5 Changes between 0.9.8zb and 0.9.8zc [15 Oct 2014]
7 *) Session Ticket Memory Leak.
9 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
10 integrity of that ticket is first verified. In the event of a session
11 ticket integrity check failing, OpenSSL will fail to free memory
12 causing a memory leak. By sending a large number of invalid session
13 tickets an attacker could exploit this issue in a Denial Of Service
18 *) Build option no-ssl3 is incomplete.
20 When OpenSSL is configured with "no-ssl3" as a build option, servers
21 could accept and complete a SSL 3.0 handshake, and clients could be
22 configured to send them.
24 [Akamai and the OpenSSL team]
26 *) Add support for TLS_FALLBACK_SCSV.
27 Client applications doing fallback retries should call
28 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
30 [Adam Langley, Bodo Moeller]
32 *) Add additional DigestInfo checks.
34 Reencode DigestInto in DER and check against the original when
35 verifying RSA signature: this will reject any improperly encoded
36 DigestInfo structures.
38 Note: this is a precautionary measure and no attacks are currently known.
42 Changes between 0.9.8za and 0.9.8zb [6 Aug 2014]
44 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
45 to a denial of service attack. A malicious server can crash the client
46 with a null pointer dereference (read) by specifying an anonymous (EC)DH
47 ciphersuite and sending carefully crafted handshake messages.
49 Thanks to Felix Gröbert (Google) for discovering and researching this
54 *) By sending carefully crafted DTLS packets an attacker could cause openssl
55 to leak memory. This can be exploited through a Denial of Service attack.
56 Thanks to Adam Langley for discovering and researching this issue.
60 *) An attacker can force openssl to consume large amounts of memory whilst
61 processing DTLS handshake messages. This can be exploited through a
62 Denial of Service attack.
63 Thanks to Adam Langley for discovering and researching this issue.
67 *) An attacker can force an error condition which causes openssl to crash
68 whilst processing DTLS packets due to memory being freed twice. This
69 can be exploited through a Denial of Service attack.
70 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
75 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
76 X509_name_oneline, X509_name_print_ex et al. to leak some information
77 from the stack. Applications may be affected if they echo pretty printing
78 output to the attacker.
80 Thanks to Ivan Fratric (Google) for discovering this issue.
82 [Emilia Käsper, and Steve Henson]
84 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
85 for corner cases. (Certain input points at infinity could lead to
86 bogus results, with non-infinity inputs mapped to infinity too.)
89 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
91 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
92 handshake can force the use of weak keying material in OpenSSL
93 SSL/TLS clients and servers.
95 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
96 researching this issue. (CVE-2014-0224)
97 [KIKUCHI Masashi, Steve Henson]
99 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
100 OpenSSL DTLS client the code can be made to recurse eventually crashing
103 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
105 [Imre Rad, Steve Henson]
107 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
108 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
109 client or server. This is potentially exploitable to run arbitrary
110 code on a vulnerable client or server.
112 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
113 [Jüri Aedla, Steve Henson]
115 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
116 are subject to a denial of service attack.
118 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
119 this issue. (CVE-2014-3470)
120 [Felix Gröbert, Ivan Fratric, Steve Henson]
122 *) Fix for the attack described in the paper "Recovering OpenSSL
123 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
124 by Yuval Yarom and Naomi Benger. Details can be obtained from:
125 http://eprint.iacr.org/2014/140
127 Thanks to Yuval Yarom and Naomi Benger for discovering this
128 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
129 [Yuval Yarom and Naomi Benger]
131 Thanks to mancha for backporting the fix to the 0.9.8 branch.
133 *) Fix handling of warning-level alerts in SSL23 client mode so they
134 don't cause client-side termination (eg. on SNI unrecognized_name
135 warnings). Add client and server support for six additional alerts
136 per RFC 6066 and RFC 4279.
139 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
140 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
141 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
142 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
143 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
144 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
145 [Rob Stradling, Adam Langley]
147 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
149 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
151 This addresses the flaw in CBC record processing discovered by
152 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
153 at: http://www.isg.rhul.ac.uk/tls/
155 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
156 Security Group at Royal Holloway, University of London
157 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
158 Emilia Käsper for the initial patch.
160 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
162 *) Return an error when checking OCSP signatures when key is NULL.
163 This fixes a DoS attack. (CVE-2013-0166)
166 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
167 the right response is stapled. Also change SSL_get_certificate()
168 so it returns the certificate actually sent.
169 See http://rt.openssl.org/Ticket/Display.html?id=2836.
171 [Rob Stradling <rob.stradling@comodo.com>]
173 *) Fix possible deadlock when decoding public keys.
176 Changes between 0.9.8w and 0.9.8x [10 May 2012]
178 *) Sanity check record length before skipping explicit IV in DTLS
181 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
182 fuzzing as a service testing platform.
186 *) Initialise tkeylen properly when encrypting CMS messages.
187 Thanks to Solar Designer of Openwall for reporting this issue.
190 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
192 *) The fix for CVE-2012-2110 did not take into account that the
193 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
194 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
195 rejecting negative len parameter. (CVE-2012-2131)
196 [Tomas Hoger <thoger@redhat.com>]
198 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
200 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
201 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
202 in CRYPTO_realloc_clean.
204 Thanks to Tavis Ormandy, Google Security Team, for discovering this
205 issue and to Adam Langley <agl@chromium.org> for fixing it.
207 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
209 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
211 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
212 in CMS and PKCS7 code. When RSA decryption fails use a random key for
213 content decryption and always return the same error. Note: this attack
214 needs on average 2^20 messages so it only affects automated senders. The
215 old behaviour can be reenabled in the CMS code by setting the
216 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
217 an MMA defence is not necessary.
218 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
219 this issue. (CVE-2012-0884)
222 *) Fix CVE-2011-4619: make sure we really are receiving a
223 client hello before rejecting multiple SGC restarts. Thanks to
224 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
227 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
229 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
230 Thanks to Antonio Martin, Enterprise Secure Access Research and
231 Development, Cisco Systems, Inc. for discovering this bug and
232 preparing a fix. (CVE-2012-0050)
235 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
237 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
238 of the Vaudenay padding oracle attack on CBC mode encryption
239 which enables an efficient plaintext recovery attack against
240 the OpenSSL implementation of DTLS. Their attack exploits timing
241 differences arising during decryption processing. A research
242 paper describing this attack can be found at:
243 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
244 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
245 Security Group at Royal Holloway, University of London
246 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
247 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
248 for preparing the fix. (CVE-2011-4108)
249 [Robin Seggelmann, Michael Tuexen]
251 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
252 [Ben Laurie, Kasper <ekasper@google.com>]
254 *) Clear bytes used for block padding of SSL 3.0 records.
256 [Adam Langley (Google)]
258 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
259 Kadianakis <desnacked@gmail.com> for discovering this issue and
260 Adam Langley for preparing the fix. (CVE-2011-4619)
261 [Adam Langley (Google)]
263 *) Prevent malformed RFC3779 data triggering an assertion failure.
264 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
265 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
266 [Rob Austein <sra@hactrn.net>]
268 *) Fix ssl_ciph.c set-up race.
269 [Adam Langley (Google)]
271 *) Fix spurious failures in ecdsatest.c.
272 [Emilia Käsper (Google)]
274 *) Fix the BIO_f_buffer() implementation (which was mixing different
275 interpretations of the '..._len' fields).
276 [Adam Langley (Google)]
278 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
279 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
280 threads won't reuse the same blinding coefficients.
282 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
283 lock to call BN_BLINDING_invert_ex, and avoids one use of
284 BN_BLINDING_update for each BN_BLINDING structure (previously,
285 the last update always remained unused).
286 [Emilia Käsper (Google)]
288 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
289 for multi-threaded use of ECDH.
290 [Adam Langley (Google)]
292 *) Fix x509_name_ex_d2i memory leak on bad inputs.
295 *) Add protection against ECDSA timing attacks as mentioned in the paper
296 by Billy Bob Brumley and Nicola Tuveri, see:
298 http://eprint.iacr.org/2011/232.pdf
300 [Billy Bob Brumley and Nicola Tuveri]
302 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
304 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
305 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
307 *) Fix bug in string printing code: if *any* escaping is enabled we must
308 escape the escape character (backslash) or the resulting string is
312 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
314 *) Disable code workaround for ancient and obsolete Netscape browsers
315 and servers: an attacker can use it in a ciphersuite downgrade attack.
316 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
319 *) Fixed J-PAKE implementation error, originally discovered by
320 Sebastien Martini, further info and confirmation from Stefan
321 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
324 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
326 *) Fix extension code to avoid race conditions which can result in a buffer
327 overrun vulnerability: resumed sessions must not be modified as they can
328 be shared by multiple threads. CVE-2010-3864
331 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
334 *) Don't reencode certificate when calculating signature: cache and use
335 the original encoding instead. This makes signature verification of
336 some broken encodings work correctly.
339 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
340 is also one of the inputs.
341 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
343 *) Don't repeatedly append PBE algorithms to table if they already exist.
344 Sort table on each new add. This effectively makes the table read only
345 after all algorithms are added and subsequent calls to PKCS12_pbe_add
349 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
351 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
354 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
355 access or freeing data twice (CVE-2010-0742)
356 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
358 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
359 common in certificates and some applications which only call
360 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
364 Reduce copying into .apps and .test in makevms.com
365 Don't try to use blank CA certificate in CA.com
366 Allow use of C files from original directories in maketests.com
367 [Steven M. Schweda" <sms@antinode.info>]
369 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
371 *) When rejecting SSL/TLS records due to an incorrect version number, never
372 update s->server with a new major version number. As of
373 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
374 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
375 the previous behavior could result in a read attempt at NULL when
376 receiving specific incorrect SSL/TLS records once record payload
377 protection is active. (CVE-2010-0740)
378 [Bodo Moeller, Adam Langley <agl@chromium.org>]
380 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
381 could be crashed if the relevant tables were not present (e.g. chrooted).
382 [Tomas Hoger <thoger@redhat.com>]
384 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
386 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
387 [Martin Olsson, Neel Mehta]
389 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
390 accommodate for stack sorting, always a write lock!).
393 *) On some versions of WIN32 Heap32Next is very slow. This can cause
394 excessive delays in the RAND_poll(): over a minute. As a workaround
395 include a time check in the inner Heap32Next loop too.
398 *) The code that handled flushing of data in SSL/TLS originally used the
399 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
400 the problem outlined in PR#1949. The fix suggested there however can
401 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
402 of Apache). So instead simplify the code to flush unconditionally.
403 This should be fine since flushing with no data to flush is a no op.
406 *) Handle TLS versions 2.0 and later properly and correctly use the
407 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
408 off ancient servers have a habit of sticking around for a while...
411 *) Modify compression code so it frees up structures without using the
412 ex_data callbacks. This works around a problem where some applications
413 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
414 restarting) then use compression (e.g. SSL with compression) later.
415 This results in significant per-connection memory leaks and
416 has caused some security issues including CVE-2008-1678 and
420 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
421 change when encrypting or decrypting.
424 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
425 connect and renegotiate with servers which do not support RI.
426 Until RI is more widely deployed this option is enabled by default.
429 *) Add "missing" ssl ctrls to clear options and mode.
432 *) If client attempts to renegotiate and doesn't support RI respond with
433 a no_renegotiation alert as required by RFC5746. Some renegotiating
434 TLS clients will continue a connection gracefully when they receive
435 the alert. Unfortunately OpenSSL mishandled this alert and would hang
436 waiting for a server hello which it will never receive. Now we treat a
437 received no_renegotiation alert as a fatal error. This is because
438 applications requesting a renegotiation might well expect it to succeed
439 and would have no code in place to handle the server denying it so the
440 only safe thing to do is to terminate the connection.
443 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
444 peer supports secure renegotiation and 0 otherwise. Print out peer
445 renegotiation support in s_client/s_server.
448 *) Replace the highly broken and deprecated SPKAC certification method with
449 the updated NID creation version. This should correctly handle UTF8.
452 *) Implement RFC5746. Re-enable renegotiation but require the extension
453 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
454 turns out to be a bad idea. It has been replaced by
455 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
456 SSL_CTX_set_options(). This is really not recommended unless you
457 know what you are doing.
458 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
460 *) Fixes to stateless session resumption handling. Use initial_ctx when
461 issuing and attempting to decrypt tickets in case it has changed during
462 servername handling. Use a non-zero length session ID when attempting
463 stateless session resumption: this makes it possible to determine if
464 a resumption has occurred immediately after receiving server hello
465 (several places in OpenSSL subtly assume this) instead of later in
469 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
470 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
471 fixes for a few places where the return code is not checked
473 [Julia Lawall <julia@diku.dk>]
475 *) Add --strict-warnings option to Configure script to include devteam
476 warnings in other configurations.
479 *) Add support for --libdir option and LIBDIR variable in makefiles. This
480 makes it possible to install openssl libraries in locations which
481 have names other than "lib", for example "/usr/lib64" which some
483 [Steve Henson, based on patch from Jeremy Utley]
485 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
486 X690 8.9.12 and can produce some misleading textual output of OIDs.
487 [Steve Henson, reported by Dan Kaminsky]
489 *) Delete MD2 from algorithm tables. This follows the recommendation in
490 several standards that it is not used in new applications due to
491 several cryptographic weaknesses. For binary compatibility reasons
492 the MD2 API is still compiled in by default.
495 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
499 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
500 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
502 [Guenter <lists@gknw.net>]
504 *) Fix the server certificate chain building code to use X509_verify_cert(),
505 it used to have an ad-hoc builder which was unable to cope with anything
506 other than a simple chain.
507 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
509 *) Don't check self signed certificate signatures in X509_verify_cert()
510 by default (a flag can override this): it just wastes time without
511 adding any security. As a useful side effect self signed root CAs
512 with non-FIPS digests are now usable in FIPS mode.
515 *) In dtls1_process_out_of_seq_message() the check if the current message
516 is already buffered was missing. For every new message was memory
517 allocated, allowing an attacker to perform an denial of service attack
518 with sending out of seq handshake messages until there is no memory
519 left. Additionally every future messege was buffered, even if the
520 sequence number made no sense and would be part of another handshake.
521 So only messages with sequence numbers less than 10 in advance will be
522 buffered. (CVE-2009-1378)
523 [Robin Seggelmann, discovered by Daniel Mentz]
525 *) Records are buffered if they arrive with a future epoch to be
526 processed after finishing the corresponding handshake. There is
527 currently no limitation to this buffer allowing an attacker to perform
528 a DOS attack with sending records with future epochs until there is no
529 memory left. This patch adds the pqueue_size() function to detemine
530 the size of a buffer and limits the record buffer to 100 entries.
532 [Robin Seggelmann, discovered by Daniel Mentz]
534 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
535 parent structure is freed. (CVE-2009-1379)
538 *) Handle non-blocking I/O properly in SSL_shutdown() call.
539 [Darryl Miles <darryl-mailinglists@netbauds.net>]
542 [Ilya O. <vrghost@gmail.com>]
544 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
546 *) Disable renegotiation completely - this fixes a severe security
547 problem (CVE-2009-3555) at the cost of breaking all
548 renegotiation. Renegotiation can be re-enabled by setting
549 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
550 run-time. This is really not recommended unless you know what
554 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
556 *) Don't set val to NULL when freeing up structures, it is freed up by
557 underlying code. If sizeof(void *) > sizeof(long) this can result in
558 zeroing past the valid field. (CVE-2009-0789)
559 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
561 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
562 checked correctly. This would allow some invalid signed attributes to
563 appear to verify correctly. (CVE-2009-0591)
564 [Ivan Nestlerode <inestlerode@us.ibm.com>]
566 *) Reject UniversalString and BMPString types with invalid lengths. This
567 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
568 a legal length. (CVE-2009-0590)
571 *) Set S/MIME signing as the default purpose rather than setting it
572 unconditionally. This allows applications to override it at the store
576 *) Permit restricted recursion of ASN1 strings. This is needed in practice
577 to handle some structures.
580 *) Improve efficiency of mem_gets: don't search whole buffer each time
582 [Jeremy Shapiro <jnshapir@us.ibm.com>]
584 *) New -hex option for openssl rand.
587 *) Print out UTF8String and NumericString when parsing ASN1.
590 *) Support NumericString type for name components.
593 *) Allow CC in the environment to override the automatically chosen
594 compiler. Note that nothing is done to ensure flags work with the
598 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
600 *) Properly check EVP_VerifyFinal() and similar return values
602 [Ben Laurie, Bodo Moeller, Google Security Team]
604 *) Enable TLS extensions by default.
607 *) Allow the CHIL engine to be loaded, whether the application is
608 multithreaded or not. (This does not release the developer from the
609 obligation to set up the dynamic locking callbacks.)
610 [Sander Temme <sander@temme.net>]
612 *) Use correct exit code if there is an error in dgst command.
613 [Steve Henson; problem pointed out by Roland Dirlewanger]
615 *) Tweak Configure so that you need to say "experimental-jpake" to enable
616 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
619 *) Add experimental JPAKE support, including demo authentication in
620 s_client and s_server.
623 *) Set the comparison function in v3_addr_canonize().
624 [Rob Austein <sra@hactrn.net>]
626 *) Add support for XMPP STARTTLS in s_client.
627 [Philip Paeps <philip@freebsd.org>]
629 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
630 to ensure that even with this option, only ciphersuites in the
631 server's preference list will be accepted. (Note that the option
632 applies only when resuming a session, so the earlier behavior was
633 just about the algorithm choice for symmetric cryptography.)
636 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
638 *) Fix NULL pointer dereference if a DTLS server received
639 ChangeCipherSpec as first record (CVE-2009-1386).
642 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
643 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
646 *) The fix in 0.9.8c that supposedly got rid of unsafe
647 double-checked locking was incomplete for RSA blinding,
648 addressing just one layer of what turns out to have been
649 doubly unsafe triple-checked locking.
651 So now fix this for real by retiring the MONT_HELPER macro
652 in crypto/rsa/rsa_eay.c.
654 [Bodo Moeller; problem pointed out by Marius Schilder]
656 *) Various precautionary measures:
658 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
660 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
661 (NB: This would require knowledge of the secret session ticket key
662 to exploit, in which case you'd be SOL either way.)
664 - Change bn_nist.c so that it will properly handle input BIGNUMs
665 outside the expected range.
667 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
670 [Neel Mehta, Bodo Moeller]
672 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
673 the load fails. Useful for distros.
674 [Ben Laurie and the FreeBSD team]
676 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
679 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
682 *) Expand ENGINE to support engine supplied SSL client certificate functions.
684 This work was sponsored by Logica.
687 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
688 keystores. Support for SSL/TLS client authentication too.
689 Not compiled unless enable-capieng specified to Configure.
691 This work was sponsored by Logica.
694 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
695 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
696 attribute creation routines such as certifcate requests and PKCS#12
700 Changes between 0.9.8g and 0.9.8h [28 May 2008]
702 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
703 handshake which could lead to a cilent crash as found using the
704 Codenomicon TLS test suite (CVE-2008-1672)
705 [Steve Henson, Mark Cox]
707 *) Fix double free in TLS server name extensions which could lead to
708 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
711 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
713 Clear the error queue to ensure that error entries left from
714 older function calls do not interfere with the correct operation.
715 [Lutz Jaenicke, Erik de Castro Lopo]
717 *) Remove root CA certificates of commercial CAs:
719 The OpenSSL project does not recommend any specific CA and does not
720 have any policy with respect to including or excluding any CA.
721 Therefore it does not make any sense to ship an arbitrary selection
722 of root CA certificates with the OpenSSL software.
725 *) RSA OAEP patches to fix two separate invalid memory reads.
726 The first one involves inputs when 'lzero' is greater than
727 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
728 before the beginning of from). The second one involves inputs where
729 the 'db' section contains nothing but zeroes (there is a one-byte
730 invalid read after the end of 'db').
731 [Ivan Nestlerode <inestlerode@us.ibm.com>]
733 *) Partial backport from 0.9.9-dev:
735 Introduce bn_mul_mont (dedicated Montgomery multiplication
736 procedure) as a candidate for BIGNUM assembler implementation.
737 While 0.9.9-dev uses assembler for various architectures, only
738 x86_64 is available by default here in the 0.9.8 branch, and
739 32-bit x86 is available through a compile-time setting.
741 To try the 32-bit x86 assembler implementation, use Configure
742 option "enable-montasm" (which exists only for this backport).
744 As "enable-montasm" for 32-bit x86 disclaims code stability
745 anyway, in this constellation we activate additional code
746 backported from 0.9.9-dev for further performance improvements,
747 namely BN_from_montgomery_word. (To enable this otherwise,
748 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
750 [Andy Polyakov (backport partially by Bodo Moeller)]
752 *) Add TLS session ticket callback. This allows an application to set
753 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
754 values. This is useful for key rollover for example where several key
755 sets may exist with different names.
758 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
759 This was broken until now in 0.9.8 releases, such that the only way
760 a registered ENGINE could be used (assuming it initialises
761 successfully on the host) was to explicitly set it as the default
762 for the relevant algorithms. This is in contradiction with 0.9.7
763 behaviour and the documentation. With this fix, when an ENGINE is
764 registered into a given algorithm's table of implementations, the
765 'uptodate' flag is reset so that auto-discovery will be used next
766 time a new context for that algorithm attempts to select an
768 [Ian Lister (tweaked by Geoff Thorpe)]
770 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
771 implemention in the following ways:
773 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
776 Lack of BER streaming support means one pass streaming processing is
777 only supported if data is detached: setting the streaming flag is
778 ignored for embedded content.
780 CMS support is disabled by default and must be explicitly enabled
781 with the enable-cms configuration option.
784 *) Update the GMP engine glue to do direct copies between BIGNUM and
785 mpz_t when openssl and GMP use the same limb size. Otherwise the
786 existing "conversion via a text string export" trick is still used.
787 [Paul Sheer <paulsheer@gmail.com>]
789 *) Zlib compression BIO. This is a filter BIO which compressed and
790 uncompresses any data passed through it.
793 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
794 RFC3394 compatible AES key wrapping.
797 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
798 sets string data without copying. X509_ALGOR_set0() and
799 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
800 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
801 from an X509_ATTRIBUTE structure optionally checking it occurs only
802 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
806 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
807 to get the expected BN_FLG_CONSTTIME behavior.
808 [Bodo Moeller (Google)]
812 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
813 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
814 - added some more tests to do_tests.pl
815 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
816 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
817 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
818 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
819 - various changes to netware.pl to enable gcc-cross builds on Win32
821 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
822 - various changes to fix missing prototype warnings
823 - fixed x86nasm.pl to create correct asm files for NASM COFF output
824 - added AES, WHIRLPOOL and CPUID assembler code to build files
825 - added missing AES assembler make rules to mk1mf.pl
826 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
827 [Guenter Knauf <eflash@gmx.net>]
829 *) Implement certificate status request TLS extension defined in RFC3546.
830 A client can set the appropriate parameters and receive the encoded
831 OCSP response via a callback. A server can query the supplied parameters
832 and set the encoded OCSP response in the callback. Add simplified examples
833 to s_client and s_server.
836 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
839 + Binary incompatibility of ssl_ctx_st structure
840 + DTLS interoperation with non-compliant servers
841 + Don't call get_session_cb() without proposed session
842 + Fix ia64 assembler code
843 [Andy Polyakov, Steve Henson]
845 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
847 *) DTLS Handshake overhaul. There were longstanding issues with
848 OpenSSL DTLS implementation, which were making it impossible for
849 RFC 4347 compliant client to communicate with OpenSSL server.
850 Unfortunately just fixing these incompatibilities would "cut off"
851 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
852 server keeps tolerating non RFC compliant syntax. The opposite is
853 not true, 0.9.8f client can not communicate with earlier server.
854 This update even addresses CVE-2007-4995.
857 *) Changes to avoid need for function casts in OpenSSL: some compilers
858 (gcc 4.2 and later) reject their use.
859 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
862 *) Add RFC4507 support to OpenSSL. This includes the corrections in
863 RFC4507bis. The encrypted ticket format is an encrypted encoded
864 SSL_SESSION structure, that way new session features are automatically
867 If a client application caches session in an SSL_SESSION structure
868 support is transparent because tickets are now stored in the encoded
871 The SSL_CTX structure automatically generates keys for ticket
872 protection in servers so again support should be possible
873 with no application modification.
875 If a client or server wishes to disable RFC4507 support then the option
876 SSL_OP_NO_TICKET can be set.
878 Add a TLS extension debugging callback to allow the contents of any client
879 or server extensions to be examined.
881 This work was sponsored by Google.
884 *) Add initial support for TLS extensions, specifically for the server_name
885 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
886 have new members for a host name. The SSL data structure has an
887 additional member SSL_CTX *initial_ctx so that new sessions can be
888 stored in that context to allow for session resumption, even after the
889 SSL has been switched to a new SSL_CTX in reaction to a client's
890 server_name extension.
892 New functions (subject to change):
895 SSL_get_servername_type()
898 New CTRL codes and macros (subject to change):
900 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
901 - SSL_CTX_set_tlsext_servername_callback()
902 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
903 - SSL_CTX_set_tlsext_servername_arg()
904 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
906 openssl s_client has a new '-servername ...' option.
908 openssl s_server has new options '-servername_host ...', '-cert2 ...',
909 '-key2 ...', '-servername_fatal' (subject to change). This allows
910 testing the HostName extension for a specific single host name ('-cert'
911 and '-key' remain fallbacks for handshakes without HostName
912 negotiation). If the unrecogninzed_name alert has to be sent, this by
913 default is a warning; it becomes fatal with the '-servername_fatal'
916 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
918 *) Add AES and SSE2 assembly language support to VC++ build.
921 *) Mitigate attack on final subtraction in Montgomery reduction.
924 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
925 (which previously caused an internal error).
928 *) Squeeze another 10% out of IGE mode when in != out.
931 *) AES IGE mode speedup.
932 [Dean Gaudet (Google)]
934 *) Add the Korean symmetric 128-bit cipher SEED (see
935 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
936 add SEED ciphersuites from RFC 4162:
938 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
939 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
940 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
941 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
943 To minimize changes between patchlevels in the OpenSSL 0.9.8
944 series, SEED remains excluded from compilation unless OpenSSL
945 is configured with 'enable-seed'.
948 *) Mitigate branch prediction attacks, which can be practical if a
949 single processor is shared, allowing a spy process to extract
950 information. For detailed background information, see
951 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
952 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
953 and Necessary Software Countermeasures"). The core of the change
954 are new versions BN_div_no_branch() and
955 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
956 respectively, which are slower, but avoid the security-relevant
957 conditional branches. These are automatically called by BN_div()
958 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
959 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
960 remove a conditional branch.
962 BN_FLG_CONSTTIME is the new name for the previous
963 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
964 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
965 in the exponent causes BN_mod_exp_mont() to use the alternative
966 implementation in BN_mod_exp_mont_consttime().) The old name
967 remains as a deprecated alias.
969 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
970 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
971 constant-time implementations for more than just exponentiation.
972 Here too the old name is kept as a deprecated alias.
974 BN_BLINDING_new() will now use BN_dup() for the modulus so that
975 the BN_BLINDING structure gets an independent copy of the
976 modulus. This means that the previous "BIGNUM *m" argument to
977 BN_BLINDING_new() and to BN_BLINDING_create_param() now
978 essentially becomes "const BIGNUM *m", although we can't actually
979 change this in the header file before 0.9.9. It allows
980 RSA_setup_blinding() to use BN_with_flags() on the modulus to
981 enable BN_FLG_CONSTTIME.
983 [Matthew D Wood (Intel Corp)]
985 *) In the SSL/TLS server implementation, be strict about session ID
986 context matching (which matters if an application uses a single
987 external cache for different purposes). Previously,
988 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
989 set. This did ensure strict client verification, but meant that,
990 with applications using a single external cache for quite
991 different requirements, clients could circumvent ciphersuite
992 restrictions for a given session ID context by starting a session
993 in a different context.
996 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
997 a ciphersuite string such as "DEFAULT:RSA" cannot enable
998 authentication-only ciphersuites.
1001 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1002 not complete and could lead to a possible single byte overflow
1003 (CVE-2007-5135) [Ben Laurie]
1005 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1007 *) Since AES128 and AES256 (and similarly Camellia128 and
1008 Camellia256) share a single mask bit in the logic of
1009 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1010 kludge to work properly if AES128 is available and AES256 isn't
1011 (or if Camellia128 is available and Camellia256 isn't).
1014 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1015 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1016 When a point or a seed is encoded in a BIT STRING, we need to
1017 prevent the removal of trailing zero bits to get the proper DER
1018 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1019 of a NamedBitList, for which trailing 0 bits need to be removed.)
1022 *) Have SSL/TLS server implementation tolerate "mismatched" record
1023 protocol version while receiving ClientHello even if the
1024 ClientHello is fragmented. (The server can't insist on the
1025 particular protocol version it has chosen before the ServerHello
1026 message has informed the client about his choice.)
1029 *) Add RFC 3779 support.
1030 [Rob Austein for ARIN, Ben Laurie]
1032 *) Load error codes if they are not already present instead of using a
1033 static variable. This allows them to be cleanly unloaded and reloaded.
1034 Improve header file function name parsing.
1037 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1038 or CAPABILITY handshake as required by RFCs.
1041 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1043 *) Introduce limits to prevent malicious keys being able to
1044 cause a denial of service. (CVE-2006-2940)
1045 [Steve Henson, Bodo Moeller]
1047 *) Fix ASN.1 parsing of certain invalid structures that can result
1048 in a denial of service. (CVE-2006-2937) [Steve Henson]
1050 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1051 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1053 *) Fix SSL client code which could crash if connecting to a
1054 malicious SSLv2 server. (CVE-2006-4343)
1055 [Tavis Ormandy and Will Drewry, Google Security Team]
1057 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1058 match only those. Before that, "AES256-SHA" would be interpreted
1059 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1060 the same strength classification in 0.9.7h) as we currently only
1061 have a single AES bit in the ciphersuite description bitmap.
1062 That change, however, also applied to ciphersuite strings such as
1063 "RC4-MD5" that intentionally matched multiple ciphersuites --
1064 namely, SSL 2.0 ciphersuites in addition to the more common ones
1065 from SSL 3.0/TLS 1.0.
1067 So we change the selection algorithm again: Naming an explicit
1068 ciphersuite selects this one ciphersuite, and any other similar
1069 ciphersuite (same bitmap) from *other* protocol versions.
1070 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1071 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1073 Since SSL 2.0 does not have any ciphersuites for which the
1074 128/256 bit distinction would be relevant, this works for now.
1075 The proper fix will be to use different bits for AES128 and
1076 AES256, which would have avoided the problems from the beginning;
1077 however, bits are scarce, so we can only do this in a new release
1078 (not just a patchlevel) when we can change the SSL_CIPHER
1079 definition to split the single 'unsigned long mask' bitmap into
1080 multiple values to extend the available space.
1084 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1086 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1087 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1089 *) Add AES IGE and biIGE modes.
1092 *) Change the Unix randomness entropy gathering to use poll() when
1093 possible instead of select(), since the latter has some
1094 undesirable limitations.
1095 [Darryl Miles via Richard Levitte and Bodo Moeller]
1097 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1098 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1099 cannot be implicitly activated as part of, e.g., the "AES" alias.
1100 However, please upgrade to OpenSSL 0.9.9[-dev] for
1101 non-experimental use of the ECC ciphersuites to get TLS extension
1102 support, which is required for curve and point format negotiation
1103 to avoid potential handshake problems.
1106 *) Disable rogue ciphersuites:
1108 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1109 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1110 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1112 The latter two were purportedly from
1113 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1116 Also deactivate the remaining ciphersuites from
1117 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1118 unofficial, and the ID has long expired.
1121 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1122 dual-core machines) and other potential thread-safety issues.
1125 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1126 versions), which is now available for royalty-free use
1127 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1128 Also, add Camellia TLS ciphersuites from RFC 4132.
1130 To minimize changes between patchlevels in the OpenSSL 0.9.8
1131 series, Camellia remains excluded from compilation unless OpenSSL
1132 is configured with 'enable-camellia'.
1135 *) Disable the padding bug check when compression is in use. The padding
1136 bug check assumes the first packet is of even length, this is not
1137 necessarily true if compresssion is enabled and can result in false
1138 positives causing handshake failure. The actual bug test is ancient
1139 code so it is hoped that implementations will either have fixed it by
1140 now or any which still have the bug do not support compression.
1143 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1145 *) When applying a cipher rule check to see if string match is an explicit
1146 cipher suite and only match that one cipher suite if it is.
1149 *) Link in manifests for VC++ if needed.
1150 [Austin Ziegler <halostatue@gmail.com>]
1152 *) Update support for ECC-based TLS ciphersuites according to
1153 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1154 TLS extensions, which are supported starting with the 0.9.9
1155 branch, not in the OpenSSL 0.9.8 branch).
1158 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1159 opaque EVP_CIPHER_CTX handling.
1162 *) Fixes and enhancements to zlib compression code. We now only use
1163 "zlib1.dll" and use the default __cdecl calling convention on Win32
1164 to conform with the standards mentioned here:
1165 http://www.zlib.net/DLL_FAQ.txt
1166 Static zlib linking now works on Windows and the new --with-zlib-include
1167 --with-zlib-lib options to Configure can be used to supply the location
1168 of the headers and library. Gracefully handle case where zlib library
1172 *) Several fixes and enhancements to the OID generation code. The old code
1173 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1174 handle numbers larger than ULONG_MAX, truncated printing and had a
1175 non standard OBJ_obj2txt() behaviour.
1178 *) Add support for building of engines under engine/ as shared libraries
1179 under VC++ build system.
1182 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1183 Hopefully, we will not see any false combination of paths any more.
1186 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1188 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1189 (part of SSL_OP_ALL). This option used to disable the
1190 countermeasure against man-in-the-middle protocol-version
1191 rollback in the SSL 2.0 server implementation, which is a bad
1192 idea. (CVE-2005-2969)
1194 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1195 for Information Security, National Institute of Advanced Industrial
1196 Science and Technology [AIST], Japan)]
1198 *) Add two function to clear and return the verify parameter flags.
1201 *) Keep cipherlists sorted in the source instead of sorting them at
1202 runtime, thus removing the need for a lock.
1205 *) Avoid some small subgroup attacks in Diffie-Hellman.
1206 [Nick Mathewson and Ben Laurie]
1208 *) Add functions for well-known primes.
1211 *) Extended Windows CE support.
1212 [Satoshi Nakamura and Andy Polyakov]
1214 *) Initialize SSL_METHOD structures at compile time instead of during
1215 runtime, thus removing the need for a lock.
1218 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1219 attempting to decrypt each encrypted key in turn. Add support to
1223 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1225 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1228 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1231 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1232 key into the same file any more.
1235 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1238 *) Add -utf8 command line and config file option to 'ca'.
1239 [Stefan <stf@udoma.org]
1241 *) Removed the macro des_crypt(), as it seems to conflict with some
1242 libraries. Use DES_crypt().
1245 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1246 involves renaming the source and generated shared-libs for
1247 both. The engines will accept the corrected or legacy ids
1248 ('ncipher' and '4758_cca' respectively) when binding. NB,
1249 this only applies when building 'shared'.
1250 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1252 *) Add attribute functions to EVP_PKEY structure. Modify
1253 PKCS12_create() to recognize a CSP name attribute and
1254 use it. Make -CSP option work again in pkcs12 utility.
1257 *) Add new functionality to the bn blinding code:
1258 - automatic re-creation of the BN_BLINDING parameters after
1259 a fixed number of uses (currently 32)
1260 - add new function for parameter creation
1261 - introduce flags to control the update behaviour of the
1262 BN_BLINDING parameters
1263 - hide BN_BLINDING structure
1264 Add a second BN_BLINDING slot to the RSA structure to improve
1265 performance when a single RSA object is shared among several
1269 *) Add support for DTLS.
1270 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1272 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1273 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1276 *) Remove buggy and incompletet DH cert support from
1277 ssl/ssl_rsa.c and ssl/s3_both.c
1280 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1281 the apps/openssl applications.
1284 *) Compile clean with "-Wall -Wmissing-prototypes
1285 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1286 DEBUG_SAFESTACK must also be set.
1289 *) Change ./Configure so that certain algorithms can be disabled by default.
1290 The new counterpiece to "no-xxx" is "enable-xxx".
1292 The patented RC5 and MDC2 algorithms will now be disabled unless
1293 "enable-rc5" and "enable-mdc2", respectively, are specified.
1295 (IDEA remains enabled despite being patented. This is because IDEA
1296 is frequently required for interoperability, and there is no license
1297 fee for non-commercial use. As before, "no-idea" can be used to
1298 avoid this algorithm.)
1302 *) Add processing of proxy certificates (see RFC 3820). This work was
1303 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1304 EGEE (Enabling Grids for E-science in Europe).
1307 *) RC4 performance overhaul on modern architectures/implementations, such
1308 as Intel P4, IA-64 and AMD64.
1311 *) New utility extract-section.pl. This can be used specify an alternative
1312 section number in a pod file instead of having to treat each file as
1313 a separate case in Makefile. This can be done by adding two lines to the
1316 =for comment openssl_section:XXX
1318 The blank line is mandatory.
1322 *) New arguments -certform, -keyform and -pass for s_client and s_server
1323 to allow alternative format key and certificate files and passphrase
1327 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1328 update associated structures and add various utility functions.
1330 Add new policy related verify parameters, include policy checking in
1331 standard verify code. Enhance 'smime' application with extra parameters
1332 to support policy checking and print out.
1335 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1336 Nehemiah processors. These extensions support AES encryption in hardware
1337 as well as RNG (though RNG support is currently disabled).
1338 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1340 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1343 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1344 [Andy Polyakov and a number of other people]
1346 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1347 implementation contributed by IBM.
1348 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1350 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1351 exponent rather than 'unsigned long'. There is a corresponding change to
1352 the new 'rsa_keygen' element of the RSA_METHOD structure.
1353 [Jelte Jansen, Geoff Thorpe]
1355 *) Functionality for creating the initial serial number file is now
1356 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1358 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1359 number file to 1, which is bound to cause problems. To avoid
1360 the problems while respecting compatibility between different 0.9.7
1361 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1362 CA.pl for serial number initialization. With the new release 0.9.8,
1363 we can fix the problem directly in the 'ca' utility.)
1366 *) Reduced header interdepencies by declaring more opaque objects in
1367 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1368 give fewer recursive includes, which could break lazy source code - so
1369 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1370 developers should define this symbol when building and using openssl to
1371 ensure they track the recommended behaviour, interfaces, [etc], but
1372 backwards-compatible behaviour prevails when this isn't defined.
1375 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1378 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1379 This will generate a random key of the appropriate length based on the
1380 cipher context. The EVP_CIPHER can provide its own random key generation
1381 routine to support keys of a specific form. This is used in the des and
1382 3des routines to generate a key of the correct parity. Update S/MIME
1383 code to use new functions and hence generate correct parity DES keys.
1384 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1385 valid (weak or incorrect parity).
1388 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1389 as looking them up. This is useful when the verified structure may contain
1390 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1391 present unless the new PKCS7_NO_CRL flag is asserted.
1394 *) Extend ASN1 oid configuration module. It now additionally accepts the
1397 shortName = some long name, 1.2.3.4
1400 *) Reimplemented the BN_CTX implementation. There is now no more static
1401 limitation on the number of variables it can handle nor the depth of the
1402 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1403 information can now expand as required, and rather than having a single
1404 static array of bignums, BN_CTX now uses a linked-list of such arrays
1405 allowing it to expand on demand whilst maintaining the usefulness of
1406 BN_CTX's "bundling".
1409 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1410 to allow all RSA operations to function using a single BN_CTX.
1413 *) Preliminary support for certificate policy evaluation and checking. This
1414 is initially intended to pass the tests outlined in "Conformance Testing
1415 of Relying Party Client Certificate Path Processing Logic" v1.07.
1418 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1419 remained unused and not that useful. A variety of other little bignum
1420 tweaks and fixes have also been made continuing on from the audit (see
1424 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1425 associated ASN1, EVP and SSL functions and old ASN1 macros.
1428 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1429 and this should never fail. So the return value from the use of
1430 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1431 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1434 *) BN_CTX_get() should return zero-valued bignums, providing the same
1435 initialised value as BN_new().
1436 [Geoff Thorpe, suggested by Ulf Möller]
1438 *) Support for inhibitAnyPolicy certificate extension.
1441 *) An audit of the BIGNUM code is underway, for which debugging code is
1442 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1443 is considered valid when processing BIGNUMs, and causes execution to
1444 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1445 further steps are taken to deliberately pollute unused data in BIGNUM
1446 structures to try and expose faulty code further on. For now, openssl will
1447 (in its default mode of operation) continue to tolerate the inconsistent
1448 forms that it has tolerated in the past, but authors and packagers should
1449 consider trying openssl and their own applications when compiled with
1450 these debugging symbols defined. It will help highlight potential bugs in
1451 their own code, and will improve the test coverage for OpenSSL itself. At
1452 some point, these tighter rules will become openssl's default to improve
1453 maintainability, though the assert()s and other overheads will remain only
1454 in debugging configurations. See bn.h for more details.
1455 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1457 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1458 that can only be obtained through BN_CTX_new() (which implicitly
1459 initialises it). The presence of this function only made it possible
1460 to overwrite an existing structure (and cause memory leaks).
1463 *) Because of the callback-based approach for implementing LHASH as a
1464 template type, lh_insert() adds opaque objects to hash-tables and
1465 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1466 to clean up those corresponding objects before destroying the hash table
1467 (and losing the object pointers). So some over-zealous constifications in
1468 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1469 objects as "const" and the lh_doall[_arg] callback wrappers are not
1470 prototyped to have "const" restrictions on the object pointers they are
1471 given (and so aren't required to cast them away any more).
1474 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1475 (speed) prefers to use its own implementation. The two implementations
1476 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1477 its object type properly exposed (MS_TM) instead of casting to/from "char
1478 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1479 aren't necessarily the greatest nomenclatures - but this is what was used
1480 internally to the implementation so I've used that for now.
1483 *) Ensure that deprecated functions do not get compiled when
1484 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1485 the self-tests were still using deprecated key-generation functions so
1486 these have been updated also.
1489 *) Reorganise PKCS#7 code to separate the digest location functionality
1490 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1491 New function PKCS7_set_digest() to set the digest type for PKCS#7
1492 digestedData type. Add additional code to correctly generate the
1493 digestedData type and add support for this type in PKCS7 initialization
1497 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1498 structure of type "other".
1501 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1502 sure the loop does correctly stop and breaking ("division by zero")
1503 modulus operations are not performed. The (pre-generated) prime
1504 table crypto/bn/bn_prime.h was already correct, but it could not be
1505 re-generated on some platforms because of the "division by zero"
1506 situation in the script.
1507 [Ralf S. Engelschall]
1509 *) Update support for ECC-based TLS ciphersuites according to
1510 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1511 SHA-1 now is only used for "small" curves (where the
1512 representation of a field element takes up to 24 bytes); for
1513 larger curves, the field element resulting from ECDH is directly
1514 used as premaster secret.
1515 [Douglas Stebila (Sun Microsystems Laboratories)]
1517 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1518 curve secp160r1 to the tests.
1519 [Douglas Stebila (Sun Microsystems Laboratories)]
1521 *) Add the possibility to load symbols globally with DSO.
1522 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1524 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1525 control of the error stack.
1528 *) Add support for STORE in ENGINE.
1531 *) Add the STORE type. The intention is to provide a common interface
1532 to certificate and key stores, be they simple file-based stores, or
1533 HSM-type store, or LDAP stores, or...
1534 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1537 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1538 pass a list of arguments to any function as well as provide a way
1539 for a function to pass data back to the caller.
1542 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1543 works like BUF_strdup() but can be used to duplicate a portion of
1544 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1548 *) Add the function sk_find_ex() which works like sk_find(), but will
1549 return an index to an element even if an exact match couldn't be
1550 found. The index is guaranteed to point at the element where the
1551 searched-for key would be inserted to preserve sorting order.
1554 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1555 takes an extra flags argument for optional functionality. Currently,
1556 the following flags are defined:
1558 OBJ_BSEARCH_VALUE_ON_NOMATCH
1559 This one gets OBJ_bsearch_ex() to return a pointer to the first
1560 element where the comparing function returns a negative or zero
1563 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1564 This one gets OBJ_bsearch_ex() to return a pointer to the first
1565 element where the comparing function returns zero. This is useful
1566 if there are more than one element where the comparing function
1570 *) Make it possible to create self-signed certificates with 'openssl ca'
1571 in such a way that the self-signed certificate becomes part of the
1572 CA database and uses the same mechanisms for serial number generation
1573 as all other certificate signing. The new flag '-selfsign' enables
1574 this functionality. Adapt CA.sh and CA.pl.in.
1577 *) Add functionality to check the public key of a certificate request
1578 against a given private. This is useful to check that a certificate
1579 request can be signed by that key (self-signing).
1582 *) Make it possible to have multiple active certificates with the same
1583 subject in the CA index file. This is done only if the keyword
1584 'unique_subject' is set to 'no' in the main CA section (default
1585 if 'CA_default') of the configuration file. The value is saved
1586 with the database itself in a separate index attribute file,
1587 named like the index file with '.attr' appended to the name.
1590 *) Generate muti valued AVAs using '+' notation in config files for
1594 *) Support for nameConstraints certificate extension.
1597 *) Support for policyConstraints certificate extension.
1600 *) Support for policyMappings certificate extension.
1603 *) Make sure the default DSA_METHOD implementation only uses its
1604 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1605 and change its own handlers to be NULL so as to remove unnecessary
1606 indirection. This lets alternative implementations fallback to the
1607 default implementation more easily.
1610 *) Support for directoryName in GeneralName related extensions
1614 *) Make it possible to link applications using Makefile.shared.
1615 Make that possible even when linking against static libraries!
1618 *) Support for single pass processing for S/MIME signing. This now
1619 means that S/MIME signing can be done from a pipe, in addition
1620 cleartext signing (multipart/signed type) is effectively streaming
1621 and the signed data does not need to be all held in memory.
1623 This is done with a new flag PKCS7_STREAM. When this flag is set
1624 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1625 is done after the data is output (and digests calculated) in
1626 SMIME_write_PKCS7().
1629 *) Add full support for -rpath/-R, both in shared libraries and
1630 applications, at least on the platforms where it's known how
1634 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1635 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1636 will now compute a table of multiples of the generator that
1637 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1638 faster (notably in the case of a single point multiplication,
1639 scalar * generator).
1640 [Nils Larsch, Bodo Moeller]
1642 *) IPv6 support for certificate extensions. The various extensions
1643 which use the IP:a.b.c.d can now take IPv6 addresses using the
1644 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1648 *) Added an ENGINE that implements RSA by performing private key
1649 exponentiations with the GMP library. The conversions to and from
1650 GMP's mpz_t format aren't optimised nor are any montgomery forms
1651 cached, and on x86 it appears OpenSSL's own performance has caught up.
1652 However there are likely to be other architectures where GMP could
1653 provide a boost. This ENGINE is not built in by default, but it can be
1654 specified at Configure time and should be accompanied by the necessary
1655 linker additions, eg;
1656 ./config -DOPENSSL_USE_GMP -lgmp
1659 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1660 testing availability of engines with "-t" - the old behaviour is
1661 produced by increasing the feature's verbosity with "-tt".
1664 *) ECDSA routines: under certain error conditions uninitialized BN objects
1665 could be freed. Solution: make sure initialization is performed early
1666 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1670 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1671 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1672 software implementations. For DSA and DH, parameter generation can
1673 also be overriden by providing the appropriate method callbacks.
1676 *) Change the "progress" mechanism used in key-generation and
1677 primality testing to functions that take a new BN_GENCB pointer in
1678 place of callback/argument pairs. The new API functions have "_ex"
1679 postfixes and the older functions are reimplemented as wrappers for
1680 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1681 declarations of the old functions to help (graceful) attempts to
1682 migrate to the new functions. Also, the new key-generation API
1683 functions operate on a caller-supplied key-structure and return
1684 success/failure rather than returning a key or NULL - this is to
1685 help make "keygen" another member function of RSA_METHOD etc.
1687 Example for using the new callback interface:
1689 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1693 BN_GENCB_set(&my_cb, my_callback, my_arg);
1695 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1696 /* For the meaning of a, b in calls to my_callback(), see the
1697 * documentation of the function that calls the callback.
1698 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1699 * my_callback should return 1 if it wants BN_is_prime_ex()
1700 * to continue, or 0 to stop.
1705 *) Change the ZLIB compression method to be stateful, and make it
1706 available to TLS with the number defined in
1707 draft-ietf-tls-compression-04.txt.
1710 *) Add the ASN.1 structures and functions for CertificatePair, which
1711 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1713 CertificatePair ::= SEQUENCE {
1714 forward [0] Certificate OPTIONAL,
1715 reverse [1] Certificate OPTIONAL,
1716 -- at least one of the pair shall be present -- }
1718 Also implement the PEM functions to read and write certificate
1719 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1721 This needed to be defined, mostly for the sake of the LDAP
1722 attribute crossCertificatePair, but may prove useful elsewhere as
1726 *) Make it possible to inhibit symlinking of shared libraries in
1727 Makefile.shared, for Cygwin's sake.
1730 *) Extend the BIGNUM API by creating a function
1731 void BN_set_negative(BIGNUM *a, int neg);
1732 and a macro that behave like
1733 int BN_is_negative(const BIGNUM *a);
1735 to avoid the need to access 'a->neg' directly in applications.
1738 *) Implement fast modular reduction for pseudo-Mersenne primes
1739 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1740 EC_GROUP_new_curve_GFp() will now automatically use this
1742 [Nils Larsch <nla@trustcenter.de>]
1744 *) Add new lock type (CRYPTO_LOCK_BN).
1747 *) Change the ENGINE framework to automatically load engines
1748 dynamically from specific directories unless they could be
1749 found to already be built in or loaded. Move all the
1750 current engines except for the cryptodev one to a new
1752 The engines in engines/ are built as shared libraries if
1753 the "shared" options was given to ./Configure or ./config.
1754 Otherwise, they are inserted in libcrypto.a.
1755 /usr/local/ssl/engines is the default directory for dynamic
1756 engines, but that can be overriden at configure time through
1757 the usual use of --prefix and/or --openssldir, and at run
1758 time with the environment variable OPENSSL_ENGINES.
1759 [Geoff Thorpe and Richard Levitte]
1761 *) Add Makefile.shared, a helper makefile to build shared
1762 libraries. Addapt Makefile.org.
1765 *) Add version info to Win32 DLLs.
1766 [Peter 'Luna' Runestig" <peter@runestig.com>]
1768 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1769 can be added using this API to created arbitrary PKCS#12
1770 files while avoiding the low level API.
1772 New options to PKCS12_create(), key or cert can be NULL and
1773 will then be omitted from the output file. The encryption
1774 algorithm NIDs can be set to -1 for no encryption, the mac
1775 iteration count can be set to 0 to omit the mac.
1777 Enhance pkcs12 utility by making the -nokeys and -nocerts
1778 options work when creating a PKCS#12 file. New option -nomac
1779 to omit the mac, NONE can be set for an encryption algorithm.
1780 New code is modified to use the enhanced PKCS12_create()
1781 instead of the low level API.
1784 *) Extend ASN1 encoder to support indefinite length constructed
1785 encoding. This can output sequences tags and octet strings in
1786 this form. Modify pk7_asn1.c to support indefinite length
1787 encoding. This is experimental and needs additional code to
1788 be useful, such as an ASN1 bio and some enhanced streaming
1791 Extend template encode functionality so that tagging is passed
1792 down to the template encoder.
1795 *) Let 'openssl req' fail if an argument to '-newkey' is not
1796 recognized instead of using RSA as a default.
1799 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1800 As these are not official, they are not included in "ALL";
1801 the "ECCdraft" ciphersuite group alias can be used to select them.
1802 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1804 *) Add ECDH engine support.
1805 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1807 *) Add ECDH in new directory crypto/ecdh/.
1808 [Douglas Stebila (Sun Microsystems Laboratories)]
1810 *) Let BN_rand_range() abort with an error after 100 iterations
1811 without success (which indicates a broken PRNG).
1814 *) Change BN_mod_sqrt() so that it verifies that the input value
1815 is really the square of the return value. (Previously,
1816 BN_mod_sqrt would show GIGO behaviour.)
1819 *) Add named elliptic curves over binary fields from X9.62, SECG,
1820 and WAP/WTLS; add OIDs that were still missing.
1822 [Sheueling Chang Shantz and Douglas Stebila
1823 (Sun Microsystems Laboratories)]
1825 *) Extend the EC library for elliptic curves over binary fields
1826 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1829 EC_GF2m_simple_method
1833 EC_GROUP_new_curve_GF2m
1834 EC_GROUP_set_curve_GF2m
1835 EC_GROUP_get_curve_GF2m
1836 EC_POINT_set_affine_coordinates_GF2m
1837 EC_POINT_get_affine_coordinates_GF2m
1838 EC_POINT_set_compressed_coordinates_GF2m
1840 Point compression for binary fields is disabled by default for
1841 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1844 As binary polynomials are represented as BIGNUMs, various members
1845 of the EC_GROUP and EC_POINT data structures can be shared
1846 between the implementations for prime fields and binary fields;
1847 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1848 are essentially identical to their ..._GFp counterparts.
1849 (For simplicity, the '..._GFp' prefix has been dropped from
1850 various internal method names.)
1852 An internal 'field_div' method (similar to 'field_mul' and
1853 'field_sqr') has been added; this is used only for binary fields.
1855 [Sheueling Chang Shantz and Douglas Stebila
1856 (Sun Microsystems Laboratories)]
1858 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1859 through methods ('mul', 'precompute_mult').
1861 The generic implementations (now internally called 'ec_wNAF_mul'
1862 and 'ec_wNAF_precomputed_mult') remain the default if these
1863 methods are undefined.
1865 [Sheueling Chang Shantz and Douglas Stebila
1866 (Sun Microsystems Laboratories)]
1868 *) New function EC_GROUP_get_degree, which is defined through
1869 EC_METHOD. For curves over prime fields, this returns the bit
1870 length of the modulus.
1872 [Sheueling Chang Shantz and Douglas Stebila
1873 (Sun Microsystems Laboratories)]
1875 *) New functions EC_GROUP_dup, EC_POINT_dup.
1876 (These simply call ..._new and ..._copy).
1878 [Sheueling Chang Shantz and Douglas Stebila
1879 (Sun Microsystems Laboratories)]
1881 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1882 Polynomials are represented as BIGNUMs (where the sign bit is not
1883 used) in the following functions [macros]:
1886 BN_GF2m_sub [= BN_GF2m_add]
1887 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1888 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1889 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1891 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1892 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1893 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1894 BN_GF2m_cmp [= BN_ucmp]
1896 (Note that only the 'mod' functions are actually for fields GF(2^m).
1897 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1899 For some functions, an the irreducible polynomial defining a
1900 field can be given as an 'unsigned int[]' with strictly
1901 decreasing elements giving the indices of those bits that are set;
1902 i.e., p[] represents the polynomial
1903 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1905 p[0] > p[1] > ... > p[k] = 0.
1906 This applies to the following functions:
1911 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1912 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1914 BN_GF2m_mod_sqrt_arr
1915 BN_GF2m_mod_solve_quad_arr
1919 Conversion can be performed by the following functions:
1924 bntest.c has additional tests for binary polynomial arithmetic.
1926 Two implementations for BN_GF2m_mod_div() are available.
1927 The default algorithm simply uses BN_GF2m_mod_inv() and
1928 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1929 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1930 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1932 [Sheueling Chang Shantz and Douglas Stebila
1933 (Sun Microsystems Laboratories)]
1935 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1936 functionality is disabled at compile-time.
1937 [Douglas Stebila <douglas.stebila@sun.com>]
1939 *) Change default behaviour of 'openssl asn1parse' so that more
1940 information is visible when viewing, e.g., a certificate:
1942 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1943 mode the content of non-printable OCTET STRINGs is output in a
1944 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1945 avoid the appearance of a printable string.
1946 [Nils Larsch <nla@trustcenter.de>]
1948 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1950 EC_GROUP_set_asn1_flag()
1951 EC_GROUP_get_asn1_flag()
1952 EC_GROUP_set_point_conversion_form()
1953 EC_GROUP_get_point_conversion_form()
1954 These control ASN1 encoding details:
1955 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1956 has been set to OPENSSL_EC_NAMED_CURVE.
1957 - Points are encoded in uncompressed form by default; options for
1958 asn1_for are as for point2oct, namely
1959 POINT_CONVERSION_COMPRESSED
1960 POINT_CONVERSION_UNCOMPRESSED
1961 POINT_CONVERSION_HYBRID
1963 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1966 EC_GROUP_get0_seed()
1967 EC_GROUP_get_seed_len()
1968 This is used only for ASN1 purposes (so far).
1969 [Nils Larsch <nla@trustcenter.de>]
1971 *) Add 'field_type' member to EC_METHOD, which holds the NID
1972 of the appropriate field type OID. The new function
1973 EC_METHOD_get_field_type() returns this value.
1974 [Nils Larsch <nla@trustcenter.de>]
1979 EC_POINT_point2hex()
1980 EC_POINT_hex2point()
1981 providing useful interfaces to EC_POINT_point2oct() and
1982 EC_POINT_oct2point().
1983 [Nils Larsch <nla@trustcenter.de>]
1985 *) Change internals of the EC library so that the functions
1986 EC_GROUP_set_generator()
1987 EC_GROUP_get_generator()
1988 EC_GROUP_get_order()
1989 EC_GROUP_get_cofactor()
1990 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1991 to methods, which would lead to unnecessary code duplication when
1992 adding different types of curves.
1993 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1995 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1996 arithmetic, and such that modified wNAFs are generated
1997 (which avoid length expansion in many cases).
2000 *) Add a function EC_GROUP_check_discriminant() (defined via
2001 EC_METHOD) that verifies that the curve discriminant is non-zero.
2003 Add a function EC_GROUP_check() that makes some sanity tests
2004 on a EC_GROUP, its generator and order. This includes
2005 EC_GROUP_check_discriminant().
2006 [Nils Larsch <nla@trustcenter.de>]
2008 *) Add ECDSA in new directory crypto/ecdsa/.
2010 Add applications 'openssl ecparam' and 'openssl ecdsa'
2011 (these are based on 'openssl dsaparam' and 'openssl dsa').
2013 ECDSA support is also included in various other files across the
2014 library. Most notably,
2015 - 'openssl req' now has a '-newkey ecdsa:file' option;
2016 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2017 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2018 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2019 them suitable for ECDSA where domain parameters must be
2020 extracted before the specific public key;
2021 - ECDSA engine support has been added.
2022 [Nils Larsch <nla@trustcenter.de>]
2024 *) Include some named elliptic curves, and add OIDs from X9.62,
2025 SECG, and WAP/WTLS. Each curve can be obtained from the new
2027 EC_GROUP_new_by_curve_name(),
2028 and the list of available named curves can be obtained with
2029 EC_get_builtin_curves().
2030 Also add a 'curve_name' member to EC_GROUP objects, which can be
2032 EC_GROUP_set_curve_name()
2033 EC_GROUP_get_curve_name()
2034 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2036 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2037 was actually never needed) and in BN_mul(). The removal in BN_mul()
2038 required a small change in bn_mul_part_recursive() and the addition
2039 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2040 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2041 bn_sub_words() and bn_add_words() except they take arrays with
2045 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2047 *) Cleanse PEM buffers before freeing them since they may contain
2049 [Benjamin Bennett <ben@psc.edu>]
2051 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2052 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2053 authentication-only ciphersuites.
2056 *) Since AES128 and AES256 share a single mask bit in the logic of
2057 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2058 kludge to work properly if AES128 is available and AES256 isn't.
2061 *) Expand security boundary to match 1.1.1 module.
2064 *) Remove redundant features: hash file source, editing of test vectors
2065 modify fipsld to use external fips_premain.c signature.
2068 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2069 run algorithm test programs.
2072 *) Make algorithm test programs more tolerant of whitespace.
2075 *) Have SSL/TLS server implementation tolerate "mismatched" record
2076 protocol version while receiving ClientHello even if the
2077 ClientHello is fragmented. (The server can't insist on the
2078 particular protocol version it has chosen before the ServerHello
2079 message has informed the client about his choice.)
2082 *) Load error codes if they are not already present instead of using a
2083 static variable. This allows them to be cleanly unloaded and reloaded.
2086 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2088 *) Introduce limits to prevent malicious keys being able to
2089 cause a denial of service. (CVE-2006-2940)
2090 [Steve Henson, Bodo Moeller]
2092 *) Fix ASN.1 parsing of certain invalid structures that can result
2093 in a denial of service. (CVE-2006-2937) [Steve Henson]
2095 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2096 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2098 *) Fix SSL client code which could crash if connecting to a
2099 malicious SSLv2 server. (CVE-2006-4343)
2100 [Tavis Ormandy and Will Drewry, Google Security Team]
2102 *) Change ciphersuite string processing so that an explicit
2103 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2104 will no longer include "AES128-SHA"), and any other similar
2105 ciphersuite (same bitmap) from *other* protocol versions (so that
2106 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2107 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2108 changes from 0.9.8b and 0.9.8d.
2111 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2113 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2114 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2116 *) Change the Unix randomness entropy gathering to use poll() when
2117 possible instead of select(), since the latter has some
2118 undesirable limitations.
2119 [Darryl Miles via Richard Levitte and Bodo Moeller]
2121 *) Disable rogue ciphersuites:
2123 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2124 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2125 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2127 The latter two were purportedly from
2128 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2131 Also deactive the remaining ciphersuites from
2132 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2133 unofficial, and the ID has long expired.
2136 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2137 dual-core machines) and other potential thread-safety issues.
2140 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2142 *) Adapt fipsld and the build system to link against the validated FIPS
2143 module in FIPS mode.
2146 *) Fixes for VC++ 2005 build under Windows.
2149 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2150 from a Windows bash shell such as MSYS. It is autodetected from the
2151 "config" script when run from a VC++ environment. Modify standard VC++
2152 build to use fipscanister.o from the GNU make build.
2155 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2157 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2158 The value now differs depending on if you build for FIPS or not.
2159 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2160 safely run with a non-FIPSed libcrypto, as it may crash because of
2161 the difference induced by this change.
2164 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2166 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2167 (part of SSL_OP_ALL). This option used to disable the
2168 countermeasure against man-in-the-middle protocol-version
2169 rollback in the SSL 2.0 server implementation, which is a bad
2170 idea. (CVE-2005-2969)
2172 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2173 for Information Security, National Institute of Advanced Industrial
2174 Science and Technology [AIST], Japan)]
2176 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2177 mainly for FIPS compliance and not fully integrated at this stage.
2180 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2181 the exponentiation using a fixed-length exponent. (Otherwise,
2182 the information leaked through timing could expose the secret key
2183 after many signatures; cf. Bleichenbacher's attack on DSA with
2187 *) Make a new fixed-window mod_exp implementation the default for
2188 RSA, DSA, and DH private-key operations so that the sequence of
2189 squares and multiplies and the memory access pattern are
2190 independent of the particular secret key. This will mitigate
2191 cache-timing and potential related attacks.
2193 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2194 and this is automatically used by BN_mod_exp_mont() if the new flag
2195 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2196 will use this BN flag for private exponents unless the flag
2197 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2198 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2200 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2202 *) Change the client implementation for SSLv23_method() and
2203 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2204 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2205 (Previously, the SSL 2.0 backwards compatible Client Hello
2206 message format would be used even with SSL_OP_NO_SSLv2.)
2209 *) Add support for smime-type MIME parameter in S/MIME messages which some
2213 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2214 a threadsafe manner. Modify rsa code to use new function and add calls
2215 to dsa and dh code (which had race conditions before).
2218 *) Include the fixed error library code in the C error file definitions
2219 instead of fixing them up at runtime. This keeps the error code
2220 structures constant.
2223 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2225 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2228 *) Fixes for newer kerberos headers. NB: the casts are needed because
2229 the 'length' field is signed on one version and unsigned on another
2230 with no (?) obvious way to tell the difference, without these VC++
2231 complains. Also the "definition" of FAR (blank) is no longer included
2232 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2233 some needed definitions.
2236 *) Undo Cygwin change.
2239 *) Added support for proxy certificates according to RFC 3820.
2240 Because they may be a security thread to unaware applications,
2241 they must be explicitely allowed in run-time. See
2242 docs/HOWTO/proxy_certificates.txt for further information.
2245 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2247 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2248 server and client random values. Previously
2249 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2250 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2252 This change has negligible security impact because:
2254 1. Server and client random values still have 24 bytes of pseudo random
2257 2. Server and client random values are sent in the clear in the initial
2260 3. The master secret is derived using the premaster secret (48 bytes in
2261 size for static RSA ciphersuites) as well as client server and random
2264 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2267 [Stephen Henson, reported by UK NISCC]
2269 *) Use Windows randomness collection on Cygwin.
2272 *) Fix hang in EGD/PRNGD query when communication socket is closed
2273 prematurely by EGD/PRNGD.
2274 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2276 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2279 *) Back-port of selected performance improvements from development
2280 branch, as well as improved support for PowerPC platforms.
2283 *) Add lots of checks for memory allocation failure, error codes to indicate
2284 failure and freeing up memory if a failure occurs.
2285 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2287 *) Add new -passin argument to dgst.
2290 *) Perform some character comparisons of different types in X509_NAME_cmp:
2291 this is needed for some certificates that reencode DNs into UTF8Strings
2292 (in violation of RFC3280) and can't or wont issue name rollover
2296 *) Make an explicit check during certificate validation to see that
2297 the CA setting in each certificate on the chain is correct. As a
2298 side effect always do the following basic checks on extensions,
2299 not just when there's an associated purpose to the check:
2301 - if there is an unhandled critical extension (unless the user
2302 has chosen to ignore this fault)
2303 - if the path length has been exceeded (if one is set at all)
2304 - that certain extensions fit the associated purpose (if one has
2308 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2310 *) Avoid a race condition when CRLs are checked in a multi threaded
2311 environment. This would happen due to the reordering of the revoked
2312 entries during signature checking and serial number lookup. Now the
2313 encoding is cached and the serial number sort performed under a lock.
2314 Add new STACK function sk_is_sorted().
2317 *) Add Delta CRL to the extension code.
2320 *) Various fixes to s3_pkt.c so alerts are sent properly.
2321 [David Holmes <d.holmes@f5.com>]
2323 *) Reduce the chances of duplicate issuer name and serial numbers (in
2324 violation of RFC3280) using the OpenSSL certificate creation utilities.
2325 This is done by creating a random 64 bit value for the initial serial
2326 number when a serial number file is created or when a self signed
2327 certificate is created using 'openssl req -x509'. The initial serial
2328 number file is created using 'openssl x509 -next_serial' in CA.pl
2329 rather than being initialized to 1.
2332 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2334 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2335 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2336 [Joe Orton, Steve Henson]
2338 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2340 [Joe Orton, Steve Henson]
2342 *) Make it possible to have multiple active certificates with the same
2343 subject in the CA index file. This is done only if the keyword
2344 'unique_subject' is set to 'no' in the main CA section (default
2345 if 'CA_default') of the configuration file. The value is saved
2346 with the database itself in a separate index attribute file,
2347 named like the index file with '.attr' appended to the name.
2350 *) X509 verify fixes. Disable broken certificate workarounds when
2351 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2352 keyUsage extension present. Don't accept CRLs with unhandled critical
2353 extensions: since verify currently doesn't process CRL extensions this
2354 rejects a CRL with *any* critical extensions. Add new verify error codes
2358 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2359 A clarification of RFC2560 will require the use of OCTET STRINGs and
2360 some implementations cannot handle the current raw format. Since OpenSSL
2361 copies and compares OCSP nonces as opaque blobs without any attempt at
2362 parsing them this should not create any compatibility issues.
2365 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2366 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2367 this HMAC (and other) operations are several times slower than OpenSSL
2371 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2372 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2374 *) Use the correct content when signing type "other".
2377 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2379 *) Fix various bugs revealed by running the NISCC test suite:
2381 Stop out of bounds reads in the ASN1 code when presented with
2382 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2384 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2386 If verify callback ignores invalid public key errors don't try to check
2387 certificate signature with the NULL public key.
2391 *) New -ignore_err option in ocsp application to stop the server
2392 exiting on the first error in a request.
2395 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2396 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2400 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2401 extra data after the compression methods not only for TLS 1.0
2402 but also for SSL 3.0 (as required by the specification).
2403 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2405 *) Change X509_certificate_type() to mark the key as exported/exportable
2406 when it's 512 *bits* long, not 512 bytes.
2409 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2410 blocks during encryption.
2413 *) Various fixes to base64 BIO and non blocking I/O. On write
2414 flushes were not handled properly if the BIO retried. On read
2415 data was not being buffered properly and had various logic bugs.
2416 This also affects blocking I/O when the data being decoded is a
2420 *) Various S/MIME bugfixes and compatibility changes:
2421 output correct application/pkcs7 MIME type if
2422 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2423 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2424 of files as .eml work). Correctly handle very long lines in MIME
2428 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2430 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2431 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2432 a protocol version number mismatch like a decryption error
2433 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2436 *) Turn on RSA blinding by default in the default implementation
2437 to avoid a timing attack. Applications that don't want it can call
2438 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2439 They would be ill-advised to do so in most cases.
2440 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2442 *) Change RSA blinding code so that it works when the PRNG is not
2443 seeded (in this case, the secret RSA exponent is abused as
2444 an unpredictable seed -- if it is not unpredictable, there
2445 is no point in blinding anyway). Make RSA blinding thread-safe
2446 by remembering the creator's thread ID in rsa->blinding and
2447 having all other threads use local one-time blinding factors
2448 (this requires more computation than sharing rsa->blinding, but
2449 avoids excessive locking; and if an RSA object is not shared
2450 between threads, blinding will still be very fast).
2453 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2454 ENGINE as defaults for all supported algorithms irrespective of
2455 the 'flags' parameter. 'flags' is now honoured, so applications
2456 should make sure they are passing it correctly.
2459 *) Target "mingw" now allows native Windows code to be generated in
2460 the Cygwin environment as well as with the MinGW compiler.
2463 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2465 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2466 via timing by performing a MAC computation even if incorrrect
2467 block cipher padding has been found. This is a countermeasure
2468 against active attacks where the attacker has to distinguish
2469 between bad padding and a MAC verification error. (CVE-2003-0078)
2471 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2472 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2473 Martin Vuagnoux (EPFL, Ilion)]
2475 *) Make the no-err option work as intended. The intention with no-err
2476 is not to have the whole error stack handling routines removed from
2477 libcrypto, it's only intended to remove all the function name and
2478 reason texts, thereby removing some of the footprint that may not
2479 be interesting if those errors aren't displayed anyway.
2481 NOTE: it's still possible for any application or module to have it's
2482 own set of error texts inserted. The routines are there, just not
2483 used by default when no-err is given.
2486 *) Add support for FreeBSD on IA64.
2487 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2489 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2490 Kerberos function mit_des_cbc_cksum(). Before this change,
2491 the value returned by DES_cbc_cksum() was like the one from
2492 mit_des_cbc_cksum(), except the bytes were swapped.
2493 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2495 *) Allow an application to disable the automatic SSL chain building.
2496 Before this a rather primitive chain build was always performed in
2497 ssl3_output_cert_chain(): an application had no way to send the
2498 correct chain if the automatic operation produced an incorrect result.
2500 Now the chain builder is disabled if either:
2502 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2504 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2506 The reasoning behind this is that an application would not want the
2507 auto chain building to take place if extra chain certificates are
2508 present and it might also want a means of sending no additional
2509 certificates (for example the chain has two certificates and the
2513 *) Add the possibility to build without the ENGINE framework.
2514 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2516 *) Under Win32 gmtime() can return NULL: check return value in
2517 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2520 *) DSA routines: under certain error conditions uninitialized BN objects
2521 could be freed. Solution: make sure initialization is performed early
2522 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2523 Nils Larsch <nla@trustcenter.de> via PR#459)
2526 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2527 checked on reconnect on the client side, therefore session resumption
2528 could still fail with a "ssl session id is different" error. This
2529 behaviour is masked when SSL_OP_ALL is used due to
2530 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2531 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2532 followup to PR #377.
2535 *) IA-32 assembler support enhancements: unified ELF targets, support
2536 for SCO/Caldera platforms, fix for Cygwin shared build.
2539 *) Add support for FreeBSD on sparc64. As a consequence, support for
2540 FreeBSD on non-x86 processors is separate from x86 processors on
2541 the config script, much like the NetBSD support.
2542 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2544 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2546 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2549 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2550 code (06) was taken as the first octet of the session ID and the last
2551 octet was ignored consequently. As a result SSLv2 client side session
2552 caching could not have worked due to the session ID mismatch between
2554 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2558 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2559 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2563 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2564 seems that in spite of existing for more than a year, many application
2565 author have done nothing to provide the necessary callbacks, which
2566 means that this particular engine will not work properly anywhere.
2567 This is a very unfortunate situation which forces us, in the name
2568 of usability, to give the hw_ncipher.c a static lock, which is part
2570 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2571 appear in 0.9.8 or later. We EXPECT application authors to have
2572 dealt properly with this when 0.9.8 is released (unless we actually
2573 make such changes in the libcrypto locking code that changes will
2574 have to be made anyway).
2577 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2578 octets have been read, EOF or an error occurs. Without this change
2579 some truncated ASN1 structures will not produce an error.
2582 *) Disable Heimdal support, since it hasn't been fully implemented.
2583 Still give the possibility to force the use of Heimdal, but with
2584 warnings and a request that patches get sent to openssl-dev.
2587 *) Add the VC-CE target, introduce the WINCE sysname, and add
2588 INSTALL.WCE and appropriate conditionals to make it build.
2589 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2591 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2592 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2593 edit numbers of the version.
2594 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2596 *) Introduce safe string copy and catenation functions
2597 (BUF_strlcpy() and BUF_strlcat()).
2598 [Ben Laurie (CHATS) and Richard Levitte]
2600 *) Avoid using fixed-size buffers for one-line DNs.
2601 [Ben Laurie (CHATS)]
2603 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2604 resizing buffers containing secrets, and use where appropriate.
2605 [Ben Laurie (CHATS)]
2607 *) Avoid using fixed size buffers for configuration file location.
2608 [Ben Laurie (CHATS)]
2610 *) Avoid filename truncation for various CA files.
2611 [Ben Laurie (CHATS)]
2613 *) Use sizeof in preference to magic numbers.
2614 [Ben Laurie (CHATS)]
2616 *) Avoid filename truncation in cert requests.
2617 [Ben Laurie (CHATS)]
2619 *) Add assertions to check for (supposedly impossible) buffer
2621 [Ben Laurie (CHATS)]
2623 *) Don't cache truncated DNS entries in the local cache (this could
2624 potentially lead to a spoofing attack).
2625 [Ben Laurie (CHATS)]
2627 *) Fix various buffers to be large enough for hex/decimal
2628 representations in a platform independent manner.
2629 [Ben Laurie (CHATS)]
2631 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2632 resizing buffers containing secrets, and use where appropriate.
2633 [Ben Laurie (CHATS)]
2635 *) Add BIO_indent() to avoid much slightly worrying code to do
2637 [Ben Laurie (CHATS)]
2639 *) Convert sprintf()/BIO_puts() to BIO_printf().
2640 [Ben Laurie (CHATS)]
2642 *) buffer_gets() could terminate with the buffer only half
2644 [Ben Laurie (CHATS)]
2646 *) Add assertions to prevent user-supplied crypto functions from
2647 overflowing internal buffers by having large block sizes, etc.
2648 [Ben Laurie (CHATS)]
2650 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2652 [Ben Laurie (CHATS)]
2654 *) Eliminate unused copy of key in RC4.
2655 [Ben Laurie (CHATS)]
2657 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2658 [Ben Laurie (CHATS)]
2660 *) Fix off-by-one error in EGD path.
2661 [Ben Laurie (CHATS)]
2663 *) If RANDFILE path is too long, ignore instead of truncating.
2664 [Ben Laurie (CHATS)]
2666 *) Eliminate unused and incorrectly sized X.509 structure
2668 [Ben Laurie (CHATS)]
2670 *) Eliminate unused and dangerous function knumber().
2671 [Ben Laurie (CHATS)]
2673 *) Eliminate unused and dangerous structure, KSSL_ERR.
2674 [Ben Laurie (CHATS)]
2676 *) Protect against overlong session ID context length in an encoded
2677 session object. Since these are local, this does not appear to be
2679 [Ben Laurie (CHATS)]
2681 *) Change from security patch (see 0.9.6e below) that did not affect
2682 the 0.9.6 release series:
2684 Remote buffer overflow in SSL3 protocol - an attacker could
2685 supply an oversized master key in Kerberos-enabled versions.
2687 [Ben Laurie (CHATS)]
2689 *) Change the SSL kerb5 codes to match RFC 2712.
2692 *) Make -nameopt work fully for req and add -reqopt switch.
2693 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2695 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2696 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2698 *) Make sure tests can be performed even if the corresponding algorithms
2699 have been removed entirely. This was also the last step to make
2700 OpenSSL compilable with DJGPP under all reasonable conditions.
2701 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2703 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2704 to allow version independent disabling of normally unselected ciphers,
2705 which may be activated as a side-effect of selecting a single cipher.
2707 (E.g., cipher list string "RSA" enables ciphersuites that are left
2708 out of "ALL" because they do not provide symmetric encryption.
2709 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2710 [Lutz Jaenicke, Bodo Moeller]
2712 *) Add appropriate support for separate platform-dependent build
2713 directories. The recommended way to make a platform-dependent
2714 build directory is the following (tested on Linux), maybe with
2717 # Place yourself outside of the OpenSSL source tree. In
2718 # this example, the environment variable OPENSSL_SOURCE
2719 # is assumed to contain the absolute OpenSSL source directory.
2720 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2721 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2722 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2723 mkdir -p `dirname $F`
2724 ln -s $OPENSSL_SOURCE/$F $F
2727 To be absolutely sure not to disturb the source tree, a "make clean"
2728 is a good thing. If it isn't successfull, don't worry about it,
2729 it probably means the source directory is very clean.
2732 *) Make sure any ENGINE control commands make local copies of string
2733 pointers passed to them whenever necessary. Otherwise it is possible
2734 the caller may have overwritten (or deallocated) the original string
2735 data when a later ENGINE operation tries to use the stored values.
2736 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2738 *) Improve diagnostics in file reading and command-line digests.
2739 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2741 *) Add AES modes CFB and OFB to the object database. Correct an
2742 error in AES-CFB decryption.
2745 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2746 allows existing EVP_CIPHER_CTX structures to be reused after
2747 calling EVP_*Final(). This behaviour is used by encryption
2748 BIOs and some applications. This has the side effect that
2749 applications must explicitly clean up cipher contexts with
2750 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2753 *) Check the values of dna and dnb in bn_mul_recursive before calling
2754 bn_mul_comba (a non zero value means the a or b arrays do not contain
2755 n2 elements) and fallback to bn_mul_normal if either is not zero.
2758 *) Fix escaping of non-ASCII characters when using the -subj option
2759 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2762 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2763 form for "surname", serialNumber has no short form.
2764 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2765 therefore remove "mail" short name for "internet 7".
2766 The OID for unique identifiers in X509 certificates is
2767 x500UniqueIdentifier, not uniqueIdentifier.
2768 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2771 *) Add an "init" command to the ENGINE config module and auto initialize
2772 ENGINEs. Without any "init" command the ENGINE will be initialized
2773 after all ctrl commands have been executed on it. If init=1 the
2774 ENGINE is initailized at that point (ctrls before that point are run
2775 on the uninitialized ENGINE and after on the initialized one). If
2776 init=0 then the ENGINE will not be iniatialized at all.
2779 *) Fix the 'app_verify_callback' interface so that the user-defined
2780 argument is actually passed to the callback: In the
2781 SSL_CTX_set_cert_verify_callback() prototype, the callback
2782 declaration has been changed from
2785 int (*cb)(X509_STORE_CTX *,void *);
2786 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2787 i=s->ctx->app_verify_callback(&ctx)
2788 has been changed into
2789 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2791 To update applications using SSL_CTX_set_cert_verify_callback(),
2792 a dummy argument can be added to their callback functions.
2793 [D. K. Smetters <smetters@parc.xerox.com>]
2795 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2796 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2798 *) Add and OPENSSL_LOAD_CONF define which will cause
2799 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2800 This allows older applications to transparently support certain
2801 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2802 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2803 load the config file and OPENSSL_add_all_algorithms_conf() which will
2804 always load it have also been added.
2807 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2808 Adjust NIDs and EVP layer.
2809 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2811 *) Config modules support in openssl utility.
2813 Most commands now load modules from the config file,
2814 though in a few (such as version) this isn't done
2815 because it couldn't be used for anything.
2817 In the case of ca and req the config file used is
2818 the same as the utility itself: that is the -config
2819 command line option can be used to specify an
2823 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2824 use "openssl_conf" if filename is NULL use default openssl config file.
2827 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2828 config section name. Add a new flag to tolerate a missing config file
2829 and move code to CONF_modules_load_file().
2832 *) Support for crypto accelerator cards from Accelerated Encryption
2833 Processing, www.aep.ie. (Use engine 'aep')
2834 The support was copied from 0.9.6c [engine] and adapted/corrected
2835 to work with the new engine framework.
2836 [AEP Inc. and Richard Levitte]
2838 *) Support for SureWare crypto accelerator cards from Baltimore
2839 Technologies. (Use engine 'sureware')
2840 The support was copied from 0.9.6c [engine] and adapted
2841 to work with the new engine framework.
2844 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2845 make the newer ENGINE framework commands for the CHIL engine work.
2846 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2848 *) Make it possible to produce shared libraries on ReliantUNIX.
2849 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2851 *) Add the configuration target debug-linux-ppro.
2852 Make 'openssl rsa' use the general key loading routines
2853 implemented in apps.c, and make those routines able to
2854 handle the key format FORMAT_NETSCAPE and the variant
2856 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2858 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2859 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2861 *) Add -keyform to rsautl, and document -engine.
2862 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2864 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2865 BIO_R_NO_SUCH_FILE error code rather than the generic
2866 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2869 *) Add new functions
2871 ERR_peek_last_error_line
2872 ERR_peek_last_error_line_data.
2873 These are similar to
2876 ERR_peek_error_line_data,
2877 but report on the latest error recorded rather than the first one
2878 still in the error queue.
2879 [Ben Laurie, Bodo Moeller]
2881 *) default_algorithms option in ENGINE config module. This allows things
2883 default_algorithms = ALL
2884 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2887 *) Prelminary ENGINE config module.
2890 *) New experimental application configuration code.
2893 *) Change the AES code to follow the same name structure as all other
2894 symmetric ciphers, and behave the same way. Move everything to
2895 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2896 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2898 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2899 [Ben Laurie and Theo de Raadt]
2901 *) Add option to output public keys in req command.
2902 [Massimiliano Pala madwolf@openca.org]
2904 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2905 (up to about 10% better than before for P-192 and P-224).
2908 *) New functions/macros
2910 SSL_CTX_set_msg_callback(ctx, cb)
2911 SSL_CTX_set_msg_callback_arg(ctx, arg)
2912 SSL_set_msg_callback(ssl, cb)
2913 SSL_set_msg_callback_arg(ssl, arg)
2915 to request calling a callback function
2917 void cb(int write_p, int version, int content_type,
2918 const void *buf, size_t len, SSL *ssl, void *arg)
2920 whenever a protocol message has been completely received
2921 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2922 protocol version according to which the SSL library interprets
2923 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2924 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2925 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2926 specification (change_cipher_spec(20), alert(21), handshake(22)).
2927 'buf' and 'len' point to the actual message, 'ssl' to the
2928 SSL object, and 'arg' is the application-defined value set by
2929 SSL[_CTX]_set_msg_callback_arg().
2931 'openssl s_client' and 'openssl s_server' have new '-msg' options
2932 to enable a callback that displays all protocol messages.
2935 *) Change the shared library support so shared libraries are built as
2936 soon as the corresponding static library is finished, and thereby get
2937 openssl and the test programs linked against the shared library.
2938 This still only happens when the keyword "shard" has been given to
2939 the configuration scripts.
2941 NOTE: shared library support is still an experimental thing, and
2942 backward binary compatibility is still not guaranteed.
2943 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2945 *) Add support for Subject Information Access extension.
2946 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2948 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2949 additional bytes when new memory had to be allocated, not just
2950 when reusing an existing buffer.
2953 *) New command line and configuration option 'utf8' for the req command.
2954 This allows field values to be specified as UTF8 strings.
2957 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2958 runs for the former and machine-readable output for the latter.
2961 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2962 of the e-mail address in the DN (i.e., it will go into a certificate
2963 extension only). The new configuration file option 'email_in_dn = no'
2964 has the same effect.
2965 [Massimiliano Pala madwolf@openca.org]
2967 *) Change all functions with names starting with des_ to be starting
2968 with DES_ instead. Add wrappers that are compatible with libdes,
2969 but are named _ossl_old_des_*. Finally, add macros that map the
2970 des_* symbols to the corresponding _ossl_old_des_* if libdes
2971 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2972 desired, the des_* symbols will be mapped to DES_*, with one
2975 Since we provide two compatibility mappings, the user needs to
2976 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2977 compatibility is desired. The default (i.e., when that macro
2978 isn't defined) is OpenSSL 0.9.6c compatibility.
2980 There are also macros that enable and disable the support of old
2981 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2982 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2983 are defined, the default will apply: to support the old des routines.
2985 In either case, one must include openssl/des.h to get the correct
2986 definitions. Do not try to just include openssl/des_old.h, that
2989 NOTE: This is a major break of an old API into a new one. Software
2990 authors are encouraged to switch to the DES_ style functions. Some
2991 time in the future, des_old.h and the libdes compatibility functions
2992 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2993 default), and then completely removed.
2996 *) Test for certificates which contain unsupported critical extensions.
2997 If such a certificate is found during a verify operation it is
2998 rejected by default: this behaviour can be overridden by either
2999 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3000 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3001 X509_supported_extension() has also been added which returns 1 if a
3002 particular extension is supported.
3005 *) Modify the behaviour of EVP cipher functions in similar way to digests
3006 to retain compatibility with existing code.
3009 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3010 compatibility with existing code. In particular the 'ctx' parameter does
3011 not have to be to be initialized before the call to EVP_DigestInit() and
3012 it is tidied up after a call to EVP_DigestFinal(). New function
3013 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3014 EVP_MD_CTX_copy() changed to not require the destination to be
3015 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3016 requires the destination to be valid.
3018 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3019 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3022 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3023 so that complete 'Handshake' protocol structures are kept in memory
3024 instead of overwriting 'msg_type' and 'length' with 'body' data.
3027 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3028 [Massimo Santin via Richard Levitte]
3030 *) Major restructuring to the underlying ENGINE code. This includes
3031 reduction of linker bloat, separation of pure "ENGINE" manipulation
3032 (initialisation, etc) from functionality dealing with implementations
3033 of specific crypto iterfaces. This change also introduces integrated
3034 support for symmetric ciphers and digest implementations - so ENGINEs
3035 can now accelerate these by providing EVP_CIPHER and EVP_MD
3036 implementations of their own. This is detailed in crypto/engine/README
3037 as it couldn't be adequately described here. However, there are a few
3038 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3039 were changed in the original introduction of ENGINE code have now
3040 reverted back - the hooking from this code to ENGINE is now a good
3041 deal more passive and at run-time, operations deal directly with
3042 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3043 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3044 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3045 they were not being used by the framework as there is no concept of a
3046 BIGNUM_METHOD and they could not be generalised to the new
3047 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3048 ENGINE_cpy() has been removed as it cannot be consistently defined in
3052 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3055 *) Change mkdef.pl to sort symbols that get the same entry number,
3056 and make sure the automatically generated functions ERR_load_*
3057 become part of libeay.num as well.
3060 *) New function SSL_renegotiate_pending(). This returns true once
3061 renegotiation has been requested (either SSL_renegotiate() call
3062 or HelloRequest/ClientHello receveived from the peer) and becomes
3063 false once a handshake has been completed.
3064 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3065 sends a HelloRequest, but does not ensure that a handshake takes
3066 place. SSL_renegotiate_pending() is useful for checking if the
3067 client has followed the request.)
3070 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3071 By default, clients may request session resumption even during
3072 renegotiation (if session ID contexts permit); with this option,
3073 session resumption is possible only in the first handshake.
3075 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3076 more bits available for options that should not be part of
3077 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3080 *) Add some demos for certificate and certificate request creation.
3083 *) Make maximum certificate chain size accepted from the peer application
3084 settable (SSL*_get/set_max_cert_list()), as proposed by
3085 "Douglas E. Engert" <deengert@anl.gov>.
3088 *) Add support for shared libraries for Unixware-7
3089 (Boyd Lynn Gerber <gerberb@zenez.com>).
3092 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3093 be done prior to destruction. Use this to unload error strings from
3094 ENGINEs that load their own error strings. NB: This adds two new API
3095 functions to "get" and "set" this destroy handler in an ENGINE.
3098 *) Alter all existing ENGINE implementations (except "openssl" and
3099 "openbsd") to dynamically instantiate their own error strings. This
3100 makes them more flexible to be built both as statically-linked ENGINEs
3101 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3102 Also, add stub code to each that makes building them as self-contained
3103 shared-libraries easier (see README.ENGINE).
3106 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3107 implementations into applications that are completely implemented in
3108 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3109 commands that can be used to configure what shared-library to load and
3110 to control aspects of the way it is handled. Also, made an update to
3111 the README.ENGINE file that brings its information up-to-date and
3112 provides some information and instructions on the "dynamic" ENGINE
3113 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3116 *) Make it possible to unload ranges of ERR strings with a new
3117 "ERR_unload_strings" function.
3120 *) Add a copy() function to EVP_MD.
3123 *) Make EVP_MD routines take a context pointer instead of just the
3124 md_data void pointer.
3127 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3128 that the digest can only process a single chunk of data
3129 (typically because it is provided by a piece of
3130 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3131 is only going to provide a single chunk of data, and hence the
3132 framework needn't accumulate the data for oneshot drivers.
3135 *) As with "ERR", make it possible to replace the underlying "ex_data"
3136 functions. This change also alters the storage and management of global
3137 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3138 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3139 index counters. The API functions that use this state have been changed
3140 to take a "class_index" rather than pointers to the class's local STACK
3141 and counter, and there is now an API function to dynamically create new
3142 classes. This centralisation allows us to (a) plug a lot of the
3143 thread-safety problems that existed, and (b) makes it possible to clean
3144 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3145 such data would previously have always leaked in application code and
3146 workarounds were in place to make the memory debugging turn a blind eye
3147 to it. Application code that doesn't use this new function will still
3148 leak as before, but their memory debugging output will announce it now
3149 rather than letting it slide.
3151 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3152 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3153 has a return value to indicate success or failure.
3156 *) Make it possible to replace the underlying "ERR" functions such that the
3157 global state (2 LHASH tables and 2 locks) is only used by the "default"
3158 implementation. This change also adds two functions to "get" and "set"
3159 the implementation prior to it being automatically set the first time
3160 any other ERR function takes place. Ie. an application can call "get",
3161 pass the return value to a module it has just loaded, and that module
3162 can call its own "set" function using that value. This means the
3163 module's "ERR" operations will use (and modify) the error state in the
3164 application and not in its own statically linked copy of OpenSSL code.
3167 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3168 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3169 the operation, and provides a more encapsulated way for external code
3170 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3171 to use these functions rather than manually incrementing the counts.
3173 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3176 *) Add EVP test program.
3179 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3182 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3183 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3184 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3185 These allow a CRL to be built without having to access X509_CRL fields
3186 directly. Modify 'ca' application to use new functions.
3189 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3190 bug workarounds. Rollback attack detection is a security feature.
3191 The problem will only arise on OpenSSL servers when TLSv1 is not
3192 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3193 Software authors not wanting to support TLSv1 will have special reasons
3194 for their choice and can explicitly enable this option.
3195 [Bodo Moeller, Lutz Jaenicke]
3197 *) Rationalise EVP so it can be extended: don't include a union of
3198 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3199 (similar to those existing for EVP_CIPHER_CTX).
3204 EVP_MD_CTX_init(&md); /* new function call */
3205 EVP_DigestInit(&md, EVP_sha1());
3206 EVP_DigestUpdate(&md, in, len);
3207 EVP_DigestFinal(&md, out, NULL);
3208 EVP_MD_CTX_cleanup(&md); /* new function call */
3212 *) Make DES key schedule conform to the usual scheme, as well as
3213 correcting its structure. This means that calls to DES functions
3214 now have to pass a pointer to a des_key_schedule instead of a
3215 plain des_key_schedule (which was actually always a pointer
3218 des_key_schedule ks;
3220 des_set_key_checked(..., &ks);
3221 des_ncbc_encrypt(..., &ks, ...);
3223 (Note that a later change renames 'des_...' into 'DES_...'.)
3226 *) Initial reduction of linker bloat: the use of some functions, such as
3227 PEM causes large amounts of unused functions to be linked in due to
3228 poor organisation. For example pem_all.c contains every PEM function
3229 which has a knock on effect of linking in large amounts of (unused)
3230 ASN1 code. Grouping together similar functions and splitting unrelated
3231 functions prevents this.
3234 *) Cleanup of EVP macros.
3237 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3238 correct _ecb suffix.
3241 *) Add initial OCSP responder support to ocsp application. The
3242 revocation information is handled using the text based index
3243 use by the ca application. The responder can either handle
3244 requests generated internally, supplied in files (for example
3245 via a CGI script) or using an internal minimal server.
3248 *) Add configuration choices to get zlib compression for TLS.
3251 *) Changes to Kerberos SSL for RFC 2712 compliance:
3252 1. Implemented real KerberosWrapper, instead of just using
3253 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3254 2. Implemented optional authenticator field of KerberosWrapper.
3256 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3257 and authenticator structs; see crypto/krb5/.
3259 Generalized Kerberos calls to support multiple Kerberos libraries.
3260 [Vern Staats <staatsvr@asc.hpc.mil>,
3261 Jeffrey Altman <jaltman@columbia.edu>
3262 via Richard Levitte]
3264 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3265 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3266 values for each of the key sizes rather than having just
3267 parameters (and 'speed' generating keys each time).
3270 *) Speed up EVP routines.
3273 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3274 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3275 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3276 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3278 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3279 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3280 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3283 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3285 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3288 *) Added the OS2-EMX target.
3289 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3291 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3292 to support NCONF routines in extension code. New function CONF_set_nconf()
3293 to allow functions which take an NCONF to also handle the old LHASH
3294 structure: this means that the old CONF compatible routines can be
3295 retained (in particular wrt extensions) without having to duplicate the
3296 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3299 *) Enhance the general user interface with mechanisms for inner control
3300 and with possibilities to have yes/no kind of prompts.
3303 *) Change all calls to low level digest routines in the library and
3304 applications to use EVP. Add missing calls to HMAC_cleanup() and
3305 don't assume HMAC_CTX can be copied using memcpy().
3306 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3308 *) Add the possibility to control engines through control names but with
3309 arbitrary arguments instead of just a string.
3310 Change the key loaders to take a UI_METHOD instead of a callback
3311 function pointer. NOTE: this breaks binary compatibility with earlier
3312 versions of OpenSSL [engine].
3313 Adapt the nCipher code for these new conditions and add a card insertion
3317 *) Enhance the general user interface with mechanisms to better support
3318 dialog box interfaces, application-defined prompts, the possibility
3319 to use defaults (for example default passwords from somewhere else)
3320 and interrupts/cancellations.
3323 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3324 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3327 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3328 tidy up some unnecessarily weird code in 'sk_new()').
3329 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3331 *) Change the key loading routines for ENGINEs to use the same kind
3332 callback (pem_password_cb) as all other routines that need this
3336 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3337 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3338 than this minimum value is recommended.
3341 *) New random seeder for OpenVMS, using the system process statistics
3342 that are easily reachable.
3345 *) Windows apparently can't transparently handle global
3346 variables defined in DLLs. Initialisations such as:
3348 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3350 wont compile. This is used by the any applications that need to
3351 declare their own ASN1 modules. This was fixed by adding the option
3352 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3353 needed for static libraries under Win32.
3356 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3357 setting of purpose and trust fields. New X509_STORE trust and
3358 purpose functions and tidy up setting in other SSL functions.
3361 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3362 structure. These are inherited by X509_STORE_CTX when it is
3363 initialised. This allows various defaults to be set in the
3364 X509_STORE structure (such as flags for CRL checking and custom
3365 purpose or trust settings) for functions which only use X509_STORE_CTX
3366 internally such as S/MIME.
3368 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3369 trust settings if they are not set in X509_STORE. This allows X509_STORE
3370 purposes and trust (in S/MIME for example) to override any set by default.
3372 Add command line options for CRL checking to smime, s_client and s_server
3376 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3377 are set then the CRL is looked up in the X509_STORE structure and
3378 its validity and signature checked, then if the certificate is found
3379 in the CRL the verify fails with a revoked error.
3381 Various new CRL related callbacks added to X509_STORE_CTX structure.
3383 Command line options added to 'verify' application to support this.
3385 This needs some additional work, such as being able to handle multiple
3386 CRLs with different times, extension based lookup (rather than just
3387 by subject name) and ultimately more complete V2 CRL extension
3391 *) Add a general user interface API (crypto/ui/). This is designed
3392 to replace things like des_read_password and friends (backward
3393 compatibility functions using this new API are provided).
3394 The purpose is to remove prompting functions from the DES code
3395 section as well as provide for prompting through dialog boxes in
3396 a window system and the like.
3399 *) Add "ex_data" support to ENGINE so implementations can add state at a
3400 per-structure level rather than having to store it globally.
3403 *) Make it possible for ENGINE structures to be copied when retrieved by
3404 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3405 This causes the "original" ENGINE structure to act like a template,
3406 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3407 operational state can be localised to each ENGINE structure, despite the
3408 fact they all share the same "methods". New ENGINE structures returned in
3409 this case have no functional references and the return value is the single
3410 structural reference. This matches the single structural reference returned
3411 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3415 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3416 needs to match any other type at all we need to manually clear the
3420 *) Changes to the "openssl engine" utility to include;
3421 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3422 about an ENGINE's available control commands.
3423 - executing control commands from command line arguments using the
3424 '-pre' and '-post' switches. '-post' is only used if '-t' is
3425 specified and the ENGINE is successfully initialised. The syntax for
3426 the individual commands are colon-separated, for example;
3427 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3430 *) New dynamic control command support for ENGINEs. ENGINEs can now
3431 declare their own commands (numbers), names (strings), descriptions,
3432 and input types for run-time discovery by calling applications. A
3433 subset of these commands are implicitly classed as "executable"
3434 depending on their input type, and only these can be invoked through
3435 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3436 can be based on user input, config files, etc). The distinction is
3437 that "executable" commands cannot return anything other than a boolean
3438 result and can only support numeric or string input, whereas some
3439 discoverable commands may only be for direct use through
3440 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3441 pointers, or other custom uses. The "executable" commands are to
3442 support parameterisations of ENGINE behaviour that can be
3443 unambiguously defined by ENGINEs and used consistently across any
3444 OpenSSL-based application. Commands have been added to all the
3445 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3446 control over shared-library paths without source code alterations.
3449 *) Changed all ENGINE implementations to dynamically allocate their
3450 ENGINEs rather than declaring them statically. Apart from this being
3451 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3452 this also allows the implementations to compile without using the
3453 internal engine_int.h header.
3456 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3457 'const' value. Any code that should be able to modify a RAND_METHOD
3458 should already have non-const pointers to it (ie. they should only
3459 modify their own ones).
3462 *) Made a variety of little tweaks to the ENGINE code.
3463 - "atalla" and "ubsec" string definitions were moved from header files
3464 to C code. "nuron" string definitions were placed in variables
3465 rather than hard-coded - allowing parameterisation of these values
3466 later on via ctrl() commands.
3467 - Removed unused "#if 0"'d code.
3468 - Fixed engine list iteration code so it uses ENGINE_free() to release
3469 structural references.
3470 - Constified the RAND_METHOD element of ENGINE structures.
3471 - Constified various get/set functions as appropriate and added
3472 missing functions (including a catch-all ENGINE_cpy that duplicates
3473 all ENGINE values onto a new ENGINE except reference counts/state).
3474 - Removed NULL parameter checks in get/set functions. Setting a method
3475 or function to NULL is a way of cancelling out a previously set
3476 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3477 and doesn't justify the extra error symbols and code.
3478 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3479 flags from engine_int.h to engine.h.
3480 - Changed prototypes for ENGINE handler functions (init(), finish(),
3481 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3484 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3485 to the algorithm using long division. The binary algorithm can be
3486 used only if the modulus is odd. On 32-bit systems, it is faster
3487 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3488 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3489 up to 450 bits. In 64-bit environments, the binary algorithm
3490 appears to be advantageous for much longer moduli; here we use it
3491 for moduli up to 2048 bits.
3494 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3495 could not support the combine flag in choice fields.
3498 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3499 extensions from a certificate request to the certificate.
3502 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3503 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3504 file: this allows the display of the certificate about to be
3505 signed to be customised, to allow certain fields to be included
3506 or excluded and extension details. The old system didn't display
3507 multicharacter strings properly, omitted fields not in the policy
3508 and couldn't display additional details such as extensions.
3511 *) Function EC_POINTs_mul for multiple scalar multiplication
3512 of an arbitrary number of elliptic curve points
3513 \sum scalars[i]*points[i],
3514 optionally including the generator defined for the EC_GROUP:
3515 scalar*generator + \sum scalars[i]*points[i].
3517 EC_POINT_mul is a simple wrapper function for the typical case
3518 that the point list has just one item (besides the optional
3522 *) First EC_METHODs for curves over GF(p):
3524 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3525 operations and provides various method functions that can also
3526 operate with faster implementations of modular arithmetic.
3528 EC_GFp_mont_method() reuses most functions that are part of
3529 EC_GFp_simple_method, but uses Montgomery arithmetic.
3531 [Bodo Moeller; point addition and point doubling
3532 implementation directly derived from source code provided by
3533 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3535 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3536 crypto/ec/ec_lib.c):
3538 Curves are EC_GROUP objects (with an optional group generator)
3539 based on EC_METHODs that are built into the library.
3541 Points are EC_POINT objects based on EC_GROUP objects.
3543 Most of the framework would be able to handle curves over arbitrary
3544 finite fields, but as there are no obvious types for fields other
3545 than GF(p), some functions are limited to that for now.
3548 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3549 that the file contains a complete HTTP response.
3552 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3553 change the def and num file printf format specifier from "%-40sXXX"
3554 to "%-39s XXX". The latter will always guarantee a space after the
3555 field while the former will cause them to run together if the field
3556 is 40 of more characters long.
3559 *) Constify the cipher and digest 'method' functions and structures
3560 and modify related functions to take constant EVP_MD and EVP_CIPHER
3564 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3565 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3568 *) Modify EVP_Digest*() routines so they now return values. Although the
3569 internal software routines can never fail additional hardware versions
3573 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3575 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3576 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3579 ERR_R_NESTED_ASN1_ERROR
3581 ERR_R_MISSING_ASN1_EOS
3582 were 4 .. 9, conflicting with
3583 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3585 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3586 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3588 Add new error code 'ERR_R_INTERNAL_ERROR'.
3591 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3595 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3596 sets the subject name for a new request or supersedes the
3597 subject name in a given request. Formats that can be parsed are
3598 'CN=Some Name, OU=myOU, C=IT'
3600 'CN=Some Name/OU=myOU/C=IT'.
3602 Add options '-batch' and '-verbose' to 'openssl req'.
3603 [Massimiliano Pala <madwolf@hackmasters.net>]
3605 *) Introduce the possibility to access global variables through
3606 functions on platform were that's the best way to handle exporting
3607 global variables in shared libraries. To enable this functionality,
3608 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3609 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3610 is normally done by Configure or something similar).
3612 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3613 in the source file (foo.c) like this:
3615 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3616 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3618 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3619 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3621 OPENSSL_DECLARE_GLOBAL(int,foo);
3622 #define foo OPENSSL_GLOBAL_REF(foo)
3623 OPENSSL_DECLARE_GLOBAL(double,bar);
3624 #define bar OPENSSL_GLOBAL_REF(bar)
3626 The #defines are very important, and therefore so is including the
3627 header file everywhere where the defined globals are used.
3629 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3630 of ASN.1 items, but that structure is a bit different.
3632 The largest change is in util/mkdef.pl which has been enhanced with
3633 better and easier to understand logic to choose which symbols should
3634 go into the Windows .def files as well as a number of fixes and code
3635 cleanup (among others, algorithm keywords are now sorted
3636 lexicographically to avoid constant rewrites).
3639 *) In BN_div() keep a copy of the sign of 'num' before writing the
3640 result to 'rm' because if rm==num the value will be overwritten
3641 and produce the wrong result if 'num' is negative: this caused
3642 problems with BN_mod() and BN_nnmod().
3645 *) Function OCSP_request_verify(). This checks the signature on an
3646 OCSP request and verifies the signer certificate. The signer
3647 certificate is just checked for a generic purpose and OCSP request
3651 *) Add OCSP_check_validity() function to check the validity of OCSP
3652 responses. OCSP responses are prepared in real time and may only
3653 be a few seconds old. Simply checking that the current time lies
3654 between thisUpdate and nextUpdate max reject otherwise valid responses
3655 caused by either OCSP responder or client clock inaccuracy. Instead
3656 we allow thisUpdate and nextUpdate to fall within a certain period of
3657 the current time. The age of the response can also optionally be
3658 checked. Two new options -validity_period and -status_age added to
3662 *) If signature or public key algorithm is unrecognized print out its
3663 OID rather that just UNKNOWN.
3666 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3667 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3668 ID to be generated from the issuer certificate alone which can then be
3669 passed to OCSP_id_issuer_cmp().
3672 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3673 ASN1 modules to export functions returning ASN1_ITEM pointers
3674 instead of the ASN1_ITEM structures themselves. This adds several
3675 new macros which allow the underlying ASN1 function/structure to
3676 be accessed transparently. As a result code should not use ASN1_ITEM
3677 references directly (such as &X509_it) but instead use the relevant
3678 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3679 use of the new ASN1 code on platforms where exporting structures
3680 is problematical (for example in shared libraries) but exporting
3681 functions returning pointers to structures is not.
3684 *) Add support for overriding the generation of SSL/TLS session IDs.
3685 These callbacks can be registered either in an SSL_CTX or per SSL.
3686 The purpose of this is to allow applications to control, if they wish,
3687 the arbitrary values chosen for use as session IDs, particularly as it
3688 can be useful for session caching in multiple-server environments. A
3689 command-line switch for testing this (and any client code that wishes
3690 to use such a feature) has been added to "s_server".
3691 [Geoff Thorpe, Lutz Jaenicke]
3693 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3694 of the form '#if defined(...) || defined(...) || ...' and
3695 '#if !defined(...) && !defined(...) && ...'. This also avoids
3696 the growing number of special cases it was previously handling.
3699 *) Make all configuration macros available for application by making
3700 sure they are available in opensslconf.h, by giving them names starting
3701 with "OPENSSL_" to avoid conflicts with other packages and by making
3702 sure e_os2.h will cover all platform-specific cases together with
3704 Additionally, it is now possible to define configuration/platform-
3705 specific names (called "system identities"). In the C code, these
3706 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3707 macro with the name beginning with "OPENSSL_SYS_", which is determined
3708 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3712 *) New option -set_serial to 'req' and 'x509' this allows the serial
3713 number to use to be specified on the command line. Previously self
3714 signed certificates were hard coded with serial number 0 and the
3715 CA options of 'x509' had to use a serial number in a file which was
3719 *) New options to 'ca' utility to support V2 CRL entry extensions.
3720 Currently CRL reason, invalidity date and hold instruction are
3721 supported. Add new CRL extensions to V3 code and some new objects.
3724 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3725 disable standard block padding (aka PKCS#5 padding) in the EVP
3726 API, which was previously mandatory. This means that the data is
3727 not padded in any way and so the total length much be a multiple
3728 of the block size, otherwise an error occurs.
3731 *) Initial (incomplete) OCSP SSL support.
3734 *) New function OCSP_parse_url(). This splits up a URL into its host,
3735 port and path components: primarily to parse OCSP URLs. New -url
3736 option to ocsp utility.
3739 *) New nonce behavior. The return value of OCSP_check_nonce() now
3740 reflects the various checks performed. Applications can decide
3741 whether to tolerate certain situations such as an absent nonce
3742 in a response when one was present in a request: the ocsp application
3743 just prints out a warning. New function OCSP_add1_basic_nonce()
3744 this is to allow responders to include a nonce in a response even if
3745 the request is nonce-less.
3748 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3749 skipped when using openssl x509 multiple times on a single input file,
3750 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3753 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3754 set string type: to handle setting ASN1_TIME structures. Fix ca
3755 utility to correctly initialize revocation date of CRLs.
3758 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3759 the clients preferred ciphersuites and rather use its own preferences.
3760 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3761 Internet Explorer by ensuring unchanged hash method during stepup.
3762 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3765 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3766 to aes and add a new 'exist' option to print out symbols that don't
3770 *) Additional options to ocsp utility to allow flags to be set and
3771 additional certificates supplied.
3774 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3775 OCSP client a number of certificate to only verify the response
3779 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3780 handle the new API. Currently only ECB, CBC modes supported. Add new
3783 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3784 Encryption Standard (AES) Ciphersuites for Transport Layer
3785 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3786 not enabled by default and were not part of the "ALL" ciphersuite
3787 alias because they were not yet official; they could be
3788 explicitly requested by specifying the "AESdraft" ciphersuite
3789 group alias. In the final release of OpenSSL 0.9.7, the group
3790 alias is called "AES" and is part of "ALL".)
3791 [Ben Laurie, Steve Henson, Bodo Moeller]
3793 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3794 request to response.
3797 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3798 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3799 extract information from a certificate request. OCSP_response_create()
3800 creates a response and optionally adds a basic response structure.
3801 OCSP_basic_add1_status() adds a complete single response to a basic
3802 response and returns the OCSP_SINGLERESP structure just added (to allow
3803 extensions to be included for example). OCSP_basic_add1_cert() adds a
3804 certificate to a basic response and OCSP_basic_sign() signs a basic
3805 response with various flags. New helper functions ASN1_TIME_check()
3806 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3807 (converts ASN1_TIME to GeneralizedTime).
3810 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3811 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3812 structure from a certificate. X509_pubkey_digest() digests the public_key
3813 contents: this is used in various key identifiers.
3816 *) Make sk_sort() tolerate a NULL argument.
3817 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3819 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3820 passed by the function are trusted implicitly. If any of them signed the
3821 response then it is assumed to be valid and is not verified.
3824 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3825 to data. This was previously part of the PKCS7 ASN1 code. This
3826 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3827 [Steve Henson, reported by Kenneth R. Robinette
3828 <support@securenetterm.com>]
3830 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3831 routines: without these tracing memory leaks is very painful.
3832 Fix leaks in PKCS12 and PKCS7 routines.
3835 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3836 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3837 effectively meant GeneralizedTime would never be used. Now it
3838 is initialised to -1 but X509_time_adj() now has to check the value
3839 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3840 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3841 [Steve Henson, reported by Kenneth R. Robinette
3842 <support@securenetterm.com>]
3844 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3845 result in a zero length in the ASN1_INTEGER structure which was
3846 not consistent with the structure when d2i_ASN1_INTEGER() was used
3847 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3848 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3849 where it did not print out a minus for negative ASN1_INTEGER.
3852 *) Add summary printout to ocsp utility. The various functions which
3853 convert status values to strings have been renamed to:
3854 OCSP_response_status_str(), OCSP_cert_status_str() and
3855 OCSP_crl_reason_str() and are no longer static. New options
3856 to verify nonce values and to disable verification. OCSP response
3857 printout format cleaned up.
3860 *) Add additional OCSP certificate checks. These are those specified
3861 in RFC2560. This consists of two separate checks: the CA of the
3862 certificate being checked must either be the OCSP signer certificate
3863 or the issuer of the OCSP signer certificate. In the latter case the
3864 OCSP signer certificate must contain the OCSP signing extended key
3865 usage. This check is performed by attempting to match the OCSP
3866 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3867 in the OCSP_CERTID structures of the response.
3870 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3871 and related routines. This uses the standard OpenSSL certificate
3872 verify routines to perform initial checks (just CA validity) and
3873 to obtain the certificate chain. Then additional checks will be
3874 performed on the chain. Currently the root CA is checked to see
3875 if it is explicitly trusted for OCSP signing. This is used to set
3876 a root CA as a global signing root: that is any certificate that
3877 chains to that CA is an acceptable OCSP signing certificate.
3880 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3881 extensions from a separate configuration file.
3882 As when reading extensions from the main configuration file,
3883 the '-extensions ...' option may be used for specifying the
3885 [Massimiliano Pala <madwolf@comune.modena.it>]
3887 *) New OCSP utility. Allows OCSP requests to be generated or
3888 read. The request can be sent to a responder and the output
3889 parsed, outputed or printed in text form. Not complete yet:
3890 still needs to check the OCSP response validity.
3893 *) New subcommands for 'openssl ca':
3894 'openssl ca -status <serial>' prints the status of the cert with
3895 the given serial number (according to the index file).
3896 'openssl ca -updatedb' updates the expiry status of certificates
3898 [Massimiliano Pala <madwolf@comune.modena.it>]
3900 *) New '-newreq-nodes' command option to CA.pl. This is like
3901 '-newreq', but calls 'openssl req' with the '-nodes' option
3902 so that the resulting key is not encrypted.
3903 [Damien Miller <djm@mindrot.org>]
3905 *) New configuration for the GNU Hurd.
3906 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3908 *) Initial code to implement OCSP basic response verify. This
3909 is currently incomplete. Currently just finds the signer's
3910 certificate and verifies the signature on the response.
3913 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3914 value of OPENSSLDIR. This is available via the new '-d' option
3915 to 'openssl version', and is also included in 'openssl version -a'.
3918 *) Allowing defining memory allocation callbacks that will be given
3919 file name and line number information in additional arguments
3920 (a const char* and an int). The basic functionality remains, as
3921 well as the original possibility to just replace malloc(),
3922 realloc() and free() by functions that do not know about these
3923 additional arguments. To register and find out the current
3924 settings for extended allocation functions, the following
3925 functions are provided:
3927 CRYPTO_set_mem_ex_functions
3928 CRYPTO_set_locked_mem_ex_functions
3929 CRYPTO_get_mem_ex_functions
3930 CRYPTO_get_locked_mem_ex_functions
3932 These work the same way as CRYPTO_set_mem_functions and friends.
3933 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3934 extended allocation function is enabled.
3935 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3936 a conventional allocation function is enabled.
3937 [Richard Levitte, Bodo Moeller]
3939 *) Finish off removing the remaining LHASH function pointer casts.
3940 There should no longer be any prototype-casting required when using
3941 the LHASH abstraction, and any casts that remain are "bugs". See
3942 the callback types and macros at the head of lhash.h for details
3943 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3946 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3947 If /dev/[u]random devices are not available or do not return enough
3948 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3950 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3951 /etc/entropy will be queried once each in this sequence, quering stops
3952 when enough entropy was collected without querying more sockets.
3955 *) Change the Unix RAND_poll() variant to be able to poll several
3956 random devices, as specified by DEVRANDOM, until a sufficient amount
3957 of data has been collected. We spend at most 10 ms on each file
3958 (select timeout) and read in non-blocking mode. DEVRANDOM now
3959 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3960 (previously it was just the string "/dev/urandom"), so on typical
3961 platforms the 10 ms delay will never occur.
3962 Also separate out the Unix variant to its own file, rand_unix.c.
3963 For VMS, there's a currently-empty rand_vms.c.
3966 *) Move OCSP client related routines to ocsp_cl.c. These
3967 provide utility functions which an application needing
3968 to issue a request to an OCSP responder and analyse the
3969 response will typically need: as opposed to those which an
3970 OCSP responder itself would need which will be added later.
3972 OCSP_request_sign() signs an OCSP request with an API similar
3973 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3974 response. OCSP_response_get1_basic() extracts basic response
3975 from response. OCSP_resp_find_status(): finds and extracts status
3976 information from an OCSP_CERTID structure (which will be created
3977 when the request structure is built). These are built from lower
3978 level functions which work on OCSP_SINGLERESP structures but
3979 wont normally be used unless the application wishes to examine
3980 extensions in the OCSP response for example.
3982 Replace nonce routines with a pair of functions.
3983 OCSP_request_add1_nonce() adds a nonce value and optionally
3984 generates a random value. OCSP_check_nonce() checks the
3985 validity of the nonce in an OCSP response.
3988 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3989 This doesn't copy the supplied OCSP_CERTID and avoids the
3990 need to free up the newly created id. Change return type
3991 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3992 This can then be used to add extensions to the request.
3993 Deleted OCSP_request_new(), since most of its functionality
3994 is now in OCSP_REQUEST_new() (and the case insensitive name
3995 clash) apart from the ability to set the request name which
3996 will be added elsewhere.
3999 *) Update OCSP API. Remove obsolete extensions argument from
4000 various functions. Extensions are now handled using the new
4001 OCSP extension code. New simple OCSP HTTP function which
4002 can be used to send requests and parse the response.
4005 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4006 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4007 uses the special reorder version of SET OF to sort the attributes
4008 and reorder them to match the encoded order. This resolves a long
4009 standing problem: a verify on a PKCS7 structure just after signing
4010 it used to fail because the attribute order did not match the
4011 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4012 it uses the received order. This is necessary to tolerate some broken
4013 software that does not order SET OF. This is handled by encoding
4014 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4015 to produce the required SET OF.
4018 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4019 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4020 files to get correct declarations of the ASN.1 item variables.
4023 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4024 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4025 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4026 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4027 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4028 ASN1_ITEM and no wrapper functions.
4031 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4032 replace the old function pointer based I/O routines. Change most of
4033 the *_d2i_bio() and *_d2i_fp() functions to use these.
4036 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4037 lines, recognice more "algorithms" that can be deselected, and make
4038 it complain about algorithm deselection that isn't recognised.
4041 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4042 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4043 to use new functions. Add NO_ASN1_OLD which can be set to remove
4044 some old style ASN1 functions: this can be used to determine if old
4045 code will still work when these eventually go away.
4048 *) New extension functions for OCSP structures, these follow the
4049 same conventions as certificates and CRLs.
4052 *) New function X509V3_add1_i2d(). This automatically encodes and
4053 adds an extension. Its behaviour can be customised with various
4054 flags to append, replace or delete. Various wrappers added for
4055 certifcates and CRLs.
4058 *) Fix to avoid calling the underlying ASN1 print routine when
4059 an extension cannot be parsed. Correct a typo in the
4060 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4063 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4064 entries for variables.
4067 *) Add functionality to apps/openssl.c for detecting locking
4068 problems: As the program is single-threaded, all we have
4069 to do is register a locking callback using an array for
4070 storing which locks are currently held by the program.
4073 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4074 SSL_get_ex_data_X509_STORE_idx(), which is used in
4075 ssl_verify_cert_chain() and thus can be called at any time
4076 during TLS/SSL handshakes so that thread-safety is essential.
4077 Unfortunately, the ex_data design is not at all suited
4078 for multi-threaded use, so it probably should be abolished.
4081 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4082 [Broadcom, tweaked and integrated by Geoff Thorpe]
4084 *) Move common extension printing code to new function
4085 X509V3_print_extensions(). Reorganise OCSP print routines and
4086 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4089 *) New function X509_signature_print() to remove duplication in some
4093 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4094 set (this was treated exactly the same as SET OF previously). This
4095 is used to reorder the STACK representing the structure to match the
4096 encoding. This will be used to get round a problem where a PKCS7
4097 structure which was signed could not be verified because the STACK
4098 order did not reflect the encoded order.
4101 *) Reimplement the OCSP ASN1 module using the new code.
4104 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4105 for its ASN1 operations. The old style function pointers still exist
4106 for now but they will eventually go away.
4109 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4110 completely replaces the old ASN1 functionality with a table driven
4111 encoder and decoder which interprets an ASN1_ITEM structure describing
4112 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4113 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4114 has also been converted to the new form.
4117 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4118 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4119 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4120 for negative moduli.
4123 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4124 of not touching the result's sign bit.
4127 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4131 *) Changed the LHASH code to use prototypes for callbacks, and created
4132 macros to declare and implement thin (optionally static) functions
4133 that provide type-safety and avoid function pointer casting for the
4134 type-specific callbacks.
4137 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4139 [Veers Staats <staatsvr@asc.hpc.mil>,
4140 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4142 *) Reformat the FAQ so the different questions and answers can be divided
4143 in sections depending on the subject.
4146 *) Have the zlib compression code load ZLIB.DLL dynamically under
4150 *) New function BN_mod_sqrt for computing square roots modulo a prime
4151 (using the probabilistic Tonelli-Shanks algorithm unless
4152 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4153 be handled deterministically).
4154 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4156 *) Make BN_mod_inverse faster by explicitly handling small quotients
4157 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4158 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4161 *) New function BN_kronecker.
4164 *) Fix BN_gcd so that it works on negative inputs; the result is
4165 positive unless both parameters are zero.
4166 Previously something reasonably close to an infinite loop was
4167 possible because numbers could be growing instead of shrinking
4168 in the implementation of Euclid's algorithm.
4171 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4172 sign of the number in question.
4174 Fix BN_is_word(a,w) to work correctly for w == 0.
4176 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4177 because its test if the absolute value of 'a' equals 'w'.
4178 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4179 it exists mostly for use in the implementations of BN_is_zero(),
4180 BN_is_one(), and BN_is_word().
4183 *) New function BN_swap.
4186 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4187 the exponentiation functions are more likely to produce reasonable
4188 results on negative inputs.
4191 *) Change BN_mod_mul so that the result is always non-negative.
4192 Previously, it could be negative if one of the factors was negative;
4193 I don't think anyone really wanted that behaviour.
4196 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4197 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4198 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4199 and add new functions:
4208 BN_mod_lshift1_quick
4212 These functions always generate non-negative results.
4214 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4215 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4217 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4218 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4219 be reduced modulo m.
4220 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4223 The following entry accidentily appeared in the CHANGES file
4224 distributed with OpenSSL 0.9.7. The modifications described in
4225 it do *not* apply to OpenSSL 0.9.7.
4227 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4228 was actually never needed) and in BN_mul(). The removal in BN_mul()
4229 required a small change in bn_mul_part_recursive() and the addition
4230 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4231 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4232 bn_sub_words() and bn_add_words() except they take arrays with
4237 *) In 'openssl passwd', verify passwords read from the terminal
4238 unless the '-salt' option is used (which usually means that
4239 verification would just waste user's time since the resulting
4240 hash is going to be compared with some given password hash)
4241 or the new '-noverify' option is used.
4243 This is an incompatible change, but it does not affect
4244 non-interactive use of 'openssl passwd' (passwords on the command
4245 line, '-stdin' option, '-in ...' option) and thus should not
4249 *) Remove all references to RSAref, since there's no more need for it.
4252 *) Make DSO load along a path given through an environment variable
4253 (SHLIB_PATH) with shl_load().
4256 *) Constify the ENGINE code as a result of BIGNUM constification.
4257 Also constify the RSA code and most things related to it. In a
4258 few places, most notable in the depth of the ASN.1 code, ugly
4259 casts back to non-const were required (to be solved at a later
4263 *) Make it so the openssl application has all engines loaded by default.
4266 *) Constify the BIGNUM routines a little more.
4269 *) Add the following functions:
4271 ENGINE_load_cswift()
4273 ENGINE_load_atalla()
4275 ENGINE_load_builtin_engines()
4277 That way, an application can itself choose if external engines that
4278 are built-in in OpenSSL shall ever be used or not. The benefit is
4279 that applications won't have to be linked with libdl or other dso
4280 libraries unless it's really needed.
4282 Changed 'openssl engine' to load all engines on demand.
4283 Changed the engine header files to avoid the duplication of some
4284 declarations (they differed!).
4287 *) 'openssl engine' can now list capabilities.
4290 *) Better error reporting in 'openssl engine'.
4293 *) Never call load_dh_param(NULL) in s_server.
4296 *) Add engine application. It can currently list engines by name and
4297 identity, and test if they are actually available.
4300 *) Improve RPM specification file by forcing symbolic linking and making
4301 sure the installed documentation is also owned by root.root.
4302 [Damien Miller <djm@mindrot.org>]
4304 *) Give the OpenSSL applications more possibilities to make use of
4305 keys (public as well as private) handled by engines.
4308 *) Add OCSP code that comes from CertCo.
4311 *) Add VMS support for the Rijndael code.
4314 *) Added untested support for Nuron crypto accelerator.
4317 *) Add support for external cryptographic devices. This code was
4318 previously distributed separately as the "engine" branch.
4319 [Geoff Thorpe, Richard Levitte]
4321 *) Rework the filename-translation in the DSO code. It is now possible to
4322 have far greater control over how a "name" is turned into a filename
4323 depending on the operating environment and any oddities about the
4324 different shared library filenames on each system.
4327 *) Support threads on FreeBSD-elf in Configure.
4330 *) Fix for SHA1 assembly problem with MASM: it produces
4331 warnings about corrupt line number information when assembling
4332 with debugging information. This is caused by the overlapping
4334 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4337 NCONF_get_number() has no error checking at all. As a replacement,
4338 NCONF_get_number_e() is defined (_e for "error checking") and is
4339 promoted strongly. The old NCONF_get_number is kept around for
4340 binary backward compatibility.
4341 Make it possible for methods to load from something other than a BIO,
4342 by providing a function pointer that is given a name instead of a BIO.
4343 For example, this could be used to load configuration data from an
4347 *) Fix for non blocking accept BIOs. Added new I/O special reason
4348 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4349 with non blocking I/O was not possible because no retry code was
4350 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4354 *) Added the beginnings of Rijndael support.
4357 *) Fix for bug in DirectoryString mask setting. Add support for
4358 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4359 to allow certificate printing to more controllable, additional
4360 'certopt' option to 'x509' to allow new printing options to be
4364 *) Clean old EAY MD5 hack from e_os.h.
4367 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4369 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4370 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4371 [Joe Orton, Steve Henson]
4373 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4375 *) Fix additional bug revealed by the NISCC test suite:
4377 Stop bug triggering large recursion when presented with
4378 certain ASN.1 tags (CVE-2003-0851)
4381 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4383 *) Fix various bugs revealed by running the NISCC test suite:
4385 Stop out of bounds reads in the ASN1 code when presented with
4386 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4388 If verify callback ignores invalid public key errors don't try to check
4389 certificate signature with the NULL public key.
4393 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4394 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4398 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4399 extra data after the compression methods not only for TLS 1.0
4400 but also for SSL 3.0 (as required by the specification).
4401 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4403 *) Change X509_certificate_type() to mark the key as exported/exportable
4404 when it's 512 *bits* long, not 512 bytes.
4407 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4409 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4410 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4411 a protocol version number mismatch like a decryption error
4412 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4415 *) Turn on RSA blinding by default in the default implementation
4416 to avoid a timing attack. Applications that don't want it can call
4417 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4418 They would be ill-advised to do so in most cases.
4419 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4421 *) Change RSA blinding code so that it works when the PRNG is not
4422 seeded (in this case, the secret RSA exponent is abused as
4423 an unpredictable seed -- if it is not unpredictable, there
4424 is no point in blinding anyway). Make RSA blinding thread-safe
4425 by remembering the creator's thread ID in rsa->blinding and
4426 having all other threads use local one-time blinding factors
4427 (this requires more computation than sharing rsa->blinding, but
4428 avoids excessive locking; and if an RSA object is not shared
4429 between threads, blinding will still be very fast).
4432 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4434 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4435 via timing by performing a MAC computation even if incorrrect
4436 block cipher padding has been found. This is a countermeasure
4437 against active attacks where the attacker has to distinguish
4438 between bad padding and a MAC verification error. (CVE-2003-0078)
4440 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4441 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4442 Martin Vuagnoux (EPFL, Ilion)]
4444 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4446 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4447 memory from it's contents. This is done with a counter that will
4448 place alternating values in each byte. This can be used to solve
4449 two issues: 1) the removal of calls to memset() by highly optimizing
4450 compilers, and 2) cleansing with other values than 0, since those can
4451 be read through on certain media, for example a swap space on disk.
4454 *) Bugfix: client side session caching did not work with external caching,
4455 because the session->cipher setting was not restored when reloading
4456 from the external cache. This problem was masked, when
4457 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4458 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4461 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4462 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4463 [Zeev Lieber <zeev-l@yahoo.com>]
4465 *) Undo an undocumented change introduced in 0.9.6e which caused
4466 repeated calls to OpenSSL_add_all_ciphers() and
4467 OpenSSL_add_all_digests() to be ignored, even after calling
4471 *) Change the default configuration reader to deal with last line not
4472 being properly terminated.
4475 *) Change X509_NAME_cmp() so it applies the special rules on handling
4476 DN values that are of type PrintableString, as well as RDNs of type
4477 emailAddress where the value has the type ia5String.
4478 [stefank@valicert.com via Richard Levitte]
4480 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4481 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4482 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4483 the bitwise-OR of the two for use by the majority of applications
4484 wanting this behaviour, and update the docs. The documented
4485 behaviour and actual behaviour were inconsistent and had been
4486 changing anyway, so this is more a bug-fix than a behavioural
4488 [Geoff Thorpe, diagnosed by Nadav Har'El]
4490 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4491 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4494 *) Fix initialization code race conditions in
4495 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4496 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4497 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4498 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4499 ssl2_get_cipher_by_char(),
4500 ssl3_get_cipher_by_char().
4501 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4503 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4504 the cached sessions are flushed, as the remove_cb() might use ex_data
4505 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4506 (see [openssl.org #212]).
4507 [Geoff Thorpe, Lutz Jaenicke]
4509 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4510 length, instead of the encoding length to d2i_ASN1_OBJECT.
4513 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4515 *) [In 0.9.6g-engine release:]
4516 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4517 [Lynn Gazis <lgazis@rainbow.com>]
4519 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4521 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4522 and get fix the header length calculation.
4523 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4524 Alon Kantor <alonk@checkpoint.com> (and others),
4527 *) Use proper error handling instead of 'assertions' in buffer
4528 overflow checks added in 0.9.6e. This prevents DoS (the
4529 assertions could call abort()).
4530 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4532 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4534 *) Add various sanity checks to asn1_get_length() to reject
4535 the ASN1 length bytes if they exceed sizeof(long), will appear
4536 negative or the content length exceeds the length of the
4538 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4540 *) Fix cipher selection routines: ciphers without encryption had no flags
4541 for the cipher strength set and where therefore not handled correctly
4542 by the selection routines (PR #130).
4545 *) Fix EVP_dsa_sha macro.
4549 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4550 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4551 that was added in OpenSSL 0.9.6d.
4553 As the countermeasure turned out to be incompatible with some
4554 broken SSL implementations, the new option is part of SSL_OP_ALL.
4555 SSL_OP_ALL is usually employed when compatibility with weird SSL
4556 implementations is desired (e.g. '-bugs' option to 's_client' and
4557 's_server'), so the new option is automatically set in many
4561 *) Changes in security patch:
4563 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4564 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4565 Air Force Materiel Command, USAF, under agreement number
4568 *) Add various sanity checks to asn1_get_length() to reject
4569 the ASN1 length bytes if they exceed sizeof(long), will appear
4570 negative or the content length exceeds the length of the
4571 supplied buffer. (CVE-2002-0659)
4572 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4574 *) Assertions for various potential buffer overflows, not known to
4576 [Ben Laurie (CHATS)]
4578 *) Various temporary buffers to hold ASCII versions of integers were
4579 too small for 64 bit platforms. (CVE-2002-0655)
4580 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4582 *) Remote buffer overflow in SSL3 protocol - an attacker could
4583 supply an oversized session ID to a client. (CVE-2002-0656)
4584 [Ben Laurie (CHATS)]
4586 *) Remote buffer overflow in SSL2 protocol - an attacker could
4587 supply an oversized client master key. (CVE-2002-0656)
4588 [Ben Laurie (CHATS)]
4590 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4592 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4593 encoded as NULL) with id-dsa-with-sha1.
4594 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4596 *) Check various X509_...() return values in apps/req.c.
4597 [Nils Larsch <nla@trustcenter.de>]
4599 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4600 an end-of-file condition would erronously be flagged, when the CRLF
4601 was just at the end of a processed block. The bug was discovered when
4602 processing data through a buffering memory BIO handing the data to a
4603 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4604 <ptsekov@syntrex.com> and Nedelcho Stanev.
4607 *) Implement a countermeasure against a vulnerability recently found
4608 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4609 before application data chunks to avoid the use of known IVs
4610 with data potentially chosen by the attacker.
4613 *) Fix length checks in ssl3_get_client_hello().
4616 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4617 to prevent ssl3_read_internal() from incorrectly assuming that
4618 ssl3_read_bytes() found application data while handshake
4619 processing was enabled when in fact s->s3->in_read_app_data was
4620 merely automatically cleared during the initial handshake.
4621 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4623 *) Fix object definitions for Private and Enterprise: they were not
4624 recognized in their shortname (=lowercase) representation. Extend
4625 obj_dat.pl to issue an error when using undefined keywords instead
4626 of silently ignoring the problem (Svenning Sorensen
4627 <sss@sss.dnsalias.net>).
4630 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4631 generators, i.e. generators other than 2 and 5. (Previously, the
4632 code did not properly initialise the 'add' and 'rem' values to
4633 BN_generate_prime().)
4635 In the new general case, we do not insist that 'generator' is
4636 actually a primitive root: This requirement is rather pointless;
4637 a generator of the order-q subgroup is just as good, if not
4641 *) Map new X509 verification errors to alerts. Discovered and submitted by
4642 Tom Wu <tom@arcot.com>.
4645 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4646 returning non-zero before the data has been completely received
4647 when using non-blocking I/O.
4648 [Bodo Moeller; problem pointed out by John Hughes]
4650 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4651 [Ben Laurie, Lutz Jaenicke]
4653 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4654 Yoram Zahavi <YoramZ@gilian.com>).
4657 *) Add information about CygWin 1.3 and on, and preserve proper
4658 configuration for the versions before that.
4659 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4661 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4662 check whether we deal with a copy of a session and do not delete from
4663 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4664 <izhar@checkpoint.com>.
4667 *) Do not store session data into the internal session cache, if it
4668 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4669 flag is set). Proposed by Aslam <aslam@funk.com>.
4672 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4676 *) [In 0.9.6d-engine release:]
4677 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4678 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4680 *) Add the configuration target linux-s390x.
4681 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4683 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4684 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4685 variable as an indication that a ClientHello message has been
4686 received. As the flag value will be lost between multiple
4687 invocations of ssl3_accept when using non-blocking I/O, the
4688 function may not be aware that a handshake has actually taken
4689 place, thus preventing a new session from being added to the
4692 To avoid this problem, we now set s->new_session to 2 instead of
4693 using a local variable.
4694 [Lutz Jaenicke, Bodo Moeller]
4696 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4697 if the SSL_R_LENGTH_MISMATCH error is detected.
4698 [Geoff Thorpe, Bodo Moeller]
4700 *) New 'shared_ldflag' column in Configure platform table.
4703 *) Fix EVP_CIPHER_mode macro.
4704 ["Dan S. Camper" <dan@bti.net>]
4706 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4707 type, we must throw them away by setting rr->length to 0.
4708 [D P Chang <dpc@qualys.com>]
4710 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4712 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4713 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4714 worked incorrectly for those cases where range = 10..._2 and
4715 3*range is two bits longer than range.)
4718 *) Only add signing time to PKCS7 structures if it is not already
4722 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4723 OBJ_ld_ce should be OBJ_id_ce.
4724 Also some ip-pda OIDs in crypto/objects/objects.txt were
4725 incorrect (cf. RFC 3039).
4726 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4728 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4729 returns early because it has nothing to do.
4730 [Andy Schneider <andy.schneider@bjss.co.uk>]
4732 *) [In 0.9.6c-engine release:]
4733 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4734 [Andy Schneider <andy.schneider@bjss.co.uk>]
4736 *) [In 0.9.6c-engine release:]
4737 Add support for Cryptographic Appliance's keyserver technology.
4738 (Use engine 'keyclient')
4739 [Cryptographic Appliances and Geoff Thorpe]
4741 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4742 is called via tools/c89.sh because arguments have to be
4743 rearranged (all '-L' options must appear before the first object
4745 [Richard Shapiro <rshapiro@abinitio.com>]
4747 *) [In 0.9.6c-engine release:]
4748 Add support for Broadcom crypto accelerator cards, backported
4750 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4752 *) [In 0.9.6c-engine release:]
4753 Add support for SureWare crypto accelerator cards from
4754 Baltimore Technologies. (Use engine 'sureware')
4755 [Baltimore Technologies and Mark Cox]
4757 *) [In 0.9.6c-engine release:]
4758 Add support for crypto accelerator cards from Accelerated
4759 Encryption Processing, www.aep.ie. (Use engine 'aep')
4760 [AEP Inc. and Mark Cox]
4762 *) Add a configuration entry for gcc on UnixWare.
4763 [Gary Benson <gbenson@redhat.com>]
4765 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4766 messages are stored in a single piece (fixed-length part and
4767 variable-length part combined) and fix various bugs found on the way.
4770 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4771 instead. BIO_gethostbyname() does not know what timeouts are
4772 appropriate, so entries would stay in cache even when they have
4774 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4776 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4777 faced with a pathologically small ClientHello fragment that does
4778 not contain client_version: Instead of aborting with an error,
4779 simply choose the highest available protocol version (i.e.,
4780 TLS 1.0 unless it is disabled). In practice, ClientHello
4781 messages are never sent like this, but this change gives us
4782 strictly correct behaviour at least for TLS.
4785 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4786 never resets s->method to s->ctx->method when called from within
4787 one of the SSL handshake functions.
4788 [Bodo Moeller; problem pointed out by Niko Baric]
4790 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4791 (sent using the client's version number) if client_version is
4792 smaller than the protocol version in use. Also change
4793 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4794 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4795 the client will at least see that alert.
4798 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4802 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4803 client receives HelloRequest while in a handshake.
4804 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4806 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4807 should end in 'break', not 'goto end' which circuments various
4808 cleanups done in state SSL_ST_OK. But session related stuff
4809 must be disabled for SSL_ST_OK in the case that we just sent a
4812 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4813 before just sending a HelloRequest.
4814 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4816 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4817 reveal whether illegal block cipher padding was found or a MAC
4818 verification error occured. (Neither SSLerr() codes nor alerts
4819 are directly visible to potential attackers, but the information
4820 may leak via logfiles.)
4822 Similar changes are not required for the SSL 2.0 implementation
4823 because the number of padding bytes is sent in clear for SSL 2.0,
4824 and the extra bytes are just ignored. However ssl/s2_pkt.c
4825 failed to verify that the purported number of padding bytes is in
4829 *) Add OpenUNIX-8 support including shared libraries
4830 (Boyd Lynn Gerber <gerberb@zenez.com>).
4833 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4834 'wristwatch attack' using huge encoding parameters (cf.
4835 James H. Manger's CRYPTO 2001 paper). Note that the
4836 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4837 encoding parameters and hence was not vulnerable.
4840 *) BN_sqr() bug fix.
4841 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4843 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4844 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4845 followed by modular reduction.
4846 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4848 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4849 equivalent based on BN_pseudo_rand() instead of BN_rand().
4852 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4853 This function was broken, as the check for a new client hello message
4854 to handle SGC did not allow these large messages.
4855 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4858 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4861 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4862 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4865 *) Rework the configuration and shared library support for Tru64 Unix.
4866 The configuration part makes use of modern compiler features and
4867 still retains old compiler behavior for those that run older versions
4868 of the OS. The shared library support part includes a variant that
4869 uses the RPATH feature, and is available through the special
4870 configuration target "alpha-cc-rpath", which will never be selected
4872 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4874 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4875 with the same message size as in ssl3_get_certificate_request().
4876 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4877 messages might inadvertently be reject as too long.
4878 [Petr Lampa <lampa@fee.vutbr.cz>]
4880 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4883 *) Modified SSL library such that the verify_callback that has been set
4884 specificly for an SSL object with SSL_set_verify() is actually being
4885 used. Before the change, a verify_callback set with this function was
4886 ignored and the verify_callback() set in the SSL_CTX at the time of
4887 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4888 to allow the necessary settings.
4891 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4892 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4893 done automatically (in contradiction to the requirements of the C
4894 standard). This made problems when used from OpenSSH.
4897 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4898 dh->length and always used
4900 BN_rand_range(priv_key, dh->p).
4902 BN_rand_range() is not necessary for Diffie-Hellman, and this
4903 specific range makes Diffie-Hellman unnecessarily inefficient if
4904 dh->length (recommended exponent length) is much smaller than the
4905 length of dh->p. We could use BN_rand_range() if the order of
4906 the subgroup was stored in the DH structure, but we only have
4911 BN_rand(priv_key, l, ...)
4913 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4919 RSA_eay_public_encrypt
4920 RSA_eay_private_decrypt
4921 RSA_eay_private_encrypt (signing)
4922 RSA_eay_public_decrypt (signature verification)
4924 (default implementations for RSA_public_encrypt,
4925 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4926 always reject numbers >= n.
4929 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4930 to synchronize access to 'locking_thread'. This is necessary on
4931 systems where access to 'locking_thread' (an 'unsigned long'
4932 variable) is not atomic.
4935 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4936 *before* setting the 'crypto_lock_rand' flag. The previous code had
4937 a race condition if 0 is a valid thread ID.
4938 [Travis Vitek <vitek@roguewave.com>]
4940 *) Add support for shared libraries under Irix.
4941 [Albert Chin-A-Young <china@thewrittenword.com>]
4943 *) Add configuration option to build on Linux on both big-endian and
4945 [Ralf Baechle <ralf@uni-koblenz.de>]
4947 *) Add the possibility to create shared libraries on HP-UX.
4950 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4952 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4953 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4954 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4955 PRNG state recovery was possible based on the output of
4956 one PRNG request appropriately sized to gain knowledge on
4957 'md' followed by enough consecutive 1-byte PRNG requests
4958 to traverse all of 'state'.
4960 1. When updating 'md_local' (the current thread's copy of 'md')
4961 during PRNG output generation, hash all of the previous
4962 'md_local' value, not just the half used for PRNG output.
4964 2. Make the number of bytes from 'state' included into the hash
4965 independent from the number of PRNG bytes requested.
4967 The first measure alone would be sufficient to avoid
4968 Markku-Juhani's attack. (Actually it had never occurred
4969 to me that the half of 'md_local' used for chaining was the
4970 half from which PRNG output bytes were taken -- I had always
4971 assumed that the secret half would be used.) The second
4972 measure makes sure that additional data from 'state' is never
4973 mixed into 'md_local' in small portions; this heuristically
4974 further strengthens the PRNG.
4977 *) Fix crypto/bn/asm/mips3.s.
4980 *) When only the key is given to "enc", the IV is undefined. Print out
4981 an error message in this case.
4984 *) Handle special case when X509_NAME is empty in X509 printing routines.
4987 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4988 positive and less than q.
4991 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4992 used: it isn't thread safe and the add_lock_callback should handle
4994 [Paul Rose <Paul.Rose@bridge.com>]
4996 *) Verify that incoming data obeys the block size in
4997 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5001 [Ulf Möller, Bodo Möller]
5003 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5004 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5005 when fixing the server behaviour for backwards-compatible 'client
5006 hello' messages. (Note that the attack is impractical against
5007 SSL 3.0 and TLS 1.0 anyway because length and version checking
5008 means that the probability of guessing a valid ciphertext is
5009 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5012 Before 0.9.5, the countermeasure (hide the error by generating a
5013 random 'decryption result') did not work properly because
5014 ERR_clear_error() was missing, meaning that SSL_get_error() would
5015 detect the supposedly ignored error.
5017 Both problems are now fixed.
5020 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5021 (previously it was 1024).
5024 *) Fix for compatibility mode trust settings: ignore trust settings
5025 unless some valid trust or reject settings are present.
5028 *) Fix for blowfish EVP: its a variable length cipher.
5031 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5032 parameters in DSA public key structures and return an error in the
5033 DSA routines if parameters are absent.
5036 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5037 in the current directory if neither $RANDFILE nor $HOME was set.
5038 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5039 caused some confusion to Windows users who haven't defined $HOME.
5040 Thus RAND_file_name() is changed again: e_os.h can define a
5041 DEFAULT_HOME, which will be used if $HOME is not set.
5042 For Windows, we use "C:"; on other platforms, we still require
5043 environment variables.
5045 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5046 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5047 having multiple threads call RAND_poll() concurrently.
5050 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5051 combination of a flag and a thread ID variable.
5052 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5053 flag), *other* threads can enter ssleay_add_bytes without obeying
5054 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5055 that they do not hold after the first thread unsets add_do_not_lock).
5058 *) Change bctest again: '-x' expressions are not available in all
5062 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5064 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5065 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5067 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5068 the default extension for executables, if any. Also, make the perl
5069 scripts that use symlink() to test if it really exists and use "cp"
5070 if it doesn't. All this made OpenSSL compilable and installable in
5074 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5075 If SEQUENCE is length is indefinite just set c->slen to the total
5076 amount of data available.
5077 [Steve Henson, reported by shige@FreeBSD.org]
5078 [This change does not apply to 0.9.7.]
5080 *) Change bctest to avoid here-documents inside command substitution
5081 (workaround for FreeBSD /bin/sh bug).
5082 For compatibility with Ultrix, avoid shell functions (introduced
5083 in the bctest version that searches along $PATH).
5086 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5087 with des_encrypt() defined on some operating systems, like Solaris
5091 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5092 On the Importance of Eliminating Errors in Cryptographic
5093 Computations, J. Cryptology 14 (2001) 2, 101-119,
5094 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5097 *) MIPS assembler BIGNUM division bug fix.
5100 *) Disabled incorrect Alpha assembler code.
5103 *) Fix PKCS#7 decode routines so they correctly update the length
5104 after reading an EOC for the EXPLICIT tag.
5106 [This change does not apply to 0.9.7.]
5108 *) Fix bug in PKCS#12 key generation routines. This was triggered
5109 if a 3DES key was generated with a 0 initial byte. Include
5110 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5111 (but broken) behaviour.
5114 *) Enhance bctest to search for a working bc along $PATH and print
5116 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5118 *) Fix memory leaks in err.c: free err_data string if necessary;
5119 don't write to the wrong index in ERR_set_error_data.
5122 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5126 *) Replace rdtsc with _emit statements for VC++ version 5.
5127 [Jeremy Cooper <jeremy@baymoo.org>]
5129 *) Make it possible to reuse SSLv2 sessions.
5132 *) In copy_email() check for >= 0 as a return value for
5133 X509_NAME_get_index_by_NID() since 0 is a valid index.
5134 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5136 *) Avoid coredump with unsupported or invalid public keys by checking if
5137 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5138 PKCS7_verify() fails with non detached data.
5141 *) Don't use getenv in library functions when run as setuid/setgid.
5142 New function OPENSSL_issetugid().
5145 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5146 due to incorrect handling of multi-threading:
5148 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5150 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5152 3. Count how many times MemCheck_off() has been called so that
5153 nested use can be treated correctly. This also avoids
5154 inband-signalling in the previous code (which relied on the
5155 assumption that thread ID 0 is impossible).
5158 *) Add "-rand" option also to s_client and s_server.
5161 *) Fix CPU detection on Irix 6.x.
5162 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5163 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5165 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5168 [This change does not apply to 0.9.7.]
5170 *) Use the cached encoding of an X509_NAME structure rather than
5171 copying it. This is apparently the reason for the libsafe "errors"
5172 but the code is actually correct.
5175 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5176 Bleichenbacher's DSA attack.
5177 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5178 to be set and top=0 forces the highest bit to be set; top=-1 is new
5179 and leaves the highest bit random.
5180 [Ulf Moeller, Bodo Moeller]
5182 *) In the NCONF_...-based implementations for CONF_... queries
5183 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5184 a temporary CONF structure with the data component set to NULL
5185 (which gives segmentation faults in lh_retrieve).
5186 Instead, use NULL for the CONF pointer in CONF_get_string and
5187 CONF_get_number (which may use environment variables) and directly
5188 return NULL from CONF_get_section.
5191 *) Fix potential buffer overrun for EBCDIC.
5194 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5195 keyUsage if basicConstraints absent for a CA.
5198 *) Make SMIME_write_PKCS7() write mail header values with a format that
5199 is more generally accepted (no spaces before the semicolon), since
5200 some programs can't parse those values properly otherwise. Also make
5201 sure BIO's that break lines after each write do not create invalid
5205 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5206 macros previously used would not encode an empty SEQUENCE OF
5207 and break the signature.
5209 [This change does not apply to 0.9.7.]
5211 *) Zero the premaster secret after deriving the master secret in
5215 *) Add some EVP_add_digest_alias registrations (as found in
5216 OpenSSL_add_all_digests()) to SSL_library_init()
5217 aka OpenSSL_add_ssl_algorithms(). This provides improved
5218 compatibility with peers using X.509 certificates
5219 with unconventional AlgorithmIdentifier OIDs.
5222 *) Fix for Irix with NO_ASM.
5223 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5225 *) ./config script fixes.
5226 [Ulf Moeller, Richard Levitte]
5228 *) Fix 'openssl passwd -1'.
5231 *) Change PKCS12_key_gen_asc() so it can cope with non null
5232 terminated strings whose length is passed in the passlen
5233 parameter, for example from PEM callbacks. This was done
5234 by adding an extra length parameter to asc2uni().
5235 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5237 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5238 call failed, free the DSA structure.
5241 *) Fix to uni2asc() to cope with zero length Unicode strings.
5242 These are present in some PKCS#12 files.
5245 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5246 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5247 when writing a 32767 byte record.
5248 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5250 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5251 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5253 (RSA objects have a reference count access to which is protected
5254 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5255 so they are meant to be shared between threads.)
5256 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5257 "Reddie, Steven" <Steven.Reddie@ca.com>]
5259 *) Fix a deadlock in CRYPTO_mem_leaks().
5262 *) Use better test patterns in bntest.
5265 *) rand_win.c fix for Borland C.
5268 *) BN_rshift bugfix for n == 0.
5271 *) Add a 'bctest' script that checks for some known 'bc' bugs
5272 so that 'make test' does not abort just because 'bc' is broken.
5275 *) Store verify_result within SSL_SESSION also for client side to
5276 avoid potential security hole. (Re-used sessions on the client side
5277 always resulted in verify_result==X509_V_OK, not using the original
5278 result of the server certificate verification.)
5281 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5282 SSL3_RT_APPLICATION_DATA, return 0.
5283 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5287 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5288 releases, have been re-implemented by renaming the previous
5289 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5290 and ssl3_read_internal, respectively, and adding 'peek' parameters
5291 to them. The new ssl[23]_{read,peek} functions are calls to
5292 ssl[23]_read_internal with the 'peek' flag set appropriately.
5293 A 'peek' parameter has also been added to ssl3_read_bytes, which
5294 does the actual work for ssl3_read_internal.
5297 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5298 the method-specific "init()" handler. Also clean up ex_data after
5299 calling the method-specific "finish()" handler. Previously, this was
5300 happening the other way round.
5303 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5304 The previous value, 12, was not always sufficient for BN_mod_exp().
5307 *) Make sure that shared libraries get the internal name engine with
5308 the full version number and not just 0. This should mark the
5309 shared libraries as not backward compatible. Of course, this should
5310 be changed again when we can guarantee backward binary compatibility.
5313 *) Fix typo in get_cert_by_subject() in by_dir.c
5314 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5316 *) Rework the system to generate shared libraries:
5318 - Make note of the expected extension for the shared libraries and
5319 if there is a need for symbolic links from for example libcrypto.so.0
5320 to libcrypto.so.0.9.7. There is extended info in Configure for
5323 - Make as few rebuilds of the shared libraries as possible.
5325 - Still avoid linking the OpenSSL programs with the shared libraries.
5327 - When installing, install the shared libraries separately from the
5331 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5333 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5334 and not in SSL_clear because the latter is also used by the
5335 accept/connect functions; previously, the settings made by
5336 SSL_set_read_ahead would be lost during the handshake.
5337 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5339 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5340 Previously, it would create entries for disableed algorithms no
5344 *) Added several new manual pages for SSL_* function.
5347 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5349 *) In ssl23_get_client_hello, generate an error message when faced
5350 with an initial SSL 3.0/TLS record that is too small to contain the
5351 first two bytes of the ClientHello message, i.e. client_version.
5352 (Note that this is a pathologic case that probably has never happened
5353 in real life.) The previous approach was to use the version number
5354 from the record header as a substitute; but our protocol choice
5355 should not depend on that one because it is not authenticated
5356 by the Finished messages.
5359 *) More robust randomness gathering functions for Windows.
5360 [Jeffrey Altman <jaltman@columbia.edu>]
5362 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5363 not set then we don't setup the error code for issuer check errors
5364 to avoid possibly overwriting other errors which the callback does
5365 handle. If an application does set the flag then we assume it knows
5366 what it is doing and can handle the new informational codes
5370 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5371 a general "ANY" type, as such it should be able to decode anything
5372 including tagged types. However it didn't check the class so it would
5373 wrongly interpret tagged types in the same way as their universal
5374 counterpart and unknown types were just rejected. Changed so that the
5375 tagged and unknown types are handled in the same way as a SEQUENCE:
5376 that is the encoding is stored intact. There is also a new type
5377 "V_ASN1_OTHER" which is used when the class is not universal, in this
5378 case we have no idea what the actual type is so we just lump them all
5382 *) On VMS, stdout may very well lead to a file that is written to
5383 in a record-oriented fashion. That means that every write() will
5384 write a separate record, which will be read separately by the
5385 programs trying to read from it. This can be very confusing.
5387 The solution is to put a BIO filter in the way that will buffer
5388 text until a linefeed is reached, and then write everything a
5389 line at a time, so every record written will be an actual line,
5390 not chunks of lines and not (usually doesn't happen, but I've
5391 seen it once) several lines in one record. BIO_f_linebuffer() is
5394 Currently, it's a VMS-only method, because that's where it has
5395 been tested well enough.
5398 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5399 it can return incorrect results.
5400 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5401 but it was in 0.9.6-beta[12].)
5404 *) Disable the check for content being present when verifying detached
5405 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5406 include zero length content when signing messages.
5409 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5410 BIO_ctrl (for BIO pairs).
5413 *) Add DSO method for VMS.
5416 *) Bug fix: Montgomery multiplication could produce results with the
5420 *) Add RPM specification openssl.spec and modify it to build three
5421 packages. The default package contains applications, application
5422 documentation and run-time libraries. The devel package contains
5423 include files, static libraries and function documentation. The
5424 doc package contains the contents of the doc directory. The original
5425 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5428 *) Add a large number of documentation files for many SSL routines.
5429 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5431 *) Add a configuration entry for Sony News 4.
5432 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5434 *) Don't set the two most significant bits to one when generating a
5435 random number < q in the DSA library.
5438 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5439 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5440 the underlying transport is blocking) if a handshake took place.
5441 (The default behaviour is needed by applications such as s_client
5442 and s_server that use select() to determine when to use SSL_read;
5443 but for applications that know in advance when to expect data, it
5444 just makes things more complicated.)
5447 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5451 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5452 work better on such systems.
5453 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5455 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5456 Update PKCS12_parse() so it copies the friendlyName and the
5457 keyid to the certificates aux info.
5460 *) Fix bug in PKCS7_verify() which caused an infinite loop
5461 if there was more than one signature.
5462 [Sven Uszpelkat <su@celocom.de>]
5464 *) Major change in util/mkdef.pl to include extra information
5465 about each symbol, as well as presentig variables as well
5466 as functions. This change means that there's n more need
5467 to rebuild the .num files when some algorithms are excluded.
5470 *) Allow the verify time to be set by an application,
5471 rather than always using the current time.
5474 *) Phase 2 verify code reorganisation. The certificate
5475 verify code now looks up an issuer certificate by a
5476 number of criteria: subject name, authority key id
5477 and key usage. It also verifies self signed certificates
5478 by the same criteria. The main comparison function is
5479 X509_check_issued() which performs these checks.
5481 Lot of changes were necessary in order to support this
5482 without completely rewriting the lookup code.
5484 Authority and subject key identifier are now cached.
5486 The LHASH 'certs' is X509_STORE has now been replaced
5487 by a STACK_OF(X509_OBJECT). This is mainly because an
5488 LHASH can't store or retrieve multiple objects with
5489 the same hash value.
5491 As a result various functions (which were all internal
5492 use only) have changed to handle the new X509_STORE
5493 structure. This will break anything that messed round
5494 with X509_STORE internally.
5496 The functions X509_STORE_add_cert() now checks for an
5497 exact match, rather than just subject name.
5499 The X509_STORE API doesn't directly support the retrieval
5500 of multiple certificates matching a given criteria, however
5501 this can be worked round by performing a lookup first
5502 (which will fill the cache with candidate certificates)
5503 and then examining the cache for matches. This is probably
5504 the best we can do without throwing out X509_LOOKUP
5505 entirely (maybe later...).
5507 The X509_VERIFY_CTX structure has been enhanced considerably.
5509 All certificate lookup operations now go via a get_issuer()
5510 callback. Although this currently uses an X509_STORE it
5511 can be replaced by custom lookups. This is a simple way
5512 to bypass the X509_STORE hackery necessary to make this
5513 work and makes it possible to use more efficient techniques
5514 in future. A very simple version which uses a simple
5515 STACK for its trusted certificate store is also provided
5516 using X509_STORE_CTX_trusted_stack().
5518 The verify_cb() and verify() callbacks now have equivalents
5519 in the X509_STORE_CTX structure.
5521 X509_STORE_CTX also has a 'flags' field which can be used
5522 to customise the verify behaviour.
5525 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5526 excludes S/MIME capabilities.
5529 *) When a certificate request is read in keep a copy of the
5530 original encoding of the signed data and use it when outputing
5531 again. Signatures then use the original encoding rather than
5532 a decoded, encoded version which may cause problems if the
5533 request is improperly encoded.
5536 *) For consistency with other BIO_puts implementations, call
5537 buffer_write(b, ...) directly in buffer_puts instead of calling
5540 In BIO_puts, increment b->num_write as in BIO_write.
5541 [Peter.Sylvester@EdelWeb.fr]
5543 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5544 BN_zero, we may not return a BIGNUM with an array consisting of
5548 *) Avoid calling abort() from within the library when problems are
5549 detected, except if preprocessor symbols have been defined
5550 (such as REF_CHECK, BN_DEBUG etc.).
5553 *) New openssl application 'rsautl'. This utility can be
5554 used for low level RSA operations. DER public key
5555 BIO/fp routines also added.
5558 *) New Configure entry and patches for compiling on QNX 4.
5559 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5561 *) A demo state-machine implementation was sponsored by
5562 Nuron (http://www.nuron.com/) and is now available in
5563 demos/state_machine.
5566 *) New options added to the 'dgst' utility for signature
5567 generation and verification.
5570 *) Unrecognized PKCS#7 content types are now handled via a
5571 catch all ASN1_TYPE structure. This allows unsupported
5572 types to be stored as a "blob" and an application can
5573 encode and decode it manually.
5576 *) Fix various signed/unsigned issues to make a_strex.c
5578 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5580 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5581 length if passed a buffer. ASN1_INTEGER_to_BN failed
5582 if passed a NULL BN and its argument was negative.
5583 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5585 *) Modification to PKCS#7 encoding routines to output definite
5586 length encoding. Since currently the whole structures are in
5587 memory there's not real point in using indefinite length
5588 constructed encoding. However if OpenSSL is compiled with
5589 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5592 *) Added BIO_vprintf() and BIO_vsnprintf().
5595 *) Added more prefixes to parse for in the the strings written
5596 through a logging bio, to cover all the levels that are available
5597 through syslog. The prefixes are now:
5599 PANIC, EMERG, EMR => LOG_EMERG
5600 ALERT, ALR => LOG_ALERT
5601 CRIT, CRI => LOG_CRIT
5602 ERROR, ERR => LOG_ERR
5603 WARNING, WARN, WAR => LOG_WARNING
5604 NOTICE, NOTE, NOT => LOG_NOTICE
5605 INFO, INF => LOG_INFO
5606 DEBUG, DBG => LOG_DEBUG
5608 and as before, if none of those prefixes are present at the
5609 beginning of the string, LOG_ERR is chosen.
5611 On Win32, the LOG_* levels are mapped according to this:
5613 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5614 LOG_WARNING => EVENTLOG_WARNING_TYPE
5615 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5619 *) Made it possible to reconfigure with just the configuration
5620 argument "reconf" or "reconfigure". The command line arguments
5621 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5622 and are retrieved from there when reconfiguring.
5626 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5628 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5631 *) The obj_dat.pl script was messing up the sorting of object
5632 names. The reason was that it compared the quoted version
5633 of strings as a result "OCSP" > "OCSP Signing" because
5634 " > SPACE. Changed script to store unquoted versions of
5635 names and add quotes on output. It was also omitting some
5636 names from the lookup table if they were given a default
5637 value (that is if SN is missing it is given the same
5638 value as LN and vice versa), these are now added on the
5639 grounds that if an object has a name we should be able to
5640 look it up. Finally added warning output when duplicate
5641 short or long names are found.
5644 *) Changes needed for Tandem NSK.
5645 [Scott Uroff <scott@xypro.com>]
5647 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5648 RSA_padding_check_SSLv23(), special padding was never detected
5649 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5650 version rollback attacks was not effective.
5652 In s23_clnt.c, don't use special rollback-attack detection padding
5653 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5654 client; similarly, in s23_srvr.c, don't do the rollback check if
5655 SSL 2.0 is the only protocol enabled in the server.
5658 *) Make it possible to get hexdumps of unprintable data with 'openssl
5659 asn1parse'. By implication, the functions ASN1_parse_dump() and
5660 BIO_dump_indent() are added.
5663 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5664 these print out strings and name structures based on various
5665 flags including RFC2253 support and proper handling of
5666 multibyte characters. Added options to the 'x509' utility
5667 to allow the various flags to be set.
5670 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5671 Also change the functions X509_cmp_current_time() and
5672 X509_gmtime_adj() work with an ASN1_TIME structure,
5673 this will enable certificates using GeneralizedTime in validity
5674 dates to be checked.
5677 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5678 negative public key encodings) on by default,
5679 NO_NEG_PUBKEY_BUG can be set to disable it.
5682 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5683 content octets. An i2c_ASN1_OBJECT is unnecessary because
5684 the encoding can be trivially obtained from the structure.
5687 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5688 not read locks (CRYPTO_r_[un]lock).
5691 *) A first attempt at creating official support for shared
5692 libraries through configuration. I've kept it so the
5693 default is static libraries only, and the OpenSSL programs
5694 are always statically linked for now, but there are
5695 preparations for dynamic linking in place.
5696 This has been tested on Linux and Tru64.
5699 *) Randomness polling function for Win9x, as described in:
5700 Peter Gutmann, Software Generation of Practically Strong
5704 *) Fix so PRNG is seeded in req if using an already existing
5708 *) New options to smime application. -inform and -outform
5709 allow alternative formats for the S/MIME message including
5710 PEM and DER. The -content option allows the content to be
5711 specified separately. This should allow things like Netscape
5712 form signing output easier to verify.
5715 *) Fix the ASN1 encoding of tags using the 'long form'.
5718 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5719 STRING types. These convert content octets to and from the
5720 underlying type. The actual tag and length octets are
5721 already assumed to have been read in and checked. These
5722 are needed because all other string types have virtually
5723 identical handling apart from the tag. By having versions
5724 of the ASN1 functions that just operate on content octets
5725 IMPLICIT tagging can be handled properly. It also allows
5726 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5727 and ASN1_INTEGER are identical apart from the tag.
5730 *) Change the handling of OID objects as follows:
5732 - New object identifiers are inserted in objects.txt, following
5733 the syntax given in objects.README.
5734 - objects.pl is used to process obj_mac.num and create a new
5736 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5739 This is currently kind of a hack, and the perl code in objects.pl
5740 isn't very elegant, but it works as I intended. The simplest way
5741 to check that it worked correctly is to look in obj_dat.h and
5742 check the array nid_objs and make sure the objects haven't moved
5743 around (this is important!). Additions are OK, as well as
5744 consistent name changes.
5747 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5750 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5751 The given file adds to whatever has already been seeded into the
5752 random pool through the RANDFILE configuration file option or
5753 environment variable, or the default random state file.
5756 *) mkstack.pl now sorts each macro group into lexical order.
5757 Previously the output order depended on the order the files
5758 appeared in the directory, resulting in needless rewriting
5762 *) Patches to make OpenSSL compile under Win32 again. Mostly
5763 work arounds for the VC++ problem that it treats func() as
5764 func(void). Also stripped out the parts of mkdef.pl that
5765 added extra typesafe functions: these no longer exist.
5768 *) Reorganisation of the stack code. The macros are now all
5769 collected in safestack.h . Each macro is defined in terms of
5770 a "stack macro" of the form SKM_<name>(type, a, b). The
5771 DEBUG_SAFESTACK is now handled in terms of function casts,
5772 this has the advantage of retaining type safety without the
5773 use of additional functions. If DEBUG_SAFESTACK is not defined
5774 then the non typesafe macros are used instead. Also modified the
5775 mkstack.pl script to handle the new form. Needs testing to see
5776 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5777 the default if no major problems. Similar behaviour for ASN1_SET_OF
5778 and PKCS12_STACK_OF.
5781 *) When some versions of IIS use the 'NET' form of private key the
5782 key derivation algorithm is different. Normally MD5(password) is
5783 used as a 128 bit RC4 key. In the modified case
5784 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5785 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5786 as the old Netscape_RSA functions except they have an additional
5787 'sgckey' parameter which uses the modified algorithm. Also added
5788 an -sgckey command line option to the rsa utility. Thanks to
5789 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5790 algorithm to openssl-dev.
5793 *) The evp_local.h macros were using 'c.##kname' which resulted in
5794 invalid expansion on some systems (SCO 5.0.5 for example).
5795 Corrected to 'c.kname'.
5796 [Phillip Porch <root@theporch.com>]
5798 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5799 a STACK of email addresses from a certificate or request, these look
5800 in the subject name and the subject alternative name extensions and
5801 omit any duplicate addresses.
5804 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5805 This makes DSA verification about 2 % faster.
5808 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5809 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5810 plus overhead for 1024 bit moduli).
5811 This makes exponentiations about 0.5 % faster for 1024 bit
5812 exponents (as measured by "openssl speed rsa2048").
5815 *) Rename memory handling macros to avoid conflicts with other
5817 Malloc => OPENSSL_malloc
5818 Malloc_locked => OPENSSL_malloc_locked
5819 Realloc => OPENSSL_realloc
5820 Free => OPENSSL_free
5823 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5824 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5827 *) CygWin32 support.
5828 [John Jarvie <jjarvie@newsguy.com>]
5830 *) The type-safe stack code has been rejigged. It is now only compiled
5831 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5832 by default all type-specific stack functions are "#define"d back to
5833 standard stack functions. This results in more streamlined output
5834 but retains the type-safety checking possibilities of the original
5838 *) The STACK code has been cleaned up, and certain type declarations
5839 that didn't make a lot of sense have been brought in line. This has
5840 also involved a cleanup of sorts in safestack.h to more correctly
5841 map type-safe stack functions onto their plain stack counterparts.
5842 This work has also resulted in a variety of "const"ifications of
5843 lots of the code, especially "_cmp" operations which should normally
5844 be prototyped with "const" parameters anyway.
5847 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5848 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5849 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5850 where all of 'md' is used each time the PRNG is used, but 'state'
5851 is used only indexed by a cyclic counter. As entropy may not be
5852 well distributed from the beginning, 'md' is important as a
5853 chaining variable. However, the output function chains only half
5854 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5855 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5856 in all of 'state' being rewritten, with the new values depending
5857 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5860 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5861 the handshake is continued after ssl_verify_cert_chain();
5862 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5863 can lead to 'unexplainable' connection aborts later.
5864 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5866 *) Major EVP API cipher revision.
5867 Add hooks for extra EVP features. This allows various cipher
5868 parameters to be set in the EVP interface. Support added for variable
5869 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5870 setting of RC2 and RC5 parameters.
5872 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5875 Remove lots of duplicated code from the EVP library. For example *every*
5876 cipher init() function handles the 'iv' in the same way according to the
5877 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5878 for CFB and OFB modes they zero ctx->num.
5880 New functionality allows removal of S/MIME code RC2 hack.
5882 Most of the routines have the same form and so can be declared in terms
5885 By shifting this to the top level EVP_CipherInit() it can be removed from
5886 all individual ciphers. If the cipher wants to handle IVs or keys
5887 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5890 Change lots of functions like EVP_EncryptUpdate() to now return a
5891 value: although software versions of the algorithms cannot fail
5892 any installed hardware versions can.
5895 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5896 this option is set, tolerate broken clients that send the negotiated
5897 protocol version number instead of the requested protocol version
5901 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5902 i.e. non-zero for export ciphersuites, zero otherwise.
5903 Previous versions had this flag inverted, inconsistent with
5904 rsa_tmp_cb (..._TMP_RSA_CB).
5905 [Bodo Moeller; problem reported by Amit Chopra]
5907 *) Add missing DSA library text string. Work around for some IIS
5908 key files with invalid SEQUENCE encoding.
5911 *) Add a document (doc/standards.txt) that list all kinds of standards
5912 and so on that are implemented in OpenSSL.
5915 *) Enhance c_rehash script. Old version would mishandle certificates
5916 with the same subject name hash and wouldn't handle CRLs at all.
5917 Added -fingerprint option to crl utility, to support new c_rehash
5921 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5924 *) Fix for SSL server purpose checking. Server checking was
5925 rejecting certificates which had extended key usage present
5926 but no ssl client purpose.
5927 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5929 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5930 is a little unclear about how a blank password is handled.
5931 Since the password in encoded as a BMPString with terminating
5932 double NULL a zero length password would end up as just the
5933 double NULL. However no password at all is different and is
5934 handled differently in the PKCS#12 key generation code. NS
5935 treats a blank password as zero length. MSIE treats it as no
5936 password on export: but it will try both on import. We now do
5937 the same: PKCS12_parse() tries zero length and no password if
5938 the password is set to "" or NULL (NULL is now a valid password:
5939 it wasn't before) as does the pkcs12 application.
5942 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5943 perror when PEM_read_bio_X509_REQ fails, the error message must
5944 be obtained from the error queue.
5947 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5948 it in ERR_remove_state if appropriate, and change ERR_get_state
5949 accordingly to avoid race conditions (this is necessary because
5950 thread_hash is no longer constant once set).
5953 *) Bugfix for linux-elf makefile.one.
5956 *) RSA_get_default_method() will now cause a default
5957 RSA_METHOD to be chosen if one doesn't exist already.
5958 Previously this was only set during a call to RSA_new()
5959 or RSA_new_method(NULL) meaning it was possible for
5960 RSA_get_default_method() to return NULL.
5963 *) Added native name translation to the existing DSO code
5964 that will convert (if the flag to do so is set) filenames
5965 that are sufficiently small and have no path information
5966 into a canonical native form. Eg. "blah" converted to
5967 "libblah.so" or "blah.dll" etc.
5970 *) New function ERR_error_string_n(e, buf, len) which is like
5971 ERR_error_string(e, buf), but writes at most 'len' bytes
5972 including the 0 terminator. For ERR_error_string_n, 'buf'
5974 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5976 *) CONF library reworked to become more general. A new CONF
5977 configuration file reader "class" is implemented as well as a
5978 new functions (NCONF_*, for "New CONF") to handle it. The now
5979 old CONF_* functions are still there, but are reimplemented to
5980 work in terms of the new functions. Also, a set of functions
5981 to handle the internal storage of the configuration data is
5982 provided to make it easier to write new configuration file
5983 reader "classes" (I can definitely see something reading a
5984 configuration file in XML format, for example), called _CONF_*,
5985 or "the configuration storage API"...
5987 The new configuration file reading functions are:
5989 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5990 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5992 NCONF_default, NCONF_WIN32
5994 NCONF_dump_fp, NCONF_dump_bio
5996 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5997 NCONF_new creates a new CONF object. This works in the same way
5998 as other interfaces in OpenSSL, like the BIO interface.
5999 NCONF_dump_* dump the internal storage of the configuration file,
6000 which is useful for debugging. All other functions take the same
6001 arguments as the old CONF_* functions wth the exception of the
6002 first that must be a `CONF *' instead of a `LHASH *'.
6004 To make it easer to use the new classes with the old CONF_* functions,
6005 the function CONF_set_default_method is provided.
6008 *) Add '-tls1' option to 'openssl ciphers', which was already
6009 mentioned in the documentation but had not been implemented.
6010 (This option is not yet really useful because even the additional
6011 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6014 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6015 OpenSSL-based applications) load shared libraries and bind to
6016 them in a portable way.
6017 [Geoff Thorpe, with contributions from Richard Levitte]
6019 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6021 *) Make sure _lrotl and _lrotr are only used with MSVC.
6023 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6024 (the default implementation of RAND_status).
6026 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6027 to '-clrext' (= clear extensions), as intended and documented.
6028 [Bodo Moeller; inconsistency pointed out by Michael Attili
6031 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6032 was larger than the MD block size.
6033 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6035 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6036 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6037 using the passed key: if the passed key was a private key the result
6038 of X509_print(), for example, would be to print out all the private key
6042 *) des_quad_cksum() byte order bug fix.
6043 [Ulf Möller, using the problem description in krb4-0.9.7, where
6044 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6046 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6048 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6050 *) For easily testing in shell scripts whether some command
6051 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6052 returns with exit code 0 iff no command of the given name is available.
6053 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6054 the output goes to stdout and nothing is printed to stderr.
6055 Additional arguments are always ignored.
6057 Since for each cipher there is a command of the same name,
6058 the 'no-cipher' compilation switches can be tested this way.
6060 ('openssl no-XXX' is not able to detect pseudo-commands such
6061 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6064 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6067 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6068 is set; it will be thrown away anyway because each handshake creates
6070 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6071 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6072 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6073 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6076 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6077 'Q' and 'R' lose their special meanings (quit/renegotiate).
6078 This is part of what -quiet does; unlike -quiet, -ign_eof
6079 does not suppress any output.
6082 *) Add compatibility options to the purpose and trust code. The
6083 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6084 accepts a certificate or CA, this was the previous behaviour,
6085 with all the associated security issues.
6087 X509_TRUST_COMPAT is the old trust behaviour: only and
6088 automatically trust self signed roots in certificate store. A
6089 new trust setting X509_TRUST_DEFAULT is used to specify that
6090 a purpose has no associated trust setting and it should instead
6091 use the value in the default purpose.
6094 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6095 and fix a memory leak.
6098 *) In util/mkerr.pl (which implements 'make errors'), preserve
6099 reason strings from the previous version of the .c file, as
6100 the default to have only downcase letters (and digits) in
6101 automatically generated reasons codes is not always appropriate.
6104 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6105 using strerror. Previously, ERR_reason_error_string() returned
6106 library names as reason strings for SYSerr; but SYSerr is a special
6107 case where small numbers are errno values, not library numbers.
6110 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6111 converts DSA parameters into DH parameters. (When creating parameters,
6112 DSA_generate_parameters is used.)
6115 *) Include 'length' (recommended exponent length) in C code generated
6116 by 'openssl dhparam -C'.
6119 *) The second argument to set_label in perlasm was already being used
6120 so couldn't be used as a "file scope" flag. Moved to third argument
6124 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6125 instead of RAND_bytes for encryption IVs and salts.
6128 *) Include RAND_status() into RAND_METHOD instead of implementing
6129 it only for md_rand.c Otherwise replacing the PRNG by calling
6130 RAND_set_rand_method would be impossible.
6133 *) Don't let DSA_generate_key() enter an infinite loop if the random
6134 number generation fails.
6137 *) New 'rand' application for creating pseudo-random output.
6140 *) Added configuration support for Linux/IA64
6141 [Rolf Haberrecker <rolf@suse.de>]
6143 *) Assembler module support for Mingw32.
6146 *) Shared library support for HPUX (in shlib/).
6147 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6149 *) Shared library support for Solaris gcc.
6150 [Lutz Behnke <behnke@trustcenter.de>]
6152 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6154 *) PKCS7_encrypt() was adding text MIME headers twice because they
6155 were added manually and by SMIME_crlf_copy().
6158 *) In bntest.c don't call BN_rand with zero bits argument.
6159 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6161 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6162 case was implemented. This caused BN_div_recp() to fail occasionally.
6165 *) Add an optional second argument to the set_label() in the perl
6166 assembly language builder. If this argument exists and is set
6167 to 1 it signals that the assembler should use a symbol whose
6168 scope is the entire file, not just the current function. This
6169 is needed with MASM which uses the format label:: for this scope.
6170 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6172 *) Change the ASN1 types so they are typedefs by default. Before
6173 almost all types were #define'd to ASN1_STRING which was causing
6174 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6178 *) Change names of new functions to the new get1/get0 naming
6179 convention: After 'get1', the caller owns a reference count
6180 and has to call ..._free; 'get0' returns a pointer to some
6181 data structure without incrementing reference counters.
6182 (Some of the existing 'get' functions increment a reference
6183 counter, some don't.)
6184 Similarly, 'set1' and 'add1' functions increase reference
6185 counters or duplicate objects.
6188 *) Allow for the possibility of temp RSA key generation failure:
6189 the code used to assume it always worked and crashed on failure.
6192 *) Fix potential buffer overrun problem in BIO_printf().
6193 [Ulf Möller, using public domain code by Patrick Powell; problem
6194 pointed out by David Sacerdote <das33@cornell.edu>]
6196 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6197 RAND_egd() and RAND_status(). In the command line application,
6198 the EGD socket can be specified like a seed file using RANDFILE
6202 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6203 Some CAs (e.g. Verisign) distribute certificates in this form.
6206 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6207 list to exclude them. This means that no special compilation option
6208 is needed to use anonymous DH: it just needs to be included in the
6212 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6213 EVP_MD_type. The old functionality is available in a new macro called
6214 EVP_MD_md(). Change code that uses it and update docs.
6217 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6218 where the 'void *' argument is replaced by a function pointer argument.
6219 Previously 'void *' was abused to point to functions, which works on
6220 many platforms, but is not correct. As these functions are usually
6221 called by macros defined in OpenSSL header files, most source code
6222 should work without changes.
6225 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6226 sections with information on -D... compiler switches used for
6227 compiling the library so that applications can see them. To enable
6228 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6229 must be defined. E.g.,
6230 #define OPENSSL_ALGORITHM_DEFINES
6231 #include <openssl/opensslconf.h>
6232 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6233 [Richard Levitte, Ulf and Bodo Möller]
6235 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6239 *) Change the 'other' type in certificate aux info to a STACK_OF
6240 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6241 the required ASN1 format: arbitrary types determined by an OID.
6244 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6245 argument to 'req'. This is not because the function is newer or
6246 better than others it just uses the work 'NEW' in the certificate
6247 request header lines. Some software needs this.
6250 *) Reorganise password command line arguments: now passwords can be
6251 obtained from various sources. Delete the PEM_cb function and make
6252 it the default behaviour: i.e. if the callback is NULL and the
6253 usrdata argument is not NULL interpret it as a null terminated pass
6254 phrase. If usrdata and the callback are NULL then the pass phrase
6255 is prompted for as usual.
6258 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6259 the support is automatically enabled. The resulting binaries will
6260 autodetect the card and use it if present.
6261 [Ben Laurie and Compaq Inc.]
6263 *) Work around for Netscape hang bug. This sends certificate request
6264 and server done in one record. Since this is perfectly legal in the
6265 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6266 the bugs/SSLv3 entry for more info.
6269 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6272 *) Add -rand argument to smime and pkcs12 applications and read/write
6276 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6279 *) Add command line password options to the remaining applications.
6282 *) Bug fix for BN_div_recp() for numerators with an even number of
6286 *) More tests in bntest.c, and changed test_bn output.
6289 *) ./config recognizes MacOS X now.
6292 *) Bug fix for BN_div() when the first words of num and divsor are
6293 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6296 *) Add support for various broken PKCS#8 formats, and command line
6297 options to produce them.
6300 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6301 get temporary BIGNUMs from a BN_CTX.
6304 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6308 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6309 include a #define from the old name to the new. The original intent
6310 was that statically linked binaries could for example just call
6311 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6312 link with digests. This never worked becayse SSLeay_add_all_digests()
6313 and SSLeay_add_all_ciphers() were in the same source file so calling
6314 one would link with the other. They are now in separate source files.
6317 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6320 *) Use a less unusual form of the Miller-Rabin primality test (it used
6321 a binary algorithm for exponentiation integrated into the Miller-Rabin
6322 loop, our standard modexp algorithms are faster).
6325 *) Support for the EBCDIC character set completed.
6326 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6328 *) Source code cleanups: use const where appropriate, eliminate casts,
6329 use void * instead of char * in lhash.
6332 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6333 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6334 this the server could overwrite ephemeral keys that the client
6338 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6339 using 50 iterations of the Rabin-Miller test.
6341 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6342 iterations of the Rabin-Miller test as required by the appendix
6343 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6344 As BN_is_prime_fasttest includes trial division, DSA parameter
6345 generation becomes much faster.
6347 This implies a change for the callback functions in DSA_is_prime
6348 and DSA_generate_parameters: The callback function is called once
6349 for each positive witness in the Rabin-Miller test, not just
6350 occasionally in the inner loop; and the parameters to the
6351 callback function now provide an iteration count for the outer
6352 loop rather than for the current invocation of the inner loop.
6353 DSA_generate_parameters additionally can call the callback
6354 function with an 'iteration count' of -1, meaning that a
6355 candidate has passed the trial division test (when q is generated
6356 from an application-provided seed, trial division is skipped).
6359 *) New function BN_is_prime_fasttest that optionally does trial
6360 division before starting the Rabin-Miller test and has
6361 an additional BN_CTX * argument (whereas BN_is_prime always
6362 has to allocate at least one BN_CTX).
6363 'callback(1, -1, cb_arg)' is called when a number has passed the
6364 trial division stage.
6367 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6371 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6374 *) New function BN_pseudo_rand().
6377 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6378 bignum version of BN_from_montgomery() with the working code from
6379 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6383 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6384 made it impossible to use the same SSL_SESSION data structure in
6385 SSL2 clients in multiple threads.
6388 *) The return value of RAND_load_file() no longer counts bytes obtained
6389 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6390 to seed the PRNG (previously an explicit byte count was required).
6391 [Ulf Möller, Bodo Möller]
6393 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6394 used (char *) instead of (void *) and had casts all over the place.
6397 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6400 *) Retain source code compatibility for BN_prime_checks macro:
6401 BN_is_prime(..., BN_prime_checks, ...) now uses
6402 BN_prime_checks_for_size to determine the appropriate number of
6403 Rabin-Miller iterations.
6406 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6407 DH_CHECK_P_NOT_SAFE_PRIME.
6408 (Check if this is true? OpenPGP calls them "strong".)
6411 *) Merge the functionality of "dh" and "gendh" programs into a new program
6412 "dhparam". The old programs are retained for now but will handle DH keys
6413 (instead of parameters) in future.
6416 *) Make the ciphers, s_server and s_client programs check the return values
6417 when a new cipher list is set.
6420 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6421 ciphers. Before when the 56bit ciphers were enabled the sorting was
6424 The syntax for the cipher sorting has been extended to support sorting by
6425 cipher-strength (using the strength_bits hard coded in the tables).
6426 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6428 Fix a bug in the cipher-command parser: when supplying a cipher command
6429 string with an "undefined" symbol (neither command nor alphanumeric
6430 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6431 an error is flagged.
6433 Due to the strength-sorting extension, the code of the
6434 ssl_create_cipher_list() function was completely rearranged. I hope that
6435 the readability was also increased :-)
6436 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6438 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6439 for the first serial number and places 2 in the serial number file. This
6440 avoids problems when the root CA is created with serial number zero and
6441 the first user certificate has the same issuer name and serial number
6445 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6446 the new code. Add documentation for this stuff.
6449 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6450 X509_*() to X509at_*() on the grounds that they don't handle X509
6451 structures and behave in an analagous way to the X509v3 functions:
6452 they shouldn't be called directly but wrapper functions should be used
6455 So we also now have some wrapper functions that call the X509at functions
6456 when passed certificate requests. (TO DO: similar things can be done with
6457 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6458 things. Some of these need some d2i or i2d and print functionality
6459 because they handle more complex structures.)
6462 *) Add missing #ifndefs that caused missing symbols when building libssl
6463 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6464 NO_RSA in ssl/s2*.c.
6465 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6467 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6468 has a return value which indicates the quality of the random data
6469 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6470 error queue. New function RAND_pseudo_bytes() generates output that is
6471 guaranteed to be unique but not unpredictable. RAND_add is like
6472 RAND_seed, but takes an extra argument for an entropy estimate
6473 (RAND_seed always assumes full entropy).
6476 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6477 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6478 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6479 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6480 false-positive rate of at most 2^-80 for random input.
6483 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6486 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6487 in the 0.9.5 release), this returns the chain
6488 from an X509_CTX structure with a dup of the stack and all
6489 the X509 reference counts upped: so the stack will exist
6490 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6493 Also make SSL_SESSION_print() print out the verify return
6497 *) Add manpage for the pkcs12 command. Also change the default
6498 behaviour so MAC iteration counts are used unless the new
6499 -nomaciter option is used. This improves file security and
6500 only older versions of MSIE (4.0 for example) need it.
6503 *) Honor the no-xxx Configure options when creating .DEF files.
6506 *) Add PKCS#10 attributes to field table: challengePassword,
6507 unstructuredName and unstructuredAddress. These are taken from
6508 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6509 international characters are used.
6511 More changes to X509_ATTRIBUTE code: allow the setting of types
6512 based on strings. Remove the 'loc' parameter when adding
6513 attributes because these will be a SET OF encoding which is sorted
6517 *) Initial changes to the 'req' utility to allow request generation
6518 automation. This will allow an application to just generate a template
6519 file containing all the field values and have req construct the
6522 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6523 used all over the place including certificate requests and PKCS#7
6524 structures. They are currently handled manually where necessary with
6525 some primitive wrappers for PKCS#7. The new functions behave in a
6526 manner analogous to the X509 extension functions: they allow
6527 attributes to be looked up by NID and added.
6529 Later something similar to the X509V3 code would be desirable to
6530 automatically handle the encoding, decoding and printing of the
6531 more complex types. The string types like challengePassword can
6532 be handled by the string table functions.
6534 Also modified the multi byte string table handling. Now there is
6535 a 'global mask' which masks out certain types. The table itself
6536 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6537 is useful when for example there is only one permissible type
6538 (as in countryName) and using the mask might result in no valid
6542 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6543 SSL_get_peer_finished to allow applications to obtain the latest
6544 Finished messages sent to the peer or expected from the peer,
6545 respectively. (SSL_get_peer_finished is usually the Finished message
6546 actually received from the peer, otherwise the protocol will be aborted.)
6548 As the Finished message are message digests of the complete handshake
6549 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6550 be used for external authentication procedures when the authentication
6551 provided by SSL/TLS is not desired or is not enough.
6554 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6555 the host supports BWX extension and if Compaq C is present on the
6556 $PATH. Just exploiting of the BWX extension results in 20-30%
6557 performance kick for some algorithms, e.g. DES and RC4 to mention
6558 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6562 *) Add support for MS "fast SGC". This is arguably a violation of the
6563 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6564 weak crypto and after checking the certificate is SGC a second one
6565 with strong crypto. MS SGC stops the first handshake after receiving
6566 the server certificate message and sends a second client hello. Since
6567 a server will typically do all the time consuming operations before
6568 expecting any further messages from the client (server key exchange
6569 is the most expensive) there is little difference between the two.
6571 To get OpenSSL to support MS SGC we have to permit a second client
6572 hello message after we have sent server done. In addition we have to
6573 reset the MAC if we do get this second client hello.
6576 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6577 if a DER encoded private key is RSA or DSA traditional format. Changed
6578 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6579 format DER encoded private key. Newer code should use PKCS#8 format which
6580 has the key type encoded in the ASN1 structure. Added DER private key
6581 support to pkcs8 application.
6584 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6585 ciphersuites has been selected (as required by the SSL 3/TLS 1
6586 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6587 is set, we interpret this as a request to violate the specification
6588 (the worst that can happen is a handshake failure, and 'correct'
6589 behaviour would result in a handshake failure anyway).
6592 *) In SSL_CTX_add_session, take into account that there might be multiple
6593 SSL_SESSION structures with the same session ID (e.g. when two threads
6594 concurrently obtain them from an external cache).
6595 The internal cache can handle only one SSL_SESSION with a given ID,
6596 so if there's a conflict, we now throw out the old one to achieve
6600 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6601 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6602 some routines that use cipher OIDs: some ciphers do not have OIDs
6603 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6607 *) Simplify the trust setting structure and code. Now we just have
6608 two sequences of OIDs for trusted and rejected settings. These will
6609 typically have values the same as the extended key usage extension
6610 and any application specific purposes.
6612 The trust checking code now has a default behaviour: it will just
6613 check for an object with the same NID as the passed id. Functions can
6614 be provided to override either the default behaviour or the behaviour
6615 for a given id. SSL client, server and email already have functions
6616 in place for compatibility: they check the NID and also return "trusted"
6617 if the certificate is self signed.
6620 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6621 traditional format into an EVP_PKEY structure.
6624 *) Add a password callback function PEM_cb() which either prompts for
6625 a password if usr_data is NULL or otherwise assumes it is a null
6626 terminated password. Allow passwords to be passed on command line
6627 environment or config files in a few more utilities.
6630 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6631 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6632 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6633 Update documentation.
6636 *) Support for ASN1 "NULL" type. This could be handled before by using
6637 ASN1_TYPE but there wasn't any function that would try to read a NULL
6638 and produce an error if it couldn't. For compatibility we also have
6639 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6640 don't allocate anything because they don't need to.
6643 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6645 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6647 *) Rebuild of the memory allocation routines used by OpenSSL code and
6648 possibly others as well. The purpose is to make an interface that
6649 provide hooks so anyone can build a separate set of allocation and
6650 deallocation routines to be used by OpenSSL, for example memory
6651 pool implementations, or something else, which was previously hard
6652 since Malloc(), Realloc() and Free() were defined as macros having
6653 the values malloc, realloc and free, respectively (except for Win32
6654 compilations). The same is provided for memory debugging code.
6655 OpenSSL already comes with functionality to find memory leaks, but
6656 this gives people a chance to debug other memory problems.
6658 With these changes, a new set of functions and macros have appeared:
6660 CRYPTO_set_mem_debug_functions() [F]
6661 CRYPTO_get_mem_debug_functions() [F]
6662 CRYPTO_dbg_set_options() [F]
6663 CRYPTO_dbg_get_options() [F]
6664 CRYPTO_malloc_debug_init() [M]
6666 The memory debug functions are NULL by default, unless the library
6667 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6668 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6669 gives the standard debugging functions that come with OpenSSL) or
6670 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6671 provided by the library user) must be used. When the standard
6672 debugging functions are used, CRYPTO_dbg_set_options can be used to
6673 request additional information:
6674 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6675 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6677 Also, things like CRYPTO_set_mem_functions will always give the
6678 expected result (the new set of functions is used for allocation
6679 and deallocation) at all times, regardless of platform and compiler
6682 To finish it up, some functions that were never use in any other
6683 way than through macros have a new API and new semantic:
6686 CRYPTO_dbg_realloc()
6689 All macros of value have retained their old syntax.
6690 [Richard Levitte and Bodo Moeller]
6692 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6693 ordering of SMIMECapabilities wasn't in "strength order" and there
6694 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6698 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6699 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6700 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6702 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6703 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6704 functionality to handle multipart/signed properly) and a utility
6705 called 'smime' to call all this stuff. This is based on code I
6706 originally wrote for Celo who have kindly allowed it to be
6707 included in OpenSSL.
6710 *) Add variants des_set_key_checked and des_set_key_unchecked of
6711 des_set_key (aka des_key_sched). Global variable des_check_key
6712 decides which of these is called by des_set_key; this way
6713 des_check_key behaves as it always did, but applications and
6714 the library itself, which was buggy for des_check_key == 1,
6715 have a cleaner way to pick the version they need.
6718 *) New function PKCS12_newpass() which changes the password of a
6722 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6723 dynamic mix. In both cases the ids can be used as an index into the
6724 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6725 functions so they accept a list of the field values and the
6726 application doesn't need to directly manipulate the X509_TRUST
6730 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6734 *) Modify the way the V3 extension code looks up extensions. This now
6735 works in a similar way to the object code: we have some "standard"
6736 extensions in a static table which is searched with OBJ_bsearch()
6737 and the application can add dynamic ones if needed. The file
6738 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6739 updated whenever a new extension is added to the core code and kept
6740 in ext_nid order. There is a simple program 'tabtest.c' which checks
6741 this. New extensions are not added too often so this file can readily
6742 be maintained manually.
6744 There are two big advantages in doing things this way. The extensions
6745 can be looked up immediately and no longer need to be "added" using
6746 X509V3_add_standard_extensions(): this function now does nothing.
6747 [Side note: I get *lots* of email saying the extension code doesn't
6748 work because people forget to call this function]
6749 Also no dynamic allocation is done unless new extensions are added:
6750 so if we don't add custom extensions there is no need to call
6751 X509V3_EXT_cleanup().
6754 *) Modify enc utility's salting as follows: make salting the default. Add a
6755 magic header, so unsalted files fail gracefully instead of just decrypting
6756 to garbage. This is because not salting is a big security hole, so people
6757 should be discouraged from doing it.
6760 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6761 digest to be passed on the command line but it only used this
6762 parameter when signing a certificate. Modified so all relevant
6763 operations are affected by the digest parameter including the
6764 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6765 DSA key was used because it didn't fix the digest.
6768 *) Initial certificate chain verify code. Currently tests the untrusted
6769 certificates for consistency with the verify purpose (which is set
6770 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6772 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6773 this is because it will reject chains with invalid extensions whereas
6774 every previous version of OpenSSL and SSLeay made no checks at all.
6776 Trust code: checks the root CA for the relevant trust settings. Trust
6777 settings have an initial value consistent with the verify purpose: e.g.
6778 if the verify purpose is for SSL client use it expects the CA to be
6779 trusted for SSL client use. However the default value can be changed to
6780 permit custom trust settings: one example of this would be to only trust
6781 certificates from a specific "secure" set of CAs.
6783 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6784 which should be used for version portability: especially since the
6785 verify structure is likely to change more often now.
6787 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6788 to set them. If not set then assume SSL clients will verify SSL servers
6791 Two new options to the verify program: -untrusted allows a set of
6792 untrusted certificates to be passed in and -purpose which sets the
6793 intended purpose of the certificate. If a purpose is set then the
6794 new chain verify code is used to check extension consistency.
6797 *) Support for the authority information access extension.
6800 *) Modify RSA and DSA PEM read routines to transparently handle
6801 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6802 public keys in a format compatible with certificate
6803 SubjectPublicKeyInfo structures. Unfortunately there were already
6804 functions called *_PublicKey_* which used various odd formats so
6805 these are retained for compatibility: however the DSA variants were
6806 never in a public release so they have been deleted. Changed dsa/rsa
6807 utilities to handle the new format: note no releases ever handled public
6808 keys so we should be OK.
6810 The primary motivation for this change is to avoid the same fiasco
6811 that dogs private keys: there are several incompatible private key
6812 formats some of which are standard and some OpenSSL specific and
6813 require various evil hacks to allow partial transparent handling and
6814 even then it doesn't work with DER formats. Given the option anything
6815 other than PKCS#8 should be dumped: but the other formats have to
6816 stay in the name of compatibility.
6818 With public keys and the benefit of hindsight one standard format
6819 is used which works with EVP_PKEY, RSA or DSA structures: though
6820 it clearly returns an error if you try to read the wrong kind of key.
6822 Added a -pubkey option to the 'x509' utility to output the public key.
6823 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6824 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6825 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6826 that do the same as the EVP_PKEY_assign_*() except they up the
6827 reference count of the added key (they don't "swallow" the
6831 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6832 CRLs would fail if the file contained no certificates or no CRLs:
6833 added a new function to read in both types and return the number
6834 read: this means that if none are read it will be an error. The
6835 DER versions of the certificate and CRL reader would always fail
6836 because it isn't possible to mix certificates and CRLs in DER format
6837 without choking one or the other routine. Changed this to just read
6838 a certificate: this is the best we can do. Also modified the code
6839 in apps/verify.c to take notice of return codes: it was previously
6840 attempting to read in certificates from NULL pointers and ignoring
6841 any errors: this is one reason why the cert and CRL reader seemed
6842 to work. It doesn't check return codes from the default certificate
6843 routines: these may well fail if the certificates aren't installed.
6846 *) Code to support otherName option in GeneralName.
6849 *) First update to verify code. Change the verify utility
6850 so it warns if it is passed a self signed certificate:
6851 for consistency with the normal behaviour. X509_verify
6852 has been modified to it will now verify a self signed
6853 certificate if *exactly* the same certificate appears
6854 in the store: it was previously impossible to trust a
6855 single self signed certificate. This means that:
6856 openssl verify ss.pem
6857 now gives a warning about a self signed certificate but
6858 openssl verify -CAfile ss.pem ss.pem
6862 *) For servers, store verify_result in SSL_SESSION data structure
6863 (and add it to external session representation).
6864 This is needed when client certificate verifications fails,
6865 but an application-provided verification callback (set by
6866 SSL_CTX_set_cert_verify_callback) allows accepting the session
6867 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6868 but returns 1): When the session is reused, we have to set
6869 ssl->verify_result to the appropriate error code to avoid
6871 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6873 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6874 case in PKCS7_dataInit() where the signed PKCS7 structure
6875 didn't contain any existing data because it was being created.
6876 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6878 *) Add a salt to the key derivation routines in enc.c. This
6879 forms the first 8 bytes of the encrypted file. Also add a
6880 -S option to allow a salt to be input on the command line.
6883 *) New function X509_cmp(). Oddly enough there wasn't a function
6884 to compare two certificates. We do this by working out the SHA1
6885 hash and comparing that. X509_cmp() will be needed by the trust
6889 *) SSL_get1_session() is like SSL_get_session(), but increments
6890 the reference count in the SSL_SESSION returned.
6891 [Geoff Thorpe <geoff@eu.c2.net>]
6893 *) Fix for 'req': it was adding a null to request attributes.
6894 Also change the X509_LOOKUP and X509_INFO code to handle
6895 certificate auxiliary information.
6898 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6902 *) Add the possibility to add extra information to the memory leak
6903 detecting output, to form tracebacks, showing from where each
6904 allocation was originated: CRYPTO_push_info("constant string") adds
6905 the string plus current file name and line number to a per-thread
6906 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6907 is like calling CYRPTO_pop_info() until the stack is empty.
6908 Also updated memory leak detection code to be multi-thread-safe.
6911 *) Add options -text and -noout to pkcs7 utility and delete the
6912 encryption options which never did anything. Update docs.
6915 *) Add options to some of the utilities to allow the pass phrase
6916 to be included on either the command line (not recommended on
6917 OSes like Unix) or read from the environment. Update the
6918 manpages and fix a few bugs.
6921 *) Add a few manpages for some of the openssl commands.
6924 *) Fix the -revoke option in ca. It was freeing up memory twice,
6925 leaking and not finding already revoked certificates.
6928 *) Extensive changes to support certificate auxiliary information.
6929 This involves the use of X509_CERT_AUX structure and X509_AUX
6930 functions. An X509_AUX function such as PEM_read_X509_AUX()
6931 can still read in a certificate file in the usual way but it
6932 will also read in any additional "auxiliary information". By
6933 doing things this way a fair degree of compatibility can be
6934 retained: existing certificates can have this information added
6935 using the new 'x509' options.
6937 Current auxiliary information includes an "alias" and some trust
6938 settings. The trust settings will ultimately be used in enhanced
6939 certificate chain verification routines: currently a certificate
6940 can only be trusted if it is self signed and then it is trusted
6944 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6945 The problem was that one of the replacement routines had not been working
6946 since SSLeay releases. For now the offending routine has been replaced
6947 with non-optimised assembler. Even so, this now gives around 95%
6948 performance improvement for 1024 bit RSA signs.
6951 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6952 handling. Most clients have the effective key size in bits equal to
6953 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6954 A few however don't do this and instead use the size of the decrypted key
6955 to determine the RC2 key length and the AlgorithmIdentifier to determine
6956 the effective key length. In this case the effective key length can still
6957 be 40 bits but the key length can be 168 bits for example. This is fixed
6958 by manually forcing an RC2 key into the EVP_PKEY structure because the
6959 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6960 the key length and effective key length are equal.
6963 *) Add a bunch of functions that should simplify the creation of
6964 X509_NAME structures. Now you should be able to do:
6965 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6966 and have it automatically work out the correct field type and fill in
6967 the structures. The more adventurous can try:
6968 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6969 and it will (hopefully) work out the correct multibyte encoding.
6972 *) Change the 'req' utility to use the new field handling and multibyte
6973 copy routines. Before the DN field creation was handled in an ad hoc
6974 way in req, ca, and x509 which was rather broken and didn't support
6975 BMPStrings or UTF8Strings. Since some software doesn't implement
6976 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6977 using the dirstring_type option. See the new comment in the default
6978 openssl.cnf for more info.
6981 *) Make crypto/rand/md_rand.c more robust:
6982 - Assure unique random numbers after fork().
6983 - Make sure that concurrent threads access the global counter and
6984 md serializably so that we never lose entropy in them
6985 or use exactly the same state in multiple threads.
6986 Access to the large state is not always serializable because
6987 the additional locking could be a performance killer, and
6988 md should be large enough anyway.
6991 *) New file apps/app_rand.c with commonly needed functionality
6992 for handling the random seed file.
6994 Use the random seed file in some applications that previously did not:
6996 dsaparam -genkey (which also ignored its '-rand' option),
6999 x509 (when signing).
7000 Except on systems with /dev/urandom, it is crucial to have a random
7001 seed file at least for key creation, DSA signing, and for DH exchanges;
7002 for RSA signatures we could do without one.
7004 gendh and gendsa (unlike genrsa) used to read only the first byte
7005 of each file listed in the '-rand' option. The function as previously
7006 found in genrsa is now in app_rand.c and is used by all programs
7007 that support '-rand'.
7010 *) In RAND_write_file, use mode 0600 for creating files;
7011 don't just chmod when it may be too late.
7014 *) Report an error from X509_STORE_load_locations
7015 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7018 *) New function ASN1_mbstring_copy() this copies a string in either
7019 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7020 into an ASN1_STRING type. A mask of permissible types is passed
7021 and it chooses the "minimal" type to use or an error if not type
7025 *) Add function equivalents to the various macros in asn1.h. The old
7026 macros are retained with an M_ prefix. Code inside the library can
7027 use the M_ macros. External code (including the openssl utility)
7028 should *NOT* in order to be "shared library friendly".
7031 *) Add various functions that can check a certificate's extensions
7032 to see if it usable for various purposes such as SSL client,
7033 server or S/MIME and CAs of these types. This is currently
7034 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7035 verification. Also added a -purpose flag to x509 utility to
7036 print out all the purposes.
7039 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7043 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7044 for, obtain and decode and extension and obtain its critical flag.
7045 This allows all the necessary extension code to be handled in a
7046 single function call.
7049 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7050 platforms. See crypto/rc4/rc4_enc.c for further details.
7053 *) New -noout option to asn1parse. This causes no output to be produced
7054 its main use is when combined with -strparse and -out to extract data
7055 from a file (which may not be in ASN.1 format).
7058 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7059 when producing the local key id.
7060 [Richard Levitte <levitte@stacken.kth.se>]
7062 *) New option -dhparam in s_server. This allows a DH parameter file to be
7063 stated explicitly. If it is not stated then it tries the first server
7064 certificate file. The previous behaviour hard coded the filename
7068 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7069 a public key to be input or output. For example:
7070 openssl rsa -in key.pem -pubout -out pubkey.pem
7071 Also added necessary DSA public key functions to handle this.
7074 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7075 in the message. This was handled by allowing
7076 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7077 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7079 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7080 to the end of the strings whereas this didn't. This would cause problems
7081 if strings read with d2i_ASN1_bytes() were later modified.
7082 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7084 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7085 data and it contains EOF it will end up returning an error. This is
7086 caused by input 46 bytes long. The cause is due to the way base64
7087 BIOs find the start of base64 encoded data. They do this by trying a
7088 trial decode on each line until they find one that works. When they
7089 do a flag is set and it starts again knowing it can pass all the
7090 data directly through the decoder. Unfortunately it doesn't reset
7091 the context it uses. This means that if EOF is reached an attempt
7092 is made to pass two EOFs through the context and this causes the
7093 resulting error. This can also cause other problems as well. As is
7094 usual with these problems it takes *ages* to find and the fix is
7095 trivial: move one line.
7096 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7098 *) Ugly workaround to get s_client and s_server working under Windows. The
7099 old code wouldn't work because it needed to select() on sockets and the
7100 tty (for keypresses and to see if data could be written). Win32 only
7101 supports select() on sockets so we select() with a 1s timeout on the
7102 sockets and then see if any characters are waiting to be read, if none
7103 are present then we retry, we also assume we can always write data to
7104 the tty. This isn't nice because the code then blocks until we've
7105 received a complete line of data and it is effectively polling the
7106 keyboard at 1s intervals: however it's quite a bit better than not
7107 working at all :-) A dedicated Windows application might handle this
7108 with an event loop for example.
7111 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7112 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7113 will be called when RSA_sign() and RSA_verify() are used. This is useful
7114 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7115 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7116 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7117 This necessitated the support of an extra signature type NID_md5_sha1
7118 for SSL signatures and modifications to the SSL library to use it instead
7119 of calling RSA_public_decrypt() and RSA_private_encrypt().
7122 *) Add new -verify -CAfile and -CApath options to the crl program, these
7123 will lookup a CRL issuers certificate and verify the signature in a
7124 similar way to the verify program. Tidy up the crl program so it
7125 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7126 less strict. It will now permit CRL extensions even if it is not
7127 a V2 CRL: this will allow it to tolerate some broken CRLs.
7130 *) Initialize all non-automatic variables each time one of the openssl
7131 sub-programs is started (this is necessary as they may be started
7132 multiple times from the "OpenSSL>" prompt).
7133 [Lennart Bang, Bodo Moeller]
7135 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7136 removing all other RSA functionality (this is what NO_RSA does). This
7137 is so (for example) those in the US can disable those operations covered
7138 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7142 *) Non-copying interface to BIO pairs.
7143 (still largely untested)
7146 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7147 ASCII string. This was handled independently in various places before.
7150 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7151 UTF8 strings a character at a time.
7154 *) Use client_version from client hello to select the protocol
7155 (s23_srvr.c) and for RSA client key exchange verification
7156 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7159 *) Add various utility functions to handle SPKACs, these were previously
7160 handled by poking round in the structure internals. Added new function
7161 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7162 print, verify and generate SPKACs. Based on an original idea from
7163 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7166 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7169 *) Allow the config file extension section to be overwritten on the
7170 command line. Based on an original idea from Massimiliano Pala
7171 <madwolf@comune.modena.it>. The new option is called -extensions
7172 and can be applied to ca, req and x509. Also -reqexts to override
7173 the request extensions in req and -crlexts to override the crl extensions
7177 *) Add new feature to the SPKAC handling in ca. Now you can include
7178 the same field multiple times by preceding it by "XXXX." for example:
7181 this is the same syntax as used in the req config file.
7184 *) Allow certificate extensions to be added to certificate requests. These
7185 are specified in a 'req_extensions' option of the req section of the
7186 config file. They can be printed out with the -text option to req but
7187 are otherwise ignored at present.
7190 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7191 data read consists of only the final block it would not decrypted because
7192 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7193 A misplaced 'break' also meant the decrypted final block might not be
7194 copied until the next read.
7197 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7198 a few extra parameters to the DH structure: these will be useful if
7199 for example we want the value of 'q' or implement X9.42 DH.
7202 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7203 provides hooks that allow the default DSA functions or functions on a
7204 "per key" basis to be replaced. This allows hardware acceleration and
7205 hardware key storage to be handled without major modification to the
7206 library. Also added low level modexp hooks and CRYPTO_EX structure and
7207 associated functions.
7210 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7211 as "read only": it can't be written to and the buffer it points to will
7212 not be freed. Reading from a read only BIO is much more efficient than
7213 a normal memory BIO. This was added because there are several times when
7214 an area of memory needs to be read from a BIO. The previous method was
7215 to create a memory BIO and write the data to it, this results in two
7216 copies of the data and an O(n^2) reading algorithm. There is a new
7217 function BIO_new_mem_buf() which creates a read only memory BIO from
7218 an area of memory. Also modified the PKCS#7 routines to use read only
7222 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7223 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7224 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7225 but a retry condition occured while trying to read the rest.
7228 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7229 NID_pkcs7_encrypted by default: this was wrong since this should almost
7230 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7231 the encrypted data type: this is a more sensible place to put it and it
7232 allows the PKCS#12 code to be tidied up that duplicated this
7236 *) Changed obj_dat.pl script so it takes its input and output files on
7237 the command line. This should avoid shell escape redirection problems
7241 *) Initial support for certificate extension requests, these are included
7242 in things like Xenroll certificate requests. Included functions to allow
7243 extensions to be obtained and added.
7246 *) -crlf option to s_client and s_server for sending newlines as
7247 CRLF (as required by many protocols).
7250 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7252 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7253 [Ralf S. Engelschall]
7255 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7256 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7258 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7262 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7263 DH parameters/keys (q is lost during that conversion, but the resulting
7264 DH parameters contain its length).
7266 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7267 much faster than DH_generate_parameters (which creates parameters
7268 where p = 2*q + 1), and also the smaller q makes DH computations
7269 much more efficient (160-bit exponentiation instead of 1024-bit
7270 exponentiation); so this provides a convenient way to support DHE
7271 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7272 utter importance to use
7273 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7275 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7276 when such DH parameters are used, because otherwise small subgroup
7277 attacks may become possible!
7280 *) Avoid memory leak in i2d_DHparams.
7283 *) Allow the -k option to be used more than once in the enc program:
7284 this allows the same encrypted message to be read by multiple recipients.
7287 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7288 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7289 it will always use the numerical form of the OID, even if it has a short
7293 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7294 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7295 otherwise bn_mod_exp was called. In the case of hardware keys for example
7296 no private key components need be present and it might store extra data
7297 in the RSA structure, which cannot be accessed from bn_mod_exp.
7298 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7299 private key operations.
7302 *) Added support for SPARC Linux.
7305 *) pem_password_cb function type incompatibly changed from
7306 typedef int pem_password_cb(char *buf, int size, int rwflag);
7308 ....(char *buf, int size, int rwflag, void *userdata);
7309 so that applications can pass data to their callbacks:
7310 The PEM[_ASN1]_{read,write}... functions and macros now take an
7311 additional void * argument, which is just handed through whenever
7312 the password callback is called.
7313 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7315 New function SSL_CTX_set_default_passwd_cb_userdata.
7317 Compatibility note: As many C implementations push function arguments
7318 onto the stack in reverse order, the new library version is likely to
7319 interoperate with programs that have been compiled with the old
7320 pem_password_cb definition (PEM_whatever takes some data that
7321 happens to be on the stack as its last argument, and the callback
7322 just ignores this garbage); but there is no guarantee whatsoever that
7325 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7326 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7327 problems not only on Windows, but also on some Unix platforms.
7328 To avoid problematic command lines, these definitions are now in an
7329 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7330 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7333 *) MIPS III/IV assembler module is reimplemented.
7336 *) More DES library cleanups: remove references to srand/rand and
7337 delete an unused file.
7340 *) Add support for the the free Netwide assembler (NASM) under Win32,
7341 since not many people have MASM (ml) and it can be hard to obtain.
7342 This is currently experimental but it seems to work OK and pass all
7343 the tests. Check out INSTALL.W32 for info.
7346 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7347 without temporary keys kept an extra copy of the server key,
7348 and connections with temporary keys did not free everything in case
7352 *) New function RSA_check_key and new openssl rsa option -check
7353 for verifying the consistency of RSA keys.
7354 [Ulf Moeller, Bodo Moeller]
7356 *) Various changes to make Win32 compile work:
7357 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7358 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7359 comparison" warnings.
7360 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7363 *) Add a debugging option to PKCS#5 v2 key generation function: when
7364 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7365 derived keys are printed to stderr.
7368 *) Copy the flags in ASN1_STRING_dup().
7369 [Roman E. Pavlov <pre@mo.msk.ru>]
7371 *) The x509 application mishandled signing requests containing DSA
7372 keys when the signing key was also DSA and the parameters didn't match.
7374 It was supposed to omit the parameters when they matched the signing key:
7375 the verifying software was then supposed to automatically use the CA's
7376 parameters if they were absent from the end user certificate.
7378 Omitting parameters is no longer recommended. The test was also
7379 the wrong way round! This was probably due to unusual behaviour in
7380 EVP_cmp_parameters() which returns 1 if the parameters match.
7381 This meant that parameters were omitted when they *didn't* match and
7382 the certificate was useless. Certificates signed with 'ca' didn't have
7384 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7386 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7387 The interface is as follows:
7388 Applications can use
7389 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7390 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7391 "off" is now the default.
7392 The library internally uses
7393 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7394 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7395 to disable memory-checking temporarily.
7397 Some inconsistent states that previously were possible (and were
7398 even the default) are now avoided.
7400 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7401 with each memory chunk allocated; this is occasionally more helpful
7402 than just having a counter.
7404 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7406 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7410 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7411 which largely parallels "options", but is for changing API behaviour,
7412 whereas "options" are about protocol behaviour.
7413 Initial "mode" flags are:
7415 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7416 a single record has been written.
7417 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7418 retries use the same buffer location.
7419 (But all of the contents must be
7423 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7426 *) Fix problems with no-hmac etc.
7427 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7429 *) New functions RSA_get_default_method(), RSA_set_method() and
7430 RSA_get_method(). These allows replacement of RSA_METHODs without having
7431 to mess around with the internals of an RSA structure.
7434 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7435 Also really enable memory leak checks in openssl.c and in some
7437 [Chad C. Mulligan, Bodo Moeller]
7439 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7440 up the length of negative integers. This has now been simplified to just
7441 store the length when it is first determined and use it later, rather
7442 than trying to keep track of where data is copied and updating it to
7444 [Steve Henson, reported by Brien Wheeler
7445 <bwheeler@authentica-security.com>]
7447 *) Add a new function PKCS7_signatureVerify. This allows the verification
7448 of a PKCS#7 signature but with the signing certificate passed to the
7449 function itself. This contrasts with PKCS7_dataVerify which assumes the
7450 certificate is present in the PKCS#7 structure. This isn't always the
7451 case: certificates can be omitted from a PKCS#7 structure and be
7452 distributed by "out of band" means (such as a certificate database).
7455 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7456 function prototypes in pem.h, also change util/mkdef.pl to add the
7457 necessary function names.
7460 *) mk1mf.pl (used by Windows builds) did not properly read the
7461 options set by Configure in the top level Makefile, and Configure
7462 was not even able to write more than one option correctly.
7463 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7466 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7467 file to be loaded from a BIO or FILE pointer. The BIO version will
7468 for example allow memory BIOs to contain config info.
7471 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7472 Whoever hopes to achieve shared-library compatibility across versions
7473 must use this, not the compile-time macro.
7474 (Exercise 0.9.4: Which is the minimum library version required by
7476 Note: All this applies only to multi-threaded programs, others don't
7480 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7481 through a BIO pair triggered the default case, i.e.
7482 SSLerr(...,SSL_R_UNKNOWN_STATE).
7485 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7486 can use the SSL library even if none of the specific BIOs is
7490 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7491 for the encoded length.
7492 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7494 *) Add initial documentation of the X509V3 functions.
7497 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7498 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7499 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7500 secure PKCS#8 private key format with a high iteration count.
7503 *) Fix determination of Perl interpreter: A perl or perl5
7504 _directory_ in $PATH was also accepted as the interpreter.
7505 [Ralf S. Engelschall]
7507 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7508 wrong with it but it was very old and did things like calling
7509 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7513 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7514 to use the new extension code.
7517 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7518 with macros. This should make it easier to change their form, add extra
7519 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7523 *) Add to configuration table a new entry that can specify an alternative
7524 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7525 according to Mark Crispin <MRC@Panda.COM>.
7529 *) DES CBC did not update the IV. Weird.
7532 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7533 Changing the behaviour of the former might break existing programs --
7534 where IV updating is needed, des_ncbc_encrypt can be used.
7537 *) When bntest is run from "make test" it drives bc to check its
7538 calculations, as well as internally checking them. If an internal check
7539 fails, it needs to cause bc to give a non-zero result or make test carries
7540 on without noticing the failure. Fixed.
7543 *) DES library cleanups.
7546 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7547 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7548 ciphers. NOTE: although the key derivation function has been verified
7549 against some published test vectors it has not been extensively tested
7550 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7554 *) Instead of "mkdir -p", which is not fully portable, use new
7555 Perl script "util/mkdir-p.pl".
7558 *) Rewrite the way password based encryption (PBE) is handled. It used to
7559 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7560 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7561 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7562 the 'parameter' field of the AlgorithmIdentifier is passed to the
7563 underlying key generation function so it must do its own ASN1 parsing.
7564 This has also changed the EVP_PBE_CipherInit() function which now has a
7565 'parameter' argument instead of literal salt and iteration count values
7566 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7569 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7570 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7571 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7572 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7573 value was just used as a "magic string" and not used directly its
7574 value doesn't matter.
7577 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7581 *) "linux-sparc64" configuration (ultrapenguin).
7582 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7583 "linux-sparc" configuration.
7584 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7586 *) config now generates no-xxx options for missing ciphers.
7589 *) Support the EBCDIC character set (work in progress).
7590 File ebcdic.c not yet included because it has a different license.
7591 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7593 *) Support BS2000/OSD-POSIX.
7594 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7596 *) Make callbacks for key generation use void * instead of char *.
7599 *) Make S/MIME samples compile (not yet tested).
7602 *) Additional typesafe stacks.
7605 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7609 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7611 *) New configuration variant "sco5-gcc".
7613 *) Updated some demos.
7614 [Sean O Riordain, Wade Scholine]
7616 *) Add missing BIO_free at exit of pkcs12 application.
7619 *) Fix memory leak in conf.c.
7622 *) Updates for Win32 to assembler version of MD5.
7625 *) Set #! path to perl in apps/der_chop to where we found it
7626 instead of using a fixed path.
7629 *) SHA library changes for irix64-mips4-cc.
7632 *) Improvements for VMS support.
7636 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7638 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7639 This also avoids the problems with SC4.2 and unpatched SC5.
7640 [Andy Polyakov <appro@fy.chalmers.se>]
7642 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7643 These are required because of the typesafe stack would otherwise break
7644 existing code. If old code used a structure member which used to be STACK
7645 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7646 sk_num or sk_value it would produce an error because the num, data members
7647 are not present in STACK_OF. Now it just produces a warning. sk_set
7648 replaces the old method of assigning a value to sk_value
7649 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7650 that does this will no longer work (and should use sk_set instead) but
7651 this could be regarded as a "questionable" behaviour anyway.
7654 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7655 correctly handle encrypted S/MIME data.
7658 *) Change type of various DES function arguments from des_cblock
7659 (which means, in function argument declarations, pointer to char)
7660 to des_cblock * (meaning pointer to array with 8 char elements),
7661 which allows the compiler to do more typechecking; it was like
7662 that back in SSLeay, but with lots of ugly casts.
7664 Introduce new type const_des_cblock.
7667 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7668 problems: find RecipientInfo structure that matches recipient certificate
7669 and initialise the ASN1 structures properly based on passed cipher.
7672 *) Belatedly make the BN tests actually check the results.
7675 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7676 to and from BNs: it was completely broken. New compilation option
7677 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7678 key elements as negative integers.
7681 *) Reorganize and speed up MD5.
7682 [Andy Polyakov <appro@fy.chalmers.se>]
7685 [Richard Levitte <richard@levitte.org>]
7687 *) New option -out to asn1parse to allow the parsed structure to be
7688 output to a file. This is most useful when combined with the -strparse
7689 option to examine the output of things like OCTET STRINGS.
7692 *) Make SSL library a little more fool-proof by not requiring any longer
7693 that SSL_set_{accept,connect}_state be called before
7694 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7695 in many applications because usually everything *appeared* to work as
7696 intended anyway -- now it really works as intended).
7699 *) Move openssl.cnf out of lib/.
7702 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7703 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7704 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7705 [Ralf S. Engelschall]
7707 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7708 handle PKCS#7 enveloped data properly.
7709 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7711 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7712 copying pointers. The cert_st handling is changed by this in
7713 various ways (and thus what used to be known as ctx->default_cert
7714 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7715 any longer when s->cert does not give us what we need).
7716 ssl_cert_instantiate becomes obsolete by this change.
7717 As soon as we've got the new code right (possibly it already is?),
7718 we have solved a couple of bugs of the earlier code where s->cert
7719 was used as if it could not have been shared with other SSL structures.
7721 Note that using the SSL API in certain dirty ways now will result
7722 in different behaviour than observed with earlier library versions:
7723 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7724 does not influence s as it used to.
7726 In order to clean up things more thoroughly, inside SSL_SESSION
7727 we don't use CERT any longer, but a new structure SESS_CERT
7728 that holds per-session data (if available); currently, this is
7729 the peer's certificate chain and, for clients, the server's certificate
7730 and temporary key. CERT holds only those values that can have
7731 meaningful defaults in an SSL_CTX.
7734 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7735 from the internal representation. Various PKCS#7 fixes: remove some
7736 evil casts and set the enc_dig_alg field properly based on the signing
7740 *) Allow PKCS#12 password to be set from the command line or the
7741 environment. Let 'ca' get its config file name from the environment
7742 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7746 *) Allow certificate policies extension to use an IA5STRING for the
7747 organization field. This is contrary to the PKIX definition but
7748 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7752 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7753 without disallowing inline assembler and the like for non-pedantic builds.
7756 *) Support Borland C++ builder.
7757 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7762 *) SHA-1 cleanups and performance enhancements.
7763 [Andy Polyakov <appro@fy.chalmers.se>]
7765 *) Sparc v8plus assembler for the bignum library.
7766 [Andy Polyakov <appro@fy.chalmers.se>]
7768 *) Accept any -xxx and +xxx compiler options in Configure.
7771 *) Update HPUX configuration.
7774 *) Add missing sk_<type>_unshift() function to safestack.h
7775 [Ralf S. Engelschall]
7777 *) New function SSL_CTX_use_certificate_chain_file that sets the
7778 "extra_cert"s in addition to the certificate. (This makes sense
7779 only for "PEM" format files, as chains as a whole are not
7783 *) Support verify_depth from the SSL API.
7784 x509_vfy.c had what can be considered an off-by-one-error:
7785 Its depth (which was not part of the external interface)
7786 was actually counting the number of certificates in a chain;
7787 now it really counts the depth.
7790 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7791 instead of X509err, which often resulted in confusing error
7792 messages since the error codes are not globally unique
7793 (e.g. an alleged error in ssl3_accept when a certificate
7794 didn't match the private key).
7796 *) New function SSL_CTX_set_session_id_context that allows to set a default
7797 value (so that you don't need SSL_set_session_id_context for each
7798 connection using the SSL_CTX).
7801 *) OAEP decoding bug fix.
7804 *) Support INSTALL_PREFIX for package builders, as proposed by
7808 *) New Configure options "threads" and "no-threads". For systems
7809 where the proper compiler options are known (currently Solaris
7810 and Linux), "threads" is the default.
7813 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7816 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7817 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7818 such as /usr/local/bin.
7821 *) "make linux-shared" to build shared libraries.
7822 [Niels Poppe <niels@netbox.org>]
7824 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7827 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7828 extension adding in x509 utility.
7831 *) Remove NOPROTO sections and error code comments.
7834 *) Partial rewrite of the DEF file generator to now parse the ANSI
7838 *) New Configure options --prefix=DIR and --openssldir=DIR.
7841 *) Complete rewrite of the error code script(s). It is all now handled
7842 by one script at the top level which handles error code gathering,
7843 header rewriting and C source file generation. It should be much better
7844 than the old method: it now uses a modified version of Ulf's parser to
7845 read the ANSI prototypes in all header files (thus the old K&R definitions
7846 aren't needed for error creation any more) and do a better job of
7847 translating function codes into names. The old 'ASN1 error code imbedded
7848 in a comment' is no longer necessary and it doesn't use .err files which
7849 have now been deleted. Also the error code call doesn't have to appear all
7850 on one line (which resulted in some large lines...).
7853 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7856 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7857 0 (which usually indicates a closed connection), but continue reading.
7860 *) Fix some race conditions.
7863 *) Add support for CRL distribution points extension. Add Certificate
7864 Policies and CRL distribution points documentation.
7867 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7870 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7871 8 of keying material. Merlin has also confirmed interop with this fix
7872 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7873 [Merlin Hughes <merlin@baltimore.ie>]
7875 *) Fix lots of warnings.
7876 [Richard Levitte <levitte@stacken.kth.se>]
7878 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7879 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7880 [Richard Levitte <levitte@stacken.kth.se>]
7882 *) Fix problems with sizeof(long) == 8.
7883 [Andy Polyakov <appro@fy.chalmers.se>]
7885 *) Change functions to ANSI C.
7888 *) Fix typos in error codes.
7889 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7891 *) Remove defunct assembler files from Configure.
7894 *) SPARC v8 assembler BIGNUM implementation.
7895 [Andy Polyakov <appro@fy.chalmers.se>]
7897 *) Support for Certificate Policies extension: both print and set.
7898 Various additions to support the r2i method this uses.
7901 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7902 return a const string when you are expecting an allocated buffer.
7905 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7906 types DirectoryString and DisplayText.
7909 *) Add code to allow r2i extensions to access the configuration database,
7910 add an LHASH database driver and add several ctx helper functions.
7913 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7914 fail when they extended the size of a BIGNUM.
7917 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7918 support typesafe stack.
7921 *) Fix typo in SSL_[gs]et_options().
7922 [Nils Frostberg <nils@medcom.se>]
7924 *) Delete various functions and files that belonged to the (now obsolete)
7925 old X509V3 handling code.
7928 *) New Configure option "rsaref".
7931 *) Don't auto-generate pem.h.
7934 *) Introduce type-safe ASN.1 SETs.
7937 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7938 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7940 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7941 that links with OpenSSL (well at least cause lots of warnings), but fear
7942 not: the conversion is trivial, and it eliminates loads of evil casts. A
7943 few STACKed things have been converted already. Feel free to convert more.
7944 In the fullness of time, I'll do away with the STACK type altogether.
7947 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7948 specified in <certfile> by updating the entry in the index.txt file.
7949 This way one no longer has to edit the index.txt file manually for
7950 revoking a certificate. The -revoke option does the gory details now.
7951 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7953 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7954 `-text' option at all and this way the `-noout -text' combination was
7955 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7956 [Ralf S. Engelschall]
7958 *) Make sure a corresponding plain text error message exists for the
7959 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7960 verify callback function determined that a certificate was revoked.
7961 [Ralf S. Engelschall]
7963 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7964 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7965 all available cipers including rc5, which was forgotten until now.
7966 In order to let the testing shell script know which algorithms
7967 are available, a new (up to now undocumented) command
7968 "openssl list-cipher-commands" is used.
7971 *) Bugfix: s_client occasionally would sleep in select() when
7972 it should have checked SSL_pending() first.
7975 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7976 the raw DSA values prior to ASN.1 encoding.
7979 *) Tweaks to Configure
7980 [Niels Poppe <niels@netbox.org>]
7982 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7986 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7989 *) New config option to avoid instructions that are illegal on the 80386.
7990 The default code is faster, but requires at least a 486.
7993 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7994 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7995 same as SSL2_VERSION anyway.
7998 *) New "-showcerts" option for s_client.
8001 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8002 application. Various cleanups and fixes.
8005 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8006 modify error routines to work internally. Add error codes and PBE init
8007 to library startup routines.
8010 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8011 packing functions to asn1 and evp. Changed function names and error
8012 codes along the way.
8015 *) PKCS12 integration: and so it begins... First of several patches to
8016 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8017 objects to objects.h
8020 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8021 and display support for Thawte strong extranet extension.
8024 *) Add LinuxPPC support.
8025 [Jeff Dubrule <igor@pobox.org>]
8027 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8028 bn_div_words in alpha.s.
8029 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8031 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8032 OAEP isn't supported when OpenSSL is built with RSAref.
8033 [Ulf Moeller <ulf@fitug.de>]
8035 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8036 so they no longer are missing under -DNOPROTO.
8037 [Soren S. Jorvang <soren@t.dk>]
8040 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8042 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8043 doesn't work when the session is reused. Coming soon!
8046 *) Fix a security hole, that allows sessions to be reused in the wrong
8047 context thus bypassing client cert protection! All software that uses
8048 client certs and session caches in multiple contexts NEEDS PATCHING to
8049 allow session reuse! A fuller solution is in the works.
8050 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8052 *) Some more source tree cleanups (removed obsolete files
8053 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8054 permission on "config" script to be executable) and a fix for the INSTALL
8056 [Ulf Moeller <ulf@fitug.de>]
8058 *) Remove some legacy and erroneous uses of malloc, free instead of
8060 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8062 *) Make rsa_oaep_test return non-zero on error.
8063 [Ulf Moeller <ulf@fitug.de>]
8065 *) Add support for native Solaris shared libraries. Configure
8066 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8067 if someone would make that last step automatic.
8068 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8070 *) ctx_size was not built with the right compiler during "make links". Fixed.
8073 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8074 except NULL ciphers". This means the default cipher list will no longer
8075 enable NULL ciphers. They need to be specifically enabled e.g. with
8076 the string "DEFAULT:eNULL".
8079 *) Fix to RSA private encryption routines: if p < q then it would
8080 occasionally produce an invalid result. This will only happen with
8081 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8084 *) Be less restrictive and allow also `perl util/perlpath.pl
8085 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8086 because this way one can also use an interpreter named `perl5' (which is
8087 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8088 installed as `perl').
8089 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8091 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8092 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8094 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8095 advapi32.lib to Win32 build and change the pem test comparision
8096 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8097 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8098 and crypto/des/ede_cbcm_enc.c.
8101 *) DES quad checksum was broken on big-endian architectures. Fixed.
8104 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8105 Win32 test batch file so it (might) work again. The Win32 test batch file
8106 is horrible: I feel ill....
8109 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8110 in e_os.h. Audit of header files to check ANSI and non ANSI
8111 sections: 10 functions were absent from non ANSI section and not exported
8112 from Windows DLLs. Fixed up libeay.num for new functions.
8115 *) Make `openssl version' output lines consistent.
8116 [Ralf S. Engelschall]
8118 *) Fix Win32 symbol export lists for BIO functions: Added
8119 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8120 to ms/libeay{16,32}.def.
8121 [Ralf S. Engelschall]
8123 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8124 fine under Unix and passes some trivial tests I've now added. But the
8125 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8126 added to make sure no one expects that this stuff really works in the
8127 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8128 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8130 [Ralf S. Engelschall]
8132 *) Fix the generation of two part addresses in perl.
8133 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8135 *) Add config entry for Linux on MIPS.
8136 [John Tobey <jtobey@channel1.com>]
8138 *) Make links whenever Configure is run, unless we are on Windoze.
8141 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8142 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8146 *) Add a useful kludge to allow package maintainers to specify compiler and
8147 other platforms details on the command line without having to patch the
8148 Configure script everytime: One now can use ``perl Configure
8149 <id>:<details>'', i.e. platform ids are allowed to have details appended
8150 to them (seperated by colons). This is treated as there would be a static
8151 pre-configured entry in Configure's %table under key <id> with value
8152 <details> and ``perl Configure <id>'' is called. So, when you want to
8153 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8154 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8155 now, which overrides the FreeBSD-elf entry on-the-fly.
8156 [Ralf S. Engelschall]
8158 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8161 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8162 on the `perl Configure ...' command line. This way one can compile
8163 OpenSSL libraries with Position Independent Code (PIC) which is needed
8164 for linking it into DSOs.
8165 [Ralf S. Engelschall]
8167 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8171 *) Cleaned up the LICENSE document: The official contact for any license
8172 questions now is the OpenSSL core team under openssl-core@openssl.org.
8173 And add a paragraph about the dual-license situation to make sure people
8174 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8175 to the OpenSSL toolkit.
8176 [Ralf S. Engelschall]
8178 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8179 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8180 Additonally cleaned up the `make links' target: Remove unnecessary
8181 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8182 to speed processing and no longer clutter the display with confusing
8183 stuff. Instead only the actually done links are displayed.
8184 [Ralf S. Engelschall]
8186 *) Permit null encryption ciphersuites, used for authentication only. It used
8187 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8188 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8192 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8193 signed attributes when verifying signatures (this would break them),
8194 the detached data encoding was wrong and public keys obtained using
8195 X509_get_pubkey() weren't freed.
8198 *) Add text documentation for the BUFFER functions. Also added a work around
8199 to a Win95 console bug. This was triggered by the password read stuff: the
8200 last character typed gets carried over to the next fread(). If you were
8201 generating a new cert request using 'req' for example then the last
8202 character of the passphrase would be CR which would then enter the first
8206 *) Added the new `Includes OpenSSL Cryptography Software' button as
8207 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8208 button and can be used by applications based on OpenSSL to show the
8209 relationship to the OpenSSL project.
8210 [Ralf S. Engelschall]
8212 *) Remove confusing variables in function signatures in files
8213 ssl/ssl_lib.c and ssl/ssl.h.
8214 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8216 *) Don't install bss_file.c under PREFIX/include/
8217 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8219 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8220 functions that return function pointers and has support for NT specific
8221 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8222 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8223 unsigned to signed types: this was killing the Win32 compile.
8226 *) Add new certificate file to stack functions,
8227 SSL_add_dir_cert_subjects_to_stack() and
8228 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8229 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8230 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8231 This means that Apache-SSL and similar packages don't have to mess around
8232 to add as many CAs as they want to the preferred list.
8235 *) Experiment with doxygen documentation. Currently only partially applied to
8237 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8238 openssl.doxy as the configuration file.
8241 *) Get rid of remaining C++-style comments which strict C compilers hate.
8242 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8244 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8245 compiled in by default: it has problems with large keys.
8248 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8249 DH private keys and/or callback functions which directly correspond to
8250 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8251 is needed for applications which have to configure certificates on a
8252 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8254 For the RSA certificate situation is makes no difference, but
8255 for the DSA certificate situation this fixes the "no shared cipher"
8256 problem where the OpenSSL cipher selection procedure failed because the
8257 temporary keys were not overtaken from the context and the API provided
8258 no way to reconfigure them.
8259 The new functions now let applications reconfigure the stuff and they
8260 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8261 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8262 non-public-API function ssl_cert_instantiate() is used as a helper
8263 function and also to reduce code redundancy inside ssl_rsa.c.
8264 [Ralf S. Engelschall]
8266 *) Move s_server -dcert and -dkey options out of the undocumented feature
8267 area because they are useful for the DSA situation and should be
8268 recognized by the users.
8269 [Ralf S. Engelschall]
8271 *) Fix the cipher decision scheme for export ciphers: the export bits are
8272 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8273 SSL_EXP_MASK. So, the original variable has to be used instead of the
8274 already masked variable.
8275 [Richard Levitte <levitte@stacken.kth.se>]
8277 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8278 [Richard Levitte <levitte@stacken.kth.se>]
8280 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8281 from `int' to `unsigned int' because it's a length and initialized by
8282 EVP_DigestFinal() which expects an `unsigned int *'.
8283 [Richard Levitte <levitte@stacken.kth.se>]
8285 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8286 script. Instead use the usual Shell->Perl transition trick.
8287 [Ralf S. Engelschall]
8289 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8290 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8291 -noout -modulus' as it's already the case for `openssl rsa -noout
8292 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8293 currently the public key is printed (a decision which was already done by
8294 `openssl dsa -modulus' in the past) which serves a similar purpose.
8295 Additionally the NO_RSA no longer completely removes the whole -modulus
8296 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8298 [Ralf S. Engelschall]
8300 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8301 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8302 [Arne Ansper <arne@ats.cyber.ee>]
8304 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8305 to be added. Now both 'req' and 'ca' can use new objects defined in the
8309 *) Add cool BIO that does syslog (or event log on NT).
8310 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8312 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8313 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8314 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8315 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8318 *) Add preliminary config info for new extension code.
8321 *) Make RSA_NO_PADDING really use no padding.
8322 [Ulf Moeller <ulf@fitug.de>]
8324 *) Generate errors when private/public key check is done.
8327 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8328 for some CRL extensions and new objects added.
8331 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8332 key usage extension and fuller support for authority key id.
8335 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8336 padding method for RSA, which is recommended for new applications in PKCS
8337 #1 v2.0 (RFC 2437, October 1998).
8338 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8339 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8340 against Bleichbacher's attack on RSA.
8341 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8344 *) Updates to the new SSL compression code
8345 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8347 *) Fix so that the version number in the master secret, when passed
8348 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8349 (because the server will not accept higher), that the version number
8350 is 0x03,0x01, not 0x03,0x00
8351 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8353 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8354 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8355 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8358 *) Support for RAW extensions where an arbitrary extension can be
8359 created by including its DER encoding. See apps/openssl.cnf for
8363 *) Make sure latest Perl versions don't interpret some generated C array
8364 code as Perl array code in the crypto/err/err_genc.pl script.
8365 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8367 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8368 not many people have the assembler. Various Win32 compilation fixes and
8369 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8373 *) Modify configure script 'Configure' to automatically create crypto/date.h
8374 file under Win32 and also build pem.h from pem.org. New script
8375 util/mkfiles.pl to create the MINFO file on environments that can't do a
8376 'make files': perl util/mkfiles.pl >MINFO should work.
8379 *) Major rework of DES function declarations, in the pursuit of correctness
8380 and purity. As a result, many evil casts evaporated, and some weirdness,
8381 too. You may find this causes warnings in your code. Zapping your evil
8382 casts will probably fix them. Mostly.
8385 *) Fix for a typo in asn1.h. Bug fix to object creation script
8386 obj_dat.pl. It considered a zero in an object definition to mean
8387 "end of object": none of the objects in objects.h have any zeros
8388 so it wasn't spotted.
8389 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8391 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8392 Masking (CBCM). In the absence of test vectors, the best I have been able
8393 to do is check that the decrypt undoes the encrypt, so far. Send me test
8394 vectors if you have them.
8397 *) Correct calculation of key length for export ciphers (too much space was
8398 allocated for null ciphers). This has not been tested!
8401 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8402 message is now correct (it understands "crypto" and "ssl" on its
8403 command line). There is also now an "update" option. This will update
8404 the util/ssleay.num and util/libeay.num files with any new functions.
8406 perl util/mkdef.pl crypto ssl update
8407 it will update them.
8410 *) Overhauled the Perl interface (perl/*):
8411 - ported BN stuff to OpenSSL's different BN library
8412 - made the perl/ source tree CVS-aware
8413 - renamed the package from SSLeay to OpenSSL (the files still contain
8414 their history because I've copied them in the repository)
8415 - removed obsolete files (the test scripts will be replaced
8416 by better Test::Harness variants in the future)
8417 [Ralf S. Engelschall]
8419 *) First cut for a very conservative source tree cleanup:
8420 1. merge various obsolete readme texts into doc/ssleay.txt
8421 where we collect the old documents and readme texts.
8422 2. remove the first part of files where I'm already sure that we no
8423 longer need them because of three reasons: either they are just temporary
8424 files which were left by Eric or they are preserved original files where
8425 I've verified that the diff is also available in the CVS via "cvs diff
8426 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8427 the crypto/md/ stuff).
8428 [Ralf S. Engelschall]
8430 *) More extension code. Incomplete support for subject and issuer alt
8431 name, issuer and authority key id. Change the i2v function parameters
8432 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8433 what that's for :-) Fix to ASN1 macro which messed up
8434 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8437 *) Preliminary support for ENUMERATED type. This is largely copied from the
8441 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8442 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8444 *) Make sure `make rehash' target really finds the `openssl' program.
8445 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8447 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8448 like to hear about it if this slows down other processors.
8451 *) Add CygWin32 platform information to Configure script.
8452 [Alan Batie <batie@aahz.jf.intel.com>]
8454 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8455 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8457 *) New program nseq to manipulate netscape certificate sequences
8460 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8464 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8465 but the BN code had some problems that would cause failures when
8466 doing certificate verification and some other functions.
8467 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8469 *) Add ASN1 and PEM code to support netscape certificate sequences.
8472 *) Add ASN1 and PEM code to support netscape certificate sequences.
8475 *) Add several PKIX and private extended key usage OIDs.
8478 *) Modify the 'ca' program to handle the new extension code. Modify
8479 openssl.cnf for new extension format, add comments.
8482 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8483 and add a sample to openssl.cnf so req -x509 now adds appropriate
8487 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8488 error code, add initial support to X509_print() and x509 application.
8491 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8492 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8493 stuff is currently isolated and isn't even compiled yet.
8496 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8497 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8498 Removed the versions check from X509 routines when loading extensions:
8499 this allows certain broken certificates that don't set the version
8500 properly to be processed.
8503 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8504 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8505 can still be regenerated with "make depend".
8508 *) Spelling mistake in C version of CAST-128.
8509 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8511 *) Changes to the error generation code. The perl script err-code.pl
8512 now reads in the old error codes and retains the old numbers, only
8513 adding new ones if necessary. It also only changes the .err files if new
8514 codes are added. The makefiles have been modified to only insert errors
8515 when needed (to avoid needlessly modifying header files). This is done
8516 by only inserting errors if the .err file is newer than the auto generated
8517 C file. To rebuild all the error codes from scratch (the old behaviour)
8518 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8519 or delete all the .err files.
8522 *) CAST-128 was incorrectly implemented for short keys. The C version has
8523 been fixed, but is untested. The assembler versions are also fixed, but
8524 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8525 to regenerate it if needed.
8526 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8527 Hagino <itojun@kame.net>]
8529 *) File was opened incorrectly in randfile.c.
8530 [Ulf Möller <ulf@fitug.de>]
8532 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8533 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8534 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8535 al: it's just almost always a UTCTime. Note this patch adds new error
8536 codes so do a "make errors" if there are problems.
8539 *) Correct Linux 1 recognition in config.
8540 [Ulf Möller <ulf@fitug.de>]
8542 *) Remove pointless MD5 hash when using DSA keys in ca.
8543 [Anonymous <nobody@replay.com>]
8545 *) Generate an error if given an empty string as a cert directory. Also
8546 generate an error if handed NULL (previously returned 0 to indicate an
8547 error, but didn't set one).
8548 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8550 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8553 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8554 parameters. This was causing a warning which killed off the Win32 compile.
8557 *) Remove C++ style comments from crypto/bn/bn_local.h.
8558 [Neil Costigan <neil.costigan@celocom.com>]
8560 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8561 based on a text string, looking up short and long names and finally
8562 "dot" format. The "dot" format stuff didn't work. Added new function
8563 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8564 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8565 OID is not part of the table.
8568 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8569 X509_LOOKUP_by_alias().
8572 *) Sort openssl functions by name.
8575 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8576 encryption from sample DSA keys (in case anyone is interested the password
8580 *) Make _all_ *_free functions accept a NULL pointer.
8581 [Frans Heymans <fheymans@isaserver.be>]
8583 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8585 [Anonymous <nobody@replay.com>]
8587 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8588 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8590 *) Don't blow it for numeric -newkey arguments to apps/req.
8591 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8593 *) Temp key "for export" tests were wrong in s3_srvr.c.
8594 [Anonymous <nobody@replay.com>]
8596 *) Add prototype for temp key callback functions
8597 SSL_CTX_set_tmp_{rsa,dh}_callback().
8600 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8601 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8604 *) X509_name_add_entry() freed the wrong thing after an error.
8605 [Arne Ansper <arne@ats.cyber.ee>]
8607 *) rsa_eay.c would attempt to free a NULL context.
8608 [Arne Ansper <arne@ats.cyber.ee>]
8610 *) BIO_s_socket() had a broken should_retry() on Windoze.
8611 [Arne Ansper <arne@ats.cyber.ee>]
8613 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8614 [Arne Ansper <arne@ats.cyber.ee>]
8616 *) Make sure the already existing X509_STORE->depth variable is initialized
8617 in X509_STORE_new(), but document the fact that this variable is still
8618 unused in the certificate verification process.
8619 [Ralf S. Engelschall]
8621 *) Fix the various library and apps files to free up pkeys obtained from
8622 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8625 *) Fix reference counting in X509_PUBKEY_get(). This makes
8626 demos/maurice/example2.c work, amongst others, probably.
8627 [Steve Henson and Ben Laurie]
8629 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8630 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8631 are no longer created. This way we have a single and consistent command
8632 line interface `openssl <command>', similar to `cvs <command>'.
8633 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8635 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8636 BIT STRING wrapper always have zero unused bits.
8639 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8642 *) Make the top-level INSTALL documentation easier to understand.
8645 *) Makefiles updated to exit if an error occurs in a sub-directory
8646 make (including if user presses ^C) [Paul Sutton]
8648 *) Make Montgomery context stuff explicit in RSA data structure.
8651 *) Fix build order of pem and err to allow for generated pem.h.
8654 *) Fix renumbering bug in X509_NAME_delete_entry().
8657 *) Enhanced the err-ins.pl script so it makes the error library number
8658 global and can add a library name. This is needed for external ASN1 and
8659 other error libraries.
8662 *) Fixed sk_insert which never worked properly.
8665 *) Fix ASN1 macros so they can handle indefinite length construted
8666 EXPLICIT tags. Some non standard certificates use these: they can now
8670 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8671 into a single doc/ssleay.txt bundle. This way the information is still
8672 preserved but no longer messes up this directory. Now it's new room for
8673 the new set of documenation files.
8674 [Ralf S. Engelschall]
8676 *) SETs were incorrectly DER encoded. This was a major pain, because they
8677 shared code with SEQUENCEs, which aren't coded the same. This means that
8678 almost everything to do with SETs or SEQUENCEs has either changed name or
8679 number of arguments.
8680 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8682 *) Fix test data to work with the above.
8685 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8686 was already fixed by Eric for 0.9.1 it seems.
8687 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8689 *) Autodetect FreeBSD3.
8692 *) Fix various bugs in Configure. This affects the following platforms:
8696 unixware-2.0-pentium
8700 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8701 before they are needed.
8704 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8708 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8710 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8711 changed SSLeay to OpenSSL in version strings.
8712 [Ralf S. Engelschall]
8714 *) Some fixups to the top-level documents.
8717 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8718 because the symlink to include/ was missing.
8719 [Ralf S. Engelschall]
8721 *) Incorporated the popular no-RSA/DSA-only patches
8722 which allow to compile a RSA-free SSLeay.
8723 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8725 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8726 when "ssleay" is still not found.
8727 [Ralf S. Engelschall]
8729 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8730 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8732 *) Updated the README file.
8733 [Ralf S. Engelschall]
8735 *) Added various .cvsignore files in the CVS repository subdirs
8736 to make a "cvs update" really silent.
8737 [Ralf S. Engelschall]
8739 *) Recompiled the error-definition header files and added
8740 missing symbols to the Win32 linker tables.
8741 [Ralf S. Engelschall]
8743 *) Cleaned up the top-level documents;
8744 o new files: CHANGES and LICENSE
8745 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8746 o merged COPYRIGHT into LICENSE
8747 o removed obsolete TODO file
8748 o renamed MICROSOFT to INSTALL.W32
8749 [Ralf S. Engelschall]
8751 *) Removed dummy files from the 0.9.1b source tree:
8752 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8753 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8754 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8755 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8756 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8757 [Ralf S. Engelschall]
8759 *) Added various platform portability fixes.
8762 *) The Genesis of the OpenSSL rpject:
8763 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8764 Young and Tim J. Hudson created while they were working for C2Net until
8766 [The OpenSSL Project]
8769 Changes between 0.9.0b and 0.9.1b [not released]
8771 *) Updated a few CA certificates under certs/
8774 *) Changed some BIGNUM api stuff.
8777 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8778 DGUX x86, Linux Alpha, etc.
8781 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8782 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8786 *) Add -strparse option to asn1pars program which parses nested
8788 [Dr Stephen Henson <shenson@bigfoot.com>]
8790 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8793 *) DSA fix for "ca" program.
8796 *) Added "-genkey" option to "dsaparam" program.
8799 *) Added RIPE MD160 (rmd160) message digest.
8802 *) Added -a (all) option to "ssleay version" command.
8805 *) Added PLATFORM define which is the id given to Configure.
8808 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8811 *) Extended the ASN.1 parser routines.
8814 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8817 *) Added a BN_CTX to the BN library.
8820 *) Fixed the weak key values in DES library
8823 *) Changed API in EVP library for cipher aliases.
8826 *) Added support for RC2/64bit cipher.
8829 *) Converted the lhash library to the crypto/mem.c functions.
8832 *) Added more recognized ASN.1 object ids.
8835 *) Added more RSA padding checks for SSL/TLS.
8838 *) Added BIO proxy/filter functionality.
8841 *) Added extra_certs to SSL_CTX which can be used
8842 send extra CA certificates to the client in the CA cert chain sending
8843 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8846 *) Now Fortezza is denied in the authentication phase because
8847 this is key exchange mechanism is not supported by SSLeay at all.
8850 *) Additional PKCS1 checks.
8853 *) Support the string "TLSv1" for all TLS v1 ciphers.
8856 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8857 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8860 *) Fixed a few memory leaks.
8863 *) Fixed various code and comment typos.
8866 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8867 bytes sent in the client random.
8868 [Edward Bishop <ebishop@spyglass.com>]