5 Changes between 0.9.8zd and 0.9.8ze [xx XXX xxxx]
9 Changes between 0.9.8zc and 0.9.8zd [8 Jan 2015]
11 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
12 message can cause a segmentation fault in OpenSSL due to a NULL pointer
13 dereference. This could lead to a Denial Of Service attack. Thanks to
14 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
18 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
19 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
20 method would be set to NULL which could later result in a NULL pointer
21 dereference. Thanks to Frank Schmirler for reporting this issue.
25 *) Abort handshake if server key exchange message is omitted for ephemeral
28 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
33 *) Remove non-export ephemeral RSA code on client and server. This code
34 violated the TLS standard by allowing the use of temporary RSA keys in
35 non-export ciphersuites and could be used by a server to effectively
36 downgrade the RSA key length used to a value smaller than the server
37 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
38 INRIA or reporting this issue.
42 *) Fix various certificate fingerprint issues.
44 By using non-DER or invalid encodings outside the signed portion of a
45 certificate the fingerprint can be changed without breaking the signature.
46 Although no details of the signed portion of the certificate can be changed
47 this can cause problems with some applications: e.g. those using the
48 certificate fingerprint for blacklists.
50 1. Reject signatures with non zero unused bits.
52 If the BIT STRING containing the signature has non zero unused bits reject
53 the signature. All current signature algorithms require zero unused bits.
55 2. Check certificate algorithm consistency.
57 Check the AlgorithmIdentifier inside TBS matches the one in the
58 certificate signature. NB: this will result in signature failure
59 errors for some broken certificates.
61 Thanks to Konrad Kraszewski from Google for reporting this issue.
63 3. Check DSA/ECDSA signatures use DER.
65 Reencode DSA/ECDSA signatures and compare with the original received
66 signature. Return an error if there is a mismatch.
68 This will reject various cases including garbage after signature
69 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
70 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
71 (negative or with leading zeroes).
73 Further analysis was conducted and fixes were developed by Stephen Henson
74 of the OpenSSL core team.
79 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
80 results on some platforms, including x86_64. This bug occurs at random
81 with a very low probability, and is not known to be exploitable in any
82 way, though its exact impact is difficult to determine. Thanks to Pieter
83 Wuille (Blockstream) who reported this issue and also suggested an initial
84 fix. Further analysis was conducted by the OpenSSL development team and
85 Adam Langley of Google. The final fix was developed by Andy Polyakov of
86 the OpenSSL core team.
90 Changes between 0.9.8zb and 0.9.8zc [15 Oct 2014]
92 *) Session Ticket Memory Leak.
94 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
95 integrity of that ticket is first verified. In the event of a session
96 ticket integrity check failing, OpenSSL will fail to free memory
97 causing a memory leak. By sending a large number of invalid session
98 tickets an attacker could exploit this issue in a Denial Of Service
103 *) Build option no-ssl3 is incomplete.
105 When OpenSSL is configured with "no-ssl3" as a build option, servers
106 could accept and complete a SSL 3.0 handshake, and clients could be
107 configured to send them.
109 [Akamai and the OpenSSL team]
111 *) Add support for TLS_FALLBACK_SCSV.
112 Client applications doing fallback retries should call
113 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
115 [Adam Langley, Bodo Moeller]
117 *) Add additional DigestInfo checks.
119 Reencode DigestInto in DER and check against the original when
120 verifying RSA signature: this will reject any improperly encoded
121 DigestInfo structures.
123 Note: this is a precautionary measure and no attacks are currently known.
127 Changes between 0.9.8za and 0.9.8zb [6 Aug 2014]
129 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
130 to a denial of service attack. A malicious server can crash the client
131 with a null pointer dereference (read) by specifying an anonymous (EC)DH
132 ciphersuite and sending carefully crafted handshake messages.
134 Thanks to Felix Gröbert (Google) for discovering and researching this
139 *) By sending carefully crafted DTLS packets an attacker could cause openssl
140 to leak memory. This can be exploited through a Denial of Service attack.
141 Thanks to Adam Langley for discovering and researching this issue.
145 *) An attacker can force openssl to consume large amounts of memory whilst
146 processing DTLS handshake messages. This can be exploited through a
147 Denial of Service attack.
148 Thanks to Adam Langley for discovering and researching this issue.
152 *) An attacker can force an error condition which causes openssl to crash
153 whilst processing DTLS packets due to memory being freed twice. This
154 can be exploited through a Denial of Service attack.
155 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
160 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
161 X509_name_oneline, X509_name_print_ex et al. to leak some information
162 from the stack. Applications may be affected if they echo pretty printing
163 output to the attacker.
165 Thanks to Ivan Fratric (Google) for discovering this issue.
167 [Emilia Käsper, and Steve Henson]
169 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
170 for corner cases. (Certain input points at infinity could lead to
171 bogus results, with non-infinity inputs mapped to infinity too.)
174 Changes between 0.9.8y and 0.9.8za [5 Jun 2014]
176 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
177 handshake can force the use of weak keying material in OpenSSL
178 SSL/TLS clients and servers.
180 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
181 researching this issue. (CVE-2014-0224)
182 [KIKUCHI Masashi, Steve Henson]
184 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
185 OpenSSL DTLS client the code can be made to recurse eventually crashing
188 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
190 [Imre Rad, Steve Henson]
192 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
193 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
194 client or server. This is potentially exploitable to run arbitrary
195 code on a vulnerable client or server.
197 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
198 [Jüri Aedla, Steve Henson]
200 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
201 are subject to a denial of service attack.
203 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
204 this issue. (CVE-2014-3470)
205 [Felix Gröbert, Ivan Fratric, Steve Henson]
207 *) Fix for the attack described in the paper "Recovering OpenSSL
208 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
209 by Yuval Yarom and Naomi Benger. Details can be obtained from:
210 http://eprint.iacr.org/2014/140
212 Thanks to Yuval Yarom and Naomi Benger for discovering this
213 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
214 [Yuval Yarom and Naomi Benger]
216 Thanks to mancha for backporting the fix to the 0.9.8 branch.
218 *) Fix handling of warning-level alerts in SSL23 client mode so they
219 don't cause client-side termination (eg. on SNI unrecognized_name
220 warnings). Add client and server support for six additional alerts
221 per RFC 6066 and RFC 4279.
224 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
225 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
226 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
227 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
228 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
229 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
230 [Rob Stradling, Adam Langley]
232 Changes between 0.9.8x and 0.9.8y [5 Feb 2013]
234 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
236 This addresses the flaw in CBC record processing discovered by
237 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
238 at: http://www.isg.rhul.ac.uk/tls/
240 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
241 Security Group at Royal Holloway, University of London
242 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
243 Emilia Käsper for the initial patch.
245 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
247 *) Return an error when checking OCSP signatures when key is NULL.
248 This fixes a DoS attack. (CVE-2013-0166)
251 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
252 the right response is stapled. Also change SSL_get_certificate()
253 so it returns the certificate actually sent.
254 See http://rt.openssl.org/Ticket/Display.html?id=2836.
256 [Rob Stradling <rob.stradling@comodo.com>]
258 *) Fix possible deadlock when decoding public keys.
261 Changes between 0.9.8w and 0.9.8x [10 May 2012]
263 *) Sanity check record length before skipping explicit IV in DTLS
266 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
267 fuzzing as a service testing platform.
271 *) Initialise tkeylen properly when encrypting CMS messages.
272 Thanks to Solar Designer of Openwall for reporting this issue.
275 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
277 *) The fix for CVE-2012-2110 did not take into account that the
278 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
279 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
280 rejecting negative len parameter. (CVE-2012-2131)
281 [Tomas Hoger <thoger@redhat.com>]
283 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
285 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
286 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
287 in CRYPTO_realloc_clean.
289 Thanks to Tavis Ormandy, Google Security Team, for discovering this
290 issue and to Adam Langley <agl@chromium.org> for fixing it.
292 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
294 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
296 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
297 in CMS and PKCS7 code. When RSA decryption fails use a random key for
298 content decryption and always return the same error. Note: this attack
299 needs on average 2^20 messages so it only affects automated senders. The
300 old behaviour can be reenabled in the CMS code by setting the
301 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
302 an MMA defence is not necessary.
303 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
304 this issue. (CVE-2012-0884)
307 *) Fix CVE-2011-4619: make sure we really are receiving a
308 client hello before rejecting multiple SGC restarts. Thanks to
309 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
312 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
314 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
315 Thanks to Antonio Martin, Enterprise Secure Access Research and
316 Development, Cisco Systems, Inc. for discovering this bug and
317 preparing a fix. (CVE-2012-0050)
320 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
322 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
323 of the Vaudenay padding oracle attack on CBC mode encryption
324 which enables an efficient plaintext recovery attack against
325 the OpenSSL implementation of DTLS. Their attack exploits timing
326 differences arising during decryption processing. A research
327 paper describing this attack can be found at:
328 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
329 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
330 Security Group at Royal Holloway, University of London
331 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
332 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
333 for preparing the fix. (CVE-2011-4108)
334 [Robin Seggelmann, Michael Tuexen]
336 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
337 [Ben Laurie, Kasper <ekasper@google.com>]
339 *) Clear bytes used for block padding of SSL 3.0 records.
341 [Adam Langley (Google)]
343 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
344 Kadianakis <desnacked@gmail.com> for discovering this issue and
345 Adam Langley for preparing the fix. (CVE-2011-4619)
346 [Adam Langley (Google)]
348 *) Prevent malformed RFC3779 data triggering an assertion failure.
349 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
350 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
351 [Rob Austein <sra@hactrn.net>]
353 *) Fix ssl_ciph.c set-up race.
354 [Adam Langley (Google)]
356 *) Fix spurious failures in ecdsatest.c.
357 [Emilia Käsper (Google)]
359 *) Fix the BIO_f_buffer() implementation (which was mixing different
360 interpretations of the '..._len' fields).
361 [Adam Langley (Google)]
363 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
364 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
365 threads won't reuse the same blinding coefficients.
367 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
368 lock to call BN_BLINDING_invert_ex, and avoids one use of
369 BN_BLINDING_update for each BN_BLINDING structure (previously,
370 the last update always remained unused).
371 [Emilia Käsper (Google)]
373 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
374 for multi-threaded use of ECDH.
375 [Adam Langley (Google)]
377 *) Fix x509_name_ex_d2i memory leak on bad inputs.
380 *) Add protection against ECDSA timing attacks as mentioned in the paper
381 by Billy Bob Brumley and Nicola Tuveri, see:
383 http://eprint.iacr.org/2011/232.pdf
385 [Billy Bob Brumley and Nicola Tuveri]
387 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
389 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
390 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
392 *) Fix bug in string printing code: if *any* escaping is enabled we must
393 escape the escape character (backslash) or the resulting string is
397 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
399 *) Disable code workaround for ancient and obsolete Netscape browsers
400 and servers: an attacker can use it in a ciphersuite downgrade attack.
401 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
404 *) Fixed J-PAKE implementation error, originally discovered by
405 Sebastien Martini, further info and confirmation from Stefan
406 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
409 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
411 *) Fix extension code to avoid race conditions which can result in a buffer
412 overrun vulnerability: resumed sessions must not be modified as they can
413 be shared by multiple threads. CVE-2010-3864
416 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
419 *) Don't reencode certificate when calculating signature: cache and use
420 the original encoding instead. This makes signature verification of
421 some broken encodings work correctly.
424 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
425 is also one of the inputs.
426 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
428 *) Don't repeatedly append PBE algorithms to table if they already exist.
429 Sort table on each new add. This effectively makes the table read only
430 after all algorithms are added and subsequent calls to PKCS12_pbe_add
434 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
436 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
439 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
440 access or freeing data twice (CVE-2010-0742)
441 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
443 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
444 common in certificates and some applications which only call
445 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
449 Reduce copying into .apps and .test in makevms.com
450 Don't try to use blank CA certificate in CA.com
451 Allow use of C files from original directories in maketests.com
452 [Steven M. Schweda" <sms@antinode.info>]
454 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
456 *) When rejecting SSL/TLS records due to an incorrect version number, never
457 update s->server with a new major version number. As of
458 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
459 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
460 the previous behavior could result in a read attempt at NULL when
461 receiving specific incorrect SSL/TLS records once record payload
462 protection is active. (CVE-2010-0740)
463 [Bodo Moeller, Adam Langley <agl@chromium.org>]
465 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
466 could be crashed if the relevant tables were not present (e.g. chrooted).
467 [Tomas Hoger <thoger@redhat.com>]
469 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
471 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
472 [Martin Olsson, Neel Mehta]
474 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
475 accommodate for stack sorting, always a write lock!).
478 *) On some versions of WIN32 Heap32Next is very slow. This can cause
479 excessive delays in the RAND_poll(): over a minute. As a workaround
480 include a time check in the inner Heap32Next loop too.
483 *) The code that handled flushing of data in SSL/TLS originally used the
484 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
485 the problem outlined in PR#1949. The fix suggested there however can
486 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
487 of Apache). So instead simplify the code to flush unconditionally.
488 This should be fine since flushing with no data to flush is a no op.
491 *) Handle TLS versions 2.0 and later properly and correctly use the
492 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
493 off ancient servers have a habit of sticking around for a while...
496 *) Modify compression code so it frees up structures without using the
497 ex_data callbacks. This works around a problem where some applications
498 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
499 restarting) then use compression (e.g. SSL with compression) later.
500 This results in significant per-connection memory leaks and
501 has caused some security issues including CVE-2008-1678 and
505 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
506 change when encrypting or decrypting.
509 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
510 connect and renegotiate with servers which do not support RI.
511 Until RI is more widely deployed this option is enabled by default.
514 *) Add "missing" ssl ctrls to clear options and mode.
517 *) If client attempts to renegotiate and doesn't support RI respond with
518 a no_renegotiation alert as required by RFC5746. Some renegotiating
519 TLS clients will continue a connection gracefully when they receive
520 the alert. Unfortunately OpenSSL mishandled this alert and would hang
521 waiting for a server hello which it will never receive. Now we treat a
522 received no_renegotiation alert as a fatal error. This is because
523 applications requesting a renegotiation might well expect it to succeed
524 and would have no code in place to handle the server denying it so the
525 only safe thing to do is to terminate the connection.
528 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
529 peer supports secure renegotiation and 0 otherwise. Print out peer
530 renegotiation support in s_client/s_server.
533 *) Replace the highly broken and deprecated SPKAC certification method with
534 the updated NID creation version. This should correctly handle UTF8.
537 *) Implement RFC5746. Re-enable renegotiation but require the extension
538 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
539 turns out to be a bad idea. It has been replaced by
540 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
541 SSL_CTX_set_options(). This is really not recommended unless you
542 know what you are doing.
543 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
545 *) Fixes to stateless session resumption handling. Use initial_ctx when
546 issuing and attempting to decrypt tickets in case it has changed during
547 servername handling. Use a non-zero length session ID when attempting
548 stateless session resumption: this makes it possible to determine if
549 a resumption has occurred immediately after receiving server hello
550 (several places in OpenSSL subtly assume this) instead of later in
554 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
555 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
556 fixes for a few places where the return code is not checked
558 [Julia Lawall <julia@diku.dk>]
560 *) Add --strict-warnings option to Configure script to include devteam
561 warnings in other configurations.
564 *) Add support for --libdir option and LIBDIR variable in makefiles. This
565 makes it possible to install openssl libraries in locations which
566 have names other than "lib", for example "/usr/lib64" which some
568 [Steve Henson, based on patch from Jeremy Utley]
570 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
571 X690 8.9.12 and can produce some misleading textual output of OIDs.
572 [Steve Henson, reported by Dan Kaminsky]
574 *) Delete MD2 from algorithm tables. This follows the recommendation in
575 several standards that it is not used in new applications due to
576 several cryptographic weaknesses. For binary compatibility reasons
577 the MD2 API is still compiled in by default.
580 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
584 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
585 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
587 [Guenter <lists@gknw.net>]
589 *) Fix the server certificate chain building code to use X509_verify_cert(),
590 it used to have an ad-hoc builder which was unable to cope with anything
591 other than a simple chain.
592 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
594 *) Don't check self signed certificate signatures in X509_verify_cert()
595 by default (a flag can override this): it just wastes time without
596 adding any security. As a useful side effect self signed root CAs
597 with non-FIPS digests are now usable in FIPS mode.
600 *) In dtls1_process_out_of_seq_message() the check if the current message
601 is already buffered was missing. For every new message was memory
602 allocated, allowing an attacker to perform an denial of service attack
603 with sending out of seq handshake messages until there is no memory
604 left. Additionally every future messege was buffered, even if the
605 sequence number made no sense and would be part of another handshake.
606 So only messages with sequence numbers less than 10 in advance will be
607 buffered. (CVE-2009-1378)
608 [Robin Seggelmann, discovered by Daniel Mentz]
610 *) Records are buffered if they arrive with a future epoch to be
611 processed after finishing the corresponding handshake. There is
612 currently no limitation to this buffer allowing an attacker to perform
613 a DOS attack with sending records with future epochs until there is no
614 memory left. This patch adds the pqueue_size() function to detemine
615 the size of a buffer and limits the record buffer to 100 entries.
617 [Robin Seggelmann, discovered by Daniel Mentz]
619 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
620 parent structure is freed. (CVE-2009-1379)
623 *) Handle non-blocking I/O properly in SSL_shutdown() call.
624 [Darryl Miles <darryl-mailinglists@netbauds.net>]
627 [Ilya O. <vrghost@gmail.com>]
629 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
631 *) Disable renegotiation completely - this fixes a severe security
632 problem (CVE-2009-3555) at the cost of breaking all
633 renegotiation. Renegotiation can be re-enabled by setting
634 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
635 run-time. This is really not recommended unless you know what
639 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
641 *) Don't set val to NULL when freeing up structures, it is freed up by
642 underlying code. If sizeof(void *) > sizeof(long) this can result in
643 zeroing past the valid field. (CVE-2009-0789)
644 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
646 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
647 checked correctly. This would allow some invalid signed attributes to
648 appear to verify correctly. (CVE-2009-0591)
649 [Ivan Nestlerode <inestlerode@us.ibm.com>]
651 *) Reject UniversalString and BMPString types with invalid lengths. This
652 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
653 a legal length. (CVE-2009-0590)
656 *) Set S/MIME signing as the default purpose rather than setting it
657 unconditionally. This allows applications to override it at the store
661 *) Permit restricted recursion of ASN1 strings. This is needed in practice
662 to handle some structures.
665 *) Improve efficiency of mem_gets: don't search whole buffer each time
667 [Jeremy Shapiro <jnshapir@us.ibm.com>]
669 *) New -hex option for openssl rand.
672 *) Print out UTF8String and NumericString when parsing ASN1.
675 *) Support NumericString type for name components.
678 *) Allow CC in the environment to override the automatically chosen
679 compiler. Note that nothing is done to ensure flags work with the
683 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
685 *) Properly check EVP_VerifyFinal() and similar return values
687 [Ben Laurie, Bodo Moeller, Google Security Team]
689 *) Enable TLS extensions by default.
692 *) Allow the CHIL engine to be loaded, whether the application is
693 multithreaded or not. (This does not release the developer from the
694 obligation to set up the dynamic locking callbacks.)
695 [Sander Temme <sander@temme.net>]
697 *) Use correct exit code if there is an error in dgst command.
698 [Steve Henson; problem pointed out by Roland Dirlewanger]
700 *) Tweak Configure so that you need to say "experimental-jpake" to enable
701 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
704 *) Add experimental JPAKE support, including demo authentication in
705 s_client and s_server.
708 *) Set the comparison function in v3_addr_canonize().
709 [Rob Austein <sra@hactrn.net>]
711 *) Add support for XMPP STARTTLS in s_client.
712 [Philip Paeps <philip@freebsd.org>]
714 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
715 to ensure that even with this option, only ciphersuites in the
716 server's preference list will be accepted. (Note that the option
717 applies only when resuming a session, so the earlier behavior was
718 just about the algorithm choice for symmetric cryptography.)
721 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
723 *) Fix NULL pointer dereference if a DTLS server received
724 ChangeCipherSpec as first record (CVE-2009-1386).
727 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
728 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
731 *) The fix in 0.9.8c that supposedly got rid of unsafe
732 double-checked locking was incomplete for RSA blinding,
733 addressing just one layer of what turns out to have been
734 doubly unsafe triple-checked locking.
736 So now fix this for real by retiring the MONT_HELPER macro
737 in crypto/rsa/rsa_eay.c.
739 [Bodo Moeller; problem pointed out by Marius Schilder]
741 *) Various precautionary measures:
743 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
745 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
746 (NB: This would require knowledge of the secret session ticket key
747 to exploit, in which case you'd be SOL either way.)
749 - Change bn_nist.c so that it will properly handle input BIGNUMs
750 outside the expected range.
752 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
755 [Neel Mehta, Bodo Moeller]
757 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
758 the load fails. Useful for distros.
759 [Ben Laurie and the FreeBSD team]
761 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
764 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
767 *) Expand ENGINE to support engine supplied SSL client certificate functions.
769 This work was sponsored by Logica.
772 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
773 keystores. Support for SSL/TLS client authentication too.
774 Not compiled unless enable-capieng specified to Configure.
776 This work was sponsored by Logica.
779 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
780 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
781 attribute creation routines such as certifcate requests and PKCS#12
785 Changes between 0.9.8g and 0.9.8h [28 May 2008]
787 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
788 handshake which could lead to a cilent crash as found using the
789 Codenomicon TLS test suite (CVE-2008-1672)
790 [Steve Henson, Mark Cox]
792 *) Fix double free in TLS server name extensions which could lead to
793 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
796 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
798 Clear the error queue to ensure that error entries left from
799 older function calls do not interfere with the correct operation.
800 [Lutz Jaenicke, Erik de Castro Lopo]
802 *) Remove root CA certificates of commercial CAs:
804 The OpenSSL project does not recommend any specific CA and does not
805 have any policy with respect to including or excluding any CA.
806 Therefore it does not make any sense to ship an arbitrary selection
807 of root CA certificates with the OpenSSL software.
810 *) RSA OAEP patches to fix two separate invalid memory reads.
811 The first one involves inputs when 'lzero' is greater than
812 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
813 before the beginning of from). The second one involves inputs where
814 the 'db' section contains nothing but zeroes (there is a one-byte
815 invalid read after the end of 'db').
816 [Ivan Nestlerode <inestlerode@us.ibm.com>]
818 *) Partial backport from 0.9.9-dev:
820 Introduce bn_mul_mont (dedicated Montgomery multiplication
821 procedure) as a candidate for BIGNUM assembler implementation.
822 While 0.9.9-dev uses assembler for various architectures, only
823 x86_64 is available by default here in the 0.9.8 branch, and
824 32-bit x86 is available through a compile-time setting.
826 To try the 32-bit x86 assembler implementation, use Configure
827 option "enable-montasm" (which exists only for this backport).
829 As "enable-montasm" for 32-bit x86 disclaims code stability
830 anyway, in this constellation we activate additional code
831 backported from 0.9.9-dev for further performance improvements,
832 namely BN_from_montgomery_word. (To enable this otherwise,
833 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
835 [Andy Polyakov (backport partially by Bodo Moeller)]
837 *) Add TLS session ticket callback. This allows an application to set
838 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
839 values. This is useful for key rollover for example where several key
840 sets may exist with different names.
843 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
844 This was broken until now in 0.9.8 releases, such that the only way
845 a registered ENGINE could be used (assuming it initialises
846 successfully on the host) was to explicitly set it as the default
847 for the relevant algorithms. This is in contradiction with 0.9.7
848 behaviour and the documentation. With this fix, when an ENGINE is
849 registered into a given algorithm's table of implementations, the
850 'uptodate' flag is reset so that auto-discovery will be used next
851 time a new context for that algorithm attempts to select an
853 [Ian Lister (tweaked by Geoff Thorpe)]
855 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
856 implemention in the following ways:
858 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
861 Lack of BER streaming support means one pass streaming processing is
862 only supported if data is detached: setting the streaming flag is
863 ignored for embedded content.
865 CMS support is disabled by default and must be explicitly enabled
866 with the enable-cms configuration option.
869 *) Update the GMP engine glue to do direct copies between BIGNUM and
870 mpz_t when openssl and GMP use the same limb size. Otherwise the
871 existing "conversion via a text string export" trick is still used.
872 [Paul Sheer <paulsheer@gmail.com>]
874 *) Zlib compression BIO. This is a filter BIO which compressed and
875 uncompresses any data passed through it.
878 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
879 RFC3394 compatible AES key wrapping.
882 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
883 sets string data without copying. X509_ALGOR_set0() and
884 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
885 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
886 from an X509_ATTRIBUTE structure optionally checking it occurs only
887 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
891 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
892 to get the expected BN_FLG_CONSTTIME behavior.
893 [Bodo Moeller (Google)]
897 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
898 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
899 - added some more tests to do_tests.pl
900 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
901 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
902 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
903 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
904 - various changes to netware.pl to enable gcc-cross builds on Win32
906 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
907 - various changes to fix missing prototype warnings
908 - fixed x86nasm.pl to create correct asm files for NASM COFF output
909 - added AES, WHIRLPOOL and CPUID assembler code to build files
910 - added missing AES assembler make rules to mk1mf.pl
911 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
912 [Guenter Knauf <eflash@gmx.net>]
914 *) Implement certificate status request TLS extension defined in RFC3546.
915 A client can set the appropriate parameters and receive the encoded
916 OCSP response via a callback. A server can query the supplied parameters
917 and set the encoded OCSP response in the callback. Add simplified examples
918 to s_client and s_server.
921 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
924 + Binary incompatibility of ssl_ctx_st structure
925 + DTLS interoperation with non-compliant servers
926 + Don't call get_session_cb() without proposed session
927 + Fix ia64 assembler code
928 [Andy Polyakov, Steve Henson]
930 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
932 *) DTLS Handshake overhaul. There were longstanding issues with
933 OpenSSL DTLS implementation, which were making it impossible for
934 RFC 4347 compliant client to communicate with OpenSSL server.
935 Unfortunately just fixing these incompatibilities would "cut off"
936 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
937 server keeps tolerating non RFC compliant syntax. The opposite is
938 not true, 0.9.8f client can not communicate with earlier server.
939 This update even addresses CVE-2007-4995.
942 *) Changes to avoid need for function casts in OpenSSL: some compilers
943 (gcc 4.2 and later) reject their use.
944 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
947 *) Add RFC4507 support to OpenSSL. This includes the corrections in
948 RFC4507bis. The encrypted ticket format is an encrypted encoded
949 SSL_SESSION structure, that way new session features are automatically
952 If a client application caches session in an SSL_SESSION structure
953 support is transparent because tickets are now stored in the encoded
956 The SSL_CTX structure automatically generates keys for ticket
957 protection in servers so again support should be possible
958 with no application modification.
960 If a client or server wishes to disable RFC4507 support then the option
961 SSL_OP_NO_TICKET can be set.
963 Add a TLS extension debugging callback to allow the contents of any client
964 or server extensions to be examined.
966 This work was sponsored by Google.
969 *) Add initial support for TLS extensions, specifically for the server_name
970 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
971 have new members for a host name. The SSL data structure has an
972 additional member SSL_CTX *initial_ctx so that new sessions can be
973 stored in that context to allow for session resumption, even after the
974 SSL has been switched to a new SSL_CTX in reaction to a client's
975 server_name extension.
977 New functions (subject to change):
980 SSL_get_servername_type()
983 New CTRL codes and macros (subject to change):
985 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
986 - SSL_CTX_set_tlsext_servername_callback()
987 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
988 - SSL_CTX_set_tlsext_servername_arg()
989 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
991 openssl s_client has a new '-servername ...' option.
993 openssl s_server has new options '-servername_host ...', '-cert2 ...',
994 '-key2 ...', '-servername_fatal' (subject to change). This allows
995 testing the HostName extension for a specific single host name ('-cert'
996 and '-key' remain fallbacks for handshakes without HostName
997 negotiation). If the unrecogninzed_name alert has to be sent, this by
998 default is a warning; it becomes fatal with the '-servername_fatal'
1001 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
1003 *) Add AES and SSE2 assembly language support to VC++ build.
1006 *) Mitigate attack on final subtraction in Montgomery reduction.
1009 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
1010 (which previously caused an internal error).
1013 *) Squeeze another 10% out of IGE mode when in != out.
1016 *) AES IGE mode speedup.
1017 [Dean Gaudet (Google)]
1019 *) Add the Korean symmetric 128-bit cipher SEED (see
1020 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
1021 add SEED ciphersuites from RFC 4162:
1023 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
1024 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
1025 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
1026 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
1028 To minimize changes between patchlevels in the OpenSSL 0.9.8
1029 series, SEED remains excluded from compilation unless OpenSSL
1030 is configured with 'enable-seed'.
1031 [KISA, Bodo Moeller]
1033 *) Mitigate branch prediction attacks, which can be practical if a
1034 single processor is shared, allowing a spy process to extract
1035 information. For detailed background information, see
1036 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
1037 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
1038 and Necessary Software Countermeasures"). The core of the change
1039 are new versions BN_div_no_branch() and
1040 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
1041 respectively, which are slower, but avoid the security-relevant
1042 conditional branches. These are automatically called by BN_div()
1043 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
1044 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
1045 remove a conditional branch.
1047 BN_FLG_CONSTTIME is the new name for the previous
1048 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
1049 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
1050 in the exponent causes BN_mod_exp_mont() to use the alternative
1051 implementation in BN_mod_exp_mont_consttime().) The old name
1052 remains as a deprecated alias.
1054 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
1055 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
1056 constant-time implementations for more than just exponentiation.
1057 Here too the old name is kept as a deprecated alias.
1059 BN_BLINDING_new() will now use BN_dup() for the modulus so that
1060 the BN_BLINDING structure gets an independent copy of the
1061 modulus. This means that the previous "BIGNUM *m" argument to
1062 BN_BLINDING_new() and to BN_BLINDING_create_param() now
1063 essentially becomes "const BIGNUM *m", although we can't actually
1064 change this in the header file before 0.9.9. It allows
1065 RSA_setup_blinding() to use BN_with_flags() on the modulus to
1066 enable BN_FLG_CONSTTIME.
1068 [Matthew D Wood (Intel Corp)]
1070 *) In the SSL/TLS server implementation, be strict about session ID
1071 context matching (which matters if an application uses a single
1072 external cache for different purposes). Previously,
1073 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
1074 set. This did ensure strict client verification, but meant that,
1075 with applications using a single external cache for quite
1076 different requirements, clients could circumvent ciphersuite
1077 restrictions for a given session ID context by starting a session
1078 in a different context.
1081 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1082 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1083 authentication-only ciphersuites.
1086 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
1087 not complete and could lead to a possible single byte overflow
1088 (CVE-2007-5135) [Ben Laurie]
1090 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
1092 *) Since AES128 and AES256 (and similarly Camellia128 and
1093 Camellia256) share a single mask bit in the logic of
1094 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1095 kludge to work properly if AES128 is available and AES256 isn't
1096 (or if Camellia128 is available and Camellia256 isn't).
1099 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
1100 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
1101 When a point or a seed is encoded in a BIT STRING, we need to
1102 prevent the removal of trailing zero bits to get the proper DER
1103 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
1104 of a NamedBitList, for which trailing 0 bits need to be removed.)
1107 *) Have SSL/TLS server implementation tolerate "mismatched" record
1108 protocol version while receiving ClientHello even if the
1109 ClientHello is fragmented. (The server can't insist on the
1110 particular protocol version it has chosen before the ServerHello
1111 message has informed the client about his choice.)
1114 *) Add RFC 3779 support.
1115 [Rob Austein for ARIN, Ben Laurie]
1117 *) Load error codes if they are not already present instead of using a
1118 static variable. This allows them to be cleanly unloaded and reloaded.
1119 Improve header file function name parsing.
1122 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
1123 or CAPABILITY handshake as required by RFCs.
1126 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
1128 *) Introduce limits to prevent malicious keys being able to
1129 cause a denial of service. (CVE-2006-2940)
1130 [Steve Henson, Bodo Moeller]
1132 *) Fix ASN.1 parsing of certain invalid structures that can result
1133 in a denial of service. (CVE-2006-2937) [Steve Henson]
1135 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1136 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1138 *) Fix SSL client code which could crash if connecting to a
1139 malicious SSLv2 server. (CVE-2006-4343)
1140 [Tavis Ormandy and Will Drewry, Google Security Team]
1142 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
1143 match only those. Before that, "AES256-SHA" would be interpreted
1144 as a pattern and match "AES128-SHA" too (since AES128-SHA got
1145 the same strength classification in 0.9.7h) as we currently only
1146 have a single AES bit in the ciphersuite description bitmap.
1147 That change, however, also applied to ciphersuite strings such as
1148 "RC4-MD5" that intentionally matched multiple ciphersuites --
1149 namely, SSL 2.0 ciphersuites in addition to the more common ones
1150 from SSL 3.0/TLS 1.0.
1152 So we change the selection algorithm again: Naming an explicit
1153 ciphersuite selects this one ciphersuite, and any other similar
1154 ciphersuite (same bitmap) from *other* protocol versions.
1155 Thus, "RC4-MD5" again will properly select both the SSL 2.0
1156 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
1158 Since SSL 2.0 does not have any ciphersuites for which the
1159 128/256 bit distinction would be relevant, this works for now.
1160 The proper fix will be to use different bits for AES128 and
1161 AES256, which would have avoided the problems from the beginning;
1162 however, bits are scarce, so we can only do this in a new release
1163 (not just a patchlevel) when we can change the SSL_CIPHER
1164 definition to split the single 'unsigned long mask' bitmap into
1165 multiple values to extend the available space.
1169 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
1171 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1172 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1174 *) Add AES IGE and biIGE modes.
1177 *) Change the Unix randomness entropy gathering to use poll() when
1178 possible instead of select(), since the latter has some
1179 undesirable limitations.
1180 [Darryl Miles via Richard Levitte and Bodo Moeller]
1182 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
1183 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
1184 cannot be implicitly activated as part of, e.g., the "AES" alias.
1185 However, please upgrade to OpenSSL 0.9.9[-dev] for
1186 non-experimental use of the ECC ciphersuites to get TLS extension
1187 support, which is required for curve and point format negotiation
1188 to avoid potential handshake problems.
1191 *) Disable rogue ciphersuites:
1193 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1194 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1195 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1197 The latter two were purportedly from
1198 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1201 Also deactivate the remaining ciphersuites from
1202 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1203 unofficial, and the ID has long expired.
1206 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1207 dual-core machines) and other potential thread-safety issues.
1210 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
1211 versions), which is now available for royalty-free use
1212 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
1213 Also, add Camellia TLS ciphersuites from RFC 4132.
1215 To minimize changes between patchlevels in the OpenSSL 0.9.8
1216 series, Camellia remains excluded from compilation unless OpenSSL
1217 is configured with 'enable-camellia'.
1220 *) Disable the padding bug check when compression is in use. The padding
1221 bug check assumes the first packet is of even length, this is not
1222 necessarily true if compresssion is enabled and can result in false
1223 positives causing handshake failure. The actual bug test is ancient
1224 code so it is hoped that implementations will either have fixed it by
1225 now or any which still have the bug do not support compression.
1228 Changes between 0.9.8a and 0.9.8b [04 May 2006]
1230 *) When applying a cipher rule check to see if string match is an explicit
1231 cipher suite and only match that one cipher suite if it is.
1234 *) Link in manifests for VC++ if needed.
1235 [Austin Ziegler <halostatue@gmail.com>]
1237 *) Update support for ECC-based TLS ciphersuites according to
1238 draft-ietf-tls-ecc-12.txt with proposed changes (but without
1239 TLS extensions, which are supported starting with the 0.9.9
1240 branch, not in the OpenSSL 0.9.8 branch).
1243 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
1244 opaque EVP_CIPHER_CTX handling.
1247 *) Fixes and enhancements to zlib compression code. We now only use
1248 "zlib1.dll" and use the default __cdecl calling convention on Win32
1249 to conform with the standards mentioned here:
1250 http://www.zlib.net/DLL_FAQ.txt
1251 Static zlib linking now works on Windows and the new --with-zlib-include
1252 --with-zlib-lib options to Configure can be used to supply the location
1253 of the headers and library. Gracefully handle case where zlib library
1257 *) Several fixes and enhancements to the OID generation code. The old code
1258 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1259 handle numbers larger than ULONG_MAX, truncated printing and had a
1260 non standard OBJ_obj2txt() behaviour.
1263 *) Add support for building of engines under engine/ as shared libraries
1264 under VC++ build system.
1267 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1268 Hopefully, we will not see any false combination of paths any more.
1271 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1273 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1274 (part of SSL_OP_ALL). This option used to disable the
1275 countermeasure against man-in-the-middle protocol-version
1276 rollback in the SSL 2.0 server implementation, which is a bad
1277 idea. (CVE-2005-2969)
1279 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1280 for Information Security, National Institute of Advanced Industrial
1281 Science and Technology [AIST], Japan)]
1283 *) Add two function to clear and return the verify parameter flags.
1286 *) Keep cipherlists sorted in the source instead of sorting them at
1287 runtime, thus removing the need for a lock.
1290 *) Avoid some small subgroup attacks in Diffie-Hellman.
1291 [Nick Mathewson and Ben Laurie]
1293 *) Add functions for well-known primes.
1296 *) Extended Windows CE support.
1297 [Satoshi Nakamura and Andy Polyakov]
1299 *) Initialize SSL_METHOD structures at compile time instead of during
1300 runtime, thus removing the need for a lock.
1303 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1304 attempting to decrypt each encrypted key in turn. Add support to
1308 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1310 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1313 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1316 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1317 key into the same file any more.
1320 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1323 *) Add -utf8 command line and config file option to 'ca'.
1324 [Stefan <stf@udoma.org]
1326 *) Removed the macro des_crypt(), as it seems to conflict with some
1327 libraries. Use DES_crypt().
1330 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1331 involves renaming the source and generated shared-libs for
1332 both. The engines will accept the corrected or legacy ids
1333 ('ncipher' and '4758_cca' respectively) when binding. NB,
1334 this only applies when building 'shared'.
1335 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1337 *) Add attribute functions to EVP_PKEY structure. Modify
1338 PKCS12_create() to recognize a CSP name attribute and
1339 use it. Make -CSP option work again in pkcs12 utility.
1342 *) Add new functionality to the bn blinding code:
1343 - automatic re-creation of the BN_BLINDING parameters after
1344 a fixed number of uses (currently 32)
1345 - add new function for parameter creation
1346 - introduce flags to control the update behaviour of the
1347 BN_BLINDING parameters
1348 - hide BN_BLINDING structure
1349 Add a second BN_BLINDING slot to the RSA structure to improve
1350 performance when a single RSA object is shared among several
1354 *) Add support for DTLS.
1355 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1357 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1358 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1361 *) Remove buggy and incompletet DH cert support from
1362 ssl/ssl_rsa.c and ssl/s3_both.c
1365 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1366 the apps/openssl applications.
1369 *) Compile clean with "-Wall -Wmissing-prototypes
1370 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1371 DEBUG_SAFESTACK must also be set.
1374 *) Change ./Configure so that certain algorithms can be disabled by default.
1375 The new counterpiece to "no-xxx" is "enable-xxx".
1377 The patented RC5 and MDC2 algorithms will now be disabled unless
1378 "enable-rc5" and "enable-mdc2", respectively, are specified.
1380 (IDEA remains enabled despite being patented. This is because IDEA
1381 is frequently required for interoperability, and there is no license
1382 fee for non-commercial use. As before, "no-idea" can be used to
1383 avoid this algorithm.)
1387 *) Add processing of proxy certificates (see RFC 3820). This work was
1388 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1389 EGEE (Enabling Grids for E-science in Europe).
1392 *) RC4 performance overhaul on modern architectures/implementations, such
1393 as Intel P4, IA-64 and AMD64.
1396 *) New utility extract-section.pl. This can be used specify an alternative
1397 section number in a pod file instead of having to treat each file as
1398 a separate case in Makefile. This can be done by adding two lines to the
1401 =for comment openssl_section:XXX
1403 The blank line is mandatory.
1407 *) New arguments -certform, -keyform and -pass for s_client and s_server
1408 to allow alternative format key and certificate files and passphrase
1412 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1413 update associated structures and add various utility functions.
1415 Add new policy related verify parameters, include policy checking in
1416 standard verify code. Enhance 'smime' application with extra parameters
1417 to support policy checking and print out.
1420 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1421 Nehemiah processors. These extensions support AES encryption in hardware
1422 as well as RNG (though RNG support is currently disabled).
1423 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1425 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1428 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1429 [Andy Polyakov and a number of other people]
1431 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1432 implementation contributed by IBM.
1433 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1435 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1436 exponent rather than 'unsigned long'. There is a corresponding change to
1437 the new 'rsa_keygen' element of the RSA_METHOD structure.
1438 [Jelte Jansen, Geoff Thorpe]
1440 *) Functionality for creating the initial serial number file is now
1441 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1443 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1444 number file to 1, which is bound to cause problems. To avoid
1445 the problems while respecting compatibility between different 0.9.7
1446 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1447 CA.pl for serial number initialization. With the new release 0.9.8,
1448 we can fix the problem directly in the 'ca' utility.)
1451 *) Reduced header interdepencies by declaring more opaque objects in
1452 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1453 give fewer recursive includes, which could break lazy source code - so
1454 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1455 developers should define this symbol when building and using openssl to
1456 ensure they track the recommended behaviour, interfaces, [etc], but
1457 backwards-compatible behaviour prevails when this isn't defined.
1460 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1463 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1464 This will generate a random key of the appropriate length based on the
1465 cipher context. The EVP_CIPHER can provide its own random key generation
1466 routine to support keys of a specific form. This is used in the des and
1467 3des routines to generate a key of the correct parity. Update S/MIME
1468 code to use new functions and hence generate correct parity DES keys.
1469 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1470 valid (weak or incorrect parity).
1473 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1474 as looking them up. This is useful when the verified structure may contain
1475 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1476 present unless the new PKCS7_NO_CRL flag is asserted.
1479 *) Extend ASN1 oid configuration module. It now additionally accepts the
1482 shortName = some long name, 1.2.3.4
1485 *) Reimplemented the BN_CTX implementation. There is now no more static
1486 limitation on the number of variables it can handle nor the depth of the
1487 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1488 information can now expand as required, and rather than having a single
1489 static array of bignums, BN_CTX now uses a linked-list of such arrays
1490 allowing it to expand on demand whilst maintaining the usefulness of
1491 BN_CTX's "bundling".
1494 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1495 to allow all RSA operations to function using a single BN_CTX.
1498 *) Preliminary support for certificate policy evaluation and checking. This
1499 is initially intended to pass the tests outlined in "Conformance Testing
1500 of Relying Party Client Certificate Path Processing Logic" v1.07.
1503 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1504 remained unused and not that useful. A variety of other little bignum
1505 tweaks and fixes have also been made continuing on from the audit (see
1509 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1510 associated ASN1, EVP and SSL functions and old ASN1 macros.
1513 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1514 and this should never fail. So the return value from the use of
1515 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1516 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1519 *) BN_CTX_get() should return zero-valued bignums, providing the same
1520 initialised value as BN_new().
1521 [Geoff Thorpe, suggested by Ulf Möller]
1523 *) Support for inhibitAnyPolicy certificate extension.
1526 *) An audit of the BIGNUM code is underway, for which debugging code is
1527 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1528 is considered valid when processing BIGNUMs, and causes execution to
1529 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1530 further steps are taken to deliberately pollute unused data in BIGNUM
1531 structures to try and expose faulty code further on. For now, openssl will
1532 (in its default mode of operation) continue to tolerate the inconsistent
1533 forms that it has tolerated in the past, but authors and packagers should
1534 consider trying openssl and their own applications when compiled with
1535 these debugging symbols defined. It will help highlight potential bugs in
1536 their own code, and will improve the test coverage for OpenSSL itself. At
1537 some point, these tighter rules will become openssl's default to improve
1538 maintainability, though the assert()s and other overheads will remain only
1539 in debugging configurations. See bn.h for more details.
1540 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1542 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1543 that can only be obtained through BN_CTX_new() (which implicitly
1544 initialises it). The presence of this function only made it possible
1545 to overwrite an existing structure (and cause memory leaks).
1548 *) Because of the callback-based approach for implementing LHASH as a
1549 template type, lh_insert() adds opaque objects to hash-tables and
1550 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1551 to clean up those corresponding objects before destroying the hash table
1552 (and losing the object pointers). So some over-zealous constifications in
1553 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1554 objects as "const" and the lh_doall[_arg] callback wrappers are not
1555 prototyped to have "const" restrictions on the object pointers they are
1556 given (and so aren't required to cast them away any more).
1559 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1560 (speed) prefers to use its own implementation. The two implementations
1561 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1562 its object type properly exposed (MS_TM) instead of casting to/from "char
1563 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1564 aren't necessarily the greatest nomenclatures - but this is what was used
1565 internally to the implementation so I've used that for now.
1568 *) Ensure that deprecated functions do not get compiled when
1569 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1570 the self-tests were still using deprecated key-generation functions so
1571 these have been updated also.
1574 *) Reorganise PKCS#7 code to separate the digest location functionality
1575 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1576 New function PKCS7_set_digest() to set the digest type for PKCS#7
1577 digestedData type. Add additional code to correctly generate the
1578 digestedData type and add support for this type in PKCS7 initialization
1582 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1583 structure of type "other".
1586 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1587 sure the loop does correctly stop and breaking ("division by zero")
1588 modulus operations are not performed. The (pre-generated) prime
1589 table crypto/bn/bn_prime.h was already correct, but it could not be
1590 re-generated on some platforms because of the "division by zero"
1591 situation in the script.
1592 [Ralf S. Engelschall]
1594 *) Update support for ECC-based TLS ciphersuites according to
1595 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1596 SHA-1 now is only used for "small" curves (where the
1597 representation of a field element takes up to 24 bytes); for
1598 larger curves, the field element resulting from ECDH is directly
1599 used as premaster secret.
1600 [Douglas Stebila (Sun Microsystems Laboratories)]
1602 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1603 curve secp160r1 to the tests.
1604 [Douglas Stebila (Sun Microsystems Laboratories)]
1606 *) Add the possibility to load symbols globally with DSO.
1607 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1609 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1610 control of the error stack.
1613 *) Add support for STORE in ENGINE.
1616 *) Add the STORE type. The intention is to provide a common interface
1617 to certificate and key stores, be they simple file-based stores, or
1618 HSM-type store, or LDAP stores, or...
1619 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1622 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1623 pass a list of arguments to any function as well as provide a way
1624 for a function to pass data back to the caller.
1627 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1628 works like BUF_strdup() but can be used to duplicate a portion of
1629 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1633 *) Add the function sk_find_ex() which works like sk_find(), but will
1634 return an index to an element even if an exact match couldn't be
1635 found. The index is guaranteed to point at the element where the
1636 searched-for key would be inserted to preserve sorting order.
1639 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1640 takes an extra flags argument for optional functionality. Currently,
1641 the following flags are defined:
1643 OBJ_BSEARCH_VALUE_ON_NOMATCH
1644 This one gets OBJ_bsearch_ex() to return a pointer to the first
1645 element where the comparing function returns a negative or zero
1648 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1649 This one gets OBJ_bsearch_ex() to return a pointer to the first
1650 element where the comparing function returns zero. This is useful
1651 if there are more than one element where the comparing function
1655 *) Make it possible to create self-signed certificates with 'openssl ca'
1656 in such a way that the self-signed certificate becomes part of the
1657 CA database and uses the same mechanisms for serial number generation
1658 as all other certificate signing. The new flag '-selfsign' enables
1659 this functionality. Adapt CA.sh and CA.pl.in.
1662 *) Add functionality to check the public key of a certificate request
1663 against a given private. This is useful to check that a certificate
1664 request can be signed by that key (self-signing).
1667 *) Make it possible to have multiple active certificates with the same
1668 subject in the CA index file. This is done only if the keyword
1669 'unique_subject' is set to 'no' in the main CA section (default
1670 if 'CA_default') of the configuration file. The value is saved
1671 with the database itself in a separate index attribute file,
1672 named like the index file with '.attr' appended to the name.
1675 *) Generate muti valued AVAs using '+' notation in config files for
1679 *) Support for nameConstraints certificate extension.
1682 *) Support for policyConstraints certificate extension.
1685 *) Support for policyMappings certificate extension.
1688 *) Make sure the default DSA_METHOD implementation only uses its
1689 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1690 and change its own handlers to be NULL so as to remove unnecessary
1691 indirection. This lets alternative implementations fallback to the
1692 default implementation more easily.
1695 *) Support for directoryName in GeneralName related extensions
1699 *) Make it possible to link applications using Makefile.shared.
1700 Make that possible even when linking against static libraries!
1703 *) Support for single pass processing for S/MIME signing. This now
1704 means that S/MIME signing can be done from a pipe, in addition
1705 cleartext signing (multipart/signed type) is effectively streaming
1706 and the signed data does not need to be all held in memory.
1708 This is done with a new flag PKCS7_STREAM. When this flag is set
1709 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1710 is done after the data is output (and digests calculated) in
1711 SMIME_write_PKCS7().
1714 *) Add full support for -rpath/-R, both in shared libraries and
1715 applications, at least on the platforms where it's known how
1719 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1720 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1721 will now compute a table of multiples of the generator that
1722 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1723 faster (notably in the case of a single point multiplication,
1724 scalar * generator).
1725 [Nils Larsch, Bodo Moeller]
1727 *) IPv6 support for certificate extensions. The various extensions
1728 which use the IP:a.b.c.d can now take IPv6 addresses using the
1729 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1733 *) Added an ENGINE that implements RSA by performing private key
1734 exponentiations with the GMP library. The conversions to and from
1735 GMP's mpz_t format aren't optimised nor are any montgomery forms
1736 cached, and on x86 it appears OpenSSL's own performance has caught up.
1737 However there are likely to be other architectures where GMP could
1738 provide a boost. This ENGINE is not built in by default, but it can be
1739 specified at Configure time and should be accompanied by the necessary
1740 linker additions, eg;
1741 ./config -DOPENSSL_USE_GMP -lgmp
1744 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1745 testing availability of engines with "-t" - the old behaviour is
1746 produced by increasing the feature's verbosity with "-tt".
1749 *) ECDSA routines: under certain error conditions uninitialized BN objects
1750 could be freed. Solution: make sure initialization is performed early
1751 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1755 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1756 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1757 software implementations. For DSA and DH, parameter generation can
1758 also be overriden by providing the appropriate method callbacks.
1761 *) Change the "progress" mechanism used in key-generation and
1762 primality testing to functions that take a new BN_GENCB pointer in
1763 place of callback/argument pairs. The new API functions have "_ex"
1764 postfixes and the older functions are reimplemented as wrappers for
1765 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1766 declarations of the old functions to help (graceful) attempts to
1767 migrate to the new functions. Also, the new key-generation API
1768 functions operate on a caller-supplied key-structure and return
1769 success/failure rather than returning a key or NULL - this is to
1770 help make "keygen" another member function of RSA_METHOD etc.
1772 Example for using the new callback interface:
1774 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1778 BN_GENCB_set(&my_cb, my_callback, my_arg);
1780 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1781 /* For the meaning of a, b in calls to my_callback(), see the
1782 * documentation of the function that calls the callback.
1783 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1784 * my_callback should return 1 if it wants BN_is_prime_ex()
1785 * to continue, or 0 to stop.
1790 *) Change the ZLIB compression method to be stateful, and make it
1791 available to TLS with the number defined in
1792 draft-ietf-tls-compression-04.txt.
1795 *) Add the ASN.1 structures and functions for CertificatePair, which
1796 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1798 CertificatePair ::= SEQUENCE {
1799 forward [0] Certificate OPTIONAL,
1800 reverse [1] Certificate OPTIONAL,
1801 -- at least one of the pair shall be present -- }
1803 Also implement the PEM functions to read and write certificate
1804 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1806 This needed to be defined, mostly for the sake of the LDAP
1807 attribute crossCertificatePair, but may prove useful elsewhere as
1811 *) Make it possible to inhibit symlinking of shared libraries in
1812 Makefile.shared, for Cygwin's sake.
1815 *) Extend the BIGNUM API by creating a function
1816 void BN_set_negative(BIGNUM *a, int neg);
1817 and a macro that behave like
1818 int BN_is_negative(const BIGNUM *a);
1820 to avoid the need to access 'a->neg' directly in applications.
1823 *) Implement fast modular reduction for pseudo-Mersenne primes
1824 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1825 EC_GROUP_new_curve_GFp() will now automatically use this
1827 [Nils Larsch <nla@trustcenter.de>]
1829 *) Add new lock type (CRYPTO_LOCK_BN).
1832 *) Change the ENGINE framework to automatically load engines
1833 dynamically from specific directories unless they could be
1834 found to already be built in or loaded. Move all the
1835 current engines except for the cryptodev one to a new
1837 The engines in engines/ are built as shared libraries if
1838 the "shared" options was given to ./Configure or ./config.
1839 Otherwise, they are inserted in libcrypto.a.
1840 /usr/local/ssl/engines is the default directory for dynamic
1841 engines, but that can be overriden at configure time through
1842 the usual use of --prefix and/or --openssldir, and at run
1843 time with the environment variable OPENSSL_ENGINES.
1844 [Geoff Thorpe and Richard Levitte]
1846 *) Add Makefile.shared, a helper makefile to build shared
1847 libraries. Addapt Makefile.org.
1850 *) Add version info to Win32 DLLs.
1851 [Peter 'Luna' Runestig" <peter@runestig.com>]
1853 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1854 can be added using this API to created arbitrary PKCS#12
1855 files while avoiding the low level API.
1857 New options to PKCS12_create(), key or cert can be NULL and
1858 will then be omitted from the output file. The encryption
1859 algorithm NIDs can be set to -1 for no encryption, the mac
1860 iteration count can be set to 0 to omit the mac.
1862 Enhance pkcs12 utility by making the -nokeys and -nocerts
1863 options work when creating a PKCS#12 file. New option -nomac
1864 to omit the mac, NONE can be set for an encryption algorithm.
1865 New code is modified to use the enhanced PKCS12_create()
1866 instead of the low level API.
1869 *) Extend ASN1 encoder to support indefinite length constructed
1870 encoding. This can output sequences tags and octet strings in
1871 this form. Modify pk7_asn1.c to support indefinite length
1872 encoding. This is experimental and needs additional code to
1873 be useful, such as an ASN1 bio and some enhanced streaming
1876 Extend template encode functionality so that tagging is passed
1877 down to the template encoder.
1880 *) Let 'openssl req' fail if an argument to '-newkey' is not
1881 recognized instead of using RSA as a default.
1884 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1885 As these are not official, they are not included in "ALL";
1886 the "ECCdraft" ciphersuite group alias can be used to select them.
1887 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1889 *) Add ECDH engine support.
1890 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1892 *) Add ECDH in new directory crypto/ecdh/.
1893 [Douglas Stebila (Sun Microsystems Laboratories)]
1895 *) Let BN_rand_range() abort with an error after 100 iterations
1896 without success (which indicates a broken PRNG).
1899 *) Change BN_mod_sqrt() so that it verifies that the input value
1900 is really the square of the return value. (Previously,
1901 BN_mod_sqrt would show GIGO behaviour.)
1904 *) Add named elliptic curves over binary fields from X9.62, SECG,
1905 and WAP/WTLS; add OIDs that were still missing.
1907 [Sheueling Chang Shantz and Douglas Stebila
1908 (Sun Microsystems Laboratories)]
1910 *) Extend the EC library for elliptic curves over binary fields
1911 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1914 EC_GF2m_simple_method
1918 EC_GROUP_new_curve_GF2m
1919 EC_GROUP_set_curve_GF2m
1920 EC_GROUP_get_curve_GF2m
1921 EC_POINT_set_affine_coordinates_GF2m
1922 EC_POINT_get_affine_coordinates_GF2m
1923 EC_POINT_set_compressed_coordinates_GF2m
1925 Point compression for binary fields is disabled by default for
1926 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1929 As binary polynomials are represented as BIGNUMs, various members
1930 of the EC_GROUP and EC_POINT data structures can be shared
1931 between the implementations for prime fields and binary fields;
1932 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1933 are essentially identical to their ..._GFp counterparts.
1934 (For simplicity, the '..._GFp' prefix has been dropped from
1935 various internal method names.)
1937 An internal 'field_div' method (similar to 'field_mul' and
1938 'field_sqr') has been added; this is used only for binary fields.
1940 [Sheueling Chang Shantz and Douglas Stebila
1941 (Sun Microsystems Laboratories)]
1943 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1944 through methods ('mul', 'precompute_mult').
1946 The generic implementations (now internally called 'ec_wNAF_mul'
1947 and 'ec_wNAF_precomputed_mult') remain the default if these
1948 methods are undefined.
1950 [Sheueling Chang Shantz and Douglas Stebila
1951 (Sun Microsystems Laboratories)]
1953 *) New function EC_GROUP_get_degree, which is defined through
1954 EC_METHOD. For curves over prime fields, this returns the bit
1955 length of the modulus.
1957 [Sheueling Chang Shantz and Douglas Stebila
1958 (Sun Microsystems Laboratories)]
1960 *) New functions EC_GROUP_dup, EC_POINT_dup.
1961 (These simply call ..._new and ..._copy).
1963 [Sheueling Chang Shantz and Douglas Stebila
1964 (Sun Microsystems Laboratories)]
1966 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1967 Polynomials are represented as BIGNUMs (where the sign bit is not
1968 used) in the following functions [macros]:
1971 BN_GF2m_sub [= BN_GF2m_add]
1972 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1973 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1974 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1976 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1977 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1978 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1979 BN_GF2m_cmp [= BN_ucmp]
1981 (Note that only the 'mod' functions are actually for fields GF(2^m).
1982 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1984 For some functions, an the irreducible polynomial defining a
1985 field can be given as an 'unsigned int[]' with strictly
1986 decreasing elements giving the indices of those bits that are set;
1987 i.e., p[] represents the polynomial
1988 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1990 p[0] > p[1] > ... > p[k] = 0.
1991 This applies to the following functions:
1996 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1997 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1999 BN_GF2m_mod_sqrt_arr
2000 BN_GF2m_mod_solve_quad_arr
2004 Conversion can be performed by the following functions:
2009 bntest.c has additional tests for binary polynomial arithmetic.
2011 Two implementations for BN_GF2m_mod_div() are available.
2012 The default algorithm simply uses BN_GF2m_mod_inv() and
2013 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
2014 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
2015 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
2017 [Sheueling Chang Shantz and Douglas Stebila
2018 (Sun Microsystems Laboratories)]
2020 *) Add new error code 'ERR_R_DISABLED' that can be used when some
2021 functionality is disabled at compile-time.
2022 [Douglas Stebila <douglas.stebila@sun.com>]
2024 *) Change default behaviour of 'openssl asn1parse' so that more
2025 information is visible when viewing, e.g., a certificate:
2027 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
2028 mode the content of non-printable OCTET STRINGs is output in a
2029 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
2030 avoid the appearance of a printable string.
2031 [Nils Larsch <nla@trustcenter.de>]
2033 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
2035 EC_GROUP_set_asn1_flag()
2036 EC_GROUP_get_asn1_flag()
2037 EC_GROUP_set_point_conversion_form()
2038 EC_GROUP_get_point_conversion_form()
2039 These control ASN1 encoding details:
2040 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
2041 has been set to OPENSSL_EC_NAMED_CURVE.
2042 - Points are encoded in uncompressed form by default; options for
2043 asn1_for are as for point2oct, namely
2044 POINT_CONVERSION_COMPRESSED
2045 POINT_CONVERSION_UNCOMPRESSED
2046 POINT_CONVERSION_HYBRID
2048 Also add 'seed' and 'seed_len' members to EC_GROUP with access
2051 EC_GROUP_get0_seed()
2052 EC_GROUP_get_seed_len()
2053 This is used only for ASN1 purposes (so far).
2054 [Nils Larsch <nla@trustcenter.de>]
2056 *) Add 'field_type' member to EC_METHOD, which holds the NID
2057 of the appropriate field type OID. The new function
2058 EC_METHOD_get_field_type() returns this value.
2059 [Nils Larsch <nla@trustcenter.de>]
2064 EC_POINT_point2hex()
2065 EC_POINT_hex2point()
2066 providing useful interfaces to EC_POINT_point2oct() and
2067 EC_POINT_oct2point().
2068 [Nils Larsch <nla@trustcenter.de>]
2070 *) Change internals of the EC library so that the functions
2071 EC_GROUP_set_generator()
2072 EC_GROUP_get_generator()
2073 EC_GROUP_get_order()
2074 EC_GROUP_get_cofactor()
2075 are implemented directly in crypto/ec/ec_lib.c and not dispatched
2076 to methods, which would lead to unnecessary code duplication when
2077 adding different types of curves.
2078 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
2080 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
2081 arithmetic, and such that modified wNAFs are generated
2082 (which avoid length expansion in many cases).
2085 *) Add a function EC_GROUP_check_discriminant() (defined via
2086 EC_METHOD) that verifies that the curve discriminant is non-zero.
2088 Add a function EC_GROUP_check() that makes some sanity tests
2089 on a EC_GROUP, its generator and order. This includes
2090 EC_GROUP_check_discriminant().
2091 [Nils Larsch <nla@trustcenter.de>]
2093 *) Add ECDSA in new directory crypto/ecdsa/.
2095 Add applications 'openssl ecparam' and 'openssl ecdsa'
2096 (these are based on 'openssl dsaparam' and 'openssl dsa').
2098 ECDSA support is also included in various other files across the
2099 library. Most notably,
2100 - 'openssl req' now has a '-newkey ecdsa:file' option;
2101 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
2102 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
2103 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
2104 them suitable for ECDSA where domain parameters must be
2105 extracted before the specific public key;
2106 - ECDSA engine support has been added.
2107 [Nils Larsch <nla@trustcenter.de>]
2109 *) Include some named elliptic curves, and add OIDs from X9.62,
2110 SECG, and WAP/WTLS. Each curve can be obtained from the new
2112 EC_GROUP_new_by_curve_name(),
2113 and the list of available named curves can be obtained with
2114 EC_get_builtin_curves().
2115 Also add a 'curve_name' member to EC_GROUP objects, which can be
2117 EC_GROUP_set_curve_name()
2118 EC_GROUP_get_curve_name()
2119 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
2121 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
2122 was actually never needed) and in BN_mul(). The removal in BN_mul()
2123 required a small change in bn_mul_part_recursive() and the addition
2124 of the functions bn_cmp_part_words(), bn_sub_part_words() and
2125 bn_add_part_words(), which do the same thing as bn_cmp_words(),
2126 bn_sub_words() and bn_add_words() except they take arrays with
2130 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
2132 *) Cleanse PEM buffers before freeing them since they may contain
2134 [Benjamin Bennett <ben@psc.edu>]
2136 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
2137 a ciphersuite string such as "DEFAULT:RSA" cannot enable
2138 authentication-only ciphersuites.
2141 *) Since AES128 and AES256 share a single mask bit in the logic of
2142 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
2143 kludge to work properly if AES128 is available and AES256 isn't.
2146 *) Expand security boundary to match 1.1.1 module.
2149 *) Remove redundant features: hash file source, editing of test vectors
2150 modify fipsld to use external fips_premain.c signature.
2153 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
2154 run algorithm test programs.
2157 *) Make algorithm test programs more tolerant of whitespace.
2160 *) Have SSL/TLS server implementation tolerate "mismatched" record
2161 protocol version while receiving ClientHello even if the
2162 ClientHello is fragmented. (The server can't insist on the
2163 particular protocol version it has chosen before the ServerHello
2164 message has informed the client about his choice.)
2167 *) Load error codes if they are not already present instead of using a
2168 static variable. This allows them to be cleanly unloaded and reloaded.
2171 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
2173 *) Introduce limits to prevent malicious keys being able to
2174 cause a denial of service. (CVE-2006-2940)
2175 [Steve Henson, Bodo Moeller]
2177 *) Fix ASN.1 parsing of certain invalid structures that can result
2178 in a denial of service. (CVE-2006-2937) [Steve Henson]
2180 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
2181 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
2183 *) Fix SSL client code which could crash if connecting to a
2184 malicious SSLv2 server. (CVE-2006-4343)
2185 [Tavis Ormandy and Will Drewry, Google Security Team]
2187 *) Change ciphersuite string processing so that an explicit
2188 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
2189 will no longer include "AES128-SHA"), and any other similar
2190 ciphersuite (same bitmap) from *other* protocol versions (so that
2191 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
2192 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
2193 changes from 0.9.8b and 0.9.8d.
2196 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
2198 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
2199 (CVE-2006-4339) [Ben Laurie and Google Security Team]
2201 *) Change the Unix randomness entropy gathering to use poll() when
2202 possible instead of select(), since the latter has some
2203 undesirable limitations.
2204 [Darryl Miles via Richard Levitte and Bodo Moeller]
2206 *) Disable rogue ciphersuites:
2208 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
2209 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
2210 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
2212 The latter two were purportedly from
2213 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
2216 Also deactive the remaining ciphersuites from
2217 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
2218 unofficial, and the ID has long expired.
2221 *) Fix RSA blinding Heisenbug (problems sometimes occured on
2222 dual-core machines) and other potential thread-safety issues.
2225 Changes between 0.9.7i and 0.9.7j [04 May 2006]
2227 *) Adapt fipsld and the build system to link against the validated FIPS
2228 module in FIPS mode.
2231 *) Fixes for VC++ 2005 build under Windows.
2234 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
2235 from a Windows bash shell such as MSYS. It is autodetected from the
2236 "config" script when run from a VC++ environment. Modify standard VC++
2237 build to use fipscanister.o from the GNU make build.
2240 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
2242 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
2243 The value now differs depending on if you build for FIPS or not.
2244 BEWARE! A program linked with a shared FIPSed libcrypto can't be
2245 safely run with a non-FIPSed libcrypto, as it may crash because of
2246 the difference induced by this change.
2249 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
2251 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2252 (part of SSL_OP_ALL). This option used to disable the
2253 countermeasure against man-in-the-middle protocol-version
2254 rollback in the SSL 2.0 server implementation, which is a bad
2255 idea. (CVE-2005-2969)
2257 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2258 for Information Security, National Institute of Advanced Industrial
2259 Science and Technology [AIST], Japan)]
2261 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2262 mainly for FIPS compliance and not fully integrated at this stage.
2265 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2266 the exponentiation using a fixed-length exponent. (Otherwise,
2267 the information leaked through timing could expose the secret key
2268 after many signatures; cf. Bleichenbacher's attack on DSA with
2272 *) Make a new fixed-window mod_exp implementation the default for
2273 RSA, DSA, and DH private-key operations so that the sequence of
2274 squares and multiplies and the memory access pattern are
2275 independent of the particular secret key. This will mitigate
2276 cache-timing and potential related attacks.
2278 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2279 and this is automatically used by BN_mod_exp_mont() if the new flag
2280 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2281 will use this BN flag for private exponents unless the flag
2282 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2283 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2285 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2287 *) Change the client implementation for SSLv23_method() and
2288 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2289 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2290 (Previously, the SSL 2.0 backwards compatible Client Hello
2291 message format would be used even with SSL_OP_NO_SSLv2.)
2294 *) Add support for smime-type MIME parameter in S/MIME messages which some
2298 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2299 a threadsafe manner. Modify rsa code to use new function and add calls
2300 to dsa and dh code (which had race conditions before).
2303 *) Include the fixed error library code in the C error file definitions
2304 instead of fixing them up at runtime. This keeps the error code
2305 structures constant.
2308 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2310 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2313 *) Fixes for newer kerberos headers. NB: the casts are needed because
2314 the 'length' field is signed on one version and unsigned on another
2315 with no (?) obvious way to tell the difference, without these VC++
2316 complains. Also the "definition" of FAR (blank) is no longer included
2317 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2318 some needed definitions.
2321 *) Undo Cygwin change.
2324 *) Added support for proxy certificates according to RFC 3820.
2325 Because they may be a security thread to unaware applications,
2326 they must be explicitely allowed in run-time. See
2327 docs/HOWTO/proxy_certificates.txt for further information.
2330 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2332 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2333 server and client random values. Previously
2334 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2335 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2337 This change has negligible security impact because:
2339 1. Server and client random values still have 24 bytes of pseudo random
2342 2. Server and client random values are sent in the clear in the initial
2345 3. The master secret is derived using the premaster secret (48 bytes in
2346 size for static RSA ciphersuites) as well as client server and random
2349 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2352 [Stephen Henson, reported by UK NISCC]
2354 *) Use Windows randomness collection on Cygwin.
2357 *) Fix hang in EGD/PRNGD query when communication socket is closed
2358 prematurely by EGD/PRNGD.
2359 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2361 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2364 *) Back-port of selected performance improvements from development
2365 branch, as well as improved support for PowerPC platforms.
2368 *) Add lots of checks for memory allocation failure, error codes to indicate
2369 failure and freeing up memory if a failure occurs.
2370 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2372 *) Add new -passin argument to dgst.
2375 *) Perform some character comparisons of different types in X509_NAME_cmp:
2376 this is needed for some certificates that reencode DNs into UTF8Strings
2377 (in violation of RFC3280) and can't or wont issue name rollover
2381 *) Make an explicit check during certificate validation to see that
2382 the CA setting in each certificate on the chain is correct. As a
2383 side effect always do the following basic checks on extensions,
2384 not just when there's an associated purpose to the check:
2386 - if there is an unhandled critical extension (unless the user
2387 has chosen to ignore this fault)
2388 - if the path length has been exceeded (if one is set at all)
2389 - that certain extensions fit the associated purpose (if one has
2393 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2395 *) Avoid a race condition when CRLs are checked in a multi threaded
2396 environment. This would happen due to the reordering of the revoked
2397 entries during signature checking and serial number lookup. Now the
2398 encoding is cached and the serial number sort performed under a lock.
2399 Add new STACK function sk_is_sorted().
2402 *) Add Delta CRL to the extension code.
2405 *) Various fixes to s3_pkt.c so alerts are sent properly.
2406 [David Holmes <d.holmes@f5.com>]
2408 *) Reduce the chances of duplicate issuer name and serial numbers (in
2409 violation of RFC3280) using the OpenSSL certificate creation utilities.
2410 This is done by creating a random 64 bit value for the initial serial
2411 number when a serial number file is created or when a self signed
2412 certificate is created using 'openssl req -x509'. The initial serial
2413 number file is created using 'openssl x509 -next_serial' in CA.pl
2414 rather than being initialized to 1.
2417 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2419 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2420 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2421 [Joe Orton, Steve Henson]
2423 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2425 [Joe Orton, Steve Henson]
2427 *) Make it possible to have multiple active certificates with the same
2428 subject in the CA index file. This is done only if the keyword
2429 'unique_subject' is set to 'no' in the main CA section (default
2430 if 'CA_default') of the configuration file. The value is saved
2431 with the database itself in a separate index attribute file,
2432 named like the index file with '.attr' appended to the name.
2435 *) X509 verify fixes. Disable broken certificate workarounds when
2436 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2437 keyUsage extension present. Don't accept CRLs with unhandled critical
2438 extensions: since verify currently doesn't process CRL extensions this
2439 rejects a CRL with *any* critical extensions. Add new verify error codes
2443 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2444 A clarification of RFC2560 will require the use of OCTET STRINGs and
2445 some implementations cannot handle the current raw format. Since OpenSSL
2446 copies and compares OCSP nonces as opaque blobs without any attempt at
2447 parsing them this should not create any compatibility issues.
2450 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2451 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2452 this HMAC (and other) operations are several times slower than OpenSSL
2456 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2457 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2459 *) Use the correct content when signing type "other".
2462 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2464 *) Fix various bugs revealed by running the NISCC test suite:
2466 Stop out of bounds reads in the ASN1 code when presented with
2467 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2469 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2471 If verify callback ignores invalid public key errors don't try to check
2472 certificate signature with the NULL public key.
2476 *) New -ignore_err option in ocsp application to stop the server
2477 exiting on the first error in a request.
2480 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2481 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2485 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2486 extra data after the compression methods not only for TLS 1.0
2487 but also for SSL 3.0 (as required by the specification).
2488 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2490 *) Change X509_certificate_type() to mark the key as exported/exportable
2491 when it's 512 *bits* long, not 512 bytes.
2494 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2495 blocks during encryption.
2498 *) Various fixes to base64 BIO and non blocking I/O. On write
2499 flushes were not handled properly if the BIO retried. On read
2500 data was not being buffered properly and had various logic bugs.
2501 This also affects blocking I/O when the data being decoded is a
2505 *) Various S/MIME bugfixes and compatibility changes:
2506 output correct application/pkcs7 MIME type if
2507 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2508 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2509 of files as .eml work). Correctly handle very long lines in MIME
2513 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2515 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2516 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2517 a protocol version number mismatch like a decryption error
2518 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2521 *) Turn on RSA blinding by default in the default implementation
2522 to avoid a timing attack. Applications that don't want it can call
2523 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2524 They would be ill-advised to do so in most cases.
2525 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2527 *) Change RSA blinding code so that it works when the PRNG is not
2528 seeded (in this case, the secret RSA exponent is abused as
2529 an unpredictable seed -- if it is not unpredictable, there
2530 is no point in blinding anyway). Make RSA blinding thread-safe
2531 by remembering the creator's thread ID in rsa->blinding and
2532 having all other threads use local one-time blinding factors
2533 (this requires more computation than sharing rsa->blinding, but
2534 avoids excessive locking; and if an RSA object is not shared
2535 between threads, blinding will still be very fast).
2538 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2539 ENGINE as defaults for all supported algorithms irrespective of
2540 the 'flags' parameter. 'flags' is now honoured, so applications
2541 should make sure they are passing it correctly.
2544 *) Target "mingw" now allows native Windows code to be generated in
2545 the Cygwin environment as well as with the MinGW compiler.
2548 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2550 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2551 via timing by performing a MAC computation even if incorrrect
2552 block cipher padding has been found. This is a countermeasure
2553 against active attacks where the attacker has to distinguish
2554 between bad padding and a MAC verification error. (CVE-2003-0078)
2556 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2557 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2558 Martin Vuagnoux (EPFL, Ilion)]
2560 *) Make the no-err option work as intended. The intention with no-err
2561 is not to have the whole error stack handling routines removed from
2562 libcrypto, it's only intended to remove all the function name and
2563 reason texts, thereby removing some of the footprint that may not
2564 be interesting if those errors aren't displayed anyway.
2566 NOTE: it's still possible for any application or module to have it's
2567 own set of error texts inserted. The routines are there, just not
2568 used by default when no-err is given.
2571 *) Add support for FreeBSD on IA64.
2572 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2574 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2575 Kerberos function mit_des_cbc_cksum(). Before this change,
2576 the value returned by DES_cbc_cksum() was like the one from
2577 mit_des_cbc_cksum(), except the bytes were swapped.
2578 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2580 *) Allow an application to disable the automatic SSL chain building.
2581 Before this a rather primitive chain build was always performed in
2582 ssl3_output_cert_chain(): an application had no way to send the
2583 correct chain if the automatic operation produced an incorrect result.
2585 Now the chain builder is disabled if either:
2587 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2589 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2591 The reasoning behind this is that an application would not want the
2592 auto chain building to take place if extra chain certificates are
2593 present and it might also want a means of sending no additional
2594 certificates (for example the chain has two certificates and the
2598 *) Add the possibility to build without the ENGINE framework.
2599 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2601 *) Under Win32 gmtime() can return NULL: check return value in
2602 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2605 *) DSA routines: under certain error conditions uninitialized BN objects
2606 could be freed. Solution: make sure initialization is performed early
2607 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2608 Nils Larsch <nla@trustcenter.de> via PR#459)
2611 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2612 checked on reconnect on the client side, therefore session resumption
2613 could still fail with a "ssl session id is different" error. This
2614 behaviour is masked when SSL_OP_ALL is used due to
2615 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2616 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2617 followup to PR #377.
2620 *) IA-32 assembler support enhancements: unified ELF targets, support
2621 for SCO/Caldera platforms, fix for Cygwin shared build.
2624 *) Add support for FreeBSD on sparc64. As a consequence, support for
2625 FreeBSD on non-x86 processors is separate from x86 processors on
2626 the config script, much like the NetBSD support.
2627 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2629 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2631 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2634 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2635 code (06) was taken as the first octet of the session ID and the last
2636 octet was ignored consequently. As a result SSLv2 client side session
2637 caching could not have worked due to the session ID mismatch between
2639 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2643 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2644 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2648 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2649 seems that in spite of existing for more than a year, many application
2650 author have done nothing to provide the necessary callbacks, which
2651 means that this particular engine will not work properly anywhere.
2652 This is a very unfortunate situation which forces us, in the name
2653 of usability, to give the hw_ncipher.c a static lock, which is part
2655 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2656 appear in 0.9.8 or later. We EXPECT application authors to have
2657 dealt properly with this when 0.9.8 is released (unless we actually
2658 make such changes in the libcrypto locking code that changes will
2659 have to be made anyway).
2662 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2663 octets have been read, EOF or an error occurs. Without this change
2664 some truncated ASN1 structures will not produce an error.
2667 *) Disable Heimdal support, since it hasn't been fully implemented.
2668 Still give the possibility to force the use of Heimdal, but with
2669 warnings and a request that patches get sent to openssl-dev.
2672 *) Add the VC-CE target, introduce the WINCE sysname, and add
2673 INSTALL.WCE and appropriate conditionals to make it build.
2674 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2676 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2677 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2678 edit numbers of the version.
2679 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2681 *) Introduce safe string copy and catenation functions
2682 (BUF_strlcpy() and BUF_strlcat()).
2683 [Ben Laurie (CHATS) and Richard Levitte]
2685 *) Avoid using fixed-size buffers for one-line DNs.
2686 [Ben Laurie (CHATS)]
2688 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2689 resizing buffers containing secrets, and use where appropriate.
2690 [Ben Laurie (CHATS)]
2692 *) Avoid using fixed size buffers for configuration file location.
2693 [Ben Laurie (CHATS)]
2695 *) Avoid filename truncation for various CA files.
2696 [Ben Laurie (CHATS)]
2698 *) Use sizeof in preference to magic numbers.
2699 [Ben Laurie (CHATS)]
2701 *) Avoid filename truncation in cert requests.
2702 [Ben Laurie (CHATS)]
2704 *) Add assertions to check for (supposedly impossible) buffer
2706 [Ben Laurie (CHATS)]
2708 *) Don't cache truncated DNS entries in the local cache (this could
2709 potentially lead to a spoofing attack).
2710 [Ben Laurie (CHATS)]
2712 *) Fix various buffers to be large enough for hex/decimal
2713 representations in a platform independent manner.
2714 [Ben Laurie (CHATS)]
2716 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2717 resizing buffers containing secrets, and use where appropriate.
2718 [Ben Laurie (CHATS)]
2720 *) Add BIO_indent() to avoid much slightly worrying code to do
2722 [Ben Laurie (CHATS)]
2724 *) Convert sprintf()/BIO_puts() to BIO_printf().
2725 [Ben Laurie (CHATS)]
2727 *) buffer_gets() could terminate with the buffer only half
2729 [Ben Laurie (CHATS)]
2731 *) Add assertions to prevent user-supplied crypto functions from
2732 overflowing internal buffers by having large block sizes, etc.
2733 [Ben Laurie (CHATS)]
2735 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2737 [Ben Laurie (CHATS)]
2739 *) Eliminate unused copy of key in RC4.
2740 [Ben Laurie (CHATS)]
2742 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2743 [Ben Laurie (CHATS)]
2745 *) Fix off-by-one error in EGD path.
2746 [Ben Laurie (CHATS)]
2748 *) If RANDFILE path is too long, ignore instead of truncating.
2749 [Ben Laurie (CHATS)]
2751 *) Eliminate unused and incorrectly sized X.509 structure
2753 [Ben Laurie (CHATS)]
2755 *) Eliminate unused and dangerous function knumber().
2756 [Ben Laurie (CHATS)]
2758 *) Eliminate unused and dangerous structure, KSSL_ERR.
2759 [Ben Laurie (CHATS)]
2761 *) Protect against overlong session ID context length in an encoded
2762 session object. Since these are local, this does not appear to be
2764 [Ben Laurie (CHATS)]
2766 *) Change from security patch (see 0.9.6e below) that did not affect
2767 the 0.9.6 release series:
2769 Remote buffer overflow in SSL3 protocol - an attacker could
2770 supply an oversized master key in Kerberos-enabled versions.
2772 [Ben Laurie (CHATS)]
2774 *) Change the SSL kerb5 codes to match RFC 2712.
2777 *) Make -nameopt work fully for req and add -reqopt switch.
2778 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2780 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2781 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2783 *) Make sure tests can be performed even if the corresponding algorithms
2784 have been removed entirely. This was also the last step to make
2785 OpenSSL compilable with DJGPP under all reasonable conditions.
2786 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2788 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2789 to allow version independent disabling of normally unselected ciphers,
2790 which may be activated as a side-effect of selecting a single cipher.
2792 (E.g., cipher list string "RSA" enables ciphersuites that are left
2793 out of "ALL" because they do not provide symmetric encryption.
2794 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2795 [Lutz Jaenicke, Bodo Moeller]
2797 *) Add appropriate support for separate platform-dependent build
2798 directories. The recommended way to make a platform-dependent
2799 build directory is the following (tested on Linux), maybe with
2802 # Place yourself outside of the OpenSSL source tree. In
2803 # this example, the environment variable OPENSSL_SOURCE
2804 # is assumed to contain the absolute OpenSSL source directory.
2805 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2806 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2807 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2808 mkdir -p `dirname $F`
2809 ln -s $OPENSSL_SOURCE/$F $F
2812 To be absolutely sure not to disturb the source tree, a "make clean"
2813 is a good thing. If it isn't successfull, don't worry about it,
2814 it probably means the source directory is very clean.
2817 *) Make sure any ENGINE control commands make local copies of string
2818 pointers passed to them whenever necessary. Otherwise it is possible
2819 the caller may have overwritten (or deallocated) the original string
2820 data when a later ENGINE operation tries to use the stored values.
2821 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2823 *) Improve diagnostics in file reading and command-line digests.
2824 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2826 *) Add AES modes CFB and OFB to the object database. Correct an
2827 error in AES-CFB decryption.
2830 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2831 allows existing EVP_CIPHER_CTX structures to be reused after
2832 calling EVP_*Final(). This behaviour is used by encryption
2833 BIOs and some applications. This has the side effect that
2834 applications must explicitly clean up cipher contexts with
2835 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2838 *) Check the values of dna and dnb in bn_mul_recursive before calling
2839 bn_mul_comba (a non zero value means the a or b arrays do not contain
2840 n2 elements) and fallback to bn_mul_normal if either is not zero.
2843 *) Fix escaping of non-ASCII characters when using the -subj option
2844 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2847 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2848 form for "surname", serialNumber has no short form.
2849 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2850 therefore remove "mail" short name for "internet 7".
2851 The OID for unique identifiers in X509 certificates is
2852 x500UniqueIdentifier, not uniqueIdentifier.
2853 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2856 *) Add an "init" command to the ENGINE config module and auto initialize
2857 ENGINEs. Without any "init" command the ENGINE will be initialized
2858 after all ctrl commands have been executed on it. If init=1 the
2859 ENGINE is initailized at that point (ctrls before that point are run
2860 on the uninitialized ENGINE and after on the initialized one). If
2861 init=0 then the ENGINE will not be iniatialized at all.
2864 *) Fix the 'app_verify_callback' interface so that the user-defined
2865 argument is actually passed to the callback: In the
2866 SSL_CTX_set_cert_verify_callback() prototype, the callback
2867 declaration has been changed from
2870 int (*cb)(X509_STORE_CTX *,void *);
2871 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2872 i=s->ctx->app_verify_callback(&ctx)
2873 has been changed into
2874 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2876 To update applications using SSL_CTX_set_cert_verify_callback(),
2877 a dummy argument can be added to their callback functions.
2878 [D. K. Smetters <smetters@parc.xerox.com>]
2880 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2881 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2883 *) Add and OPENSSL_LOAD_CONF define which will cause
2884 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2885 This allows older applications to transparently support certain
2886 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2887 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2888 load the config file and OPENSSL_add_all_algorithms_conf() which will
2889 always load it have also been added.
2892 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2893 Adjust NIDs and EVP layer.
2894 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2896 *) Config modules support in openssl utility.
2898 Most commands now load modules from the config file,
2899 though in a few (such as version) this isn't done
2900 because it couldn't be used for anything.
2902 In the case of ca and req the config file used is
2903 the same as the utility itself: that is the -config
2904 command line option can be used to specify an
2908 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2909 use "openssl_conf" if filename is NULL use default openssl config file.
2912 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2913 config section name. Add a new flag to tolerate a missing config file
2914 and move code to CONF_modules_load_file().
2917 *) Support for crypto accelerator cards from Accelerated Encryption
2918 Processing, www.aep.ie. (Use engine 'aep')
2919 The support was copied from 0.9.6c [engine] and adapted/corrected
2920 to work with the new engine framework.
2921 [AEP Inc. and Richard Levitte]
2923 *) Support for SureWare crypto accelerator cards from Baltimore
2924 Technologies. (Use engine 'sureware')
2925 The support was copied from 0.9.6c [engine] and adapted
2926 to work with the new engine framework.
2929 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2930 make the newer ENGINE framework commands for the CHIL engine work.
2931 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2933 *) Make it possible to produce shared libraries on ReliantUNIX.
2934 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2936 *) Add the configuration target debug-linux-ppro.
2937 Make 'openssl rsa' use the general key loading routines
2938 implemented in apps.c, and make those routines able to
2939 handle the key format FORMAT_NETSCAPE and the variant
2941 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2943 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2944 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2946 *) Add -keyform to rsautl, and document -engine.
2947 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2949 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2950 BIO_R_NO_SUCH_FILE error code rather than the generic
2951 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2954 *) Add new functions
2956 ERR_peek_last_error_line
2957 ERR_peek_last_error_line_data.
2958 These are similar to
2961 ERR_peek_error_line_data,
2962 but report on the latest error recorded rather than the first one
2963 still in the error queue.
2964 [Ben Laurie, Bodo Moeller]
2966 *) default_algorithms option in ENGINE config module. This allows things
2968 default_algorithms = ALL
2969 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2972 *) Prelminary ENGINE config module.
2975 *) New experimental application configuration code.
2978 *) Change the AES code to follow the same name structure as all other
2979 symmetric ciphers, and behave the same way. Move everything to
2980 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2981 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2983 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2984 [Ben Laurie and Theo de Raadt]
2986 *) Add option to output public keys in req command.
2987 [Massimiliano Pala madwolf@openca.org]
2989 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2990 (up to about 10% better than before for P-192 and P-224).
2993 *) New functions/macros
2995 SSL_CTX_set_msg_callback(ctx, cb)
2996 SSL_CTX_set_msg_callback_arg(ctx, arg)
2997 SSL_set_msg_callback(ssl, cb)
2998 SSL_set_msg_callback_arg(ssl, arg)
3000 to request calling a callback function
3002 void cb(int write_p, int version, int content_type,
3003 const void *buf, size_t len, SSL *ssl, void *arg)
3005 whenever a protocol message has been completely received
3006 (write_p == 0) or sent (write_p == 1). Here 'version' is the
3007 protocol version according to which the SSL library interprets
3008 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
3009 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
3010 the content type as defined in the SSL 3.0/TLS 1.0 protocol
3011 specification (change_cipher_spec(20), alert(21), handshake(22)).
3012 'buf' and 'len' point to the actual message, 'ssl' to the
3013 SSL object, and 'arg' is the application-defined value set by
3014 SSL[_CTX]_set_msg_callback_arg().
3016 'openssl s_client' and 'openssl s_server' have new '-msg' options
3017 to enable a callback that displays all protocol messages.
3020 *) Change the shared library support so shared libraries are built as
3021 soon as the corresponding static library is finished, and thereby get
3022 openssl and the test programs linked against the shared library.
3023 This still only happens when the keyword "shard" has been given to
3024 the configuration scripts.
3026 NOTE: shared library support is still an experimental thing, and
3027 backward binary compatibility is still not guaranteed.
3028 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
3030 *) Add support for Subject Information Access extension.
3031 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
3033 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
3034 additional bytes when new memory had to be allocated, not just
3035 when reusing an existing buffer.
3038 *) New command line and configuration option 'utf8' for the req command.
3039 This allows field values to be specified as UTF8 strings.
3042 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
3043 runs for the former and machine-readable output for the latter.
3046 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
3047 of the e-mail address in the DN (i.e., it will go into a certificate
3048 extension only). The new configuration file option 'email_in_dn = no'
3049 has the same effect.
3050 [Massimiliano Pala madwolf@openca.org]
3052 *) Change all functions with names starting with des_ to be starting
3053 with DES_ instead. Add wrappers that are compatible with libdes,
3054 but are named _ossl_old_des_*. Finally, add macros that map the
3055 des_* symbols to the corresponding _ossl_old_des_* if libdes
3056 compatibility is desired. If OpenSSL 0.9.6c compatibility is
3057 desired, the des_* symbols will be mapped to DES_*, with one
3060 Since we provide two compatibility mappings, the user needs to
3061 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
3062 compatibility is desired. The default (i.e., when that macro
3063 isn't defined) is OpenSSL 0.9.6c compatibility.
3065 There are also macros that enable and disable the support of old
3066 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
3067 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
3068 are defined, the default will apply: to support the old des routines.
3070 In either case, one must include openssl/des.h to get the correct
3071 definitions. Do not try to just include openssl/des_old.h, that
3074 NOTE: This is a major break of an old API into a new one. Software
3075 authors are encouraged to switch to the DES_ style functions. Some
3076 time in the future, des_old.h and the libdes compatibility functions
3077 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
3078 default), and then completely removed.
3081 *) Test for certificates which contain unsupported critical extensions.
3082 If such a certificate is found during a verify operation it is
3083 rejected by default: this behaviour can be overridden by either
3084 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
3085 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
3086 X509_supported_extension() has also been added which returns 1 if a
3087 particular extension is supported.
3090 *) Modify the behaviour of EVP cipher functions in similar way to digests
3091 to retain compatibility with existing code.
3094 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
3095 compatibility with existing code. In particular the 'ctx' parameter does
3096 not have to be to be initialized before the call to EVP_DigestInit() and
3097 it is tidied up after a call to EVP_DigestFinal(). New function
3098 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
3099 EVP_MD_CTX_copy() changed to not require the destination to be
3100 initialized valid and new function EVP_MD_CTX_copy_ex() added which
3101 requires the destination to be valid.
3103 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
3104 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
3107 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
3108 so that complete 'Handshake' protocol structures are kept in memory
3109 instead of overwriting 'msg_type' and 'length' with 'body' data.
3112 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
3113 [Massimo Santin via Richard Levitte]
3115 *) Major restructuring to the underlying ENGINE code. This includes
3116 reduction of linker bloat, separation of pure "ENGINE" manipulation
3117 (initialisation, etc) from functionality dealing with implementations
3118 of specific crypto iterfaces. This change also introduces integrated
3119 support for symmetric ciphers and digest implementations - so ENGINEs
3120 can now accelerate these by providing EVP_CIPHER and EVP_MD
3121 implementations of their own. This is detailed in crypto/engine/README
3122 as it couldn't be adequately described here. However, there are a few
3123 API changes worth noting - some RSA, DSA, DH, and RAND functions that
3124 were changed in the original introduction of ENGINE code have now
3125 reverted back - the hooking from this code to ENGINE is now a good
3126 deal more passive and at run-time, operations deal directly with
3127 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
3128 dereferencing through an ENGINE pointer any more. Also, the ENGINE
3129 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
3130 they were not being used by the framework as there is no concept of a
3131 BIGNUM_METHOD and they could not be generalised to the new
3132 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
3133 ENGINE_cpy() has been removed as it cannot be consistently defined in
3137 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
3140 *) Change mkdef.pl to sort symbols that get the same entry number,
3141 and make sure the automatically generated functions ERR_load_*
3142 become part of libeay.num as well.
3145 *) New function SSL_renegotiate_pending(). This returns true once
3146 renegotiation has been requested (either SSL_renegotiate() call
3147 or HelloRequest/ClientHello receveived from the peer) and becomes
3148 false once a handshake has been completed.
3149 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
3150 sends a HelloRequest, but does not ensure that a handshake takes
3151 place. SSL_renegotiate_pending() is useful for checking if the
3152 client has followed the request.)
3155 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
3156 By default, clients may request session resumption even during
3157 renegotiation (if session ID contexts permit); with this option,
3158 session resumption is possible only in the first handshake.
3160 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
3161 more bits available for options that should not be part of
3162 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
3165 *) Add some demos for certificate and certificate request creation.
3168 *) Make maximum certificate chain size accepted from the peer application
3169 settable (SSL*_get/set_max_cert_list()), as proposed by
3170 "Douglas E. Engert" <deengert@anl.gov>.
3173 *) Add support for shared libraries for Unixware-7
3174 (Boyd Lynn Gerber <gerberb@zenez.com>).
3177 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
3178 be done prior to destruction. Use this to unload error strings from
3179 ENGINEs that load their own error strings. NB: This adds two new API
3180 functions to "get" and "set" this destroy handler in an ENGINE.
3183 *) Alter all existing ENGINE implementations (except "openssl" and
3184 "openbsd") to dynamically instantiate their own error strings. This
3185 makes them more flexible to be built both as statically-linked ENGINEs
3186 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
3187 Also, add stub code to each that makes building them as self-contained
3188 shared-libraries easier (see README.ENGINE).
3191 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
3192 implementations into applications that are completely implemented in
3193 self-contained shared-libraries. The "dynamic" ENGINE exposes control
3194 commands that can be used to configure what shared-library to load and
3195 to control aspects of the way it is handled. Also, made an update to
3196 the README.ENGINE file that brings its information up-to-date and
3197 provides some information and instructions on the "dynamic" ENGINE
3198 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
3201 *) Make it possible to unload ranges of ERR strings with a new
3202 "ERR_unload_strings" function.
3205 *) Add a copy() function to EVP_MD.
3208 *) Make EVP_MD routines take a context pointer instead of just the
3209 md_data void pointer.
3212 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
3213 that the digest can only process a single chunk of data
3214 (typically because it is provided by a piece of
3215 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
3216 is only going to provide a single chunk of data, and hence the
3217 framework needn't accumulate the data for oneshot drivers.
3220 *) As with "ERR", make it possible to replace the underlying "ex_data"
3221 functions. This change also alters the storage and management of global
3222 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
3223 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
3224 index counters. The API functions that use this state have been changed
3225 to take a "class_index" rather than pointers to the class's local STACK
3226 and counter, and there is now an API function to dynamically create new
3227 classes. This centralisation allows us to (a) plug a lot of the
3228 thread-safety problems that existed, and (b) makes it possible to clean
3229 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
3230 such data would previously have always leaked in application code and
3231 workarounds were in place to make the memory debugging turn a blind eye
3232 to it. Application code that doesn't use this new function will still
3233 leak as before, but their memory debugging output will announce it now
3234 rather than letting it slide.
3236 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
3237 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
3238 has a return value to indicate success or failure.
3241 *) Make it possible to replace the underlying "ERR" functions such that the
3242 global state (2 LHASH tables and 2 locks) is only used by the "default"
3243 implementation. This change also adds two functions to "get" and "set"
3244 the implementation prior to it being automatically set the first time
3245 any other ERR function takes place. Ie. an application can call "get",
3246 pass the return value to a module it has just loaded, and that module
3247 can call its own "set" function using that value. This means the
3248 module's "ERR" operations will use (and modify) the error state in the
3249 application and not in its own statically linked copy of OpenSSL code.
3252 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3253 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3254 the operation, and provides a more encapsulated way for external code
3255 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3256 to use these functions rather than manually incrementing the counts.
3258 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3261 *) Add EVP test program.
3264 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3267 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3268 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3269 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3270 These allow a CRL to be built without having to access X509_CRL fields
3271 directly. Modify 'ca' application to use new functions.
3274 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3275 bug workarounds. Rollback attack detection is a security feature.
3276 The problem will only arise on OpenSSL servers when TLSv1 is not
3277 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3278 Software authors not wanting to support TLSv1 will have special reasons
3279 for their choice and can explicitly enable this option.
3280 [Bodo Moeller, Lutz Jaenicke]
3282 *) Rationalise EVP so it can be extended: don't include a union of
3283 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3284 (similar to those existing for EVP_CIPHER_CTX).
3289 EVP_MD_CTX_init(&md); /* new function call */
3290 EVP_DigestInit(&md, EVP_sha1());
3291 EVP_DigestUpdate(&md, in, len);
3292 EVP_DigestFinal(&md, out, NULL);
3293 EVP_MD_CTX_cleanup(&md); /* new function call */
3297 *) Make DES key schedule conform to the usual scheme, as well as
3298 correcting its structure. This means that calls to DES functions
3299 now have to pass a pointer to a des_key_schedule instead of a
3300 plain des_key_schedule (which was actually always a pointer
3303 des_key_schedule ks;
3305 des_set_key_checked(..., &ks);
3306 des_ncbc_encrypt(..., &ks, ...);
3308 (Note that a later change renames 'des_...' into 'DES_...'.)
3311 *) Initial reduction of linker bloat: the use of some functions, such as
3312 PEM causes large amounts of unused functions to be linked in due to
3313 poor organisation. For example pem_all.c contains every PEM function
3314 which has a knock on effect of linking in large amounts of (unused)
3315 ASN1 code. Grouping together similar functions and splitting unrelated
3316 functions prevents this.
3319 *) Cleanup of EVP macros.
3322 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3323 correct _ecb suffix.
3326 *) Add initial OCSP responder support to ocsp application. The
3327 revocation information is handled using the text based index
3328 use by the ca application. The responder can either handle
3329 requests generated internally, supplied in files (for example
3330 via a CGI script) or using an internal minimal server.
3333 *) Add configuration choices to get zlib compression for TLS.
3336 *) Changes to Kerberos SSL for RFC 2712 compliance:
3337 1. Implemented real KerberosWrapper, instead of just using
3338 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3339 2. Implemented optional authenticator field of KerberosWrapper.
3341 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3342 and authenticator structs; see crypto/krb5/.
3344 Generalized Kerberos calls to support multiple Kerberos libraries.
3345 [Vern Staats <staatsvr@asc.hpc.mil>,
3346 Jeffrey Altman <jaltman@columbia.edu>
3347 via Richard Levitte]
3349 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3350 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3351 values for each of the key sizes rather than having just
3352 parameters (and 'speed' generating keys each time).
3355 *) Speed up EVP routines.
3358 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3359 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3360 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3361 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3363 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3364 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3365 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3368 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3370 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3373 *) Added the OS2-EMX target.
3374 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3376 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3377 to support NCONF routines in extension code. New function CONF_set_nconf()
3378 to allow functions which take an NCONF to also handle the old LHASH
3379 structure: this means that the old CONF compatible routines can be
3380 retained (in particular wrt extensions) without having to duplicate the
3381 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3384 *) Enhance the general user interface with mechanisms for inner control
3385 and with possibilities to have yes/no kind of prompts.
3388 *) Change all calls to low level digest routines in the library and
3389 applications to use EVP. Add missing calls to HMAC_cleanup() and
3390 don't assume HMAC_CTX can be copied using memcpy().
3391 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3393 *) Add the possibility to control engines through control names but with
3394 arbitrary arguments instead of just a string.
3395 Change the key loaders to take a UI_METHOD instead of a callback
3396 function pointer. NOTE: this breaks binary compatibility with earlier
3397 versions of OpenSSL [engine].
3398 Adapt the nCipher code for these new conditions and add a card insertion
3402 *) Enhance the general user interface with mechanisms to better support
3403 dialog box interfaces, application-defined prompts, the possibility
3404 to use defaults (for example default passwords from somewhere else)
3405 and interrupts/cancellations.
3408 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3409 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3412 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3413 tidy up some unnecessarily weird code in 'sk_new()').
3414 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3416 *) Change the key loading routines for ENGINEs to use the same kind
3417 callback (pem_password_cb) as all other routines that need this
3421 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3422 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3423 than this minimum value is recommended.
3426 *) New random seeder for OpenVMS, using the system process statistics
3427 that are easily reachable.
3430 *) Windows apparently can't transparently handle global
3431 variables defined in DLLs. Initialisations such as:
3433 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3435 wont compile. This is used by the any applications that need to
3436 declare their own ASN1 modules. This was fixed by adding the option
3437 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3438 needed for static libraries under Win32.
3441 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3442 setting of purpose and trust fields. New X509_STORE trust and
3443 purpose functions and tidy up setting in other SSL functions.
3446 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3447 structure. These are inherited by X509_STORE_CTX when it is
3448 initialised. This allows various defaults to be set in the
3449 X509_STORE structure (such as flags for CRL checking and custom
3450 purpose or trust settings) for functions which only use X509_STORE_CTX
3451 internally such as S/MIME.
3453 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3454 trust settings if they are not set in X509_STORE. This allows X509_STORE
3455 purposes and trust (in S/MIME for example) to override any set by default.
3457 Add command line options for CRL checking to smime, s_client and s_server
3461 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3462 are set then the CRL is looked up in the X509_STORE structure and
3463 its validity and signature checked, then if the certificate is found
3464 in the CRL the verify fails with a revoked error.
3466 Various new CRL related callbacks added to X509_STORE_CTX structure.
3468 Command line options added to 'verify' application to support this.
3470 This needs some additional work, such as being able to handle multiple
3471 CRLs with different times, extension based lookup (rather than just
3472 by subject name) and ultimately more complete V2 CRL extension
3476 *) Add a general user interface API (crypto/ui/). This is designed
3477 to replace things like des_read_password and friends (backward
3478 compatibility functions using this new API are provided).
3479 The purpose is to remove prompting functions from the DES code
3480 section as well as provide for prompting through dialog boxes in
3481 a window system and the like.
3484 *) Add "ex_data" support to ENGINE so implementations can add state at a
3485 per-structure level rather than having to store it globally.
3488 *) Make it possible for ENGINE structures to be copied when retrieved by
3489 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3490 This causes the "original" ENGINE structure to act like a template,
3491 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3492 operational state can be localised to each ENGINE structure, despite the
3493 fact they all share the same "methods". New ENGINE structures returned in
3494 this case have no functional references and the return value is the single
3495 structural reference. This matches the single structural reference returned
3496 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3500 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3501 needs to match any other type at all we need to manually clear the
3505 *) Changes to the "openssl engine" utility to include;
3506 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3507 about an ENGINE's available control commands.
3508 - executing control commands from command line arguments using the
3509 '-pre' and '-post' switches. '-post' is only used if '-t' is
3510 specified and the ENGINE is successfully initialised. The syntax for
3511 the individual commands are colon-separated, for example;
3512 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3515 *) New dynamic control command support for ENGINEs. ENGINEs can now
3516 declare their own commands (numbers), names (strings), descriptions,
3517 and input types for run-time discovery by calling applications. A
3518 subset of these commands are implicitly classed as "executable"
3519 depending on their input type, and only these can be invoked through
3520 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3521 can be based on user input, config files, etc). The distinction is
3522 that "executable" commands cannot return anything other than a boolean
3523 result and can only support numeric or string input, whereas some
3524 discoverable commands may only be for direct use through
3525 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3526 pointers, or other custom uses. The "executable" commands are to
3527 support parameterisations of ENGINE behaviour that can be
3528 unambiguously defined by ENGINEs and used consistently across any
3529 OpenSSL-based application. Commands have been added to all the
3530 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3531 control over shared-library paths without source code alterations.
3534 *) Changed all ENGINE implementations to dynamically allocate their
3535 ENGINEs rather than declaring them statically. Apart from this being
3536 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3537 this also allows the implementations to compile without using the
3538 internal engine_int.h header.
3541 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3542 'const' value. Any code that should be able to modify a RAND_METHOD
3543 should already have non-const pointers to it (ie. they should only
3544 modify their own ones).
3547 *) Made a variety of little tweaks to the ENGINE code.
3548 - "atalla" and "ubsec" string definitions were moved from header files
3549 to C code. "nuron" string definitions were placed in variables
3550 rather than hard-coded - allowing parameterisation of these values
3551 later on via ctrl() commands.
3552 - Removed unused "#if 0"'d code.
3553 - Fixed engine list iteration code so it uses ENGINE_free() to release
3554 structural references.
3555 - Constified the RAND_METHOD element of ENGINE structures.
3556 - Constified various get/set functions as appropriate and added
3557 missing functions (including a catch-all ENGINE_cpy that duplicates
3558 all ENGINE values onto a new ENGINE except reference counts/state).
3559 - Removed NULL parameter checks in get/set functions. Setting a method
3560 or function to NULL is a way of cancelling out a previously set
3561 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3562 and doesn't justify the extra error symbols and code.
3563 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3564 flags from engine_int.h to engine.h.
3565 - Changed prototypes for ENGINE handler functions (init(), finish(),
3566 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3569 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3570 to the algorithm using long division. The binary algorithm can be
3571 used only if the modulus is odd. On 32-bit systems, it is faster
3572 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3573 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3574 up to 450 bits. In 64-bit environments, the binary algorithm
3575 appears to be advantageous for much longer moduli; here we use it
3576 for moduli up to 2048 bits.
3579 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3580 could not support the combine flag in choice fields.
3583 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3584 extensions from a certificate request to the certificate.
3587 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3588 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3589 file: this allows the display of the certificate about to be
3590 signed to be customised, to allow certain fields to be included
3591 or excluded and extension details. The old system didn't display
3592 multicharacter strings properly, omitted fields not in the policy
3593 and couldn't display additional details such as extensions.
3596 *) Function EC_POINTs_mul for multiple scalar multiplication
3597 of an arbitrary number of elliptic curve points
3598 \sum scalars[i]*points[i],
3599 optionally including the generator defined for the EC_GROUP:
3600 scalar*generator + \sum scalars[i]*points[i].
3602 EC_POINT_mul is a simple wrapper function for the typical case
3603 that the point list has just one item (besides the optional
3607 *) First EC_METHODs for curves over GF(p):
3609 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3610 operations and provides various method functions that can also
3611 operate with faster implementations of modular arithmetic.
3613 EC_GFp_mont_method() reuses most functions that are part of
3614 EC_GFp_simple_method, but uses Montgomery arithmetic.
3616 [Bodo Moeller; point addition and point doubling
3617 implementation directly derived from source code provided by
3618 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3620 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3621 crypto/ec/ec_lib.c):
3623 Curves are EC_GROUP objects (with an optional group generator)
3624 based on EC_METHODs that are built into the library.
3626 Points are EC_POINT objects based on EC_GROUP objects.
3628 Most of the framework would be able to handle curves over arbitrary
3629 finite fields, but as there are no obvious types for fields other
3630 than GF(p), some functions are limited to that for now.
3633 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3634 that the file contains a complete HTTP response.
3637 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3638 change the def and num file printf format specifier from "%-40sXXX"
3639 to "%-39s XXX". The latter will always guarantee a space after the
3640 field while the former will cause them to run together if the field
3641 is 40 of more characters long.
3644 *) Constify the cipher and digest 'method' functions and structures
3645 and modify related functions to take constant EVP_MD and EVP_CIPHER
3649 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3650 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3653 *) Modify EVP_Digest*() routines so they now return values. Although the
3654 internal software routines can never fail additional hardware versions
3658 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3660 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3661 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3664 ERR_R_NESTED_ASN1_ERROR
3666 ERR_R_MISSING_ASN1_EOS
3667 were 4 .. 9, conflicting with
3668 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3670 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3671 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3673 Add new error code 'ERR_R_INTERNAL_ERROR'.
3676 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3680 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3681 sets the subject name for a new request or supersedes the
3682 subject name in a given request. Formats that can be parsed are
3683 'CN=Some Name, OU=myOU, C=IT'
3685 'CN=Some Name/OU=myOU/C=IT'.
3687 Add options '-batch' and '-verbose' to 'openssl req'.
3688 [Massimiliano Pala <madwolf@hackmasters.net>]
3690 *) Introduce the possibility to access global variables through
3691 functions on platform were that's the best way to handle exporting
3692 global variables in shared libraries. To enable this functionality,
3693 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3694 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3695 is normally done by Configure or something similar).
3697 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3698 in the source file (foo.c) like this:
3700 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3701 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3703 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3704 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3706 OPENSSL_DECLARE_GLOBAL(int,foo);
3707 #define foo OPENSSL_GLOBAL_REF(foo)
3708 OPENSSL_DECLARE_GLOBAL(double,bar);
3709 #define bar OPENSSL_GLOBAL_REF(bar)
3711 The #defines are very important, and therefore so is including the
3712 header file everywhere where the defined globals are used.
3714 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3715 of ASN.1 items, but that structure is a bit different.
3717 The largest change is in util/mkdef.pl which has been enhanced with
3718 better and easier to understand logic to choose which symbols should
3719 go into the Windows .def files as well as a number of fixes and code
3720 cleanup (among others, algorithm keywords are now sorted
3721 lexicographically to avoid constant rewrites).
3724 *) In BN_div() keep a copy of the sign of 'num' before writing the
3725 result to 'rm' because if rm==num the value will be overwritten
3726 and produce the wrong result if 'num' is negative: this caused
3727 problems with BN_mod() and BN_nnmod().
3730 *) Function OCSP_request_verify(). This checks the signature on an
3731 OCSP request and verifies the signer certificate. The signer
3732 certificate is just checked for a generic purpose and OCSP request
3736 *) Add OCSP_check_validity() function to check the validity of OCSP
3737 responses. OCSP responses are prepared in real time and may only
3738 be a few seconds old. Simply checking that the current time lies
3739 between thisUpdate and nextUpdate max reject otherwise valid responses
3740 caused by either OCSP responder or client clock inaccuracy. Instead
3741 we allow thisUpdate and nextUpdate to fall within a certain period of
3742 the current time. The age of the response can also optionally be
3743 checked. Two new options -validity_period and -status_age added to
3747 *) If signature or public key algorithm is unrecognized print out its
3748 OID rather that just UNKNOWN.
3751 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3752 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3753 ID to be generated from the issuer certificate alone which can then be
3754 passed to OCSP_id_issuer_cmp().
3757 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3758 ASN1 modules to export functions returning ASN1_ITEM pointers
3759 instead of the ASN1_ITEM structures themselves. This adds several
3760 new macros which allow the underlying ASN1 function/structure to
3761 be accessed transparently. As a result code should not use ASN1_ITEM
3762 references directly (such as &X509_it) but instead use the relevant
3763 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3764 use of the new ASN1 code on platforms where exporting structures
3765 is problematical (for example in shared libraries) but exporting
3766 functions returning pointers to structures is not.
3769 *) Add support for overriding the generation of SSL/TLS session IDs.
3770 These callbacks can be registered either in an SSL_CTX or per SSL.
3771 The purpose of this is to allow applications to control, if they wish,
3772 the arbitrary values chosen for use as session IDs, particularly as it
3773 can be useful for session caching in multiple-server environments. A
3774 command-line switch for testing this (and any client code that wishes
3775 to use such a feature) has been added to "s_server".
3776 [Geoff Thorpe, Lutz Jaenicke]
3778 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3779 of the form '#if defined(...) || defined(...) || ...' and
3780 '#if !defined(...) && !defined(...) && ...'. This also avoids
3781 the growing number of special cases it was previously handling.
3784 *) Make all configuration macros available for application by making
3785 sure they are available in opensslconf.h, by giving them names starting
3786 with "OPENSSL_" to avoid conflicts with other packages and by making
3787 sure e_os2.h will cover all platform-specific cases together with
3789 Additionally, it is now possible to define configuration/platform-
3790 specific names (called "system identities"). In the C code, these
3791 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3792 macro with the name beginning with "OPENSSL_SYS_", which is determined
3793 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3797 *) New option -set_serial to 'req' and 'x509' this allows the serial
3798 number to use to be specified on the command line. Previously self
3799 signed certificates were hard coded with serial number 0 and the
3800 CA options of 'x509' had to use a serial number in a file which was
3804 *) New options to 'ca' utility to support V2 CRL entry extensions.
3805 Currently CRL reason, invalidity date and hold instruction are
3806 supported. Add new CRL extensions to V3 code and some new objects.
3809 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3810 disable standard block padding (aka PKCS#5 padding) in the EVP
3811 API, which was previously mandatory. This means that the data is
3812 not padded in any way and so the total length much be a multiple
3813 of the block size, otherwise an error occurs.
3816 *) Initial (incomplete) OCSP SSL support.
3819 *) New function OCSP_parse_url(). This splits up a URL into its host,
3820 port and path components: primarily to parse OCSP URLs. New -url
3821 option to ocsp utility.
3824 *) New nonce behavior. The return value of OCSP_check_nonce() now
3825 reflects the various checks performed. Applications can decide
3826 whether to tolerate certain situations such as an absent nonce
3827 in a response when one was present in a request: the ocsp application
3828 just prints out a warning. New function OCSP_add1_basic_nonce()
3829 this is to allow responders to include a nonce in a response even if
3830 the request is nonce-less.
3833 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3834 skipped when using openssl x509 multiple times on a single input file,
3835 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3838 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3839 set string type: to handle setting ASN1_TIME structures. Fix ca
3840 utility to correctly initialize revocation date of CRLs.
3843 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3844 the clients preferred ciphersuites and rather use its own preferences.
3845 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3846 Internet Explorer by ensuring unchanged hash method during stepup.
3847 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3850 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3851 to aes and add a new 'exist' option to print out symbols that don't
3855 *) Additional options to ocsp utility to allow flags to be set and
3856 additional certificates supplied.
3859 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3860 OCSP client a number of certificate to only verify the response
3864 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3865 handle the new API. Currently only ECB, CBC modes supported. Add new
3868 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3869 Encryption Standard (AES) Ciphersuites for Transport Layer
3870 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3871 not enabled by default and were not part of the "ALL" ciphersuite
3872 alias because they were not yet official; they could be
3873 explicitly requested by specifying the "AESdraft" ciphersuite
3874 group alias. In the final release of OpenSSL 0.9.7, the group
3875 alias is called "AES" and is part of "ALL".)
3876 [Ben Laurie, Steve Henson, Bodo Moeller]
3878 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3879 request to response.
3882 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3883 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3884 extract information from a certificate request. OCSP_response_create()
3885 creates a response and optionally adds a basic response structure.
3886 OCSP_basic_add1_status() adds a complete single response to a basic
3887 response and returns the OCSP_SINGLERESP structure just added (to allow
3888 extensions to be included for example). OCSP_basic_add1_cert() adds a
3889 certificate to a basic response and OCSP_basic_sign() signs a basic
3890 response with various flags. New helper functions ASN1_TIME_check()
3891 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3892 (converts ASN1_TIME to GeneralizedTime).
3895 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3896 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3897 structure from a certificate. X509_pubkey_digest() digests the public_key
3898 contents: this is used in various key identifiers.
3901 *) Make sk_sort() tolerate a NULL argument.
3902 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3904 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3905 passed by the function are trusted implicitly. If any of them signed the
3906 response then it is assumed to be valid and is not verified.
3909 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3910 to data. This was previously part of the PKCS7 ASN1 code. This
3911 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3912 [Steve Henson, reported by Kenneth R. Robinette
3913 <support@securenetterm.com>]
3915 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3916 routines: without these tracing memory leaks is very painful.
3917 Fix leaks in PKCS12 and PKCS7 routines.
3920 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3921 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3922 effectively meant GeneralizedTime would never be used. Now it
3923 is initialised to -1 but X509_time_adj() now has to check the value
3924 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3925 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3926 [Steve Henson, reported by Kenneth R. Robinette
3927 <support@securenetterm.com>]
3929 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3930 result in a zero length in the ASN1_INTEGER structure which was
3931 not consistent with the structure when d2i_ASN1_INTEGER() was used
3932 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3933 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3934 where it did not print out a minus for negative ASN1_INTEGER.
3937 *) Add summary printout to ocsp utility. The various functions which
3938 convert status values to strings have been renamed to:
3939 OCSP_response_status_str(), OCSP_cert_status_str() and
3940 OCSP_crl_reason_str() and are no longer static. New options
3941 to verify nonce values and to disable verification. OCSP response
3942 printout format cleaned up.
3945 *) Add additional OCSP certificate checks. These are those specified
3946 in RFC2560. This consists of two separate checks: the CA of the
3947 certificate being checked must either be the OCSP signer certificate
3948 or the issuer of the OCSP signer certificate. In the latter case the
3949 OCSP signer certificate must contain the OCSP signing extended key
3950 usage. This check is performed by attempting to match the OCSP
3951 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3952 in the OCSP_CERTID structures of the response.
3955 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3956 and related routines. This uses the standard OpenSSL certificate
3957 verify routines to perform initial checks (just CA validity) and
3958 to obtain the certificate chain. Then additional checks will be
3959 performed on the chain. Currently the root CA is checked to see
3960 if it is explicitly trusted for OCSP signing. This is used to set
3961 a root CA as a global signing root: that is any certificate that
3962 chains to that CA is an acceptable OCSP signing certificate.
3965 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3966 extensions from a separate configuration file.
3967 As when reading extensions from the main configuration file,
3968 the '-extensions ...' option may be used for specifying the
3970 [Massimiliano Pala <madwolf@comune.modena.it>]
3972 *) New OCSP utility. Allows OCSP requests to be generated or
3973 read. The request can be sent to a responder and the output
3974 parsed, outputed or printed in text form. Not complete yet:
3975 still needs to check the OCSP response validity.
3978 *) New subcommands for 'openssl ca':
3979 'openssl ca -status <serial>' prints the status of the cert with
3980 the given serial number (according to the index file).
3981 'openssl ca -updatedb' updates the expiry status of certificates
3983 [Massimiliano Pala <madwolf@comune.modena.it>]
3985 *) New '-newreq-nodes' command option to CA.pl. This is like
3986 '-newreq', but calls 'openssl req' with the '-nodes' option
3987 so that the resulting key is not encrypted.
3988 [Damien Miller <djm@mindrot.org>]
3990 *) New configuration for the GNU Hurd.
3991 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3993 *) Initial code to implement OCSP basic response verify. This
3994 is currently incomplete. Currently just finds the signer's
3995 certificate and verifies the signature on the response.
3998 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3999 value of OPENSSLDIR. This is available via the new '-d' option
4000 to 'openssl version', and is also included in 'openssl version -a'.
4003 *) Allowing defining memory allocation callbacks that will be given
4004 file name and line number information in additional arguments
4005 (a const char* and an int). The basic functionality remains, as
4006 well as the original possibility to just replace malloc(),
4007 realloc() and free() by functions that do not know about these
4008 additional arguments. To register and find out the current
4009 settings for extended allocation functions, the following
4010 functions are provided:
4012 CRYPTO_set_mem_ex_functions
4013 CRYPTO_set_locked_mem_ex_functions
4014 CRYPTO_get_mem_ex_functions
4015 CRYPTO_get_locked_mem_ex_functions
4017 These work the same way as CRYPTO_set_mem_functions and friends.
4018 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
4019 extended allocation function is enabled.
4020 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
4021 a conventional allocation function is enabled.
4022 [Richard Levitte, Bodo Moeller]
4024 *) Finish off removing the remaining LHASH function pointer casts.
4025 There should no longer be any prototype-casting required when using
4026 the LHASH abstraction, and any casts that remain are "bugs". See
4027 the callback types and macros at the head of lhash.h for details
4028 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
4031 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
4032 If /dev/[u]random devices are not available or do not return enough
4033 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
4035 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
4036 /etc/entropy will be queried once each in this sequence, quering stops
4037 when enough entropy was collected without querying more sockets.
4040 *) Change the Unix RAND_poll() variant to be able to poll several
4041 random devices, as specified by DEVRANDOM, until a sufficient amount
4042 of data has been collected. We spend at most 10 ms on each file
4043 (select timeout) and read in non-blocking mode. DEVRANDOM now
4044 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
4045 (previously it was just the string "/dev/urandom"), so on typical
4046 platforms the 10 ms delay will never occur.
4047 Also separate out the Unix variant to its own file, rand_unix.c.
4048 For VMS, there's a currently-empty rand_vms.c.
4051 *) Move OCSP client related routines to ocsp_cl.c. These
4052 provide utility functions which an application needing
4053 to issue a request to an OCSP responder and analyse the
4054 response will typically need: as opposed to those which an
4055 OCSP responder itself would need which will be added later.
4057 OCSP_request_sign() signs an OCSP request with an API similar
4058 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
4059 response. OCSP_response_get1_basic() extracts basic response
4060 from response. OCSP_resp_find_status(): finds and extracts status
4061 information from an OCSP_CERTID structure (which will be created
4062 when the request structure is built). These are built from lower
4063 level functions which work on OCSP_SINGLERESP structures but
4064 wont normally be used unless the application wishes to examine
4065 extensions in the OCSP response for example.
4067 Replace nonce routines with a pair of functions.
4068 OCSP_request_add1_nonce() adds a nonce value and optionally
4069 generates a random value. OCSP_check_nonce() checks the
4070 validity of the nonce in an OCSP response.
4073 *) Change function OCSP_request_add() to OCSP_request_add0_id().
4074 This doesn't copy the supplied OCSP_CERTID and avoids the
4075 need to free up the newly created id. Change return type
4076 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
4077 This can then be used to add extensions to the request.
4078 Deleted OCSP_request_new(), since most of its functionality
4079 is now in OCSP_REQUEST_new() (and the case insensitive name
4080 clash) apart from the ability to set the request name which
4081 will be added elsewhere.
4084 *) Update OCSP API. Remove obsolete extensions argument from
4085 various functions. Extensions are now handled using the new
4086 OCSP extension code. New simple OCSP HTTP function which
4087 can be used to send requests and parse the response.
4090 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
4091 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
4092 uses the special reorder version of SET OF to sort the attributes
4093 and reorder them to match the encoded order. This resolves a long
4094 standing problem: a verify on a PKCS7 structure just after signing
4095 it used to fail because the attribute order did not match the
4096 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
4097 it uses the received order. This is necessary to tolerate some broken
4098 software that does not order SET OF. This is handled by encoding
4099 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
4100 to produce the required SET OF.
4103 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
4104 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
4105 files to get correct declarations of the ASN.1 item variables.
4108 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
4109 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
4110 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
4111 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
4112 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
4113 ASN1_ITEM and no wrapper functions.
4116 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
4117 replace the old function pointer based I/O routines. Change most of
4118 the *_d2i_bio() and *_d2i_fp() functions to use these.
4121 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
4122 lines, recognice more "algorithms" that can be deselected, and make
4123 it complain about algorithm deselection that isn't recognised.
4126 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
4127 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
4128 to use new functions. Add NO_ASN1_OLD which can be set to remove
4129 some old style ASN1 functions: this can be used to determine if old
4130 code will still work when these eventually go away.
4133 *) New extension functions for OCSP structures, these follow the
4134 same conventions as certificates and CRLs.
4137 *) New function X509V3_add1_i2d(). This automatically encodes and
4138 adds an extension. Its behaviour can be customised with various
4139 flags to append, replace or delete. Various wrappers added for
4140 certifcates and CRLs.
4143 *) Fix to avoid calling the underlying ASN1 print routine when
4144 an extension cannot be parsed. Correct a typo in the
4145 OCSP_SERVICELOC extension. Tidy up print OCSP format.
4148 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
4149 entries for variables.
4152 *) Add functionality to apps/openssl.c for detecting locking
4153 problems: As the program is single-threaded, all we have
4154 to do is register a locking callback using an array for
4155 storing which locks are currently held by the program.
4158 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
4159 SSL_get_ex_data_X509_STORE_idx(), which is used in
4160 ssl_verify_cert_chain() and thus can be called at any time
4161 during TLS/SSL handshakes so that thread-safety is essential.
4162 Unfortunately, the ex_data design is not at all suited
4163 for multi-threaded use, so it probably should be abolished.
4166 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
4167 [Broadcom, tweaked and integrated by Geoff Thorpe]
4169 *) Move common extension printing code to new function
4170 X509V3_print_extensions(). Reorganise OCSP print routines and
4171 implement some needed OCSP ASN1 functions. Add OCSP extensions.
4174 *) New function X509_signature_print() to remove duplication in some
4178 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
4179 set (this was treated exactly the same as SET OF previously). This
4180 is used to reorder the STACK representing the structure to match the
4181 encoding. This will be used to get round a problem where a PKCS7
4182 structure which was signed could not be verified because the STACK
4183 order did not reflect the encoded order.
4186 *) Reimplement the OCSP ASN1 module using the new code.
4189 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
4190 for its ASN1 operations. The old style function pointers still exist
4191 for now but they will eventually go away.
4194 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
4195 completely replaces the old ASN1 functionality with a table driven
4196 encoder and decoder which interprets an ASN1_ITEM structure describing
4197 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
4198 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
4199 has also been converted to the new form.
4202 *) Change BN_mod_exp_recp so that negative moduli are tolerated
4203 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
4204 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
4205 for negative moduli.
4208 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
4209 of not touching the result's sign bit.
4212 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
4216 *) Changed the LHASH code to use prototypes for callbacks, and created
4217 macros to declare and implement thin (optionally static) functions
4218 that provide type-safety and avoid function pointer casting for the
4219 type-specific callbacks.
4222 *) Added Kerberos Cipher Suites to be used with TLS, as written in
4224 [Veers Staats <staatsvr@asc.hpc.mil>,
4225 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
4227 *) Reformat the FAQ so the different questions and answers can be divided
4228 in sections depending on the subject.
4231 *) Have the zlib compression code load ZLIB.DLL dynamically under
4235 *) New function BN_mod_sqrt for computing square roots modulo a prime
4236 (using the probabilistic Tonelli-Shanks algorithm unless
4237 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
4238 be handled deterministically).
4239 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4241 *) Make BN_mod_inverse faster by explicitly handling small quotients
4242 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
4243 512 bits], about 30% for larger ones [1024 or 2048 bits].)
4246 *) New function BN_kronecker.
4249 *) Fix BN_gcd so that it works on negative inputs; the result is
4250 positive unless both parameters are zero.
4251 Previously something reasonably close to an infinite loop was
4252 possible because numbers could be growing instead of shrinking
4253 in the implementation of Euclid's algorithm.
4256 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4257 sign of the number in question.
4259 Fix BN_is_word(a,w) to work correctly for w == 0.
4261 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4262 because its test if the absolute value of 'a' equals 'w'.
4263 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4264 it exists mostly for use in the implementations of BN_is_zero(),
4265 BN_is_one(), and BN_is_word().
4268 *) New function BN_swap.
4271 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4272 the exponentiation functions are more likely to produce reasonable
4273 results on negative inputs.
4276 *) Change BN_mod_mul so that the result is always non-negative.
4277 Previously, it could be negative if one of the factors was negative;
4278 I don't think anyone really wanted that behaviour.
4281 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4282 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4283 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4284 and add new functions:
4293 BN_mod_lshift1_quick
4297 These functions always generate non-negative results.
4299 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4300 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4302 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4303 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4304 be reduced modulo m.
4305 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4308 The following entry accidentily appeared in the CHANGES file
4309 distributed with OpenSSL 0.9.7. The modifications described in
4310 it do *not* apply to OpenSSL 0.9.7.
4312 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4313 was actually never needed) and in BN_mul(). The removal in BN_mul()
4314 required a small change in bn_mul_part_recursive() and the addition
4315 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4316 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4317 bn_sub_words() and bn_add_words() except they take arrays with
4322 *) In 'openssl passwd', verify passwords read from the terminal
4323 unless the '-salt' option is used (which usually means that
4324 verification would just waste user's time since the resulting
4325 hash is going to be compared with some given password hash)
4326 or the new '-noverify' option is used.
4328 This is an incompatible change, but it does not affect
4329 non-interactive use of 'openssl passwd' (passwords on the command
4330 line, '-stdin' option, '-in ...' option) and thus should not
4334 *) Remove all references to RSAref, since there's no more need for it.
4337 *) Make DSO load along a path given through an environment variable
4338 (SHLIB_PATH) with shl_load().
4341 *) Constify the ENGINE code as a result of BIGNUM constification.
4342 Also constify the RSA code and most things related to it. In a
4343 few places, most notable in the depth of the ASN.1 code, ugly
4344 casts back to non-const were required (to be solved at a later
4348 *) Make it so the openssl application has all engines loaded by default.
4351 *) Constify the BIGNUM routines a little more.
4354 *) Add the following functions:
4356 ENGINE_load_cswift()
4358 ENGINE_load_atalla()
4360 ENGINE_load_builtin_engines()
4362 That way, an application can itself choose if external engines that
4363 are built-in in OpenSSL shall ever be used or not. The benefit is
4364 that applications won't have to be linked with libdl or other dso
4365 libraries unless it's really needed.
4367 Changed 'openssl engine' to load all engines on demand.
4368 Changed the engine header files to avoid the duplication of some
4369 declarations (they differed!).
4372 *) 'openssl engine' can now list capabilities.
4375 *) Better error reporting in 'openssl engine'.
4378 *) Never call load_dh_param(NULL) in s_server.
4381 *) Add engine application. It can currently list engines by name and
4382 identity, and test if they are actually available.
4385 *) Improve RPM specification file by forcing symbolic linking and making
4386 sure the installed documentation is also owned by root.root.
4387 [Damien Miller <djm@mindrot.org>]
4389 *) Give the OpenSSL applications more possibilities to make use of
4390 keys (public as well as private) handled by engines.
4393 *) Add OCSP code that comes from CertCo.
4396 *) Add VMS support for the Rijndael code.
4399 *) Added untested support for Nuron crypto accelerator.
4402 *) Add support for external cryptographic devices. This code was
4403 previously distributed separately as the "engine" branch.
4404 [Geoff Thorpe, Richard Levitte]
4406 *) Rework the filename-translation in the DSO code. It is now possible to
4407 have far greater control over how a "name" is turned into a filename
4408 depending on the operating environment and any oddities about the
4409 different shared library filenames on each system.
4412 *) Support threads on FreeBSD-elf in Configure.
4415 *) Fix for SHA1 assembly problem with MASM: it produces
4416 warnings about corrupt line number information when assembling
4417 with debugging information. This is caused by the overlapping
4419 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4422 NCONF_get_number() has no error checking at all. As a replacement,
4423 NCONF_get_number_e() is defined (_e for "error checking") and is
4424 promoted strongly. The old NCONF_get_number is kept around for
4425 binary backward compatibility.
4426 Make it possible for methods to load from something other than a BIO,
4427 by providing a function pointer that is given a name instead of a BIO.
4428 For example, this could be used to load configuration data from an
4432 *) Fix for non blocking accept BIOs. Added new I/O special reason
4433 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4434 with non blocking I/O was not possible because no retry code was
4435 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4439 *) Added the beginnings of Rijndael support.
4442 *) Fix for bug in DirectoryString mask setting. Add support for
4443 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4444 to allow certificate printing to more controllable, additional
4445 'certopt' option to 'x509' to allow new printing options to be
4449 *) Clean old EAY MD5 hack from e_os.h.
4452 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4454 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4455 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4456 [Joe Orton, Steve Henson]
4458 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4460 *) Fix additional bug revealed by the NISCC test suite:
4462 Stop bug triggering large recursion when presented with
4463 certain ASN.1 tags (CVE-2003-0851)
4466 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4468 *) Fix various bugs revealed by running the NISCC test suite:
4470 Stop out of bounds reads in the ASN1 code when presented with
4471 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4473 If verify callback ignores invalid public key errors don't try to check
4474 certificate signature with the NULL public key.
4478 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4479 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4483 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4484 extra data after the compression methods not only for TLS 1.0
4485 but also for SSL 3.0 (as required by the specification).
4486 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4488 *) Change X509_certificate_type() to mark the key as exported/exportable
4489 when it's 512 *bits* long, not 512 bytes.
4492 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4494 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4495 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4496 a protocol version number mismatch like a decryption error
4497 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4500 *) Turn on RSA blinding by default in the default implementation
4501 to avoid a timing attack. Applications that don't want it can call
4502 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4503 They would be ill-advised to do so in most cases.
4504 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4506 *) Change RSA blinding code so that it works when the PRNG is not
4507 seeded (in this case, the secret RSA exponent is abused as
4508 an unpredictable seed -- if it is not unpredictable, there
4509 is no point in blinding anyway). Make RSA blinding thread-safe
4510 by remembering the creator's thread ID in rsa->blinding and
4511 having all other threads use local one-time blinding factors
4512 (this requires more computation than sharing rsa->blinding, but
4513 avoids excessive locking; and if an RSA object is not shared
4514 between threads, blinding will still be very fast).
4517 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4519 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4520 via timing by performing a MAC computation even if incorrrect
4521 block cipher padding has been found. This is a countermeasure
4522 against active attacks where the attacker has to distinguish
4523 between bad padding and a MAC verification error. (CVE-2003-0078)
4525 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4526 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4527 Martin Vuagnoux (EPFL, Ilion)]
4529 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4531 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4532 memory from it's contents. This is done with a counter that will
4533 place alternating values in each byte. This can be used to solve
4534 two issues: 1) the removal of calls to memset() by highly optimizing
4535 compilers, and 2) cleansing with other values than 0, since those can
4536 be read through on certain media, for example a swap space on disk.
4539 *) Bugfix: client side session caching did not work with external caching,
4540 because the session->cipher setting was not restored when reloading
4541 from the external cache. This problem was masked, when
4542 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4543 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4546 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4547 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4548 [Zeev Lieber <zeev-l@yahoo.com>]
4550 *) Undo an undocumented change introduced in 0.9.6e which caused
4551 repeated calls to OpenSSL_add_all_ciphers() and
4552 OpenSSL_add_all_digests() to be ignored, even after calling
4556 *) Change the default configuration reader to deal with last line not
4557 being properly terminated.
4560 *) Change X509_NAME_cmp() so it applies the special rules on handling
4561 DN values that are of type PrintableString, as well as RDNs of type
4562 emailAddress where the value has the type ia5String.
4563 [stefank@valicert.com via Richard Levitte]
4565 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4566 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4567 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4568 the bitwise-OR of the two for use by the majority of applications
4569 wanting this behaviour, and update the docs. The documented
4570 behaviour and actual behaviour were inconsistent and had been
4571 changing anyway, so this is more a bug-fix than a behavioural
4573 [Geoff Thorpe, diagnosed by Nadav Har'El]
4575 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4576 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4579 *) Fix initialization code race conditions in
4580 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4581 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4582 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4583 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4584 ssl2_get_cipher_by_char(),
4585 ssl3_get_cipher_by_char().
4586 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4588 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4589 the cached sessions are flushed, as the remove_cb() might use ex_data
4590 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4591 (see [openssl.org #212]).
4592 [Geoff Thorpe, Lutz Jaenicke]
4594 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4595 length, instead of the encoding length to d2i_ASN1_OBJECT.
4598 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4600 *) [In 0.9.6g-engine release:]
4601 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4602 [Lynn Gazis <lgazis@rainbow.com>]
4604 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4606 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4607 and get fix the header length calculation.
4608 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4609 Alon Kantor <alonk@checkpoint.com> (and others),
4612 *) Use proper error handling instead of 'assertions' in buffer
4613 overflow checks added in 0.9.6e. This prevents DoS (the
4614 assertions could call abort()).
4615 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4617 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4619 *) Add various sanity checks to asn1_get_length() to reject
4620 the ASN1 length bytes if they exceed sizeof(long), will appear
4621 negative or the content length exceeds the length of the
4623 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4625 *) Fix cipher selection routines: ciphers without encryption had no flags
4626 for the cipher strength set and where therefore not handled correctly
4627 by the selection routines (PR #130).
4630 *) Fix EVP_dsa_sha macro.
4634 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4635 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4636 that was added in OpenSSL 0.9.6d.
4638 As the countermeasure turned out to be incompatible with some
4639 broken SSL implementations, the new option is part of SSL_OP_ALL.
4640 SSL_OP_ALL is usually employed when compatibility with weird SSL
4641 implementations is desired (e.g. '-bugs' option to 's_client' and
4642 's_server'), so the new option is automatically set in many
4646 *) Changes in security patch:
4648 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4649 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4650 Air Force Materiel Command, USAF, under agreement number
4653 *) Add various sanity checks to asn1_get_length() to reject
4654 the ASN1 length bytes if they exceed sizeof(long), will appear
4655 negative or the content length exceeds the length of the
4656 supplied buffer. (CVE-2002-0659)
4657 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4659 *) Assertions for various potential buffer overflows, not known to
4661 [Ben Laurie (CHATS)]
4663 *) Various temporary buffers to hold ASCII versions of integers were
4664 too small for 64 bit platforms. (CVE-2002-0655)
4665 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4667 *) Remote buffer overflow in SSL3 protocol - an attacker could
4668 supply an oversized session ID to a client. (CVE-2002-0656)
4669 [Ben Laurie (CHATS)]
4671 *) Remote buffer overflow in SSL2 protocol - an attacker could
4672 supply an oversized client master key. (CVE-2002-0656)
4673 [Ben Laurie (CHATS)]
4675 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4677 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4678 encoded as NULL) with id-dsa-with-sha1.
4679 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4681 *) Check various X509_...() return values in apps/req.c.
4682 [Nils Larsch <nla@trustcenter.de>]
4684 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4685 an end-of-file condition would erronously be flagged, when the CRLF
4686 was just at the end of a processed block. The bug was discovered when
4687 processing data through a buffering memory BIO handing the data to a
4688 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4689 <ptsekov@syntrex.com> and Nedelcho Stanev.
4692 *) Implement a countermeasure against a vulnerability recently found
4693 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4694 before application data chunks to avoid the use of known IVs
4695 with data potentially chosen by the attacker.
4698 *) Fix length checks in ssl3_get_client_hello().
4701 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4702 to prevent ssl3_read_internal() from incorrectly assuming that
4703 ssl3_read_bytes() found application data while handshake
4704 processing was enabled when in fact s->s3->in_read_app_data was
4705 merely automatically cleared during the initial handshake.
4706 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4708 *) Fix object definitions for Private and Enterprise: they were not
4709 recognized in their shortname (=lowercase) representation. Extend
4710 obj_dat.pl to issue an error when using undefined keywords instead
4711 of silently ignoring the problem (Svenning Sorensen
4712 <sss@sss.dnsalias.net>).
4715 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4716 generators, i.e. generators other than 2 and 5. (Previously, the
4717 code did not properly initialise the 'add' and 'rem' values to
4718 BN_generate_prime().)
4720 In the new general case, we do not insist that 'generator' is
4721 actually a primitive root: This requirement is rather pointless;
4722 a generator of the order-q subgroup is just as good, if not
4726 *) Map new X509 verification errors to alerts. Discovered and submitted by
4727 Tom Wu <tom@arcot.com>.
4730 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4731 returning non-zero before the data has been completely received
4732 when using non-blocking I/O.
4733 [Bodo Moeller; problem pointed out by John Hughes]
4735 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4736 [Ben Laurie, Lutz Jaenicke]
4738 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4739 Yoram Zahavi <YoramZ@gilian.com>).
4742 *) Add information about CygWin 1.3 and on, and preserve proper
4743 configuration for the versions before that.
4744 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4746 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4747 check whether we deal with a copy of a session and do not delete from
4748 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4749 <izhar@checkpoint.com>.
4752 *) Do not store session data into the internal session cache, if it
4753 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4754 flag is set). Proposed by Aslam <aslam@funk.com>.
4757 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4761 *) [In 0.9.6d-engine release:]
4762 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4763 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4765 *) Add the configuration target linux-s390x.
4766 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4768 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4769 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4770 variable as an indication that a ClientHello message has been
4771 received. As the flag value will be lost between multiple
4772 invocations of ssl3_accept when using non-blocking I/O, the
4773 function may not be aware that a handshake has actually taken
4774 place, thus preventing a new session from being added to the
4777 To avoid this problem, we now set s->new_session to 2 instead of
4778 using a local variable.
4779 [Lutz Jaenicke, Bodo Moeller]
4781 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4782 if the SSL_R_LENGTH_MISMATCH error is detected.
4783 [Geoff Thorpe, Bodo Moeller]
4785 *) New 'shared_ldflag' column in Configure platform table.
4788 *) Fix EVP_CIPHER_mode macro.
4789 ["Dan S. Camper" <dan@bti.net>]
4791 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4792 type, we must throw them away by setting rr->length to 0.
4793 [D P Chang <dpc@qualys.com>]
4795 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4797 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4798 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4799 worked incorrectly for those cases where range = 10..._2 and
4800 3*range is two bits longer than range.)
4803 *) Only add signing time to PKCS7 structures if it is not already
4807 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4808 OBJ_ld_ce should be OBJ_id_ce.
4809 Also some ip-pda OIDs in crypto/objects/objects.txt were
4810 incorrect (cf. RFC 3039).
4811 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4813 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4814 returns early because it has nothing to do.
4815 [Andy Schneider <andy.schneider@bjss.co.uk>]
4817 *) [In 0.9.6c-engine release:]
4818 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4819 [Andy Schneider <andy.schneider@bjss.co.uk>]
4821 *) [In 0.9.6c-engine release:]
4822 Add support for Cryptographic Appliance's keyserver technology.
4823 (Use engine 'keyclient')
4824 [Cryptographic Appliances and Geoff Thorpe]
4826 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4827 is called via tools/c89.sh because arguments have to be
4828 rearranged (all '-L' options must appear before the first object
4830 [Richard Shapiro <rshapiro@abinitio.com>]
4832 *) [In 0.9.6c-engine release:]
4833 Add support for Broadcom crypto accelerator cards, backported
4835 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4837 *) [In 0.9.6c-engine release:]
4838 Add support for SureWare crypto accelerator cards from
4839 Baltimore Technologies. (Use engine 'sureware')
4840 [Baltimore Technologies and Mark Cox]
4842 *) [In 0.9.6c-engine release:]
4843 Add support for crypto accelerator cards from Accelerated
4844 Encryption Processing, www.aep.ie. (Use engine 'aep')
4845 [AEP Inc. and Mark Cox]
4847 *) Add a configuration entry for gcc on UnixWare.
4848 [Gary Benson <gbenson@redhat.com>]
4850 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4851 messages are stored in a single piece (fixed-length part and
4852 variable-length part combined) and fix various bugs found on the way.
4855 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4856 instead. BIO_gethostbyname() does not know what timeouts are
4857 appropriate, so entries would stay in cache even when they have
4859 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4861 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4862 faced with a pathologically small ClientHello fragment that does
4863 not contain client_version: Instead of aborting with an error,
4864 simply choose the highest available protocol version (i.e.,
4865 TLS 1.0 unless it is disabled). In practice, ClientHello
4866 messages are never sent like this, but this change gives us
4867 strictly correct behaviour at least for TLS.
4870 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4871 never resets s->method to s->ctx->method when called from within
4872 one of the SSL handshake functions.
4873 [Bodo Moeller; problem pointed out by Niko Baric]
4875 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4876 (sent using the client's version number) if client_version is
4877 smaller than the protocol version in use. Also change
4878 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4879 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4880 the client will at least see that alert.
4883 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4887 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4888 client receives HelloRequest while in a handshake.
4889 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4891 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4892 should end in 'break', not 'goto end' which circuments various
4893 cleanups done in state SSL_ST_OK. But session related stuff
4894 must be disabled for SSL_ST_OK in the case that we just sent a
4897 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4898 before just sending a HelloRequest.
4899 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4901 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4902 reveal whether illegal block cipher padding was found or a MAC
4903 verification error occured. (Neither SSLerr() codes nor alerts
4904 are directly visible to potential attackers, but the information
4905 may leak via logfiles.)
4907 Similar changes are not required for the SSL 2.0 implementation
4908 because the number of padding bytes is sent in clear for SSL 2.0,
4909 and the extra bytes are just ignored. However ssl/s2_pkt.c
4910 failed to verify that the purported number of padding bytes is in
4914 *) Add OpenUNIX-8 support including shared libraries
4915 (Boyd Lynn Gerber <gerberb@zenez.com>).
4918 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4919 'wristwatch attack' using huge encoding parameters (cf.
4920 James H. Manger's CRYPTO 2001 paper). Note that the
4921 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4922 encoding parameters and hence was not vulnerable.
4925 *) BN_sqr() bug fix.
4926 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4928 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4929 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4930 followed by modular reduction.
4931 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4933 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4934 equivalent based on BN_pseudo_rand() instead of BN_rand().
4937 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4938 This function was broken, as the check for a new client hello message
4939 to handle SGC did not allow these large messages.
4940 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4943 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4946 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4947 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4950 *) Rework the configuration and shared library support for Tru64 Unix.
4951 The configuration part makes use of modern compiler features and
4952 still retains old compiler behavior for those that run older versions
4953 of the OS. The shared library support part includes a variant that
4954 uses the RPATH feature, and is available through the special
4955 configuration target "alpha-cc-rpath", which will never be selected
4957 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4959 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4960 with the same message size as in ssl3_get_certificate_request().
4961 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4962 messages might inadvertently be reject as too long.
4963 [Petr Lampa <lampa@fee.vutbr.cz>]
4965 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4968 *) Modified SSL library such that the verify_callback that has been set
4969 specificly for an SSL object with SSL_set_verify() is actually being
4970 used. Before the change, a verify_callback set with this function was
4971 ignored and the verify_callback() set in the SSL_CTX at the time of
4972 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4973 to allow the necessary settings.
4976 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4977 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4978 done automatically (in contradiction to the requirements of the C
4979 standard). This made problems when used from OpenSSH.
4982 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4983 dh->length and always used
4985 BN_rand_range(priv_key, dh->p).
4987 BN_rand_range() is not necessary for Diffie-Hellman, and this
4988 specific range makes Diffie-Hellman unnecessarily inefficient if
4989 dh->length (recommended exponent length) is much smaller than the
4990 length of dh->p. We could use BN_rand_range() if the order of
4991 the subgroup was stored in the DH structure, but we only have
4996 BN_rand(priv_key, l, ...)
4998 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
5004 RSA_eay_public_encrypt
5005 RSA_eay_private_decrypt
5006 RSA_eay_private_encrypt (signing)
5007 RSA_eay_public_decrypt (signature verification)
5009 (default implementations for RSA_public_encrypt,
5010 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
5011 always reject numbers >= n.
5014 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
5015 to synchronize access to 'locking_thread'. This is necessary on
5016 systems where access to 'locking_thread' (an 'unsigned long'
5017 variable) is not atomic.
5020 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
5021 *before* setting the 'crypto_lock_rand' flag. The previous code had
5022 a race condition if 0 is a valid thread ID.
5023 [Travis Vitek <vitek@roguewave.com>]
5025 *) Add support for shared libraries under Irix.
5026 [Albert Chin-A-Young <china@thewrittenword.com>]
5028 *) Add configuration option to build on Linux on both big-endian and
5030 [Ralf Baechle <ralf@uni-koblenz.de>]
5032 *) Add the possibility to create shared libraries on HP-UX.
5035 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
5037 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
5038 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
5039 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
5040 PRNG state recovery was possible based on the output of
5041 one PRNG request appropriately sized to gain knowledge on
5042 'md' followed by enough consecutive 1-byte PRNG requests
5043 to traverse all of 'state'.
5045 1. When updating 'md_local' (the current thread's copy of 'md')
5046 during PRNG output generation, hash all of the previous
5047 'md_local' value, not just the half used for PRNG output.
5049 2. Make the number of bytes from 'state' included into the hash
5050 independent from the number of PRNG bytes requested.
5052 The first measure alone would be sufficient to avoid
5053 Markku-Juhani's attack. (Actually it had never occurred
5054 to me that the half of 'md_local' used for chaining was the
5055 half from which PRNG output bytes were taken -- I had always
5056 assumed that the secret half would be used.) The second
5057 measure makes sure that additional data from 'state' is never
5058 mixed into 'md_local' in small portions; this heuristically
5059 further strengthens the PRNG.
5062 *) Fix crypto/bn/asm/mips3.s.
5065 *) When only the key is given to "enc", the IV is undefined. Print out
5066 an error message in this case.
5069 *) Handle special case when X509_NAME is empty in X509 printing routines.
5072 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
5073 positive and less than q.
5076 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
5077 used: it isn't thread safe and the add_lock_callback should handle
5079 [Paul Rose <Paul.Rose@bridge.com>]
5081 *) Verify that incoming data obeys the block size in
5082 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
5086 [Ulf Möller, Bodo Möller]
5088 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
5089 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
5090 when fixing the server behaviour for backwards-compatible 'client
5091 hello' messages. (Note that the attack is impractical against
5092 SSL 3.0 and TLS 1.0 anyway because length and version checking
5093 means that the probability of guessing a valid ciphertext is
5094 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
5097 Before 0.9.5, the countermeasure (hide the error by generating a
5098 random 'decryption result') did not work properly because
5099 ERR_clear_error() was missing, meaning that SSL_get_error() would
5100 detect the supposedly ignored error.
5102 Both problems are now fixed.
5105 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
5106 (previously it was 1024).
5109 *) Fix for compatibility mode trust settings: ignore trust settings
5110 unless some valid trust or reject settings are present.
5113 *) Fix for blowfish EVP: its a variable length cipher.
5116 *) Fix various bugs related to DSA S/MIME verification. Handle missing
5117 parameters in DSA public key structures and return an error in the
5118 DSA routines if parameters are absent.
5121 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
5122 in the current directory if neither $RANDFILE nor $HOME was set.
5123 RAND_file_name() in 0.9.6a returned NULL in this case. This has
5124 caused some confusion to Windows users who haven't defined $HOME.
5125 Thus RAND_file_name() is changed again: e_os.h can define a
5126 DEFAULT_HOME, which will be used if $HOME is not set.
5127 For Windows, we use "C:"; on other platforms, we still require
5128 environment variables.
5130 *) Move 'if (!initialized) RAND_poll()' into regions protected by
5131 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
5132 having multiple threads call RAND_poll() concurrently.
5135 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
5136 combination of a flag and a thread ID variable.
5137 Otherwise while one thread is in ssleay_rand_bytes (which sets the
5138 flag), *other* threads can enter ssleay_add_bytes without obeying
5139 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
5140 that they do not hold after the first thread unsets add_do_not_lock).
5143 *) Change bctest again: '-x' expressions are not available in all
5147 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
5149 *) Fix a couple of memory leaks in PKCS7_dataDecode()
5150 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
5152 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
5153 the default extension for executables, if any. Also, make the perl
5154 scripts that use symlink() to test if it really exists and use "cp"
5155 if it doesn't. All this made OpenSSL compilable and installable in
5159 *) Fix for asn1_GetSequence() for indefinite length constructed data.
5160 If SEQUENCE is length is indefinite just set c->slen to the total
5161 amount of data available.
5162 [Steve Henson, reported by shige@FreeBSD.org]
5163 [This change does not apply to 0.9.7.]
5165 *) Change bctest to avoid here-documents inside command substitution
5166 (workaround for FreeBSD /bin/sh bug).
5167 For compatibility with Ultrix, avoid shell functions (introduced
5168 in the bctest version that searches along $PATH).
5171 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
5172 with des_encrypt() defined on some operating systems, like Solaris
5176 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
5177 On the Importance of Eliminating Errors in Cryptographic
5178 Computations, J. Cryptology 14 (2001) 2, 101-119,
5179 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
5182 *) MIPS assembler BIGNUM division bug fix.
5185 *) Disabled incorrect Alpha assembler code.
5188 *) Fix PKCS#7 decode routines so they correctly update the length
5189 after reading an EOC for the EXPLICIT tag.
5191 [This change does not apply to 0.9.7.]
5193 *) Fix bug in PKCS#12 key generation routines. This was triggered
5194 if a 3DES key was generated with a 0 initial byte. Include
5195 PKCS12_BROKEN_KEYGEN compilation option to retain the old
5196 (but broken) behaviour.
5199 *) Enhance bctest to search for a working bc along $PATH and print
5201 [Tim Rice <tim@multitalents.net> via Richard Levitte]
5203 *) Fix memory leaks in err.c: free err_data string if necessary;
5204 don't write to the wrong index in ERR_set_error_data.
5207 *) Implement ssl23_peek (analogous to ssl23_read), which previously
5211 *) Replace rdtsc with _emit statements for VC++ version 5.
5212 [Jeremy Cooper <jeremy@baymoo.org>]
5214 *) Make it possible to reuse SSLv2 sessions.
5217 *) In copy_email() check for >= 0 as a return value for
5218 X509_NAME_get_index_by_NID() since 0 is a valid index.
5219 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
5221 *) Avoid coredump with unsupported or invalid public keys by checking if
5222 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
5223 PKCS7_verify() fails with non detached data.
5226 *) Don't use getenv in library functions when run as setuid/setgid.
5227 New function OPENSSL_issetugid().
5230 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
5231 due to incorrect handling of multi-threading:
5233 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
5235 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
5237 3. Count how many times MemCheck_off() has been called so that
5238 nested use can be treated correctly. This also avoids
5239 inband-signalling in the previous code (which relied on the
5240 assumption that thread ID 0 is impossible).
5243 *) Add "-rand" option also to s_client and s_server.
5246 *) Fix CPU detection on Irix 6.x.
5247 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
5248 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5250 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5253 [This change does not apply to 0.9.7.]
5255 *) Use the cached encoding of an X509_NAME structure rather than
5256 copying it. This is apparently the reason for the libsafe "errors"
5257 but the code is actually correct.
5260 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5261 Bleichenbacher's DSA attack.
5262 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5263 to be set and top=0 forces the highest bit to be set; top=-1 is new
5264 and leaves the highest bit random.
5265 [Ulf Moeller, Bodo Moeller]
5267 *) In the NCONF_...-based implementations for CONF_... queries
5268 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5269 a temporary CONF structure with the data component set to NULL
5270 (which gives segmentation faults in lh_retrieve).
5271 Instead, use NULL for the CONF pointer in CONF_get_string and
5272 CONF_get_number (which may use environment variables) and directly
5273 return NULL from CONF_get_section.
5276 *) Fix potential buffer overrun for EBCDIC.
5279 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5280 keyUsage if basicConstraints absent for a CA.
5283 *) Make SMIME_write_PKCS7() write mail header values with a format that
5284 is more generally accepted (no spaces before the semicolon), since
5285 some programs can't parse those values properly otherwise. Also make
5286 sure BIO's that break lines after each write do not create invalid
5290 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5291 macros previously used would not encode an empty SEQUENCE OF
5292 and break the signature.
5294 [This change does not apply to 0.9.7.]
5296 *) Zero the premaster secret after deriving the master secret in
5300 *) Add some EVP_add_digest_alias registrations (as found in
5301 OpenSSL_add_all_digests()) to SSL_library_init()
5302 aka OpenSSL_add_ssl_algorithms(). This provides improved
5303 compatibility with peers using X.509 certificates
5304 with unconventional AlgorithmIdentifier OIDs.
5307 *) Fix for Irix with NO_ASM.
5308 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5310 *) ./config script fixes.
5311 [Ulf Moeller, Richard Levitte]
5313 *) Fix 'openssl passwd -1'.
5316 *) Change PKCS12_key_gen_asc() so it can cope with non null
5317 terminated strings whose length is passed in the passlen
5318 parameter, for example from PEM callbacks. This was done
5319 by adding an extra length parameter to asc2uni().
5320 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5322 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5323 call failed, free the DSA structure.
5326 *) Fix to uni2asc() to cope with zero length Unicode strings.
5327 These are present in some PKCS#12 files.
5330 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5331 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5332 when writing a 32767 byte record.
5333 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5335 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5336 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5338 (RSA objects have a reference count access to which is protected
5339 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5340 so they are meant to be shared between threads.)
5341 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5342 "Reddie, Steven" <Steven.Reddie@ca.com>]
5344 *) Fix a deadlock in CRYPTO_mem_leaks().
5347 *) Use better test patterns in bntest.
5350 *) rand_win.c fix for Borland C.
5353 *) BN_rshift bugfix for n == 0.
5356 *) Add a 'bctest' script that checks for some known 'bc' bugs
5357 so that 'make test' does not abort just because 'bc' is broken.
5360 *) Store verify_result within SSL_SESSION also for client side to
5361 avoid potential security hole. (Re-used sessions on the client side
5362 always resulted in verify_result==X509_V_OK, not using the original
5363 result of the server certificate verification.)
5366 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5367 SSL3_RT_APPLICATION_DATA, return 0.
5368 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5372 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5373 releases, have been re-implemented by renaming the previous
5374 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5375 and ssl3_read_internal, respectively, and adding 'peek' parameters
5376 to them. The new ssl[23]_{read,peek} functions are calls to
5377 ssl[23]_read_internal with the 'peek' flag set appropriately.
5378 A 'peek' parameter has also been added to ssl3_read_bytes, which
5379 does the actual work for ssl3_read_internal.
5382 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5383 the method-specific "init()" handler. Also clean up ex_data after
5384 calling the method-specific "finish()" handler. Previously, this was
5385 happening the other way round.
5388 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5389 The previous value, 12, was not always sufficient for BN_mod_exp().
5392 *) Make sure that shared libraries get the internal name engine with
5393 the full version number and not just 0. This should mark the
5394 shared libraries as not backward compatible. Of course, this should
5395 be changed again when we can guarantee backward binary compatibility.
5398 *) Fix typo in get_cert_by_subject() in by_dir.c
5399 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5401 *) Rework the system to generate shared libraries:
5403 - Make note of the expected extension for the shared libraries and
5404 if there is a need for symbolic links from for example libcrypto.so.0
5405 to libcrypto.so.0.9.7. There is extended info in Configure for
5408 - Make as few rebuilds of the shared libraries as possible.
5410 - Still avoid linking the OpenSSL programs with the shared libraries.
5412 - When installing, install the shared libraries separately from the
5416 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5418 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5419 and not in SSL_clear because the latter is also used by the
5420 accept/connect functions; previously, the settings made by
5421 SSL_set_read_ahead would be lost during the handshake.
5422 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5424 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5425 Previously, it would create entries for disableed algorithms no
5429 *) Added several new manual pages for SSL_* function.
5432 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5434 *) In ssl23_get_client_hello, generate an error message when faced
5435 with an initial SSL 3.0/TLS record that is too small to contain the
5436 first two bytes of the ClientHello message, i.e. client_version.
5437 (Note that this is a pathologic case that probably has never happened
5438 in real life.) The previous approach was to use the version number
5439 from the record header as a substitute; but our protocol choice
5440 should not depend on that one because it is not authenticated
5441 by the Finished messages.
5444 *) More robust randomness gathering functions for Windows.
5445 [Jeffrey Altman <jaltman@columbia.edu>]
5447 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5448 not set then we don't setup the error code for issuer check errors
5449 to avoid possibly overwriting other errors which the callback does
5450 handle. If an application does set the flag then we assume it knows
5451 what it is doing and can handle the new informational codes
5455 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5456 a general "ANY" type, as such it should be able to decode anything
5457 including tagged types. However it didn't check the class so it would
5458 wrongly interpret tagged types in the same way as their universal
5459 counterpart and unknown types were just rejected. Changed so that the
5460 tagged and unknown types are handled in the same way as a SEQUENCE:
5461 that is the encoding is stored intact. There is also a new type
5462 "V_ASN1_OTHER" which is used when the class is not universal, in this
5463 case we have no idea what the actual type is so we just lump them all
5467 *) On VMS, stdout may very well lead to a file that is written to
5468 in a record-oriented fashion. That means that every write() will
5469 write a separate record, which will be read separately by the
5470 programs trying to read from it. This can be very confusing.
5472 The solution is to put a BIO filter in the way that will buffer
5473 text until a linefeed is reached, and then write everything a
5474 line at a time, so every record written will be an actual line,
5475 not chunks of lines and not (usually doesn't happen, but I've
5476 seen it once) several lines in one record. BIO_f_linebuffer() is
5479 Currently, it's a VMS-only method, because that's where it has
5480 been tested well enough.
5483 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5484 it can return incorrect results.
5485 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5486 but it was in 0.9.6-beta[12].)
5489 *) Disable the check for content being present when verifying detached
5490 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5491 include zero length content when signing messages.
5494 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5495 BIO_ctrl (for BIO pairs).
5498 *) Add DSO method for VMS.
5501 *) Bug fix: Montgomery multiplication could produce results with the
5505 *) Add RPM specification openssl.spec and modify it to build three
5506 packages. The default package contains applications, application
5507 documentation and run-time libraries. The devel package contains
5508 include files, static libraries and function documentation. The
5509 doc package contains the contents of the doc directory. The original
5510 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5513 *) Add a large number of documentation files for many SSL routines.
5514 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5516 *) Add a configuration entry for Sony News 4.
5517 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5519 *) Don't set the two most significant bits to one when generating a
5520 random number < q in the DSA library.
5523 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5524 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5525 the underlying transport is blocking) if a handshake took place.
5526 (The default behaviour is needed by applications such as s_client
5527 and s_server that use select() to determine when to use SSL_read;
5528 but for applications that know in advance when to expect data, it
5529 just makes things more complicated.)
5532 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5536 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5537 work better on such systems.
5538 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5540 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5541 Update PKCS12_parse() so it copies the friendlyName and the
5542 keyid to the certificates aux info.
5545 *) Fix bug in PKCS7_verify() which caused an infinite loop
5546 if there was more than one signature.
5547 [Sven Uszpelkat <su@celocom.de>]
5549 *) Major change in util/mkdef.pl to include extra information
5550 about each symbol, as well as presentig variables as well
5551 as functions. This change means that there's n more need
5552 to rebuild the .num files when some algorithms are excluded.
5555 *) Allow the verify time to be set by an application,
5556 rather than always using the current time.
5559 *) Phase 2 verify code reorganisation. The certificate
5560 verify code now looks up an issuer certificate by a
5561 number of criteria: subject name, authority key id
5562 and key usage. It also verifies self signed certificates
5563 by the same criteria. The main comparison function is
5564 X509_check_issued() which performs these checks.
5566 Lot of changes were necessary in order to support this
5567 without completely rewriting the lookup code.
5569 Authority and subject key identifier are now cached.
5571 The LHASH 'certs' is X509_STORE has now been replaced
5572 by a STACK_OF(X509_OBJECT). This is mainly because an
5573 LHASH can't store or retrieve multiple objects with
5574 the same hash value.
5576 As a result various functions (which were all internal
5577 use only) have changed to handle the new X509_STORE
5578 structure. This will break anything that messed round
5579 with X509_STORE internally.
5581 The functions X509_STORE_add_cert() now checks for an
5582 exact match, rather than just subject name.
5584 The X509_STORE API doesn't directly support the retrieval
5585 of multiple certificates matching a given criteria, however
5586 this can be worked round by performing a lookup first
5587 (which will fill the cache with candidate certificates)
5588 and then examining the cache for matches. This is probably
5589 the best we can do without throwing out X509_LOOKUP
5590 entirely (maybe later...).
5592 The X509_VERIFY_CTX structure has been enhanced considerably.
5594 All certificate lookup operations now go via a get_issuer()
5595 callback. Although this currently uses an X509_STORE it
5596 can be replaced by custom lookups. This is a simple way
5597 to bypass the X509_STORE hackery necessary to make this
5598 work and makes it possible to use more efficient techniques
5599 in future. A very simple version which uses a simple
5600 STACK for its trusted certificate store is also provided
5601 using X509_STORE_CTX_trusted_stack().
5603 The verify_cb() and verify() callbacks now have equivalents
5604 in the X509_STORE_CTX structure.
5606 X509_STORE_CTX also has a 'flags' field which can be used
5607 to customise the verify behaviour.
5610 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5611 excludes S/MIME capabilities.
5614 *) When a certificate request is read in keep a copy of the
5615 original encoding of the signed data and use it when outputing
5616 again. Signatures then use the original encoding rather than
5617 a decoded, encoded version which may cause problems if the
5618 request is improperly encoded.
5621 *) For consistency with other BIO_puts implementations, call
5622 buffer_write(b, ...) directly in buffer_puts instead of calling
5625 In BIO_puts, increment b->num_write as in BIO_write.
5626 [Peter.Sylvester@EdelWeb.fr]
5628 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5629 BN_zero, we may not return a BIGNUM with an array consisting of
5633 *) Avoid calling abort() from within the library when problems are
5634 detected, except if preprocessor symbols have been defined
5635 (such as REF_CHECK, BN_DEBUG etc.).
5638 *) New openssl application 'rsautl'. This utility can be
5639 used for low level RSA operations. DER public key
5640 BIO/fp routines also added.
5643 *) New Configure entry and patches for compiling on QNX 4.
5644 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5646 *) A demo state-machine implementation was sponsored by
5647 Nuron (http://www.nuron.com/) and is now available in
5648 demos/state_machine.
5651 *) New options added to the 'dgst' utility for signature
5652 generation and verification.
5655 *) Unrecognized PKCS#7 content types are now handled via a
5656 catch all ASN1_TYPE structure. This allows unsupported
5657 types to be stored as a "blob" and an application can
5658 encode and decode it manually.
5661 *) Fix various signed/unsigned issues to make a_strex.c
5663 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5665 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5666 length if passed a buffer. ASN1_INTEGER_to_BN failed
5667 if passed a NULL BN and its argument was negative.
5668 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5670 *) Modification to PKCS#7 encoding routines to output definite
5671 length encoding. Since currently the whole structures are in
5672 memory there's not real point in using indefinite length
5673 constructed encoding. However if OpenSSL is compiled with
5674 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5677 *) Added BIO_vprintf() and BIO_vsnprintf().
5680 *) Added more prefixes to parse for in the the strings written
5681 through a logging bio, to cover all the levels that are available
5682 through syslog. The prefixes are now:
5684 PANIC, EMERG, EMR => LOG_EMERG
5685 ALERT, ALR => LOG_ALERT
5686 CRIT, CRI => LOG_CRIT
5687 ERROR, ERR => LOG_ERR
5688 WARNING, WARN, WAR => LOG_WARNING
5689 NOTICE, NOTE, NOT => LOG_NOTICE
5690 INFO, INF => LOG_INFO
5691 DEBUG, DBG => LOG_DEBUG
5693 and as before, if none of those prefixes are present at the
5694 beginning of the string, LOG_ERR is chosen.
5696 On Win32, the LOG_* levels are mapped according to this:
5698 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5699 LOG_WARNING => EVENTLOG_WARNING_TYPE
5700 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5704 *) Made it possible to reconfigure with just the configuration
5705 argument "reconf" or "reconfigure". The command line arguments
5706 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5707 and are retrieved from there when reconfiguring.
5711 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5713 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5716 *) The obj_dat.pl script was messing up the sorting of object
5717 names. The reason was that it compared the quoted version
5718 of strings as a result "OCSP" > "OCSP Signing" because
5719 " > SPACE. Changed script to store unquoted versions of
5720 names and add quotes on output. It was also omitting some
5721 names from the lookup table if they were given a default
5722 value (that is if SN is missing it is given the same
5723 value as LN and vice versa), these are now added on the
5724 grounds that if an object has a name we should be able to
5725 look it up. Finally added warning output when duplicate
5726 short or long names are found.
5729 *) Changes needed for Tandem NSK.
5730 [Scott Uroff <scott@xypro.com>]
5732 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5733 RSA_padding_check_SSLv23(), special padding was never detected
5734 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5735 version rollback attacks was not effective.
5737 In s23_clnt.c, don't use special rollback-attack detection padding
5738 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5739 client; similarly, in s23_srvr.c, don't do the rollback check if
5740 SSL 2.0 is the only protocol enabled in the server.
5743 *) Make it possible to get hexdumps of unprintable data with 'openssl
5744 asn1parse'. By implication, the functions ASN1_parse_dump() and
5745 BIO_dump_indent() are added.
5748 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5749 these print out strings and name structures based on various
5750 flags including RFC2253 support and proper handling of
5751 multibyte characters. Added options to the 'x509' utility
5752 to allow the various flags to be set.
5755 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5756 Also change the functions X509_cmp_current_time() and
5757 X509_gmtime_adj() work with an ASN1_TIME structure,
5758 this will enable certificates using GeneralizedTime in validity
5759 dates to be checked.
5762 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5763 negative public key encodings) on by default,
5764 NO_NEG_PUBKEY_BUG can be set to disable it.
5767 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5768 content octets. An i2c_ASN1_OBJECT is unnecessary because
5769 the encoding can be trivially obtained from the structure.
5772 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5773 not read locks (CRYPTO_r_[un]lock).
5776 *) A first attempt at creating official support for shared
5777 libraries through configuration. I've kept it so the
5778 default is static libraries only, and the OpenSSL programs
5779 are always statically linked for now, but there are
5780 preparations for dynamic linking in place.
5781 This has been tested on Linux and Tru64.
5784 *) Randomness polling function for Win9x, as described in:
5785 Peter Gutmann, Software Generation of Practically Strong
5789 *) Fix so PRNG is seeded in req if using an already existing
5793 *) New options to smime application. -inform and -outform
5794 allow alternative formats for the S/MIME message including
5795 PEM and DER. The -content option allows the content to be
5796 specified separately. This should allow things like Netscape
5797 form signing output easier to verify.
5800 *) Fix the ASN1 encoding of tags using the 'long form'.
5803 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5804 STRING types. These convert content octets to and from the
5805 underlying type. The actual tag and length octets are
5806 already assumed to have been read in and checked. These
5807 are needed because all other string types have virtually
5808 identical handling apart from the tag. By having versions
5809 of the ASN1 functions that just operate on content octets
5810 IMPLICIT tagging can be handled properly. It also allows
5811 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5812 and ASN1_INTEGER are identical apart from the tag.
5815 *) Change the handling of OID objects as follows:
5817 - New object identifiers are inserted in objects.txt, following
5818 the syntax given in objects.README.
5819 - objects.pl is used to process obj_mac.num and create a new
5821 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5824 This is currently kind of a hack, and the perl code in objects.pl
5825 isn't very elegant, but it works as I intended. The simplest way
5826 to check that it worked correctly is to look in obj_dat.h and
5827 check the array nid_objs and make sure the objects haven't moved
5828 around (this is important!). Additions are OK, as well as
5829 consistent name changes.
5832 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5835 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5836 The given file adds to whatever has already been seeded into the
5837 random pool through the RANDFILE configuration file option or
5838 environment variable, or the default random state file.
5841 *) mkstack.pl now sorts each macro group into lexical order.
5842 Previously the output order depended on the order the files
5843 appeared in the directory, resulting in needless rewriting
5847 *) Patches to make OpenSSL compile under Win32 again. Mostly
5848 work arounds for the VC++ problem that it treats func() as
5849 func(void). Also stripped out the parts of mkdef.pl that
5850 added extra typesafe functions: these no longer exist.
5853 *) Reorganisation of the stack code. The macros are now all
5854 collected in safestack.h . Each macro is defined in terms of
5855 a "stack macro" of the form SKM_<name>(type, a, b). The
5856 DEBUG_SAFESTACK is now handled in terms of function casts,
5857 this has the advantage of retaining type safety without the
5858 use of additional functions. If DEBUG_SAFESTACK is not defined
5859 then the non typesafe macros are used instead. Also modified the
5860 mkstack.pl script to handle the new form. Needs testing to see
5861 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5862 the default if no major problems. Similar behaviour for ASN1_SET_OF
5863 and PKCS12_STACK_OF.
5866 *) When some versions of IIS use the 'NET' form of private key the
5867 key derivation algorithm is different. Normally MD5(password) is
5868 used as a 128 bit RC4 key. In the modified case
5869 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5870 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5871 as the old Netscape_RSA functions except they have an additional
5872 'sgckey' parameter which uses the modified algorithm. Also added
5873 an -sgckey command line option to the rsa utility. Thanks to
5874 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5875 algorithm to openssl-dev.
5878 *) The evp_local.h macros were using 'c.##kname' which resulted in
5879 invalid expansion on some systems (SCO 5.0.5 for example).
5880 Corrected to 'c.kname'.
5881 [Phillip Porch <root@theporch.com>]
5883 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5884 a STACK of email addresses from a certificate or request, these look
5885 in the subject name and the subject alternative name extensions and
5886 omit any duplicate addresses.
5889 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5890 This makes DSA verification about 2 % faster.
5893 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5894 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5895 plus overhead for 1024 bit moduli).
5896 This makes exponentiations about 0.5 % faster for 1024 bit
5897 exponents (as measured by "openssl speed rsa2048").
5900 *) Rename memory handling macros to avoid conflicts with other
5902 Malloc => OPENSSL_malloc
5903 Malloc_locked => OPENSSL_malloc_locked
5904 Realloc => OPENSSL_realloc
5905 Free => OPENSSL_free
5908 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5909 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5912 *) CygWin32 support.
5913 [John Jarvie <jjarvie@newsguy.com>]
5915 *) The type-safe stack code has been rejigged. It is now only compiled
5916 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5917 by default all type-specific stack functions are "#define"d back to
5918 standard stack functions. This results in more streamlined output
5919 but retains the type-safety checking possibilities of the original
5923 *) The STACK code has been cleaned up, and certain type declarations
5924 that didn't make a lot of sense have been brought in line. This has
5925 also involved a cleanup of sorts in safestack.h to more correctly
5926 map type-safe stack functions onto their plain stack counterparts.
5927 This work has also resulted in a variety of "const"ifications of
5928 lots of the code, especially "_cmp" operations which should normally
5929 be prototyped with "const" parameters anyway.
5932 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5933 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5934 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5935 where all of 'md' is used each time the PRNG is used, but 'state'
5936 is used only indexed by a cyclic counter. As entropy may not be
5937 well distributed from the beginning, 'md' is important as a
5938 chaining variable. However, the output function chains only half
5939 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5940 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5941 in all of 'state' being rewritten, with the new values depending
5942 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5945 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5946 the handshake is continued after ssl_verify_cert_chain();
5947 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5948 can lead to 'unexplainable' connection aborts later.
5949 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5951 *) Major EVP API cipher revision.
5952 Add hooks for extra EVP features. This allows various cipher
5953 parameters to be set in the EVP interface. Support added for variable
5954 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5955 setting of RC2 and RC5 parameters.
5957 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5960 Remove lots of duplicated code from the EVP library. For example *every*
5961 cipher init() function handles the 'iv' in the same way according to the
5962 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5963 for CFB and OFB modes they zero ctx->num.
5965 New functionality allows removal of S/MIME code RC2 hack.
5967 Most of the routines have the same form and so can be declared in terms
5970 By shifting this to the top level EVP_CipherInit() it can be removed from
5971 all individual ciphers. If the cipher wants to handle IVs or keys
5972 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5975 Change lots of functions like EVP_EncryptUpdate() to now return a
5976 value: although software versions of the algorithms cannot fail
5977 any installed hardware versions can.
5980 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5981 this option is set, tolerate broken clients that send the negotiated
5982 protocol version number instead of the requested protocol version
5986 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5987 i.e. non-zero for export ciphersuites, zero otherwise.
5988 Previous versions had this flag inverted, inconsistent with
5989 rsa_tmp_cb (..._TMP_RSA_CB).
5990 [Bodo Moeller; problem reported by Amit Chopra]
5992 *) Add missing DSA library text string. Work around for some IIS
5993 key files with invalid SEQUENCE encoding.
5996 *) Add a document (doc/standards.txt) that list all kinds of standards
5997 and so on that are implemented in OpenSSL.
6000 *) Enhance c_rehash script. Old version would mishandle certificates
6001 with the same subject name hash and wouldn't handle CRLs at all.
6002 Added -fingerprint option to crl utility, to support new c_rehash
6006 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
6009 *) Fix for SSL server purpose checking. Server checking was
6010 rejecting certificates which had extended key usage present
6011 but no ssl client purpose.
6012 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
6014 *) Make PKCS#12 code work with no password. The PKCS#12 spec
6015 is a little unclear about how a blank password is handled.
6016 Since the password in encoded as a BMPString with terminating
6017 double NULL a zero length password would end up as just the
6018 double NULL. However no password at all is different and is
6019 handled differently in the PKCS#12 key generation code. NS
6020 treats a blank password as zero length. MSIE treats it as no
6021 password on export: but it will try both on import. We now do
6022 the same: PKCS12_parse() tries zero length and no password if
6023 the password is set to "" or NULL (NULL is now a valid password:
6024 it wasn't before) as does the pkcs12 application.
6027 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
6028 perror when PEM_read_bio_X509_REQ fails, the error message must
6029 be obtained from the error queue.
6032 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
6033 it in ERR_remove_state if appropriate, and change ERR_get_state
6034 accordingly to avoid race conditions (this is necessary because
6035 thread_hash is no longer constant once set).
6038 *) Bugfix for linux-elf makefile.one.
6041 *) RSA_get_default_method() will now cause a default
6042 RSA_METHOD to be chosen if one doesn't exist already.
6043 Previously this was only set during a call to RSA_new()
6044 or RSA_new_method(NULL) meaning it was possible for
6045 RSA_get_default_method() to return NULL.
6048 *) Added native name translation to the existing DSO code
6049 that will convert (if the flag to do so is set) filenames
6050 that are sufficiently small and have no path information
6051 into a canonical native form. Eg. "blah" converted to
6052 "libblah.so" or "blah.dll" etc.
6055 *) New function ERR_error_string_n(e, buf, len) which is like
6056 ERR_error_string(e, buf), but writes at most 'len' bytes
6057 including the 0 terminator. For ERR_error_string_n, 'buf'
6059 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
6061 *) CONF library reworked to become more general. A new CONF
6062 configuration file reader "class" is implemented as well as a
6063 new functions (NCONF_*, for "New CONF") to handle it. The now
6064 old CONF_* functions are still there, but are reimplemented to
6065 work in terms of the new functions. Also, a set of functions
6066 to handle the internal storage of the configuration data is
6067 provided to make it easier to write new configuration file
6068 reader "classes" (I can definitely see something reading a
6069 configuration file in XML format, for example), called _CONF_*,
6070 or "the configuration storage API"...
6072 The new configuration file reading functions are:
6074 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
6075 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
6077 NCONF_default, NCONF_WIN32
6079 NCONF_dump_fp, NCONF_dump_bio
6081 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
6082 NCONF_new creates a new CONF object. This works in the same way
6083 as other interfaces in OpenSSL, like the BIO interface.
6084 NCONF_dump_* dump the internal storage of the configuration file,
6085 which is useful for debugging. All other functions take the same
6086 arguments as the old CONF_* functions wth the exception of the
6087 first that must be a `CONF *' instead of a `LHASH *'.
6089 To make it easer to use the new classes with the old CONF_* functions,
6090 the function CONF_set_default_method is provided.
6093 *) Add '-tls1' option to 'openssl ciphers', which was already
6094 mentioned in the documentation but had not been implemented.
6095 (This option is not yet really useful because even the additional
6096 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
6099 *) Initial DSO code added into libcrypto for letting OpenSSL (and
6100 OpenSSL-based applications) load shared libraries and bind to
6101 them in a portable way.
6102 [Geoff Thorpe, with contributions from Richard Levitte]
6104 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
6106 *) Make sure _lrotl and _lrotr are only used with MSVC.
6108 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
6109 (the default implementation of RAND_status).
6111 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
6112 to '-clrext' (= clear extensions), as intended and documented.
6113 [Bodo Moeller; inconsistency pointed out by Michael Attili
6116 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
6117 was larger than the MD block size.
6118 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
6120 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
6121 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
6122 using the passed key: if the passed key was a private key the result
6123 of X509_print(), for example, would be to print out all the private key
6127 *) des_quad_cksum() byte order bug fix.
6128 [Ulf Möller, using the problem description in krb4-0.9.7, where
6129 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
6131 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
6133 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
6135 *) For easily testing in shell scripts whether some command
6136 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
6137 returns with exit code 0 iff no command of the given name is available.
6138 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
6139 the output goes to stdout and nothing is printed to stderr.
6140 Additional arguments are always ignored.
6142 Since for each cipher there is a command of the same name,
6143 the 'no-cipher' compilation switches can be tested this way.
6145 ('openssl no-XXX' is not able to detect pseudo-commands such
6146 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
6149 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
6152 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
6153 is set; it will be thrown away anyway because each handshake creates
6155 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
6156 to parameters -- in previous versions (since OpenSSL 0.9.3) the
6157 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
6158 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
6161 *) New s_client option -ign_eof: EOF at stdin is ignored, and
6162 'Q' and 'R' lose their special meanings (quit/renegotiate).
6163 This is part of what -quiet does; unlike -quiet, -ign_eof
6164 does not suppress any output.
6167 *) Add compatibility options to the purpose and trust code. The
6168 purpose X509_PURPOSE_ANY is "any purpose" which automatically
6169 accepts a certificate or CA, this was the previous behaviour,
6170 with all the associated security issues.
6172 X509_TRUST_COMPAT is the old trust behaviour: only and
6173 automatically trust self signed roots in certificate store. A
6174 new trust setting X509_TRUST_DEFAULT is used to specify that
6175 a purpose has no associated trust setting and it should instead
6176 use the value in the default purpose.
6179 *) Fix the PKCS#8 DSA private key code so it decodes keys again
6180 and fix a memory leak.
6183 *) In util/mkerr.pl (which implements 'make errors'), preserve
6184 reason strings from the previous version of the .c file, as
6185 the default to have only downcase letters (and digits) in
6186 automatically generated reasons codes is not always appropriate.
6189 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
6190 using strerror. Previously, ERR_reason_error_string() returned
6191 library names as reason strings for SYSerr; but SYSerr is a special
6192 case where small numbers are errno values, not library numbers.
6195 *) Add '-dsaparam' option to 'openssl dhparam' application. This
6196 converts DSA parameters into DH parameters. (When creating parameters,
6197 DSA_generate_parameters is used.)
6200 *) Include 'length' (recommended exponent length) in C code generated
6201 by 'openssl dhparam -C'.
6204 *) The second argument to set_label in perlasm was already being used
6205 so couldn't be used as a "file scope" flag. Moved to third argument
6209 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
6210 instead of RAND_bytes for encryption IVs and salts.
6213 *) Include RAND_status() into RAND_METHOD instead of implementing
6214 it only for md_rand.c Otherwise replacing the PRNG by calling
6215 RAND_set_rand_method would be impossible.
6218 *) Don't let DSA_generate_key() enter an infinite loop if the random
6219 number generation fails.
6222 *) New 'rand' application for creating pseudo-random output.
6225 *) Added configuration support for Linux/IA64
6226 [Rolf Haberrecker <rolf@suse.de>]
6228 *) Assembler module support for Mingw32.
6231 *) Shared library support for HPUX (in shlib/).
6232 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
6234 *) Shared library support for Solaris gcc.
6235 [Lutz Behnke <behnke@trustcenter.de>]
6237 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
6239 *) PKCS7_encrypt() was adding text MIME headers twice because they
6240 were added manually and by SMIME_crlf_copy().
6243 *) In bntest.c don't call BN_rand with zero bits argument.
6244 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
6246 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
6247 case was implemented. This caused BN_div_recp() to fail occasionally.
6250 *) Add an optional second argument to the set_label() in the perl
6251 assembly language builder. If this argument exists and is set
6252 to 1 it signals that the assembler should use a symbol whose
6253 scope is the entire file, not just the current function. This
6254 is needed with MASM which uses the format label:: for this scope.
6255 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6257 *) Change the ASN1 types so they are typedefs by default. Before
6258 almost all types were #define'd to ASN1_STRING which was causing
6259 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6263 *) Change names of new functions to the new get1/get0 naming
6264 convention: After 'get1', the caller owns a reference count
6265 and has to call ..._free; 'get0' returns a pointer to some
6266 data structure without incrementing reference counters.
6267 (Some of the existing 'get' functions increment a reference
6268 counter, some don't.)
6269 Similarly, 'set1' and 'add1' functions increase reference
6270 counters or duplicate objects.
6273 *) Allow for the possibility of temp RSA key generation failure:
6274 the code used to assume it always worked and crashed on failure.
6277 *) Fix potential buffer overrun problem in BIO_printf().
6278 [Ulf Möller, using public domain code by Patrick Powell; problem
6279 pointed out by David Sacerdote <das33@cornell.edu>]
6281 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6282 RAND_egd() and RAND_status(). In the command line application,
6283 the EGD socket can be specified like a seed file using RANDFILE
6287 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6288 Some CAs (e.g. Verisign) distribute certificates in this form.
6291 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6292 list to exclude them. This means that no special compilation option
6293 is needed to use anonymous DH: it just needs to be included in the
6297 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6298 EVP_MD_type. The old functionality is available in a new macro called
6299 EVP_MD_md(). Change code that uses it and update docs.
6302 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6303 where the 'void *' argument is replaced by a function pointer argument.
6304 Previously 'void *' was abused to point to functions, which works on
6305 many platforms, but is not correct. As these functions are usually
6306 called by macros defined in OpenSSL header files, most source code
6307 should work without changes.
6310 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6311 sections with information on -D... compiler switches used for
6312 compiling the library so that applications can see them. To enable
6313 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6314 must be defined. E.g.,
6315 #define OPENSSL_ALGORITHM_DEFINES
6316 #include <openssl/opensslconf.h>
6317 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6318 [Richard Levitte, Ulf and Bodo Möller]
6320 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6324 *) Change the 'other' type in certificate aux info to a STACK_OF
6325 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6326 the required ASN1 format: arbitrary types determined by an OID.
6329 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6330 argument to 'req'. This is not because the function is newer or
6331 better than others it just uses the work 'NEW' in the certificate
6332 request header lines. Some software needs this.
6335 *) Reorganise password command line arguments: now passwords can be
6336 obtained from various sources. Delete the PEM_cb function and make
6337 it the default behaviour: i.e. if the callback is NULL and the
6338 usrdata argument is not NULL interpret it as a null terminated pass
6339 phrase. If usrdata and the callback are NULL then the pass phrase
6340 is prompted for as usual.
6343 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6344 the support is automatically enabled. The resulting binaries will
6345 autodetect the card and use it if present.
6346 [Ben Laurie and Compaq Inc.]
6348 *) Work around for Netscape hang bug. This sends certificate request
6349 and server done in one record. Since this is perfectly legal in the
6350 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6351 the bugs/SSLv3 entry for more info.
6354 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6357 *) Add -rand argument to smime and pkcs12 applications and read/write
6361 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6364 *) Add command line password options to the remaining applications.
6367 *) Bug fix for BN_div_recp() for numerators with an even number of
6371 *) More tests in bntest.c, and changed test_bn output.
6374 *) ./config recognizes MacOS X now.
6377 *) Bug fix for BN_div() when the first words of num and divsor are
6378 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6381 *) Add support for various broken PKCS#8 formats, and command line
6382 options to produce them.
6385 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6386 get temporary BIGNUMs from a BN_CTX.
6389 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6393 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6394 include a #define from the old name to the new. The original intent
6395 was that statically linked binaries could for example just call
6396 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6397 link with digests. This never worked becayse SSLeay_add_all_digests()
6398 and SSLeay_add_all_ciphers() were in the same source file so calling
6399 one would link with the other. They are now in separate source files.
6402 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6405 *) Use a less unusual form of the Miller-Rabin primality test (it used
6406 a binary algorithm for exponentiation integrated into the Miller-Rabin
6407 loop, our standard modexp algorithms are faster).
6410 *) Support for the EBCDIC character set completed.
6411 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6413 *) Source code cleanups: use const where appropriate, eliminate casts,
6414 use void * instead of char * in lhash.
6417 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6418 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6419 this the server could overwrite ephemeral keys that the client
6423 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6424 using 50 iterations of the Rabin-Miller test.
6426 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6427 iterations of the Rabin-Miller test as required by the appendix
6428 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6429 As BN_is_prime_fasttest includes trial division, DSA parameter
6430 generation becomes much faster.
6432 This implies a change for the callback functions in DSA_is_prime
6433 and DSA_generate_parameters: The callback function is called once
6434 for each positive witness in the Rabin-Miller test, not just
6435 occasionally in the inner loop; and the parameters to the
6436 callback function now provide an iteration count for the outer
6437 loop rather than for the current invocation of the inner loop.
6438 DSA_generate_parameters additionally can call the callback
6439 function with an 'iteration count' of -1, meaning that a
6440 candidate has passed the trial division test (when q is generated
6441 from an application-provided seed, trial division is skipped).
6444 *) New function BN_is_prime_fasttest that optionally does trial
6445 division before starting the Rabin-Miller test and has
6446 an additional BN_CTX * argument (whereas BN_is_prime always
6447 has to allocate at least one BN_CTX).
6448 'callback(1, -1, cb_arg)' is called when a number has passed the
6449 trial division stage.
6452 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6456 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6459 *) New function BN_pseudo_rand().
6462 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6463 bignum version of BN_from_montgomery() with the working code from
6464 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6468 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6469 made it impossible to use the same SSL_SESSION data structure in
6470 SSL2 clients in multiple threads.
6473 *) The return value of RAND_load_file() no longer counts bytes obtained
6474 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6475 to seed the PRNG (previously an explicit byte count was required).
6476 [Ulf Möller, Bodo Möller]
6478 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6479 used (char *) instead of (void *) and had casts all over the place.
6482 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6485 *) Retain source code compatibility for BN_prime_checks macro:
6486 BN_is_prime(..., BN_prime_checks, ...) now uses
6487 BN_prime_checks_for_size to determine the appropriate number of
6488 Rabin-Miller iterations.
6491 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6492 DH_CHECK_P_NOT_SAFE_PRIME.
6493 (Check if this is true? OpenPGP calls them "strong".)
6496 *) Merge the functionality of "dh" and "gendh" programs into a new program
6497 "dhparam". The old programs are retained for now but will handle DH keys
6498 (instead of parameters) in future.
6501 *) Make the ciphers, s_server and s_client programs check the return values
6502 when a new cipher list is set.
6505 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6506 ciphers. Before when the 56bit ciphers were enabled the sorting was
6509 The syntax for the cipher sorting has been extended to support sorting by
6510 cipher-strength (using the strength_bits hard coded in the tables).
6511 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6513 Fix a bug in the cipher-command parser: when supplying a cipher command
6514 string with an "undefined" symbol (neither command nor alphanumeric
6515 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6516 an error is flagged.
6518 Due to the strength-sorting extension, the code of the
6519 ssl_create_cipher_list() function was completely rearranged. I hope that
6520 the readability was also increased :-)
6521 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6523 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6524 for the first serial number and places 2 in the serial number file. This
6525 avoids problems when the root CA is created with serial number zero and
6526 the first user certificate has the same issuer name and serial number
6530 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6531 the new code. Add documentation for this stuff.
6534 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6535 X509_*() to X509at_*() on the grounds that they don't handle X509
6536 structures and behave in an analagous way to the X509v3 functions:
6537 they shouldn't be called directly but wrapper functions should be used
6540 So we also now have some wrapper functions that call the X509at functions
6541 when passed certificate requests. (TO DO: similar things can be done with
6542 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6543 things. Some of these need some d2i or i2d and print functionality
6544 because they handle more complex structures.)
6547 *) Add missing #ifndefs that caused missing symbols when building libssl
6548 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6549 NO_RSA in ssl/s2*.c.
6550 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6552 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6553 has a return value which indicates the quality of the random data
6554 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6555 error queue. New function RAND_pseudo_bytes() generates output that is
6556 guaranteed to be unique but not unpredictable. RAND_add is like
6557 RAND_seed, but takes an extra argument for an entropy estimate
6558 (RAND_seed always assumes full entropy).
6561 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6562 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6563 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6564 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6565 false-positive rate of at most 2^-80 for random input.
6568 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6571 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6572 in the 0.9.5 release), this returns the chain
6573 from an X509_CTX structure with a dup of the stack and all
6574 the X509 reference counts upped: so the stack will exist
6575 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6578 Also make SSL_SESSION_print() print out the verify return
6582 *) Add manpage for the pkcs12 command. Also change the default
6583 behaviour so MAC iteration counts are used unless the new
6584 -nomaciter option is used. This improves file security and
6585 only older versions of MSIE (4.0 for example) need it.
6588 *) Honor the no-xxx Configure options when creating .DEF files.
6591 *) Add PKCS#10 attributes to field table: challengePassword,
6592 unstructuredName and unstructuredAddress. These are taken from
6593 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6594 international characters are used.
6596 More changes to X509_ATTRIBUTE code: allow the setting of types
6597 based on strings. Remove the 'loc' parameter when adding
6598 attributes because these will be a SET OF encoding which is sorted
6602 *) Initial changes to the 'req' utility to allow request generation
6603 automation. This will allow an application to just generate a template
6604 file containing all the field values and have req construct the
6607 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6608 used all over the place including certificate requests and PKCS#7
6609 structures. They are currently handled manually where necessary with
6610 some primitive wrappers for PKCS#7. The new functions behave in a
6611 manner analogous to the X509 extension functions: they allow
6612 attributes to be looked up by NID and added.
6614 Later something similar to the X509V3 code would be desirable to
6615 automatically handle the encoding, decoding and printing of the
6616 more complex types. The string types like challengePassword can
6617 be handled by the string table functions.
6619 Also modified the multi byte string table handling. Now there is
6620 a 'global mask' which masks out certain types. The table itself
6621 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6622 is useful when for example there is only one permissible type
6623 (as in countryName) and using the mask might result in no valid
6627 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6628 SSL_get_peer_finished to allow applications to obtain the latest
6629 Finished messages sent to the peer or expected from the peer,
6630 respectively. (SSL_get_peer_finished is usually the Finished message
6631 actually received from the peer, otherwise the protocol will be aborted.)
6633 As the Finished message are message digests of the complete handshake
6634 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6635 be used for external authentication procedures when the authentication
6636 provided by SSL/TLS is not desired or is not enough.
6639 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6640 the host supports BWX extension and if Compaq C is present on the
6641 $PATH. Just exploiting of the BWX extension results in 20-30%
6642 performance kick for some algorithms, e.g. DES and RC4 to mention
6643 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6647 *) Add support for MS "fast SGC". This is arguably a violation of the
6648 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6649 weak crypto and after checking the certificate is SGC a second one
6650 with strong crypto. MS SGC stops the first handshake after receiving
6651 the server certificate message and sends a second client hello. Since
6652 a server will typically do all the time consuming operations before
6653 expecting any further messages from the client (server key exchange
6654 is the most expensive) there is little difference between the two.
6656 To get OpenSSL to support MS SGC we have to permit a second client
6657 hello message after we have sent server done. In addition we have to
6658 reset the MAC if we do get this second client hello.
6661 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6662 if a DER encoded private key is RSA or DSA traditional format. Changed
6663 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6664 format DER encoded private key. Newer code should use PKCS#8 format which
6665 has the key type encoded in the ASN1 structure. Added DER private key
6666 support to pkcs8 application.
6669 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6670 ciphersuites has been selected (as required by the SSL 3/TLS 1
6671 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6672 is set, we interpret this as a request to violate the specification
6673 (the worst that can happen is a handshake failure, and 'correct'
6674 behaviour would result in a handshake failure anyway).
6677 *) In SSL_CTX_add_session, take into account that there might be multiple
6678 SSL_SESSION structures with the same session ID (e.g. when two threads
6679 concurrently obtain them from an external cache).
6680 The internal cache can handle only one SSL_SESSION with a given ID,
6681 so if there's a conflict, we now throw out the old one to achieve
6685 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6686 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6687 some routines that use cipher OIDs: some ciphers do not have OIDs
6688 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6692 *) Simplify the trust setting structure and code. Now we just have
6693 two sequences of OIDs for trusted and rejected settings. These will
6694 typically have values the same as the extended key usage extension
6695 and any application specific purposes.
6697 The trust checking code now has a default behaviour: it will just
6698 check for an object with the same NID as the passed id. Functions can
6699 be provided to override either the default behaviour or the behaviour
6700 for a given id. SSL client, server and email already have functions
6701 in place for compatibility: they check the NID and also return "trusted"
6702 if the certificate is self signed.
6705 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6706 traditional format into an EVP_PKEY structure.
6709 *) Add a password callback function PEM_cb() which either prompts for
6710 a password if usr_data is NULL or otherwise assumes it is a null
6711 terminated password. Allow passwords to be passed on command line
6712 environment or config files in a few more utilities.
6715 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6716 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6717 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6718 Update documentation.
6721 *) Support for ASN1 "NULL" type. This could be handled before by using
6722 ASN1_TYPE but there wasn't any function that would try to read a NULL
6723 and produce an error if it couldn't. For compatibility we also have
6724 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6725 don't allocate anything because they don't need to.
6728 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6730 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6732 *) Rebuild of the memory allocation routines used by OpenSSL code and
6733 possibly others as well. The purpose is to make an interface that
6734 provide hooks so anyone can build a separate set of allocation and
6735 deallocation routines to be used by OpenSSL, for example memory
6736 pool implementations, or something else, which was previously hard
6737 since Malloc(), Realloc() and Free() were defined as macros having
6738 the values malloc, realloc and free, respectively (except for Win32
6739 compilations). The same is provided for memory debugging code.
6740 OpenSSL already comes with functionality to find memory leaks, but
6741 this gives people a chance to debug other memory problems.
6743 With these changes, a new set of functions and macros have appeared:
6745 CRYPTO_set_mem_debug_functions() [F]
6746 CRYPTO_get_mem_debug_functions() [F]
6747 CRYPTO_dbg_set_options() [F]
6748 CRYPTO_dbg_get_options() [F]
6749 CRYPTO_malloc_debug_init() [M]
6751 The memory debug functions are NULL by default, unless the library
6752 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6753 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6754 gives the standard debugging functions that come with OpenSSL) or
6755 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6756 provided by the library user) must be used. When the standard
6757 debugging functions are used, CRYPTO_dbg_set_options can be used to
6758 request additional information:
6759 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6760 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6762 Also, things like CRYPTO_set_mem_functions will always give the
6763 expected result (the new set of functions is used for allocation
6764 and deallocation) at all times, regardless of platform and compiler
6767 To finish it up, some functions that were never use in any other
6768 way than through macros have a new API and new semantic:
6771 CRYPTO_dbg_realloc()
6774 All macros of value have retained their old syntax.
6775 [Richard Levitte and Bodo Moeller]
6777 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6778 ordering of SMIMECapabilities wasn't in "strength order" and there
6779 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6783 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6784 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6785 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6787 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6788 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6789 functionality to handle multipart/signed properly) and a utility
6790 called 'smime' to call all this stuff. This is based on code I
6791 originally wrote for Celo who have kindly allowed it to be
6792 included in OpenSSL.
6795 *) Add variants des_set_key_checked and des_set_key_unchecked of
6796 des_set_key (aka des_key_sched). Global variable des_check_key
6797 decides which of these is called by des_set_key; this way
6798 des_check_key behaves as it always did, but applications and
6799 the library itself, which was buggy for des_check_key == 1,
6800 have a cleaner way to pick the version they need.
6803 *) New function PKCS12_newpass() which changes the password of a
6807 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6808 dynamic mix. In both cases the ids can be used as an index into the
6809 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6810 functions so they accept a list of the field values and the
6811 application doesn't need to directly manipulate the X509_TRUST
6815 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6819 *) Modify the way the V3 extension code looks up extensions. This now
6820 works in a similar way to the object code: we have some "standard"
6821 extensions in a static table which is searched with OBJ_bsearch()
6822 and the application can add dynamic ones if needed. The file
6823 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6824 updated whenever a new extension is added to the core code and kept
6825 in ext_nid order. There is a simple program 'tabtest.c' which checks
6826 this. New extensions are not added too often so this file can readily
6827 be maintained manually.
6829 There are two big advantages in doing things this way. The extensions
6830 can be looked up immediately and no longer need to be "added" using
6831 X509V3_add_standard_extensions(): this function now does nothing.
6832 [Side note: I get *lots* of email saying the extension code doesn't
6833 work because people forget to call this function]
6834 Also no dynamic allocation is done unless new extensions are added:
6835 so if we don't add custom extensions there is no need to call
6836 X509V3_EXT_cleanup().
6839 *) Modify enc utility's salting as follows: make salting the default. Add a
6840 magic header, so unsalted files fail gracefully instead of just decrypting
6841 to garbage. This is because not salting is a big security hole, so people
6842 should be discouraged from doing it.
6845 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6846 digest to be passed on the command line but it only used this
6847 parameter when signing a certificate. Modified so all relevant
6848 operations are affected by the digest parameter including the
6849 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6850 DSA key was used because it didn't fix the digest.
6853 *) Initial certificate chain verify code. Currently tests the untrusted
6854 certificates for consistency with the verify purpose (which is set
6855 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6857 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6858 this is because it will reject chains with invalid extensions whereas
6859 every previous version of OpenSSL and SSLeay made no checks at all.
6861 Trust code: checks the root CA for the relevant trust settings. Trust
6862 settings have an initial value consistent with the verify purpose: e.g.
6863 if the verify purpose is for SSL client use it expects the CA to be
6864 trusted for SSL client use. However the default value can be changed to
6865 permit custom trust settings: one example of this would be to only trust
6866 certificates from a specific "secure" set of CAs.
6868 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6869 which should be used for version portability: especially since the
6870 verify structure is likely to change more often now.
6872 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6873 to set them. If not set then assume SSL clients will verify SSL servers
6876 Two new options to the verify program: -untrusted allows a set of
6877 untrusted certificates to be passed in and -purpose which sets the
6878 intended purpose of the certificate. If a purpose is set then the
6879 new chain verify code is used to check extension consistency.
6882 *) Support for the authority information access extension.
6885 *) Modify RSA and DSA PEM read routines to transparently handle
6886 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6887 public keys in a format compatible with certificate
6888 SubjectPublicKeyInfo structures. Unfortunately there were already
6889 functions called *_PublicKey_* which used various odd formats so
6890 these are retained for compatibility: however the DSA variants were
6891 never in a public release so they have been deleted. Changed dsa/rsa
6892 utilities to handle the new format: note no releases ever handled public
6893 keys so we should be OK.
6895 The primary motivation for this change is to avoid the same fiasco
6896 that dogs private keys: there are several incompatible private key
6897 formats some of which are standard and some OpenSSL specific and
6898 require various evil hacks to allow partial transparent handling and
6899 even then it doesn't work with DER formats. Given the option anything
6900 other than PKCS#8 should be dumped: but the other formats have to
6901 stay in the name of compatibility.
6903 With public keys and the benefit of hindsight one standard format
6904 is used which works with EVP_PKEY, RSA or DSA structures: though
6905 it clearly returns an error if you try to read the wrong kind of key.
6907 Added a -pubkey option to the 'x509' utility to output the public key.
6908 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6909 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6910 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6911 that do the same as the EVP_PKEY_assign_*() except they up the
6912 reference count of the added key (they don't "swallow" the
6916 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6917 CRLs would fail if the file contained no certificates or no CRLs:
6918 added a new function to read in both types and return the number
6919 read: this means that if none are read it will be an error. The
6920 DER versions of the certificate and CRL reader would always fail
6921 because it isn't possible to mix certificates and CRLs in DER format
6922 without choking one or the other routine. Changed this to just read
6923 a certificate: this is the best we can do. Also modified the code
6924 in apps/verify.c to take notice of return codes: it was previously
6925 attempting to read in certificates from NULL pointers and ignoring
6926 any errors: this is one reason why the cert and CRL reader seemed
6927 to work. It doesn't check return codes from the default certificate
6928 routines: these may well fail if the certificates aren't installed.
6931 *) Code to support otherName option in GeneralName.
6934 *) First update to verify code. Change the verify utility
6935 so it warns if it is passed a self signed certificate:
6936 for consistency with the normal behaviour. X509_verify
6937 has been modified to it will now verify a self signed
6938 certificate if *exactly* the same certificate appears
6939 in the store: it was previously impossible to trust a
6940 single self signed certificate. This means that:
6941 openssl verify ss.pem
6942 now gives a warning about a self signed certificate but
6943 openssl verify -CAfile ss.pem ss.pem
6947 *) For servers, store verify_result in SSL_SESSION data structure
6948 (and add it to external session representation).
6949 This is needed when client certificate verifications fails,
6950 but an application-provided verification callback (set by
6951 SSL_CTX_set_cert_verify_callback) allows accepting the session
6952 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6953 but returns 1): When the session is reused, we have to set
6954 ssl->verify_result to the appropriate error code to avoid
6956 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6958 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6959 case in PKCS7_dataInit() where the signed PKCS7 structure
6960 didn't contain any existing data because it was being created.
6961 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6963 *) Add a salt to the key derivation routines in enc.c. This
6964 forms the first 8 bytes of the encrypted file. Also add a
6965 -S option to allow a salt to be input on the command line.
6968 *) New function X509_cmp(). Oddly enough there wasn't a function
6969 to compare two certificates. We do this by working out the SHA1
6970 hash and comparing that. X509_cmp() will be needed by the trust
6974 *) SSL_get1_session() is like SSL_get_session(), but increments
6975 the reference count in the SSL_SESSION returned.
6976 [Geoff Thorpe <geoff@eu.c2.net>]
6978 *) Fix for 'req': it was adding a null to request attributes.
6979 Also change the X509_LOOKUP and X509_INFO code to handle
6980 certificate auxiliary information.
6983 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6987 *) Add the possibility to add extra information to the memory leak
6988 detecting output, to form tracebacks, showing from where each
6989 allocation was originated: CRYPTO_push_info("constant string") adds
6990 the string plus current file name and line number to a per-thread
6991 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6992 is like calling CYRPTO_pop_info() until the stack is empty.
6993 Also updated memory leak detection code to be multi-thread-safe.
6996 *) Add options -text and -noout to pkcs7 utility and delete the
6997 encryption options which never did anything. Update docs.
7000 *) Add options to some of the utilities to allow the pass phrase
7001 to be included on either the command line (not recommended on
7002 OSes like Unix) or read from the environment. Update the
7003 manpages and fix a few bugs.
7006 *) Add a few manpages for some of the openssl commands.
7009 *) Fix the -revoke option in ca. It was freeing up memory twice,
7010 leaking and not finding already revoked certificates.
7013 *) Extensive changes to support certificate auxiliary information.
7014 This involves the use of X509_CERT_AUX structure and X509_AUX
7015 functions. An X509_AUX function such as PEM_read_X509_AUX()
7016 can still read in a certificate file in the usual way but it
7017 will also read in any additional "auxiliary information". By
7018 doing things this way a fair degree of compatibility can be
7019 retained: existing certificates can have this information added
7020 using the new 'x509' options.
7022 Current auxiliary information includes an "alias" and some trust
7023 settings. The trust settings will ultimately be used in enhanced
7024 certificate chain verification routines: currently a certificate
7025 can only be trusted if it is self signed and then it is trusted
7029 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
7030 The problem was that one of the replacement routines had not been working
7031 since SSLeay releases. For now the offending routine has been replaced
7032 with non-optimised assembler. Even so, this now gives around 95%
7033 performance improvement for 1024 bit RSA signs.
7036 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
7037 handling. Most clients have the effective key size in bits equal to
7038 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
7039 A few however don't do this and instead use the size of the decrypted key
7040 to determine the RC2 key length and the AlgorithmIdentifier to determine
7041 the effective key length. In this case the effective key length can still
7042 be 40 bits but the key length can be 168 bits for example. This is fixed
7043 by manually forcing an RC2 key into the EVP_PKEY structure because the
7044 EVP code can't currently handle unusual RC2 key sizes: it always assumes
7045 the key length and effective key length are equal.
7048 *) Add a bunch of functions that should simplify the creation of
7049 X509_NAME structures. Now you should be able to do:
7050 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
7051 and have it automatically work out the correct field type and fill in
7052 the structures. The more adventurous can try:
7053 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
7054 and it will (hopefully) work out the correct multibyte encoding.
7057 *) Change the 'req' utility to use the new field handling and multibyte
7058 copy routines. Before the DN field creation was handled in an ad hoc
7059 way in req, ca, and x509 which was rather broken and didn't support
7060 BMPStrings or UTF8Strings. Since some software doesn't implement
7061 BMPStrings or UTF8Strings yet, they can be enabled using the config file
7062 using the dirstring_type option. See the new comment in the default
7063 openssl.cnf for more info.
7066 *) Make crypto/rand/md_rand.c more robust:
7067 - Assure unique random numbers after fork().
7068 - Make sure that concurrent threads access the global counter and
7069 md serializably so that we never lose entropy in them
7070 or use exactly the same state in multiple threads.
7071 Access to the large state is not always serializable because
7072 the additional locking could be a performance killer, and
7073 md should be large enough anyway.
7076 *) New file apps/app_rand.c with commonly needed functionality
7077 for handling the random seed file.
7079 Use the random seed file in some applications that previously did not:
7081 dsaparam -genkey (which also ignored its '-rand' option),
7084 x509 (when signing).
7085 Except on systems with /dev/urandom, it is crucial to have a random
7086 seed file at least for key creation, DSA signing, and for DH exchanges;
7087 for RSA signatures we could do without one.
7089 gendh and gendsa (unlike genrsa) used to read only the first byte
7090 of each file listed in the '-rand' option. The function as previously
7091 found in genrsa is now in app_rand.c and is used by all programs
7092 that support '-rand'.
7095 *) In RAND_write_file, use mode 0600 for creating files;
7096 don't just chmod when it may be too late.
7099 *) Report an error from X509_STORE_load_locations
7100 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
7103 *) New function ASN1_mbstring_copy() this copies a string in either
7104 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
7105 into an ASN1_STRING type. A mask of permissible types is passed
7106 and it chooses the "minimal" type to use or an error if not type
7110 *) Add function equivalents to the various macros in asn1.h. The old
7111 macros are retained with an M_ prefix. Code inside the library can
7112 use the M_ macros. External code (including the openssl utility)
7113 should *NOT* in order to be "shared library friendly".
7116 *) Add various functions that can check a certificate's extensions
7117 to see if it usable for various purposes such as SSL client,
7118 server or S/MIME and CAs of these types. This is currently
7119 VERY EXPERIMENTAL but will ultimately be used for certificate chain
7120 verification. Also added a -purpose flag to x509 utility to
7121 print out all the purposes.
7124 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
7128 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
7129 for, obtain and decode and extension and obtain its critical flag.
7130 This allows all the necessary extension code to be handled in a
7131 single function call.
7134 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
7135 platforms. See crypto/rc4/rc4_enc.c for further details.
7138 *) New -noout option to asn1parse. This causes no output to be produced
7139 its main use is when combined with -strparse and -out to extract data
7140 from a file (which may not be in ASN.1 format).
7143 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
7144 when producing the local key id.
7145 [Richard Levitte <levitte@stacken.kth.se>]
7147 *) New option -dhparam in s_server. This allows a DH parameter file to be
7148 stated explicitly. If it is not stated then it tries the first server
7149 certificate file. The previous behaviour hard coded the filename
7153 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
7154 a public key to be input or output. For example:
7155 openssl rsa -in key.pem -pubout -out pubkey.pem
7156 Also added necessary DSA public key functions to handle this.
7159 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
7160 in the message. This was handled by allowing
7161 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
7162 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
7164 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
7165 to the end of the strings whereas this didn't. This would cause problems
7166 if strings read with d2i_ASN1_bytes() were later modified.
7167 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
7169 *) Fix for base64 decode bug. When a base64 bio reads only one line of
7170 data and it contains EOF it will end up returning an error. This is
7171 caused by input 46 bytes long. The cause is due to the way base64
7172 BIOs find the start of base64 encoded data. They do this by trying a
7173 trial decode on each line until they find one that works. When they
7174 do a flag is set and it starts again knowing it can pass all the
7175 data directly through the decoder. Unfortunately it doesn't reset
7176 the context it uses. This means that if EOF is reached an attempt
7177 is made to pass two EOFs through the context and this causes the
7178 resulting error. This can also cause other problems as well. As is
7179 usual with these problems it takes *ages* to find and the fix is
7180 trivial: move one line.
7181 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
7183 *) Ugly workaround to get s_client and s_server working under Windows. The
7184 old code wouldn't work because it needed to select() on sockets and the
7185 tty (for keypresses and to see if data could be written). Win32 only
7186 supports select() on sockets so we select() with a 1s timeout on the
7187 sockets and then see if any characters are waiting to be read, if none
7188 are present then we retry, we also assume we can always write data to
7189 the tty. This isn't nice because the code then blocks until we've
7190 received a complete line of data and it is effectively polling the
7191 keyboard at 1s intervals: however it's quite a bit better than not
7192 working at all :-) A dedicated Windows application might handle this
7193 with an event loop for example.
7196 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
7197 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
7198 will be called when RSA_sign() and RSA_verify() are used. This is useful
7199 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
7200 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
7201 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
7202 This necessitated the support of an extra signature type NID_md5_sha1
7203 for SSL signatures and modifications to the SSL library to use it instead
7204 of calling RSA_public_decrypt() and RSA_private_encrypt().
7207 *) Add new -verify -CAfile and -CApath options to the crl program, these
7208 will lookup a CRL issuers certificate and verify the signature in a
7209 similar way to the verify program. Tidy up the crl program so it
7210 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
7211 less strict. It will now permit CRL extensions even if it is not
7212 a V2 CRL: this will allow it to tolerate some broken CRLs.
7215 *) Initialize all non-automatic variables each time one of the openssl
7216 sub-programs is started (this is necessary as they may be started
7217 multiple times from the "OpenSSL>" prompt).
7218 [Lennart Bang, Bodo Moeller]
7220 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
7221 removing all other RSA functionality (this is what NO_RSA does). This
7222 is so (for example) those in the US can disable those operations covered
7223 by the RSA patent while allowing storage and parsing of RSA keys and RSA
7227 *) Non-copying interface to BIO pairs.
7228 (still largely untested)
7231 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
7232 ASCII string. This was handled independently in various places before.
7235 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
7236 UTF8 strings a character at a time.
7239 *) Use client_version from client hello to select the protocol
7240 (s23_srvr.c) and for RSA client key exchange verification
7241 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
7244 *) Add various utility functions to handle SPKACs, these were previously
7245 handled by poking round in the structure internals. Added new function
7246 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
7247 print, verify and generate SPKACs. Based on an original idea from
7248 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
7251 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7254 *) Allow the config file extension section to be overwritten on the
7255 command line. Based on an original idea from Massimiliano Pala
7256 <madwolf@comune.modena.it>. The new option is called -extensions
7257 and can be applied to ca, req and x509. Also -reqexts to override
7258 the request extensions in req and -crlexts to override the crl extensions
7262 *) Add new feature to the SPKAC handling in ca. Now you can include
7263 the same field multiple times by preceding it by "XXXX." for example:
7266 this is the same syntax as used in the req config file.
7269 *) Allow certificate extensions to be added to certificate requests. These
7270 are specified in a 'req_extensions' option of the req section of the
7271 config file. They can be printed out with the -text option to req but
7272 are otherwise ignored at present.
7275 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7276 data read consists of only the final block it would not decrypted because
7277 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7278 A misplaced 'break' also meant the decrypted final block might not be
7279 copied until the next read.
7282 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7283 a few extra parameters to the DH structure: these will be useful if
7284 for example we want the value of 'q' or implement X9.42 DH.
7287 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7288 provides hooks that allow the default DSA functions or functions on a
7289 "per key" basis to be replaced. This allows hardware acceleration and
7290 hardware key storage to be handled without major modification to the
7291 library. Also added low level modexp hooks and CRYPTO_EX structure and
7292 associated functions.
7295 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7296 as "read only": it can't be written to and the buffer it points to will
7297 not be freed. Reading from a read only BIO is much more efficient than
7298 a normal memory BIO. This was added because there are several times when
7299 an area of memory needs to be read from a BIO. The previous method was
7300 to create a memory BIO and write the data to it, this results in two
7301 copies of the data and an O(n^2) reading algorithm. There is a new
7302 function BIO_new_mem_buf() which creates a read only memory BIO from
7303 an area of memory. Also modified the PKCS#7 routines to use read only
7307 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7308 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7309 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7310 but a retry condition occured while trying to read the rest.
7313 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7314 NID_pkcs7_encrypted by default: this was wrong since this should almost
7315 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7316 the encrypted data type: this is a more sensible place to put it and it
7317 allows the PKCS#12 code to be tidied up that duplicated this
7321 *) Changed obj_dat.pl script so it takes its input and output files on
7322 the command line. This should avoid shell escape redirection problems
7326 *) Initial support for certificate extension requests, these are included
7327 in things like Xenroll certificate requests. Included functions to allow
7328 extensions to be obtained and added.
7331 *) -crlf option to s_client and s_server for sending newlines as
7332 CRLF (as required by many protocols).
7335 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7337 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7338 [Ralf S. Engelschall]
7340 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7341 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7343 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7347 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7348 DH parameters/keys (q is lost during that conversion, but the resulting
7349 DH parameters contain its length).
7351 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7352 much faster than DH_generate_parameters (which creates parameters
7353 where p = 2*q + 1), and also the smaller q makes DH computations
7354 much more efficient (160-bit exponentiation instead of 1024-bit
7355 exponentiation); so this provides a convenient way to support DHE
7356 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7357 utter importance to use
7358 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7360 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7361 when such DH parameters are used, because otherwise small subgroup
7362 attacks may become possible!
7365 *) Avoid memory leak in i2d_DHparams.
7368 *) Allow the -k option to be used more than once in the enc program:
7369 this allows the same encrypted message to be read by multiple recipients.
7372 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7373 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7374 it will always use the numerical form of the OID, even if it has a short
7378 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7379 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7380 otherwise bn_mod_exp was called. In the case of hardware keys for example
7381 no private key components need be present and it might store extra data
7382 in the RSA structure, which cannot be accessed from bn_mod_exp.
7383 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7384 private key operations.
7387 *) Added support for SPARC Linux.
7390 *) pem_password_cb function type incompatibly changed from
7391 typedef int pem_password_cb(char *buf, int size, int rwflag);
7393 ....(char *buf, int size, int rwflag, void *userdata);
7394 so that applications can pass data to their callbacks:
7395 The PEM[_ASN1]_{read,write}... functions and macros now take an
7396 additional void * argument, which is just handed through whenever
7397 the password callback is called.
7398 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7400 New function SSL_CTX_set_default_passwd_cb_userdata.
7402 Compatibility note: As many C implementations push function arguments
7403 onto the stack in reverse order, the new library version is likely to
7404 interoperate with programs that have been compiled with the old
7405 pem_password_cb definition (PEM_whatever takes some data that
7406 happens to be on the stack as its last argument, and the callback
7407 just ignores this garbage); but there is no guarantee whatsoever that
7410 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7411 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7412 problems not only on Windows, but also on some Unix platforms.
7413 To avoid problematic command lines, these definitions are now in an
7414 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7415 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7418 *) MIPS III/IV assembler module is reimplemented.
7421 *) More DES library cleanups: remove references to srand/rand and
7422 delete an unused file.
7425 *) Add support for the the free Netwide assembler (NASM) under Win32,
7426 since not many people have MASM (ml) and it can be hard to obtain.
7427 This is currently experimental but it seems to work OK and pass all
7428 the tests. Check out INSTALL.W32 for info.
7431 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7432 without temporary keys kept an extra copy of the server key,
7433 and connections with temporary keys did not free everything in case
7437 *) New function RSA_check_key and new openssl rsa option -check
7438 for verifying the consistency of RSA keys.
7439 [Ulf Moeller, Bodo Moeller]
7441 *) Various changes to make Win32 compile work:
7442 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7443 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7444 comparison" warnings.
7445 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7448 *) Add a debugging option to PKCS#5 v2 key generation function: when
7449 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7450 derived keys are printed to stderr.
7453 *) Copy the flags in ASN1_STRING_dup().
7454 [Roman E. Pavlov <pre@mo.msk.ru>]
7456 *) The x509 application mishandled signing requests containing DSA
7457 keys when the signing key was also DSA and the parameters didn't match.
7459 It was supposed to omit the parameters when they matched the signing key:
7460 the verifying software was then supposed to automatically use the CA's
7461 parameters if they were absent from the end user certificate.
7463 Omitting parameters is no longer recommended. The test was also
7464 the wrong way round! This was probably due to unusual behaviour in
7465 EVP_cmp_parameters() which returns 1 if the parameters match.
7466 This meant that parameters were omitted when they *didn't* match and
7467 the certificate was useless. Certificates signed with 'ca' didn't have
7469 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7471 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7472 The interface is as follows:
7473 Applications can use
7474 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7475 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7476 "off" is now the default.
7477 The library internally uses
7478 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7479 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7480 to disable memory-checking temporarily.
7482 Some inconsistent states that previously were possible (and were
7483 even the default) are now avoided.
7485 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7486 with each memory chunk allocated; this is occasionally more helpful
7487 than just having a counter.
7489 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7491 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7495 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7496 which largely parallels "options", but is for changing API behaviour,
7497 whereas "options" are about protocol behaviour.
7498 Initial "mode" flags are:
7500 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7501 a single record has been written.
7502 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7503 retries use the same buffer location.
7504 (But all of the contents must be
7508 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7511 *) Fix problems with no-hmac etc.
7512 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7514 *) New functions RSA_get_default_method(), RSA_set_method() and
7515 RSA_get_method(). These allows replacement of RSA_METHODs without having
7516 to mess around with the internals of an RSA structure.
7519 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7520 Also really enable memory leak checks in openssl.c and in some
7522 [Chad C. Mulligan, Bodo Moeller]
7524 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7525 up the length of negative integers. This has now been simplified to just
7526 store the length when it is first determined and use it later, rather
7527 than trying to keep track of where data is copied and updating it to
7529 [Steve Henson, reported by Brien Wheeler
7530 <bwheeler@authentica-security.com>]
7532 *) Add a new function PKCS7_signatureVerify. This allows the verification
7533 of a PKCS#7 signature but with the signing certificate passed to the
7534 function itself. This contrasts with PKCS7_dataVerify which assumes the
7535 certificate is present in the PKCS#7 structure. This isn't always the
7536 case: certificates can be omitted from a PKCS#7 structure and be
7537 distributed by "out of band" means (such as a certificate database).
7540 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7541 function prototypes in pem.h, also change util/mkdef.pl to add the
7542 necessary function names.
7545 *) mk1mf.pl (used by Windows builds) did not properly read the
7546 options set by Configure in the top level Makefile, and Configure
7547 was not even able to write more than one option correctly.
7548 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7551 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7552 file to be loaded from a BIO or FILE pointer. The BIO version will
7553 for example allow memory BIOs to contain config info.
7556 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7557 Whoever hopes to achieve shared-library compatibility across versions
7558 must use this, not the compile-time macro.
7559 (Exercise 0.9.4: Which is the minimum library version required by
7561 Note: All this applies only to multi-threaded programs, others don't
7565 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7566 through a BIO pair triggered the default case, i.e.
7567 SSLerr(...,SSL_R_UNKNOWN_STATE).
7570 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7571 can use the SSL library even if none of the specific BIOs is
7575 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7576 for the encoded length.
7577 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7579 *) Add initial documentation of the X509V3 functions.
7582 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7583 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7584 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7585 secure PKCS#8 private key format with a high iteration count.
7588 *) Fix determination of Perl interpreter: A perl or perl5
7589 _directory_ in $PATH was also accepted as the interpreter.
7590 [Ralf S. Engelschall]
7592 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7593 wrong with it but it was very old and did things like calling
7594 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7598 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7599 to use the new extension code.
7602 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7603 with macros. This should make it easier to change their form, add extra
7604 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7608 *) Add to configuration table a new entry that can specify an alternative
7609 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7610 according to Mark Crispin <MRC@Panda.COM>.
7614 *) DES CBC did not update the IV. Weird.
7617 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7618 Changing the behaviour of the former might break existing programs --
7619 where IV updating is needed, des_ncbc_encrypt can be used.
7622 *) When bntest is run from "make test" it drives bc to check its
7623 calculations, as well as internally checking them. If an internal check
7624 fails, it needs to cause bc to give a non-zero result or make test carries
7625 on without noticing the failure. Fixed.
7628 *) DES library cleanups.
7631 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7632 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7633 ciphers. NOTE: although the key derivation function has been verified
7634 against some published test vectors it has not been extensively tested
7635 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7639 *) Instead of "mkdir -p", which is not fully portable, use new
7640 Perl script "util/mkdir-p.pl".
7643 *) Rewrite the way password based encryption (PBE) is handled. It used to
7644 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7645 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7646 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7647 the 'parameter' field of the AlgorithmIdentifier is passed to the
7648 underlying key generation function so it must do its own ASN1 parsing.
7649 This has also changed the EVP_PBE_CipherInit() function which now has a
7650 'parameter' argument instead of literal salt and iteration count values
7651 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7654 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7655 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7656 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7657 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7658 value was just used as a "magic string" and not used directly its
7659 value doesn't matter.
7662 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7666 *) "linux-sparc64" configuration (ultrapenguin).
7667 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7668 "linux-sparc" configuration.
7669 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7671 *) config now generates no-xxx options for missing ciphers.
7674 *) Support the EBCDIC character set (work in progress).
7675 File ebcdic.c not yet included because it has a different license.
7676 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7678 *) Support BS2000/OSD-POSIX.
7679 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7681 *) Make callbacks for key generation use void * instead of char *.
7684 *) Make S/MIME samples compile (not yet tested).
7687 *) Additional typesafe stacks.
7690 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7694 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7696 *) New configuration variant "sco5-gcc".
7698 *) Updated some demos.
7699 [Sean O Riordain, Wade Scholine]
7701 *) Add missing BIO_free at exit of pkcs12 application.
7704 *) Fix memory leak in conf.c.
7707 *) Updates for Win32 to assembler version of MD5.
7710 *) Set #! path to perl in apps/der_chop to where we found it
7711 instead of using a fixed path.
7714 *) SHA library changes for irix64-mips4-cc.
7717 *) Improvements for VMS support.
7721 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7723 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7724 This also avoids the problems with SC4.2 and unpatched SC5.
7725 [Andy Polyakov <appro@fy.chalmers.se>]
7727 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7728 These are required because of the typesafe stack would otherwise break
7729 existing code. If old code used a structure member which used to be STACK
7730 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7731 sk_num or sk_value it would produce an error because the num, data members
7732 are not present in STACK_OF. Now it just produces a warning. sk_set
7733 replaces the old method of assigning a value to sk_value
7734 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7735 that does this will no longer work (and should use sk_set instead) but
7736 this could be regarded as a "questionable" behaviour anyway.
7739 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7740 correctly handle encrypted S/MIME data.
7743 *) Change type of various DES function arguments from des_cblock
7744 (which means, in function argument declarations, pointer to char)
7745 to des_cblock * (meaning pointer to array with 8 char elements),
7746 which allows the compiler to do more typechecking; it was like
7747 that back in SSLeay, but with lots of ugly casts.
7749 Introduce new type const_des_cblock.
7752 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7753 problems: find RecipientInfo structure that matches recipient certificate
7754 and initialise the ASN1 structures properly based on passed cipher.
7757 *) Belatedly make the BN tests actually check the results.
7760 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7761 to and from BNs: it was completely broken. New compilation option
7762 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7763 key elements as negative integers.
7766 *) Reorganize and speed up MD5.
7767 [Andy Polyakov <appro@fy.chalmers.se>]
7770 [Richard Levitte <richard@levitte.org>]
7772 *) New option -out to asn1parse to allow the parsed structure to be
7773 output to a file. This is most useful when combined with the -strparse
7774 option to examine the output of things like OCTET STRINGS.
7777 *) Make SSL library a little more fool-proof by not requiring any longer
7778 that SSL_set_{accept,connect}_state be called before
7779 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7780 in many applications because usually everything *appeared* to work as
7781 intended anyway -- now it really works as intended).
7784 *) Move openssl.cnf out of lib/.
7787 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7788 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7789 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7790 [Ralf S. Engelschall]
7792 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7793 handle PKCS#7 enveloped data properly.
7794 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7796 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7797 copying pointers. The cert_st handling is changed by this in
7798 various ways (and thus what used to be known as ctx->default_cert
7799 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7800 any longer when s->cert does not give us what we need).
7801 ssl_cert_instantiate becomes obsolete by this change.
7802 As soon as we've got the new code right (possibly it already is?),
7803 we have solved a couple of bugs of the earlier code where s->cert
7804 was used as if it could not have been shared with other SSL structures.
7806 Note that using the SSL API in certain dirty ways now will result
7807 in different behaviour than observed with earlier library versions:
7808 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7809 does not influence s as it used to.
7811 In order to clean up things more thoroughly, inside SSL_SESSION
7812 we don't use CERT any longer, but a new structure SESS_CERT
7813 that holds per-session data (if available); currently, this is
7814 the peer's certificate chain and, for clients, the server's certificate
7815 and temporary key. CERT holds only those values that can have
7816 meaningful defaults in an SSL_CTX.
7819 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7820 from the internal representation. Various PKCS#7 fixes: remove some
7821 evil casts and set the enc_dig_alg field properly based on the signing
7825 *) Allow PKCS#12 password to be set from the command line or the
7826 environment. Let 'ca' get its config file name from the environment
7827 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7831 *) Allow certificate policies extension to use an IA5STRING for the
7832 organization field. This is contrary to the PKIX definition but
7833 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7837 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7838 without disallowing inline assembler and the like for non-pedantic builds.
7841 *) Support Borland C++ builder.
7842 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7847 *) SHA-1 cleanups and performance enhancements.
7848 [Andy Polyakov <appro@fy.chalmers.se>]
7850 *) Sparc v8plus assembler for the bignum library.
7851 [Andy Polyakov <appro@fy.chalmers.se>]
7853 *) Accept any -xxx and +xxx compiler options in Configure.
7856 *) Update HPUX configuration.
7859 *) Add missing sk_<type>_unshift() function to safestack.h
7860 [Ralf S. Engelschall]
7862 *) New function SSL_CTX_use_certificate_chain_file that sets the
7863 "extra_cert"s in addition to the certificate. (This makes sense
7864 only for "PEM" format files, as chains as a whole are not
7868 *) Support verify_depth from the SSL API.
7869 x509_vfy.c had what can be considered an off-by-one-error:
7870 Its depth (which was not part of the external interface)
7871 was actually counting the number of certificates in a chain;
7872 now it really counts the depth.
7875 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7876 instead of X509err, which often resulted in confusing error
7877 messages since the error codes are not globally unique
7878 (e.g. an alleged error in ssl3_accept when a certificate
7879 didn't match the private key).
7881 *) New function SSL_CTX_set_session_id_context that allows to set a default
7882 value (so that you don't need SSL_set_session_id_context for each
7883 connection using the SSL_CTX).
7886 *) OAEP decoding bug fix.
7889 *) Support INSTALL_PREFIX for package builders, as proposed by
7893 *) New Configure options "threads" and "no-threads". For systems
7894 where the proper compiler options are known (currently Solaris
7895 and Linux), "threads" is the default.
7898 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7901 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7902 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7903 such as /usr/local/bin.
7906 *) "make linux-shared" to build shared libraries.
7907 [Niels Poppe <niels@netbox.org>]
7909 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7912 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7913 extension adding in x509 utility.
7916 *) Remove NOPROTO sections and error code comments.
7919 *) Partial rewrite of the DEF file generator to now parse the ANSI
7923 *) New Configure options --prefix=DIR and --openssldir=DIR.
7926 *) Complete rewrite of the error code script(s). It is all now handled
7927 by one script at the top level which handles error code gathering,
7928 header rewriting and C source file generation. It should be much better
7929 than the old method: it now uses a modified version of Ulf's parser to
7930 read the ANSI prototypes in all header files (thus the old K&R definitions
7931 aren't needed for error creation any more) and do a better job of
7932 translating function codes into names. The old 'ASN1 error code imbedded
7933 in a comment' is no longer necessary and it doesn't use .err files which
7934 have now been deleted. Also the error code call doesn't have to appear all
7935 on one line (which resulted in some large lines...).
7938 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7941 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7942 0 (which usually indicates a closed connection), but continue reading.
7945 *) Fix some race conditions.
7948 *) Add support for CRL distribution points extension. Add Certificate
7949 Policies and CRL distribution points documentation.
7952 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7955 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7956 8 of keying material. Merlin has also confirmed interop with this fix
7957 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7958 [Merlin Hughes <merlin@baltimore.ie>]
7960 *) Fix lots of warnings.
7961 [Richard Levitte <levitte@stacken.kth.se>]
7963 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7964 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7965 [Richard Levitte <levitte@stacken.kth.se>]
7967 *) Fix problems with sizeof(long) == 8.
7968 [Andy Polyakov <appro@fy.chalmers.se>]
7970 *) Change functions to ANSI C.
7973 *) Fix typos in error codes.
7974 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7976 *) Remove defunct assembler files from Configure.
7979 *) SPARC v8 assembler BIGNUM implementation.
7980 [Andy Polyakov <appro@fy.chalmers.se>]
7982 *) Support for Certificate Policies extension: both print and set.
7983 Various additions to support the r2i method this uses.
7986 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7987 return a const string when you are expecting an allocated buffer.
7990 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7991 types DirectoryString and DisplayText.
7994 *) Add code to allow r2i extensions to access the configuration database,
7995 add an LHASH database driver and add several ctx helper functions.
7998 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7999 fail when they extended the size of a BIGNUM.
8002 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
8003 support typesafe stack.
8006 *) Fix typo in SSL_[gs]et_options().
8007 [Nils Frostberg <nils@medcom.se>]
8009 *) Delete various functions and files that belonged to the (now obsolete)
8010 old X509V3 handling code.
8013 *) New Configure option "rsaref".
8016 *) Don't auto-generate pem.h.
8019 *) Introduce type-safe ASN.1 SETs.
8022 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
8023 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
8025 *) Introduce type-safe STACKs. This will almost certainly break lots of code
8026 that links with OpenSSL (well at least cause lots of warnings), but fear
8027 not: the conversion is trivial, and it eliminates loads of evil casts. A
8028 few STACKed things have been converted already. Feel free to convert more.
8029 In the fullness of time, I'll do away with the STACK type altogether.
8032 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
8033 specified in <certfile> by updating the entry in the index.txt file.
8034 This way one no longer has to edit the index.txt file manually for
8035 revoking a certificate. The -revoke option does the gory details now.
8036 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
8038 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
8039 `-text' option at all and this way the `-noout -text' combination was
8040 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
8041 [Ralf S. Engelschall]
8043 *) Make sure a corresponding plain text error message exists for the
8044 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
8045 verify callback function determined that a certificate was revoked.
8046 [Ralf S. Engelschall]
8048 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
8049 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
8050 all available cipers including rc5, which was forgotten until now.
8051 In order to let the testing shell script know which algorithms
8052 are available, a new (up to now undocumented) command
8053 "openssl list-cipher-commands" is used.
8056 *) Bugfix: s_client occasionally would sleep in select() when
8057 it should have checked SSL_pending() first.
8060 *) New functions DSA_do_sign and DSA_do_verify to provide access to
8061 the raw DSA values prior to ASN.1 encoding.
8064 *) Tweaks to Configure
8065 [Niels Poppe <niels@netbox.org>]
8067 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
8071 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
8074 *) New config option to avoid instructions that are illegal on the 80386.
8075 The default code is faster, but requires at least a 486.
8078 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
8079 SSL2_SERVER_VERSION (not used at all) macros, which are now the
8080 same as SSL2_VERSION anyway.
8083 *) New "-showcerts" option for s_client.
8086 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
8087 application. Various cleanups and fixes.
8090 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
8091 modify error routines to work internally. Add error codes and PBE init
8092 to library startup routines.
8095 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
8096 packing functions to asn1 and evp. Changed function names and error
8097 codes along the way.
8100 *) PKCS12 integration: and so it begins... First of several patches to
8101 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
8102 objects to objects.h
8105 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
8106 and display support for Thawte strong extranet extension.
8109 *) Add LinuxPPC support.
8110 [Jeff Dubrule <igor@pobox.org>]
8112 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
8113 bn_div_words in alpha.s.
8114 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
8116 *) Make sure the RSA OAEP test is skipped under -DRSAref because
8117 OAEP isn't supported when OpenSSL is built with RSAref.
8118 [Ulf Moeller <ulf@fitug.de>]
8120 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
8121 so they no longer are missing under -DNOPROTO.
8122 [Soren S. Jorvang <soren@t.dk>]
8125 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
8127 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
8128 doesn't work when the session is reused. Coming soon!
8131 *) Fix a security hole, that allows sessions to be reused in the wrong
8132 context thus bypassing client cert protection! All software that uses
8133 client certs and session caches in multiple contexts NEEDS PATCHING to
8134 allow session reuse! A fuller solution is in the works.
8135 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
8137 *) Some more source tree cleanups (removed obsolete files
8138 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
8139 permission on "config" script to be executable) and a fix for the INSTALL
8141 [Ulf Moeller <ulf@fitug.de>]
8143 *) Remove some legacy and erroneous uses of malloc, free instead of
8145 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
8147 *) Make rsa_oaep_test return non-zero on error.
8148 [Ulf Moeller <ulf@fitug.de>]
8150 *) Add support for native Solaris shared libraries. Configure
8151 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
8152 if someone would make that last step automatic.
8153 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
8155 *) ctx_size was not built with the right compiler during "make links". Fixed.
8158 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
8159 except NULL ciphers". This means the default cipher list will no longer
8160 enable NULL ciphers. They need to be specifically enabled e.g. with
8161 the string "DEFAULT:eNULL".
8164 *) Fix to RSA private encryption routines: if p < q then it would
8165 occasionally produce an invalid result. This will only happen with
8166 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
8169 *) Be less restrictive and allow also `perl util/perlpath.pl
8170 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
8171 because this way one can also use an interpreter named `perl5' (which is
8172 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
8173 installed as `perl').
8174 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8176 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
8177 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8179 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
8180 advapi32.lib to Win32 build and change the pem test comparision
8181 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
8182 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
8183 and crypto/des/ede_cbcm_enc.c.
8186 *) DES quad checksum was broken on big-endian architectures. Fixed.
8189 *) Comment out two functions in bio.h that aren't implemented. Fix up the
8190 Win32 test batch file so it (might) work again. The Win32 test batch file
8191 is horrible: I feel ill....
8194 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
8195 in e_os.h. Audit of header files to check ANSI and non ANSI
8196 sections: 10 functions were absent from non ANSI section and not exported
8197 from Windows DLLs. Fixed up libeay.num for new functions.
8200 *) Make `openssl version' output lines consistent.
8201 [Ralf S. Engelschall]
8203 *) Fix Win32 symbol export lists for BIO functions: Added
8204 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
8205 to ms/libeay{16,32}.def.
8206 [Ralf S. Engelschall]
8208 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
8209 fine under Unix and passes some trivial tests I've now added. But the
8210 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
8211 added to make sure no one expects that this stuff really works in the
8212 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
8213 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
8215 [Ralf S. Engelschall]
8217 *) Fix the generation of two part addresses in perl.
8218 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
8220 *) Add config entry for Linux on MIPS.
8221 [John Tobey <jtobey@channel1.com>]
8223 *) Make links whenever Configure is run, unless we are on Windoze.
8226 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
8227 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
8231 *) Add a useful kludge to allow package maintainers to specify compiler and
8232 other platforms details on the command line without having to patch the
8233 Configure script everytime: One now can use ``perl Configure
8234 <id>:<details>'', i.e. platform ids are allowed to have details appended
8235 to them (seperated by colons). This is treated as there would be a static
8236 pre-configured entry in Configure's %table under key <id> with value
8237 <details> and ``perl Configure <id>'' is called. So, when you want to
8238 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
8239 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
8240 now, which overrides the FreeBSD-elf entry on-the-fly.
8241 [Ralf S. Engelschall]
8243 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
8246 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
8247 on the `perl Configure ...' command line. This way one can compile
8248 OpenSSL libraries with Position Independent Code (PIC) which is needed
8249 for linking it into DSOs.
8250 [Ralf S. Engelschall]
8252 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8256 *) Cleaned up the LICENSE document: The official contact for any license
8257 questions now is the OpenSSL core team under openssl-core@openssl.org.
8258 And add a paragraph about the dual-license situation to make sure people
8259 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8260 to the OpenSSL toolkit.
8261 [Ralf S. Engelschall]
8263 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8264 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8265 Additonally cleaned up the `make links' target: Remove unnecessary
8266 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8267 to speed processing and no longer clutter the display with confusing
8268 stuff. Instead only the actually done links are displayed.
8269 [Ralf S. Engelschall]
8271 *) Permit null encryption ciphersuites, used for authentication only. It used
8272 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8273 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8277 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8278 signed attributes when verifying signatures (this would break them),
8279 the detached data encoding was wrong and public keys obtained using
8280 X509_get_pubkey() weren't freed.
8283 *) Add text documentation for the BUFFER functions. Also added a work around
8284 to a Win95 console bug. This was triggered by the password read stuff: the
8285 last character typed gets carried over to the next fread(). If you were
8286 generating a new cert request using 'req' for example then the last
8287 character of the passphrase would be CR which would then enter the first
8291 *) Added the new `Includes OpenSSL Cryptography Software' button as
8292 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8293 button and can be used by applications based on OpenSSL to show the
8294 relationship to the OpenSSL project.
8295 [Ralf S. Engelschall]
8297 *) Remove confusing variables in function signatures in files
8298 ssl/ssl_lib.c and ssl/ssl.h.
8299 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8301 *) Don't install bss_file.c under PREFIX/include/
8302 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8304 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8305 functions that return function pointers and has support for NT specific
8306 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8307 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8308 unsigned to signed types: this was killing the Win32 compile.
8311 *) Add new certificate file to stack functions,
8312 SSL_add_dir_cert_subjects_to_stack() and
8313 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8314 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8315 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8316 This means that Apache-SSL and similar packages don't have to mess around
8317 to add as many CAs as they want to the preferred list.
8320 *) Experiment with doxygen documentation. Currently only partially applied to
8322 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8323 openssl.doxy as the configuration file.
8326 *) Get rid of remaining C++-style comments which strict C compilers hate.
8327 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8329 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8330 compiled in by default: it has problems with large keys.
8333 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8334 DH private keys and/or callback functions which directly correspond to
8335 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8336 is needed for applications which have to configure certificates on a
8337 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8339 For the RSA certificate situation is makes no difference, but
8340 for the DSA certificate situation this fixes the "no shared cipher"
8341 problem where the OpenSSL cipher selection procedure failed because the
8342 temporary keys were not overtaken from the context and the API provided
8343 no way to reconfigure them.
8344 The new functions now let applications reconfigure the stuff and they
8345 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8346 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8347 non-public-API function ssl_cert_instantiate() is used as a helper
8348 function and also to reduce code redundancy inside ssl_rsa.c.
8349 [Ralf S. Engelschall]
8351 *) Move s_server -dcert and -dkey options out of the undocumented feature
8352 area because they are useful for the DSA situation and should be
8353 recognized by the users.
8354 [Ralf S. Engelschall]
8356 *) Fix the cipher decision scheme for export ciphers: the export bits are
8357 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8358 SSL_EXP_MASK. So, the original variable has to be used instead of the
8359 already masked variable.
8360 [Richard Levitte <levitte@stacken.kth.se>]
8362 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8363 [Richard Levitte <levitte@stacken.kth.se>]
8365 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8366 from `int' to `unsigned int' because it's a length and initialized by
8367 EVP_DigestFinal() which expects an `unsigned int *'.
8368 [Richard Levitte <levitte@stacken.kth.se>]
8370 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8371 script. Instead use the usual Shell->Perl transition trick.
8372 [Ralf S. Engelschall]
8374 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8375 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8376 -noout -modulus' as it's already the case for `openssl rsa -noout
8377 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8378 currently the public key is printed (a decision which was already done by
8379 `openssl dsa -modulus' in the past) which serves a similar purpose.
8380 Additionally the NO_RSA no longer completely removes the whole -modulus
8381 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8383 [Ralf S. Engelschall]
8385 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8386 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8387 [Arne Ansper <arne@ats.cyber.ee>]
8389 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8390 to be added. Now both 'req' and 'ca' can use new objects defined in the
8394 *) Add cool BIO that does syslog (or event log on NT).
8395 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8397 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8398 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8399 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8400 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8403 *) Add preliminary config info for new extension code.
8406 *) Make RSA_NO_PADDING really use no padding.
8407 [Ulf Moeller <ulf@fitug.de>]
8409 *) Generate errors when private/public key check is done.
8412 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8413 for some CRL extensions and new objects added.
8416 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8417 key usage extension and fuller support for authority key id.
8420 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8421 padding method for RSA, which is recommended for new applications in PKCS
8422 #1 v2.0 (RFC 2437, October 1998).
8423 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8424 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8425 against Bleichbacher's attack on RSA.
8426 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8429 *) Updates to the new SSL compression code
8430 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8432 *) Fix so that the version number in the master secret, when passed
8433 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8434 (because the server will not accept higher), that the version number
8435 is 0x03,0x01, not 0x03,0x00
8436 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8438 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8439 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8440 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8443 *) Support for RAW extensions where an arbitrary extension can be
8444 created by including its DER encoding. See apps/openssl.cnf for
8448 *) Make sure latest Perl versions don't interpret some generated C array
8449 code as Perl array code in the crypto/err/err_genc.pl script.
8450 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8452 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8453 not many people have the assembler. Various Win32 compilation fixes and
8454 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8458 *) Modify configure script 'Configure' to automatically create crypto/date.h
8459 file under Win32 and also build pem.h from pem.org. New script
8460 util/mkfiles.pl to create the MINFO file on environments that can't do a
8461 'make files': perl util/mkfiles.pl >MINFO should work.
8464 *) Major rework of DES function declarations, in the pursuit of correctness
8465 and purity. As a result, many evil casts evaporated, and some weirdness,
8466 too. You may find this causes warnings in your code. Zapping your evil
8467 casts will probably fix them. Mostly.
8470 *) Fix for a typo in asn1.h. Bug fix to object creation script
8471 obj_dat.pl. It considered a zero in an object definition to mean
8472 "end of object": none of the objects in objects.h have any zeros
8473 so it wasn't spotted.
8474 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8476 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8477 Masking (CBCM). In the absence of test vectors, the best I have been able
8478 to do is check that the decrypt undoes the encrypt, so far. Send me test
8479 vectors if you have them.
8482 *) Correct calculation of key length for export ciphers (too much space was
8483 allocated for null ciphers). This has not been tested!
8486 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8487 message is now correct (it understands "crypto" and "ssl" on its
8488 command line). There is also now an "update" option. This will update
8489 the util/ssleay.num and util/libeay.num files with any new functions.
8491 perl util/mkdef.pl crypto ssl update
8492 it will update them.
8495 *) Overhauled the Perl interface (perl/*):
8496 - ported BN stuff to OpenSSL's different BN library
8497 - made the perl/ source tree CVS-aware
8498 - renamed the package from SSLeay to OpenSSL (the files still contain
8499 their history because I've copied them in the repository)
8500 - removed obsolete files (the test scripts will be replaced
8501 by better Test::Harness variants in the future)
8502 [Ralf S. Engelschall]
8504 *) First cut for a very conservative source tree cleanup:
8505 1. merge various obsolete readme texts into doc/ssleay.txt
8506 where we collect the old documents and readme texts.
8507 2. remove the first part of files where I'm already sure that we no
8508 longer need them because of three reasons: either they are just temporary
8509 files which were left by Eric or they are preserved original files where
8510 I've verified that the diff is also available in the CVS via "cvs diff
8511 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8512 the crypto/md/ stuff).
8513 [Ralf S. Engelschall]
8515 *) More extension code. Incomplete support for subject and issuer alt
8516 name, issuer and authority key id. Change the i2v function parameters
8517 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8518 what that's for :-) Fix to ASN1 macro which messed up
8519 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8522 *) Preliminary support for ENUMERATED type. This is largely copied from the
8526 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8527 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8529 *) Make sure `make rehash' target really finds the `openssl' program.
8530 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8532 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8533 like to hear about it if this slows down other processors.
8536 *) Add CygWin32 platform information to Configure script.
8537 [Alan Batie <batie@aahz.jf.intel.com>]
8539 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8540 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8542 *) New program nseq to manipulate netscape certificate sequences
8545 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8549 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8550 but the BN code had some problems that would cause failures when
8551 doing certificate verification and some other functions.
8552 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8554 *) Add ASN1 and PEM code to support netscape certificate sequences.
8557 *) Add ASN1 and PEM code to support netscape certificate sequences.
8560 *) Add several PKIX and private extended key usage OIDs.
8563 *) Modify the 'ca' program to handle the new extension code. Modify
8564 openssl.cnf for new extension format, add comments.
8567 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8568 and add a sample to openssl.cnf so req -x509 now adds appropriate
8572 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8573 error code, add initial support to X509_print() and x509 application.
8576 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8577 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8578 stuff is currently isolated and isn't even compiled yet.
8581 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8582 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8583 Removed the versions check from X509 routines when loading extensions:
8584 this allows certain broken certificates that don't set the version
8585 properly to be processed.
8588 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8589 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8590 can still be regenerated with "make depend".
8593 *) Spelling mistake in C version of CAST-128.
8594 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8596 *) Changes to the error generation code. The perl script err-code.pl
8597 now reads in the old error codes and retains the old numbers, only
8598 adding new ones if necessary. It also only changes the .err files if new
8599 codes are added. The makefiles have been modified to only insert errors
8600 when needed (to avoid needlessly modifying header files). This is done
8601 by only inserting errors if the .err file is newer than the auto generated
8602 C file. To rebuild all the error codes from scratch (the old behaviour)
8603 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8604 or delete all the .err files.
8607 *) CAST-128 was incorrectly implemented for short keys. The C version has
8608 been fixed, but is untested. The assembler versions are also fixed, but
8609 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8610 to regenerate it if needed.
8611 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8612 Hagino <itojun@kame.net>]
8614 *) File was opened incorrectly in randfile.c.
8615 [Ulf Möller <ulf@fitug.de>]
8617 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8618 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8619 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8620 al: it's just almost always a UTCTime. Note this patch adds new error
8621 codes so do a "make errors" if there are problems.
8624 *) Correct Linux 1 recognition in config.
8625 [Ulf Möller <ulf@fitug.de>]
8627 *) Remove pointless MD5 hash when using DSA keys in ca.
8628 [Anonymous <nobody@replay.com>]
8630 *) Generate an error if given an empty string as a cert directory. Also
8631 generate an error if handed NULL (previously returned 0 to indicate an
8632 error, but didn't set one).
8633 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8635 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8638 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8639 parameters. This was causing a warning which killed off the Win32 compile.
8642 *) Remove C++ style comments from crypto/bn/bn_local.h.
8643 [Neil Costigan <neil.costigan@celocom.com>]
8645 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8646 based on a text string, looking up short and long names and finally
8647 "dot" format. The "dot" format stuff didn't work. Added new function
8648 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8649 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8650 OID is not part of the table.
8653 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8654 X509_LOOKUP_by_alias().
8657 *) Sort openssl functions by name.
8660 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8661 encryption from sample DSA keys (in case anyone is interested the password
8665 *) Make _all_ *_free functions accept a NULL pointer.
8666 [Frans Heymans <fheymans@isaserver.be>]
8668 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8670 [Anonymous <nobody@replay.com>]
8672 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8673 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8675 *) Don't blow it for numeric -newkey arguments to apps/req.
8676 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8678 *) Temp key "for export" tests were wrong in s3_srvr.c.
8679 [Anonymous <nobody@replay.com>]
8681 *) Add prototype for temp key callback functions
8682 SSL_CTX_set_tmp_{rsa,dh}_callback().
8685 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8686 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8689 *) X509_name_add_entry() freed the wrong thing after an error.
8690 [Arne Ansper <arne@ats.cyber.ee>]
8692 *) rsa_eay.c would attempt to free a NULL context.
8693 [Arne Ansper <arne@ats.cyber.ee>]
8695 *) BIO_s_socket() had a broken should_retry() on Windoze.
8696 [Arne Ansper <arne@ats.cyber.ee>]
8698 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8699 [Arne Ansper <arne@ats.cyber.ee>]
8701 *) Make sure the already existing X509_STORE->depth variable is initialized
8702 in X509_STORE_new(), but document the fact that this variable is still
8703 unused in the certificate verification process.
8704 [Ralf S. Engelschall]
8706 *) Fix the various library and apps files to free up pkeys obtained from
8707 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8710 *) Fix reference counting in X509_PUBKEY_get(). This makes
8711 demos/maurice/example2.c work, amongst others, probably.
8712 [Steve Henson and Ben Laurie]
8714 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8715 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8716 are no longer created. This way we have a single and consistent command
8717 line interface `openssl <command>', similar to `cvs <command>'.
8718 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8720 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8721 BIT STRING wrapper always have zero unused bits.
8724 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8727 *) Make the top-level INSTALL documentation easier to understand.
8730 *) Makefiles updated to exit if an error occurs in a sub-directory
8731 make (including if user presses ^C) [Paul Sutton]
8733 *) Make Montgomery context stuff explicit in RSA data structure.
8736 *) Fix build order of pem and err to allow for generated pem.h.
8739 *) Fix renumbering bug in X509_NAME_delete_entry().
8742 *) Enhanced the err-ins.pl script so it makes the error library number
8743 global and can add a library name. This is needed for external ASN1 and
8744 other error libraries.
8747 *) Fixed sk_insert which never worked properly.
8750 *) Fix ASN1 macros so they can handle indefinite length construted
8751 EXPLICIT tags. Some non standard certificates use these: they can now
8755 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8756 into a single doc/ssleay.txt bundle. This way the information is still
8757 preserved but no longer messes up this directory. Now it's new room for
8758 the new set of documenation files.
8759 [Ralf S. Engelschall]
8761 *) SETs were incorrectly DER encoded. This was a major pain, because they
8762 shared code with SEQUENCEs, which aren't coded the same. This means that
8763 almost everything to do with SETs or SEQUENCEs has either changed name or
8764 number of arguments.
8765 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8767 *) Fix test data to work with the above.
8770 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8771 was already fixed by Eric for 0.9.1 it seems.
8772 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8774 *) Autodetect FreeBSD3.
8777 *) Fix various bugs in Configure. This affects the following platforms:
8781 unixware-2.0-pentium
8785 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8786 before they are needed.
8789 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8793 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8795 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8796 changed SSLeay to OpenSSL in version strings.
8797 [Ralf S. Engelschall]
8799 *) Some fixups to the top-level documents.
8802 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8803 because the symlink to include/ was missing.
8804 [Ralf S. Engelschall]
8806 *) Incorporated the popular no-RSA/DSA-only patches
8807 which allow to compile a RSA-free SSLeay.
8808 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8810 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8811 when "ssleay" is still not found.
8812 [Ralf S. Engelschall]
8814 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8815 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8817 *) Updated the README file.
8818 [Ralf S. Engelschall]
8820 *) Added various .cvsignore files in the CVS repository subdirs
8821 to make a "cvs update" really silent.
8822 [Ralf S. Engelschall]
8824 *) Recompiled the error-definition header files and added
8825 missing symbols to the Win32 linker tables.
8826 [Ralf S. Engelschall]
8828 *) Cleaned up the top-level documents;
8829 o new files: CHANGES and LICENSE
8830 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8831 o merged COPYRIGHT into LICENSE
8832 o removed obsolete TODO file
8833 o renamed MICROSOFT to INSTALL.W32
8834 [Ralf S. Engelschall]
8836 *) Removed dummy files from the 0.9.1b source tree:
8837 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8838 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8839 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8840 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8841 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8842 [Ralf S. Engelschall]
8844 *) Added various platform portability fixes.
8847 *) The Genesis of the OpenSSL rpject:
8848 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8849 Young and Tim J. Hudson created while they were working for C2Net until
8851 [The OpenSSL Project]
8854 Changes between 0.9.0b and 0.9.1b [not released]
8856 *) Updated a few CA certificates under certs/
8859 *) Changed some BIGNUM api stuff.
8862 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8863 DGUX x86, Linux Alpha, etc.
8866 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8867 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8871 *) Add -strparse option to asn1pars program which parses nested
8873 [Dr Stephen Henson <shenson@bigfoot.com>]
8875 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8878 *) DSA fix for "ca" program.
8881 *) Added "-genkey" option to "dsaparam" program.
8884 *) Added RIPE MD160 (rmd160) message digest.
8887 *) Added -a (all) option to "ssleay version" command.
8890 *) Added PLATFORM define which is the id given to Configure.
8893 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8896 *) Extended the ASN.1 parser routines.
8899 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8902 *) Added a BN_CTX to the BN library.
8905 *) Fixed the weak key values in DES library
8908 *) Changed API in EVP library for cipher aliases.
8911 *) Added support for RC2/64bit cipher.
8914 *) Converted the lhash library to the crypto/mem.c functions.
8917 *) Added more recognized ASN.1 object ids.
8920 *) Added more RSA padding checks for SSL/TLS.
8923 *) Added BIO proxy/filter functionality.
8926 *) Added extra_certs to SSL_CTX which can be used
8927 send extra CA certificates to the client in the CA cert chain sending
8928 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8931 *) Now Fortezza is denied in the authentication phase because
8932 this is key exchange mechanism is not supported by SSLeay at all.
8935 *) Additional PKCS1 checks.
8938 *) Support the string "TLSv1" for all TLS v1 ciphers.
8941 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8942 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8945 *) Fixed a few memory leaks.
8948 *) Fixed various code and comment typos.
8951 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8952 bytes sent in the client random.
8953 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob