,
1479 the following flags are defined:
1481 OBJ_BSEARCH_VALUE_ON_NOMATCH
1482 This one gets OBJ_bsearch_ex() to return a pointer to the first
1483 element where the comparing function returns a negative or zero
1486 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1487 This one gets OBJ_bsearch_ex() to return a pointer to the first
1488 element where the comparing function returns zero. This is useful
1489 if there are more than one element where the comparing function
1493 *) Make it possible to create self-signed certificates with 'openssl ca'
1494 in such a way that the self-signed certificate becomes part of the
1495 CA database and uses the same mechanisms for serial number generation
1496 as all other certificate signing. The new flag '-selfsign' enables
1497 this functionality. Adapt CA.sh and CA.pl.in.
1500 *) Add functionality to check the public key of a certificate request
1501 against a given private. This is useful to check that a certificate
1502 request can be signed by that key (self-signing).
1505 *) Make it possible to have multiple active certificates with the same
1506 subject in the CA index file. This is done only if the keyword
1507 'unique_subject' is set to 'no' in the main CA section (default
1508 if 'CA_default') of the configuration file. The value is saved
1509 with the database itself in a separate index attribute file,
1510 named like the index file with '.attr' appended to the name.
1513 *) Generate muti valued AVAs using '+' notation in config files for
1517 *) Support for nameConstraints certificate extension.
1520 *) Support for policyConstraints certificate extension.
1523 *) Support for policyMappings certificate extension.
1526 *) Make sure the default DSA_METHOD implementation only uses its
1527 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1528 and change its own handlers to be NULL so as to remove unnecessary
1529 indirection. This lets alternative implementations fallback to the
1530 default implementation more easily.
1533 *) Support for directoryName in GeneralName related extensions
1537 *) Make it possible to link applications using Makefile.shared.
1538 Make that possible even when linking against static libraries!
1541 *) Support for single pass processing for S/MIME signing. This now
1542 means that S/MIME signing can be done from a pipe, in addition
1543 cleartext signing (multipart/signed type) is effectively streaming
1544 and the signed data does not need to be all held in memory.
1546 This is done with a new flag PKCS7_STREAM. When this flag is set
1547 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1548 is done after the data is output (and digests calculated) in
1549 SMIME_write_PKCS7().
1552 *) Add full support for -rpath/-R, both in shared libraries and
1553 applications, at least on the platforms where it's known how
1557 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1558 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1559 will now compute a table of multiples of the generator that
1560 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1561 faster (notably in the case of a single point multiplication,
1562 scalar * generator).
1563 [Nils Larsch, Bodo Moeller]
1565 *) IPv6 support for certificate extensions. The various extensions
1566 which use the IP:a.b.c.d can now take IPv6 addresses using the
1567 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1571 *) Added an ENGINE that implements RSA by performing private key
1572 exponentiations with the GMP library. The conversions to and from
1573 GMP's mpz_t format aren't optimised nor are any montgomery forms
1574 cached, and on x86 it appears OpenSSL's own performance has caught up.
1575 However there are likely to be other architectures where GMP could
1576 provide a boost. This ENGINE is not built in by default, but it can be
1577 specified at Configure time and should be accompanied by the necessary
1578 linker additions, eg;
1579 ./config -DOPENSSL_USE_GMP -lgmp
1582 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1583 testing availability of engines with "-t" - the old behaviour is
1584 produced by increasing the feature's verbosity with "-tt".
1587 *) ECDSA routines: under certain error conditions uninitialized BN objects
1588 could be freed. Solution: make sure initialization is performed early
1589 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1593 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1594 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1595 software implementations. For DSA and DH, parameter generation can
1596 also be overriden by providing the appropriate method callbacks.
1599 *) Change the "progress" mechanism used in key-generation and
1600 primality testing to functions that take a new BN_GENCB pointer in
1601 place of callback/argument pairs. The new API functions have "_ex"
1602 postfixes and the older functions are reimplemented as wrappers for
1603 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1604 declarations of the old functions to help (graceful) attempts to
1605 migrate to the new functions. Also, the new key-generation API
1606 functions operate on a caller-supplied key-structure and return
1607 success/failure rather than returning a key or NULL - this is to
1608 help make "keygen" another member function of RSA_METHOD etc.
1610 Example for using the new callback interface:
1612 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1616 BN_GENCB_set(&my_cb, my_callback, my_arg);
1618 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1619 /* For the meaning of a, b in calls to my_callback(), see the
1620 * documentation of the function that calls the callback.
1621 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1622 * my_callback should return 1 if it wants BN_is_prime_ex()
1623 * to continue, or 0 to stop.
1628 *) Change the ZLIB compression method to be stateful, and make it
1629 available to TLS with the number defined in
1630 draft-ietf-tls-compression-04.txt.
1633 *) Add the ASN.1 structures and functions for CertificatePair, which
1634 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1636 CertificatePair ::= SEQUENCE {
1637 forward [0] Certificate OPTIONAL,
1638 reverse [1] Certificate OPTIONAL,
1639 -- at least one of the pair shall be present -- }
1641 Also implement the PEM functions to read and write certificate
1642 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1644 This needed to be defined, mostly for the sake of the LDAP
1645 attribute crossCertificatePair, but may prove useful elsewhere as
1649 *) Make it possible to inhibit symlinking of shared libraries in
1650 Makefile.shared, for Cygwin's sake.
1653 *) Extend the BIGNUM API by creating a function
1654 void BN_set_negative(BIGNUM *a, int neg);
1655 and a macro that behave like
1656 int BN_is_negative(const BIGNUM *a);
1658 to avoid the need to access 'a->neg' directly in applications.
1661 *) Implement fast modular reduction for pseudo-Mersenne primes
1662 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1663 EC_GROUP_new_curve_GFp() will now automatically use this
1665 [Nils Larsch <nla@trustcenter.de>]
1667 *) Add new lock type (CRYPTO_LOCK_BN).
1670 *) Change the ENGINE framework to automatically load engines
1671 dynamically from specific directories unless they could be
1672 found to already be built in or loaded. Move all the
1673 current engines except for the cryptodev one to a new
1675 The engines in engines/ are built as shared libraries if
1676 the "shared" options was given to ./Configure or ./config.
1677 Otherwise, they are inserted in libcrypto.a.
1678 /usr/local/ssl/engines is the default directory for dynamic
1679 engines, but that can be overriden at configure time through
1680 the usual use of --prefix and/or --openssldir, and at run
1681 time with the environment variable OPENSSL_ENGINES.
1682 [Geoff Thorpe and Richard Levitte]
1684 *) Add Makefile.shared, a helper makefile to build shared
1685 libraries. Addapt Makefile.org.
1688 *) Add version info to Win32 DLLs.
1689 [Peter 'Luna' Runestig" <peter@runestig.com>]
1691 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1692 can be added using this API to created arbitrary PKCS#12
1693 files while avoiding the low level API.
1695 New options to PKCS12_create(), key or cert can be NULL and
1696 will then be omitted from the output file. The encryption
1697 algorithm NIDs can be set to -1 for no encryption, the mac
1698 iteration count can be set to 0 to omit the mac.
1700 Enhance pkcs12 utility by making the -nokeys and -nocerts
1701 options work when creating a PKCS#12 file. New option -nomac
1702 to omit the mac, NONE can be set for an encryption algorithm.
1703 New code is modified to use the enhanced PKCS12_create()
1704 instead of the low level API.
1707 *) Extend ASN1 encoder to support indefinite length constructed
1708 encoding. This can output sequences tags and octet strings in
1709 this form. Modify pk7_asn1.c to support indefinite length
1710 encoding. This is experimental and needs additional code to
1711 be useful, such as an ASN1 bio and some enhanced streaming
1714 Extend template encode functionality so that tagging is passed
1715 down to the template encoder.
1718 *) Let 'openssl req' fail if an argument to '-newkey' is not
1719 recognized instead of using RSA as a default.
1722 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1723 As these are not official, they are not included in "ALL";
1724 the "ECCdraft" ciphersuite group alias can be used to select them.
1725 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1727 *) Add ECDH engine support.
1728 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1730 *) Add ECDH in new directory crypto/ecdh/.
1731 [Douglas Stebila (Sun Microsystems Laboratories)]
1733 *) Let BN_rand_range() abort with an error after 100 iterations
1734 without success (which indicates a broken PRNG).
1737 *) Change BN_mod_sqrt() so that it verifies that the input value
1738 is really the square of the return value. (Previously,
1739 BN_mod_sqrt would show GIGO behaviour.)
1742 *) Add named elliptic curves over binary fields from X9.62, SECG,
1743 and WAP/WTLS; add OIDs that were still missing.
1745 [Sheueling Chang Shantz and Douglas Stebila
1746 (Sun Microsystems Laboratories)]
1748 *) Extend the EC library for elliptic curves over binary fields
1749 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1752 EC_GF2m_simple_method
1756 EC_GROUP_new_curve_GF2m
1757 EC_GROUP_set_curve_GF2m
1758 EC_GROUP_get_curve_GF2m
1759 EC_POINT_set_affine_coordinates_GF2m
1760 EC_POINT_get_affine_coordinates_GF2m
1761 EC_POINT_set_compressed_coordinates_GF2m
1763 Point compression for binary fields is disabled by default for
1764 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1767 As binary polynomials are represented as BIGNUMs, various members
1768 of the EC_GROUP and EC_POINT data structures can be shared
1769 between the implementations for prime fields and binary fields;
1770 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1771 are essentially identical to their ..._GFp counterparts.
1772 (For simplicity, the '..._GFp' prefix has been dropped from
1773 various internal method names.)
1775 An internal 'field_div' method (similar to 'field_mul' and
1776 'field_sqr') has been added; this is used only for binary fields.
1778 [Sheueling Chang Shantz and Douglas Stebila
1779 (Sun Microsystems Laboratories)]
1781 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1782 through methods ('mul', 'precompute_mult').
1784 The generic implementations (now internally called 'ec_wNAF_mul'
1785 and 'ec_wNAF_precomputed_mult') remain the default if these
1786 methods are undefined.
1788 [Sheueling Chang Shantz and Douglas Stebila
1789 (Sun Microsystems Laboratories)]
1791 *) New function EC_GROUP_get_degree, which is defined through
1792 EC_METHOD. For curves over prime fields, this returns the bit
1793 length of the modulus.
1795 [Sheueling Chang Shantz and Douglas Stebila
1796 (Sun Microsystems Laboratories)]
1798 *) New functions EC_GROUP_dup, EC_POINT_dup.
1799 (These simply call ..._new and ..._copy).
1801 [Sheueling Chang Shantz and Douglas Stebila
1802 (Sun Microsystems Laboratories)]
1804 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1805 Polynomials are represented as BIGNUMs (where the sign bit is not
1806 used) in the following functions [macros]:
1809 BN_GF2m_sub [= BN_GF2m_add]
1810 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1811 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1812 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1814 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1815 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1816 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1817 BN_GF2m_cmp [= BN_ucmp]
1819 (Note that only the 'mod' functions are actually for fields GF(2^m).
1820 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1822 For some functions, an the irreducible polynomial defining a
1823 field can be given as an 'unsigned int[]' with strictly
1824 decreasing elements giving the indices of those bits that are set;
1825 i.e., p[] represents the polynomial
1826 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1828 p[0] > p[1] > ... > p[k] = 0.
1829 This applies to the following functions:
1834 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1835 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1837 BN_GF2m_mod_sqrt_arr
1838 BN_GF2m_mod_solve_quad_arr
1842 Conversion can be performed by the following functions:
1847 bntest.c has additional tests for binary polynomial arithmetic.
1849 Two implementations for BN_GF2m_mod_div() are available.
1850 The default algorithm simply uses BN_GF2m_mod_inv() and
1851 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1852 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1853 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1855 [Sheueling Chang Shantz and Douglas Stebila
1856 (Sun Microsystems Laboratories)]
1858 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1859 functionality is disabled at compile-time.
1860 [Douglas Stebila <douglas.stebila@sun.com>]
1862 *) Change default behaviour of 'openssl asn1parse' so that more
1863 information is visible when viewing, e.g., a certificate:
1865 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1866 mode the content of non-printable OCTET STRINGs is output in a
1867 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1868 avoid the appearance of a printable string.
1869 [Nils Larsch <nla@trustcenter.de>]
1871 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1873 EC_GROUP_set_asn1_flag()
1874 EC_GROUP_get_asn1_flag()
1875 EC_GROUP_set_point_conversion_form()
1876 EC_GROUP_get_point_conversion_form()
1877 These control ASN1 encoding details:
1878 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1879 has been set to OPENSSL_EC_NAMED_CURVE.
1880 - Points are encoded in uncompressed form by default; options for
1881 asn1_for are as for point2oct, namely
1882 POINT_CONVERSION_COMPRESSED
1883 POINT_CONVERSION_UNCOMPRESSED
1884 POINT_CONVERSION_HYBRID
1886 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1889 EC_GROUP_get0_seed()
1890 EC_GROUP_get_seed_len()
1891 This is used only for ASN1 purposes (so far).
1892 [Nils Larsch <nla@trustcenter.de>]
1894 *) Add 'field_type' member to EC_METHOD, which holds the NID
1895 of the appropriate field type OID. The new function
1896 EC_METHOD_get_field_type() returns this value.
1897 [Nils Larsch <nla@trustcenter.de>]
1902 EC_POINT_point2hex()
1903 EC_POINT_hex2point()
1904 providing useful interfaces to EC_POINT_point2oct() and
1905 EC_POINT_oct2point().
1906 [Nils Larsch <nla@trustcenter.de>]
1908 *) Change internals of the EC library so that the functions
1909 EC_GROUP_set_generator()
1910 EC_GROUP_get_generator()
1911 EC_GROUP_get_order()
1912 EC_GROUP_get_cofactor()
1913 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1914 to methods, which would lead to unnecessary code duplication when
1915 adding different types of curves.
1916 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1918 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1919 arithmetic, and such that modified wNAFs are generated
1920 (which avoid length expansion in many cases).
1923 *) Add a function EC_GROUP_check_discriminant() (defined via
1924 EC_METHOD) that verifies that the curve discriminant is non-zero.
1926 Add a function EC_GROUP_check() that makes some sanity tests
1927 on a EC_GROUP, its generator and order. This includes
1928 EC_GROUP_check_discriminant().
1929 [Nils Larsch <nla@trustcenter.de>]
1931 *) Add ECDSA in new directory crypto/ecdsa/.
1933 Add applications 'openssl ecparam' and 'openssl ecdsa'
1934 (these are based on 'openssl dsaparam' and 'openssl dsa').
1936 ECDSA support is also included in various other files across the
1937 library. Most notably,
1938 - 'openssl req' now has a '-newkey ecdsa:file' option;
1939 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1940 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1941 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1942 them suitable for ECDSA where domain parameters must be
1943 extracted before the specific public key;
1944 - ECDSA engine support has been added.
1945 [Nils Larsch <nla@trustcenter.de>]
1947 *) Include some named elliptic curves, and add OIDs from X9.62,
1948 SECG, and WAP/WTLS. Each curve can be obtained from the new
1950 EC_GROUP_new_by_curve_name(),
1951 and the list of available named curves can be obtained with
1952 EC_get_builtin_curves().
1953 Also add a 'curve_name' member to EC_GROUP objects, which can be
1955 EC_GROUP_set_curve_name()
1956 EC_GROUP_get_curve_name()
1957 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1959 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1960 was actually never needed) and in BN_mul(). The removal in BN_mul()
1961 required a small change in bn_mul_part_recursive() and the addition
1962 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1963 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1964 bn_sub_words() and bn_add_words() except they take arrays with
1968 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1970 *) Cleanse PEM buffers before freeing them since they may contain
1972 [Benjamin Bennett <ben@psc.edu>]
1974 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1975 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1976 authentication-only ciphersuites.
1979 *) Since AES128 and AES256 share a single mask bit in the logic of
1980 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1981 kludge to work properly if AES128 is available and AES256 isn't.
1984 *) Expand security boundary to match 1.1.1 module.
1987 *) Remove redundant features: hash file source, editing of test vectors
1988 modify fipsld to use external fips_premain.c signature.
1991 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1992 run algorithm test programs.
1995 *) Make algorithm test programs more tolerant of whitespace.
1998 *) Have SSL/TLS server implementation tolerate "mismatched" record
1999 protocol version while receiving ClientHello even if the
2000 ClientHello is fragmented. (The server can't insist on the
2001 particular protocol version it has chosen before the ServerHello
2002 message has informed the client about his choice.)
2005 *) Load error codes if they are not already present instead of using a
2006 static variable. This allows them to be cleanly unloaded and reloaded.
2009 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2011 *) Introduce limits to prevent malicious keys being able to
2012 cause a denial of service. (CVE-2006-2940)
2013 [Steve Henson, Bodo Moeller]
2015 *) Fix ASN.1 parsing of certain invalid structures that can result
2016 in a denial of service. (CVE-2006-2937) [Steve Henson]
2018 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2019 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2021 *) Fix SSL client code which could crash if connecting to a
2022 malicious SSLv2 server. (CVE-2006-4343)
2023 [Tavis Ormandy and Will Drewry, Google Security Team]
2025 *) Change ciphersuite string processing so that an explicit
2026 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2027 will no longer include "AES128-SHA"), and any other similar
2028 ciphersuite (same bitmap) from *other* protocol versions (so that
2029 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2030 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2031 changes from 0.9.8b and 0.9.8d.
2034 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2036 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2037 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2039 *) Change the Unix randomness entropy gathering to use poll() when
2040 possible instead of select(), since the latter has some
2041 undesirable limitations.
2042 [Darryl Miles via Richard Levitte and Bodo Moeller]
2044 *) Disable rogue ciphersuites:
2046 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2047 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2048 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2050 The latter two were purportedly from
2051 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2054 Also deactive the remaining ciphersuites from
2055 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2056 unofficial, and the ID has long expired.
2059 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2060 dual-core machines) and other potential thread-safety issues.
2063 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2065 *) Adapt fipsld and the build system to link against the validated FIPS
2066 module in FIPS mode.
2069 *) Fixes for VC++ 2005 build under Windows.
2072 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2073 from a Windows bash shell such as MSYS. It is autodetected from the
2074 "config" script when run from a VC++ environment. Modify standard VC++
2075 build to use fipscanister.o from the GNU make build.
2078 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2080 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2081 The value now differs depending on if you build for FIPS or not.
2082 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2083 safely run with a non-FIPSed libcrypto, as it may crash because of
2084 the difference induced by this change.
2087 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2089 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2090 (part of SSL_OP_ALL). This option used to disable the
2091 countermeasure against man-in-the-middle protocol-version
2092 rollback in the SSL 2.0 server implementation, which is a bad
2093 idea. (CVE-2005-2969)
2095 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2096 for Information Security, National Institute of Advanced Industrial
2097 Science and Technology [AIST], Japan)]
2099 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2100 mainly for FIPS compliance and not fully integrated at this stage.
2103 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2104 the exponentiation using a fixed-length exponent. (Otherwise,
2105 the information leaked through timing could expose the secret key
2106 after many signatures; cf. Bleichenbacher's attack on DSA with
2110 *) Make a new fixed-window mod_exp implementation the default for
2111 RSA, DSA, and DH private-key operations so that the sequence of
2112 squares and multiplies and the memory access pattern are
2113 independent of the particular secret key. This will mitigate
2114 cache-timing and potential related attacks.
2116 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2117 and this is automatically used by BN_mod_exp_mont() if the new flag
2118 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2119 will use this BN flag for private exponents unless the flag
2120 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2121 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2123 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2125 *) Change the client implementation for SSLv23_method() and
2126 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2127 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2128 (Previously, the SSL 2.0 backwards compatible Client Hello
2129 message format would be used even with SSL_OP_NO_SSLv2.)
2132 *) Add support for smime-type MIME parameter in S/MIME messages which some
2136 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2137 a threadsafe manner. Modify rsa code to use new function and add calls
2138 to dsa and dh code (which had race conditions before).
2141 *) Include the fixed error library code in the C error file definitions
2142 instead of fixing them up at runtime. This keeps the error code
2143 structures constant.
2146 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2148 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2151 *) Fixes for newer kerberos headers. NB: the casts are needed because
2152 the 'length' field is signed on one version and unsigned on another
2153 with no (?) obvious way to tell the difference, without these VC++
2154 complains. Also the "definition" of FAR (blank) is no longer included
2155 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2156 some needed definitions.
2159 *) Undo Cygwin change.
2162 *) Added support for proxy certificates according to RFC 3820.
2163 Because they may be a security thread to unaware applications,
2164 they must be explicitely allowed in run-time. See
2165 docs/HOWTO/proxy_certificates.txt for further information.
2168 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2170 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2171 server and client random values. Previously
2172 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2173 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2175 This change has negligible security impact because:
2177 1. Server and client random values still have 24 bytes of pseudo random
2180 2. Server and client random values are sent in the clear in the initial
2183 3. The master secret is derived using the premaster secret (48 bytes in
2184 size for static RSA ciphersuites) as well as client server and random
2187 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2190 [Stephen Henson, reported by UK NISCC]
2192 *) Use Windows randomness collection on Cygwin.
2195 *) Fix hang in EGD/PRNGD query when communication socket is closed
2196 prematurely by EGD/PRNGD.
2197 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2199 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2202 *) Back-port of selected performance improvements from development
2203 branch, as well as improved support for PowerPC platforms.
2206 *) Add lots of checks for memory allocation failure, error codes to indicate
2207 failure and freeing up memory if a failure occurs.
2208 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2210 *) Add new -passin argument to dgst.
2213 *) Perform some character comparisons of different types in X509_NAME_cmp:
2214 this is needed for some certificates that reencode DNs into UTF8Strings
2215 (in violation of RFC3280) and can't or wont issue name rollover
2219 *) Make an explicit check during certificate validation to see that
2220 the CA setting in each certificate on the chain is correct. As a
2221 side effect always do the following basic checks on extensions,
2222 not just when there's an associated purpose to the check:
2224 - if there is an unhandled critical extension (unless the user
2225 has chosen to ignore this fault)
2226 - if the path length has been exceeded (if one is set at all)
2227 - that certain extensions fit the associated purpose (if one has
2231 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2233 *) Avoid a race condition when CRLs are checked in a multi threaded
2234 environment. This would happen due to the reordering of the revoked
2235 entries during signature checking and serial number lookup. Now the
2236 encoding is cached and the serial number sort performed under a lock.
2237 Add new STACK function sk_is_sorted().
2240 *) Add Delta CRL to the extension code.
2243 *) Various fixes to s3_pkt.c so alerts are sent properly.
2244 [David Holmes <d.holmes@f5.com>]
2246 *) Reduce the chances of duplicate issuer name and serial numbers (in
2247 violation of RFC3280) using the OpenSSL certificate creation utilities.
2248 This is done by creating a random 64 bit value for the initial serial
2249 number when a serial number file is created or when a self signed
2250 certificate is created using 'openssl req -x509'. The initial serial
2251 number file is created using 'openssl x509 -next_serial' in CA.pl
2252 rather than being initialized to 1.
2255 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2257 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2258 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2259 [Joe Orton, Steve Henson]
2261 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2263 [Joe Orton, Steve Henson]
2265 *) Make it possible to have multiple active certificates with the same
2266 subject in the CA index file. This is done only if the keyword
2267 'unique_subject' is set to 'no' in the main CA section (default
2268 if 'CA_default') of the configuration file. The value is saved
2269 with the database itself in a separate index attribute file,
2270 named like the index file with '.attr' appended to the name.
2273 *) X509 verify fixes. Disable broken certificate workarounds when
2274 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2275 keyUsage extension present. Don't accept CRLs with unhandled critical
2276 extensions: since verify currently doesn't process CRL extensions this
2277 rejects a CRL with *any* critical extensions. Add new verify error codes
2281 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2282 A clarification of RFC2560 will require the use of OCTET STRINGs and
2283 some implementations cannot handle the current raw format. Since OpenSSL
2284 copies and compares OCSP nonces as opaque blobs without any attempt at
2285 parsing them this should not create any compatibility issues.
2288 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2289 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2290 this HMAC (and other) operations are several times slower than OpenSSL
2294 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2295 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2297 *) Use the correct content when signing type "other".
2300 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2302 *) Fix various bugs revealed by running the NISCC test suite:
2304 Stop out of bounds reads in the ASN1 code when presented with
2305 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2307 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2309 If verify callback ignores invalid public key errors don't try to check
2310 certificate signature with the NULL public key.
2314 *) New -ignore_err option in ocsp application to stop the server
2315 exiting on the first error in a request.
2318 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2319 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2323 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2324 extra data after the compression methods not only for TLS 1.0
2325 but also for SSL 3.0 (as required by the specification).
2326 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2328 *) Change X509_certificate_type() to mark the key as exported/exportable
2329 when it's 512 *bits* long, not 512 bytes.
2332 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2333 blocks during encryption.
2336 *) Various fixes to base64 BIO and non blocking I/O. On write
2337 flushes were not handled properly if the BIO retried. On read
2338 data was not being buffered properly and had various logic bugs.
2339 This also affects blocking I/O when the data being decoded is a
2343 *) Various S/MIME bugfixes and compatibility changes:
2344 output correct application/pkcs7 MIME type if
2345 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2346 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2347 of files as .eml work). Correctly handle very long lines in MIME
2351 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2353 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2354 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2355 a protocol version number mismatch like a decryption error
2356 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2359 *) Turn on RSA blinding by default in the default implementation
2360 to avoid a timing attack. Applications that don't want it can call
2361 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2362 They would be ill-advised to do so in most cases.
2363 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2365 *) Change RSA blinding code so that it works when the PRNG is not
2366 seeded (in this case, the secret RSA exponent is abused as
2367 an unpredictable seed -- if it is not unpredictable, there
2368 is no point in blinding anyway). Make RSA blinding thread-safe
2369 by remembering the creator's thread ID in rsa->blinding and
2370 having all other threads use local one-time blinding factors
2371 (this requires more computation than sharing rsa->blinding, but
2372 avoids excessive locking; and if an RSA object is not shared
2373 between threads, blinding will still be very fast).
2376 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2377 ENGINE as defaults for all supported algorithms irrespective of
2378 the 'flags' parameter. 'flags' is now honoured, so applications
2379 should make sure they are passing it correctly.
2382 *) Target "mingw" now allows native Windows code to be generated in
2383 the Cygwin environment as well as with the MinGW compiler.
2386 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2388 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2389 via timing by performing a MAC computation even if incorrrect
2390 block cipher padding has been found. This is a countermeasure
2391 against active attacks where the attacker has to distinguish
2392 between bad padding and a MAC verification error. (CVE-2003-0078)
2394 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2395 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2396 Martin Vuagnoux (EPFL, Ilion)]
2398 *) Make the no-err option work as intended. The intention with no-err
2399 is not to have the whole error stack handling routines removed from
2400 libcrypto, it's only intended to remove all the function name and
2401 reason texts, thereby removing some of the footprint that may not
2402 be interesting if those errors aren't displayed anyway.
2404 NOTE: it's still possible for any application or module to have it's
2405 own set of error texts inserted. The routines are there, just not
2406 used by default when no-err is given.
2409 *) Add support for FreeBSD on IA64.
2410 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2412 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2413 Kerberos function mit_des_cbc_cksum(). Before this change,
2414 the value returned by DES_cbc_cksum() was like the one from
2415 mit_des_cbc_cksum(), except the bytes were swapped.
2416 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2418 *) Allow an application to disable the automatic SSL chain building.
2419 Before this a rather primitive chain build was always performed in
2420 ssl3_output_cert_chain(): an application had no way to send the
2421 correct chain if the automatic operation produced an incorrect result.
2423 Now the chain builder is disabled if either:
2425 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2427 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2429 The reasoning behind this is that an application would not want the
2430 auto chain building to take place if extra chain certificates are
2431 present and it might also want a means of sending no additional
2432 certificates (for example the chain has two certificates and the
2436 *) Add the possibility to build without the ENGINE framework.
2437 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2439 *) Under Win32 gmtime() can return NULL: check return value in
2440 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2443 *) DSA routines: under certain error conditions uninitialized BN objects
2444 could be freed. Solution: make sure initialization is performed early
2445 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2446 Nils Larsch <nla@trustcenter.de> via PR#459)
2449 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2450 checked on reconnect on the client side, therefore session resumption
2451 could still fail with a "ssl session id is different" error. This
2452 behaviour is masked when SSL_OP_ALL is used due to
2453 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2454 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2455 followup to PR #377.
2458 *) IA-32 assembler support enhancements: unified ELF targets, support
2459 for SCO/Caldera platforms, fix for Cygwin shared build.
2462 *) Add support for FreeBSD on sparc64. As a consequence, support for
2463 FreeBSD on non-x86 processors is separate from x86 processors on
2464 the config script, much like the NetBSD support.
2465 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2467 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2469 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2472 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2473 code (06) was taken as the first octet of the session ID and the last
2474 octet was ignored consequently. As a result SSLv2 client side session
2475 caching could not have worked due to the session ID mismatch between
2477 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2481 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2482 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2486 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2487 seems that in spite of existing for more than a year, many application
2488 author have done nothing to provide the necessary callbacks, which
2489 means that this particular engine will not work properly anywhere.
2490 This is a very unfortunate situation which forces us, in the name
2491 of usability, to give the hw_ncipher.c a static lock, which is part
2493 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2494 appear in 0.9.8 or later. We EXPECT application authors to have
2495 dealt properly with this when 0.9.8 is released (unless we actually
2496 make such changes in the libcrypto locking code that changes will
2497 have to be made anyway).
2500 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2501 octets have been read, EOF or an error occurs. Without this change
2502 some truncated ASN1 structures will not produce an error.
2505 *) Disable Heimdal support, since it hasn't been fully implemented.
2506 Still give the possibility to force the use of Heimdal, but with
2507 warnings and a request that patches get sent to openssl-dev.
2510 *) Add the VC-CE target, introduce the WINCE sysname, and add
2511 INSTALL.WCE and appropriate conditionals to make it build.
2512 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2514 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2515 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2516 edit numbers of the version.
2517 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2519 *) Introduce safe string copy and catenation functions
2520 (BUF_strlcpy() and BUF_strlcat()).
2521 [Ben Laurie (CHATS) and Richard Levitte]
2523 *) Avoid using fixed-size buffers for one-line DNs.
2524 [Ben Laurie (CHATS)]
2526 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2527 resizing buffers containing secrets, and use where appropriate.
2528 [Ben Laurie (CHATS)]
2530 *) Avoid using fixed size buffers for configuration file location.
2531 [Ben Laurie (CHATS)]
2533 *) Avoid filename truncation for various CA files.
2534 [Ben Laurie (CHATS)]
2536 *) Use sizeof in preference to magic numbers.
2537 [Ben Laurie (CHATS)]
2539 *) Avoid filename truncation in cert requests.
2540 [Ben Laurie (CHATS)]
2542 *) Add assertions to check for (supposedly impossible) buffer
2544 [Ben Laurie (CHATS)]
2546 *) Don't cache truncated DNS entries in the local cache (this could
2547 potentially lead to a spoofing attack).
2548 [Ben Laurie (CHATS)]
2550 *) Fix various buffers to be large enough for hex/decimal
2551 representations in a platform independent manner.
2552 [Ben Laurie (CHATS)]
2554 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2555 resizing buffers containing secrets, and use where appropriate.
2556 [Ben Laurie (CHATS)]
2558 *) Add BIO_indent() to avoid much slightly worrying code to do
2560 [Ben Laurie (CHATS)]
2562 *) Convert sprintf()/BIO_puts() to BIO_printf().
2563 [Ben Laurie (CHATS)]
2565 *) buffer_gets() could terminate with the buffer only half
2567 [Ben Laurie (CHATS)]
2569 *) Add assertions to prevent user-supplied crypto functions from
2570 overflowing internal buffers by having large block sizes, etc.
2571 [Ben Laurie (CHATS)]
2573 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2575 [Ben Laurie (CHATS)]
2577 *) Eliminate unused copy of key in RC4.
2578 [Ben Laurie (CHATS)]
2580 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2581 [Ben Laurie (CHATS)]
2583 *) Fix off-by-one error in EGD path.
2584 [Ben Laurie (CHATS)]
2586 *) If RANDFILE path is too long, ignore instead of truncating.
2587 [Ben Laurie (CHATS)]
2589 *) Eliminate unused and incorrectly sized X.509 structure
2591 [Ben Laurie (CHATS)]
2593 *) Eliminate unused and dangerous function knumber().
2594 [Ben Laurie (CHATS)]
2596 *) Eliminate unused and dangerous structure, KSSL_ERR.
2597 [Ben Laurie (CHATS)]
2599 *) Protect against overlong session ID context length in an encoded
2600 session object. Since these are local, this does not appear to be
2602 [Ben Laurie (CHATS)]
2604 *) Change from security patch (see 0.9.6e below) that did not affect
2605 the 0.9.6 release series:
2607 Remote buffer overflow in SSL3 protocol - an attacker could
2608 supply an oversized master key in Kerberos-enabled versions.
2610 [Ben Laurie (CHATS)]
2612 *) Change the SSL kerb5 codes to match RFC 2712.
2615 *) Make -nameopt work fully for req and add -reqopt switch.
2616 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2618 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2619 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2621 *) Make sure tests can be performed even if the corresponding algorithms
2622 have been removed entirely. This was also the last step to make
2623 OpenSSL compilable with DJGPP under all reasonable conditions.
2624 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2626 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2627 to allow version independent disabling of normally unselected ciphers,
2628 which may be activated as a side-effect of selecting a single cipher.
2630 (E.g., cipher list string "RSA" enables ciphersuites that are left
2631 out of "ALL" because they do not provide symmetric encryption.
2632 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2633 [Lutz Jaenicke, Bodo Moeller]
2635 *) Add appropriate support for separate platform-dependent build
2636 directories. The recommended way to make a platform-dependent
2637 build directory is the following (tested on Linux), maybe with
2640 # Place yourself outside of the OpenSSL source tree. In
2641 # this example, the environment variable OPENSSL_SOURCE
2642 # is assumed to contain the absolute OpenSSL source directory.
2643 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2644 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2645 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2646 mkdir -p `dirname $F`
2647 ln -s $OPENSSL_SOURCE/$F $F
2650 To be absolutely sure not to disturb the source tree, a "make clean"
2651 is a good thing. If it isn't successfull, don't worry about it,
2652 it probably means the source directory is very clean.
2655 *) Make sure any ENGINE control commands make local copies of string
2656 pointers passed to them whenever necessary. Otherwise it is possible
2657 the caller may have overwritten (or deallocated) the original string
2658 data when a later ENGINE operation tries to use the stored values.
2659 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2661 *) Improve diagnostics in file reading and command-line digests.
2662 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2664 *) Add AES modes CFB and OFB to the object database. Correct an
2665 error in AES-CFB decryption.
2668 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2669 allows existing EVP_CIPHER_CTX structures to be reused after
2670 calling EVP_*Final(). This behaviour is used by encryption
2671 BIOs and some applications. This has the side effect that
2672 applications must explicitly clean up cipher contexts with
2673 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2676 *) Check the values of dna and dnb in bn_mul_recursive before calling
2677 bn_mul_comba (a non zero value means the a or b arrays do not contain
2678 n2 elements) and fallback to bn_mul_normal if either is not zero.
2681 *) Fix escaping of non-ASCII characters when using the -subj option
2682 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2685 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2686 form for "surname", serialNumber has no short form.
2687 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2688 therefore remove "mail" short name for "internet 7".
2689 The OID for unique identifiers in X509 certificates is
2690 x500UniqueIdentifier, not uniqueIdentifier.
2691 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2694 *) Add an "init" command to the ENGINE config module and auto initialize
2695 ENGINEs. Without any "init" command the ENGINE will be initialized
2696 after all ctrl commands have been executed on it. If init=1 the
2697 ENGINE is initailized at that point (ctrls before that point are run
2698 on the uninitialized ENGINE and after on the initialized one). If
2699 init=0 then the ENGINE will not be iniatialized at all.
2702 *) Fix the 'app_verify_callback' interface so that the user-defined
2703 argument is actually passed to the callback: In the
2704 SSL_CTX_set_cert_verify_callback() prototype, the callback
2705 declaration has been changed from
2708 int (*cb)(X509_STORE_CTX *,void *);
2709 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2710 i=s->ctx->app_verify_callback(&ctx)
2711 has been changed into
2712 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2714 To update applications using SSL_CTX_set_cert_verify_callback(),
2715 a dummy argument can be added to their callback functions.
2716 [D. K. Smetters <smetters@parc.xerox.com>]
2718 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2719 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2721 *) Add and OPENSSL_LOAD_CONF define which will cause
2722 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2723 This allows older applications to transparently support certain
2724 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2725 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2726 load the config file and OPENSSL_add_all_algorithms_conf() which will
2727 always load it have also been added.
2730 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2731 Adjust NIDs and EVP layer.
2732 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2734 *) Config modules support in openssl utility.
2736 Most commands now load modules from the config file,
2737 though in a few (such as version) this isn't done
2738 because it couldn't be used for anything.
2740 In the case of ca and req the config file used is
2741 the same as the utility itself: that is the -config
2742 command line option can be used to specify an
2746 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2747 use "openssl_conf" if filename is NULL use default openssl config file.
2750 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2751 config section name. Add a new flag to tolerate a missing config file
2752 and move code to CONF_modules_load_file().
2755 *) Support for crypto accelerator cards from Accelerated Encryption
2756 Processing, www.aep.ie. (Use engine 'aep')
2757 The support was copied from 0.9.6c [engine] and adapted/corrected
2758 to work with the new engine framework.
2759 [AEP Inc. and Richard Levitte]
2761 *) Support for SureWare crypto accelerator cards from Baltimore
2762 Technologies. (Use engine 'sureware')
2763 The support was copied from 0.9.6c [engine] and adapted
2764 to work with the new engine framework.
2767 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2768 make the newer ENGINE framework commands for the CHIL engine work.
2769 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2771 *) Make it possible to produce shared libraries on ReliantUNIX.
2772 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2774 *) Add the configuration target debug-linux-ppro.
2775 Make 'openssl rsa' use the general key loading routines
2776 implemented in apps.c, and make those routines able to
2777 handle the key format FORMAT_NETSCAPE and the variant
2779 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2781 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2782 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2784 *) Add -keyform to rsautl, and document -engine.
2785 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2787 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2788 BIO_R_NO_SUCH_FILE error code rather than the generic
2789 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2792 *) Add new functions
2794 ERR_peek_last_error_line
2795 ERR_peek_last_error_line_data.
2796 These are similar to
2799 ERR_peek_error_line_data,
2800 but report on the latest error recorded rather than the first one
2801 still in the error queue.
2802 [Ben Laurie, Bodo Moeller]
2804 *) default_algorithms option in ENGINE config module. This allows things
2806 default_algorithms = ALL
2807 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2810 *) Prelminary ENGINE config module.
2813 *) New experimental application configuration code.
2816 *) Change the AES code to follow the same name structure as all other
2817 symmetric ciphers, and behave the same way. Move everything to
2818 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2819 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2821 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2822 [Ben Laurie and Theo de Raadt]
2824 *) Add option to output public keys in req command.
2825 [Massimiliano Pala madwolf@openca.org]
2827 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2828 (up to about 10% better than before for P-192 and P-224).
2831 *) New functions/macros
2833 SSL_CTX_set_msg_callback(ctx, cb)
2834 SSL_CTX_set_msg_callback_arg(ctx, arg)
2835 SSL_set_msg_callback(ssl, cb)
2836 SSL_set_msg_callback_arg(ssl, arg)
2838 to request calling a callback function
2840 void cb(int write_p, int version, int content_type,
2841 const void *buf, size_t len, SSL *ssl, void *arg)
2843 whenever a protocol message has been completely received
2844 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2845 protocol version according to which the SSL library interprets
2846 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2847 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2848 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2849 specification (change_cipher_spec(20), alert(21), handshake(22)).
2850 'buf' and 'len' point to the actual message, 'ssl' to the
2851 SSL object, and 'arg' is the application-defined value set by
2852 SSL[_CTX]_set_msg_callback_arg().
2854 'openssl s_client' and 'openssl s_server' have new '-msg' options
2855 to enable a callback that displays all protocol messages.
2858 *) Change the shared library support so shared libraries are built as
2859 soon as the corresponding static library is finished, and thereby get
2860 openssl and the test programs linked against the shared library.
2861 This still only happens when the keyword "shard" has been given to
2862 the configuration scripts.
2864 NOTE: shared library support is still an experimental thing, and
2865 backward binary compatibility is still not guaranteed.
2866 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2868 *) Add support for Subject Information Access extension.
2869 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2871 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2872 additional bytes when new memory had to be allocated, not just
2873 when reusing an existing buffer.
2876 *) New command line and configuration option 'utf8' for the req command.
2877 This allows field values to be specified as UTF8 strings.
2880 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2881 runs for the former and machine-readable output for the latter.
2884 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2885 of the e-mail address in the DN (i.e., it will go into a certificate
2886 extension only). The new configuration file option 'email_in_dn = no'
2887 has the same effect.
2888 [Massimiliano Pala madwolf@openca.org]
2890 *) Change all functions with names starting with des_ to be starting
2891 with DES_ instead. Add wrappers that are compatible with libdes,
2892 but are named _ossl_old_des_*. Finally, add macros that map the
2893 des_* symbols to the corresponding _ossl_old_des_* if libdes
2894 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2895 desired, the des_* symbols will be mapped to DES_*, with one
2898 Since we provide two compatibility mappings, the user needs to
2899 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2900 compatibility is desired. The default (i.e., when that macro
2901 isn't defined) is OpenSSL 0.9.6c compatibility.
2903 There are also macros that enable and disable the support of old
2904 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2905 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2906 are defined, the default will apply: to support the old des routines.
2908 In either case, one must include openssl/des.h to get the correct
2909 definitions. Do not try to just include openssl/des_old.h, that
2912 NOTE: This is a major break of an old API into a new one. Software
2913 authors are encouraged to switch to the DES_ style functions. Some
2914 time in the future, des_old.h and the libdes compatibility functions
2915 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2916 default), and then completely removed.
2919 *) Test for certificates which contain unsupported critical extensions.
2920 If such a certificate is found during a verify operation it is
2921 rejected by default: this behaviour can be overridden by either
2922 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2923 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2924 X509_supported_extension() has also been added which returns 1 if a
2925 particular extension is supported.
2928 *) Modify the behaviour of EVP cipher functions in similar way to digests
2929 to retain compatibility with existing code.
2932 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2933 compatibility with existing code. In particular the 'ctx' parameter does
2934 not have to be to be initialized before the call to EVP_DigestInit() and
2935 it is tidied up after a call to EVP_DigestFinal(). New function
2936 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2937 EVP_MD_CTX_copy() changed to not require the destination to be
2938 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2939 requires the destination to be valid.
2941 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2942 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2945 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2946 so that complete 'Handshake' protocol structures are kept in memory
2947 instead of overwriting 'msg_type' and 'length' with 'body' data.
2950 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2951 [Massimo Santin via Richard Levitte]
2953 *) Major restructuring to the underlying ENGINE code. This includes
2954 reduction of linker bloat, separation of pure "ENGINE" manipulation
2955 (initialisation, etc) from functionality dealing with implementations
2956 of specific crypto iterfaces. This change also introduces integrated
2957 support for symmetric ciphers and digest implementations - so ENGINEs
2958 can now accelerate these by providing EVP_CIPHER and EVP_MD
2959 implementations of their own. This is detailed in crypto/engine/README
2960 as it couldn't be adequately described here. However, there are a few
2961 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2962 were changed in the original introduction of ENGINE code have now
2963 reverted back - the hooking from this code to ENGINE is now a good
2964 deal more passive and at run-time, operations deal directly with
2965 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2966 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2967 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2968 they were not being used by the framework as there is no concept of a
2969 BIGNUM_METHOD and they could not be generalised to the new
2970 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2971 ENGINE_cpy() has been removed as it cannot be consistently defined in
2975 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2978 *) Change mkdef.pl to sort symbols that get the same entry number,
2979 and make sure the automatically generated functions ERR_load_*
2980 become part of libeay.num as well.
2983 *) New function SSL_renegotiate_pending(). This returns true once
2984 renegotiation has been requested (either SSL_renegotiate() call
2985 or HelloRequest/ClientHello receveived from the peer) and becomes
2986 false once a handshake has been completed.
2987 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2988 sends a HelloRequest, but does not ensure that a handshake takes
2989 place. SSL_renegotiate_pending() is useful for checking if the
2990 client has followed the request.)
2993 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2994 By default, clients may request session resumption even during
2995 renegotiation (if session ID contexts permit); with this option,
2996 session resumption is possible only in the first handshake.
2998 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2999 more bits available for options that should not be part of
3000 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3003 *) Add some demos for certificate and certificate request creation.
3006 *) Make maximum certificate chain size accepted from the peer application
3007 settable (SSL*_get/set_max_cert_list()), as proposed by
3008 "Douglas E. Engert" <deengert@anl.gov>.
3011 *) Add support for shared libraries for Unixware-7
3012 (Boyd Lynn Gerber <gerberb@zenez.com>).
3015 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3016 be done prior to destruction. Use this to unload error strings from
3017 ENGINEs that load their own error strings. NB: This adds two new API
3018 functions to "get" and "set" this destroy handler in an ENGINE.
3021 *) Alter all existing ENGINE implementations (except "openssl" and
3022 "openbsd") to dynamically instantiate their own error strings. This
3023 makes them more flexible to be built both as statically-linked ENGINEs
3024 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3025 Also, add stub code to each that makes building them as self-contained
3026 shared-libraries easier (see README.ENGINE).
3029 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3030 implementations into applications that are completely implemented in
3031 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3032 commands that can be used to configure what shared-library to load and
3033 to control aspects of the way it is handled. Also, made an update to
3034 the README.ENGINE file that brings its information up-to-date and
3035 provides some information and instructions on the "dynamic" ENGINE
3036 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3039 *) Make it possible to unload ranges of ERR strings with a new
3040 "ERR_unload_strings" function.
3043 *) Add a copy() function to EVP_MD.
3046 *) Make EVP_MD routines take a context pointer instead of just the
3047 md_data void pointer.
3050 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3051 that the digest can only process a single chunk of data
3052 (typically because it is provided by a piece of
3053 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3054 is only going to provide a single chunk of data, and hence the
3055 framework needn't accumulate the data for oneshot drivers.
3058 *) As with "ERR", make it possible to replace the underlying "ex_data"
3059 functions. This change also alters the storage and management of global
3060 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3061 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3062 index counters. The API functions that use this state have been changed
3063 to take a "class_index" rather than pointers to the class's local STACK
3064 and counter, and there is now an API function to dynamically create new
3065 classes. This centralisation allows us to (a) plug a lot of the
3066 thread-safety problems that existed, and (b) makes it possible to clean
3067 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3068 such data would previously have always leaked in application code and
3069 workarounds were in place to make the memory debugging turn a blind eye
3070 to it. Application code that doesn't use this new function will still
3071 leak as before, but their memory debugging output will announce it now
3072 rather than letting it slide.
3074 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3075 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3076 has a return value to indicate success or failure.
3079 *) Make it possible to replace the underlying "ERR" functions such that the
3080 global state (2 LHASH tables and 2 locks) is only used by the "default"
3081 implementation. This change also adds two functions to "get" and "set"
3082 the implementation prior to it being automatically set the first time
3083 any other ERR function takes place. Ie. an application can call "get",
3084 pass the return value to a module it has just loaded, and that module
3085 can call its own "set" function using that value. This means the
3086 module's "ERR" operations will use (and modify) the error state in the
3087 application and not in its own statically linked copy of OpenSSL code.
3090 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3091 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3092 the operation, and provides a more encapsulated way for external code
3093 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3094 to use these functions rather than manually incrementing the counts.
3096 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3099 *) Add EVP test program.
3102 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3105 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3106 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3107 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3108 These allow a CRL to be built without having to access X509_CRL fields
3109 directly. Modify 'ca' application to use new functions.
3112 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3113 bug workarounds. Rollback attack detection is a security feature.
3114 The problem will only arise on OpenSSL servers when TLSv1 is not
3115 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3116 Software authors not wanting to support TLSv1 will have special reasons
3117 for their choice and can explicitly enable this option.
3118 [Bodo Moeller, Lutz Jaenicke]
3120 *) Rationalise EVP so it can be extended: don't include a union of
3121 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3122 (similar to those existing for EVP_CIPHER_CTX).
3127 EVP_MD_CTX_init(&md); /* new function call */
3128 EVP_DigestInit(&md, EVP_sha1());
3129 EVP_DigestUpdate(&md, in, len);
3130 EVP_DigestFinal(&md, out, NULL);
3131 EVP_MD_CTX_cleanup(&md); /* new function call */
3135 *) Make DES key schedule conform to the usual scheme, as well as
3136 correcting its structure. This means that calls to DES functions
3137 now have to pass a pointer to a des_key_schedule instead of a
3138 plain des_key_schedule (which was actually always a pointer
3141 des_key_schedule ks;
3143 des_set_key_checked(..., &ks);
3144 des_ncbc_encrypt(..., &ks, ...);
3146 (Note that a later change renames 'des_...' into 'DES_...'.)
3149 *) Initial reduction of linker bloat: the use of some functions, such as
3150 PEM causes large amounts of unused functions to be linked in due to
3151 poor organisation. For example pem_all.c contains every PEM function
3152 which has a knock on effect of linking in large amounts of (unused)
3153 ASN1 code. Grouping together similar functions and splitting unrelated
3154 functions prevents this.
3157 *) Cleanup of EVP macros.
3160 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3161 correct _ecb suffix.
3164 *) Add initial OCSP responder support to ocsp application. The
3165 revocation information is handled using the text based index
3166 use by the ca application. The responder can either handle
3167 requests generated internally, supplied in files (for example
3168 via a CGI script) or using an internal minimal server.
3171 *) Add configuration choices to get zlib compression for TLS.
3174 *) Changes to Kerberos SSL for RFC 2712 compliance:
3175 1. Implemented real KerberosWrapper, instead of just using
3176 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3177 2. Implemented optional authenticator field of KerberosWrapper.
3179 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3180 and authenticator structs; see crypto/krb5/.
3182 Generalized Kerberos calls to support multiple Kerberos libraries.
3183 [Vern Staats <staatsvr@asc.hpc.mil>,
3184 Jeffrey Altman <jaltman@columbia.edu>
3185 via Richard Levitte]
3187 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3188 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3189 values for each of the key sizes rather than having just
3190 parameters (and 'speed' generating keys each time).
3193 *) Speed up EVP routines.
3196 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3197 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3198 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3199 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3201 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3202 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3203 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3206 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3208 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3211 *) Added the OS2-EMX target.
3212 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3214 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3215 to support NCONF routines in extension code. New function CONF_set_nconf()
3216 to allow functions which take an NCONF to also handle the old LHASH
3217 structure: this means that the old CONF compatible routines can be
3218 retained (in particular wrt extensions) without having to duplicate the
3219 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3222 *) Enhance the general user interface with mechanisms for inner control
3223 and with possibilities to have yes/no kind of prompts.
3226 *) Change all calls to low level digest routines in the library and
3227 applications to use EVP. Add missing calls to HMAC_cleanup() and
3228 don't assume HMAC_CTX can be copied using memcpy().
3229 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3231 *) Add the possibility to control engines through control names but with
3232 arbitrary arguments instead of just a string.
3233 Change the key loaders to take a UI_METHOD instead of a callback
3234 function pointer. NOTE: this breaks binary compatibility with earlier
3235 versions of OpenSSL [engine].
3236 Adapt the nCipher code for these new conditions and add a card insertion
3240 *) Enhance the general user interface with mechanisms to better support
3241 dialog box interfaces, application-defined prompts, the possibility
3242 to use defaults (for example default passwords from somewhere else)
3243 and interrupts/cancellations.
3246 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3247 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3250 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3251 tidy up some unnecessarily weird code in 'sk_new()').
3252 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3254 *) Change the key loading routines for ENGINEs to use the same kind
3255 callback (pem_password_cb) as all other routines that need this
3259 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3260 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3261 than this minimum value is recommended.
3264 *) New random seeder for OpenVMS, using the system process statistics
3265 that are easily reachable.
3268 *) Windows apparently can't transparently handle global
3269 variables defined in DLLs. Initialisations such as:
3271 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3273 wont compile. This is used by the any applications that need to
3274 declare their own ASN1 modules. This was fixed by adding the option
3275 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3276 needed for static libraries under Win32.
3279 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3280 setting of purpose and trust fields. New X509_STORE trust and
3281 purpose functions and tidy up setting in other SSL functions.
3284 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3285 structure. These are inherited by X509_STORE_CTX when it is
3286 initialised. This allows various defaults to be set in the
3287 X509_STORE structure (such as flags for CRL checking and custom
3288 purpose or trust settings) for functions which only use X509_STORE_CTX
3289 internally such as S/MIME.
3291 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3292 trust settings if they are not set in X509_STORE. This allows X509_STORE
3293 purposes and trust (in S/MIME for example) to override any set by default.
3295 Add command line options for CRL checking to smime, s_client and s_server
3299 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3300 are set then the CRL is looked up in the X509_STORE structure and
3301 its validity and signature checked, then if the certificate is found
3302 in the CRL the verify fails with a revoked error.
3304 Various new CRL related callbacks added to X509_STORE_CTX structure.
3306 Command line options added to 'verify' application to support this.
3308 This needs some additional work, such as being able to handle multiple
3309 CRLs with different times, extension based lookup (rather than just
3310 by subject name) and ultimately more complete V2 CRL extension
3314 *) Add a general user interface API (crypto/ui/). This is designed
3315 to replace things like des_read_password and friends (backward
3316 compatibility functions using this new API are provided).
3317 The purpose is to remove prompting functions from the DES code
3318 section as well as provide for prompting through dialog boxes in
3319 a window system and the like.
3322 *) Add "ex_data" support to ENGINE so implementations can add state at a
3323 per-structure level rather than having to store it globally.
3326 *) Make it possible for ENGINE structures to be copied when retrieved by
3327 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3328 This causes the "original" ENGINE structure to act like a template,
3329 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3330 operational state can be localised to each ENGINE structure, despite the
3331 fact they all share the same "methods". New ENGINE structures returned in
3332 this case have no functional references and the return value is the single
3333 structural reference. This matches the single structural reference returned
3334 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3338 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3339 needs to match any other type at all we need to manually clear the
3343 *) Changes to the "openssl engine" utility to include;
3344 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3345 about an ENGINE's available control commands.
3346 - executing control commands from command line arguments using the
3347 '-pre' and '-post' switches. '-post' is only used if '-t' is
3348 specified and the ENGINE is successfully initialised. The syntax for
3349 the individual commands are colon-separated, for example;
3350 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3353 *) New dynamic control command support for ENGINEs. ENGINEs can now
3354 declare their own commands (numbers), names (strings), descriptions,
3355 and input types for run-time discovery by calling applications. A
3356 subset of these commands are implicitly classed as "executable"
3357 depending on their input type, and only these can be invoked through
3358 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3359 can be based on user input, config files, etc). The distinction is
3360 that "executable" commands cannot return anything other than a boolean
3361 result and can only support numeric or string input, whereas some
3362 discoverable commands may only be for direct use through
3363 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3364 pointers, or other custom uses. The "executable" commands are to
3365 support parameterisations of ENGINE behaviour that can be
3366 unambiguously defined by ENGINEs and used consistently across any
3367 OpenSSL-based application. Commands have been added to all the
3368 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3369 control over shared-library paths without source code alterations.
3372 *) Changed all ENGINE implementations to dynamically allocate their
3373 ENGINEs rather than declaring them statically. Apart from this being
3374 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3375 this also allows the implementations to compile without using the
3376 internal engine_int.h header.
3379 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3380 'const' value. Any code that should be able to modify a RAND_METHOD
3381 should already have non-const pointers to it (ie. they should only
3382 modify their own ones).
3385 *) Made a variety of little tweaks to the ENGINE code.
3386 - "atalla" and "ubsec" string definitions were moved from header files
3387 to C code. "nuron" string definitions were placed in variables
3388 rather than hard-coded - allowing parameterisation of these values
3389 later on via ctrl() commands.
3390 - Removed unused "#if 0"'d code.
3391 - Fixed engine list iteration code so it uses ENGINE_free() to release
3392 structural references.
3393 - Constified the RAND_METHOD element of ENGINE structures.
3394 - Constified various get/set functions as appropriate and added
3395 missing functions (including a catch-all ENGINE_cpy that duplicates
3396 all ENGINE values onto a new ENGINE except reference counts/state).
3397 - Removed NULL parameter checks in get/set functions. Setting a method
3398 or function to NULL is a way of cancelling out a previously set
3399 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3400 and doesn't justify the extra error symbols and code.
3401 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3402 flags from engine_int.h to engine.h.
3403 - Changed prototypes for ENGINE handler functions (init(), finish(),
3404 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3407 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3408 to the algorithm using long division. The binary algorithm can be
3409 used only if the modulus is odd. On 32-bit systems, it is faster
3410 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3411 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3412 up to 450 bits. In 64-bit environments, the binary algorithm
3413 appears to be advantageous for much longer moduli; here we use it
3414 for moduli up to 2048 bits.
3417 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3418 could not support the combine flag in choice fields.
3421 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3422 extensions from a certificate request to the certificate.
3425 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3426 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3427 file: this allows the display of the certificate about to be
3428 signed to be customised, to allow certain fields to be included
3429 or excluded and extension details. The old system didn't display
3430 multicharacter strings properly, omitted fields not in the policy
3431 and couldn't display additional details such as extensions.
3434 *) Function EC_POINTs_mul for multiple scalar multiplication
3435 of an arbitrary number of elliptic curve points
3436 \sum scalars[i]*points[i],
3437 optionally including the generator defined for the EC_GROUP:
3438 scalar*generator + \sum scalars[i]*points[i].
3440 EC_POINT_mul is a simple wrapper function for the typical case
3441 that the point list has just one item (besides the optional
3445 *) First EC_METHODs for curves over GF(p):
3447 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3448 operations and provides various method functions that can also
3449 operate with faster implementations of modular arithmetic.
3451 EC_GFp_mont_method() reuses most functions that are part of
3452 EC_GFp_simple_method, but uses Montgomery arithmetic.
3454 [Bodo Moeller; point addition and point doubling
3455 implementation directly derived from source code provided by
3456 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3458 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3459 crypto/ec/ec_lib.c):
3461 Curves are EC_GROUP objects (with an optional group generator)
3462 based on EC_METHODs that are built into the library.
3464 Points are EC_POINT objects based on EC_GROUP objects.
3466 Most of the framework would be able to handle curves over arbitrary
3467 finite fields, but as there are no obvious types for fields other
3468 than GF(p), some functions are limited to that for now.
3471 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3472 that the file contains a complete HTTP response.
3475 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3476 change the def and num file printf format specifier from "%-40sXXX"
3477 to "%-39s XXX". The latter will always guarantee a space after the
3478 field while the former will cause them to run together if the field
3479 is 40 of more characters long.
3482 *) Constify the cipher and digest 'method' functions and structures
3483 and modify related functions to take constant EVP_MD and EVP_CIPHER
3487 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3488 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3491 *) Modify EVP_Digest*() routines so they now return values. Although the
3492 internal software routines can never fail additional hardware versions
3496 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3498 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3499 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3502 ERR_R_NESTED_ASN1_ERROR
3504 ERR_R_MISSING_ASN1_EOS
3505 were 4 .. 9, conflicting with
3506 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3508 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3509 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3511 Add new error code 'ERR_R_INTERNAL_ERROR'.
3514 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3518 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3519 sets the subject name for a new request or supersedes the
3520 subject name in a given request. Formats that can be parsed are
3521 'CN=Some Name, OU=myOU, C=IT'
3523 'CN=Some Name/OU=myOU/C=IT'.
3525 Add options '-batch' and '-verbose' to 'openssl req'.
3526 [Massimiliano Pala <madwolf@hackmasters.net>]
3528 *) Introduce the possibility to access global variables through
3529 functions on platform were that's the best way to handle exporting
3530 global variables in shared libraries. To enable this functionality,
3531 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3532 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3533 is normally done by Configure or something similar).
3535 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3536 in the source file (foo.c) like this:
3538 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3539 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3541 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3542 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3544 OPENSSL_DECLARE_GLOBAL(int,foo);
3545 #define foo OPENSSL_GLOBAL_REF(foo)
3546 OPENSSL_DECLARE_GLOBAL(double,bar);
3547 #define bar OPENSSL_GLOBAL_REF(bar)
3549 The #defines are very important, and therefore so is including the
3550 header file everywhere where the defined globals are used.
3552 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3553 of ASN.1 items, but that structure is a bit different.
3555 The largest change is in util/mkdef.pl which has been enhanced with
3556 better and easier to understand logic to choose which symbols should
3557 go into the Windows .def files as well as a number of fixes and code
3558 cleanup (among others, algorithm keywords are now sorted
3559 lexicographically to avoid constant rewrites).
3562 *) In BN_div() keep a copy of the sign of 'num' before writing the
3563 result to 'rm' because if rm==num the value will be overwritten
3564 and produce the wrong result if 'num' is negative: this caused
3565 problems with BN_mod() and BN_nnmod().
3568 *) Function OCSP_request_verify(). This checks the signature on an
3569 OCSP request and verifies the signer certificate. The signer
3570 certificate is just checked for a generic purpose and OCSP request
3574 *) Add OCSP_check_validity() function to check the validity of OCSP
3575 responses. OCSP responses are prepared in real time and may only
3576 be a few seconds old. Simply checking that the current time lies
3577 between thisUpdate and nextUpdate max reject otherwise valid responses
3578 caused by either OCSP responder or client clock inaccuracy. Instead
3579 we allow thisUpdate and nextUpdate to fall within a certain period of
3580 the current time. The age of the response can also optionally be
3581 checked. Two new options -validity_period and -status_age added to
3585 *) If signature or public key algorithm is unrecognized print out its
3586 OID rather that just UNKNOWN.
3589 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3590 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3591 ID to be generated from the issuer certificate alone which can then be
3592 passed to OCSP_id_issuer_cmp().
3595 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3596 ASN1 modules to export functions returning ASN1_ITEM pointers
3597 instead of the ASN1_ITEM structures themselves. This adds several
3598 new macros which allow the underlying ASN1 function/structure to
3599 be accessed transparently. As a result code should not use ASN1_ITEM
3600 references directly (such as &X509_it) but instead use the relevant
3601 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3602 use of the new ASN1 code on platforms where exporting structures
3603 is problematical (for example in shared libraries) but exporting
3604 functions returning pointers to structures is not.
3607 *) Add support for overriding the generation of SSL/TLS session IDs.
3608 These callbacks can be registered either in an SSL_CTX or per SSL.
3609 The purpose of this is to allow applications to control, if they wish,
3610 the arbitrary values chosen for use as session IDs, particularly as it
3611 can be useful for session caching in multiple-server environments. A
3612 command-line switch for testing this (and any client code that wishes
3613 to use such a feature) has been added to "s_server".
3614 [Geoff Thorpe, Lutz Jaenicke]
3616 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3617 of the form '#if defined(...) || defined(...) || ...' and
3618 '#if !defined(...) && !defined(...) && ...'. This also avoids
3619 the growing number of special cases it was previously handling.
3622 *) Make all configuration macros available for application by making
3623 sure they are available in opensslconf.h, by giving them names starting
3624 with "OPENSSL_" to avoid conflicts with other packages and by making
3625 sure e_os2.h will cover all platform-specific cases together with
3627 Additionally, it is now possible to define configuration/platform-
3628 specific names (called "system identities"). In the C code, these
3629 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3630 macro with the name beginning with "OPENSSL_SYS_", which is determined
3631 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3635 *) New option -set_serial to 'req' and 'x509' this allows the serial
3636 number to use to be specified on the command line. Previously self
3637 signed certificates were hard coded with serial number 0 and the
3638 CA options of 'x509' had to use a serial number in a file which was
3642 *) New options to 'ca' utility to support V2 CRL entry extensions.
3643 Currently CRL reason, invalidity date and hold instruction are
3644 supported. Add new CRL extensions to V3 code and some new objects.
3647 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3648 disable standard block padding (aka PKCS#5 padding) in the EVP
3649 API, which was previously mandatory. This means that the data is
3650 not padded in any way and so the total length much be a multiple
3651 of the block size, otherwise an error occurs.
3654 *) Initial (incomplete) OCSP SSL support.
3657 *) New function OCSP_parse_url(). This splits up a URL into its host,
3658 port and path components: primarily to parse OCSP URLs. New -url
3659 option to ocsp utility.
3662 *) New nonce behavior. The return value of OCSP_check_nonce() now
3663 reflects the various checks performed. Applications can decide
3664 whether to tolerate certain situations such as an absent nonce
3665 in a response when one was present in a request: the ocsp application
3666 just prints out a warning. New function OCSP_add1_basic_nonce()
3667 this is to allow responders to include a nonce in a response even if
3668 the request is nonce-less.
3671 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3672 skipped when using openssl x509 multiple times on a single input file,
3673 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3676 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3677 set string type: to handle setting ASN1_TIME structures. Fix ca
3678 utility to correctly initialize revocation date of CRLs.
3681 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3682 the clients preferred ciphersuites and rather use its own preferences.
3683 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3684 Internet Explorer by ensuring unchanged hash method during stepup.
3685 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3688 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3689 to aes and add a new 'exist' option to print out symbols that don't
3693 *) Additional options to ocsp utility to allow flags to be set and
3694 additional certificates supplied.
3697 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3698 OCSP client a number of certificate to only verify the response
3702 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3703 handle the new API. Currently only ECB, CBC modes supported. Add new
3706 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3707 Encryption Standard (AES) Ciphersuites for Transport Layer
3708 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3709 not enabled by default and were not part of the "ALL" ciphersuite
3710 alias because they were not yet official; they could be
3711 explicitly requested by specifying the "AESdraft" ciphersuite
3712 group alias. In the final release of OpenSSL 0.9.7, the group
3713 alias is called "AES" and is part of "ALL".)
3714 [Ben Laurie, Steve Henson, Bodo Moeller]
3716 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3717 request to response.
3720 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3721 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3722 extract information from a certificate request. OCSP_response_create()
3723 creates a response and optionally adds a basic response structure.
3724 OCSP_basic_add1_status() adds a complete single response to a basic
3725 response and returns the OCSP_SINGLERESP structure just added (to allow
3726 extensions to be included for example). OCSP_basic_add1_cert() adds a
3727 certificate to a basic response and OCSP_basic_sign() signs a basic
3728 response with various flags. New helper functions ASN1_TIME_check()
3729 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3730 (converts ASN1_TIME to GeneralizedTime).
3733 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3734 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3735 structure from a certificate. X509_pubkey_digest() digests the public_key
3736 contents: this is used in various key identifiers.
3739 *) Make sk_sort() tolerate a NULL argument.
3740 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3742 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3743 passed by the function are trusted implicitly. If any of them signed the
3744 response then it is assumed to be valid and is not verified.
3747 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3748 to data. This was previously part of the PKCS7 ASN1 code. This
3749 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3750 [Steve Henson, reported by Kenneth R. Robinette
3751 <support@securenetterm.com>]
3753 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3754 routines: without these tracing memory leaks is very painful.
3755 Fix leaks in PKCS12 and PKCS7 routines.
3758 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3759 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3760 effectively meant GeneralizedTime would never be used. Now it
3761 is initialised to -1 but X509_time_adj() now has to check the value
3762 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3763 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3764 [Steve Henson, reported by Kenneth R. Robinette
3765 <support@securenetterm.com>]
3767 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3768 result in a zero length in the ASN1_INTEGER structure which was
3769 not consistent with the structure when d2i_ASN1_INTEGER() was used
3770 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3771 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3772 where it did not print out a minus for negative ASN1_INTEGER.
3775 *) Add summary printout to ocsp utility. The various functions which
3776 convert status values to strings have been renamed to:
3777 OCSP_response_status_str(), OCSP_cert_status_str() and
3778 OCSP_crl_reason_str() and are no longer static. New options
3779 to verify nonce values and to disable verification. OCSP response
3780 printout format cleaned up.
3783 *) Add additional OCSP certificate checks. These are those specified
3784 in RFC2560. This consists of two separate checks: the CA of the
3785 certificate being checked must either be the OCSP signer certificate
3786 or the issuer of the OCSP signer certificate. In the latter case the
3787 OCSP signer certificate must contain the OCSP signing extended key
3788 usage. This check is performed by attempting to match the OCSP
3789 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3790 in the OCSP_CERTID structures of the response.
3793 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3794 and related routines. This uses the standard OpenSSL certificate
3795 verify routines to perform initial checks (just CA validity) and
3796 to obtain the certificate chain. Then additional checks will be
3797 performed on the chain. Currently the root CA is checked to see
3798 if it is explicitly trusted for OCSP signing. This is used to set
3799 a root CA as a global signing root: that is any certificate that
3800 chains to that CA is an acceptable OCSP signing certificate.
3803 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3804 extensions from a separate configuration file.
3805 As when reading extensions from the main configuration file,
3806 the '-extensions ...' option may be used for specifying the
3808 [Massimiliano Pala <madwolf@comune.modena.it>]
3810 *) New OCSP utility. Allows OCSP requests to be generated or
3811 read. The request can be sent to a responder and the output
3812 parsed, outputed or printed in text form. Not complete yet:
3813 still needs to check the OCSP response validity.
3816 *) New subcommands for 'openssl ca':
3817 'openssl ca -status <serial>' prints the status of the cert with
3818 the given serial number (according to the index file).
3819 'openssl ca -updatedb' updates the expiry status of certificates
3821 [Massimiliano Pala <madwolf@comune.modena.it>]
3823 *) New '-newreq-nodes' command option to CA.pl. This is like
3824 '-newreq', but calls 'openssl req' with the '-nodes' option
3825 so that the resulting key is not encrypted.
3826 [Damien Miller <djm@mindrot.org>]
3828 *) New configuration for the GNU Hurd.
3829 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3831 *) Initial code to implement OCSP basic response verify. This
3832 is currently incomplete. Currently just finds the signer's
3833 certificate and verifies the signature on the response.
3836 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3837 value of OPENSSLDIR. This is available via the new '-d' option
3838 to 'openssl version', and is also included in 'openssl version -a'.
3841 *) Allowing defining memory allocation callbacks that will be given
3842 file name and line number information in additional arguments
3843 (a const char* and an int). The basic functionality remains, as
3844 well as the original possibility to just replace malloc(),
3845 realloc() and free() by functions that do not know about these
3846 additional arguments. To register and find out the current
3847 settings for extended allocation functions, the following
3848 functions are provided:
3850 CRYPTO_set_mem_ex_functions
3851 CRYPTO_set_locked_mem_ex_functions
3852 CRYPTO_get_mem_ex_functions
3853 CRYPTO_get_locked_mem_ex_functions
3855 These work the same way as CRYPTO_set_mem_functions and friends.
3856 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3857 extended allocation function is enabled.
3858 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3859 a conventional allocation function is enabled.
3860 [Richard Levitte, Bodo Moeller]
3862 *) Finish off removing the remaining LHASH function pointer casts.
3863 There should no longer be any prototype-casting required when using
3864 the LHASH abstraction, and any casts that remain are "bugs". See
3865 the callback types and macros at the head of lhash.h for details
3866 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3869 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3870 If /dev/[u]random devices are not available or do not return enough
3871 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3873 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3874 /etc/entropy will be queried once each in this sequence, quering stops
3875 when enough entropy was collected without querying more sockets.
3878 *) Change the Unix RAND_poll() variant to be able to poll several
3879 random devices, as specified by DEVRANDOM, until a sufficient amount
3880 of data has been collected. We spend at most 10 ms on each file
3881 (select timeout) and read in non-blocking mode. DEVRANDOM now
3882 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3883 (previously it was just the string "/dev/urandom"), so on typical
3884 platforms the 10 ms delay will never occur.
3885 Also separate out the Unix variant to its own file, rand_unix.c.
3886 For VMS, there's a currently-empty rand_vms.c.
3889 *) Move OCSP client related routines to ocsp_cl.c. These
3890 provide utility functions which an application needing
3891 to issue a request to an OCSP responder and analyse the
3892 response will typically need: as opposed to those which an
3893 OCSP responder itself would need which will be added later.
3895 OCSP_request_sign() signs an OCSP request with an API similar
3896 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3897 response. OCSP_response_get1_basic() extracts basic response
3898 from response. OCSP_resp_find_status(): finds and extracts status
3899 information from an OCSP_CERTID structure (which will be created
3900 when the request structure is built). These are built from lower
3901 level functions which work on OCSP_SINGLERESP structures but
3902 wont normally be used unless the application wishes to examine
3903 extensions in the OCSP response for example.
3905 Replace nonce routines with a pair of functions.
3906 OCSP_request_add1_nonce() adds a nonce value and optionally
3907 generates a random value. OCSP_check_nonce() checks the
3908 validity of the nonce in an OCSP response.
3911 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3912 This doesn't copy the supplied OCSP_CERTID and avoids the
3913 need to free up the newly created id. Change return type
3914 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3915 This can then be used to add extensions to the request.
3916 Deleted OCSP_request_new(), since most of its functionality
3917 is now in OCSP_REQUEST_new() (and the case insensitive name
3918 clash) apart from the ability to set the request name which
3919 will be added elsewhere.
3922 *) Update OCSP API. Remove obsolete extensions argument from
3923 various functions. Extensions are now handled using the new
3924 OCSP extension code. New simple OCSP HTTP function which
3925 can be used to send requests and parse the response.
3928 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3929 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3930 uses the special reorder version of SET OF to sort the attributes
3931 and reorder them to match the encoded order. This resolves a long
3932 standing problem: a verify on a PKCS7 structure just after signing
3933 it used to fail because the attribute order did not match the
3934 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3935 it uses the received order. This is necessary to tolerate some broken
3936 software that does not order SET OF. This is handled by encoding
3937 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3938 to produce the required SET OF.
3941 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3942 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3943 files to get correct declarations of the ASN.1 item variables.
3946 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3947 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3948 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3949 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3950 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3951 ASN1_ITEM and no wrapper functions.
3954 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3955 replace the old function pointer based I/O routines. Change most of
3956 the *_d2i_bio() and *_d2i_fp() functions to use these.
3959 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3960 lines, recognice more "algorithms" that can be deselected, and make
3961 it complain about algorithm deselection that isn't recognised.
3964 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3965 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3966 to use new functions. Add NO_ASN1_OLD which can be set to remove
3967 some old style ASN1 functions: this can be used to determine if old
3968 code will still work when these eventually go away.
3971 *) New extension functions for OCSP structures, these follow the
3972 same conventions as certificates and CRLs.
3975 *) New function X509V3_add1_i2d(). This automatically encodes and
3976 adds an extension. Its behaviour can be customised with various
3977 flags to append, replace or delete. Various wrappers added for
3978 certifcates and CRLs.
3981 *) Fix to avoid calling the underlying ASN1 print routine when
3982 an extension cannot be parsed. Correct a typo in the
3983 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3986 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3987 entries for variables.
3990 *) Add functionality to apps/openssl.c for detecting locking
3991 problems: As the program is single-threaded, all we have
3992 to do is register a locking callback using an array for
3993 storing which locks are currently held by the program.
3996 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3997 SSL_get_ex_data_X509_STORE_idx(), which is used in
3998 ssl_verify_cert_chain() and thus can be called at any time
3999 during TLS/SSL handshakes so that thread-safety is essential.
4000 Unfortunately, the ex_data design is not at all suited
4001 for multi-threaded use, so it probably should be abolished.
4004 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4005 [Broadcom, tweaked and integrated by Geoff Thorpe]
4007 *) Move common extension printing code to new function
4008 X509V3_print_extensions(). Reorganise OCSP print routines and
4009 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4012 *) New function X509_signature_print() to remove duplication in some
4016 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4017 set (this was treated exactly the same as SET OF previously). This
4018 is used to reorder the STACK representing the structure to match the
4019 encoding. This will be used to get round a problem where a PKCS7
4020 structure which was signed could not be verified because the STACK
4021 order did not reflect the encoded order.
4024 *) Reimplement the OCSP ASN1 module using the new code.
4027 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4028 for its ASN1 operations. The old style function pointers still exist
4029 for now but they will eventually go away.
4032 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4033 completely replaces the old ASN1 functionality with a table driven
4034 encoder and decoder which interprets an ASN1_ITEM structure describing
4035 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4036 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4037 has also been converted to the new form.
4040 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4041 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4042 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4043 for negative moduli.
4046 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4047 of not touching the result's sign bit.
4050 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4054 *) Changed the LHASH code to use prototypes for callbacks, and created
4055 macros to declare and implement thin (optionally static) functions
4056 that provide type-safety and avoid function pointer casting for the
4057 type-specific callbacks.
4060 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4062 [Veers Staats <staatsvr@asc.hpc.mil>,
4063 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4065 *) Reformat the FAQ so the different questions and answers can be divided
4066 in sections depending on the subject.
4069 *) Have the zlib compression code load ZLIB.DLL dynamically under
4073 *) New function BN_mod_sqrt for computing square roots modulo a prime
4074 (using the probabilistic Tonelli-Shanks algorithm unless
4075 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4076 be handled deterministically).
4077 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4079 *) Make BN_mod_inverse faster by explicitly handling small quotients
4080 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4081 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4084 *) New function BN_kronecker.
4087 *) Fix BN_gcd so that it works on negative inputs; the result is
4088 positive unless both parameters are zero.
4089 Previously something reasonably close to an infinite loop was
4090 possible because numbers could be growing instead of shrinking
4091 in the implementation of Euclid's algorithm.
4094 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4095 sign of the number in question.
4097 Fix BN_is_word(a,w) to work correctly for w == 0.
4099 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4100 because its test if the absolute value of 'a' equals 'w'.
4101 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4102 it exists mostly for use in the implementations of BN_is_zero(),
4103 BN_is_one(), and BN_is_word().
4106 *) New function BN_swap.
4109 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4110 the exponentiation functions are more likely to produce reasonable
4111 results on negative inputs.
4114 *) Change BN_mod_mul so that the result is always non-negative.
4115 Previously, it could be negative if one of the factors was negative;
4116 I don't think anyone really wanted that behaviour.
4119 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4120 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4121 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4122 and add new functions:
4131 BN_mod_lshift1_quick
4135 These functions always generate non-negative results.
4137 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4138 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4140 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4141 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4142 be reduced modulo m.
4143 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4146 The following entry accidentily appeared in the CHANGES file
4147 distributed with OpenSSL 0.9.7. The modifications described in
4148 it do *not* apply to OpenSSL 0.9.7.
4150 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4151 was actually never needed) and in BN_mul(). The removal in BN_mul()
4152 required a small change in bn_mul_part_recursive() and the addition
4153 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4154 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4155 bn_sub_words() and bn_add_words() except they take arrays with
4160 *) In 'openssl passwd', verify passwords read from the terminal
4161 unless the '-salt' option is used (which usually means that
4162 verification would just waste user's time since the resulting
4163 hash is going to be compared with some given password hash)
4164 or the new '-noverify' option is used.
4166 This is an incompatible change, but it does not affect
4167 non-interactive use of 'openssl passwd' (passwords on the command
4168 line, '-stdin' option, '-in ...' option) and thus should not
4172 *) Remove all references to RSAref, since there's no more need for it.
4175 *) Make DSO load along a path given through an environment variable
4176 (SHLIB_PATH) with shl_load().
4179 *) Constify the ENGINE code as a result of BIGNUM constification.
4180 Also constify the RSA code and most things related to it. In a
4181 few places, most notable in the depth of the ASN.1 code, ugly
4182 casts back to non-const were required (to be solved at a later
4186 *) Make it so the openssl application has all engines loaded by default.
4189 *) Constify the BIGNUM routines a little more.
4192 *) Add the following functions:
4194 ENGINE_load_cswift()
4196 ENGINE_load_atalla()
4198 ENGINE_load_builtin_engines()
4200 That way, an application can itself choose if external engines that
4201 are built-in in OpenSSL shall ever be used or not. The benefit is
4202 that applications won't have to be linked with libdl or other dso
4203 libraries unless it's really needed.
4205 Changed 'openssl engine' to load all engines on demand.
4206 Changed the engine header files to avoid the duplication of some
4207 declarations (they differed!).
4210 *) 'openssl engine' can now list capabilities.
4213 *) Better error reporting in 'openssl engine'.
4216 *) Never call load_dh_param(NULL) in s_server.
4219 *) Add engine application. It can currently list engines by name and
4220 identity, and test if they are actually available.
4223 *) Improve RPM specification file by forcing symbolic linking and making
4224 sure the installed documentation is also owned by root.root.
4225 [Damien Miller <djm@mindrot.org>]
4227 *) Give the OpenSSL applications more possibilities to make use of
4228 keys (public as well as private) handled by engines.
4231 *) Add OCSP code that comes from CertCo.
4234 *) Add VMS support for the Rijndael code.
4237 *) Added untested support for Nuron crypto accelerator.
4240 *) Add support for external cryptographic devices. This code was
4241 previously distributed separately as the "engine" branch.
4242 [Geoff Thorpe, Richard Levitte]
4244 *) Rework the filename-translation in the DSO code. It is now possible to
4245 have far greater control over how a "name" is turned into a filename
4246 depending on the operating environment and any oddities about the
4247 different shared library filenames on each system.
4250 *) Support threads on FreeBSD-elf in Configure.
4253 *) Fix for SHA1 assembly problem with MASM: it produces
4254 warnings about corrupt line number information when assembling
4255 with debugging information. This is caused by the overlapping
4257 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4260 NCONF_get_number() has no error checking at all. As a replacement,
4261 NCONF_get_number_e() is defined (_e for "error checking") and is
4262 promoted strongly. The old NCONF_get_number is kept around for
4263 binary backward compatibility.
4264 Make it possible for methods to load from something other than a BIO,
4265 by providing a function pointer that is given a name instead of a BIO.
4266 For example, this could be used to load configuration data from an
4270 *) Fix for non blocking accept BIOs. Added new I/O special reason
4271 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4272 with non blocking I/O was not possible because no retry code was
4273 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4277 *) Added the beginnings of Rijndael support.
4280 *) Fix for bug in DirectoryString mask setting. Add support for
4281 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4282 to allow certificate printing to more controllable, additional
4283 'certopt' option to 'x509' to allow new printing options to be
4287 *) Clean old EAY MD5 hack from e_os.h.
4290 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4292 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4293 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4294 [Joe Orton, Steve Henson]
4296 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4298 *) Fix additional bug revealed by the NISCC test suite:
4300 Stop bug triggering large recursion when presented with
4301 certain ASN.1 tags (CVE-2003-0851)
4304 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4306 *) Fix various bugs revealed by running the NISCC test suite:
4308 Stop out of bounds reads in the ASN1 code when presented with
4309 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4311 If verify callback ignores invalid public key errors don't try to check
4312 certificate signature with the NULL public key.
4316 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4317 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4321 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4322 extra data after the compression methods not only for TLS 1.0
4323 but also for SSL 3.0 (as required by the specification).
4324 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4326 *) Change X509_certificate_type() to mark the key as exported/exportable
4327 when it's 512 *bits* long, not 512 bytes.
4330 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4332 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4333 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4334 a protocol version number mismatch like a decryption error
4335 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4338 *) Turn on RSA blinding by default in the default implementation
4339 to avoid a timing attack. Applications that don't want it can call
4340 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4341 They would be ill-advised to do so in most cases.
4342 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4344 *) Change RSA blinding code so that it works when the PRNG is not
4345 seeded (in this case, the secret RSA exponent is abused as
4346 an unpredictable seed -- if it is not unpredictable, there
4347 is no point in blinding anyway). Make RSA blinding thread-safe
4348 by remembering the creator's thread ID in rsa->blinding and
4349 having all other threads use local one-time blinding factors
4350 (this requires more computation than sharing rsa->blinding, but
4351 avoids excessive locking; and if an RSA object is not shared
4352 between threads, blinding will still be very fast).
4355 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4357 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4358 via timing by performing a MAC computation even if incorrrect
4359 block cipher padding has been found. This is a countermeasure
4360 against active attacks where the attacker has to distinguish
4361 between bad padding and a MAC verification error. (CVE-2003-0078)
4363 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4364 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4365 Martin Vuagnoux (EPFL, Ilion)]
4367 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4369 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4370 memory from it's contents. This is done with a counter that will
4371 place alternating values in each byte. This can be used to solve
4372 two issues: 1) the removal of calls to memset() by highly optimizing
4373 compilers, and 2) cleansing with other values than 0, since those can
4374 be read through on certain media, for example a swap space on disk.
4377 *) Bugfix: client side session caching did not work with external caching,
4378 because the session->cipher setting was not restored when reloading
4379 from the external cache. This problem was masked, when
4380 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4381 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4384 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4385 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4386 [Zeev Lieber <zeev-l@yahoo.com>]
4388 *) Undo an undocumented change introduced in 0.9.6e which caused
4389 repeated calls to OpenSSL_add_all_ciphers() and
4390 OpenSSL_add_all_digests() to be ignored, even after calling
4394 *) Change the default configuration reader to deal with last line not
4395 being properly terminated.
4398 *) Change X509_NAME_cmp() so it applies the special rules on handling
4399 DN values that are of type PrintableString, as well as RDNs of type
4400 emailAddress where the value has the type ia5String.
4401 [stefank@valicert.com via Richard Levitte]
4403 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4404 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4405 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4406 the bitwise-OR of the two for use by the majority of applications
4407 wanting this behaviour, and update the docs. The documented
4408 behaviour and actual behaviour were inconsistent and had been
4409 changing anyway, so this is more a bug-fix than a behavioural
4411 [Geoff Thorpe, diagnosed by Nadav Har'El]
4413 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4414 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4417 *) Fix initialization code race conditions in
4418 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4419 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4420 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4421 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4422 ssl2_get_cipher_by_char(),
4423 ssl3_get_cipher_by_char().
4424 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4426 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4427 the cached sessions are flushed, as the remove_cb() might use ex_data
4428 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4429 (see [openssl.org #212]).
4430 [Geoff Thorpe, Lutz Jaenicke]
4432 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4433 length, instead of the encoding length to d2i_ASN1_OBJECT.
4436 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4438 *) [In 0.9.6g-engine release:]
4439 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4440 [Lynn Gazis <lgazis@rainbow.com>]
4442 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4444 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4445 and get fix the header length calculation.
4446 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4447 Alon Kantor <alonk@checkpoint.com> (and others),
4450 *) Use proper error handling instead of 'assertions' in buffer
4451 overflow checks added in 0.9.6e. This prevents DoS (the
4452 assertions could call abort()).
4453 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4455 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4457 *) Add various sanity checks to asn1_get_length() to reject
4458 the ASN1 length bytes if they exceed sizeof(long), will appear
4459 negative or the content length exceeds the length of the
4461 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4463 *) Fix cipher selection routines: ciphers without encryption had no flags
4464 for the cipher strength set and where therefore not handled correctly
4465 by the selection routines (PR #130).
4468 *) Fix EVP_dsa_sha macro.
4472 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4473 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4474 that was added in OpenSSL 0.9.6d.
4476 As the countermeasure turned out to be incompatible with some
4477 broken SSL implementations, the new option is part of SSL_OP_ALL.
4478 SSL_OP_ALL is usually employed when compatibility with weird SSL
4479 implementations is desired (e.g. '-bugs' option to 's_client' and
4480 's_server'), so the new option is automatically set in many
4484 *) Changes in security patch:
4486 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4487 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4488 Air Force Materiel Command, USAF, under agreement number
4491 *) Add various sanity checks to asn1_get_length() to reject
4492 the ASN1 length bytes if they exceed sizeof(long), will appear
4493 negative or the content length exceeds the length of the
4494 supplied buffer. (CVE-2002-0659)
4495 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4497 *) Assertions for various potential buffer overflows, not known to
4499 [Ben Laurie (CHATS)]
4501 *) Various temporary buffers to hold ASCII versions of integers were
4502 too small for 64 bit platforms. (CVE-2002-0655)
4503 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4505 *) Remote buffer overflow in SSL3 protocol - an attacker could
4506 supply an oversized session ID to a client. (CVE-2002-0656)
4507 [Ben Laurie (CHATS)]
4509 *) Remote buffer overflow in SSL2 protocol - an attacker could
4510 supply an oversized client master key. (CVE-2002-0656)
4511 [Ben Laurie (CHATS)]
4513 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4515 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4516 encoded as NULL) with id-dsa-with-sha1.
4517 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4519 *) Check various X509_...() return values in apps/req.c.
4520 [Nils Larsch <nla@trustcenter.de>]
4522 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4523 an end-of-file condition would erronously be flagged, when the CRLF
4524 was just at the end of a processed block. The bug was discovered when
4525 processing data through a buffering memory BIO handing the data to a
4526 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4527 <ptsekov@syntrex.com> and Nedelcho Stanev.
4530 *) Implement a countermeasure against a vulnerability recently found
4531 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4532 before application data chunks to avoid the use of known IVs
4533 with data potentially chosen by the attacker.
4536 *) Fix length checks in ssl3_get_client_hello().
4539 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4540 to prevent ssl3_read_internal() from incorrectly assuming that
4541 ssl3_read_bytes() found application data while handshake
4542 processing was enabled when in fact s->s3->in_read_app_data was
4543 merely automatically cleared during the initial handshake.
4544 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4546 *) Fix object definitions for Private and Enterprise: they were not
4547 recognized in their shortname (=lowercase) representation. Extend
4548 obj_dat.pl to issue an error when using undefined keywords instead
4549 of silently ignoring the problem (Svenning Sorensen
4550 <sss@sss.dnsalias.net>).
4553 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4554 generators, i.e. generators other than 2 and 5. (Previously, the
4555 code did not properly initialise the 'add' and 'rem' values to
4556 BN_generate_prime().)
4558 In the new general case, we do not insist that 'generator' is
4559 actually a primitive root: This requirement is rather pointless;
4560 a generator of the order-q subgroup is just as good, if not
4564 *) Map new X509 verification errors to alerts. Discovered and submitted by
4565 Tom Wu <tom@arcot.com>.
4568 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4569 returning non-zero before the data has been completely received
4570 when using non-blocking I/O.
4571 [Bodo Moeller; problem pointed out by John Hughes]
4573 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4574 [Ben Laurie, Lutz Jaenicke]
4576 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4577 Yoram Zahavi <YoramZ@gilian.com>).
4580 *) Add information about CygWin 1.3 and on, and preserve proper
4581 configuration for the versions before that.
4582 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4584 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4585 check whether we deal with a copy of a session and do not delete from
4586 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4587 <izhar@checkpoint.com>.
4590 *) Do not store session data into the internal session cache, if it
4591 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4592 flag is set). Proposed by Aslam <aslam@funk.com>.
4595 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4599 *) [In 0.9.6d-engine release:]
4600 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4601 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4603 *) Add the configuration target linux-s390x.
4604 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4606 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4607 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4608 variable as an indication that a ClientHello message has been
4609 received. As the flag value will be lost between multiple
4610 invocations of ssl3_accept when using non-blocking I/O, the
4611 function may not be aware that a handshake has actually taken
4612 place, thus preventing a new session from being added to the
4615 To avoid this problem, we now set s->new_session to 2 instead of
4616 using a local variable.
4617 [Lutz Jaenicke, Bodo Moeller]
4619 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4620 if the SSL_R_LENGTH_MISMATCH error is detected.
4621 [Geoff Thorpe, Bodo Moeller]
4623 *) New 'shared_ldflag' column in Configure platform table.
4626 *) Fix EVP_CIPHER_mode macro.
4627 ["Dan S. Camper" <dan@bti.net>]
4629 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4630 type, we must throw them away by setting rr->length to 0.
4631 [D P Chang <dpc@qualys.com>]
4633 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4635 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4636 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4637 worked incorrectly for those cases where range = 10..._2 and
4638 3*range is two bits longer than range.)
4641 *) Only add signing time to PKCS7 structures if it is not already
4645 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4646 OBJ_ld_ce should be OBJ_id_ce.
4647 Also some ip-pda OIDs in crypto/objects/objects.txt were
4648 incorrect (cf. RFC 3039).
4649 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4651 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4652 returns early because it has nothing to do.
4653 [Andy Schneider <andy.schneider@bjss.co.uk>]
4655 *) [In 0.9.6c-engine release:]
4656 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4657 [Andy Schneider <andy.schneider@bjss.co.uk>]
4659 *) [In 0.9.6c-engine release:]
4660 Add support for Cryptographic Appliance's keyserver technology.
4661 (Use engine 'keyclient')
4662 [Cryptographic Appliances and Geoff Thorpe]
4664 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4665 is called via tools/c89.sh because arguments have to be
4666 rearranged (all '-L' options must appear before the first object
4668 [Richard Shapiro <rshapiro@abinitio.com>]
4670 *) [In 0.9.6c-engine release:]
4671 Add support for Broadcom crypto accelerator cards, backported
4673 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4675 *) [In 0.9.6c-engine release:]
4676 Add support for SureWare crypto accelerator cards from
4677 Baltimore Technologies. (Use engine 'sureware')
4678 [Baltimore Technologies and Mark Cox]
4680 *) [In 0.9.6c-engine release:]
4681 Add support for crypto accelerator cards from Accelerated
4682 Encryption Processing, www.aep.ie. (Use engine 'aep')
4683 [AEP Inc. and Mark Cox]
4685 *) Add a configuration entry for gcc on UnixWare.
4686 [Gary Benson <gbenson@redhat.com>]
4688 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4689 messages are stored in a single piece (fixed-length part and
4690 variable-length part combined) and fix various bugs found on the way.
4693 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4694 instead. BIO_gethostbyname() does not know what timeouts are
4695 appropriate, so entries would stay in cache even when they have
4697 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4699 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4700 faced with a pathologically small ClientHello fragment that does
4701 not contain client_version: Instead of aborting with an error,
4702 simply choose the highest available protocol version (i.e.,
4703 TLS 1.0 unless it is disabled). In practice, ClientHello
4704 messages are never sent like this, but this change gives us
4705 strictly correct behaviour at least for TLS.
4708 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4709 never resets s->method to s->ctx->method when called from within
4710 one of the SSL handshake functions.
4711 [Bodo Moeller; problem pointed out by Niko Baric]
4713 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4714 (sent using the client's version number) if client_version is
4715 smaller than the protocol version in use. Also change
4716 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4717 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4718 the client will at least see that alert.
4721 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4725 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4726 client receives HelloRequest while in a handshake.
4727 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4729 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4730 should end in 'break', not 'goto end' which circuments various
4731 cleanups done in state SSL_ST_OK. But session related stuff
4732 must be disabled for SSL_ST_OK in the case that we just sent a
4735 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4736 before just sending a HelloRequest.
4737 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4739 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4740 reveal whether illegal block cipher padding was found or a MAC
4741 verification error occured. (Neither SSLerr() codes nor alerts
4742 are directly visible to potential attackers, but the information
4743 may leak via logfiles.)
4745 Similar changes are not required for the SSL 2.0 implementation
4746 because the number of padding bytes is sent in clear for SSL 2.0,
4747 and the extra bytes are just ignored. However ssl/s2_pkt.c
4748 failed to verify that the purported number of padding bytes is in
4752 *) Add OpenUNIX-8 support including shared libraries
4753 (Boyd Lynn Gerber <gerberb@zenez.com>).
4756 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4757 'wristwatch attack' using huge encoding parameters (cf.
4758 James H. Manger's CRYPTO 2001 paper). Note that the
4759 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4760 encoding parameters and hence was not vulnerable.
4763 *) BN_sqr() bug fix.
4764 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4766 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4767 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4768 followed by modular reduction.
4769 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4771 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4772 equivalent based on BN_pseudo_rand() instead of BN_rand().
4775 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4776 This function was broken, as the check for a new client hello message
4777 to handle SGC did not allow these large messages.
4778 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4781 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4784 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4785 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4788 *) Rework the configuration and shared library support for Tru64 Unix.
4789 The configuration part makes use of modern compiler features and
4790 still retains old compiler behavior for those that run older versions
4791 of the OS. The shared library support part includes a variant that
4792 uses the RPATH feature, and is available through the special
4793 configuration target "alpha-cc-rpath", which will never be selected
4795 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4797 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4798 with the same message size as in ssl3_get_certificate_request().
4799 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4800 messages might inadvertently be reject as too long.
4801 [Petr Lampa <lampa@fee.vutbr.cz>]
4803 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4806 *) Modified SSL library such that the verify_callback that has been set
4807 specificly for an SSL object with SSL_set_verify() is actually being
4808 used. Before the change, a verify_callback set with this function was
4809 ignored and the verify_callback() set in the SSL_CTX at the time of
4810 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4811 to allow the necessary settings.
4814 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4815 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4816 done automatically (in contradiction to the requirements of the C
4817 standard). This made problems when used from OpenSSH.
4820 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4821 dh->length and always used
4823 BN_rand_range(priv_key, dh->p).
4825 BN_rand_range() is not necessary for Diffie-Hellman, and this
4826 specific range makes Diffie-Hellman unnecessarily inefficient if
4827 dh->length (recommended exponent length) is much smaller than the
4828 length of dh->p. We could use BN_rand_range() if the order of
4829 the subgroup was stored in the DH structure, but we only have
4834 BN_rand(priv_key, l, ...)
4836 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4842 RSA_eay_public_encrypt
4843 RSA_eay_private_decrypt
4844 RSA_eay_private_encrypt (signing)
4845 RSA_eay_public_decrypt (signature verification)
4847 (default implementations for RSA_public_encrypt,
4848 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4849 always reject numbers >= n.
4852 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4853 to synchronize access to 'locking_thread'. This is necessary on
4854 systems where access to 'locking_thread' (an 'unsigned long'
4855 variable) is not atomic.
4858 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4859 *before* setting the 'crypto_lock_rand' flag. The previous code had
4860 a race condition if 0 is a valid thread ID.
4861 [Travis Vitek <vitek@roguewave.com>]
4863 *) Add support for shared libraries under Irix.
4864 [Albert Chin-A-Young <china@thewrittenword.com>]
4866 *) Add configuration option to build on Linux on both big-endian and
4868 [Ralf Baechle <ralf@uni-koblenz.de>]
4870 *) Add the possibility to create shared libraries on HP-UX.
4873 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4875 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4876 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4877 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4878 PRNG state recovery was possible based on the output of
4879 one PRNG request appropriately sized to gain knowledge on
4880 'md' followed by enough consecutive 1-byte PRNG requests
4881 to traverse all of 'state'.
4883 1. When updating 'md_local' (the current thread's copy of 'md')
4884 during PRNG output generation, hash all of the previous
4885 'md_local' value, not just the half used for PRNG output.
4887 2. Make the number of bytes from 'state' included into the hash
4888 independent from the number of PRNG bytes requested.
4890 The first measure alone would be sufficient to avoid
4891 Markku-Juhani's attack. (Actually it had never occurred
4892 to me that the half of 'md_local' used for chaining was the
4893 half from which PRNG output bytes were taken -- I had always
4894 assumed that the secret half would be used.) The second
4895 measure makes sure that additional data from 'state' is never
4896 mixed into 'md_local' in small portions; this heuristically
4897 further strengthens the PRNG.
4900 *) Fix crypto/bn/asm/mips3.s.
4903 *) When only the key is given to "enc", the IV is undefined. Print out
4904 an error message in this case.
4907 *) Handle special case when X509_NAME is empty in X509 printing routines.
4910 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4911 positive and less than q.
4914 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4915 used: it isn't thread safe and the add_lock_callback should handle
4917 [Paul Rose <Paul.Rose@bridge.com>]
4919 *) Verify that incoming data obeys the block size in
4920 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4924 [Ulf Möller, Bodo Möller]
4926 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4927 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4928 when fixing the server behaviour for backwards-compatible 'client
4929 hello' messages. (Note that the attack is impractical against
4930 SSL 3.0 and TLS 1.0 anyway because length and version checking
4931 means that the probability of guessing a valid ciphertext is
4932 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4935 Before 0.9.5, the countermeasure (hide the error by generating a
4936 random 'decryption result') did not work properly because
4937 ERR_clear_error() was missing, meaning that SSL_get_error() would
4938 detect the supposedly ignored error.
4940 Both problems are now fixed.
4943 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4944 (previously it was 1024).
4947 *) Fix for compatibility mode trust settings: ignore trust settings
4948 unless some valid trust or reject settings are present.
4951 *) Fix for blowfish EVP: its a variable length cipher.
4954 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4955 parameters in DSA public key structures and return an error in the
4956 DSA routines if parameters are absent.
4959 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4960 in the current directory if neither $RANDFILE nor $HOME was set.
4961 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4962 caused some confusion to Windows users who haven't defined $HOME.
4963 Thus RAND_file_name() is changed again: e_os.h can define a
4964 DEFAULT_HOME, which will be used if $HOME is not set.
4965 For Windows, we use "C:"; on other platforms, we still require
4966 environment variables.
4968 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4969 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4970 having multiple threads call RAND_poll() concurrently.
4973 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4974 combination of a flag and a thread ID variable.
4975 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4976 flag), *other* threads can enter ssleay_add_bytes without obeying
4977 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4978 that they do not hold after the first thread unsets add_do_not_lock).
4981 *) Change bctest again: '-x' expressions are not available in all
4985 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4987 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4988 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4990 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4991 the default extension for executables, if any. Also, make the perl
4992 scripts that use symlink() to test if it really exists and use "cp"
4993 if it doesn't. All this made OpenSSL compilable and installable in
4997 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4998 If SEQUENCE is length is indefinite just set c->slen to the total
4999 amount of data available.
5000 [Steve Henson, reported by shige@FreeBSD.org]
5001 [This change does not apply to 0.9.7.]
5003 *) Change bctest to avoid here-documents inside command substitution
5004 (workaround for FreeBSD /bin/sh bug).
5005 For compatibility with Ultrix, avoid shell functions (introduced
5006 in the bctest version that searches along $PATH).
5009 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5010 with des_encrypt() defined on some operating systems, like Solaris
5014 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5015 On the Importance of Eliminating Errors in Cryptographic
5016 Computations, J. Cryptology 14 (2001) 2, 101-119,
5017 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5020 *) MIPS assembler BIGNUM division bug fix.
5023 *) Disabled incorrect Alpha assembler code.
5026 *) Fix PKCS#7 decode routines so they correctly update the length
5027 after reading an EOC for the EXPLICIT tag.
5029 [This change does not apply to 0.9.7.]
5031 *) Fix bug in PKCS#12 key generation routines. This was triggered
5032 if a 3DES key was generated with a 0 initial byte. Include
5033 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5034 (but broken) behaviour.
5037 *) Enhance bctest to search for a working bc along $PATH and print
5039 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5041 *) Fix memory leaks in err.c: free err_data string if necessary;
5042 don't write to the wrong index in ERR_set_error_data.
5045 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5049 *) Replace rdtsc with _emit statements for VC++ version 5.
5050 [Jeremy Cooper <jeremy@baymoo.org>]
5052 *) Make it possible to reuse SSLv2 sessions.
5055 *) In copy_email() check for >= 0 as a return value for
5056 X509_NAME_get_index_by_NID() since 0 is a valid index.
5057 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5059 *) Avoid coredump with unsupported or invalid public keys by checking if
5060 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5061 PKCS7_verify() fails with non detached data.
5064 *) Don't use getenv in library functions when run as setuid/setgid.
5065 New function OPENSSL_issetugid().
5068 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5069 due to incorrect handling of multi-threading:
5071 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5073 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5075 3. Count how many times MemCheck_off() has been called so that
5076 nested use can be treated correctly. This also avoids
5077 inband-signalling in the previous code (which relied on the
5078 assumption that thread ID 0 is impossible).
5081 *) Add "-rand" option also to s_client and s_server.
5084 *) Fix CPU detection on Irix 6.x.
5085 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5086 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5088 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5091 [This change does not apply to 0.9.7.]
5093 *) Use the cached encoding of an X509_NAME structure rather than
5094 copying it. This is apparently the reason for the libsafe "errors"
5095 but the code is actually correct.
5098 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5099 Bleichenbacher's DSA attack.
5100 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5101 to be set and top=0 forces the highest bit to be set; top=-1 is new
5102 and leaves the highest bit random.
5103 [Ulf Moeller, Bodo Moeller]
5105 *) In the NCONF_...-based implementations for CONF_... queries
5106 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5107 a temporary CONF structure with the data component set to NULL
5108 (which gives segmentation faults in lh_retrieve).
5109 Instead, use NULL for the CONF pointer in CONF_get_string and
5110 CONF_get_number (which may use environment variables) and directly
5111 return NULL from CONF_get_section.
5114 *) Fix potential buffer overrun for EBCDIC.
5117 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5118 keyUsage if basicConstraints absent for a CA.
5121 *) Make SMIME_write_PKCS7() write mail header values with a format that
5122 is more generally accepted (no spaces before the semicolon), since
5123 some programs can't parse those values properly otherwise. Also make
5124 sure BIO's that break lines after each write do not create invalid
5128 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5129 macros previously used would not encode an empty SEQUENCE OF
5130 and break the signature.
5132 [This change does not apply to 0.9.7.]
5134 *) Zero the premaster secret after deriving the master secret in
5138 *) Add some EVP_add_digest_alias registrations (as found in
5139 OpenSSL_add_all_digests()) to SSL_library_init()
5140 aka OpenSSL_add_ssl_algorithms(). This provides improved
5141 compatibility with peers using X.509 certificates
5142 with unconventional AlgorithmIdentifier OIDs.
5145 *) Fix for Irix with NO_ASM.
5146 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5148 *) ./config script fixes.
5149 [Ulf Moeller, Richard Levitte]
5151 *) Fix 'openssl passwd -1'.
5154 *) Change PKCS12_key_gen_asc() so it can cope with non null
5155 terminated strings whose length is passed in the passlen
5156 parameter, for example from PEM callbacks. This was done
5157 by adding an extra length parameter to asc2uni().
5158 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5160 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5161 call failed, free the DSA structure.
5164 *) Fix to uni2asc() to cope with zero length Unicode strings.
5165 These are present in some PKCS#12 files.
5168 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5169 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5170 when writing a 32767 byte record.
5171 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5173 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5174 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5176 (RSA objects have a reference count access to which is protected
5177 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5178 so they are meant to be shared between threads.)
5179 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5180 "Reddie, Steven" <Steven.Reddie@ca.com>]
5182 *) Fix a deadlock in CRYPTO_mem_leaks().
5185 *) Use better test patterns in bntest.
5188 *) rand_win.c fix for Borland C.
5191 *) BN_rshift bugfix for n == 0.
5194 *) Add a 'bctest' script that checks for some known 'bc' bugs
5195 so that 'make test' does not abort just because 'bc' is broken.
5198 *) Store verify_result within SSL_SESSION also for client side to
5199 avoid potential security hole. (Re-used sessions on the client side
5200 always resulted in verify_result==X509_V_OK, not using the original
5201 result of the server certificate verification.)
5204 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5205 SSL3_RT_APPLICATION_DATA, return 0.
5206 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5210 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5211 releases, have been re-implemented by renaming the previous
5212 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5213 and ssl3_read_internal, respectively, and adding 'peek' parameters
5214 to them. The new ssl[23]_{read,peek} functions are calls to
5215 ssl[23]_read_internal with the 'peek' flag set appropriately.
5216 A 'peek' parameter has also been added to ssl3_read_bytes, which
5217 does the actual work for ssl3_read_internal.
5220 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5221 the method-specific "init()" handler. Also clean up ex_data after
5222 calling the method-specific "finish()" handler. Previously, this was
5223 happening the other way round.
5226 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5227 The previous value, 12, was not always sufficient for BN_mod_exp().
5230 *) Make sure that shared libraries get the internal name engine with
5231 the full version number and not just 0. This should mark the
5232 shared libraries as not backward compatible. Of course, this should
5233 be changed again when we can guarantee backward binary compatibility.
5236 *) Fix typo in get_cert_by_subject() in by_dir.c
5237 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5239 *) Rework the system to generate shared libraries:
5241 - Make note of the expected extension for the shared libraries and
5242 if there is a need for symbolic links from for example libcrypto.so.0
5243 to libcrypto.so.0.9.7. There is extended info in Configure for
5246 - Make as few rebuilds of the shared libraries as possible.
5248 - Still avoid linking the OpenSSL programs with the shared libraries.
5250 - When installing, install the shared libraries separately from the
5254 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5256 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5257 and not in SSL_clear because the latter is also used by the
5258 accept/connect functions; previously, the settings made by
5259 SSL_set_read_ahead would be lost during the handshake.
5260 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5262 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5263 Previously, it would create entries for disableed algorithms no
5267 *) Added several new manual pages for SSL_* function.
5270 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5272 *) In ssl23_get_client_hello, generate an error message when faced
5273 with an initial SSL 3.0/TLS record that is too small to contain the
5274 first two bytes of the ClientHello message, i.e. client_version.
5275 (Note that this is a pathologic case that probably has never happened
5276 in real life.) The previous approach was to use the version number
5277 from the record header as a substitute; but our protocol choice
5278 should not depend on that one because it is not authenticated
5279 by the Finished messages.
5282 *) More robust randomness gathering functions for Windows.
5283 [Jeffrey Altman <jaltman@columbia.edu>]
5285 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5286 not set then we don't setup the error code for issuer check errors
5287 to avoid possibly overwriting other errors which the callback does
5288 handle. If an application does set the flag then we assume it knows
5289 what it is doing and can handle the new informational codes
5293 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5294 a general "ANY" type, as such it should be able to decode anything
5295 including tagged types. However it didn't check the class so it would
5296 wrongly interpret tagged types in the same way as their universal
5297 counterpart and unknown types were just rejected. Changed so that the
5298 tagged and unknown types are handled in the same way as a SEQUENCE:
5299 that is the encoding is stored intact. There is also a new type
5300 "V_ASN1_OTHER" which is used when the class is not universal, in this
5301 case we have no idea what the actual type is so we just lump them all
5305 *) On VMS, stdout may very well lead to a file that is written to
5306 in a record-oriented fashion. That means that every write() will
5307 write a separate record, which will be read separately by the
5308 programs trying to read from it. This can be very confusing.
5310 The solution is to put a BIO filter in the way that will buffer
5311 text until a linefeed is reached, and then write everything a
5312 line at a time, so every record written will be an actual line,
5313 not chunks of lines and not (usually doesn't happen, but I've
5314 seen it once) several lines in one record. BIO_f_linebuffer() is
5317 Currently, it's a VMS-only method, because that's where it has
5318 been tested well enough.
5321 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5322 it can return incorrect results.
5323 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5324 but it was in 0.9.6-beta[12].)
5327 *) Disable the check for content being present when verifying detached
5328 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5329 include zero length content when signing messages.
5332 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5333 BIO_ctrl (for BIO pairs).
5336 *) Add DSO method for VMS.
5339 *) Bug fix: Montgomery multiplication could produce results with the
5343 *) Add RPM specification openssl.spec and modify it to build three
5344 packages. The default package contains applications, application
5345 documentation and run-time libraries. The devel package contains
5346 include files, static libraries and function documentation. The
5347 doc package contains the contents of the doc directory. The original
5348 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5351 *) Add a large number of documentation files for many SSL routines.
5352 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5354 *) Add a configuration entry for Sony News 4.
5355 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5357 *) Don't set the two most significant bits to one when generating a
5358 random number < q in the DSA library.
5361 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5362 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5363 the underlying transport is blocking) if a handshake took place.
5364 (The default behaviour is needed by applications such as s_client
5365 and s_server that use select() to determine when to use SSL_read;
5366 but for applications that know in advance when to expect data, it
5367 just makes things more complicated.)
5370 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5374 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5375 work better on such systems.
5376 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5378 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5379 Update PKCS12_parse() so it copies the friendlyName and the
5380 keyid to the certificates aux info.
5383 *) Fix bug in PKCS7_verify() which caused an infinite loop
5384 if there was more than one signature.
5385 [Sven Uszpelkat <su@celocom.de>]
5387 *) Major change in util/mkdef.pl to include extra information
5388 about each symbol, as well as presentig variables as well
5389 as functions. This change means that there's n more need
5390 to rebuild the .num files when some algorithms are excluded.
5393 *) Allow the verify time to be set by an application,
5394 rather than always using the current time.
5397 *) Phase 2 verify code reorganisation. The certificate
5398 verify code now looks up an issuer certificate by a
5399 number of criteria: subject name, authority key id
5400 and key usage. It also verifies self signed certificates
5401 by the same criteria. The main comparison function is
5402 X509_check_issued() which performs these checks.
5404 Lot of changes were necessary in order to support this
5405 without completely rewriting the lookup code.
5407 Authority and subject key identifier are now cached.
5409 The LHASH 'certs' is X509_STORE has now been replaced
5410 by a STACK_OF(X509_OBJECT). This is mainly because an
5411 LHASH can't store or retrieve multiple objects with
5412 the same hash value.
5414 As a result various functions (which were all internal
5415 use only) have changed to handle the new X509_STORE
5416 structure. This will break anything that messed round
5417 with X509_STORE internally.
5419 The functions X509_STORE_add_cert() now checks for an
5420 exact match, rather than just subject name.
5422 The X509_STORE API doesn't directly support the retrieval
5423 of multiple certificates matching a given criteria, however
5424 this can be worked round by performing a lookup first
5425 (which will fill the cache with candidate certificates)
5426 and then examining the cache for matches. This is probably
5427 the best we can do without throwing out X509_LOOKUP
5428 entirely (maybe later...).
5430 The X509_VERIFY_CTX structure has been enhanced considerably.
5432 All certificate lookup operations now go via a get_issuer()
5433 callback. Although this currently uses an X509_STORE it
5434 can be replaced by custom lookups. This is a simple way
5435 to bypass the X509_STORE hackery necessary to make this
5436 work and makes it possible to use more efficient techniques
5437 in future. A very simple version which uses a simple
5438 STACK for its trusted certificate store is also provided
5439 using X509_STORE_CTX_trusted_stack().
5441 The verify_cb() and verify() callbacks now have equivalents
5442 in the X509_STORE_CTX structure.
5444 X509_STORE_CTX also has a 'flags' field which can be used
5445 to customise the verify behaviour.
5448 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5449 excludes S/MIME capabilities.
5452 *) When a certificate request is read in keep a copy of the
5453 original encoding of the signed data and use it when outputing
5454 again. Signatures then use the original encoding rather than
5455 a decoded, encoded version which may cause problems if the
5456 request is improperly encoded.
5459 *) For consistency with other BIO_puts implementations, call
5460 buffer_write(b, ...) directly in buffer_puts instead of calling
5463 In BIO_puts, increment b->num_write as in BIO_write.
5464 [Peter.Sylvester@EdelWeb.fr]
5466 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5467 BN_zero, we may not return a BIGNUM with an array consisting of
5471 *) Avoid calling abort() from within the library when problems are
5472 detected, except if preprocessor symbols have been defined
5473 (such as REF_CHECK, BN_DEBUG etc.).
5476 *) New openssl application 'rsautl'. This utility can be
5477 used for low level RSA operations. DER public key
5478 BIO/fp routines also added.
5481 *) New Configure entry and patches for compiling on QNX 4.
5482 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5484 *) A demo state-machine implementation was sponsored by
5485 Nuron (http://www.nuron.com/) and is now available in
5486 demos/state_machine.
5489 *) New options added to the 'dgst' utility for signature
5490 generation and verification.
5493 *) Unrecognized PKCS#7 content types are now handled via a
5494 catch all ASN1_TYPE structure. This allows unsupported
5495 types to be stored as a "blob" and an application can
5496 encode and decode it manually.
5499 *) Fix various signed/unsigned issues to make a_strex.c
5501 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5503 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5504 length if passed a buffer. ASN1_INTEGER_to_BN failed
5505 if passed a NULL BN and its argument was negative.
5506 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5508 *) Modification to PKCS#7 encoding routines to output definite
5509 length encoding. Since currently the whole structures are in
5510 memory there's not real point in using indefinite length
5511 constructed encoding. However if OpenSSL is compiled with
5512 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5515 *) Added BIO_vprintf() and BIO_vsnprintf().
5518 *) Added more prefixes to parse for in the the strings written
5519 through a logging bio, to cover all the levels that are available
5520 through syslog. The prefixes are now:
5522 PANIC, EMERG, EMR => LOG_EMERG
5523 ALERT, ALR => LOG_ALERT
5524 CRIT, CRI => LOG_CRIT
5525 ERROR, ERR => LOG_ERR
5526 WARNING, WARN, WAR => LOG_WARNING
5527 NOTICE, NOTE, NOT => LOG_NOTICE
5528 INFO, INF => LOG_INFO
5529 DEBUG, DBG => LOG_DEBUG
5531 and as before, if none of those prefixes are present at the
5532 beginning of the string, LOG_ERR is chosen.
5534 On Win32, the LOG_* levels are mapped according to this:
5536 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5537 LOG_WARNING => EVENTLOG_WARNING_TYPE
5538 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5542 *) Made it possible to reconfigure with just the configuration
5543 argument "reconf" or "reconfigure". The command line arguments
5544 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5545 and are retrieved from there when reconfiguring.
5549 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5551 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5554 *) The obj_dat.pl script was messing up the sorting of object
5555 names. The reason was that it compared the quoted version
5556 of strings as a result "OCSP" > "OCSP Signing" because
5557 " > SPACE. Changed script to store unquoted versions of
5558 names and add quotes on output. It was also omitting some
5559 names from the lookup table if they were given a default
5560 value (that is if SN is missing it is given the same
5561 value as LN and vice versa), these are now added on the
5562 grounds that if an object has a name we should be able to
5563 look it up. Finally added warning output when duplicate
5564 short or long names are found.
5567 *) Changes needed for Tandem NSK.
5568 [Scott Uroff <scott@xypro.com>]
5570 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5571 RSA_padding_check_SSLv23(), special padding was never detected
5572 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5573 version rollback attacks was not effective.
5575 In s23_clnt.c, don't use special rollback-attack detection padding
5576 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5577 client; similarly, in s23_srvr.c, don't do the rollback check if
5578 SSL 2.0 is the only protocol enabled in the server.
5581 *) Make it possible to get hexdumps of unprintable data with 'openssl
5582 asn1parse'. By implication, the functions ASN1_parse_dump() and
5583 BIO_dump_indent() are added.
5586 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5587 these print out strings and name structures based on various
5588 flags including RFC2253 support and proper handling of
5589 multibyte characters. Added options to the 'x509' utility
5590 to allow the various flags to be set.
5593 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5594 Also change the functions X509_cmp_current_time() and
5595 X509_gmtime_adj() work with an ASN1_TIME structure,
5596 this will enable certificates using GeneralizedTime in validity
5597 dates to be checked.
5600 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5601 negative public key encodings) on by default,
5602 NO_NEG_PUBKEY_BUG can be set to disable it.
5605 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5606 content octets. An i2c_ASN1_OBJECT is unnecessary because
5607 the encoding can be trivially obtained from the structure.
5610 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5611 not read locks (CRYPTO_r_[un]lock).
5614 *) A first attempt at creating official support for shared
5615 libraries through configuration. I've kept it so the
5616 default is static libraries only, and the OpenSSL programs
5617 are always statically linked for now, but there are
5618 preparations for dynamic linking in place.
5619 This has been tested on Linux and Tru64.
5622 *) Randomness polling function for Win9x, as described in:
5623 Peter Gutmann, Software Generation of Practically Strong
5627 *) Fix so PRNG is seeded in req if using an already existing
5631 *) New options to smime application. -inform and -outform
5632 allow alternative formats for the S/MIME message including
5633 PEM and DER. The -content option allows the content to be
5634 specified separately. This should allow things like Netscape
5635 form signing output easier to verify.
5638 *) Fix the ASN1 encoding of tags using the 'long form'.
5641 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5642 STRING types. These convert content octets to and from the
5643 underlying type. The actual tag and length octets are
5644 already assumed to have been read in and checked. These
5645 are needed because all other string types have virtually
5646 identical handling apart from the tag. By having versions
5647 of the ASN1 functions that just operate on content octets
5648 IMPLICIT tagging can be handled properly. It also allows
5649 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5650 and ASN1_INTEGER are identical apart from the tag.
5653 *) Change the handling of OID objects as follows:
5655 - New object identifiers are inserted in objects.txt, following
5656 the syntax given in objects.README.
5657 - objects.pl is used to process obj_mac.num and create a new
5659 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5662 This is currently kind of a hack, and the perl code in objects.pl
5663 isn't very elegant, but it works as I intended. The simplest way
5664 to check that it worked correctly is to look in obj_dat.h and
5665 check the array nid_objs and make sure the objects haven't moved
5666 around (this is important!). Additions are OK, as well as
5667 consistent name changes.
5670 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5673 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5674 The given file adds to whatever has already been seeded into the
5675 random pool through the RANDFILE configuration file option or
5676 environment variable, or the default random state file.
5679 *) mkstack.pl now sorts each macro group into lexical order.
5680 Previously the output order depended on the order the files
5681 appeared in the directory, resulting in needless rewriting
5685 *) Patches to make OpenSSL compile under Win32 again. Mostly
5686 work arounds for the VC++ problem that it treats func() as
5687 func(void). Also stripped out the parts of mkdef.pl that
5688 added extra typesafe functions: these no longer exist.
5691 *) Reorganisation of the stack code. The macros are now all
5692 collected in safestack.h . Each macro is defined in terms of
5693 a "stack macro" of the form SKM_<name>(type, a, b). The
5694 DEBUG_SAFESTACK is now handled in terms of function casts,
5695 this has the advantage of retaining type safety without the
5696 use of additional functions. If DEBUG_SAFESTACK is not defined
5697 then the non typesafe macros are used instead. Also modified the
5698 mkstack.pl script to handle the new form. Needs testing to see
5699 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5700 the default if no major problems. Similar behaviour for ASN1_SET_OF
5701 and PKCS12_STACK_OF.
5704 *) When some versions of IIS use the 'NET' form of private key the
5705 key derivation algorithm is different. Normally MD5(password) is
5706 used as a 128 bit RC4 key. In the modified case
5707 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5708 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5709 as the old Netscape_RSA functions except they have an additional
5710 'sgckey' parameter which uses the modified algorithm. Also added
5711 an -sgckey command line option to the rsa utility. Thanks to
5712 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5713 algorithm to openssl-dev.
5716 *) The evp_local.h macros were using 'c.##kname' which resulted in
5717 invalid expansion on some systems (SCO 5.0.5 for example).
5718 Corrected to 'c.kname'.
5719 [Phillip Porch <root@theporch.com>]
5721 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5722 a STACK of email addresses from a certificate or request, these look
5723 in the subject name and the subject alternative name extensions and
5724 omit any duplicate addresses.
5727 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5728 This makes DSA verification about 2 % faster.
5731 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5732 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5733 plus overhead for 1024 bit moduli).
5734 This makes exponentiations about 0.5 % faster for 1024 bit
5735 exponents (as measured by "openssl speed rsa2048").
5738 *) Rename memory handling macros to avoid conflicts with other
5740 Malloc => OPENSSL_malloc
5741 Malloc_locked => OPENSSL_malloc_locked
5742 Realloc => OPENSSL_realloc
5743 Free => OPENSSL_free
5746 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5747 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5750 *) CygWin32 support.
5751 [John Jarvie <jjarvie@newsguy.com>]
5753 *) The type-safe stack code has been rejigged. It is now only compiled
5754 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5755 by default all type-specific stack functions are "#define"d back to
5756 standard stack functions. This results in more streamlined output
5757 but retains the type-safety checking possibilities of the original
5761 *) The STACK code has been cleaned up, and certain type declarations
5762 that didn't make a lot of sense have been brought in line. This has
5763 also involved a cleanup of sorts in safestack.h to more correctly
5764 map type-safe stack functions onto their plain stack counterparts.
5765 This work has also resulted in a variety of "const"ifications of
5766 lots of the code, especially "_cmp" operations which should normally
5767 be prototyped with "const" parameters anyway.
5770 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5771 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5772 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5773 where all of 'md' is used each time the PRNG is used, but 'state'
5774 is used only indexed by a cyclic counter. As entropy may not be
5775 well distributed from the beginning, 'md' is important as a
5776 chaining variable. However, the output function chains only half
5777 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5778 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5779 in all of 'state' being rewritten, with the new values depending
5780 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5783 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5784 the handshake is continued after ssl_verify_cert_chain();
5785 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5786 can lead to 'unexplainable' connection aborts later.
5787 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5789 *) Major EVP API cipher revision.
5790 Add hooks for extra EVP features. This allows various cipher
5791 parameters to be set in the EVP interface. Support added for variable
5792 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5793 setting of RC2 and RC5 parameters.
5795 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5798 Remove lots of duplicated code from the EVP library. For example *every*
5799 cipher init() function handles the 'iv' in the same way according to the
5800 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5801 for CFB and OFB modes they zero ctx->num.
5803 New functionality allows removal of S/MIME code RC2 hack.
5805 Most of the routines have the same form and so can be declared in terms
5808 By shifting this to the top level EVP_CipherInit() it can be removed from
5809 all individual ciphers. If the cipher wants to handle IVs or keys
5810 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5813 Change lots of functions like EVP_EncryptUpdate() to now return a
5814 value: although software versions of the algorithms cannot fail
5815 any installed hardware versions can.
5818 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5819 this option is set, tolerate broken clients that send the negotiated
5820 protocol version number instead of the requested protocol version
5824 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5825 i.e. non-zero for export ciphersuites, zero otherwise.
5826 Previous versions had this flag inverted, inconsistent with
5827 rsa_tmp_cb (..._TMP_RSA_CB).
5828 [Bodo Moeller; problem reported by Amit Chopra]
5830 *) Add missing DSA library text string. Work around for some IIS
5831 key files with invalid SEQUENCE encoding.
5834 *) Add a document (doc/standards.txt) that list all kinds of standards
5835 and so on that are implemented in OpenSSL.
5838 *) Enhance c_rehash script. Old version would mishandle certificates
5839 with the same subject name hash and wouldn't handle CRLs at all.
5840 Added -fingerprint option to crl utility, to support new c_rehash
5844 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5847 *) Fix for SSL server purpose checking. Server checking was
5848 rejecting certificates which had extended key usage present
5849 but no ssl client purpose.
5850 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5852 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5853 is a little unclear about how a blank password is handled.
5854 Since the password in encoded as a BMPString with terminating
5855 double NULL a zero length password would end up as just the
5856 double NULL. However no password at all is different and is
5857 handled differently in the PKCS#12 key generation code. NS
5858 treats a blank password as zero length. MSIE treats it as no
5859 password on export: but it will try both on import. We now do
5860 the same: PKCS12_parse() tries zero length and no password if
5861 the password is set to "" or NULL (NULL is now a valid password:
5862 it wasn't before) as does the pkcs12 application.
5865 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5866 perror when PEM_read_bio_X509_REQ fails, the error message must
5867 be obtained from the error queue.
5870 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5871 it in ERR_remove_state if appropriate, and change ERR_get_state
5872 accordingly to avoid race conditions (this is necessary because
5873 thread_hash is no longer constant once set).
5876 *) Bugfix for linux-elf makefile.one.
5879 *) RSA_get_default_method() will now cause a default
5880 RSA_METHOD to be chosen if one doesn't exist already.
5881 Previously this was only set during a call to RSA_new()
5882 or RSA_new_method(NULL) meaning it was possible for
5883 RSA_get_default_method() to return NULL.
5886 *) Added native name translation to the existing DSO code
5887 that will convert (if the flag to do so is set) filenames
5888 that are sufficiently small and have no path information
5889 into a canonical native form. Eg. "blah" converted to
5890 "libblah.so" or "blah.dll" etc.
5893 *) New function ERR_error_string_n(e, buf, len) which is like
5894 ERR_error_string(e, buf), but writes at most 'len' bytes
5895 including the 0 terminator. For ERR_error_string_n, 'buf'
5897 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5899 *) CONF library reworked to become more general. A new CONF
5900 configuration file reader "class" is implemented as well as a
5901 new functions (NCONF_*, for "New CONF") to handle it. The now
5902 old CONF_* functions are still there, but are reimplemented to
5903 work in terms of the new functions. Also, a set of functions
5904 to handle the internal storage of the configuration data is
5905 provided to make it easier to write new configuration file
5906 reader "classes" (I can definitely see something reading a
5907 configuration file in XML format, for example), called _CONF_*,
5908 or "the configuration storage API"...
5910 The new configuration file reading functions are:
5912 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5913 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5915 NCONF_default, NCONF_WIN32
5917 NCONF_dump_fp, NCONF_dump_bio
5919 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5920 NCONF_new creates a new CONF object. This works in the same way
5921 as other interfaces in OpenSSL, like the BIO interface.
5922 NCONF_dump_* dump the internal storage of the configuration file,
5923 which is useful for debugging. All other functions take the same
5924 arguments as the old CONF_* functions wth the exception of the
5925 first that must be a `CONF *' instead of a `LHASH *'.
5927 To make it easer to use the new classes with the old CONF_* functions,
5928 the function CONF_set_default_method is provided.
5931 *) Add '-tls1' option to 'openssl ciphers', which was already
5932 mentioned in the documentation but had not been implemented.
5933 (This option is not yet really useful because even the additional
5934 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5937 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5938 OpenSSL-based applications) load shared libraries and bind to
5939 them in a portable way.
5940 [Geoff Thorpe, with contributions from Richard Levitte]
5942 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5944 *) Make sure _lrotl and _lrotr are only used with MSVC.
5946 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5947 (the default implementation of RAND_status).
5949 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5950 to '-clrext' (= clear extensions), as intended and documented.
5951 [Bodo Moeller; inconsistency pointed out by Michael Attili
5954 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5955 was larger than the MD block size.
5956 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5958 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5959 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5960 using the passed key: if the passed key was a private key the result
5961 of X509_print(), for example, would be to print out all the private key
5965 *) des_quad_cksum() byte order bug fix.
5966 [Ulf Möller, using the problem description in krb4-0.9.7, where
5967 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5969 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5971 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5973 *) For easily testing in shell scripts whether some command
5974 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5975 returns with exit code 0 iff no command of the given name is available.
5976 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5977 the output goes to stdout and nothing is printed to stderr.
5978 Additional arguments are always ignored.
5980 Since for each cipher there is a command of the same name,
5981 the 'no-cipher' compilation switches can be tested this way.
5983 ('openssl no-XXX' is not able to detect pseudo-commands such
5984 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5987 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5990 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5991 is set; it will be thrown away anyway because each handshake creates
5993 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5994 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5995 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5996 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5999 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6000 'Q' and 'R' lose their special meanings (quit/renegotiate).
6001 This is part of what -quiet does; unlike -quiet, -ign_eof
6002 does not suppress any output.
6005 *) Add compatibility options to the purpose and trust code. The
6006 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6007 accepts a certificate or CA, this was the previous behaviour,
6008 with all the associated security issues.
6010 X509_TRUST_COMPAT is the old trust behaviour: only and
6011 automatically trust self signed roots in certificate store. A
6012 new trust setting X509_TRUST_DEFAULT is used to specify that
6013 a purpose has no associated trust setting and it should instead
6014 use the value in the default purpose.
6017 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6018 and fix a memory leak.
6021 *) In util/mkerr.pl (which implements 'make errors'), preserve
6022 reason strings from the previous version of the .c file, as
6023 the default to have only downcase letters (and digits) in
6024 automatically generated reasons codes is not always appropriate.
6027 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6028 using strerror. Previously, ERR_reason_error_string() returned
6029 library names as reason strings for SYSerr; but SYSerr is a special
6030 case where small numbers are errno values, not library numbers.
6033 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6034 converts DSA parameters into DH parameters. (When creating parameters,
6035 DSA_generate_parameters is used.)
6038 *) Include 'length' (recommended exponent length) in C code generated
6039 by 'openssl dhparam -C'.
6042 *) The second argument to set_label in perlasm was already being used
6043 so couldn't be used as a "file scope" flag. Moved to third argument
6047 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6048 instead of RAND_bytes for encryption IVs and salts.
6051 *) Include RAND_status() into RAND_METHOD instead of implementing
6052 it only for md_rand.c Otherwise replacing the PRNG by calling
6053 RAND_set_rand_method would be impossible.
6056 *) Don't let DSA_generate_key() enter an infinite loop if the random
6057 number generation fails.
6060 *) New 'rand' application for creating pseudo-random output.
6063 *) Added configuration support for Linux/IA64
6064 [Rolf Haberrecker <rolf@suse.de>]
6066 *) Assembler module support for Mingw32.
6069 *) Shared library support for HPUX (in shlib/).
6070 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6072 *) Shared library support for Solaris gcc.
6073 [Lutz Behnke <behnke@trustcenter.de>]
6075 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6077 *) PKCS7_encrypt() was adding text MIME headers twice because they
6078 were added manually and by SMIME_crlf_copy().
6081 *) In bntest.c don't call BN_rand with zero bits argument.
6082 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6084 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6085 case was implemented. This caused BN_div_recp() to fail occasionally.
6088 *) Add an optional second argument to the set_label() in the perl
6089 assembly language builder. If this argument exists and is set
6090 to 1 it signals that the assembler should use a symbol whose
6091 scope is the entire file, not just the current function. This
6092 is needed with MASM which uses the format label:: for this scope.
6093 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6095 *) Change the ASN1 types so they are typedefs by default. Before
6096 almost all types were #define'd to ASN1_STRING which was causing
6097 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6101 *) Change names of new functions to the new get1/get0 naming
6102 convention: After 'get1', the caller owns a reference count
6103 and has to call ..._free; 'get0' returns a pointer to some
6104 data structure without incrementing reference counters.
6105 (Some of the existing 'get' functions increment a reference
6106 counter, some don't.)
6107 Similarly, 'set1' and 'add1' functions increase reference
6108 counters or duplicate objects.
6111 *) Allow for the possibility of temp RSA key generation failure:
6112 the code used to assume it always worked and crashed on failure.
6115 *) Fix potential buffer overrun problem in BIO_printf().
6116 [Ulf Möller, using public domain code by Patrick Powell; problem
6117 pointed out by David Sacerdote <das33@cornell.edu>]
6119 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6120 RAND_egd() and RAND_status(). In the command line application,
6121 the EGD socket can be specified like a seed file using RANDFILE
6125 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6126 Some CAs (e.g. Verisign) distribute certificates in this form.
6129 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6130 list to exclude them. This means that no special compilation option
6131 is needed to use anonymous DH: it just needs to be included in the
6135 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6136 EVP_MD_type. The old functionality is available in a new macro called
6137 EVP_MD_md(). Change code that uses it and update docs.
6140 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6141 where the 'void *' argument is replaced by a function pointer argument.
6142 Previously 'void *' was abused to point to functions, which works on
6143 many platforms, but is not correct. As these functions are usually
6144 called by macros defined in OpenSSL header files, most source code
6145 should work without changes.
6148 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6149 sections with information on -D... compiler switches used for
6150 compiling the library so that applications can see them. To enable
6151 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6152 must be defined. E.g.,
6153 #define OPENSSL_ALGORITHM_DEFINES
6154 #include <openssl/opensslconf.h>
6155 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6156 [Richard Levitte, Ulf and Bodo Möller]
6158 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6162 *) Change the 'other' type in certificate aux info to a STACK_OF
6163 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6164 the required ASN1 format: arbitrary types determined by an OID.
6167 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6168 argument to 'req'. This is not because the function is newer or
6169 better than others it just uses the work 'NEW' in the certificate
6170 request header lines. Some software needs this.
6173 *) Reorganise password command line arguments: now passwords can be
6174 obtained from various sources. Delete the PEM_cb function and make
6175 it the default behaviour: i.e. if the callback is NULL and the
6176 usrdata argument is not NULL interpret it as a null terminated pass
6177 phrase. If usrdata and the callback are NULL then the pass phrase
6178 is prompted for as usual.
6181 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6182 the support is automatically enabled. The resulting binaries will
6183 autodetect the card and use it if present.
6184 [Ben Laurie and Compaq Inc.]
6186 *) Work around for Netscape hang bug. This sends certificate request
6187 and server done in one record. Since this is perfectly legal in the
6188 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6189 the bugs/SSLv3 entry for more info.
6192 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6195 *) Add -rand argument to smime and pkcs12 applications and read/write
6199 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6202 *) Add command line password options to the remaining applications.
6205 *) Bug fix for BN_div_recp() for numerators with an even number of
6209 *) More tests in bntest.c, and changed test_bn output.
6212 *) ./config recognizes MacOS X now.
6215 *) Bug fix for BN_div() when the first words of num and divsor are
6216 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6219 *) Add support for various broken PKCS#8 formats, and command line
6220 options to produce them.
6223 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6224 get temporary BIGNUMs from a BN_CTX.
6227 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6231 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6232 include a #define from the old name to the new. The original intent
6233 was that statically linked binaries could for example just call
6234 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6235 link with digests. This never worked becayse SSLeay_add_all_digests()
6236 and SSLeay_add_all_ciphers() were in the same source file so calling
6237 one would link with the other. They are now in separate source files.
6240 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6243 *) Use a less unusual form of the Miller-Rabin primality test (it used
6244 a binary algorithm for exponentiation integrated into the Miller-Rabin
6245 loop, our standard modexp algorithms are faster).
6248 *) Support for the EBCDIC character set completed.
6249 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6251 *) Source code cleanups: use const where appropriate, eliminate casts,
6252 use void * instead of char * in lhash.
6255 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6256 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6257 this the server could overwrite ephemeral keys that the client
6261 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6262 using 50 iterations of the Rabin-Miller test.
6264 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6265 iterations of the Rabin-Miller test as required by the appendix
6266 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6267 As BN_is_prime_fasttest includes trial division, DSA parameter
6268 generation becomes much faster.
6270 This implies a change for the callback functions in DSA_is_prime
6271 and DSA_generate_parameters: The callback function is called once
6272 for each positive witness in the Rabin-Miller test, not just
6273 occasionally in the inner loop; and the parameters to the
6274 callback function now provide an iteration count for the outer
6275 loop rather than for the current invocation of the inner loop.
6276 DSA_generate_parameters additionally can call the callback
6277 function with an 'iteration count' of -1, meaning that a
6278 candidate has passed the trial division test (when q is generated
6279 from an application-provided seed, trial division is skipped).
6282 *) New function BN_is_prime_fasttest that optionally does trial
6283 division before starting the Rabin-Miller test and has
6284 an additional BN_CTX * argument (whereas BN_is_prime always
6285 has to allocate at least one BN_CTX).
6286 'callback(1, -1, cb_arg)' is called when a number has passed the
6287 trial division stage.
6290 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6294 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6297 *) New function BN_pseudo_rand().
6300 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6301 bignum version of BN_from_montgomery() with the working code from
6302 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6306 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6307 made it impossible to use the same SSL_SESSION data structure in
6308 SSL2 clients in multiple threads.
6311 *) The return value of RAND_load_file() no longer counts bytes obtained
6312 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6313 to seed the PRNG (previously an explicit byte count was required).
6314 [Ulf Möller, Bodo Möller]
6316 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6317 used (char *) instead of (void *) and had casts all over the place.
6320 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6323 *) Retain source code compatibility for BN_prime_checks macro:
6324 BN_is_prime(..., BN_prime_checks, ...) now uses
6325 BN_prime_checks_for_size to determine the appropriate number of
6326 Rabin-Miller iterations.
6329 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6330 DH_CHECK_P_NOT_SAFE_PRIME.
6331 (Check if this is true? OpenPGP calls them "strong".)
6334 *) Merge the functionality of "dh" and "gendh" programs into a new program
6335 "dhparam". The old programs are retained for now but will handle DH keys
6336 (instead of parameters) in future.
6339 *) Make the ciphers, s_server and s_client programs check the return values
6340 when a new cipher list is set.
6343 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6344 ciphers. Before when the 56bit ciphers were enabled the sorting was
6347 The syntax for the cipher sorting has been extended to support sorting by
6348 cipher-strength (using the strength_bits hard coded in the tables).
6349 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6351 Fix a bug in the cipher-command parser: when supplying a cipher command
6352 string with an "undefined" symbol (neither command nor alphanumeric
6353 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6354 an error is flagged.
6356 Due to the strength-sorting extension, the code of the
6357 ssl_create_cipher_list() function was completely rearranged. I hope that
6358 the readability was also increased :-)
6359 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6361 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6362 for the first serial number and places 2 in the serial number file. This
6363 avoids problems when the root CA is created with serial number zero and
6364 the first user certificate has the same issuer name and serial number
6368 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6369 the new code. Add documentation for this stuff.
6372 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6373 X509_*() to X509at_*() on the grounds that they don't handle X509
6374 structures and behave in an analagous way to the X509v3 functions:
6375 they shouldn't be called directly but wrapper functions should be used
6378 So we also now have some wrapper functions that call the X509at functions
6379 when passed certificate requests. (TO DO: similar things can be done with
6380 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6381 things. Some of these need some d2i or i2d and print functionality
6382 because they handle more complex structures.)
6385 *) Add missing #ifndefs that caused missing symbols when building libssl
6386 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6387 NO_RSA in ssl/s2*.c.
6388 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6390 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6391 has a return value which indicates the quality of the random data
6392 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6393 error queue. New function RAND_pseudo_bytes() generates output that is
6394 guaranteed to be unique but not unpredictable. RAND_add is like
6395 RAND_seed, but takes an extra argument for an entropy estimate
6396 (RAND_seed always assumes full entropy).
6399 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6400 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6401 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6402 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6403 false-positive rate of at most 2^-80 for random input.
6406 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6409 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6410 in the 0.9.5 release), this returns the chain
6411 from an X509_CTX structure with a dup of the stack and all
6412 the X509 reference counts upped: so the stack will exist
6413 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6416 Also make SSL_SESSION_print() print out the verify return
6420 *) Add manpage for the pkcs12 command. Also change the default
6421 behaviour so MAC iteration counts are used unless the new
6422 -nomaciter option is used. This improves file security and
6423 only older versions of MSIE (4.0 for example) need it.
6426 *) Honor the no-xxx Configure options when creating .DEF files.
6429 *) Add PKCS#10 attributes to field table: challengePassword,
6430 unstructuredName and unstructuredAddress. These are taken from
6431 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6432 international characters are used.
6434 More changes to X509_ATTRIBUTE code: allow the setting of types
6435 based on strings. Remove the 'loc' parameter when adding
6436 attributes because these will be a SET OF encoding which is sorted
6440 *) Initial changes to the 'req' utility to allow request generation
6441 automation. This will allow an application to just generate a template
6442 file containing all the field values and have req construct the
6445 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6446 used all over the place including certificate requests and PKCS#7
6447 structures. They are currently handled manually where necessary with
6448 some primitive wrappers for PKCS#7. The new functions behave in a
6449 manner analogous to the X509 extension functions: they allow
6450 attributes to be looked up by NID and added.
6452 Later something similar to the X509V3 code would be desirable to
6453 automatically handle the encoding, decoding and printing of the
6454 more complex types. The string types like challengePassword can
6455 be handled by the string table functions.
6457 Also modified the multi byte string table handling. Now there is
6458 a 'global mask' which masks out certain types. The table itself
6459 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6460 is useful when for example there is only one permissible type
6461 (as in countryName) and using the mask might result in no valid
6465 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6466 SSL_get_peer_finished to allow applications to obtain the latest
6467 Finished messages sent to the peer or expected from the peer,
6468 respectively. (SSL_get_peer_finished is usually the Finished message
6469 actually received from the peer, otherwise the protocol will be aborted.)
6471 As the Finished message are message digests of the complete handshake
6472 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6473 be used for external authentication procedures when the authentication
6474 provided by SSL/TLS is not desired or is not enough.
6477 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6478 the host supports BWX extension and if Compaq C is present on the
6479 $PATH. Just exploiting of the BWX extension results in 20-30%
6480 performance kick for some algorithms, e.g. DES and RC4 to mention
6481 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6485 *) Add support for MS "fast SGC". This is arguably a violation of the
6486 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6487 weak crypto and after checking the certificate is SGC a second one
6488 with strong crypto. MS SGC stops the first handshake after receiving
6489 the server certificate message and sends a second client hello. Since
6490 a server will typically do all the time consuming operations before
6491 expecting any further messages from the client (server key exchange
6492 is the most expensive) there is little difference between the two.
6494 To get OpenSSL to support MS SGC we have to permit a second client
6495 hello message after we have sent server done. In addition we have to
6496 reset the MAC if we do get this second client hello.
6499 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6500 if a DER encoded private key is RSA or DSA traditional format. Changed
6501 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6502 format DER encoded private key. Newer code should use PKCS#8 format which
6503 has the key type encoded in the ASN1 structure. Added DER private key
6504 support to pkcs8 application.
6507 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6508 ciphersuites has been selected (as required by the SSL 3/TLS 1
6509 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6510 is set, we interpret this as a request to violate the specification
6511 (the worst that can happen is a handshake failure, and 'correct'
6512 behaviour would result in a handshake failure anyway).
6515 *) In SSL_CTX_add_session, take into account that there might be multiple
6516 SSL_SESSION structures with the same session ID (e.g. when two threads
6517 concurrently obtain them from an external cache).
6518 The internal cache can handle only one SSL_SESSION with a given ID,
6519 so if there's a conflict, we now throw out the old one to achieve
6523 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6524 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6525 some routines that use cipher OIDs: some ciphers do not have OIDs
6526 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6530 *) Simplify the trust setting structure and code. Now we just have
6531 two sequences of OIDs for trusted and rejected settings. These will
6532 typically have values the same as the extended key usage extension
6533 and any application specific purposes.
6535 The trust checking code now has a default behaviour: it will just
6536 check for an object with the same NID as the passed id. Functions can
6537 be provided to override either the default behaviour or the behaviour
6538 for a given id. SSL client, server and email already have functions
6539 in place for compatibility: they check the NID and also return "trusted"
6540 if the certificate is self signed.
6543 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6544 traditional format into an EVP_PKEY structure.
6547 *) Add a password callback function PEM_cb() which either prompts for
6548 a password if usr_data is NULL or otherwise assumes it is a null
6549 terminated password. Allow passwords to be passed on command line
6550 environment or config files in a few more utilities.
6553 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6554 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6555 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6556 Update documentation.
6559 *) Support for ASN1 "NULL" type. This could be handled before by using
6560 ASN1_TYPE but there wasn't any function that would try to read a NULL
6561 and produce an error if it couldn't. For compatibility we also have
6562 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6563 don't allocate anything because they don't need to.
6566 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6568 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6570 *) Rebuild of the memory allocation routines used by OpenSSL code and
6571 possibly others as well. The purpose is to make an interface that
6572 provide hooks so anyone can build a separate set of allocation and
6573 deallocation routines to be used by OpenSSL, for example memory
6574 pool implementations, or something else, which was previously hard
6575 since Malloc(), Realloc() and Free() were defined as macros having
6576 the values malloc, realloc and free, respectively (except for Win32
6577 compilations). The same is provided for memory debugging code.
6578 OpenSSL already comes with functionality to find memory leaks, but
6579 this gives people a chance to debug other memory problems.
6581 With these changes, a new set of functions and macros have appeared:
6583 CRYPTO_set_mem_debug_functions() [F]
6584 CRYPTO_get_mem_debug_functions() [F]
6585 CRYPTO_dbg_set_options() [F]
6586 CRYPTO_dbg_get_options() [F]
6587 CRYPTO_malloc_debug_init() [M]
6589 The memory debug functions are NULL by default, unless the library
6590 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6591 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6592 gives the standard debugging functions that come with OpenSSL) or
6593 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6594 provided by the library user) must be used. When the standard
6595 debugging functions are used, CRYPTO_dbg_set_options can be used to
6596 request additional information:
6597 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6598 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6600 Also, things like CRYPTO_set_mem_functions will always give the
6601 expected result (the new set of functions is used for allocation
6602 and deallocation) at all times, regardless of platform and compiler
6605 To finish it up, some functions that were never use in any other
6606 way than through macros have a new API and new semantic:
6609 CRYPTO_dbg_realloc()
6612 All macros of value have retained their old syntax.
6613 [Richard Levitte and Bodo Moeller]
6615 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6616 ordering of SMIMECapabilities wasn't in "strength order" and there
6617 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6621 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6622 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6623 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6625 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6626 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6627 functionality to handle multipart/signed properly) and a utility
6628 called 'smime' to call all this stuff. This is based on code I
6629 originally wrote for Celo who have kindly allowed it to be
6630 included in OpenSSL.
6633 *) Add variants des_set_key_checked and des_set_key_unchecked of
6634 des_set_key (aka des_key_sched). Global variable des_check_key
6635 decides which of these is called by des_set_key; this way
6636 des_check_key behaves as it always did, but applications and
6637 the library itself, which was buggy for des_check_key == 1,
6638 have a cleaner way to pick the version they need.
6641 *) New function PKCS12_newpass() which changes the password of a
6645 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6646 dynamic mix. In both cases the ids can be used as an index into the
6647 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6648 functions so they accept a list of the field values and the
6649 application doesn't need to directly manipulate the X509_TRUST
6653 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6657 *) Modify the way the V3 extension code looks up extensions. This now
6658 works in a similar way to the object code: we have some "standard"
6659 extensions in a static table which is searched with OBJ_bsearch()
6660 and the application can add dynamic ones if needed. The file
6661 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6662 updated whenever a new extension is added to the core code and kept
6663 in ext_nid order. There is a simple program 'tabtest.c' which checks
6664 this. New extensions are not added too often so this file can readily
6665 be maintained manually.
6667 There are two big advantages in doing things this way. The extensions
6668 can be looked up immediately and no longer need to be "added" using
6669 X509V3_add_standard_extensions(): this function now does nothing.
6670 [Side note: I get *lots* of email saying the extension code doesn't
6671 work because people forget to call this function]
6672 Also no dynamic allocation is done unless new extensions are added:
6673 so if we don't add custom extensions there is no need to call
6674 X509V3_EXT_cleanup().
6677 *) Modify enc utility's salting as follows: make salting the default. Add a
6678 magic header, so unsalted files fail gracefully instead of just decrypting
6679 to garbage. This is because not salting is a big security hole, so people
6680 should be discouraged from doing it.
6683 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6684 digest to be passed on the command line but it only used this
6685 parameter when signing a certificate. Modified so all relevant
6686 operations are affected by the digest parameter including the
6687 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6688 DSA key was used because it didn't fix the digest.
6691 *) Initial certificate chain verify code. Currently tests the untrusted
6692 certificates for consistency with the verify purpose (which is set
6693 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6695 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6696 this is because it will reject chains with invalid extensions whereas
6697 every previous version of OpenSSL and SSLeay made no checks at all.
6699 Trust code: checks the root CA for the relevant trust settings. Trust
6700 settings have an initial value consistent with the verify purpose: e.g.
6701 if the verify purpose is for SSL client use it expects the CA to be
6702 trusted for SSL client use. However the default value can be changed to
6703 permit custom trust settings: one example of this would be to only trust
6704 certificates from a specific "secure" set of CAs.
6706 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6707 which should be used for version portability: especially since the
6708 verify structure is likely to change more often now.
6710 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6711 to set them. If not set then assume SSL clients will verify SSL servers
6714 Two new options to the verify program: -untrusted allows a set of
6715 untrusted certificates to be passed in and -purpose which sets the
6716 intended purpose of the certificate. If a purpose is set then the
6717 new chain verify code is used to check extension consistency.
6720 *) Support for the authority information access extension.
6723 *) Modify RSA and DSA PEM read routines to transparently handle
6724 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6725 public keys in a format compatible with certificate
6726 SubjectPublicKeyInfo structures. Unfortunately there were already
6727 functions called *_PublicKey_* which used various odd formats so
6728 these are retained for compatibility: however the DSA variants were
6729 never in a public release so they have been deleted. Changed dsa/rsa
6730 utilities to handle the new format: note no releases ever handled public
6731 keys so we should be OK.
6733 The primary motivation for this change is to avoid the same fiasco
6734 that dogs private keys: there are several incompatible private key
6735 formats some of which are standard and some OpenSSL specific and
6736 require various evil hacks to allow partial transparent handling and
6737 even then it doesn't work with DER formats. Given the option anything
6738 other than PKCS#8 should be dumped: but the other formats have to
6739 stay in the name of compatibility.
6741 With public keys and the benefit of hindsight one standard format
6742 is used which works with EVP_PKEY, RSA or DSA structures: though
6743 it clearly returns an error if you try to read the wrong kind of key.
6745 Added a -pubkey option to the 'x509' utility to output the public key.
6746 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6747 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6748 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6749 that do the same as the EVP_PKEY_assign_*() except they up the
6750 reference count of the added key (they don't "swallow" the
6754 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6755 CRLs would fail if the file contained no certificates or no CRLs:
6756 added a new function to read in both types and return the number
6757 read: this means that if none are read it will be an error. The
6758 DER versions of the certificate and CRL reader would always fail
6759 because it isn't possible to mix certificates and CRLs in DER format
6760 without choking one or the other routine. Changed this to just read
6761 a certificate: this is the best we can do. Also modified the code
6762 in apps/verify.c to take notice of return codes: it was previously
6763 attempting to read in certificates from NULL pointers and ignoring
6764 any errors: this is one reason why the cert and CRL reader seemed
6765 to work. It doesn't check return codes from the default certificate
6766 routines: these may well fail if the certificates aren't installed.
6769 *) Code to support otherName option in GeneralName.
6772 *) First update to verify code. Change the verify utility
6773 so it warns if it is passed a self signed certificate:
6774 for consistency with the normal behaviour. X509_verify
6775 has been modified to it will now verify a self signed
6776 certificate if *exactly* the same certificate appears
6777 in the store: it was previously impossible to trust a
6778 single self signed certificate. This means that:
6779 openssl verify ss.pem
6780 now gives a warning about a self signed certificate but
6781 openssl verify -CAfile ss.pem ss.pem
6785 *) For servers, store verify_result in SSL_SESSION data structure
6786 (and add it to external session representation).
6787 This is needed when client certificate verifications fails,
6788 but an application-provided verification callback (set by
6789 SSL_CTX_set_cert_verify_callback) allows accepting the session
6790 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6791 but returns 1): When the session is reused, we have to set
6792 ssl->verify_result to the appropriate error code to avoid
6794 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6796 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6797 case in PKCS7_dataInit() where the signed PKCS7 structure
6798 didn't contain any existing data because it was being created.
6799 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6801 *) Add a salt to the key derivation routines in enc.c. This
6802 forms the first 8 bytes of the encrypted file. Also add a
6803 -S option to allow a salt to be input on the command line.
6806 *) New function X509_cmp(). Oddly enough there wasn't a function
6807 to compare two certificates. We do this by working out the SHA1
6808 hash and comparing that. X509_cmp() will be needed by the trust
6812 *) SSL_get1_session() is like SSL_get_session(), but increments
6813 the reference count in the SSL_SESSION returned.
6814 [Geoff Thorpe <geoff@eu.c2.net>]
6816 *) Fix for 'req': it was adding a null to request attributes.
6817 Also change the X509_LOOKUP and X509_INFO code to handle
6818 certificate auxiliary information.
6821 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6825 *) Add the possibility to add extra information to the memory leak
6826 detecting output, to form tracebacks, showing from where each
6827 allocation was originated: CRYPTO_push_info("constant string") adds
6828 the string plus current file name and line number to a per-thread
6829 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6830 is like calling CYRPTO_pop_info() until the stack is empty.
6831 Also updated memory leak detection code to be multi-thread-safe.
6834 *) Add options -text and -noout to pkcs7 utility and delete the
6835 encryption options which never did anything. Update docs.
6838 *) Add options to some of the utilities to allow the pass phrase
6839 to be included on either the command line (not recommended on
6840 OSes like Unix) or read from the environment. Update the
6841 manpages and fix a few bugs.
6844 *) Add a few manpages for some of the openssl commands.
6847 *) Fix the -revoke option in ca. It was freeing up memory twice,
6848 leaking and not finding already revoked certificates.
6851 *) Extensive changes to support certificate auxiliary information.
6852 This involves the use of X509_CERT_AUX structure and X509_AUX
6853 functions. An X509_AUX function such as PEM_read_X509_AUX()
6854 can still read in a certificate file in the usual way but it
6855 will also read in any additional "auxiliary information". By
6856 doing things this way a fair degree of compatibility can be
6857 retained: existing certificates can have this information added
6858 using the new 'x509' options.
6860 Current auxiliary information includes an "alias" and some trust
6861 settings. The trust settings will ultimately be used in enhanced
6862 certificate chain verification routines: currently a certificate
6863 can only be trusted if it is self signed and then it is trusted
6867 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6868 The problem was that one of the replacement routines had not been working
6869 since SSLeay releases. For now the offending routine has been replaced
6870 with non-optimised assembler. Even so, this now gives around 95%
6871 performance improvement for 1024 bit RSA signs.
6874 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6875 handling. Most clients have the effective key size in bits equal to
6876 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6877 A few however don't do this and instead use the size of the decrypted key
6878 to determine the RC2 key length and the AlgorithmIdentifier to determine
6879 the effective key length. In this case the effective key length can still
6880 be 40 bits but the key length can be 168 bits for example. This is fixed
6881 by manually forcing an RC2 key into the EVP_PKEY structure because the
6882 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6883 the key length and effective key length are equal.
6886 *) Add a bunch of functions that should simplify the creation of
6887 X509_NAME structures. Now you should be able to do:
6888 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6889 and have it automatically work out the correct field type and fill in
6890 the structures. The more adventurous can try:
6891 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6892 and it will (hopefully) work out the correct multibyte encoding.
6895 *) Change the 'req' utility to use the new field handling and multibyte
6896 copy routines. Before the DN field creation was handled in an ad hoc
6897 way in req, ca, and x509 which was rather broken and didn't support
6898 BMPStrings or UTF8Strings. Since some software doesn't implement
6899 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6900 using the dirstring_type option. See the new comment in the default
6901 openssl.cnf for more info.
6904 *) Make crypto/rand/md_rand.c more robust:
6905 - Assure unique random numbers after fork().
6906 - Make sure that concurrent threads access the global counter and
6907 md serializably so that we never lose entropy in them
6908 or use exactly the same state in multiple threads.
6909 Access to the large state is not always serializable because
6910 the additional locking could be a performance killer, and
6911 md should be large enough anyway.
6914 *) New file apps/app_rand.c with commonly needed functionality
6915 for handling the random seed file.
6917 Use the random seed file in some applications that previously did not:
6919 dsaparam -genkey (which also ignored its '-rand' option),
6922 x509 (when signing).
6923 Except on systems with /dev/urandom, it is crucial to have a random
6924 seed file at least for key creation, DSA signing, and for DH exchanges;
6925 for RSA signatures we could do without one.
6927 gendh and gendsa (unlike genrsa) used to read only the first byte
6928 of each file listed in the '-rand' option. The function as previously
6929 found in genrsa is now in app_rand.c and is used by all programs
6930 that support '-rand'.
6933 *) In RAND_write_file, use mode 0600 for creating files;
6934 don't just chmod when it may be too late.
6937 *) Report an error from X509_STORE_load_locations
6938 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6941 *) New function ASN1_mbstring_copy() this copies a string in either
6942 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6943 into an ASN1_STRING type. A mask of permissible types is passed
6944 and it chooses the "minimal" type to use or an error if not type
6948 *) Add function equivalents to the various macros in asn1.h. The old
6949 macros are retained with an M_ prefix. Code inside the library can
6950 use the M_ macros. External code (including the openssl utility)
6951 should *NOT* in order to be "shared library friendly".
6954 *) Add various functions that can check a certificate's extensions
6955 to see if it usable for various purposes such as SSL client,
6956 server or S/MIME and CAs of these types. This is currently
6957 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6958 verification. Also added a -purpose flag to x509 utility to
6959 print out all the purposes.
6962 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6966 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6967 for, obtain and decode and extension and obtain its critical flag.
6968 This allows all the necessary extension code to be handled in a
6969 single function call.
6972 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6973 platforms. See crypto/rc4/rc4_enc.c for further details.
6976 *) New -noout option to asn1parse. This causes no output to be produced
6977 its main use is when combined with -strparse and -out to extract data
6978 from a file (which may not be in ASN.1 format).
6981 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6982 when producing the local key id.
6983 [Richard Levitte <levitte@stacken.kth.se>]
6985 *) New option -dhparam in s_server. This allows a DH parameter file to be
6986 stated explicitly. If it is not stated then it tries the first server
6987 certificate file. The previous behaviour hard coded the filename
6991 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6992 a public key to be input or output. For example:
6993 openssl rsa -in key.pem -pubout -out pubkey.pem
6994 Also added necessary DSA public key functions to handle this.
6997 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6998 in the message. This was handled by allowing
6999 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7000 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7002 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7003 to the end of the strings whereas this didn't. This would cause problems
7004 if strings read with d2i_ASN1_bytes() were later modified.
7005 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7007 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7008 data and it contains EOF it will end up returning an error. This is
7009 caused by input 46 bytes long. The cause is due to the way base64
7010 BIOs find the start of base64 encoded data. They do this by trying a
7011 trial decode on each line until they find one that works. When they
7012 do a flag is set and it starts again knowing it can pass all the
7013 data directly through the decoder. Unfortunately it doesn't reset
7014 the context it uses. This means that if EOF is reached an attempt
7015 is made to pass two EOFs through the context and this causes the
7016 resulting error. This can also cause other problems as well. As is
7017 usual with these problems it takes *ages* to find and the fix is
7018 trivial: move one line.
7019 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7021 *) Ugly workaround to get s_client and s_server working under Windows. The
7022 old code wouldn't work because it needed to select() on sockets and the
7023 tty (for keypresses and to see if data could be written). Win32 only
7024 supports select() on sockets so we select() with a 1s timeout on the
7025 sockets and then see if any characters are waiting to be read, if none
7026 are present then we retry, we also assume we can always write data to
7027 the tty. This isn't nice because the code then blocks until we've
7028 received a complete line of data and it is effectively polling the
7029 keyboard at 1s intervals: however it's quite a bit better than not
7030 working at all :-) A dedicated Windows application might handle this
7031 with an event loop for example.
7034 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7035 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7036 will be called when RSA_sign() and RSA_verify() are used. This is useful
7037 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7038 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7039 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7040 This necessitated the support of an extra signature type NID_md5_sha1
7041 for SSL signatures and modifications to the SSL library to use it instead
7042 of calling RSA_public_decrypt() and RSA_private_encrypt().
7045 *) Add new -verify -CAfile and -CApath options to the crl program, these
7046 will lookup a CRL issuers certificate and verify the signature in a
7047 similar way to the verify program. Tidy up the crl program so it
7048 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7049 less strict. It will now permit CRL extensions even if it is not
7050 a V2 CRL: this will allow it to tolerate some broken CRLs.
7053 *) Initialize all non-automatic variables each time one of the openssl
7054 sub-programs is started (this is necessary as they may be started
7055 multiple times from the "OpenSSL>" prompt).
7056 [Lennart Bang, Bodo Moeller]
7058 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7059 removing all other RSA functionality (this is what NO_RSA does). This
7060 is so (for example) those in the US can disable those operations covered
7061 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7065 *) Non-copying interface to BIO pairs.
7066 (still largely untested)
7069 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7070 ASCII string. This was handled independently in various places before.
7073 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7074 UTF8 strings a character at a time.
7077 *) Use client_version from client hello to select the protocol
7078 (s23_srvr.c) and for RSA client key exchange verification
7079 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7082 *) Add various utility functions to handle SPKACs, these were previously
7083 handled by poking round in the structure internals. Added new function
7084 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7085 print, verify and generate SPKACs. Based on an original idea from
7086 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7089 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7092 *) Allow the config file extension section to be overwritten on the
7093 command line. Based on an original idea from Massimiliano Pala
7094 <madwolf@comune.modena.it>. The new option is called -extensions
7095 and can be applied to ca, req and x509. Also -reqexts to override
7096 the request extensions in req and -crlexts to override the crl extensions
7100 *) Add new feature to the SPKAC handling in ca. Now you can include
7101 the same field multiple times by preceding it by "XXXX." for example:
7104 this is the same syntax as used in the req config file.
7107 *) Allow certificate extensions to be added to certificate requests. These
7108 are specified in a 'req_extensions' option of the req section of the
7109 config file. They can be printed out with the -text option to req but
7110 are otherwise ignored at present.
7113 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7114 data read consists of only the final block it would not decrypted because
7115 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7116 A misplaced 'break' also meant the decrypted final block might not be
7117 copied until the next read.
7120 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7121 a few extra parameters to the DH structure: these will be useful if
7122 for example we want the value of 'q' or implement X9.42 DH.
7125 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7126 provides hooks that allow the default DSA functions or functions on a
7127 "per key" basis to be replaced. This allows hardware acceleration and
7128 hardware key storage to be handled without major modification to the
7129 library. Also added low level modexp hooks and CRYPTO_EX structure and
7130 associated functions.
7133 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7134 as "read only": it can't be written to and the buffer it points to will
7135 not be freed. Reading from a read only BIO is much more efficient than
7136 a normal memory BIO. This was added because there are several times when
7137 an area of memory needs to be read from a BIO. The previous method was
7138 to create a memory BIO and write the data to it, this results in two
7139 copies of the data and an O(n^2) reading algorithm. There is a new
7140 function BIO_new_mem_buf() which creates a read only memory BIO from
7141 an area of memory. Also modified the PKCS#7 routines to use read only
7145 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7146 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7147 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7148 but a retry condition occured while trying to read the rest.
7151 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7152 NID_pkcs7_encrypted by default: this was wrong since this should almost
7153 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7154 the encrypted data type: this is a more sensible place to put it and it
7155 allows the PKCS#12 code to be tidied up that duplicated this
7159 *) Changed obj_dat.pl script so it takes its input and output files on
7160 the command line. This should avoid shell escape redirection problems
7164 *) Initial support for certificate extension requests, these are included
7165 in things like Xenroll certificate requests. Included functions to allow
7166 extensions to be obtained and added.
7169 *) -crlf option to s_client and s_server for sending newlines as
7170 CRLF (as required by many protocols).
7173 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7175 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7176 [Ralf S. Engelschall]
7178 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7179 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7181 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7185 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7186 DH parameters/keys (q is lost during that conversion, but the resulting
7187 DH parameters contain its length).
7189 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7190 much faster than DH_generate_parameters (which creates parameters
7191 where p = 2*q + 1), and also the smaller q makes DH computations
7192 much more efficient (160-bit exponentiation instead of 1024-bit
7193 exponentiation); so this provides a convenient way to support DHE
7194 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7195 utter importance to use
7196 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7198 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7199 when such DH parameters are used, because otherwise small subgroup
7200 attacks may become possible!
7203 *) Avoid memory leak in i2d_DHparams.
7206 *) Allow the -k option to be used more than once in the enc program:
7207 this allows the same encrypted message to be read by multiple recipients.
7210 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7211 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7212 it will always use the numerical form of the OID, even if it has a short
7216 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7217 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7218 otherwise bn_mod_exp was called. In the case of hardware keys for example
7219 no private key components need be present and it might store extra data
7220 in the RSA structure, which cannot be accessed from bn_mod_exp.
7221 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7222 private key operations.
7225 *) Added support for SPARC Linux.
7228 *) pem_password_cb function type incompatibly changed from
7229 typedef int pem_password_cb(char *buf, int size, int rwflag);
7231 ....(char *buf, int size, int rwflag, void *userdata);
7232 so that applications can pass data to their callbacks:
7233 The PEM[_ASN1]_{read,write}... functions and macros now take an
7234 additional void * argument, which is just handed through whenever
7235 the password callback is called.
7236 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7238 New function SSL_CTX_set_default_passwd_cb_userdata.
7240 Compatibility note: As many C implementations push function arguments
7241 onto the stack in reverse order, the new library version is likely to
7242 interoperate with programs that have been compiled with the old
7243 pem_password_cb definition (PEM_whatever takes some data that
7244 happens to be on the stack as its last argument, and the callback
7245 just ignores this garbage); but there is no guarantee whatsoever that
7248 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7249 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7250 problems not only on Windows, but also on some Unix platforms.
7251 To avoid problematic command lines, these definitions are now in an
7252 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7253 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7256 *) MIPS III/IV assembler module is reimplemented.
7259 *) More DES library cleanups: remove references to srand/rand and
7260 delete an unused file.
7263 *) Add support for the the free Netwide assembler (NASM) under Win32,
7264 since not many people have MASM (ml) and it can be hard to obtain.
7265 This is currently experimental but it seems to work OK and pass all
7266 the tests. Check out INSTALL.W32 for info.
7269 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7270 without temporary keys kept an extra copy of the server key,
7271 and connections with temporary keys did not free everything in case
7275 *) New function RSA_check_key and new openssl rsa option -check
7276 for verifying the consistency of RSA keys.
7277 [Ulf Moeller, Bodo Moeller]
7279 *) Various changes to make Win32 compile work:
7280 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7281 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7282 comparison" warnings.
7283 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7286 *) Add a debugging option to PKCS#5 v2 key generation function: when
7287 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7288 derived keys are printed to stderr.
7291 *) Copy the flags in ASN1_STRING_dup().
7292 [Roman E. Pavlov <pre@mo.msk.ru>]
7294 *) The x509 application mishandled signing requests containing DSA
7295 keys when the signing key was also DSA and the parameters didn't match.
7297 It was supposed to omit the parameters when they matched the signing key:
7298 the verifying software was then supposed to automatically use the CA's
7299 parameters if they were absent from the end user certificate.
7301 Omitting parameters is no longer recommended. The test was also
7302 the wrong way round! This was probably due to unusual behaviour in
7303 EVP_cmp_parameters() which returns 1 if the parameters match.
7304 This meant that parameters were omitted when they *didn't* match and
7305 the certificate was useless. Certificates signed with 'ca' didn't have
7307 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7309 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7310 The interface is as follows:
7311 Applications can use
7312 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7313 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7314 "off" is now the default.
7315 The library internally uses
7316 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7317 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7318 to disable memory-checking temporarily.
7320 Some inconsistent states that previously were possible (and were
7321 even the default) are now avoided.
7323 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7324 with each memory chunk allocated; this is occasionally more helpful
7325 than just having a counter.
7327 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7329 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7333 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7334 which largely parallels "options", but is for changing API behaviour,
7335 whereas "options" are about protocol behaviour.
7336 Initial "mode" flags are:
7338 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7339 a single record has been written.
7340 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7341 retries use the same buffer location.
7342 (But all of the contents must be
7346 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7349 *) Fix problems with no-hmac etc.
7350 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7352 *) New functions RSA_get_default_method(), RSA_set_method() and
7353 RSA_get_method(). These allows replacement of RSA_METHODs without having
7354 to mess around with the internals of an RSA structure.
7357 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7358 Also really enable memory leak checks in openssl.c and in some
7360 [Chad C. Mulligan, Bodo Moeller]
7362 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7363 up the length of negative integers. This has now been simplified to just
7364 store the length when it is first determined and use it later, rather
7365 than trying to keep track of where data is copied and updating it to
7367 [Steve Henson, reported by Brien Wheeler
7368 <bwheeler@authentica-security.com>]
7370 *) Add a new function PKCS7_signatureVerify. This allows the verification
7371 of a PKCS#7 signature but with the signing certificate passed to the
7372 function itself. This contrasts with PKCS7_dataVerify which assumes the
7373 certificate is present in the PKCS#7 structure. This isn't always the
7374 case: certificates can be omitted from a PKCS#7 structure and be
7375 distributed by "out of band" means (such as a certificate database).
7378 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7379 function prototypes in pem.h, also change util/mkdef.pl to add the
7380 necessary function names.
7383 *) mk1mf.pl (used by Windows builds) did not properly read the
7384 options set by Configure in the top level Makefile, and Configure
7385 was not even able to write more than one option correctly.
7386 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7389 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7390 file to be loaded from a BIO or FILE pointer. The BIO version will
7391 for example allow memory BIOs to contain config info.
7394 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7395 Whoever hopes to achieve shared-library compatibility across versions
7396 must use this, not the compile-time macro.
7397 (Exercise 0.9.4: Which is the minimum library version required by
7399 Note: All this applies only to multi-threaded programs, others don't
7403 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7404 through a BIO pair triggered the default case, i.e.
7405 SSLerr(...,SSL_R_UNKNOWN_STATE).
7408 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7409 can use the SSL library even if none of the specific BIOs is
7413 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7414 for the encoded length.
7415 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7417 *) Add initial documentation of the X509V3 functions.
7420 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7421 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7422 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7423 secure PKCS#8 private key format with a high iteration count.
7426 *) Fix determination of Perl interpreter: A perl or perl5
7427 _directory_ in $PATH was also accepted as the interpreter.
7428 [Ralf S. Engelschall]
7430 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7431 wrong with it but it was very old and did things like calling
7432 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7436 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7437 to use the new extension code.
7440 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7441 with macros. This should make it easier to change their form, add extra
7442 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7446 *) Add to configuration table a new entry that can specify an alternative
7447 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7448 according to Mark Crispin <MRC@Panda.COM>.
7452 *) DES CBC did not update the IV. Weird.
7455 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7456 Changing the behaviour of the former might break existing programs --
7457 where IV updating is needed, des_ncbc_encrypt can be used.
7460 *) When bntest is run from "make test" it drives bc to check its
7461 calculations, as well as internally checking them. If an internal check
7462 fails, it needs to cause bc to give a non-zero result or make test carries
7463 on without noticing the failure. Fixed.
7466 *) DES library cleanups.
7469 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7470 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7471 ciphers. NOTE: although the key derivation function has been verified
7472 against some published test vectors it has not been extensively tested
7473 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7477 *) Instead of "mkdir -p", which is not fully portable, use new
7478 Perl script "util/mkdir-p.pl".
7481 *) Rewrite the way password based encryption (PBE) is handled. It used to
7482 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7483 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7484 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7485 the 'parameter' field of the AlgorithmIdentifier is passed to the
7486 underlying key generation function so it must do its own ASN1 parsing.
7487 This has also changed the EVP_PBE_CipherInit() function which now has a
7488 'parameter' argument instead of literal salt and iteration count values
7489 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7492 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7493 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7494 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7495 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7496 value was just used as a "magic string" and not used directly its
7497 value doesn't matter.
7500 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7504 *) "linux-sparc64" configuration (ultrapenguin).
7505 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7506 "linux-sparc" configuration.
7507 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7509 *) config now generates no-xxx options for missing ciphers.
7512 *) Support the EBCDIC character set (work in progress).
7513 File ebcdic.c not yet included because it has a different license.
7514 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7516 *) Support BS2000/OSD-POSIX.
7517 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7519 *) Make callbacks for key generation use void * instead of char *.
7522 *) Make S/MIME samples compile (not yet tested).
7525 *) Additional typesafe stacks.
7528 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7532 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7534 *) New configuration variant "sco5-gcc".
7536 *) Updated some demos.
7537 [Sean O Riordain, Wade Scholine]
7539 *) Add missing BIO_free at exit of pkcs12 application.
7542 *) Fix memory leak in conf.c.
7545 *) Updates for Win32 to assembler version of MD5.
7548 *) Set #! path to perl in apps/der_chop to where we found it
7549 instead of using a fixed path.
7552 *) SHA library changes for irix64-mips4-cc.
7555 *) Improvements for VMS support.
7559 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7561 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7562 This also avoids the problems with SC4.2 and unpatched SC5.
7563 [Andy Polyakov <appro@fy.chalmers.se>]
7565 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7566 These are required because of the typesafe stack would otherwise break
7567 existing code. If old code used a structure member which used to be STACK
7568 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7569 sk_num or sk_value it would produce an error because the num, data members
7570 are not present in STACK_OF. Now it just produces a warning. sk_set
7571 replaces the old method of assigning a value to sk_value
7572 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7573 that does this will no longer work (and should use sk_set instead) but
7574 this could be regarded as a "questionable" behaviour anyway.
7577 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7578 correctly handle encrypted S/MIME data.
7581 *) Change type of various DES function arguments from des_cblock
7582 (which means, in function argument declarations, pointer to char)
7583 to des_cblock * (meaning pointer to array with 8 char elements),
7584 which allows the compiler to do more typechecking; it was like
7585 that back in SSLeay, but with lots of ugly casts.
7587 Introduce new type const_des_cblock.
7590 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7591 problems: find RecipientInfo structure that matches recipient certificate
7592 and initialise the ASN1 structures properly based on passed cipher.
7595 *) Belatedly make the BN tests actually check the results.
7598 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7599 to and from BNs: it was completely broken. New compilation option
7600 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7601 key elements as negative integers.
7604 *) Reorganize and speed up MD5.
7605 [Andy Polyakov <appro@fy.chalmers.se>]
7608 [Richard Levitte <richard@levitte.org>]
7610 *) New option -out to asn1parse to allow the parsed structure to be
7611 output to a file. This is most useful when combined with the -strparse
7612 option to examine the output of things like OCTET STRINGS.
7615 *) Make SSL library a little more fool-proof by not requiring any longer
7616 that SSL_set_{accept,connect}_state be called before
7617 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7618 in many applications because usually everything *appeared* to work as
7619 intended anyway -- now it really works as intended).
7622 *) Move openssl.cnf out of lib/.
7625 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7626 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7627 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7628 [Ralf S. Engelschall]
7630 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7631 handle PKCS#7 enveloped data properly.
7632 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7634 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7635 copying pointers. The cert_st handling is changed by this in
7636 various ways (and thus what used to be known as ctx->default_cert
7637 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7638 any longer when s->cert does not give us what we need).
7639 ssl_cert_instantiate becomes obsolete by this change.
7640 As soon as we've got the new code right (possibly it already is?),
7641 we have solved a couple of bugs of the earlier code where s->cert
7642 was used as if it could not have been shared with other SSL structures.
7644 Note that using the SSL API in certain dirty ways now will result
7645 in different behaviour than observed with earlier library versions:
7646 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7647 does not influence s as it used to.
7649 In order to clean up things more thoroughly, inside SSL_SESSION
7650 we don't use CERT any longer, but a new structure SESS_CERT
7651 that holds per-session data (if available); currently, this is
7652 the peer's certificate chain and, for clients, the server's certificate
7653 and temporary key. CERT holds only those values that can have
7654 meaningful defaults in an SSL_CTX.
7657 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7658 from the internal representation. Various PKCS#7 fixes: remove some
7659 evil casts and set the enc_dig_alg field properly based on the signing
7663 *) Allow PKCS#12 password to be set from the command line or the
7664 environment. Let 'ca' get its config file name from the environment
7665 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7669 *) Allow certificate policies extension to use an IA5STRING for the
7670 organization field. This is contrary to the PKIX definition but
7671 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7675 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7676 without disallowing inline assembler and the like for non-pedantic builds.
7679 *) Support Borland C++ builder.
7680 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7685 *) SHA-1 cleanups and performance enhancements.
7686 [Andy Polyakov <appro@fy.chalmers.se>]
7688 *) Sparc v8plus assembler for the bignum library.
7689 [Andy Polyakov <appro@fy.chalmers.se>]
7691 *) Accept any -xxx and +xxx compiler options in Configure.
7694 *) Update HPUX configuration.
7697 *) Add missing sk_<type>_unshift() function to safestack.h
7698 [Ralf S. Engelschall]
7700 *) New function SSL_CTX_use_certificate_chain_file that sets the
7701 "extra_cert"s in addition to the certificate. (This makes sense
7702 only for "PEM" format files, as chains as a whole are not
7706 *) Support verify_depth from the SSL API.
7707 x509_vfy.c had what can be considered an off-by-one-error:
7708 Its depth (which was not part of the external interface)
7709 was actually counting the number of certificates in a chain;
7710 now it really counts the depth.
7713 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7714 instead of X509err, which often resulted in confusing error
7715 messages since the error codes are not globally unique
7716 (e.g. an alleged error in ssl3_accept when a certificate
7717 didn't match the private key).
7719 *) New function SSL_CTX_set_session_id_context that allows to set a default
7720 value (so that you don't need SSL_set_session_id_context for each
7721 connection using the SSL_CTX).
7724 *) OAEP decoding bug fix.
7727 *) Support INSTALL_PREFIX for package builders, as proposed by
7731 *) New Configure options "threads" and "no-threads". For systems
7732 where the proper compiler options are known (currently Solaris
7733 and Linux), "threads" is the default.
7736 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7739 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7740 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7741 such as /usr/local/bin.
7744 *) "make linux-shared" to build shared libraries.
7745 [Niels Poppe <niels@netbox.org>]
7747 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7750 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7751 extension adding in x509 utility.
7754 *) Remove NOPROTO sections and error code comments.
7757 *) Partial rewrite of the DEF file generator to now parse the ANSI
7761 *) New Configure options --prefix=DIR and --openssldir=DIR.
7764 *) Complete rewrite of the error code script(s). It is all now handled
7765 by one script at the top level which handles error code gathering,
7766 header rewriting and C source file generation. It should be much better
7767 than the old method: it now uses a modified version of Ulf's parser to
7768 read the ANSI prototypes in all header files (thus the old K&R definitions
7769 aren't needed for error creation any more) and do a better job of
7770 translating function codes into names. The old 'ASN1 error code imbedded
7771 in a comment' is no longer necessary and it doesn't use .err files which
7772 have now been deleted. Also the error code call doesn't have to appear all
7773 on one line (which resulted in some large lines...).
7776 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7779 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7780 0 (which usually indicates a closed connection), but continue reading.
7783 *) Fix some race conditions.
7786 *) Add support for CRL distribution points extension. Add Certificate
7787 Policies and CRL distribution points documentation.
7790 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7793 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7794 8 of keying material. Merlin has also confirmed interop with this fix
7795 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7796 [Merlin Hughes <merlin@baltimore.ie>]
7798 *) Fix lots of warnings.
7799 [Richard Levitte <levitte@stacken.kth.se>]
7801 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7802 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7803 [Richard Levitte <levitte@stacken.kth.se>]
7805 *) Fix problems with sizeof(long) == 8.
7806 [Andy Polyakov <appro@fy.chalmers.se>]
7808 *) Change functions to ANSI C.
7811 *) Fix typos in error codes.
7812 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7814 *) Remove defunct assembler files from Configure.
7817 *) SPARC v8 assembler BIGNUM implementation.
7818 [Andy Polyakov <appro@fy.chalmers.se>]
7820 *) Support for Certificate Policies extension: both print and set.
7821 Various additions to support the r2i method this uses.
7824 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7825 return a const string when you are expecting an allocated buffer.
7828 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7829 types DirectoryString and DisplayText.
7832 *) Add code to allow r2i extensions to access the configuration database,
7833 add an LHASH database driver and add several ctx helper functions.
7836 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7837 fail when they extended the size of a BIGNUM.
7840 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7841 support typesafe stack.
7844 *) Fix typo in SSL_[gs]et_options().
7845 [Nils Frostberg <nils@medcom.se>]
7847 *) Delete various functions and files that belonged to the (now obsolete)
7848 old X509V3 handling code.
7851 *) New Configure option "rsaref".
7854 *) Don't auto-generate pem.h.
7857 *) Introduce type-safe ASN.1 SETs.
7860 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7861 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7863 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7864 that links with OpenSSL (well at least cause lots of warnings), but fear
7865 not: the conversion is trivial, and it eliminates loads of evil casts. A
7866 few STACKed things have been converted already. Feel free to convert more.
7867 In the fullness of time, I'll do away with the STACK type altogether.
7870 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7871 specified in <certfile> by updating the entry in the index.txt file.
7872 This way one no longer has to edit the index.txt file manually for
7873 revoking a certificate. The -revoke option does the gory details now.
7874 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7876 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7877 `-text' option at all and this way the `-noout -text' combination was
7878 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7879 [Ralf S. Engelschall]
7881 *) Make sure a corresponding plain text error message exists for the
7882 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7883 verify callback function determined that a certificate was revoked.
7884 [Ralf S. Engelschall]
7886 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7887 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7888 all available cipers including rc5, which was forgotten until now.
7889 In order to let the testing shell script know which algorithms
7890 are available, a new (up to now undocumented) command
7891 "openssl list-cipher-commands" is used.
7894 *) Bugfix: s_client occasionally would sleep in select() when
7895 it should have checked SSL_pending() first.
7898 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7899 the raw DSA values prior to ASN.1 encoding.
7902 *) Tweaks to Configure
7903 [Niels Poppe <niels@netbox.org>]
7905 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7909 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7912 *) New config option to avoid instructions that are illegal on the 80386.
7913 The default code is faster, but requires at least a 486.
7916 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7917 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7918 same as SSL2_VERSION anyway.
7921 *) New "-showcerts" option for s_client.
7924 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7925 application. Various cleanups and fixes.
7928 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7929 modify error routines to work internally. Add error codes and PBE init
7930 to library startup routines.
7933 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7934 packing functions to asn1 and evp. Changed function names and error
7935 codes along the way.
7938 *) PKCS12 integration: and so it begins... First of several patches to
7939 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7940 objects to objects.h
7943 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7944 and display support for Thawte strong extranet extension.
7947 *) Add LinuxPPC support.
7948 [Jeff Dubrule <igor@pobox.org>]
7950 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7951 bn_div_words in alpha.s.
7952 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7954 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7955 OAEP isn't supported when OpenSSL is built with RSAref.
7956 [Ulf Moeller <ulf@fitug.de>]
7958 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7959 so they no longer are missing under -DNOPROTO.
7960 [Soren S. Jorvang <soren@t.dk>]
7963 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7965 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7966 doesn't work when the session is reused. Coming soon!
7969 *) Fix a security hole, that allows sessions to be reused in the wrong
7970 context thus bypassing client cert protection! All software that uses
7971 client certs and session caches in multiple contexts NEEDS PATCHING to
7972 allow session reuse! A fuller solution is in the works.
7973 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7975 *) Some more source tree cleanups (removed obsolete files
7976 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7977 permission on "config" script to be executable) and a fix for the INSTALL
7979 [Ulf Moeller <ulf@fitug.de>]
7981 *) Remove some legacy and erroneous uses of malloc, free instead of
7983 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7985 *) Make rsa_oaep_test return non-zero on error.
7986 [Ulf Moeller <ulf@fitug.de>]
7988 *) Add support for native Solaris shared libraries. Configure
7989 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7990 if someone would make that last step automatic.
7991 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7993 *) ctx_size was not built with the right compiler during "make links". Fixed.
7996 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7997 except NULL ciphers". This means the default cipher list will no longer
7998 enable NULL ciphers. They need to be specifically enabled e.g. with
7999 the string "DEFAULT:eNULL".
8002 *) Fix to RSA private encryption routines: if p < q then it would
8003 occasionally produce an invalid result. This will only happen with
8004 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8007 *) Be less restrictive and allow also `perl util/perlpath.pl
8008 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8009 because this way one can also use an interpreter named `perl5' (which is
8010 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8011 installed as `perl').
8012 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8014 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8015 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8017 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8018 advapi32.lib to Win32 build and change the pem test comparision
8019 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8020 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8021 and crypto/des/ede_cbcm_enc.c.
8024 *) DES quad checksum was broken on big-endian architectures. Fixed.
8027 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8028 Win32 test batch file so it (might) work again. The Win32 test batch file
8029 is horrible: I feel ill....
8032 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8033 in e_os.h. Audit of header files to check ANSI and non ANSI
8034 sections: 10 functions were absent from non ANSI section and not exported
8035 from Windows DLLs. Fixed up libeay.num for new functions.
8038 *) Make `openssl version' output lines consistent.
8039 [Ralf S. Engelschall]
8041 *) Fix Win32 symbol export lists for BIO functions: Added
8042 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8043 to ms/libeay{16,32}.def.
8044 [Ralf S. Engelschall]
8046 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8047 fine under Unix and passes some trivial tests I've now added. But the
8048 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8049 added to make sure no one expects that this stuff really works in the
8050 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8051 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8053 [Ralf S. Engelschall]
8055 *) Fix the generation of two part addresses in perl.
8056 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8058 *) Add config entry for Linux on MIPS.
8059 [John Tobey <jtobey@channel1.com>]
8061 *) Make links whenever Configure is run, unless we are on Windoze.
8064 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8065 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8069 *) Add a useful kludge to allow package maintainers to specify compiler and
8070 other platforms details on the command line without having to patch the
8071 Configure script everytime: One now can use ``perl Configure
8072 <id>:<details>'', i.e. platform ids are allowed to have details appended
8073 to them (seperated by colons). This is treated as there would be a static
8074 pre-configured entry in Configure's %table under key <id> with value
8075 <details> and ``perl Configure <id>'' is called. So, when you want to
8076 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8077 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8078 now, which overrides the FreeBSD-elf entry on-the-fly.
8079 [Ralf S. Engelschall]
8081 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8084 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8085 on the `perl Configure ...' command line. This way one can compile
8086 OpenSSL libraries with Position Independent Code (PIC) which is needed
8087 for linking it into DSOs.
8088 [Ralf S. Engelschall]
8090 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8094 *) Cleaned up the LICENSE document: The official contact for any license
8095 questions now is the OpenSSL core team under openssl-core@openssl.org.
8096 And add a paragraph about the dual-license situation to make sure people
8097 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8098 to the OpenSSL toolkit.
8099 [Ralf S. Engelschall]
8101 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8102 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8103 Additonally cleaned up the `make links' target: Remove unnecessary
8104 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8105 to speed processing and no longer clutter the display with confusing
8106 stuff. Instead only the actually done links are displayed.
8107 [Ralf S. Engelschall]
8109 *) Permit null encryption ciphersuites, used for authentication only. It used
8110 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8111 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8115 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8116 signed attributes when verifying signatures (this would break them),
8117 the detached data encoding was wrong and public keys obtained using
8118 X509_get_pubkey() weren't freed.
8121 *) Add text documentation for the BUFFER functions. Also added a work around
8122 to a Win95 console bug. This was triggered by the password read stuff: the
8123 last character typed gets carried over to the next fread(). If you were
8124 generating a new cert request using 'req' for example then the last
8125 character of the passphrase would be CR which would then enter the first
8129 *) Added the new `Includes OpenSSL Cryptography Software' button as
8130 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8131 button and can be used by applications based on OpenSSL to show the
8132 relationship to the OpenSSL project.
8133 [Ralf S. Engelschall]
8135 *) Remove confusing variables in function signatures in files
8136 ssl/ssl_lib.c and ssl/ssl.h.
8137 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8139 *) Don't install bss_file.c under PREFIX/include/
8140 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8142 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8143 functions that return function pointers and has support for NT specific
8144 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8145 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8146 unsigned to signed types: this was killing the Win32 compile.
8149 *) Add new certificate file to stack functions,
8150 SSL_add_dir_cert_subjects_to_stack() and
8151 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8152 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8153 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8154 This means that Apache-SSL and similar packages don't have to mess around
8155 to add as many CAs as they want to the preferred list.
8158 *) Experiment with doxygen documentation. Currently only partially applied to
8160 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8161 openssl.doxy as the configuration file.
8164 *) Get rid of remaining C++-style comments which strict C compilers hate.
8165 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8167 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8168 compiled in by default: it has problems with large keys.
8171 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8172 DH private keys and/or callback functions which directly correspond to
8173 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8174 is needed for applications which have to configure certificates on a
8175 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8177 For the RSA certificate situation is makes no difference, but
8178 for the DSA certificate situation this fixes the "no shared cipher"
8179 problem where the OpenSSL cipher selection procedure failed because the
8180 temporary keys were not overtaken from the context and the API provided
8181 no way to reconfigure them.
8182 The new functions now let applications reconfigure the stuff and they
8183 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8184 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8185 non-public-API function ssl_cert_instantiate() is used as a helper
8186 function and also to reduce code redundancy inside ssl_rsa.c.
8187 [Ralf S. Engelschall]
8189 *) Move s_server -dcert and -dkey options out of the undocumented feature
8190 area because they are useful for the DSA situation and should be
8191 recognized by the users.
8192 [Ralf S. Engelschall]
8194 *) Fix the cipher decision scheme for export ciphers: the export bits are
8195 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8196 SSL_EXP_MASK. So, the original variable has to be used instead of the
8197 already masked variable.
8198 [Richard Levitte <levitte@stacken.kth.se>]
8200 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8201 [Richard Levitte <levitte@stacken.kth.se>]
8203 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8204 from `int' to `unsigned int' because it's a length and initialized by
8205 EVP_DigestFinal() which expects an `unsigned int *'.
8206 [Richard Levitte <levitte@stacken.kth.se>]
8208 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8209 script. Instead use the usual Shell->Perl transition trick.
8210 [Ralf S. Engelschall]
8212 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8213 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8214 -noout -modulus' as it's already the case for `openssl rsa -noout
8215 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8216 currently the public key is printed (a decision which was already done by
8217 `openssl dsa -modulus' in the past) which serves a similar purpose.
8218 Additionally the NO_RSA no longer completely removes the whole -modulus
8219 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8221 [Ralf S. Engelschall]
8223 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8224 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8225 [Arne Ansper <arne@ats.cyber.ee>]
8227 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8228 to be added. Now both 'req' and 'ca' can use new objects defined in the
8232 *) Add cool BIO that does syslog (or event log on NT).
8233 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8235 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8236 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8237 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8238 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8241 *) Add preliminary config info for new extension code.
8244 *) Make RSA_NO_PADDING really use no padding.
8245 [Ulf Moeller <ulf@fitug.de>]
8247 *) Generate errors when private/public key check is done.
8250 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8251 for some CRL extensions and new objects added.
8254 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8255 key usage extension and fuller support for authority key id.
8258 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8259 padding method for RSA, which is recommended for new applications in PKCS
8260 #1 v2.0 (RFC 2437, October 1998).
8261 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8262 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8263 against Bleichbacher's attack on RSA.
8264 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8267 *) Updates to the new SSL compression code
8268 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8270 *) Fix so that the version number in the master secret, when passed
8271 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8272 (because the server will not accept higher), that the version number
8273 is 0x03,0x01, not 0x03,0x00
8274 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8276 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8277 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8278 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8281 *) Support for RAW extensions where an arbitrary extension can be
8282 created by including its DER encoding. See apps/openssl.cnf for
8286 *) Make sure latest Perl versions don't interpret some generated C array
8287 code as Perl array code in the crypto/err/err_genc.pl script.
8288 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8290 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8291 not many people have the assembler. Various Win32 compilation fixes and
8292 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8296 *) Modify configure script 'Configure' to automatically create crypto/date.h
8297 file under Win32 and also build pem.h from pem.org. New script
8298 util/mkfiles.pl to create the MINFO file on environments that can't do a
8299 'make files': perl util/mkfiles.pl >MINFO should work.
8302 *) Major rework of DES function declarations, in the pursuit of correctness
8303 and purity. As a result, many evil casts evaporated, and some weirdness,
8304 too. You may find this causes warnings in your code. Zapping your evil
8305 casts will probably fix them. Mostly.
8308 *) Fix for a typo in asn1.h. Bug fix to object creation script
8309 obj_dat.pl. It considered a zero in an object definition to mean
8310 "end of object": none of the objects in objects.h have any zeros
8311 so it wasn't spotted.
8312 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8314 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8315 Masking (CBCM). In the absence of test vectors, the best I have been able
8316 to do is check that the decrypt undoes the encrypt, so far. Send me test
8317 vectors if you have them.
8320 *) Correct calculation of key length for export ciphers (too much space was
8321 allocated for null ciphers). This has not been tested!
8324 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8325 message is now correct (it understands "crypto" and "ssl" on its
8326 command line). There is also now an "update" option. This will update
8327 the util/ssleay.num and util/libeay.num files with any new functions.
8329 perl util/mkdef.pl crypto ssl update
8330 it will update them.
8333 *) Overhauled the Perl interface (perl/*):
8334 - ported BN stuff to OpenSSL's different BN library
8335 - made the perl/ source tree CVS-aware
8336 - renamed the package from SSLeay to OpenSSL (the files still contain
8337 their history because I've copied them in the repository)
8338 - removed obsolete files (the test scripts will be replaced
8339 by better Test::Harness variants in the future)
8340 [Ralf S. Engelschall]
8342 *) First cut for a very conservative source tree cleanup:
8343 1. merge various obsolete readme texts into doc/ssleay.txt
8344 where we collect the old documents and readme texts.
8345 2. remove the first part of files where I'm already sure that we no
8346 longer need them because of three reasons: either they are just temporary
8347 files which were left by Eric or they are preserved original files where
8348 I've verified that the diff is also available in the CVS via "cvs diff
8349 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8350 the crypto/md/ stuff).
8351 [Ralf S. Engelschall]
8353 *) More extension code. Incomplete support for subject and issuer alt
8354 name, issuer and authority key id. Change the i2v function parameters
8355 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8356 what that's for :-) Fix to ASN1 macro which messed up
8357 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8360 *) Preliminary support for ENUMERATED type. This is largely copied from the
8364 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8365 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8367 *) Make sure `make rehash' target really finds the `openssl' program.
8368 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8370 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8371 like to hear about it if this slows down other processors.
8374 *) Add CygWin32 platform information to Configure script.
8375 [Alan Batie <batie@aahz.jf.intel.com>]
8377 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8378 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8380 *) New program nseq to manipulate netscape certificate sequences
8383 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8387 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8388 but the BN code had some problems that would cause failures when
8389 doing certificate verification and some other functions.
8390 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8392 *) Add ASN1 and PEM code to support netscape certificate sequences.
8395 *) Add ASN1 and PEM code to support netscape certificate sequences.
8398 *) Add several PKIX and private extended key usage OIDs.
8401 *) Modify the 'ca' program to handle the new extension code. Modify
8402 openssl.cnf for new extension format, add comments.
8405 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8406 and add a sample to openssl.cnf so req -x509 now adds appropriate
8410 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8411 error code, add initial support to X509_print() and x509 application.
8414 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8415 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8416 stuff is currently isolated and isn't even compiled yet.
8419 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8420 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8421 Removed the versions check from X509 routines when loading extensions:
8422 this allows certain broken certificates that don't set the version
8423 properly to be processed.
8426 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8427 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8428 can still be regenerated with "make depend".
8431 *) Spelling mistake in C version of CAST-128.
8432 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8434 *) Changes to the error generation code. The perl script err-code.pl
8435 now reads in the old error codes and retains the old numbers, only
8436 adding new ones if necessary. It also only changes the .err files if new
8437 codes are added. The makefiles have been modified to only insert errors
8438 when needed (to avoid needlessly modifying header files). This is done
8439 by only inserting errors if the .err file is newer than the auto generated
8440 C file. To rebuild all the error codes from scratch (the old behaviour)
8441 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8442 or delete all the .err files.
8445 *) CAST-128 was incorrectly implemented for short keys. The C version has
8446 been fixed, but is untested. The assembler versions are also fixed, but
8447 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8448 to regenerate it if needed.
8449 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8450 Hagino <itojun@kame.net>]
8452 *) File was opened incorrectly in randfile.c.
8453 [Ulf Möller <ulf@fitug.de>]
8455 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8456 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8457 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8458 al: it's just almost always a UTCTime. Note this patch adds new error
8459 codes so do a "make errors" if there are problems.
8462 *) Correct Linux 1 recognition in config.
8463 [Ulf Möller <ulf@fitug.de>]
8465 *) Remove pointless MD5 hash when using DSA keys in ca.
8466 [Anonymous <nobody@replay.com>]
8468 *) Generate an error if given an empty string as a cert directory. Also
8469 generate an error if handed NULL (previously returned 0 to indicate an
8470 error, but didn't set one).
8471 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8473 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8476 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8477 parameters. This was causing a warning which killed off the Win32 compile.
8480 *) Remove C++ style comments from crypto/bn/bn_local.h.
8481 [Neil Costigan <neil.costigan@celocom.com>]
8483 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8484 based on a text string, looking up short and long names and finally
8485 "dot" format. The "dot" format stuff didn't work. Added new function
8486 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8487 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8488 OID is not part of the table.
8491 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8492 X509_LOOKUP_by_alias().
8495 *) Sort openssl functions by name.
8498 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8499 encryption from sample DSA keys (in case anyone is interested the password
8503 *) Make _all_ *_free functions accept a NULL pointer.
8504 [Frans Heymans <fheymans@isaserver.be>]
8506 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8508 [Anonymous <nobody@replay.com>]
8510 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8511 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8513 *) Don't blow it for numeric -newkey arguments to apps/req.
8514 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8516 *) Temp key "for export" tests were wrong in s3_srvr.c.
8517 [Anonymous <nobody@replay.com>]
8519 *) Add prototype for temp key callback functions
8520 SSL_CTX_set_tmp_{rsa,dh}_callback().
8523 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8524 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8527 *) X509_name_add_entry() freed the wrong thing after an error.
8528 [Arne Ansper <arne@ats.cyber.ee>]
8530 *) rsa_eay.c would attempt to free a NULL context.
8531 [Arne Ansper <arne@ats.cyber.ee>]
8533 *) BIO_s_socket() had a broken should_retry() on Windoze.
8534 [Arne Ansper <arne@ats.cyber.ee>]
8536 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8537 [Arne Ansper <arne@ats.cyber.ee>]
8539 *) Make sure the already existing X509_STORE->depth variable is initialized
8540 in X509_STORE_new(), but document the fact that this variable is still
8541 unused in the certificate verification process.
8542 [Ralf S. Engelschall]
8544 *) Fix the various library and apps files to free up pkeys obtained from
8545 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8548 *) Fix reference counting in X509_PUBKEY_get(). This makes
8549 demos/maurice/example2.c work, amongst others, probably.
8550 [Steve Henson and Ben Laurie]
8552 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8553 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8554 are no longer created. This way we have a single and consistent command
8555 line interface `openssl <command>', similar to `cvs <command>'.
8556 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8558 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8559 BIT STRING wrapper always have zero unused bits.
8562 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8565 *) Make the top-level INSTALL documentation easier to understand.
8568 *) Makefiles updated to exit if an error occurs in a sub-directory
8569 make (including if user presses ^C) [Paul Sutton]
8571 *) Make Montgomery context stuff explicit in RSA data structure.
8574 *) Fix build order of pem and err to allow for generated pem.h.
8577 *) Fix renumbering bug in X509_NAME_delete_entry().
8580 *) Enhanced the err-ins.pl script so it makes the error library number
8581 global and can add a library name. This is needed for external ASN1 and
8582 other error libraries.
8585 *) Fixed sk_insert which never worked properly.
8588 *) Fix ASN1 macros so they can handle indefinite length construted
8589 EXPLICIT tags. Some non standard certificates use these: they can now
8593 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8594 into a single doc/ssleay.txt bundle. This way the information is still
8595 preserved but no longer messes up this directory. Now it's new room for
8596 the new set of documenation files.
8597 [Ralf S. Engelschall]
8599 *) SETs were incorrectly DER encoded. This was a major pain, because they
8600 shared code with SEQUENCEs, which aren't coded the same. This means that
8601 almost everything to do with SETs or SEQUENCEs has either changed name or
8602 number of arguments.
8603 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8605 *) Fix test data to work with the above.
8608 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8609 was already fixed by Eric for 0.9.1 it seems.
8610 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8612 *) Autodetect FreeBSD3.
8615 *) Fix various bugs in Configure. This affects the following platforms:
8619 unixware-2.0-pentium
8623 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8624 before they are needed.
8627 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8631 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8633 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8634 changed SSLeay to OpenSSL in version strings.
8635 [Ralf S. Engelschall]
8637 *) Some fixups to the top-level documents.
8640 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8641 because the symlink to include/ was missing.
8642 [Ralf S. Engelschall]
8644 *) Incorporated the popular no-RSA/DSA-only patches
8645 which allow to compile a RSA-free SSLeay.
8646 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8648 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8649 when "ssleay" is still not found.
8650 [Ralf S. Engelschall]
8652 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8653 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8655 *) Updated the README file.
8656 [Ralf S. Engelschall]
8658 *) Added various .cvsignore files in the CVS repository subdirs
8659 to make a "cvs update" really silent.
8660 [Ralf S. Engelschall]
8662 *) Recompiled the error-definition header files and added
8663 missing symbols to the Win32 linker tables.
8664 [Ralf S. Engelschall]
8666 *) Cleaned up the top-level documents;
8667 o new files: CHANGES and LICENSE
8668 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8669 o merged COPYRIGHT into LICENSE
8670 o removed obsolete TODO file
8671 o renamed MICROSOFT to INSTALL.W32
8672 [Ralf S. Engelschall]
8674 *) Removed dummy files from the 0.9.1b source tree:
8675 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8676 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8677 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8678 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8679 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8680 [Ralf S. Engelschall]
8682 *) Added various platform portability fixes.
8685 *) The Genesis of the OpenSSL rpject:
8686 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8687 Young and Tim J. Hudson created while they were working for C2Net until
8689 [The OpenSSL Project]
8692 Changes between 0.9.0b and 0.9.1b [not released]
8694 *) Updated a few CA certificates under certs/
8697 *) Changed some BIGNUM api stuff.
8700 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8701 DGUX x86, Linux Alpha, etc.
8704 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8705 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8709 *) Add -strparse option to asn1pars program which parses nested
8711 [Dr Stephen Henson <shenson@bigfoot.com>]
8713 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8716 *) DSA fix for "ca" program.
8719 *) Added "-genkey" option to "dsaparam" program.
8722 *) Added RIPE MD160 (rmd160) message digest.
8725 *) Added -a (all) option to "ssleay version" command.
8728 *) Added PLATFORM define which is the id given to Configure.
8731 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8734 *) Extended the ASN.1 parser routines.
8737 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8740 *) Added a BN_CTX to the BN library.
8743 *) Fixed the weak key values in DES library
8746 *) Changed API in EVP library for cipher aliases.
8749 *) Added support for RC2/64bit cipher.
8752 *) Converted the lhash library to the crypto/mem.c functions.
8755 *) Added more recognized ASN.1 object ids.
8758 *) Added more RSA padding checks for SSL/TLS.
8761 *) Added BIO proxy/filter functionality.
8764 *) Added extra_certs to SSL_CTX which can be used
8765 send extra CA certificates to the client in the CA cert chain sending
8766 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8769 *) Now Fortezza is denied in the authentication phase because
8770 this is key exchange mechanism is not supported by SSLeay at all.
8773 *) Additional PKCS1 checks.
8776 *) Support the string "TLSv1" for all TLS v1 ciphers.
8779 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8780 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8783 *) Fixed a few memory leaks.
8786 *) Fixed various code and comment typos.
8789 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8790 bytes sent in the client random.
8791 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob